Springer Series in Astrophysics and Cosmology

Eleonora Di Valentino
Dillon Brout Editors

The Hubble
Constant
Tension
Springer Series in Astrophysics and Cosmology

Series Editors
Cosimo Bambi, Department of Physics, Fudan University, Shanghai, China
Dipankar Bhattacharya, Inter-University Centre for Astronomy and Astrophysics,
Pune, India
Yifu Cai, Department of Astronomy, University of Science and Technology of
China, Hefei, China
Pengfei Chen, School of Astronomy and Space Science, Nanjing University,
Nanjing, China
Maurizio Falanga, (ISSI), International Space Science Institute, Bern, Bern,
Switzerland
Paolo Pani, Department of Physics, Sapienza University of Rome, Rome, Italy
Renxin Xu, Department of Astronomy, Perkings University, Beijing, China
Naoki Yoshida, University of Tokyo, Tokyo, Chiba, Japan
The series covers all areas of astrophysics and cosmology, including theory, observa-
tions, and instrumentation. It publishes monographs and edited volumes. All books
are authored or edited by leading experts in the field and are primarily intended for
researchers and graduate students.
Eleonora Di Valentino · Dillon Brout
Editors

The Hubble Constant Tension

Editors
Eleonora Di Valentino Dillon Brout
School of Mathematics and Statistics Boston University
University of Sheffield Boston, USA
Sheffield, UK

ISSN 2731-734X ISSN 2731-7358 (electronic)

Springer Series in Astrophysics and Cosmology
ISBN 978-981-99-0176-0 ISBN 978-981-99-0177-7 (eBook)
https://doi.org/10.1007/978-981-99-0177-7

© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature
Singapore Pte Ltd. 2024

This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse
of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and
transmission or information storage and retrieval, electronic adaptation, computer software, or by similar
or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or
the editors give a warranty, expressed or implied, with respect to the material contained herein or for any
errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd.
The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721,
Singapore

Paper in this product is recyclable.

Contents

Part I Intro Measurements

1 The Hubble Tension, Book Prologue: A Perspective Along
Several Axes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Adam G. Riess
2 The Hubble Constant: A Historical Review . . . . . . . . . . . . . . . . . . . . . . 7
R. Brent Tully

Part II Measurements (and Systematics)

3 Trigonometric Parallax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
Lennart Lindegren
4 Measuring H0 with H2 O Megamasers . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
James A. Braatz and Dominic W. Pesce
5 Eclipsing Binary Stars as Precise and Accurate Distance
Indicators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
Grzegorz Pietrzyński and Dariusz Graczyk
6 On Cepheid Distances in the H0 Measurement . . . . . . . . . . . . . . . . . . . 89
Richard I. Anderson
7 The Role of Type Ia Supernovae in Constraining the Hubble
Constant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
Dan Scolnic and Maria Vincenzi
8 Tip of the Red Giant Branch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
Siyang Li and Rachael L. Beaton
9 Surface Brightness Fluctuations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
Michele Cantiello and John P. Blakeslee
10 The Pursuit of the Hubble Constant Using Type II Supernovae . . . . 177
Thomas de Jaeger and Lluís Galbany

v
vi Contents

11 The Mira Distance Ladder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191

Caroline D. Huang
12 Tully–Fisher Relation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
Khaled Said
13 Determination of the Local Hubble Constant Using Giant
Extragalactic H II Regions and H II Galaxies . . . . . . . . . . . . . . . . . . . . 235
David Fernández-Arenas and Ricardo Chávez
14 Strong Lensing and H0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
Tommaso Treu and Anowar J. Shajib
15 Addressing the Hubble Tension with Cosmic Chronometers . . . . . . . 277
Michele Moresco
16 The Cosmic Microwave Background and H0 . . . . . . . . . . . . . . . . . . . . . 295
Pablo Lemos and Paul Shah
17 Measuring H0 with Spectroscopic Surveys . . . . . . . . . . . . . . . . . . . . . . . 319
Mikhail M. Ivanov and Oliver H. E. Philcox

Part III Complications with Measurements

18 Impact of Peculiar Velocities on Measurements of H0 . . . . . . . . . . . . . 341
W. D’Arcy Kenworthy and Tamara M. Davis
19 The Impact of Dust on Cepheid and Type Ia Supernova
Distances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363
Dillon Brout and Adam Riess

Part IV Theoretical Proposed Solutions

20 (Introduction to the Second Part of the Book) What About
the Solutions? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387
Eleonora Di Valentino
21 Not Empty Enough: A Local Void Cannot Solve the H0 Tension . . . 391
Dragan Huterer and Hao-Yi Wu
22 Resolving the Hubble Tension with Early Dark Energy . . . . . . . . . . . 403
Vivian Poulin and Tristan L. Smith
23 New Early Dark Energy as a Solution to the H0 and S8 Tensions . . . 431
Florian Niedermann and Martin S. Sloth
24 On the Dark Radiation Role in the Hubble Constant Tension . . . . . . 457
Stefano Gariazzo and Olga Mena
25 Decaying Dark Matter and the Hubble Tension . . . . . . . . . . . . . . . . . . 481
Andreas Nygaard, Emil Brinch Holm, Thomas Tram,
and Steen Hannestad
Contents vii

26 Signs of a Non-zero Equation-of-State for Dark Matter . . . . . . . . . . . 493

Krishna Naidoo
27 Resolving the Hubble Tension at Late Times with Dark Energy . . . . 503
Marco Raveri
28 Cosmological Models with Negative Λ . . . . . . . . . . . . . . . . . . . . . . . . . . . 517
Anjan A. Sen
29 On the Interacting Dark Energy Scenarios—The Case
for Hubble Constant Tension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 531
Supriya Pan and Weiqiang Yang
30 Modified Gravity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 553
Emmanuel N. Saridakis
31 Early-Time Modified Gravity and the Hubble Tension . . . . . . . . . . . . 575
Matteo Braglia
32 Primordial Magnetic Fields and the Hubble Tension . . . . . . . . . . . . . . 587
Karsten Jedamzik and Levon Pogosian
33 Varying Fundamental Constants Meet Hubble . . . . . . . . . . . . . . . . . . . 613
Jens Chluba and Luke Hart
34 Anomalies and Tensions in Cosmology and a Primordial
Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 639
Dhiraj Kumar Hazra and Arman Shafieloo

Part V Complications with Theoretical Solutions

35 The Tension in the Absolute Magnitude of Type Ia Supernovae . . . . 661
David Camarena and Valerio Marra
36 CMB Anomalies and the Hubble Tension . . . . . . . . . . . . . . . . . . . . . . . . 675
William Giarè
 Part I
Intro Measurements
Chapter 1
The Hubble Tension, Book Prologue:
A Perspective Along Several Axes

Adam G. Riess

It is, I think, particularly in periods of acknowledged crisis that

scientists have turned to philosophical analysis as a device for
unlocking the riddles of their field. Scientists have not generally
needed or wanted to be philosophers.
—Thomas Kuhn

In the Kuhnian stage of scientific progress called “normal science”, a model organizes
our present understanding and provides the means to predict the outcomes of new
experiments. New results which fail to conform to the model are initially considered
errors of the experiments. As anomalous results pile up or become too big to overlook,
a crisis is reached and the model must be replaced by one which can explain the old
and new data, and the cycle repeats. Cosmological models have been susceptible to
such revisions largely because they have shallow roots in fundamental physics. The
present cosmological model, LambdaCDM model looks ripe for audit in this context.
It has been good at describing what we see but it lacks fundamental comprehension
of the physics needed to explain most of the Universe (i.e., Dark Energy and Dark
Matter) and lacks the ability to answer some of our biggest questions (“How does
gravity operate on quantum scales?” “Did inflation occur and why?” “What is the
nature of neutrinos?”, the “Why now?” problem, etc.) The chapters in this book
describe why the “Hubble Tension”, now a decade old, provides a strong challenge
to LambdaCDM. The Universe appears to be expanding faster than the prediction
from LambdaCDM, calibrated by the CMB, with more than 5 sigma (>99.99998%)
confidence, the standard threshold of discovery (see attached figure). Yet there is
no consensus as to the cause of the Tension, nor on the nature of a model revision,
leaving us in a Kuhnian purgatory.
The name “Hubble Tension” was first used in the early 2010’s to describe the
significance of the discrepancy between the measured and expected value of the
Hubble constant (Physicists tend to call 2–3 sigma differences or 95% to 99.7%

A. G. Riess (B)
Johns Hopkins University and Space Telescope Science Institute, Baltimore, MD 21218, USA
e-mail: ariess@stsci.edu

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 3
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_1
4 A. G. Riess

confidence, a “Tension”). While the increased significance in the last decade should
have elevated this tension to a “Hubble Crisis” in the physicist’s vernacular, “Hubble
Tension” is an even better descriptor as the level of angst this discrepancy elicits
within the field of cosmology. That is, the interpretation of the Hubble Tension
appears to depend, to a surprising degree, on the (possibly subconscious) priors of
the beholder, or perhaps what they most value.
Let’s explore several axes along which some different perspectives on the Tension
are apparent. One such axis we might call “Modelers versus Empiricists”. For most
Empiricists, measuring the Hubble constant provides an important test of Lamb-
daCDM which they view as requiring further scrutiny due to its unsatisfactory,
phenomenological underpinnings. If the quality of the measurements is high (i.e.,
well designed and gathered, the data is publicly available, requires no significant
model dependence for interpretation, presents no significant oddities and has survived
years of scrutiny) the Tension is generally accepted as a reality with hopes for an
eventual explanation. Empiricists by definition can accept the veracity of an experi-
ment without an explanation based on the caliber of the measurements. For Modelers,
the inability to explain the Tension within the context of a cosmological model is held
as a strong prior against acceptance of the experimental results. This group is rarely
focused on the specific details of the measurements because what really bothers them
is the lack of an explanation. Therefore, the span of different, experimental results
more often nurtures doubt rather than confirmation of the Tension and the standard
of accepting the Tension soberly proffered as “extraordinary claims require extraor-
dinary evidence”. “Extraordinary” cannot be defined a priori but is weighed against
the challenge of revising LambdaCDM. Modelers may refer to “unknown errors”, an
evergreen, escape hatch, though from an Empiricist’s point of view the speculation
itself lacks enough rigor to be falsified. Empiricists and Modelers are inclined to see
the Hubble Tension as the other’s problem which produces another kind of tension.
Empiricists themselves may be divided along “Averagers” and “Analyzers”. The
former prefer to average everything in the belief that this will reduce uncertainties,
equating systematic or methodological uncertainties of different observables with
random errors. While there is some wisdom in crowd-sourcing, a shortcoming of
this approach for the Hubble Tension is that it can delay progress by decades. A
newly emerging Tension supported by new data, space missions (e.g., Gaia, Planck,
HST, JWST) or understanding will always be washed out by averaging. I would be
skeptical of analyses that use older data when an objectively better source is available.
(A close cousin of the Averagers are “Nostalgics” who rely on historic data, perhaps
data they themselves collected, rather than state-of-the-art data.) Analyzers try to
understand the differences in the hopes of progress, but must be wary of “over
analyzing” which materializes as the particle Physicists’s “Look Elsewhere Effect”.
That is, if you correlate everything against anything, you are bound to find something,
and thorough accounting is required to avoid insignificant findings, or, the bane of
astronomy, selection bias.
Another axis of Tension is the “Bettors vs the Seekers”. Bettors use the history
and their memories of past Tensions to decide what to conclude at present. Many
in this group see the present Hubble constant discrepancy of 67 vs 73 km/sec/Mpc
1 The Hubble Tension, Book Prologue: A Perspective Along Several Axes 5

as a repeat of the 50 vs 100 km/sec/Mpc difference of the 1970’s-1990’s, betting

that the extremes of measurements with small errors are wrong. Some may bet on
an answer “in between” following the example of the past. Bettors are inclined to
offer a literal wager on the final answer. The Bettors case can and often is taken
to an extreme, i.e., that all unexpected experimental results are wrong which is a
good, statistical bet for any individual example. They may advise to just double all
the uncertainties or assume “unknown unknowns” at play to cover their bets. It is
a somewhat jaded perspective; a confirmation bias against discovery. Seekers, in
contrast, are more impressed by a difference in the present Tension than the past
because the conflicting inferences of the Hubble constant are not mutually exclusive,
their comparison depends on an incomplete cosmological model. The seekers are
looking for new physics in LambdaCDM that can explain the discrepancy. Its hard
but noble work.
Finally, there is the axis dividing those whose measurements seem most at odds,
Early vs Late Universe experimentalists. It is common for those in one of these groups
to doubt the measurements in the other as they are less familiar with them and judge
the other’s tools to be inferior or complex. To a large degree ignorance of the methods
is the source of skepticism rather than any specific concern. Members of these groups
say they would like to see another, independent confirmation of the other side, but
are often slow to recognize those that already exist. I have experienced this axis first
hand as I have attended conferences of each with polls undertaken pointing to the
high expectation of an error from “the other side” but none as likely in their own.
A number of these axes overlap or may be seen simply as Optimists (i.e., Seekers,
Analyzers, Empiricists) versus Pessimists (i.e., Bettors, Averagers, Modelers) in the
way they view the Tension.
Finally, it is interesting to consider these perspectives in the context of historical
examples to “back-test” their value. There are famous results which were correct
but the explanation required decades such as the precession of Mercury, the Solar
neutrino problem, galaxy velocities in Coma, and even currently accepted results like
flat rotation curves in galaxies and the accelerating Universe have only an incomplete
explanation. We may also consider the circumstances of experimental results which
did not hold up such as Weber’s discovery of gravitational waves in the 1960’s, Arp’s
quantized quasar redshifts, or Van Manaan’s rotating galaxies of the 1920’s, or more
recent claims of faster-than-light neutrinos and finding B modes from inflation in
the CMB. Past examples show history is not a substitute for doing the science work.
For every example there is a counter example showing history is not predictive, or
as the Danish proverb says, “It is difficult to make predictions, especially about the
future.” Still, having considered a wide range of examples the biggest discriminator
of outcome is probably the time interval over which the discrepancy lasts (provided
the data has been available over that time). The scientific community has shown
itself to be very adept at finding errors rather quickly, and if the data is public, errors
(of more than a singular author) rarely last a year let alone a decade. In contrast,
anomalies which survived scrutiny over many years are the ones where something
interesting was usually going on. Judged by these standards, the Hubble Tension is
one worth paying attention to.
6 A. G. Riess

Fig. 1.1 Recent measurements of H 0 from two approaches. Direct (bottom) from distance
versus redshift and Indirect (top) from an early Universe calibration (via the CMB or BBN) of
LambdaCDM. Adapted from [1]

So with which perspective should you read the following chapters on the Hubble
Tension? I suggest curiosity (Fig. 1.1).

Reference

1. Di Valentino, E. Mon. Not. Roy. Astron. Soc. 502(2), 2065–2073 (2021)

Chapter 2
The Hubble Constant: A Historical
Review

R. Brent Tully

2.1 In the Beginning

In the first decades of the Twentieth century, evidence was accumulating that spiral
nebulae lay beyond the Milky Way as “island universes”, with discussion culminating
in the Great Debate [1, 2]. History books usually cite the convincing evidence for the
extragalactic nature of spiral nebulae to the identification of cepheid variable stars
in M31 and M33 by Hubble [3] in 1925. The relationship between pulsation period
and luminosity had been established by Leavitt [4] in 1912. Hubble determined the
distances to each of M31 and M33 to be 285 kpc. However, three years earlier, a
better distance to M31 of 450 kpc had been calculated by Öpik [5] assuming the virial
theorem relation between velocity dispersion, dimension, and mass and assuming the
solar relation between luminosity and mass. The modern value for the distance to
M31 is 740 kpc and to M33 is 850 kpc.
In the meantime, spectra of the spiral nebulae were being accumulated, mostly by
Slipher [6], and it was being remarked that almost all spectral features were displaced
to longer wavelengths from their laboratory positions. In a seminal paper [7], Hubble
demonstrated the correlation between velocity displacement, subsequently called
redshift, and galaxy distance, with the relationship between these parameters that
now bears his name. In fact, the expansion of the Universe had been anticipated
and hinted at by Lemaître [8] on dynamical grounds. The static Universe posited
by Einstein [9] with his constant .Ʌ to offset gravity was unsatisfactorily unstable.
The initial estimate of Hubble’s constant [7] was . H0 = 500 km s−1 Mpc−1 , later
increased slightly by Hubble to 530 km s−1 Mpc−1 [10]. The uncertainty was large,
as pointed out by Jan Oort who favored a lower value [11] and Arthur Eddington who
favored a higher value [12]. Eddington provided a theoretical perspective, a prelude
to an interplay that continues to be a bane of studies of the scale of the Universe.

R. B. Tully (B)
University of Hawaii, 2680 Woodlawn Dr., Honolulu, HI 96822, USA
e-mail: tully@ifa.hawaii.edu

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 7
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_2
8 R. B. Tully

Eddington claimed that the combination of relativity theory and quantum theory gave
a value of . H0 with the wonderful quote: “I have now been able to reach a solution
which I regard as definitive. The result is 865 km. per sec. per megaparsec.”

2.2 Evident Problems

If the Universe has been expanding at a uniform rate since its inception then it began
at a time ago that is the inverse of the Hubble constant. With Hubbles’s initial value
of 500 km s−1 Mpc−1 the implied age is 2 billion years. The discrete moment of
the beginning of expansion has come to be called the “Big Bang.” Already by 1931,
radioactive dating of geological rocks was giving ages of the Earth of 2–3 Gyr [13].
Surely, the Universe is older than the Earth. A plausible resolution was offered by
the steady state theory of Bondi and Gold [14] and Hoyle [15], whereby matter is
continuously created as space expands so as to keep constant conditions on average
in perpetuity.
However, evidence began to emerge that the early estimates of . H0 were too high.
Walter Baade made a breakthrough with the discovery of distinct luminosity differ-
ences between the brightest of young (Population I) stars and old (Population II)
stars when he could resolve red giant branch stars in M31, M32, and NGC 205 from
Mt. Wilson during Los Angeles blackouts during WWII [16]. The stepping stone to
the calibration of the cepheid period-luminosity relation had been through Galac-
tic globular clusters which have cepheid-like variables. It was recognized that the
brightest resolved stars in the central region of M31 and in the elliptical companions
were similar to the brightest stars in globular clusters. It became apparent that there
are two kinds of cepheid variables: Population II cepheids as found in environments
of old stars and Population I cepheids found amongst young stars that are brighter
at a given pulsation period by .∼1.5 mag [17], implying a factor 2 increase in cos-
mological distances. Baade provided a very nice history of this development in 1956
[18].

2.3 Grasping at Straws

Work on the extragalactic distance scale picked up in the 1950s with the inauguration
of the Hale 200-inch telescope on Mt. Palomar. There was the important contribution
by Humason et al. [19]. Imaging of the Virgo cluster galaxy NGC 4321 revealed that
features taken earlier to be individual stars were, in fact, HI regions. The brightest
individual stars are some 1.8 mag fainter. Humason, Mayall, and Sandage concluded
−1
. H0 ∼ 180 km s Mpc−1 . Soon after, Sandage [20] reviewed the situation. In addition
to the confusion between old and young populations with cepheids, he argued that
the apparent width of the cepheid period-luminosity relation at fixed period caused a
2 The Hubble Constant: A Historical Review 9

significant under-estimation of distances. He placed . H0 between 50 and 100 km s−1

Mpc−1 .
Over the next two decades, there was increased interest in the distance scale but
there were only marginal improvements. The problem to be addressed was clear but
the tools were blunt. Eric Holmberg tried using a correlation between the absolute
magnitude and surface brightness of galaxies [21]. Sydney van den Bergh produced a
luminosity classification for mid- to late-type spirals that could be given an absolute
calibration [22, 23]. Jose Sersic promoted the use of the diameter of HII regions
[24]. René Racine compared the globular cluster luminosity function between M31
and a galaxy in the Virgo cluster [25]. In 1970, van den Bergh summarized results
from nine methods: the diameter of HII regions, luminosity classes, brightest globu-
lar clusters, assumptions regarding mass-to-light ratios, magnitude of third brightest
galaxy cluster member, supernovae, a relation between the surface brightness and
diameter of galaxies, and the magnitude of the brightest star in a galaxy [26]. None
of these methods survives as useful today. The use of supernovae in 1970 was rudi-
mentary. The mass-to-light method was a variation of what was used by Öpik [5].
It was assumed that the total mass of a galaxy, . Mt , is given by the virial theorem,
with the dependency . Mt ∝ ∆V 2 R, where .∆V is a measure of the galaxy kinematics
and . R is an estimate of the galaxy radius. A relationship is assumed between mass
and luminosity. The kinematic data for this so-called indicative mass procedure were
provided by Morton Roberts’ observations of the linewidths of galaxies in the neutral
hydrogen 21 cm line [27] or subsequent follow up.

2.4 . H0 50 or 100?

The discovery in 1965 of the permeation of space with the cosmic microwave back-
ground provided compelling evidence of a hot Big Bang beginning of the Universe
[28, 29], and by implication a finite age. Necessarily, there should be compatibility
between the expansion age inferred from the inverse of the Hubble constant, the
ages of objects within the Universe, and a model that describes the evolution of the
Universe since its inception. Two simple models to consider are a completely empty
universe or a topologically flat universe filled with matter. With . H0 = 50 km s−1
Mpc−1 ages would be 19.3 and 13.0 Gyr, respectively, with these models. With
−1
. H0 = 100 km s Mpc−1 the respective ages would be 9.7 and 6.5 Gyr.
Ages of stars within our galaxy could be estimated through radioactive dating,
white dwarf cooling or, particularly, the main sequence turnoff magnitude within
globular clusters. Into and through the 1980s, the best estimate of the oldest ages
found in the Galaxy were .∼16 Gyr [30, 31], with the Universe presumably 1 Gyr
or so older. Evidently, the Universe is not empty. Occam’s razor favored models that
did not invoke a cosmological constant. An open universe with the density of matter
25% of the critical density to achieve closure (.Ωm = 0.25) and . H0 = 50 km s−1
Mpc−1 would be consistent with an age of 16 Gyr. The age of a closed universe
(.Ωm = 1) and . H0 = 50, at 13 Gyr, would be marginally consistent with stellar ages.
10 R. B. Tully

Such simple models were inconsistent with those stellar ages if the Hubble constant
was much above 50 km s−1 Mpc−1 .
Sandage and Gustav Tammann, in a series of papers between 1974 and 1982, pre-
sented arguments that favored low values of . H0 [32–36]. The methods they used to
acquire distances included sizes of HII regions, brightest stars, luminosity classifica-
tions, and bootstraps through statistical properties of the Virgo cluster to applications
in the field. Estimates of . H0 in these papers ranged from 50 to 57 with typical errors
.±7 or less. Gérard de Vaucouleurs, in a series of rebuttals between 1978 and 1986,
favored high values of . H0 [37–41]. He and colleagues built a ladder founded on
primary novae, cepheids, RR Lyrae, and horizontal branch stars, secondary open
and globular clusters, brightest blue and red supergiant stars, brightest HII loops or
rings, and the velocity dispersions of HII regions, then a tertiary luminosity index
calibration. They found values of . H0 between 90 and 110 km s−1 Mpc−1 . So was
born the . H0 50 or 100 controversy.

2.4.1 Better Methods

Already by the 1970s, there was a reasonable foundation for the distance scale through
parallaxes and the distances they gave to cepheids and RR Lyrae stars on the hori-
zontal branch, albeit subject to improvements. But there had been only crude ways
to estimate distances beyond the Local Group and into the domain unaffected by ran-
dom motions. It was in this time period that methodologies began to emerge that are
still retained today for the measurement of galaxy distances at significant redshifts.
In 1977, this author and Richard Fisher published two papers finding . H0 = 75 −
80 km s−1 Mpc−1 based on the correlation between the absolute luminosities of spiral
galaxies and their rotation rates [42, 43], subsequently to be referred to as the TF
relation. This procedure recalls the indicative mass method but deviates from the
parameterization of the virial theorem. It invokes only two observables, a distance
dependent luminosity and a distance independent measure of the rotation rate (taking
account of the galaxy inclination and internal obscuration). A power law was assumed
between luminosity and rotation rate, with a free index subject to empirical definition.
There is the obvious but simple explanation for the correlation that more massive
spiral galaxies tend to be more luminous and spin faster. More massive galaxies tend
to contain fractional more old stars, hence are redder. Therefore, it was not a surprise
that the free index of the TF relation increases toward longer wavelengths [44], The
methodology was more solidly grounded in 1979 by shifting from photographic blue
magnitudes to photoelectric near-infrared magnitudes, as demonstrated by Aaronson
et al. [45]. These authors demonstrated that the free index would approximate four,
. L ∼ ∆V , if spiral galaxies share dimensionless scale-length mass profiles/rotation
4

curves, have disks with a common central surface brightness, and have the same mean
mass-to-light ratio. This latter condition is more easily met in the infrared. The true
situation is more complicated. Indeed, it is mainstream to assume that dark matter
dominates the mass, so it could be considered a puzzle that the correlation between
2 The Hubble Constant: A Historical Review 11

star light and kinematics is as tight as observed. The interest of most theorists has
been on the implications for galaxy formation and evolution [46–51]. Treating the
relation as an empirical phenomenon, Aaronson and colleagues determined values of
. H0 between 90 and 95 over the years 1980–1986 [52–54]. An early detailed mapping

of the velocity field of the Local Supercluster was carried out by this collaboration
[55].
A contemporaneous discovery with the TF relation, was an analogous correla-
tion between the luminosities and central velocity dispersions of early-type galaxies,
ellipticals and lenticulars, by Faber and Jackson [56]. The two parameter formulation
for ellipticals had higher dispersion than the TF relation for spirals but in 1987 it was
found that the addition of distance-independent surface brightness as a third param-
eter provided a significant improvement [57, 58]. The procedure, with parameters
gathered into distance dependent and independent parts, resulted in what became
called the fundamental plane. Initially there was no attempt at an absolute calibra-
tion but the method was used to map relative distances with redshifts. A group soon
known as the Seven Samurai (David Burstein, Roger Davies, Alan Dressler, Sandra
Faber, Donald Lynden-Bell, Roberto Terlevich, and Gary Wegner) demonstrated the
existence of a large scale flow toward a mysterious “Great Attractor” [59–62]. Even as
late as 2001 when substantial samples had been accumulated, the fundamental plane
was still only being used to derived peculiar velocities and velocity flows based on
relative distance measurements [63–65]. However, it has found an important place
today within absolute measurements of the extragalactic distance scale.
A next contribution to the pantheon of important methodologies for the measure-
ment of galaxy distances came in 1988 with recognition that distance information
is encoded in the surface brightness fluctuation amplitudes of dominantly old stel-
lar populations [66]. The brightest stars in regions devoid of recent star formation
lie on the upper red giant branch of stellar evolution. The relative number of such
stars within an observer’s pixel increases as the square of distance, causing a pre-
dictable damping with distance. Between 1989 and 2000, John Tonry and colleagues
exploited this phenomenon to derive estimates of the Hubble constant that fell from
an initial . H0 = 88 to a later . H0 = 77 km s−1 Mpc−1 [67–70]. The study culminated
in the publication of 300 distances to galaxies within 4000 km s−1 and a study of the
flow patterns discerned from this sample [71].
The foundation of the surface brightness fluctuation technique is the constancy
of luminosities of stars at the “tip” of the red giant branch, where degenerate helium
cores reach a critical mass for the onset of burning to carbon and subsequently a
star finds itself on the horizontal branch. The phenomenon harkens back to Baade’s
identification of Population II stars in M31 and companions [16]. Jeremy Mould and
others were using observations of such stars in the 1980s to estimate distances to
galaxies within the Local Group [72]. As charge coupled device (CCD) detectors
became common-place, the so-called TRGB method became well established [73,
74]. Observations from ground-based telescopes began to push slightly beyond the
Local Group [75]. With space facilities the TRGB method would evolve to be an
important pedestal for extragalactic distances.
12 R. B. Tully

Other methods of possible merit explored in those years have not found a place as
important players today. There was evidence that the planetary nebulae luminosity
function (PNLF) had small scatter [76, 77] and was giving values of . H0 = 86 km s−1
Mpc−1 . However, the PNLF has shown to have limited usefulness. It required an
empirical calibration to a statistically significant number of galaxies, cross-referenced
with fundamental methods, but once calibrated had limited range of application. A
similar criticism can be made of the use of the Balmer spectral features in very
luminous B and A spectral class supergiants [78, 79]. Potentially, this Flux weighted
Gravity-Luminosity Relationship (FGLR) could be very accurate but its range is
limited and requires considerable effort per target. There was once hope for the
utility of the globular cluster luminosity function which gave values of . H0 in the
range 75–80 [80–82]. However it was difficult to have confidence in results when
galaxy globular cluster populations are so variable and the formation processes are
in dispute. Novae were found to give . H0 = 70 km s−1 Mpc−1 with limited usage
[83, 84].

2.4.2 Model-Based Methods

All of the methods discussed so far have been linked to the distance ladder that starts
with stellar parallaxes and builds through properties of stars in various parts of the
Hertzprung-Russell diagram like the main sequence turnoff, red giant and horizontal
branches, and variable RR Lyrae and cepheid stars. Further steps on the ladder can
proceed by association of such features in external galaxies, providing absolute scales
to the sorts of methods that have been discussed. With exceptions yet to be discussed,
these methods access limited redshifts. There are other ways of estimating galaxy
distances, though, that avoid the calibration ladder and can provide information on
the expansion scale with a long reach.
One of these involves gravitational lens time delays. Sjur Refsdal explored the
concept in 1964, well before there was observational evidence of the phenomenon
[85]. A gravitational lens caused by a mass concentration could cause light from
a background source to follow multiple paths to reach an observer, with the time
followed on each path slightly different. If the emission from the background source
is fluctuating, say because it is a supernova or active galactic nucleus, then the arrival
to the observer of that signal will be slightly different along each of the paths. Roughly,
the time delay between two paths, .∆t, is described by the formula [86]
[ ][ ]
1 + zl dl ds
.∆t ∼ 0.5 (Θ2A − Θ2B ) (2.1)
c dls

where .d are angular diameter distances, .dl from observer to the lens (at redshift .zl ),
d from observer to the source, and .dls from lens to source, and .Θ A , .Θ B are angular
. s
separations from the lens of images A and B. Each of the distances depends inversely
on . H0 , permitting a reformulation of the time delay to give a value of the expansion
2 The Hubble Constant: A Historical Review 13

parameter within an assumed world model. Early estimates of . H0 were in the range
50 to 75 and even greater than 100, with recognized large uncertainties related to the
distribution of mass around the lens [87, 88].
Another model dependent methodology developed out of a phenomenon first
discussed by Rashid Sunyaev and Yakov Zel’dovich [89]. The possible exploitation
to obtain distances was discussed by Silk and White [90] and applied by Mark
Birkinshaw and colleagues and others [91–94]. The so-called Sunyaev-Zel’dovich
(SZ) effect pertains to a dimming of the cosmic microwave background radiation
in the direction of rich clusters of galaxies due to inverse-Compton scattering of
the radiation by electrons in the hot cluster halo. Thermal bremsstrahlung flux can
also be observed from that same hot gas at X-ray wavelengths. The SZ temperature
decrement depends on the electron temperature and density along the line-of-sight
in the cluster. The X-ray flux depends on the square root of the electron temperature,
the product of the electron and proton densities (effectively the square of electron
density), the volume of the hot gas cloud, and because it is an apparent flux, inversely
on the square of distance. If all of the microwave decrement, the X-ray flux, and the
gas temperature can be measured then an estimate of distance can be derived from the
observed size and shape of the cluster and a model of the gas distribution. It is to be
assumed that the gas is in hydrostatic equilibrium, which can clearly not be the case
if a cluster is experiencing a merger event but might be a reasonable approximation
in some cases. There is the complication that the X-ray flux, depending on electron
density squared, will particularly arise in high density regions of the gas, while the
SZ effect arises from lower electron density, longer path length regions. Complex
modelling is required. The early analyses gave values of the Hubble constant in the
range 45–70 km s−1 Mpc−1 .
Today supernovae play an important role in measurements of the scale of the
Universe but in early years their contributions were ambiguous. Walter Baade made
clear the distinction between novae that can be recurring events in stars and much
brighter supernovae, and with Fritz Zwicky, attempted to estimate their absolute
magnitudes [95, 96]. The spectra of supernovae vary considerably over time within
an event and between events. There are those without, or almost without, the spectral
features of hydrogen (called Type I) and those exhibiting strong hydrogen features
(called Type II). Various sub-classes can be identified by spectral or light curve
characteristics. There are two fundamentally different paths to supernova cataclysms.
There are core collapse events in massive stars that have exhausted their nuclear
fuels, usually identified as Type II although also as Type Ib or Type Ic. Then there
are Type Ia merger or accretion events that take white dwarfs over the Chandrasekhar
limit of .1.44 Mʘ for electron degeneracy pressure support (or at least close enough
to that limit to cause ignition). The Type Ia supernovae are brighter, hence seen to
greater distances and predominant in surveys because of access to greater volumes.
Charles Kowal made an early attempt to define the absolute luminosities of the
various supernova classes [97].
Attempts were made to specify absolute magnitudes using models. In principal,
a link could be made between the distance independent expansion velocity of the
supernova photosphere and the distant dependent angular expansion as determined
14 R. B. Tully

by measurements of temperature and apparent magnitude [98–103]. The book from a

meeting in 1984 provided a good summary of the status of Hubble constant estimates
at that date from supernova considerations [104]. With 13 separate model-driven
contributions published up to 1995, the average value of . H0 was 58 km s−1 Mpc−1 .
A leader in the field who can remain anonymous once said “Either . H0 equals 50 or
we don’t understand core-collapse supernovae”. Right, one of those.

2.5 Satellites, Supernovae, and the CMB

Changes were coming in the 1990s. At the foundational level, the Hipparcos satel-
lite that operated from 1989 to 1993 greatly increased the number and improved
the quality of stellar parallaxes. Distances to globular clusters were increased and
consequently their ages decreased. By the end of the decade, it was being argued that
the oldest stars in the Galaxy had ages 12–13 Gyr [105–107].
In 1990, Hubble Space Telescope (HST) was launched. After a false start, a solu-
tion was implemented with a service mission and the addition of Wide Field/Planetary
Camera 2 (WFPC2) in 1993. The full HST bounty only began to be harvested, though,
after the third service mission in 2002 with the activation of the much more powerful
Advanced Camera for Surveys (ACS). Only hints of the considerable role that HST
was to play were revealed in the last decade of the 20th century. Even before the
correction of the optics, two teams began using HST to observe cepheid variables in
nearby galaxies, switching to the use of WFPC2 as it became available. One team
involved Abhijit Saha, Alan Sandage and collaborators [108–113]. Their results were
summarized by Tamman, Sandage, and Saha in 2000 who concluded that the dis-
tances to SNIa calibrated by cepheids gave . H0 = 58 ± 6 km s−1 Mpc−1 [114]. The
other team lead by Wendy Freedman was formed to carry out what became known
as the HST Key Project [115–119]. The Key Project study, cumulatively grounded in
cepheid, tip of the red giant branch, globular cluster, planetary nebulae, and surface
brightness fluctuation observations [120], arrived at a value of the Hubble constant
of .72 ± 8 km s−1 Mpc−1 [121].1
The effort to calibrate the SNIa scale was two-pronged. Paralleling the work
acquiring cepheid distances to hosts of SNIa were vast improvements in SNIa acqui-
sitions. The utility of the SNIa tool was much improved by the adjustments that
could be made to minimize scatter through the recognition by Mark Phillips that the
peak brightness depends on the decay rate of luminosity after peak brightness [122].
Diverse teams other than the those of the Key Project or Saha-Sandage that looked
at what could be done with SNIa in the 1990s found values of . H0 between 63 and
70 [123–127].
Throughout this period, other methodologies continued to be used, generally con-
tributing more noise than signal. Sandage and co-workers published on studies using

1 Up to this point in the discussion, values of . H0 have been given without errors. For the most part,
errors given in the last century (and often the current one) have been wildly optimistic.
2 The Hubble Constant: A Historical Review 15

a spiral luminosity function, spiral lookalikes, the globular cluster luminosity func-
tion, and their own analysis of the TF relation, finding in each case . H0 = 57 within
−1
.±5 km s Mpc−1 [128–131].
The most exciting developments with supernovae at this time did not directly
involve the Hubble constant. Observations of high redshift SNIa by teams lead by
Schmidt and Perlmutter [132, 133] provided evidence contrary to the cold dark matter
(CDM) model with sufficient density in matter for a closed universe (.Ωm = 1). The
most distant supernovae were too faint. Assuming that they had intrinsic luminosities
similar to nearby SNIa, the observations could best be fit with a model with cosmic
acceleration, such as anticipated with the inclusion of the cosmological constant, .Ʌ,
in Einstein’s equation.
In quite another domain, a satellite to study the cosmic microwave background
was to have a profound impact on our understanding of the scale of the Universe.
The Cosmic Background Explorer (COBE) that operated between 1989 and 1993
demonstrated unambiguously the black body spectrum of the cosmic microwave
background, strong proof for its origin in a hot Big Bang beginning of the Universe
[134, 135]. The strong hints of a peak in the power spectrum of fluctuations was
confirmed by the balloon experiments MAXIMA and BOOMERANG [136–138].
The angular scale of the first peak provided compelling evidence that the Universe
has a flat topology. The results from distant SNIa indicated that cold dark matter alone
could not produce a spatially flat universe and gave a preference for the existence of
vacuum energy. The microwave background and SNIa observations taken together
were consistent with a.ɅCDM model of a universe with both dark matter with density
.Ωm and dark energy with density .ΩɅ such that .Ωm + ΩɅ = 1.

There was yet another ingredient in the mix. Wide angle redshift surveys were
accumulating and more power was being seen on large scales than anticipated by
the standard CDM model with .Ωm = 1 [139–142]. The disagreements became more
accute with the additional constraints imposed by the microwave background power
spectrum results from COBE [143–145]. Models with .Ωm ∼ 0.3 were favored,
whether with or without dark energy. Perversely, models with .Ωm = 1 were still
being offered that could be reconciled with the large scale structure and microwave
background information but only if . H0 was taken to be .∼ 35 km s−1 Mpc−1 , in utter
contradiction with distance ladder observations [146–148].
Increasingly, wide field imaging and redshift surveys were becoming available as
the 20th century came to a close. Michael Hudson was reconstructing density and
velocity fields from these inputs [149, 150]. It was demonstrated that the TF relation
could give distances with comparable accuracy observing at optical . R, . I bands as
obtained in the infrared, given area photometry using CCDs [151]. Jeff Willick and
colleagues made major contributions to issues of calibration and biasing [152–157].
They came to favor a value of the Hubble constant of 85 km s−1 Mpc−1 and, from
the peculiar velocity field, matter density .Ωm ∼ 0.3 [158, 159]. Riccardo Giovanelli,
Martha Haynes, and colleagues greatly enlarged the TF relation sample sizes with
extensive 21cm observations with Arecibo Telescope. They derived . H0 = 69 km s−1
Mpc−1 [160].
16 R. B. Tully

A bit of a wild card was introduced with the introduction by Avishai Dekel and oth-
ers of an analysis tool that came to be called POTENT [161, 162]. A smoothed three-
dimensional velocity field was inferred from observed radial velocities assuming a
potential flow, consistent with peculiar velocities arising from density perturbations.
Initial conditions were linked with final positions using the Zel’dovich approximation
assuming linear theory of the growth of structure. Dekel et al. were able to generate
maps of the velocity and density field within a radius of 8,000 km s−1 that impres-
sively resemble maps being built two decades later. However, it was a considerable
distraction that values of matter density were being inferred that were .Ωm ∼ 1 or
only slightly less [163]. A similar high value was found for .Ωm from a power spec-
trum analysis of the density fluctuations arising from the Giovanelli et al. observed
peculiar velocities [164]. By contrast, an analysis based on numerical action orbit
reconstructions by Ed Shaya, Jim Peebles, and this author derived a value of matter
density of .Ωm = 0.17 ± 0.10 [165]. Igor Karachentsev and others were remarking
on the extremely cold local flow [166]. It could be seen from orbit reconstructions
that if .Ωm ∼ 1 then there would be the triple-value signatures of collapse [167] all
over the map, a situation far from actuality. The situation was nicely summarized
by contributions at a workshop organized by Stéphane Courteau in 1999 on cosmic
flows in Victoria, Canada [168].
As the century came to an end, there was a remarkable paradigm shift. A foun-
dation for it had been laid 20 years earlier by Alan Guth’s ideas of the origin of the
Universe in an inflation driven by vacuum energy [169]. It was thought natural that
this origin would result in a universe that is close to topologically flat. The power
spectrum of cosmic microwave background fluctuations with a peak on an angu-
lar scale of about a degree provided compelling evidence that this is the case. The
measurements that SNIa at .z ∼ 1 are fainter than expected if .Ωm = 1 favored the
requirement of a positive .Ʌ contribution. There was general agreement that redshift
surveys were indicating that there was too much structure on large scale and that
biasing between the distribution of galaxies and the distribution of matter was insuf-
ficient to rescue .Ωm = 1 models. Even though it was demonstrated that there are
large scale flows [59], flows on local scales are quiet, direct evidence for modest
fluctuations in matter. Then there was abundant evidence from measurements of the
Hubble parameter that, assuredly, the CDM universe with closure density in matter
was untenable and even an open universe with .ΩɅ = 0 was contradicted.
The one model that accommodated the observations posited that the Universe is
spatially flat due to the combination of a currently dominant “dark energy” and sub-
dominant matter, most of it “dark matter”. This .ɅCDM model relieved the Hubble
constant tension at the time. Any value of . H0 in the range being proposed could be
accommodated. The value from the TF relation favored by this author as of 2000 was
−1
. H0 = 77 km s Mpc−1 , lower than earlier estimates [151] primarily because of the
availability of 24 zero-point calibrators, whereas there were only 3 in 1988 [170].
The HST Key Project derived the value . H0 = 71 from their calibration of the TF
relation [171]. Overall, the Key Project best value was. H0 = 72 km s−1 Mpc−1 [121].
For a brief period, there was no Hubble constant tension. That situation would not
last.
2 The Hubble Constant: A Historical Review 17

2.6 . H0 : 67 or 75?

Increasingly detailed observations were made of the cosmic microwave background

with the Wilkinson Microwave Anisotropy Probe (WMAP) that operated between
2001 and 2010 and with the Planck satellite between 2009 and 2013. The spatial
power spectrum of fluctuations could be fit in impressive detail with a .ɅCDM model
with. H0 = 67.4 ± 0.5 km s−1 Mpc−1 and.Ωm = 0.315 ± 0.007 [172–174]. Acoustic
oscillations stimulated around dark matter fluctuations in the plasma of the hot Uni-
verse would have an imprint in baryon structures in the cold Universe on the scale of
the sound horizon at recombination. These baryon acoustic oscillations (BAO) give
rise to a feature in two-point correlation functions over a range of redshifts, obtained
in studies of large galaxy redshift surveys like the SDSS-III Baryon Oscillation Spec-
troscopic Survey (BOSS) and the Dark Energy Survey (DES) [175–177]. The angular
scale of the BAO feature closely tracks the evolution of the Hubble parameter as a
function of redshift, . H (z), as anticipated by the .ɅCDM model defined by the Planck
analysis. Similar consistent results are found from the power spectrum in the galaxy
distributions of these redshift surveys [178]. These exceptionally stringent observa-
tional constraints require that . H0 67 km s−1 Mpc−1 within .∼ 1% if the popular
.ɅCDM model correctly describes our Universe.

Meanwhile since 2000, attention has been given to the various facets of the dis-
tance scale ladder and significantly higher values of . H0 were reported, giving rise
to a new Hubble tension. Methodologies have been diverse and details regarding
each will be discussed in the ensuing chapters in this book. Only highlights of major
programs will be given attention here.
Begin at the foundation with HST. Though the number of targets is yet small,
high quality stellar parallaxes have been obtained with HST to Galactic cepheids
through a spatial scanning technique [179]. The observations are important because,
earlier, all but two of the Galactic cepheids that had good parallax distances have
shorter periods (.T < 10 days) than the cepheids studied in external galaxies. There
are to date HST observations of cepheid variables in 42 host galaxies of SNIa events.
The coupling with SNIa observations out to .z ∼ 1 yields . H0 = 73.3 ± 1.0 km s−1
Mpc−1 [180].
Taken at face value, the difference between this ladder . H0 and the early universe
.ɅCDM model . H0 is .5σ . The concern is unidentified systematics. For confirmation,
a completely independent path is needed. Such a path is being explored, although it
is not yet fully developed. It begins with the replacement of cepheids for distance
measurements to nearby galaxies with tip of the red giant branch. Currently, TRGB
absolute magnitudes are established through parallax measurements to horizontal
branch and RR Lyrae stars, giving distances to Galactic clusters and Milky Way
companion satellites, whence establishing the absolute magnitudes of the brightest
stars on the red giant branch [107, 181]. The astronomy community was startled
in 2020 when Wendy Freedman and team replaced the calibration of SNIa with
TRGB distances to host galaxies in place of cepheid distances, and found . H0 =
69.6 ± 1.7 km s−1 Mpc−1 , consistent with the early universe value [182]. A re-
18 R. B. Tully

evaluation of the study determined . H0 = 71.5 ± 1.8 km s−1 Mpc−1 , consistent but
leaning toward tension [183].
To have a completely separate path, there needs to be an alternative to SNIa
that can be applied at substantial redshifts. Core collapse supernovae (SNII) offer a
possibility. It has been demonstrated that there is a correlation between expansion
velocities and peak magnitudes, with scatter of about 15%. With still a small sample,
a calibration based on cepheids is giving . H0 = 76 ± 5 km s−1 Mpc−1 [184]. The
surface brightness fluctuation technique probably offers an even greater avenue for
progress because observations in the infrared with space telescopes give distances
with accuracies of 5% and can reach to several hundred megaparsecs. With current
applications, again grounded through cepheid calibrations, . H0 = 73.3 ± 2.4 km s−1
Mpc−1 is being found [185].
Programs to study the velocity field of galaxies have been concurrent with efforts to
define the Hubble constant. The peculiar velocity of a galaxy can be obtained indepen-
dent of an absolute calibration and hence of . H0 . Very large samples have been accu-
mulated involving the TF and fundamental plane relations, extending to.0.05c − 0.1c
and providing wide coverage of the sky. Individual distance uncertainties are 20–25%
but sample sizes are in the 10’s of thousands. There have been two noteworthy fun-
damental plane contributions arising from wide field imaging and redshift programs:
the Six Degree Field Galaxy Survey for peculiar velocities (6dFGSv) targeting 9,000
E.−S0 galaxies in the south celestial hemisphere out to 16,000 km s−1 [186, 187],
and the SDSS peculiar velocity sample of 34,000 E.−S0 galaxies in the north celes-
tial, north Galactic quadrant extending to 30,000 km s−1 [188]. The latter sample
was given a zero-point calibration ultimately based on cepheids, which resulted in
the determination . H0 = 75 ± 2 km s−1 Mpc−1 .
Complementary observations of spiral galaxies have received impetus from exten-
sive 21cm neutral hydrogen survey programs [189, 190]. Major velocity field stud-
ies have been published based on TF relation distance measurements [191–195]. It
has been suggested that the correlation with spiral galaxy rotation rate is tightened
if HI flux is added to the optical or infrared flux, creating a measure of the total
baryonic mass. The resultant correlation has been called the baryonic TF relation
(BTFR) [196, 197]. The BTFR has been used with a sample of 10,000 galaxies, with
zero point scaling set by a combination of cepheid and TRGB distances, to obtain
−1
. H0 = 75.5 ± 2.5 km s Mpc−1 [198].
It is an unfortunate feature of the distance ladder that there is no one methodology
that is useful from the nearby within the range of parallax measurements to the
redshift regime beyond peculiar velocity effects. By necessity, methodologies must
be interlinked. If the Hubble constant is to be determined robustly from the ladder,
the linkages must be robust, which is to say, they must be abundant and diverse.
However, making linkages is tricky. For example, cepheid variables are only found
in systems with young stars; likewise SNII. The use of the TF relation is restricted to
spiral galaxies. The fundamental plane and surface brightness fluctuation methods
are only used with massive early-type galaxies. The TRGB method can be used to
derive distances to essentially any galaxy but only if it is nearby; within 10 Mpc
observing with HST, unless a special effort is made. SNIa can arise in any kind of
2 The Hubble Constant: A Historical Review 19

host but they are rare nearby. There is the probability that there are subtle variations in
the luminosities of SNIa that depend on host properties, such as whether or not there
are young stars, which raises issues if calibrations only involve cepheids [199–201].
By necessity, unless targets are within the reach of cepheid or TRGB calibration,
linkages between methods must be through group affiliations. Groups can range
from clusters of many hundreds of members to pairs. Linkages between method-
ologies across this full range are useful and desirable. Good group catalogs should
have a physical basis, have a high level of completion among the galaxy samples
with distances, and have minimal contamination from interlopers. Groups considered
appropriate [202–204] were used to acquire linkages between the largest available
samples of the procedures described above in the compilation of the Cosmicflows-4
compendium of 56,000 galaxy distances [205]. Samples were brought to a common
absolute scale based on cepheid and TRGB calibrations with a Bayesian Markov
chain Monte Carlo (MCMC) analysis (where a small cepheid-TRGB scale differ-
ence was reconciled with preliminary parallax information from Gaia satellite obser-
vations). The value of the Hubble constant from the Cosmicflows-4 compilation is
arguably the best available determination from the distance ladder because it inte-
grates all the major methodologies and surveys. . H0 = 74.6 ± 0.8 km s−1 Mpc−1 is
reported [205]. This statistical error is very small because the combined sample of
56,000 galaxies in 38,000 groups is so large. It is estimated that systematic errors
could be as large as 3 km s−1 Mpc−1 .

2.6.1 . H0 from Models

A number of modeled observations that are independent of the distance ladder have
been proposed that have dependence on the Hubble constant, several of which are
discussed in this volume. In each instance, it must be asked if the model constrains
the Hubble constant or if the Hubble constant constrains the model. To this reviewer’s
awareness, the strongest case of a model constraint on a galaxy distance is provided
by the nuclear maser observations of NGC 4258 [206], because of the remarkably
fortuitous edge-on geometry, locations, and strength of the maser signals. The occur-
rence of this event so near by (7.58 Mpc) is one of extreme luck because the few
other nuclear maser cases that have been studied give distance results that are far less
certain [207].
The detached eclipsing binary measurements that give a distance to the Large
Magellanic Cloud might likewise pass the bar of informing the extragalactic distance
scale [208]. It does depend on knowledge of a surface brightness-color relation and,
more, on the locations of targets across a 5 degree field (a projection that is 9% of
the LMC distance). The claimed precision of the LMC distance is 1% but what does
that mean for an irregular system that is so large on the sky and presumably in depth?
The LMC has played an important role in the calibration of the cepheid scale but, in
the fullness of time, perhaps the two Magellanic clouds will be by-passed in distance
scale studies.
20 R. B. Tully

There are a number of physical properties of the Universe that, to the degree that
they are properly modeled, bear on the value of the Hubble constant. There are strong
gravitational lensing time delays of AGN signal variations through cluster halos
[209], gravitational waves arising from neutron star mergers [210], the Sunyaev-
Zel’dovich effect involving comparisons of the X-ray flux and microwave background
damping associated with cluster plasmas [211], cosmic chronometers giving record
to the relative passage of time with massive, passively evolving galaxies [212], and
the pair-wise galaxy separation mapping of baryon acoustic oscillation features at
various epochs [213]. Generally, the evaluation of these interesting phenomena is
done in the context of the .ɅCDM paradigm, which is at the root of the . H0 tension
controversy.
Some time ago, George Efstathiou initiated an attempt to reconcile the distance
ladder measurements with the early universe implications for . H0 [214]. He explored
possibilities of systematics that might affect ladder measurements; in other words,
relieving the Hubble tension by postulating that the ladder distances are in error.
Of course, systematic errors are possible. However, for this reviewer, the effort of
reconciling distance scale observations with a favored world model too easily brings
back memories of the end of the last century. The young turks of the time, of which
I suppose I was one, were rather convinced that, within a Friedmann-Lemaître-
Robertson-Walker framework, only a .Ʌ /= 0 cosmology model would accommodate
the distance scale measurements that used quality tools. The theoretical community
too easily discounted information on the distance scale, failing to recognize that the
influential voices of Alan Sandage and Gustav Tammann were grounded in the same
exercise of that of Efstathiou.

2.7 Future History

Readers can be forgiven for despairing of a resolution of the question of the current
expansion rate of the Universe after a hundred years of effort, though the problem
is well posed. Happily, the tools are at hand for an unambiguous resolution with
the Global Astrometric Interferometer for Astrophysics (Gaia) and the James Webb
Space Telescope (JWST). The problem can be separated into three parts: (1) the firm
calibration of the methods founded within the Milky Way, specifically cepheids and
TRGB, (2) the transfer of the calibration to methods that can be used at substantial
redshifts, particularly SNIa and surface brightness fluctuation, and (3) the acquisition
of large samples with those latter methods.
It can be anticipated that there will be serious advances on the first problem over
the next year or so. The Gaia third data release (DR3) in June, 2022, provides parallax
information on thousands of cepheids, tens of thousands of RR Lyrae stars, and hun-
dreds of thousands of red giant branch stars with corollary metallicity information.
Anything written today may be well out of date to the reader.
The way forward with problems (2) and (3) will involve paths that are indepen-
dent yet connected. There are two obvious, well-defined paths; one involving young
2 The Hubble Constant: A Historical Review 21

Population I stars and the other involving old Population II stars. The Pop I path is
already well known: cepheid variables calibrate SNIa which in turn are observed to
high redshift. The Pop II path uses TRGB to calibrate surface brightness fluctuations
in massive elliptical galaxies which in turn are observed to high redshift. These two
paths can stand alone, but there are ample opportunities for cross-links through com-
monality of calibrator hosts, SNIa in ellipticals, and group associations. TRGB can
directly calibrate SNIa in late as well as early galaxy types.
JWST is particularly well suited to exploit the Pop II path. The volume of space
opened up for TRGB observations is an order of magnitude greater with JWST
than with HST (from 10 Mpc to greater than 20 Mpc). The increase in volume for
surface brightness fluctuation measurements is at least 30-fold (from 100 Mpc to
.∼ 300 Mpc). If . H0 is to be determined to a precision of 1%, calibration samples
must have sizes of at least .102 and samples at high redshift should be significantly
larger. With JWST, it is more feasible to expect that the necessary large calibrator
samples can be acquired for TRGB distances rather than for cepheid distances.
The SNIa tool can also be improved. While there is excitement that the Vera C.
Rubin Observatory will make acquisitions of large numbers of SNIa at large red-
shifts, advancing our knowledge regarding cosmic acceleration, smaller telescopes
already in operation are sufficient for the simpler problem of the Hubble constant.
Surveys with names like Pan-STARRS, ATLAS, ZTF, and ASAS-SN are picking up
several SNIa events within .0.1c every night. If only there are the resources for man-
power, samples of .104 + could be gathered in a few years. With such large samples,
complemented by multi-epoch spectroscopy and photometry at multiple optical and
infrared bands, the information will be available to fine-tune the SNIa distance tool
technology to account for variations in the paths to supernova events.
Two other large surveys in Australia will have major impacts on studies of the
nature of large scale structure and how it formed. WALLABY is a 21cm blind survey
with the Australian Square Kilometer Array Pathfinder that is expected to provide
some 200,000 TF measures with a mean redshift of.z = 0.05. The Taipan Galaxy Sur-
vey uses a multiplexed spectrograph on the UK Schmidt Telescope at Siding Springs
Observatory and is anticipated to obtain some 50,000 fundamental plane distances
within.z = 0.1. These studies promise to make valuable contributions toward a deter-
mination of the Hubble constant. However, their uncertainties of 25% per measure-
ment will leave them more vulnerable to systematics than can be expected with SNIa
or surface brightness fluctuation measurements with 5% individual uncertainties.
As the 20th century came to an end, ladder measurements of the Hubble constant
were at odds with the favored cosmological model of the time of cold dark matter
with.Ʌ = 0. The new favorite became the.ɅCDM model with dark energy giving rise
to acceleration of space in a topologically flat universe. Yet ladder measurements,
continuously improving, create doubts that this currently favorite model is complete.
Yes, there is a Hubble tension.

Acknowledgements Two of the leaders in the quest to measure the Hubble constant, Marc Aaron-
son and Jeff Willick, died in tragic accidents in their prime.There are so many others who have
contributed, too many to be individually acknowledged so I won’t even try.
22 R. B. Tully

References

1. H. Shapley, Bull. Nat. Res. Counc. 2, 171 (1921)

2. H.D. Curtis, Bull. Nat. Res. Counc. 2, 194 (1921)
3. E.P. Hubble, The Observatory 48, 139 (1925)
4. H.S. Leavitt, E.C. Pickering, Harv.Coll. Obs. Circ. 173, 1 (1912)
5. E. Opik, ApJ 55, 406 (1922). https://doi.org/10.1086/142680
6. V.M. Slipher, Pop. Astron. 23, 21 (1915)
7. E.P. Hubble, Proc. Nat. Acad. Sci. 15, 168 (1929). https://doi.org/10.1073/pnas.15.3.168
8. G. Lemaître, Ann. Soc. Sci. Brux. 47, 49 (1927)
9. A. Einstein, Sitzungsb. König. Preuss. Akad. 142 (1917)
10. E.P. Hubble, Realm of the Nebulae (Yale University Press). ISBN: 9780300025002
11. J.H. Oort, BAN 6, 155 (1931)
12. A.S. Eddington, MNRAS 95, 636 (1935). https://doi.org/10.1093/mnras/95.8.636
13. G.B. Dalrymple, The Age of the Earth (Stanford University Press). ISBN: 0804723311
14. H. Bondi, T. Gold, MNRAS 108, 252 (1948). https://doi.org/10.1093/mnras/108.3.252
15. F. Hoyle, MNRAS 108, 372 (1948). https://doi.org/10.1093/mnras/108.5.372
16. W. Baade, ApJ 100, 137 (1944). https://doi.org/10.1086/144650
17. W. Baade, Trans. I.A.U. 8, 397 (1954)
18. W. Baade, PASP 68, 5 (1956). https://doi.org/10.1086/126870
19. M.L. Humason, N.U. Mayall, A.R. Sandage, AJ 61, 97 (1956). https://doi.org/10.1086/
107297
20. A.R. Sandage, ApJ 127, 513 (1958). https://doi.org/10.1086/146483
21. E. Holmberg, Lund Medd. Astron. Obs. Ser. II, 136 (1958)
22. S. van den Bergh, Zf. Ap 49, 198 (1960)
23. S. van den Bergh, JRASC 54, 49 (1960)
24. J.L. Sersic, Zf Ap. 50, 168 (1960)
25. R. Racine, JRASC 62, 367 (1968)
26. S. van den Bergh, Nature 225, 503 (1970). https://doi.org/10.1038/225503a0
27. M.S. Roberts, AJ 74, 859 (1969). https://doi.org/10.1086/110874
28. A.A. Penzias, R.W. Wilson, ApJ 142, 419 (1965). https://doi.org/10.1086/148307
29. R.H. Dicke et al., ApJ 142, 414 (1965). https://doi.org/10.1086/148306
30. K. Janes, P. Demarque, ApJ 264, 206 (1983). https://doi.org/10.1086/160587
31. A. Chieffi, O. Straniero, ApJS 71, 47 (1989). https://doi.org/10.1086/191364
32. A.R. Sandage, G.A. Tammann, ApJ 194, 223 (1974). https://doi.org/10.1086/153238
33. A.R. Sandage, G.A. Tammann, ApJ 194, 559 (1974). https://doi.org/10.1086/153275
34. A.R. Sandage, G.A. Tammann, ApJ 196, 313 (1975). https://doi.org/10.1086/153413
35. A.R. Sandage, G.A. Tammann, ApJ 197, 265 (1975). https://doi.org/10.1086/153510
36. A.R. Sandage, G.A. Tammann, ApJ 256, 339 (1982). https://doi.org/10.1086/159911
37. G. de Vaucouleurs, G. Bollinger, ApJ 233, 433 (1979). https://doi.org/10.1086/157405
38. G. de Vaucouleurs, W.L. Peters, ApJ 248, 395 (1981). https://doi.org/10.1086/159165
39. G. de Vaucouleurs, ApJ 268, 468 (1983). https://doi.org/10.1086/160972
40. G. de Vaucouleurs, H.G. Corwin Jr., ApJ 297, 23 (1985). https://doi.org/10.1086/163499
41. G. de Vaucouleurs, W.L. Peters, ApJ 303, 19 (1986). https://doi.org/10.1086/164048
42. R.B. Tully, J.R. Fisher, A&A 54, 66 (1977)
43. J.R. Fisher, R.B. Tully, ComAp 7, 85 (1977)
44. R.B. Tully, J.R. Mould, M. Aaronson, ApJ 257, 527 (1982). https://doi.org/10.1086/160009
45. M. Aaronson, J.P. Huchra, J.R. Mould, ApJ 229, 1 (1979). https://doi.org/10.1086/156923
46. T. Okamoto, V.R. Eke, C.S. Frenk, A. Jenkins, MNRAS 363, 1299 (2005). https://doi.org/10.
1111/j.1365-2966.2005.09525.x
47. F. Governato et al., MNRAS 374, 1479 (2007). https://doi.org/10.1111/j.1365-2966.2006.
11266.x
48. S. Trujillo-Gomez, A. Klypin, J. Primack, A.J. Romanowsky, ApJ 742, 16 (2011). https://doi.
org/10.1088/0004-637X/742/1/16
2 The Hubble Constant: A Historical Review 23

49. J. Guedes, S. Callegari, P. Madau, L. Mayer, ApJ 742, 76 (2011). https://doi.org/10.1088/

0004-637X/742/2/76
50. F. Marinacci, R. Pakmor, V. Springel, MNRAS 437, 1750 (2014). https://doi.org/10.1093/
mnras/stt2003
51. M. Glowacki, E. Elson, R. Davé, MNRAS 498, 3687 (2020). https://doi.org/10.1093/mnras/
staa2616
52. M. Aaronson et al., ApJ 239, 12 (1980). https://doi.org/10.1086/158084
53. G.D. Bothun et al., ApJ 278, 475 (1984). https://doi.org/10.1086/161814
54. M. Aaronson et al., ApJ 302, 536 (1986). https://doi.org/10.1086/164014
55. M. Aaronson et al., ApJ 258, 64 (1982). https://doi.org/10.1086/160053
56. S.M. Faber, R.E. Jackson, ApJ 204, 668 (1976). https://doi.org/10.1086/154215
57. S. Djorgovski, M. Davis, ApJ 313, 59 (1987). https://doi.org/10.1086/164948
58. A. Dressler et al., ApJ 313, 42 (1987). https://doi.org/10.1086/164947
59. A. Dressler et al., ApJ 313, L37 (1987). https://doi.org/10.1086/184827
60. D. Lynden-Bell et al., ApJ 326, 19 (1988). https://doi.org/10.1086/166066
61. D. Burstein, S.M. Faber, A. Dressler, ApJ 354, 18 (1990). https://doi.org/10.1086/168664
62. A. Dressler, S.M. Faber, ApJ 354, L45 (1990). https://doi.org/10.1086/164947
63. M. Colless et al., MNRAS 321, 277 (2001). https://doi.org/10.1046/j.1365-8711.2001.04044.
x
64. M.J. Hudson et al., MNRAS 327, 265 (2001). https://doi.org/10.1046/j.1365-8711.2001.
04786.x
65. M. Bernardi et al., AJ 123, 2990 (2002). https://doi.org/10.1086/340463
66. J.L. Tonry, D.P. Schneider, AJ 96, 807 (1988). https://doi.org/10.1086/114847
67. J.L. Tonry, E.A. Ajhar, G.A. Luppino, ApJ 346, 57 (1989). https://doi.org/10.1086/185578
68. J.L. Tonry, ApJ 373, L1 (1991). https://doi.org/10.1086/186037
69. J.L. Tonry, J.P. Blakeslee, E.A. Ajhar, A. Dressler, ApJ 475, 399 (1997). https://doi.org/10.
1086/303576
70. J.L. Tonry, J.P. Blakeslee, E.A. Ajhar, A. Dressler, ApJ 530, 625 (2000). https://doi.org/10.
1086/308409
71. J.L. Tonry et al., ApJ 546, 681 (2001). https://doi.org/10.1086/318301
72. J.R. Mould, J. Kristian, ApJ 305, 591 (1986). https://doi.org/10.1086/164273
73. M.G. Lee, W.L. Freedman, B.F. Madore, ApJ 417, 553 (1993). https://doi.org/10.1086/173334
74. G.S. Da Costa, T.E. Armandroff, AJ 100, 162 (1990). https://doi.org/10.1086/115500
75. S. Sakai, B.F. Madore, W.L. Freedman, ApJ 461, 713 (1996). https://doi.org/10.1086/177096
76. G.H. Jacoby, R. Ciardullo, H.C. Ford, ApJ 356, 332 (1990). https://doi.org/10.1086/168843
77. R.H. Mendez, R.P. Kudritzki, R. Ciardullo, G.H. Jacoby, A&A 275, 534 (1993)
78. R.B. Tully, S.C. Wolff, ApJ 281, 67 (1984). https://doi.org/10.1086/162075
79. R.P. Kudritzki et al., ApJ 788, 56 (2014). https://doi.org/10.1088/0004-637X/788/1/56
80. D.A. Hanes, MNRAS 188, 901 (1979). https://doi.org/10.1093/mnras/188.4.901
81. R. McMillan, R. Ciardullo, G.H. Jacoby, ApJ 416, 62 (1993). https://doi.org/10.1086/173215
82. B.C. Whitmore et al., ApJ 454, L73 (1995). https://doi.org/10.1086/309788
83. C.J. Pritchet, S. van den Bergh, ApJ 318, 507 (1987). https://doi.org/10.1086/165387
84. M. della Valle, M. Livio, ApJ 452, 704 (1995). https://doi.org/10.1086/176342
85. S. Refsdal, MNRAS 128, 307 (1964). https://doi.org/10.1093/mnras/128.4.307
86. C.S. Kochanek, P.L. Schechter (2003), arXiv:astro-ph/0306040
87. U. Borgeest, S. Refsdal, A&A 141, 318 (1984)
88. G. Rhee, Nature 350, 211 (1991). https://doi.org/10.1038/350211a0
89. R.A. Sunyaev, Ya.B. Zel’dovich, Comm. Ap. Space Sci. 4, 173 (1972)
90. J. Silk, S.D.M. White, ApJ 226, L103 (1978). https://doi.org/10.1086/182841
91. M. Birkinshaw, J.P. Hughes, K.A. Arnaud, ApJ 379, 466 (1991). https://doi.org/10.1086/
170522
92. M. Birkinshaw, J.P. Hughes, ApJ 420, 33 (1994). https://doi.org/10.1086/173540
93. T. Herbig, C.R. Lawrence, A.C.S. Readhead, S. Gulkis, ApJ 449, L5 (1995). https://doi.org/
10.1086/309616
24 R. B. Tully

94. Y. Rephaeli, ARA&A 33, 541 (1995). https://doi.org/10.1146/annurev.aa.33.090195.002545

95. W. Baade, ApJ 88, 285 (1938). https://doi.org/10.1086/143983
96. W. Baade, F. Zwicky, ApJ 88, 411 (1938). https://doi.org/10.1086/143996
97. C.T. Kowal, AJ 73, 1021 (1968). https://doi.org/10.1086/110763
98. D. Branch, B. Patchett, MNRAS 161, 71 (1973). https://doi.org/10.1093/mnras/161.1.71
99. R.P. Kirshner, J. Kwan, ApJ 193, 27 (1974). https://doi.org/10.1086/153123
100. D. Branch et al., ApJ 244, 780 (1981). https://doi.org/10.1086/158755
101. D. Branch, ApJ 320, L23 (1987). https://doi.org/10.1086/184970
102. B.P. Schmidt et al., ApJ 432, 42 (1994). https://doi.org/10.1086/174546
103. A. Fisher, D. Branch, P. Hoflich, A. Khokhlov, ApJ 447, L73 (1995). https://doi.org/10.1086/
309563
104. N. Bartel, Lect. Not. Phys. 224, (1985). https://doi.org/10.1007/3-540-15206-7
105. B. Chaboyer et al., ApJ 494, 96 (1998). https://doi.org/10.1086/305201
106. F. Pont, M. Mayor, C. Turon, D.A. Vandenberg, A&A 329, 87 (1998)
107. E. Carretta, R.G. Gratton, G. Clementini, F. Fusi Pecci, ApJ 533, 215 (2000). https://doi.org/
10.1086/308629
108. A. Saha et al., ApJ 425, 14 (1994). https://doi.org/10.1086/173957
109. A. Saha, A.R. Sandage et al., ApJS 107, 693 (1996). https://doi.org/10.1086/192378
110. A. Saha, A.R. Sandage et al., ApJ 486, 1 (1997). https://doi.org/10.1086/304507
111. A. Saha, A.R. Sandage et al., ApJ 522, 802 (1999). https://doi.org/10.1086/307693
112. A. Saha, A.R. Sandage et al., ApJ 562, 314 (2001). https://doi.org/10.1086/323529
113. A.R. Sandage, A. Saha et al., ApJ 460, L15 (1996). https://doi.org/10.1086/309973
114. G.A. Tammann, A.R. Sandage, A. Saha et al., Proc. HST Symp. 14, 222 (2003). https://doi.
org/10.1086/309973
115. D.D. Kelson et al., ApJ 463, 26 (1996). https://doi.org/10.1086/177221
116. B.K. Gibson et al., ApJ 529, 723 (2000). https://doi.org/10.1086/308306
117. W.L. Freedman et al., ApJ 427, 628 (1994). https://doi.org/10.1086/174172
118. W.L. Freedman et al., Nature 371, 757 (1994). https://doi.org/10.1038/371757a0
119. A. Turner et al., ApJ 505, 207 (1998). https://doi.org/10.1086/306150
120. L. Ferrarese et al., ApJS 128, 431 (2000). https://doi.org/10.1086/313391
121. W.L. Freedman et al., ApJ 553, 47 (2001). https://doi.org/10.1086/320638
122. M.M. Phillips, ApJ 413, L105 (1993). https://doi.org/10.1086/186970
123. A.G. Riess, W.P. Press, R.P. Kirshner, ApJ 438, L17 (1995). https://doi.org/10.1086/187704
124. S. Jha et al., ApJS 125, 73 (1999). https://doi.org/10.1086/313275
125. M.M. Phillips et al., AJ 118, 1766 (1999). https://doi.org/10.1086/301032
126. Y. Wang, ApJ 536, 531 (2000). https://doi.org/10.1086/308958
127. M. Hernandez et al., MNRAS 319, 223 (2000). https://doi.org/10.1046/j.1365-8711.2000.
03841.x
128. A.R. Sandage, G.A. Tammann, ApJ 464, L51 (1996). https://doi.org/10.1086/310083
129. A.R. Sandage, AJ 111, 1 (1996a). https://doi.org/10.1086/117755
130. A.R. Sandage, AJ 111, 18 (1996b). https://doi.org/10.1086/117756
131. M. Federspiel, G.A. Tammann, A.R. Sandage, ApJ 495, 115 (1998). https://doi.org/10.1086/
305263
132. B.P. Schmidt et al., ApJ 507, 46 (1998). https://doi.org/10.1086/306308
133. S. Perlmutter et al., ApJ 517, 565 (1999). https://doi.org/10.1086/307221
134. G.F. Smoot et al., ApJ 396, L1 (1992). https://doi.org/10.1086/186504
135. C.L. Bennett et al., ApJ 464, L1 (1996). https://doi.org/10.1086/310075
136. S. Hanany et al., ApJ 545, L5 (2000). https://doi.org/10.1086/317322
137. P. de Bernardis et al., Nature 404, 955 (2000). https://doi.org/10.1038/35010035
138. C.B. Netterfield et al., ApJ 571, 604 (2002). https://doi.org/10.1086/340118
139. W. Saunders et al., MNRAS 242, 318 (1990). https://doi.org/10.1093/mnras/242.3.318
140. W. Saunders et al., Nature 349, 32 (1991). https://doi.org/10.1038/349032a0
141. G. Efstathiou, W.J. Sutherland, S.J. Maddox, Nature 348, 705 (1990). https://doi.org/10.1038/
348705a0
2 The Hubble Constant: A Historical Review 25

142. J. Loveday et al., ApJ 400, L43 (1992). https://doi.org/10.1086/186645

143. E.F. Bunn, M. White, ApJ 480, 6 (1997). https://doi.org/10.1086/303955
144. J.R. Bond, A.H. Jaffe, R. Soc. London Phil. Trans. A 357, 57 (1999). https://doi.org/10.1098/
rsta.1999.0314
145. J.A. Peacock, MNRAS 284, 885 (1997). https://doi.org/10.1093/mnras/284.4.885
146. J.G. Bartlett et al., Science 267, 980 (1995). https://doi.org/10.1126/science.267.5200.980
147. M. White et al., MNRAS 276, L69 (1995). https://doi.org/10.1093/mnras/276.1.L69
148. C.H. Lineweaver et al., A&A 322, 365 (1997). https://doi.org/10.48550/arXiv.astro-ph/
9610133
149. M.J. Hudson, MNRAS 265, 43 (1993). https://doi.org/10.1093/mnras/265.1.43
150. M.J. Hudson, MNRAS 266, 475 (1994). https://doi.org/10.1093/mnras/266.2.475
151. M.J. Pierce, R.B. Tully, ApJ 330, 579 (1988). https://doi.org/10.1086/307707
152. J.A. Willick, ApJS 92, 1 (1994). https://doi.org/10.1086/191957
153. J.A. Willick et al., ApJ 446, 12 (1995). https://doi.org/10.1086/175762
154. J.A. Willick et al., ApJ 457, 460 (1996). https://doi.org/10.1086/176746
155. J.A. Willick et al., ApJS 109, 333 (1997). https://doi.org/10.1086/312983
156. J.A. Willick et al., ApJ 486, 629 (1997). https://doi.org/10.1086/304551
157. M.A. Strauss, J.A. Willick, Phys. Rep. 261, 271 (1995). https://doi.org/10.1016/0370-
1573(95)00013-7
158. J.A. Willick, M.A. Strauss, ApJ 507, 64 (1998). https://doi.org/10.1086/306314
159. J.A. Willick, P. Batra, ApJ 548, 564 (2001). https://doi.org/10.1086/319005
160. R. Giovanelli et al., ApJ 477, L1 (1997). https://doi.org/10.1086/310521
161. A. Dekel, E. Bertschinger, S.M. Faber, ApJ 364, 349 (1990). https://doi.org/10.1086/169418
162. A. Dekel et al., ApJ 412, 1 (1993). https://doi.org/10.1086/172896
163. A. Dekel et al., ApJ 522, 1 (1999). https://doi.org/10.1086/307636
164. W. Freudling et al., ApJ 523, 1 (1999). https://doi.org/10.1086/307707
165. E.J. Shaya, P.J.E. Peebles, R.B. Tully, ApJ 454, 15 (1995). https://doi.org/10.1086/176460
166. I.D. Karachentsev et al., A&A 389, 812 (2002). https://doi.org/10.1051/0004-6361:20020649
167. J.L. Tonry, M. Davis, ApJ 246, 680 (1981). https://doi.org/10.1086/158965
168. S. Courteau, J.A. Willick, ASP Conf. Ser. 201 (2000)
169. A.H. Guth, PhRvD 23, 347 (1981). https://doi.org/10.1103/PhysRevD.23.347
170. R.B. Tully, M.J. Pierce, ApJ 533, 744 (2000). https://doi.org/10.1086/308700
171. S. Sakai et al., ApJ 529, 698 (2000). https://doi.org/10.1086/308305
172. G. Efstathiou, S. Gratton, QJAp 4, 8 (2021). https://doi.org/10.21105/astro.1910.00483
173. Planck Collaboration et al., A&A 641, A6 (2020). https://doi.org/10.1051/0004-6361/
201833910
174. Planck Collaboration et al., A&A 571, A16 (2014). https://doi.org/10.1051/0004-6361/
201321591
175. E. Aubourg et al., PhRvD 92, 123516 (2015). https://doi.org/10.1103/PhysRevD.92.123516
176. E. Macaulay et al., MNRAS 486, 2184 (2019). https://doi.org/10.1093/mnras/stz978
177. P. Lemos et al., MNRAS 483, 4803 (2019). https://doi.org/10.1093/mnras/sty3082
178. G. d’Amico et al., JCAP 5, 005 (2020). https://doi.org/10.1088/1475-7516/2020/05/005
179. A.G. Riess et al., ApJ 855, 136 (2018). https://doi.org/10.3847/1538-4357/aaadb7
180. A.G. Riess et al., ApJ 934, L7 (2022). https://doi.org/10.3847/2041-8213/ac5c5b
181. A.G. Rizzi et al., ApJ 661, 815 (2007). https://doi.org/10.1086/516566
182. W.L. Freedman et al., ApJ 891, 57 (2020). https://doi.org/10.3847/1538-4357/ab7339
183. G.S. Anand et al., ApJ 932, 15 (2022). https://doi.org/10.3847/1538-4357/ac68df
184. T. de Jaeger et al., MNRAS 496, 3402 (2020). https://doi.org/10.1093/mnras/staa1801
185. J.P. Blakeslee et al., ApJ 911, 65 (2021). https://doi.org/10.3847/1538-4357/abe86a
186. C.M. Springob et al., MNRAS 445, 2677 (2014). https://doi.org/10.1093/mnras/stu1743
187. F. Qin et al., MNRAS 477, 5150 (2018). https://doi.org/10.1093/mnras/sty928
188. C. Howlett et al., MNRAS 515, 953 (2022). https://doi.org/10.1093/mnras/stac1681
189. M.P. Haynes et al., ApJ 861, 49 (2018). https://doi.org/10.3847/1538-4357/aac956
190. H.M. Courtois, R.B. Tully, MNRAS 447, 1531 (2015). https://doi.org/10.1093/mnras/stu2405
26 R. B. Tully

191. C.M. Springob et al., ApJS 172, 599 (2007). https://doi.org/10.1086/519527

192. T. Hong et al., MNRAS 487, 2061 (2019). https://doi.org/10.1093/mnras/stz1413
193. R.B. Tully et al., ApJ 676, 184 (2008). https://doi.org/10.1086/527428
194. R.B. Tully et al., AJ 146, 86 (2013). https://doi.org/10.1088/0004-6256/146/4/86
195. R.B. Tully, H.M. Courtois, J.G. Sorce, AJ 152, 50 (2016). https://doi.org/10.3847/0004-6256/
152/2/50
196. S.S. McGaugh et al., ApJ 533, L99 (2000). https://doi.org/10.1086/312628
197. F. Lelli, S.S. McGaugh, J.M. Schombert, ApJ 816, L14 (2016). https://doi.org/10.3847/2041-
8205/816/1/L14
198. E. Kourkchi et al., MNRAS 511, 6160 (2022). https://doi.org/10.1093/mnras/stac303
199. B.T. Hayden et al., ApJ 764, 191 (2013). https://doi.org/10.1088/0004-637X/764/2/191
200. M. Childress et al., ApJ 770, 108 (2013). https://doi.org/10.1088/0004-637X/770/2/108
201. M. Roman et al., A&A 615, A68 (2018). https://doi.org/10.1051/0004-6361/201731425
202. R.B. Tully, AJ 149, 171 (2015). https://doi.org/10.1088/0004-6256/149/5/171
203. E. Kourkchi, R.B. Tully, ApJ 843, 16 (2017). https://doi.org/10.3847/1538-4357/aa76db
204. E. Tempel et al., A&A 602, A100 (2017). https://doi.org/10.1051/0004-6361/201730499
205. R.B. Tully et al., ApJ 944, 94 (2023). https://doi.org/10.3847/1538-4357/ac94d8
206. M.J. Reid, D.W. Pesce, A.G. Riess, ApJ 886, L27 (2019). https://doi.org/10.3847/2041-8213/
ab552d
207. D.W. Pesce et al., ApJ 891, L1 (2020). https://doi.org/10.3847/2041-8213/ab75f0
208. G. Pietrzyński et al., ApJ 891, L1 (2019). https://doi.org/10.1038/s41586-019-0999-4
209. S. Birrer et al. (2022). arXiv:2210.10833
210. B.P. Abbott et al., ApJ 909, 218 (2021). https://doi.org/10.3847/1538-4357/abdcb7
211. A. Kozmanyan et al., A&A 621, A34 (2019). https://doi.org/10.1051/0004-6361/201833879
212. M. Moresco et al., ApJ 868, 84 (2018). https://doi.org/10.3847/1538-4357/aae829
213. S. Alam et al., PhRvD 103, 083533 (2021). https://doi.org/10.1103/PhysRevD.103.083533
214. G. Efstathiou (2020). arXiv:2007.10716v2
 Part II
Measurements (and Systematics)
Chapter 3
Trigonometric Parallax

Lennart Lindegren

3.1 Introduction

Parallax, or more precisely stellar trigonometric parallax, refers to the determination

of distances to objects beyond the solar system through accurate angular measure-
ments, using the orbit of the Earth, or of a satellite near the Earth, as baseline for
the triangulation. The term parallax is also used for a variety of other methods to
determine distances in astronomy, such as diurnal, secular, statistical, dynamical,
photometric, and spectroscopic parallax; these are not discussed here.
With the adoption of heliocentrism in the 17th century, the anticipated detection
of stellar parallax was seen as the ultimate proof of the Copernican hypothesis.
Over the following centuries, attempts to measure the effect stimulated instrument
developments and lead to important discoveries [1, 2]. Stellar parallax was finally
detected in the 1830s by Henderson, W. Struve, and, most convincingly, Bessel
(Fig. 3.1) [3]. In the subsequent 50 years the number of stars with a measured parallax
grew very slowly owing to the exceedingly tedious and difficult visual observation
process (Table 3.1). The use of photography and long-focus refractors from the late
19th century gave a burst in numbers but substantial improvements in accuracy only
came with further technological advances in plate measuring engines, telescopes,
and detectors [4]. A new era began with the launch of Hipparcos in 1989 and the
Hubble Space Telescope in 1990, and a quantum leap was taken when Gaia was
launched in 2013. In the radio domain, very long baseline interferometry (VLBI) has
seen a parallel development and is now routinely used to measure the parallaxes for
Galactic maser sources with an accuracy approaching 1 .µas [5].
Throughout the long history of parallax measurements, random and systematic
errors have been of primary concern. What are their causes, how can they be reduced

L. Lindegren (B)
Lund Observatory, Department of Physics, Lund University, 22100 Lund, Sweden
e-mail: lennart.lindegren@fysik.lu.se

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 29
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_3
30 L. Lindegren

Fig. 3.1 Bessel’s determination of the parallax of 61 Cyg [6]. The geometry of the double star
61 Cyg AB and the background stars ‘a’ (BD+37.◦ 4173) and ‘b’ (BD+37.◦ 4179) around 1838 is
shown at left; the other panels show the individual measurements of the angular separations, together
with Bessel’s fit of parallax and proper motion. Bessel’s result for the parallax,.316 ± 20 mas, differs
by only 1.5 standard deviation from the modern value 286 mas

Table 3.1 Examples of historical and current parallax programs and catalogs. The third column is
the median formal uncertainty in mas as given in the catalog, using the conversion .σ𝛡 = 1.5 × p.e.
when probable errors are given
Year No. of .σ𝛡 Reference, remark
stars
1838 1 20 [6] 61 Cyg observed by Bessel
1889 49 30 [7] compilation by Oudemans
1924 1682 13 [8] General catalogue of stellar parallaxes
1935 3933 13 [9] General catalogue of stellar parallaxes (2nd ed.)
1952 5823 13 [10] General catalogue of trigonometric stellar parallaxes (3rd ed.)
1995 8112 10 [11] General catalogue of trigonometric stellar parallaxes (4th ed.)
1997 117 955 1.1 [12] The Hipparcos Catalogue
2005 298 0.5 [13] USNO CCD parallax program (progress report)
2007 117 955 0.9 [14] Hipparcos, the New Reduction
2017 105 0.2 [15] HST/FGS astrometry (review)
2018 7 0.045 [16] HST/WFC3 spatial scanning
2021 .1468 × 0.34 [17] Gaia (E)DR3, all stars (.G 21)
106
.156 × 10 d:o, .G ≤ 17
6 0.043
.7.3 × 10 d:o, .G ≤ 13
6 0.016

by improved instruments and procedures, and how should they be dealt with in
astrophysical applications of stellar parallax? As shown in the rest of this chapter,
these questions remain highly relevant for the best modern determinations.
3 Trigonometric Parallax 31

3.2 Astrometric Model

The reduction of astrometric observations at the microarcsec level requires that the
relevant models of the object, observer, and light propagation from the object to
observer are consistently formulated in a relativistic framework. To model light
propagation from distant celestial objects, as well as the motion of bodies within
the solar system, the Barycentric Celestial Reference System (BCRS) is the rec-
ommended framework [18]. Within the BCRS, spatial coordinates are represented
as 3-vectors, here denoted . b, with origin at the solar system barycentre. The coor-
dinate axes are oriented according to the standard (ICRS) astronomical reference
system defining the celestial coordinates .α and .δ. The temporal coordinate .t is the
Barycentric Coordinate Time, TCB. The units of measurement are the SI meter and
SI second.
In the BCRS system let . bs (t) be the motion of an object (the ‘source’), and . bo (t)
the (known) motion of the observer, for example Gaia. The coordinate direction
towards the source at the time of observation, .tobs , is given by the unit vector

bs (t) − bo (tobs )
. ū(tobs ) = . (3.1)
|bs (t) − bo (tobs )|

Ideally, the time .t should be the instant when the light is emitted from the source.
For sources well beyond the solar system it is however customary to ignore most
of the light travel time (which may be many years) and only consider the variable
part caused by the motion of the observer around the solar system barycentre. That
is also the convention adopted here, achieved by using the ‘barycentric date’ .t =
tobs − ū(tobs )' bo (tobs )c−1 as the time argument in . bs (t), where .c is the speed of light.
(The prime .' denotes scalar product.)
The coordinate direction .ū(tobs ) varies according to the motions of the source
and observer, and the astrometric observations ultimately aim to reconstruct . bs (t)
from a series of such measurements.1 For non-single sources the function . bs (t) can
obviously be very complex. It turns out, however, that a linear model assuming
constant space velocity .v s ,

b (t) = bs (tep ) + (t − tep )v s ,

. s (3.2)

is adequate for the astrometric reduction of the vast majority of sources, at least over
an interval of several years, even though many of them are in reality binaries with
more complex motions [20]. (When needed, non-linear motions can of course be

1 The direction towards the source as measured in the instrument frame, that is relative to a set of
orthogonal axes fixed in the astrometric instrument, differs from the coordinate direction because
of gravitational light deflection in the solar system, stellar aberration, and the celestial orientation
(attitude) of the instrument frame. The first two effects can be accurately corrected by means of the
known motion of the observer and the known solar system ephemerides [19], and can be ignored
for the present discussion. The attitude represents a rigid-body rotation that needs to be determined
from observations; see Sect. 3.5.2.
32 L. Lindegren

included in the process of solving for the astrometric parameters of the source.) Here
t is a reference epoch, preferably chosen not far from the centre of the observed
. ep
interval. Equations (3.1) and (3.2), together with the corresponding approximations,
is known as the standard model of stellar motion (see Appendix A of [21]).
Following [19] we define the parallax of the source as2

A
.𝛡 = , (3.3)
|bs (tep )|

where . A is the astronomical unit. The parallax is the preferred way to parametrize
the distance to an extrasolar object in astrometry.3 The reason for this choice is that,
in the absence of observational biases, the random errors in parallax from a series of
astrometric measurements tend to be symmetric around the true value. This permits
a simple interpretation of the observed parallax in terms of the likelihood function
(Sect. 3.6). As defined in (3.3) the parallax is dimensionless, but it may be more
conveniently expressed as an angle, for example in mas through multiplication by
.648 000 000/π.
The kinematic model in (3.2) has six degrees of freedom. In addition to .𝛡 we
therefore need five more parameters to describe the stellar motion according to the
standard model. Two of them are the celestial coordinates .α, .δ of the barycentric
unit vector . r = bs |bs |−1 at .tep ; two more are the components .μα∗ , .μδ of the proper
motion vector .µ = r × (v s × r)|bs |−1 in the directions of increasing right ascension
and declination, respectively; and the last one is the radial velocity .vr = r ' v s . All six
parameters refer to the adopted reference epoch .tep , although we do not explicitly
write .𝛡 (tep ), etc. Formulas for changing .tep are given in Appendix A of [21].
To first order in the small quantities .𝛡 and .(t − tep )µ, Eqs. (3.1)–(3.3) give
( )
r + (t − tep )v s − bo (tobs ) 𝛡 /A
. ū(tobs ) = | ( ) | (3.4)
| r + (t − tep )v s − bo (tobs ) 𝛡 /A|
. = r + (t − tep )µ + 𝛡 f (tobs ) + O(𝛡 2 ) , (3.5)

where .µ is the previously defined proper motion vector and

. f (tobs ) = r × (r × bo (tobs )/A) (3.6)

is the parallax factor expressed as a vector perpendicular to . r. It is seen that .ū(t)

to first order does not depend on the radial velocity. The second-order terms, known
as perspective acceleration, are important for some nearby high-velocity stars [23],
and for these one must include .vr in the astrometric modelling. For most Galactic

2 The symbol .𝛡 is a variant of the Greek letter ‘pi’. It is used for the parallax to avoid confusion
with the mathematical constant .π that sometimes appears in the same formulae.
3 This definition breaks down at cosmological distances where the BCRS is no longer applicable.

See for example [22] for a discussion of cosmological parallax.

3 Trigonometric Parallax 33

stars, however, the astrometric observations are adequately modelled by just the five
astrometric parameters .α, .δ, .𝛡 , .μα∗ , and .μδ .
Although the vector form of Eq. (3.4) is convenient for numerical computation,
a variant using standard coordinates is often preferred. Standard coordinates .(ξ, η)
are obtained by gnomonic projection of the various vectors onto a tangent plane to
the unit celestial sphere (e.g. [24]). With .(αc , δc ) denoting the adopted tangent point
(normally chosen close to . r) and
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
cos δc cos αc − sin αc − sin δc cos αc
. r c = ⎣ cos δc sin αc ⎦ pc = ⎣ cos αc ⎦ q c = ⎣ − sin δc sin αc ⎦ , (3.7)
sin δc 0 cos δc

we have for the arbitrary direction .u the standard coordinates

p'c u q 'c u r c + pc ξ + q c η
ξ=
. , η = , and conversely u = . (3.8)
r 'c u r 'c u (1 + ξ 2 + η2 )1/2

From Eq. (3.4) we find the standard coordinates of .ū(tobs ),

ξ̄ (tobs ) = ξ0 + (t − tep )μξ + f ξ (tobs )𝛡

. , (3.9)
η̄(tobs ) = η0 + (t − tep )μη + f η (tobs )𝛡

where .(ξ0 , η0 ) are the standard coordinates of . r, .μξ = p'c µ/r 'c r, . f ξ = p'c f /r 'c r,
etc. Provided that the tangent point is close to . r, it can be seen that .μξ and . f ξ are
approximately, but not exactly, equal to .μα∗ and . f α∗ (the parallax factor in right
ascension), and similarly for the components in .η and .δ; strict equality holds only
when . r c = r, in which case we also have .ξ0 = η0 = 0. From a set of observed
standard coordinates .(ξ̄ , η̄), corrected for aberration etc., we can now obtain the
linearised astrometric parameters .ξ0 , .η0 , .μξ , .μη , and .𝛡 by least-squares solution of
the aggregated equations (3.9). Often the equations in .ξ and .η are treated separately,
yielding two independent estimates of the parallax,.𝛡ξ and.𝛡η . Ideally they should be
the same, which can be used as a consistency check. Finally the linearised parameters
.ξ0 , .η0 , .μξ , .μη are transformed back to the normal ones .α, .δ, .𝛡 , .μα∗ using Eq. (3.8).
Equation (3.5) is sometimes written in linearised spherical coordinates, e.g. as

ᾱ(tobs ) = α + (t − tep )μα + f α (tobs )𝛡

. , (3.10)
δ̄(tobs ) = δ + (t − tep )μδ + f δ (tobs )𝛡

where.ᾱ,.δ̄ are the celestial coordinates of.ū,. f α ,. f δ the corresponding parallax factors,
and .μα = μα∗ / cos δ. This kind of expansion should be avoided as it introduces
additional approximations proportional to .tan δ, which could become catastrophic
near the poles. This is avoided through the use of standard coordinates as described
above, or by using the vector form in Eq. (3.4).
34 L. Lindegren

3.3 Differential and Global Techniques

Stellar parallax is often described as the apparent shift, caused by the Earth’s orbital
motion, of a nearby target star against distant background sources. This actually
describes the measurement of relative parallax, that is of the difference between
the parallax of the target star and the (mean) parallax of the background sources.
Measured differential parallaxes usually need to be corrected to ‘absolute’ by adding
an estimate of the mean parallax of the background stars (Sect. 3.4.3).
Bessel’s observations from 1837–38 (Fig. 3.1) may illustrate the determination of
differential and absolute parallaxes. Bessel obtained the value .316.2 ± 20.2 mas for
the differential parallax if 61 Cyg with respect to the background stars ‘a’ and ‘b’.
He was fortunate in that both stars turn out to be distant red giants with parallaxes
(2.2 and 2.0 mas, respectively) that are almost negligible compared with the differ-
ential parallax and its uncertainty [3]. Strictly speaking, however, his measurements
thus gave an absolute parallax of about 318 mas, compared to the modern value
286.0 mas (although in 1838 the parallax was 285.0 mas, owing to the radial velocity
of .−65 km s.−1 ).
From Eq. (3.6) we see that the parallax factor depends only on the barycentric
position of the observer at the time of observation, . bo (tobs ), and the position of
the source, . r. Therefore, if two sources are observed simultaneously from the same
observatory within a small part of the sky, they will have practically the same parallax
factor . f and the difference in coordinate direction becomes

ū (tobs ) − ū1 (tobs ) = r 2 − r 1 + (t − tep )(µ2 − µ1 ) + (𝛡2 − 𝛡1 ) f (tobs ) .(3.11)

. 2

From a series of measurements of .ū2 − ū1 spread over a few years it is clearly
possible to determine the differential astrometric parameters . r 2 − r 1 , .µ2 − µ1 , and
.𝛡2 − 𝛡1 , but not the individual (absolute) values. In particular, it is not possible to

determine .𝛡1 and .𝛡2 separately. Parallax measurements confined to a small field
(.≪ 1 rad) therefore give differential parallaxes of the measured sources, that is up to
some unknown common offset.
In a classical parallax determination, as practised from Bessel’s days up to the
present time with HST and large ground-based telescopes, a target source is measured
with respect to a background of several reference sources, presumed to be much
more distant than the target. In a second step, the measured differential parallax of
the target source is corrected from relative to absolute by adding the mean parallax of
the reference stars. To avoid circularity, the reference parallaxes must be estimated
by some independent means. As discussed in [25] the main methods available are
based on statistical, spectroscopic, and photometric distance estimates. Statistical
methods attempt to determine the mean parallax of stars as a function of (primarily)
magnitude and Galactic coordinates, based on proper motion surveys, star counts,
and Galactic models. Although historically important, the statistical methods may
be too uncertain for modern parallax determinations where it has been superseded
by spectroscopic and photometric methods (e.g., [26, 27]). The spectrophotometric
3 Trigonometric Parallax 35

Fig. 3.2 Illustrating the principles of relative and absolute parallax determination. a Measuring the
small angles . A, . B between a target source and a reference (background) source gives the relative
parallax .𝛡2 − 𝛡1 . b In this configuration the measurement of the large angles . A, . B gives the
absolute parallax of the target source independent of the distance to the reference source

method is further discussed in Sect. 3.4.3 in connection with parallax measurements

with the HST.
Extragalactic objects such as quasars are in principle ideal reference sources for
parallax determination, thanks to their extremely small parallaxes (e.g. redshift .z ≥
0.01 implies .𝛡 ≤ 0.025 μas). However, in the optical domain quasars are generally
too faint and too scarce, especially in the Galactic zone, to be of much practical use.
The situation is very different at radio wavelengths. Using phase-referencing VLBI
[28, 29] the position difference between a target source and a radio-bright quasar
used as reference can be accurately measured for angular separations up to a few
degrees. Thus, absolute parallaxes for hundreds of Galactic and extragalactic sources
have been measured with accuracies reaching below 10.µas [30].
As an alternative to the methods described above, the parallax degeneracy in
Eq. (3.11) can be lifted by arranging the observations in such a way that the parallax
factors . f 1 and . f 2 are significantly different in a given observation. As seen from
Eq. (3.6), this requires that . r 1 and . r 2 are significantly different, in other words that
the sources are separated by a large angle (. 1 rad). This principle, schematically
illustrated in Fig. 3.2, was adopted for the European Space Agency missions Hippar-
cos and Gaia. Because optical telescopes cannot have a well-corrected field of view
substantially bigger than about .1◦ diameter, these missions employ two telescopes or
viewing directions, separated by a basic angle of.58◦ (Hipparcos) or.106.5◦ (Gaia). A
major challenge of the method is to determine the precise value of the basic angle, or
rather to keep it stable for a sufficiently long time that it can be accurately calibrated
(Sect. 3.5.2).
36 L. Lindegren

3.4 Parallax Measurements Using the Hubble Space

Telescope

The main instruments used for astrometry on-board the Hubble Space Telescope
(HST) are the Fine Guidance Sensors (FGS), the Wide Field and Planetary Camera 2
(WFPC2, 1993–2009), the Advanced Camera for Surveys with the High-Resolution
Channel (ACS/HRC, 2002–2006) and the Wide-Field Channel (ACS/WFC, since
2002), and the Wide Field Camera 3 (WFC3, since 2009). Here I focus on the
determination of stellar parallaxes by means of the FGS and the spatial scanning
technique using the WFC3.

3.4.1 Use of the Fine Guidance Sensors Instrument

The three Fine Guidance Sensors (FGS) of the Hubble Space Telescope are part of the
pointing control system used for guiding the telescope, but they are also a versatile
science instrument for precise relative position measurements and interferometric
studies e.g. of close binaries. Each FGS comprises a two-axis, white-light shearing
interferometer operating in its own ‘pickle’-shaped field of view (FOV) of . 4 ×
17 arcmin.2 located in the perimeter of the HST focal plane (Fig. 3.3a). An aperture
of .5 × 5 arcsec.2 (the instantaneous field of view, IFOV) can be moved to any point
within the FOV by means of a star selector assembly. The wavefront tilt of the light
transmitted by the IFOV is interferometrically sensed in two orthogonal directions by
means of Koesters prisms. For an unresolved object the resulting interferogram (know
as the transfer function) is S-shaped, with an approximately linear central segment of

Fig. 3.3 a Layout of the HST focal plane since 2009, with the Wide Field Camera 3 (WFC3),
Advanced Camera for Surveys (ACS), and the three Fine Guidance Sensors (FGS) as projected on
the sky. FGS1R is the pickle used for astrometry. (Drawing based on [31]) b Calibration of the
optical field angle distortion of the FGS1R pickle was achieved by measuring stars in the M35
calibration field with various pointing offsets and roll angles. (Reproduced from [31])
3 Trigonometric Parallax 37

.±10 mas around the zero-crossing point obtained when the star is accurately centered
in the IFOV.
Two of the three FGS are needed to fix the pointing and roll of the telescope, which
is achieved by locking their IFOV on guide stars with known positions. The third
FGS is then available for scientific observations of sources within its FOV. Initially
FGS3 was the designated ‘astrometer’, but after the refurbishing in 1997 FGS1R is
preferred. As a science instrument the FGS can observe stars in the magnitude range
.V 3 to 17.5. A detailed technical description of the FGS is in [32].
The FGS can be operated in several modes. In position mode (POS) the central
zero-crossing of the fringes is tracked to sample the position of the IFOV, in FOV
coordinates, at a rate of 40 Hz. By sequentially observing the selected objects in the
FOV (typically the target object and several reference stars), while the pointing and
roll of the telescope is held fixed, the positions of the objects in FOV coordinates are
obtained after averaging and correcting for the optical field angle distortion (OFAD)
and other calibrated effects (see below), and for differential stellar aberration. The
POS mode is normally used for parallax work, except if the target star is a binary for
which the resolved structure or orbital motion must be taken into account. In transfer
mode (TRANS) the fringes are sampled over a wider range of IFOV positions, from
which images of binaries and other non-point sources can be reconstructed with a
resolution of about 20 mas for FGS3, and down to about 10 mas for FGS1R [33].
TheOFADincludesthedesigndistortionsoftheRitchey–Chrétientelescope,which
haveanamplitudeofseveralarcsec,aswellassmallercontributionsfromalignmentand
figuring errors of the telescope and FGS optics and encoder errors of the star selectors.
There are also small but significant temporal changes, especially to the scale and low-
order distortion. For astrometric purposes the OFAD needs to be known to within a
few mas. Calibration and monitoring of the long-term stability is achieved by making
positional observations of stars in a calibration field centered on the open cluster M35.
Combining the data collected in a short time from several different telescope pointings
and roll angles (Fig. 3.3b), some 30 OFAD parameters are obtained in a robust least-
squares solution simultaneously with the attitude angles of the different pointings and
the celestial positions of the calibration stars [34]. For FGS1R the resulting OFAD
uncertainty is about 1.5 mas in each coordinate [35], and below 1 mas in the center of
the FOV [15]. Additional calibrations are needed to correct for cross-filter shifts (if a
neutral filter is used, for.V 8) and lateral color effects.
A single POS measurement normally lasts for 1–2 min, during which the position
of the object in FOV coordinates is sampled several thousand times. The adopted
position is the median of the samples, which efficiently reduces photon noise and
HST jitter. For parallax determination some 5–10 reference stars are used. Together
with the science target they are observed multiple times during a visit (lasting up to
an orbit of 90 min), which allows for intra-orbit drift correction. The complete data
set typically comprises about 10 such visits spread over 1.5–2 years, most of them
scheduled around maximum (positive and negative) parallax factor. The astrometric
model used to analyse the data is outlined in [15]. Briefly, the measured FOV coor-
dinates .(x ' , y ' ), corrected as mentioned above, are fitted by a linear model based on
Eq. (3.9),
38 L. Lindegren

ξ0 + μξ (t − tep ) + f ξ 𝛡ξ = Ax ' + By ' + C

. . (3.12)
η0 + μη (t − tep ) + f η 𝛡η = Dx ' + E y ' + F

Here . A through . F are ‘plate constants’ describing the offset, scale, rotation, and
shear of the FOV frame in a given visit; these are common for all the measurements
in the visit, while the astrometric parameters in the left members are common for
all the measurements of a given source. In view of the differential nature of the
measurements (Sect. 3.3), prior information on the astrometric parameters of the
reference stars are added, including uncertainties, to obtain a unique solution; in
particular, spectrophotometric estimates of the reference star parallaxes (Sect. 3.4.3)
are used to obtain the absolute parallax of the target object.
Astrometric results obtained with the FGS and published up to 2016 are reviewed
in [15]. They include absolute parallaxes for 17 standard candles (ten Galactic
Cepheids, five RR Lyr, and two dwarf Cepheids), with a median uncertainty of
0.18 mas (6.5% median relative uncertainty), and 88 absolute parallaxes for various
other kinds of objects (cluster stars, white dwarfs, cataclysmic variables, M dwarf
binaries, metal-poor stars, etc.). Few astrometric programs using the FGS have been
carried out in the last decade following the application of the spatial scanning tech-
nique (Sect. 3.4.2) and the launch of Gaia (Sect. 3.5).

3.4.2 Use of the Spatial Scanning Technique on WFC3

Since its installation in 2009 the Wide Field Camera 3 (WFC3) has been the HST’s
primary science instrument thanks to its high-resolution imaging over a wide wave-
length range [36]. The ultraviolet/visible channel (UVIS, 200–1000 nm) comprises
two 2k.×4k CCDs with a pixel size of .40 × 40 mas.2 ; the infrared channel (IR, 800–
1700 nm) features a 1k.×1k HgCdTe array with a pixel size of .135 × 121 mas.2 .
Although both channels have been used for astrometry, UVIS is obviously the pre-
ferred choice, except for very cool objects [37], thanks to its higher optical resolution
and smaller pixels. For high-precision astrometry and photometry one must calibrate
very carefully not only the geometric distortions (including telescope and camera
optics, filters, and CCD irregularities), but also the effective PSF [38, 39], which is
especially critical for the analysis of undersampled images [40].
The geometric distortion of UVIS is determined by means of a self-calibration
procedure similar to the one illustrated in Fig. 3.3 (right), using multiple displaced
and rotated images of a rich stellar field, in this case the core of .ω Cen. The residual
geometric distortion is about 0.01 pixel, or 0.4 mas [41], which sets the accuracy
floor for the positions of well-exposed images.4 Conventional pointed imaging with
WFC3 (and its predecessor WFCP2, and ACS) has proven immensely useful in many

4Assuming that the geometric distortion is constant during the calibration, and given a sufficient
density of sources, this procedure can determine all kinds of distortions up to a single scale factor,
which can only be fixed by means of external information.
3 Trigonometric Parallax 39

Fig. 3.4 Spatial scanning observation with WFC3. Left: Digital Sky Survey image of a . 3 ×
6 arcmin area centered on the Cepheid SY Aur. The colored frames show the position of the UVIS
field of view at the beginning and end of the spatial scan. Middle: The resulting UVIS scan image.
Right: Part of the scan image magnified by a factor seven. (Reproduced from [42] by permission of
the authors.)

investigations that benefit from the excellent cluster/field star separation that can be
achieved with proper motions over just a few years; this includes population studies,
the internal dynamics of clusters, and kinematic parallaxes. However, except for the
IR observations already mentioned, pointed imaging with the WFC3 has not been
much used for parallax determination, probably because the method would be limited
to nearby, faint stars that are already well explored with ground-based techniques
that can reach a similar precision (cf. Table 3.1).
Spatial scanning is an alternative observing strategy for WFC3/UVIS that over-
comes many of the limitations of pointed observations for high-precision astrometric
and photometric observations of bright stars [16, 42–44]. Originally motivated by
the wish to obtain time-resolved, high signal-to-noise photometric observations of
exo-planet transits [36], the spatial scanning spreads out the light of each star along
a ‘trail’ on the CCD (Fig. 3.4). This is achieved by slewing the telescope in a user-
defined manner, in the simplest case a fixed direction with uniform speed, as in
Fig. 3.4. The maximum slew rate is just under 5 arcsec s.−1 under FGS control, or
7.8 arcsec s.−1 if gyros are used, which gives a smoother but less accurate motion.
For astrometry, the main advantages of spatial scanning are: (1) that many indepen-
dent position measurements are obtained normal to the trail, which reduces random
and (some) systematic noise by a large factor; (2) that much brighter sources can be
observed without saturating the pixels; and (3) that the time-resolved measurements
can be corrected for telescope jitter that would just blur the images in pointed mode.
A simple calculation illustrates the potential gain in precision: At visual wavelengths
the optical resolution of the HST is about .(600 nm)/(2.4 m) 50 mas or 1.25 pixel,
so a trail of length 2000 contains . N ∼ 2000/1.25 = 1600 independent resolution
elements. By averaging over all the measurements normal to the trail, the result-
40 L. Lindegren
√
ing positional uncertainty could be as small as .(0.4 mas)/ N = 10 μas per scan.
In practice the precision obtained for well-sampled bright stars is in the range 20
to 80 .µas per scan [16]. To minimize effects of charge-transfer inefficiency in the
(across-trail) measurement direction, the trails are typically oriented approximately
along the parallel readout direction; they should however make a small angle to the
readout direction in order to average out sub-pixel phase errors. The analysis of the
across-trail measurements includes corrections for the time-dependent differential
aberration, frame-to-frame rotation, and roll-angle variations along the scan, as well
as the known geometric distortion and other effects calibrated from pointed mode
observations, including the effective PSF.
The main disadvantages of spatial scanning are (1) that precise positional mea-
surements are only possible in one direction, that is normal to the trail; and (2) that
some trails will overlap which may be a problem in crowded areas. The first disad-
vantage is minimal for parallax determination, if the measurement direction is chosen
roughly along the major axis of the parallax ellipse (that is, in ecliptic longitude).
To minimize overlap the length and precise direction of the scans need to be chosen
with care. Moreover, to bridge the dynamic range gap between a bright target source
and the faint fields stars used as reference objects, it may be necessary to combine
measurements from both shallow and deep exposures taken at the same epoch, by a
suitable choice of filters and scan speeds. For additional considerations and details
on the technique and data analysis, see [16, 42, 43].
Parallaxes obtained with this technique have yielded uncertainties in the range 29
to 74 .µas for bright Cepheids (.V 10) [16]; this includes the systematic uncertainty
of the correction from relative to absolute parallax, which in this case is only about
10–20% of the total uncertainty. The resulting relative uncertainty in distance is
7–17% for the individual Cepheids.
Spatial scanning was also applied to the determination of the parallax of the
globular cluster NGC 6397 [44]. The mean parallax of 39 cluster stars (.V 13–19)
was obtained with a statistical uncertainty of 13 .µas, to which should be added a
systematic uncertainty of 18 .µas for the correction from relative to absolute. The
relative uncertainty of the trigonometric distance (.∼2.4 kpc) is therefore about 5%.

3.4.3 Reference Stars and Correction to Absolute

As discussed in Sect. 3.3, parallax measurements made within a small field are dif-
ferential, and the conversion from relative to absolute parallaxes requires knowledge
of the mean parallax of the reference sources. For the FGS and WFC3 observations
described above the spectrophotometric method is used for the correction to absolute.
This aims to estimate the absolute magnitudes (. MV , say) and interstellar extinctions
(. A V ) of the individual reference sources by means of spectra and/or multi-wavelength
photometry, from which their spectrophotometric parallaxes are obtained as

.𝛡sp = (100 mas) × 10−0.2μ0 , (3.13)

3 Trigonometric Parallax 41

where .μ0 = V − MV − A V is the absorption-corrected distance modulus (which

should be independent of the wavelength band used, e.g. .V ). The combination of
spectral type (or .Teff ), luminosity class (or .log g), and colors makes it possible to
estimate . MV and . A V . Spectrophotometrically, it is difficult to estimate .μ0 with a
precision much better than about 0.3 mag. Differentiation of Eq. (3.13) gives
| |
| ∆𝛡sp |
| |
.
| 𝛡 | = 0.2(ln 10) |∆μ0 | 0.46 |∆μ0 | , (3.14)
sp

so an error of 0.3 mag in distance modulus corresponds to a relative error in parallax

(and distance) of .0.46 × 0.3 14%. Using reference sources that are expected to
have small parallaxes, such as distant giants, thus makes the uncertainty of their
parallaxes correspondingly smaller, and the correction to absolute more accurate.
The luminosity class is generally the most problematic parameter to estimate,
where a dwarf/giant misclassification can easily result in an error of 5 mag in the
distance modulus, or a factor 10 in the calculated spectrophotometric parallax. To
minimize the risk of such error, various strategies have been used. A classical device to
discriminate between luminosity classes is the reduced proper motion [45], utilizing
the fact that a nearby dwarf tends to have higher proper motion than a distant giant of
the same apparent magnitude and color. Furthermore, the handle on both extinction
and luminosity class is improved by using multiband photometry that extends into
the infrared (e.g. from 2MASS and WISE), and by fitting synthetic colors to such
data. Finally, increasing the number of reference sources per parallax field improves
not only the precision of the correction to absolute but also its robustness, where for
example a misclassified dwarf stands out more clearly as an outlier.
Recent programs have adopted a comprehensive modelling approach in which the
astrometric parameters of the target and reference sources are estimated, together with
plate constants and other local unknowns, in a global multisource, multiparameter
solution. The need to correct from relative to absolute parallax is in practice elimi-
nated by introducing the spectrophotometric parallaxes of the reference sources as
prior information in the solution [16, 44]. The approach is very flexible and allows
many other kinds of prior information to be incorporated in the solution, for example
on the proper motions of the reference sources from a kinematic Galaxy model, and
extinction maps. To avoid compromising the statistical interpretation of the resulting
parallax (Sect. 3.6), it is important that no prior is assumed on the parallax of the
target source.
A useful feature of the global modelling approach is that purely astrometric esti-
mates of the parallaxes of the reference sources can be obtained by treating each
star, in turn, as an additional target source. This is achieved in a series of separate
solutions, one for each reference source, in which its prior information is ignored.
The full solution is then cross-validated by comparing the purely spectrophotometric
parallaxes against the corresponding, independent astrometric values [16]. A sim-
ulation of the global solution for the bright Cepheid SS CMa, using 1000 random
fields drawn from the Besançon Galaxy Model [46], indicates that the correction to
absolute in that case would be determined with a typical precision of 9.µas if.∼20 ref-
42 L. Lindegren

erence sources are used, each having an uncertainty of 15% in its spectrophotometric
distance (.∼0.3 mag in distance modulus) [43].
Parallax determinations using small-field telescopes in space and on ground will
continue to have important applications after Gaia, for example in the infrared, for
very bright, faint, or complex sources, and where an even higher precision is needed.
However, the availability of accurate Gaia astrometry for virtually all potentially
useful reference sources is bound to change the way such observations are analyzed.
The most obvious implication is for the correction from relative to absolute, which
becomes very straightforward if the Gaia parallaxes can be adopted for the reference
sources. But this is not necessarily the best strategy. As an illustrative example let
us consider the 99 reference sources used for the parallaxes of bright Cepheids
observed with the spatial scanning technique [16]. The median uncertainty of their
spectrophotometric parallaxes is 30 .µas. All of them have parallaxes in Gaia DR3,
with a median formal uncertainty of 28 .µas. However, most of the statistical weight
in the correction to absolute comes from a small number of distant giants having very
small uncertainties in their spectrohotometric parallaxes. A more relevant quantity
could therefore be the uncertainty corresponding to the mean statistical weight of the
reference stars, or .⟨σ𝛡−2 ⟩−1/2 . This is 14 .µas for the spectrohotometric parallaxes of
the same stars, and 27 .µas for their Gaia DR3 parallaxes. Thus, even the final (DR5)
Gaia parallaxes, which are expected to be almost a factor two better than in DR3
(Sect. 3.5.4), will not make the mean weight of these particular reference sources
much higher than already achieved with the spectrophotometry. In a more indirect
way, Gaia could still contribute a lot. It should allow many more reference sources
to be used, with exquisite dwarf/giant separation and screening against problematic
sources, and perhaps also brighter (more nearby) sources than is possible with the
spectrophotometric method. Furthermore, the geometric calibration of the images
is likely to be improved through the incorporation of accurate positions and proper
motions from Gaia, from which the target source will also benefit.

3.5 Parallax Measurements with Gaia

The Gaia satellite was launched on 19 December 2013 and arrived three weeks
later to its destination near the Sun–Earth . L 2 Lagrangian point, 1.5 million km from
the Earth. Science operations began on 25 July 2014 and are currently (May 2023)
expected to end in 2025. Three releases of science data have been made to date.
The third data release was made in two stages: in December 2020 the early release
EDR3 of mainly astrometric and photometric data was made, and in June 2022 the
full DR3. Two more releases (DR4 and DR5) are expected in the coming years. The
nominal precision of parallaxes in Gaia (E)DR3 is typically around 20.µas for bright
stars (.G 6 to 14), 70.µas at .G 17, and 500.µas at .G 20 (Fig. 3.5). See [47]
for an overview of the mission, [17, 48] for summaries of the contents of EDR3 and
DR3, and [49] for some of the early science results.
3 Trigonometric Parallax 43

Fig. 3.5 Formal parallax uncertainty in Gaia (E) DR3 versus.G magnitude. Colors show the density
of points on a logarithmic scale, with a random subset plotted for .G > 11.5. The white curve is the
median uncertainty. The black curves show extrapolated median formal errors in DR4 and DR5. The
letters at the top show the ranges for (A) sources too bright for the on-board detection; (B) partially
saturated images; (C) calibration-limited performance; (D) photon-noise limited performance; (E)
too faint for reliable detection

3.5.1 Measurement Principle

Gaia is optimised for accurate angular measurements in the along-scan (AL) direc-
tion, while less strict requirements apply in the across-scan (AC) direction (Fig. 3.6).
This inequality is reflected in many features of the instrument and its operation.
The rectangular shape of the primary mirrors, .1.45 × 0.5 m2 (AL.×AC), results in
a point-spread function (PSF) that has three times higher angular resolution in the
AL direction than AC (Fig. 3.7a). The CCD pixels are correspondingly dimensioned,
.59 × 177 mas . The rotation of the satellite makes the source images move over the
2

CCDs at a very nearly constant AL rate of .η̇ = −60 arcsec s−1 . This is accurately
compensated by reading out the CCDs in drift-scanning mode at the same rate, thus
preserving the full optical resolution AL. The much smaller AC scan rate .ζ̇ varies
sinusoidally along the scanning path in the range .±0.173 arcsec s−1 , thus widening
the PSF in the AC direction by up to 1 pixel. The AC motion is not compensated
but analytically modelled as part of the PSF calibration (Sect. 3.5.2). Only a small
pixel window centred on each detected source is sampled on-board and transmitted
to the ground. To further reduce the telemetry rate and read-out noise, the full two-
dimensional window is only transmitted for bright (.G 13 mag) sources, while for
fainter it is marginalized AC, creating a one-dimensional AL image.
The accurate time .tobs when the image centroid passes over a fiducial ‘observation
line’, nominally at the centre of the active CCD area, constitutes the elementary
44 L. Lindegren

Fig. 3.6 Schematic picture of the 62 astrometric CCDs (small rectangles) projected through the
Gaia telescopes onto the sky, creating two fields of view (preceding, PFoV, and following, FFoV)
separated by the basic angle. Field angles .(η, ζ ) are defined separately in each field. The red curve
shows the path of a source crossing first the PFoV and then the FFoV. Along- and across-scan
directions (AL, AC) describe the apparent motions of source images on the CCDs, rather than the
actual scanning motion of the instrument. The slow precession of the spin axis (Fig. 3.7b) generates
a small positive or negative drift of the images in the AC direction. The picture shows the view from
‘outside’ the celestial sphere and is not drawn to scale

Fig. 3.7 a Example of the point spread function (PSF) of Gaia. The resolution is about 0.1 arcsec
in the along-scan (AL) direction and 0.3 arcsec across-scan (AC). b Gaia’s scanning law over four
days. During this time the spin axis (. z) traces an arc of .16.6◦ , while each of the viewing directions
(preceding P, following F) completes 16 revolutions. The . z axis makes a constant angle of .45◦ to
the Sun

astrometric measurement in the AL direction [50]. Up to nine such AL measurements

are obtained during the 43 s it takes the image to cross the astrometric field of view
(Fig. 3.6). For bright sources, less accurate measurements are also obtained in the AC
direction at the same times; these are the centroid coordinates expressed in AC pixels
(.m) and fractions thereof. Under the most favourable conditions (which occur for
.G 12) the precision of a single measurement is about 60 .µas AL and 300 .µas AC.
On average about 8000 AL and 40 AC observations are collected every second, but
3 Trigonometric Parallax 45

the rates vary considerably depending on the local density of sources on the sky. To
date more than .2 × 1012 astrometric measurements have been collected. This does
not include the skymapper (SM) observations, which are essential for the operation
of the instrument but not directly used for the astrometry, or the photometric (BP,
RP) and radial-velocity (RVS) observations using dedicated CCDs.

3.5.2 Astrometric Processing

A schematic overview of the processing chain for the main astrometric results in Gaia
(E)DR3 is shown in Fig. 3.8. I refer to the published documentation, in particular
[51] and references therein, for detailed descriptions of the various steps; here only
certain aspects of relevance for the parallaxes are discussed. The separate chains for
photometric and spectroscopic processing are not considered here, nor the special
treatment of non-single sources.
The main processing chain is optimised for sources that can be treated as point
objects (angular size . 50 mas; Fig. 3.7a) moving according to the standard astro-
metric model of Sect. 3.2. As indicated in the left part of Fig. 3.8, after an initial
analysis of the down-linked data, where all non-spurious detections are matched to
some particular source, the centroid positions AL (.tobs ) and AC (.m) and other image
parameters (including the flux for photometry) are estimated by fitting an image
profile to the CCD samples. Normally this uses a two-dimensional PSF for bright
sources (.G 13 mag) and a one-dimensional line-spread function (LSF) for fainter
sources, where pixels are binned on-board in the AC direction. Prior to the image
parameter determination the LSF and PSF were calibrated, using a subset of well-
behaved observations, as functions of the field of view (P or F), position in the field,
time, window class, gate, and color (effective wavenumber5 .νeff ).
The LSF/PSF models are linear combinations of precomputed basis functions
designed to represent the effective LSF or PSF for a range of colors and opti-
cal aberrations. The LSF/PSF calibration takes into account color-dependent image
shifts (‘chromaticity’) caused by asymmetric optical aberrations and the wavelength-
dependent diffraction. Chromaticity is therefore largely eliminated already in the
image parameter determination stage by fitting the LSF or PSF of the appropriate
color. The color information normally comes from the on-board photometric instru-
ment (BP and RP), but for some (mainly faint) sources no reliable color was avail-
able at the time of the image parameter determination, in which case the LSF/PSF
calibration for the default color .νeff = 1.43 μm−1 was used. In the subsequent astro-
metric processing such sources obtained a so-called six-parameter solution, in which
a‘pseudocolor’ was introduced as the sixth astrometric parameter (Sect. 2.3 in [51]).
The pseudocolor is therefore an estimate of .νeff obtained by utilizing the tiny spectral
dispersion of the stellar images produced by the chromaticity.

5 A useful analytical approximation relating the effective wavenumber to the Gaia color index is
.νeff 1.76 − (1.61/π )atan[0.531(G BP − G RP )] μm−1 [51].
46 L. Lindegren

Fig. 3.8 Simplified block diagram for the processing of astrometric data from Gaia

It can be noted that the LSP/PSF calibration for EDR3 was restricted to the
well-populated color interval .1.24 < νeff < 1.72 μm−1 (.3.0 G BP − G RP 0.14).
Beyond that range the calibration for the nearest color was used, leading to parallaxes
that are more uncertain.
In binaries with a projected AL separation exceeding 2–3 pixels (0.12–0.18 arcsec)
the components may produce separate peaks in the CCD samples, in which case the
centroiding algorithm normally picks up the brighter peak. For smaller projected
separations the centroid position is a complex function of the separation and flux
ratio of the components, scanning geometry, and the LSF (see [52] for a toy model),
which may create spurious signals in the astrometric results [53].
The right part of Fig. 3.8 corresponds to the processing known as the astromet-
ric global iterative solution (AGIS). This aims to combine all the measured cen-
troid positions into a globally consistent system of astrometric parameters for all the
sources [50]. The precise AL measurements .tobs are the main input to AGIS. For a
given observation .tobs depends not only on the astrometric parameters of the relevant
source, but also on a number of ‘nuisance parameters’ that describe the behaviour
of the instrument; these are the attitude, calibration, and global parameters. The
attitude parameters describe the accurate celestial pointing of the instrument as a
continuous function of time. The calibration parameters map the location of the fidu-
cial observation lines of each CCD in terms of the field angles .(η, ζ ). The global
parameters describe additional effects that are coherent on larger scales than the indi-
vidual CCDs. Following the self-calibration principle, the nuisance parameters are
estimated together with the astrometric parameters in a single least-squares solution,
using the regular observations. For practical reasons only a subset of well-behaved,
effectively single ‘primary’ sources is used to estimate the nuisance parameters. For
Gaia EDR3 some 6.5 billion CCD observations of 14 million primary sources were
used to determine 70 million source parameters, 11 million attitude parameters, and
3 million instrument (calibration and global) parameters [51].
A critical part of the data analysis is the definition of a good instrument model. This
must be flexible and detailed enough to represent the actual instrument to an accuracy
compatible with the best centroiding precision (. 60 μas), yet robust enough to
be fully constrained by the available observations. In practice the model develops
through a lengthy process of trial and error, where residual patterns in preliminary
3 Trigonometric Parallax 47

solutions may hint towards possible model improvements. A main challenge of the
approach is that certain instrumental effects may be fully or partly absorbed by the
astrometric parameters, in which case they do not show up in the residuals, or only
partially and perhaps in a very distorted form. Simulated solutions in an AGIS-
like framework are an indispensable tool for investigating possible degeneracies in
various models.
In this context basic-angle variations (BAV) are of special interest. It has been
known since the early phases of the Hipparcos project that basic-angle stability on
time scales shorter than the rotation period (6 h for Gaia) is fundamental to the
global measurement principle. In particular, a BAV proportional to .cos Ω, where
.Ω is the rotation phase of the instrument relative to the Sun (see Fig. 1 in [54]), is
tightly correlated with a global zero-point error of the measured absolute parallaxes.
Very strict engineering requirements were consequently put on short-term BAV in
Gaia. Furthermore, an internal metrology device, known as the basic-angle monitor
(BAM), was introduced to measure such variations at the sub-.μas level. During the
commissioning of Gaia it was found that the basic angle has a strong dependence on
.Ω with an amplitude of about 1 mas, two orders of magnitude greater than pre-launch
predictions [47]. The variations measured by the BAM are taken into account in the
processing of the astrometric data, but this does not completely remove the BAV
as sensed by the astrometric detectors. One possible reason for this could be that
the separate CCDs used by the BAM are located well outside the astrometric field:
because the optical field distortions also exhibit periodic variations, the BAV sensed
by the BAM will be slightly different from the effective variation in the astrometric
field. For EDR3, residual BAV are taken into account in the global model by means
of a Fourier series in .Ω, where the coefficients are slowly varying functions of
time (Sect. 3.4.3 of [51]). Although this model includes a parameter describing the
dependence on .cos Ω, its determination is very delicate owing to the strong coupling
with a global parallax offset (correlation coefficient . 0.99992), and it was not
possible to completely remove such an offset (Sect. 3.5.3).
Another instrument effect that could bias the absolute parallax measurement is
a possible dependence of the AL centroid position on the AC scan rate (.ζ̇ ; see
Appendix-B in [51]). Such an effect could be produced by CTI in the CCDs for a PSF
that is asymmetric in the AC direction. Because the CTI depends on the magnitude
of the source, the overall result of the unmodelled effect could be a magnitude-
dependent parallax bias. In the nominal scanning there is a strong coupling between
.ζ̇ and the AL parallax factor . f η (correlation coefficient . +0.985), which prevented
the effect to be calibrated for EDR3. During one year from 2019.5 the precession of
the scanning law was however reversed, leading to an equally strong anti-correlation
between .ζ̇ and . f η , which should enable the effect to be disentangled in DR4 or DR5.
In addition to the dependencies introduced already in the LSF/PSF modelling
mentioned above, the astrometric instrument model for EDR3 contains functions
depending on magnitude (only for the AC calibration), saturation, subpixel phase,
time since the last charge injection, and AC rate (only quadratic in .ζ̇ , to avoid degen-
eracy with parallax). See [51] for details.
48 L. Lindegren

3.5.3 Known Parallax Systematic in Gaia (E)DR3

Had the large BAV been ignored in the astrometric solution, the result would be a very
large (. 800 µas) overall offset in the parallaxes, with considerable (. 100 µas)
variations over the sky. Thus, corrections for the BAV, as measured by the BAM,
were introduced already for Gaia DR1, and starting with DR2 a global model rep-
resenting spin-related distortions (including BAV) was fitted in AGIS as part of the
self-calibration process. Nevertheless, residual systematics of several tens of .μas
were clearly evident [55]. It was however decided to publish the parallaxes without
applying any ad hoc corrections, and merely alert users to the known or suspected
biases. The same policy was followed for EDR3; but an accompanying paper, here
called L21 [56], examined the issue in some detail and gave a provisional recipe for
its correction as a function of magnitude (.G), color (.νeff ), and ecliptic latitude (.β).
L21 was based on an analysis of the EDR3 results for 1.1 million quasars (.G 14),
2.4 million probable LMC members (.G 19), and 121 000 wide binaries with pri-
mary .G < 14 and secondary .G < 18. Importantly, the LMC sources and binaries
were only used differentially, i.e. for mapping systematic differences depending on
magnitude and color. The estimated parallax bias is given as a function . Z (G, νeff , β),
such the corrected parallax is6

𝛡L21 = 𝛡EDR3 − Z (G, νeff , β).

. (3.15)

Since the publication of Gaia EDR3 and L21 in December 2020, a considerable
number of papers have appeared in which other methods and other kinds of objects
are explored for further quantification of the parallax bias. These include the use of
classical Cepheids in the field and in clusters [57–59], RR Lyr variables [60, 61],
asteroseismic distances [62, 63], spectrophotometric distances [64, 65], eclipsing
binaries [66–68], orbital parallaxes [69], and open and globular clusters [70–72].
Similar to L21, quasars and wide binaries have also been used [73], employing a
larger sample of resolved binaries constructed from EDR3 [74].
In general the cited papers agree that the L21 recipe removes most of the parallax
bias in (E)DR3, at least for .G > 13 or 14, but in many cases significant differences
are found. I define the residual parallax bias in (E)DR3 relative to L21 as

.∆Z = 𝛡L21 − 𝛡other , (3.16)

where .𝛡other is the parallax obtained by some other method; thus . Z + ∆Z is the
total bias in (E)DR3.7 Fig. 3.9 is an attempt to summarize the finding of the various
determinations of .∆Z to date. Several authors conclude that .∆Z = +10 to .+20 μas
for .G < 11 or 12, i.e. that L21 overcorrects the bias for the bright stars [57, 59–62,

6 L21 provides two bias functions: . Z 5 for five-parameter solutions with known .νeff , and . Z 6 for
six-parameters solutions where the pseudocolor .ν̂eff was used instead. The subsequent discussion
only concerns . Z 5 , which is applicable to most sources brighter than .G = 18.
7 Some authors define residual parallax offset in the opposite sense, i.e. as .−∆Z [57–59].
3 Trigonometric Parallax 49

Fig. 3.9 CMD showing the residual parallax bias .∆Z (in .μas relative to L21) obtained in various
studies. The result of each study, with its uncertainty when available, is plotted at the approximate
mean color and magnitude of the sample, color-coded according to the method used. The background
shows the distribution of the three main types of sources used to derive . Z in L21, namely quasars
(orange dots), binaries (gray lines connecting the components), and stars in the LMC (green contours
representing the smoothed density in steps of 1 dex). In L21 the binaries and LMC stars were used
to map the variations of . Z with magnitude, color, and position; its absolute value is anchored in the
quasar parallaxes

71, 75]. However, as indicated in the figure, perhaps this only applies to stars of
intermediate or red colors, while the reverse may be true for hotter stars (.G BP −
G RP 0.5). The variation according to position is clearly more complex than just
a function of ecliptic latitude; thus results may differ significantly depending on the
area investigated [63, 65, 71, 73]. Overall, residual systematics are found to be at
the level of about 10 .µas RMS.
Another kind of systematics of the parallaxes in Gaia (E)DR3 is the underesti-
mation of the uncertainty, as given by the formal standard error .σ𝛡 , and the spatially
correlated errors on angular scales of several degrees. The underestimation may
require both a multiplicative factor (.k 1.05) and a quadratically added contribu-
tion (.s 10 μas) representing the overall accuracy floor. The statistical correlation
of √
the errors implies that the uncertainty of the sample mean does not decrease as
.1/ n when averaging over .n sources, if they are all within a small area of the sky
(as in a cluster); both effects are crudely described by
√
σ true =
. 𝛡 (kσ𝛡 )2 /n + s 2 , (3.17)
50 L. Lindegren

where .k and .s may depend on factors such as the magnitude, local star density, or
astrometric goodness-of-fit [70, 71, 74].

3.5.4 Expected Improvements in Gaia DR4 and DR5

The astrometric solution for Gaia DR3 is based on data collected during 33 months
of observations. For DR4 the solution will use 63 months of data collected until
early 2020, and DR5 may add another five years. The formal uncertainty of parallax
determinations could therefore improve roughly by a factor 0.72 for DR4 and 0.52
for DR5, as suggested by the black curves in Fig. 3.5. The real improvement in terms
of accuracy (i.e. including systematics) depends on the extent to which the various
modelling errors can be reduced in the data processing (Fig. 3.8), which is difficult
to predict.
Most modelling errors are benign in the sense that they behave almost like uncor-
related noise; their impact on the astrometry is therefore qualitatively similar to that
of the photon noise and can be expected to improve by the same factors as mentioned
above, even if the models are not improved. For many sources there is a potential
for significantly improved calibrations, e.g. in the non-saturated bright range (C in
Fig. 3.5) and for sources of extreme colors. Similarly, the non-single star processing
[76], using more realistic source models than the standard one e.g. for astromet-
ric binaries, will normally result in improved parallaxes [77]. Furthermore, epoch
astrometry will be made available for all sources in DR4 and DR5, enabling users to
fit their own source models to the data.
The parallax bias described in Sect. 3.5.3 is a rather different story. This bias is
at least partly related to near-degeneracies with certain instrumental effects, namely
the .cos Ω variation of the basic angle and centroid shifts depending on the AC rate.
Although its variations with magnitude, color, and position are likely to become
smaller in DR4 and DR5, thanks to the general improvement of calibration models,
it may not be possible to reduce the bias to insignificant levels. Independent methods
to assess the parallax bias will therefore be needed also for the future releases.

3.6 Statistical Meaning and Use of Parallax

The linearity of the standard astrometric model, Eq. (3.5), has important practi-
cal consequences for the statistical interpretation of the astrometric parameters. Let
'
. y = [∆α∗ , ∆δ, 𝛡, μα∗ , μδ ] be the astrometric parameters obtained for a particular
source, where the position is expressed differentially to some fixed reference point.
Furthermore, let . y0 be the corresponding true parameter vector—assuming that such
values exist, in other words that the standard model is correct. Thanks to the linear
estimation method used (essentially a weighted least-squares solution) the resulting
astrometry is also unbiased, .E( y) = y0 , provided that the standard model holds and
3 Trigonometric Parallax 51

the individual observations are unbiased. As we have seen, both of these assump-
tions can be questioned, and ways to deal with possible biases are therefore needed
as already discussed (concerning biases in the Gaia DR3 proper motions, see [78]).
Assuming unbiasedness, another consequence of the linearity is that the error
distribution tends towards the multivariate normal (Gaussian), . y ∼ N ( y0 , C), where
. C is the covariance matrix of . y. That is, the probability density function is

[ ]
−m/2 −1/2 1 ' −1
. p( y| y0 ) = (2π ) det(C) exp − ( y − y0 ) C ( y − y0 ) , (3.18)
2

for .m = 5 (the dimensionality of . y). Specifically for the parallax,

[ ]
1 (𝛡 − 𝛡0 )2
. p(𝛡 |𝛡0 ) = √ exp − , (3.19)
σ𝛡 2π 2σ𝛡2

where .σ𝛡2 is the diagonal element of .C corresponding to the parallax.

In these equations. y is formally regarded as a random variable and. p is the limiting
distribution of . y when the experiment is repeated many times for fixed . y0 and .C.
In practice we only have the result from a single experiment, namely . y, and the true
parameter vector . y0 is unknown. Regarded as a function of . y0 , for the given . y, this
is known as the likelihood function: .L( y0 | y) = p( y| y0 ).
Assuming that the least-squares estimation is done with the appropriate weight-
ing of the observations, taking into account possible correlations among the data, we
obtain .C as part of the estimation process. In reality various approximations and sim-
plifications are usually made. In the Gaia data analysis, for example, the correlations
among the eight or nine CCD observations in a single transit of the astrometric field
(Fig. 3.6) are currently neglected, and no marginalisation is made over the nuisance
parameters. This leads to an underestimation of .C and .σ𝛡 , and empirical correction
factors may be required, as in Eq. (3.17).
In spite of the several caveats given above, the important fact remains that the
probability density in Eqs. (3.18) or (3.19) (i.e. the likelihood function) is generally
known to a good approximation. This provides a solid foundation for statistical
inference in most astrophysical problems where astrometric data play a major role,
for example using Bayesian inference. A simple but non-trivial example, extensively
discussed in the literature, is the estimation of distance.d = A/𝛡0 from the measured
parallax value .𝛡 and its uncertainty .σ𝛡 [79–82]. Additional considerations relevant
for the interpretation of astrometric data are discussed in [83].
The published astrometric parameters . y, together with their covariance matrix .C,
should thus be regarded as a concise specification of the likelihood function of the
true parameters . y0 , using Eq. (3.18). Under the assumptions mentioned above they
provide sufficient statistics of the observations with respect to the adopted source
model [84], as long as that model is linear in its parameters.
52 L. Lindegren

3.7 Conclusion and Future Prospects of the Parallax

Method

The parallax method is the most direct way to obtain information about stellar dis-
tances. Because it is based on simple geometry, with minimal assumptions concerning
the nature of the sources, it is sometimes claimed to be model-independent. But as we
have seen, the modelling of parallax observations is far from trivial. A basic reason is
that the parallax is very small even for the nearest stars, and that the requirements in
terms of measurement accuracy increase in proportion to the distance covered. In the
last decades the barrier to classical optical astrometry set by the atmosphere has been
overcome by the use of VLBI at radio frequencies, adaptive optics at near-infrared
wavelengths, and space techniques in the visual domain. All of these techniques now
produce parallax determinations at the microarcsecond level. For the future, at least
two space missions dedicated to astrometric surveys in the near-infrared are currently
under study, JASMINE [85] and GaiaNIR [86].
Thanks to Gaia we now have trigonometric parallaxes available for nearly two
billion sources brighter than .G 21 mag, and the next releases of Gaia data will
bring additional improvements and new kinds of data, such as the epoch astrometry.
However, results for the brightest stars (.G 6 mag) suffer from saturation of the
detectors, and the fixed observation schedule implied by the scanning law of Gaia
limits the integration time and hence the precision for faint objects. Bigger telescopes
in space, and on ground using adaptive optics, will have important advantages over
Gaia in crowded areas. Relative parallax determinations using HST and its successors
(James Webb Space Telecope, Nancy Grace Roman Space Telecope), as well as with
large ground-based facilities, will therefore remain relevant for specific targets where
Gaia cannot deliver sufficient accuracy or resolution. For the correction from relative
to absolute parallax, the statistical and spectrophotometric methods can be vastly
improved by incorporating Gaia information on the reference sources.

Acknowledgements I wish to thank Stefano Casertano for many perceptive comments on the
manuscript, and Adam Riess for important clarifications.

References

1. J.D. Fernie, JRASC 69, 153 (1975)

2. J.D. Fernie, JRASC 69, 222 (1975)
3. M.J. Reid, K.M. Menten, Astronomische Nachrichten 341, 860 (2020)
4. L. Lindegren, Direct distance determination using parallax: techniques, promises and limita-
tions, in IAU Symposium 289, Advancing the Physics of Cosmic Distances, ed. by R. de Grijs,
p. 52 (2013)
5. M.J. Reid, M. Honma, ARA&A 52, 339 (2014)
6. F.W. Bessel, Astronomische Nachrichten 16, 65 (1838)
7. J.A.C. Oudemans, Astronomische Nachrichten 122, 193 (1889)
3 Trigonometric Parallax 53

8. F. Schlesinger, M. Palmer, A. Pond, Y.U. Observatory, in General Catalogue of Stellar Paral-

laxes (Yale University Observatory, 1924)
9. F. Schlesinger, L. Jenkins, General Catalogue of Stellar Parallaxes: Compiled at Yale Univer-
sity Observatory (Yale University Observatory, 1935)
10. L.F. Jenkins, General Catalogue of Trigonometric Stellar Parallaxes (Yale University Obser-
vatory, 1952)
11. W.F. van Altena, J.T. Lee, E.D. Hoffleit, The General Catalogue of Trigonometric [Stellar]
Parallaxes (Yale University Observatory, 1995)
12. M.A.C. Perryman et al., A&A 323, L49 (1997)
13. H.C. Harris et al., Progress in Parallaxes at USNO, in Astrometry in the Age of the Next
Generation of Large Telescopes, ed. by P.K. Seidelmann, A.K.B. Monet. Astronomical Society
of the Pacific Conference Series, vol. 338, p. 122 (2005)
14. F. van Leeuwen, A&A 474, 653 (2007)
15. G.F. Benedict, B.E. McArthur, E.P. Nelan, T.E. Harrison, PASP 129, 012001 (2017)
16. A.G. Riess et al., ApJ 855, 136 (2018)
17. Gaia Collaboration, A.G.A. Brown, A. Vallenari, T. Prusti, A&A 649, A1 (2021)
18. M. Soffel et al., AJ 126, 2687 (2003)
19. S.A. Klioner, AJ 125, 1580 (2003)
20. Z. Penoyre, V. Belokurov, N. Wyn Evans, A. Everall, S.E. Koposov, MNRAS 495, 321 (2020)
21. L. Lindegren, A&A 633, A1 (2020)
22. K. Rosquist, ApJ 331, 648 (1988)
23. J.H.J. de Bruijne, A.C. Eilers, A&A 546, A61 (2012)
24. C.A. Murray, Vectorial Astrometry (Adam Hilger, 1983)
25. W.F. van Altena, AJ 79, 826 (1974)
26. W.C. Jao et al., AJ 129, 1954 (2005)
27. G.F. Benedict et al., AJ 133, 1810 (2007)
28. M.J. Rioja, R. Dodson, A&Ar 28, 6 (2020)
29. M.J. Reid, PASP 134, 123001 (2022)
30. M.J. Reid et al., ApJ 885, 131 (2019)
31. E.P. Nelan, FGS Instrument Handbook v. 28.0 (2021). https://hst-docs.stsci.edu/fgsihb
32. A. Bradley, L. Abramowicz-Reed, D. Story, G. Benedict, W. Jefferys, PASP 103, 317 (1991)
33. L. Dressel, M. Marinelli, WFC3 Instrument Handbook for Cycle 31 v. 15.0 (2023). https://hst-
docs.stsci.edu/wfc3ihb
34. W. Jefferys et al., Optical field angle distortion calibration of FGS3, in Calibrating Hubble
Space Telescope, ed. by J.C. Blades, S.J. Osmer, p. 353 (1994)
35. B.E. McArthur, G.F. Benedict, W.J. Jefferys, E. Nelan, The optical field angle distortion cal-
ibration of HST fine guidance sensors 1R and 3, in The 2005 HST Calibration Workshop:
Hubble After the Transition to Two-Gyro Mode, ed. by A.M. Koekemoer, P. Goudfrooij, L.L.
Dressel, p. 396 (2006)
36. J.W. MacKenty, Performance and calibration of the HST Wide Field Camera 3, in Space
Telescopes and Instrumentation 2012: Optical, Infrared, and Millimeter Wave, ed. by M.C.
Clampin, G.G. Fazio, H.A. MacEwen, J. Oschmann, M. Jacobus. Society of Photo-Optical
Instrumentation Engineers (SPIE) Conference Series, vol. 8442, p. 84421V (2012)
37. L.R. Bedin, C. Fontanive, MNRAS 481, 5339 (2018)
38. J. Anderson, I.R. King, PASP 112, 1360 (2000)
39. J. Anderson, L.R. Bedin, MNRAS 470, 948 (2017)
40. J. Anderson, Analyzing poorly sampled images: HST imaging astrometry, in Astrometry for
Astrophysics, ed. by W.F. van Altena, p. 257 (Cambridge University Press, 2012)
41. A. Bellini, J. Anderson, L.R. Bedin, PASP 123, 622 (2011)
42. A.G. Riess, S. Casertano, J. Anderson, J. MacKenty, A.V. Filippenko, ApJ 785, 161 (2014)
43. S. Casertano et al., ApJ 825, 11 (2016)
44. T.M. Brown et al., ApJL 856, L6 (2018)
45. E.M. Jones, ApJ 173, 671 (1972)
46. A.C. Robin, C. Reylé, S. Derrière, S. Picaud, A&A 409, 523 (2003)
54 L. Lindegren

47. Gaia Collaboration, T. Prusti, J.H.J. de Bruijne, A.G.A. Brown, A&A 595, A1 (2016)
48. Gaia Collaboration, A. Vallenari, A.G.A. Brown, T. Prusti (2022), arXiv e-prints
arXiv:2208.00211
49. A.G.A. Brown, ARA&A 59, 59 (2021)
50. L. Lindegren et al., A&A 538, A78 (2012)
51. L. Lindegren et al., A&A 649, A2 (2021)
52. L. Lindegren, Expected astrometric properties of binaries in (E)DR3, Technical note LL-136,
Lund Observatory (2022). https://dms.cosmos.esa.int/COSMOS/doc_fetch.php?id=1566327
53. B. Holl et al. (2022), arXiv e-prints : arXiv:2212.11971
54. A.G. Butkevich, S.A. Klioner, L. Lindegren, D. Hobbs, F. van Leeuwen, A&A 603, A45 (2017)
55. F. Arenou et al., A&A 616, A17 (2018)
56. L. Lindegren et al., A&A 649, A4 (2021)
57. A.G. Riess et al., ApJL 908, L6 (2021)
58. A.G. Riess et al., ApJ 938, 36 (2022)
59. R. Molinaro et al., MNRAS 520, 4154 (2023)
60. A. Bhardwaj et al., ApJ 909, 200 (2021)
61. C.K. Gilligan et al., MNRAS 503, 4719 (2021)
62. J.C. Zinn, AJ 161, 214 (2021)
63. S. Khan et al. (2023), arXiv e-prints: arXiv:2304.07158
64. Y. Huang, H. Yuan, T.C. Beers, H. Zhang, ApJL 910, L5 (2021)
65. C. Wang, H. Yuan, Y. Huang, AJ 163, 149 (2022)
66. K.G. Stassun, G. Torres, ApJL 907, L33 (2021)
67. F. Ren et al., ApJL 911, L20 (2021)
68. F. Ren et al., AJ 161, 176 (2021)
69. M.A.T. Groenewegen, A&A 669, A4 (2023)
70. E. Vasiliev, H. Baumgardt, MNRAS 505, 5978 (2021)
71. J. Maíz Apellániz, A&A 657, A130 (2022)
72. C. Flynn et al., MNRAS 509, 4276 (2022)
73. M.A.T. Groenewegen, A&A 654, A20 (2021)
74. K. El-Badry, H.W. Rix, T.M. Heintz, MNRAS 506, 2269 (2021)
75. M. Cruz Reyes, R.I. Anderson, A&A 672, A85 (2023)
76. J.L. Halbwachs et al. (2022), arXiv e-prints: arXiv:2206.05726
77. Gaia Collaboration, F. Arenou, C. Babusiaux, Barstow (2022), arXiv e-prints:
arXiv:2206.05595
78. T. Cantat-Gaudin, T.D. Brandt, A&A 649, A124 (2021)
79. J. Smith, Haywood, H. Eichhorn, MNRAS 281, 211 (1996)
80. C.A.L. Bailer-Jones, PASP 127, 994 (2015)
81. X. Luri et al., A&A 616, A9 (2018)
82. G.M. Eadie et al. (2023), arXiv e-prints: arXiv:2302.04703
83. A.G.A. Brown, Statistical astrometry, in Astrometry for Astrophysics, ed. by W.F. van Altena,
p. 242 (Cambridge University Press, 2012)
84. D.W. Hogg (2018), arXiv e-prints: arXiv:1804.07766
85. N. Gouda, Overview and recent progress of JASMINE near- infrared space astrometry mission
(2022). https://doi.org/10.5281/zenodo.7068453
86. D. Hobbs et al., Exp. Astron. 51, 783 (2021)
Chapter 4
Measuring H0 with H2 O Megamasers
. .

James A. Braatz and Dominic W. Pesce

4.1 Water Vapor Megamasers in the Nuclei of Active

Galaxies

Water vapor masers emitted at a rest frequency of 22.235080 GHz (1.35 cm) are
readily observed toward star forming regions and in the envelopes of evolved stars in
the Galaxy. These masers have typical isotropic luminosities of .∼ 10.−4 . L ʘ , though
in the brightest examples such as W49 they can emit with an isotropic luminosity
.∼ 1 . L ʘ and reach flux densities of thousands of Janskies. Extragalactic analogs to
such masers, associated with stars and star formation, have been detected in nearby
galaxies, where the integrated emission of the observed maser sources can reach
isotropic luminosities of .∼ 10 . L ʘ . Generally associated with enhanced star forming
galaxies, these systems have been called kilomasers. About 20 kilomaser galaxies
have so far been identified.
Another class of maser systems, the megamasers, are powered by active galactic
nuclei (AGN). These often have apparent isotropic luminosities of hundreds of . L ʘ ,
with the brightest among them being thousands of . L ʘ . About 160 extragalactic
masers in AGN have been detected to date1
Megamasers are highly variable over a range of timescales. Variability seen on
timescales of days and longer is likely the result of changes in the maser environment;
masers are highly beamed, nonlinear amplifiers that are sensitive both to the line-

1 A catalog of extragalactic water masers is maintained by the Megamaser Cosmology Project at

https://safe.nrao.edu/wiki/bin/view/Main/PublicWaterMaserList.

J. A. Braatz (B)
NRAO, 520 Edgemont Rd., Charlottesville, VA 22903, USA
e-mail: jbraatz@nrao.edu
D. W. Pesce
Center for Astrophysics .| Harvard & Smithsonian, 60 Garden St., Cambridge, MA 02138, USA
e-mail: dpesce@cfa.harvard.edu
© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 55
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_4
56 J. A. Braatz and D. W. Pesce

of-sight arrangement of molecular clouds and to the energy supply that produces
the pump to create the population inversion. Several have even shown variability
on timescales as short as minutes, where the rapid flux variations are thought to be
caused by interstellar scintillation (e.g. [17, 42]).
Megamasers exhibit some diversity in their spectral shapes. A few have broad
(.∼ 50–100 km s.−1 ) and highly variable profiles [3] and have been called “jet masers”
because they appear to be affiliated with strong radio continuum jets near the nucleus,
as exemplified by the maser in the radio galaxy NGC 1052 [11, 51]. Other megamaser
systems have scattered narrow components, and in some cases they appear to be
affiliated with outflows within a few parsecs of the AGN. Many megamasers have
simple profiles with emission detected only within .∼ 100 km s.−1 of the systemic
recession velocity of the host galaxy, and the nature of their emission is largely
unexplored. Others are rich with many spectral components at velocities that may be
separated from the galaxy recession velocity by 1000 km s.−1 or more.
In about 35 of the known water megamaser systems, the spectral profiles are
characterized by three distinct groups of narrow maser components, one group near
the galaxy recession velocity (systemic masers) and the other two (high-velocity
masers) red- and blue-shifted in velocity by .∼ 250–1000 km s.−1 . These masers arise
in the thin accretion disks that surround the supermassive black holes in the galaxy
nuclei, and they are called disk masers. The spectrum in Fig. 4.1 is representative of
the characteristic spectral profile of a disk maser, where in this example the recession
velocity of the host galaxy UGC 3789 is .∼3300 km s.−1 and the maser disk orbits the
supermassive black hole (SMBH) with velocities up to .∼750 km s.−1 .
Disk megamasers originate in the warm (.∼ 1000 K) molecular accretion disks
around low-luminosity AGN, typically .∼ 0.1–1 pc from the central SMBH. On these
scales, the gravitational potential is dominated by the SMBH and the gas exhibits
Keplerian orbital dynamics. The physical conditions of the molecular gas in these
disks enable a population inversion in the rotational energy levels responsible for
22 GHz water maser activity, and the maser emission is maximally amplified along
paths that (1) travel through a large volume of masing gas and (2) have small velocity
gradients along the line of sight (so as not to lose resonance for stimulated emission
by Doppler shifting). The exponential nature of maser amplification thus imposes a
strong observational selection preference to see edge-on disk systems, such that our
line of sight passes through the disk and thereby accesses the longest path lengths of
masing gas. For such an edge-on disk orientation, the line-of-sight velocity gradient
is minimized along sightlines parallel to or perpendicular to the orbits of the masing
gas, leading to the characteristic grouping of observed spectral features into systemic
and high-velocity components.
Disk masers are unique tools that have proven to be extremely productive scientific
probes. They provide a direct means to map the molecular gas within a pc of the
AGN, and thereby reveal the size and orientation of the thin accretion disk on these
scales; they orbit the supermassive black hole within its sphere of influence and
thus determine precise black hole masses; and in some cases their distances can be
measured using the geometric “megamaser technique”.
4 Measuring . H0 with H.2 O Megamasers 57

4.2 Measuring Geometric Distances to H.2 O Megamasers

Two methods have been employed to determine geometric distances to galaxies

based on observations of their accretion disk megamasers [25]. The first requires
measurements of the proper motions of systemic maser components as they orbit
the central supermassive black hole, achieved through a sequence of VLBI map-
ping observations. So far this technique has only been applied to the nearby galaxy
NGC 4258. The maser disk in this galaxy has high orbital velocities approaching
−1
.∼ 1000 km s. , and the observed proper motions are.∼ 30.μas yr.−1 , which is near the
limit of detection. Measuring such proper motions becomes prohibitively difficult
for more distant galaxies and for maser systems with smaller orbital velocities, and
we do not discuss this method further in the present contribution.
The second method (“the megamaser technique”) can be more widely applied. In
essence, the linear size of the maser disk, . R, can be determined through an analysis
of the measured centripetal accelerations (a = .v2 /R) of maser clouds combined with
the disk’s measured Keplerian rotation profile, while the angular size of the disk, .θ ,
is determined by VLBI mapping; the distance to the galaxy (D = R/.θ ) is therefore
determined geometrically [6, 25]. In practice, a 3-D model of the maser disk is fit to
the observed data to determine the galaxy distance, as described later.
The megamaser technique requires two types of observations. The first is, again,
a map of the distribution on the sky for each of the detected maser components using
VLBI. The second is a time sequence of spectral profiles usually spanning one to two
years for the purpose of measuring secular velocity drifts (i.e., line-of-sight accelera-
tions) of the individual, narrow maser components. Since each VLBI map includes a
spectral profile, all the required data could in principle be collected through a series
of VLBI observations, as in the first method. More commonly, however, spectral
monitoring has been achieved with a series of single-dish observations because they
are simpler to observe and process.
The maser targets of interest are typically faint, so observations to determine
distances have relied on the most sensitive telescopes at K-band (18–26 GHz): the
100-m Robert C. Byrd Green Bank Telescope (GBT), the Karl G. Jansky Very Large
Array (VLA), the 100-m Effelsberg Telescope, and the Very Long Baseline Array
(VLBA). Other large telescopes in the 60 to 70-m class such as the Sardinia Radio
Telescope, the Tianma Telescope, and the NASA DSN stations also make important
contributions, especially in regard to surveys and monitoring observations.

4.2.1 VLBI Maps of Maser Disks

A VLBI array observing megamaser disks typically has maximum projected E-W
and N-S baselines of .∼8500 km and .∼3500 km, respectively, making the synthesized
beam (.θb ≈ 1.2λ/D ) approximately 1.0 .× 0.4 mas at an observing frequency of
22 GHz. A representative maser disk in the Hubble flow spans an angular extent of
.∼1–2 mas (Fig. 4.1). Because the individual maser components observed in each
58 J. A. Braatz and D. W. Pesce

Fig. 4.1 Representative data from the megamaser galaxy UGC 3789. The bottom left panel shows
a GBT spectrum with the distinctive triple-peaked spectrum of a clean disk maser. The top left
panel shows the VLBI distribution of the maser spots on the sky; the colors represent line-of-sight
velocities, with red-shifted and blue-shifted masers colored appropriately to identify the rotation of
the edge-on maser disk. The top center panel shows a position-velocity diagram determined from
the VLBI map, with a Keplerian fit to the rotation shown in green. The impact parameter is the
projected angular distance from the black hole, measured along the disk diameter. The right panel
shows the velocities of the maser peaks from the systemic part of the spectrum as observed over
time, revealing the positive velocity drifts used to determine line-of-sight accelerations of maser
components on the front side of the disk

spectral channel are spatially unresolved, it is possible to determine the relative

positions of maser components within a single observation to a fraction of the beam
size by measuring the centroids of the unresolved masers with 2-D Gaussian fits. The
θb
uncertainty in relative positions of maser components is then .∆θ ≈ 2S/N , where .θb
is the VLBI synthesized beam size in one dimension and . S/N is the signal-to-noise
ratio of the detected maser component in the VLBI spectral channel. Peak maser flux
densities in the megamaser systems of interest typically reach .∼ 50–100 mJy, but the
much fainter components in the rich maser spectra must also be mapped with high
precision to optimize the model fit. Therefore the desired noise levels in VLBI maps
is .< 1 mJy beam.−1 per (representative) .∼ 0.3 km s.−1 channel. The narrow spectral
channels chosen for this noise calculation are set to have . 3 spectral channels across
the narrowest (.∼ 1 km s.−1 ) maser spectral components.
To achieve such sensitivity, VLBI mapping experiments have used the High Sen-
sitivity Array (HSA), which comprises the VLBA, the GBT, the phased VLA, and
4 Measuring . H0 with H.2 O Megamasers 59

the Effelsberg 100-m telescope. VLBI observing campaigns generally require aver-
aging multiple “tracks” of 6–9 hours each, and it is beneficial for these tracks to be
observed within a month or two of each other because the variable maser spectra
may change substantially on longer timescales, and the velocity drifts of systemic
maser lines will cause the lines to blur if the averaging time is long compared to
the timescale of the drift. When added to VLBA observations, the GBT and VLA
provide exquisitely sensitive baselines but do not affect the overall synthesized beam
size significantly, while the Effelsberg telescope can increase the E-W baselines for
high-declination targets. The observations are sensitive to weather and use globally
dispersed telescopes under uncorrelated atmospheric conditions, so they are best
done in the winter months. The sensitivity of the array enables relative positions of
maser components to be measured with a precision of a few tens of .μas. The VLBI
map in Fig. 4.1 shows a representative case in UGC 3789 with the detected masers
spanning .∼ 2 mas and showing orbits .∼ 0.1 - 0.3 pc from the supermassive black
hole.

4.2.2 Spectral Monitoring of Maser Profiles

Maser profiles are observed with periodic monitoring observations to measure secu-
lar velocity drifts of the spectral components. For these maser clouds in orbit around
the central black hole, such velocity drifts reveal the line-of-sight centripetal acceler-
ations. Either sensitive single-dish observations or interferometric observations may
be used to collect the spectral information. Ideally, a sequence of spectral monitor-
ing observations would provide measured accelerations for every one of the maser
spots identified in the VLBI map. The maps, however, use long integrations with
the HSA and are significantly more sensitive than individual epochs from the mon-
itoring observations, so the faintest maser lines mapped with VLBI may not have
corresponding detections in the monitoring sequence to measure accelerations. To
optimize the matching of VLBI-detected maser components with measured acceler-
ations, it is beneficial to monitor the spectral profiles over a time span that straddles
the VLBI observations.
Observing campaigns using GBT monitoring to measure accelerations have used
a roughly monthly cadence for observations. Because maser observations at 22 GHz
are sensitive to humidity and other weather conditions, the GBT does not typically
achieve sensitive K-band observations during the summer months. The GBT moni-
toring season will therefore span about 9 months. Spectra obtained from the VLBI
observations, ideally observed in the middle of the monitoring sequence, supple-
ment the monitoring data. Summer gaps can be filled in with observations from the
VLA, which has similar sensitivity to the GBT at K-band and generally much dryer
observing conditions in the summer.
The line widths of individual maser components in a disk maser spectral profile
are .∼1–2 km s.−1 . Each of the three groups of maser features – systemic, red-, and
blue-shifted – comprise many such lines spanning .∼50 - 250 km s.−1 in each group
60 J. A. Braatz and D. W. Pesce

(see Fig. 4.1), and the individual lines are blended and variable. Measuring individual
accelerations therefore requires a global fitting procedure to model the accelerations,
velocities, and amplitudes of each of the maser spectral components, decomposed
from each spectrum and approximated by Gaussian profiles. Sensitive spectra are
required to provide robust and precise fitting. For typical maser disk systems in the
Hubble flow, spectral monitoring requires sensitivities of .∼ 2 mJy per .∼ 0.3 km s.−1
spectral channel (three channels across the narrowest maser lines), achievable with
GBT observations of 2–3 hours per epoch.
Accelerations are determined for each of the narrow maser components observed
throughout the spectrum. In the red- and blue-shifted components, the masers are
expected to be near the midline of the maser disk, where the acceleration vectors point
in the plane of the sky toward the black hole, so the measured line-of-sight accel-
erations are generally near zero. Small measured positive and negative deviations
from zero, however, can help to place such maser clouds at small azimuthal offsets
from the midline and help constrain the disk model. The systemic maser drifts, as
shown in the right panel of Fig. 4.1, are always measured with positive accelerations
(redward velocity drifts), indicating that the systemic masers originate on the near
side of the disk, in front of the black hole. Presumably, maser clouds also exist on the
back side of the disk and their emission is either attenuated by free-free absorption
in the vicinity of the black hole or beamed away from our line of sight. Masers at
azimuthal angles intermediate between the midline and the observed line to the black
hole are fainter, if detected at all, as the line-of-sight velocity coherence of gas in the
rotating disk is reduced at those intermediate angles.

4.2.3 Modeling Maser Disks to Determine Their Distance

The heterogeneous observations used to measure maser distances – VLBI mapping

plus single-dish spectral monitoring – generate a (x,y,v,a) data set consisting of
the positions, line-of-sight velocities, and line-of-sight accelerations for each maser
spot measured from a VLBI spectral cube, along with their uncertainties. The (x,y)
on-sky relative positions of each maser component are determined by fitting an ellip-
tical Gaussian to the unresolved maser detection in each VLBI spectral channel.
Uncertainties in the sky position are taken from the formal uncertainties in the Gaus-
sian fitting procedure. Systematic uncertainties (e.g. from residual delay errors in
the VLBI calibration) may be included as well. The line-of-sight velocity for each
spot is determined from the center frequency of the VLBI spectral channel, and its
uncertainty is taken to be the half-width of the spectral channel. The line-of-sight
acceleration for each spot, and its uncertainty, are as measured from the fits to the
secular velocity drifts in the monitoring observations. Each maser spot also has a
measured flux density and a corresponding uncertainty, but these values are not used
directly in the fitting of the disk model. The flux density of each maser component,
however, determines the signal-to-noise ratio that sets the uncertainty in its measured
position and contributes to the uncertainty in its measured acceleration.
4 Measuring . H0 with H.2 O Megamasers 61

The model fit to these data is a three-dimensional, thin, warped disk representing
the geometry of the gas in orbit around the central supermassive black hole. While
models allowing for confocal elliptical orbits have been investigated, all maser disks
studied to date have been well fit by concentric circular orbits that are allowed to
tilt in position angle and inclination angle to account for the warp. The warp is
therefore parameterized by the position angle .Ω0 of the orbital ring at a reference
radius (which may be zero) and its derivative with angular radius, .dΩ/dr , as well
as the inclination angle at the reference radius, .i 0 and its radial derivative .di/dr .
Second order derivatives may be applied in modeling these warp parameters as well.
Each maser spot observed with VLBI is therefore placed on the disk at a modeled
azimuthal angle and radius from the disk center. As described above, every VLBI
maser component may not have a measured acceleration assigned to it, but spatial
and velocity data alone can still help to constrain the fit, in the context of the broader
model. High velocity components with unmeasured line-of-sight accelerations are
assigned to fall near the disk midline (where the line-of-sight acceleration is zero)
with appropriate azimuthal and radial uncertainties.
The full set of global model parameters, then, in a representative fit to a disk maser
system consists of the position (.x0 , y0 ) of the dynamical center (i.e., the position
of the supermassive black hole for relaxed Keplerian systems), the line-of-sight
recession velocity .v0 of the black hole, the position angle and inclination angle
warp coefficients, the mass . Mbh of the supermassive black hole, and the angular-size
distance . D to the maser disk. The Hubble Constant is thus determined completely
from each fit maser disk. The uncertainty in . H0 for each galaxy’s measurement
must account not only for the statistical uncertainties from the model fit but also
an estimate of the (unknown) peculiar velocity, .v p , of the galaxy, since the Hubble
recession velocity of each galaxy is .vr ec = v0 + v p .
Early papers utilized a random walk Metropolis-Hastings Markov Chain Monte
Carlo (MCMC) algorithm to sample the posterior probability distributions [48] for
the global parameters of the model. The MCMC algorithm incorporates an estimate
of the “error floor” on the input data parameters, tunable by the user, to account for
systematic uncertainties in the measurements. Conservatively applied, these error
floor estimations limit the precision of H.0 measurements. An update to the fitting
procedure [44] applies a more efficient Hamiltonian Monte Carlo (HMC) sampler and
fits for the error floor parameters directly, thereby reducing the systematic uncertainty
and improving the overall fit uncertainties on the global parameters. Results presented
in this chapter refer to fits using the HMC sampler.

4.3 The Maser Distance to NGC 4258 as an Anchor

to Distance Ladders

The archetypal H.2 O megamaser disk is found in the nucleus of the low-luminosity
Seyfert 2 galaxy NGC 4258. Fortuitously nearby at a distance of 7.6 Mpc, the maser
was first discovered [9] during a survey of 29 late-type galaxies with the OVRO 40-m
62 J. A. Braatz and D. W. Pesce

telescope and using a 40 MHz bandwidth spectrometer that covered .∼ 500 km s.−1 in
velocity, centered on the galaxy’s recession velocity. The discovery observation thus
detected only the systemic maser emission, with an isotropic luminosity of .∼120 L.ʘ ,
much greater than the luminosities of masers associated with star formation. Subse-
quent VLA observations [10] of NGC 4258 and the other newly detected megamaser
in NGC 1068 (450 L.ʘ ) localized the maser emission to the central .∼1 pc of the AGN
and clearly identified the maser emission with nuclear gas. The discovery of high-
velocity maser emission in NGC 4258 [40], Doppler-shifted by .∼ .±1000 km s.−1 ,
was achieved during testing of a new 200 MHz spectrometer on the Nobeyama 45-m
telescope. Early spectral monitoring of the systemic masers [16, 21] identified the
secular velocity drifts of the systemic masers, now known to be affiliated with the
orbital acceleration.
The first VLBI map to include the high-velocity lines [39] revealed the edge-on,
warped maser disk in Keplerian rotation and secured the identification of the central
object as a supermassive black hole with a mass of .∼4 × 107 M.ʘ . The method to
determine the geometric distance to the maser [25] based on the observed orbital
properties of the maser clouds was finally established and the distance measured
through a series of VLBI monitoring observations. The initial work applied the two
independent techniques to measure the maser distance, the first based on the measured
proper motions, and the second using the aforementioned measurements of line-of-
sight accelerations. As the proper motions of the fast-rotating systemic masers in
nearby NGC 4258 are only .∼30 .μas yr.−1 , just at the detection threshold for VLBI
monitoring, all subsequent high-precision distance measurements of masers in this
galaxy, as well as all the more distant megamasers, rely on the acceleration method.
NGC 4258 provides a critical anchor for calibrating both the Cepheid period-
luminosity relationship and the tip of the red giant branch (TRGB) [49]. Because the
metallicity of NGC 4258 is similar to that of many of the Cepheid-host galaxies it
is used to calibrate, and the metallicity gradient in the galaxy itself is small, NGC
4258 reduces the systematic uncertainty introduced using other anchors such as the
LMC. Likewise, the Cepheids in NGC 4258 have a similar mean level of “crowded
backgrounds” as SN Ia host systems, further reducing systematics.
To refine the measured maser distance to NGC 4258, an exquisite set of maser
monitoring data were obtained through 18 epochs over 3 years [1] and initially were
fit with the warped disk model using tunable error floors for the input parameters,
conservatively set, to obtain a distance 7.60 .± 0.17 (stat) .± 0.15 (sys) Mpc [26].
The same data were later re-analyzed, including the error floors as part of the formal
fit, using both a modified Metropolis-Hastings MCMC approach as well as a Hamil-
tonian MCMC approach, to determine the distance 7.576 .± 0.082 (stat) .± 0.076
(sys) Mpc [49]. Upon applying only the NGC 4258 maser distance as the anchor to
the Cepheid-to-SN Ia ladder, the Hubble Constant was determined to be
H.0 = 72.0 .± 1.9 km s.−1 Mpc.−1 . Applying the maser anchor to the TRGB method
results in H.0 = 71.1 .± 1.9 km s.−1 Mpc.−1 [49]. Along with Gaia/HST parallaxes to
MW Cepheids and measurements of Detached Eclipsing Binaries in the LMC/SMC,
the maser distance to NGC 4258 marks one of the three geometric anchors to the
“baseline” measurement of the S. H0 ES team: H.0 = 73.04 .± 1.04 km s.−1 Mpc.−1 [50].
4 Measuring . H0 with H.2 O Megamasers 63

4.4 Direct Measurement of H.0 from Megamaser Distances

in the Hubble Flow

NGC 4258 is too near to provide a precise, direct measurement of . H0 on its own,
as its unknown peculiar velocity may be a large fraction of its observed recession
velocity. At the time its maser distance was first measured, it was the only known
maser disk galaxy for which the megamaser method could be applied. With peak
flux densities reaching several Jy, it is by far the brightest of the “clean” disk masers
in Keplerian rotation around the central supermassive black hole. The maser also
extends some .∼15 mas in angular extent, much larger than the typical 0.4 .× 1.0 mas
VLBI beam, thus making it the largest known megamaser disk in angular size on the
sky by close to an order of magnitude. Some initial speculation was that the maser
was a rare system with an extremely fortunate placement in the universe at only
7.6 Mpc distance. If at 100 Mpc, the peak maser flux in NGC 4258 would be
.∼20 mJy, detectable only by 100-m class telescopes. Given the opportunity to mea-

sure direct geometric distances in the Hubble flow, the remarkable maser system
in NGC 4258 therefore inspired new surveys to identify similar systems at larger
distances that could be measured using the same megamaser technique.

4.4.1 Early Surveys to Detect Megamasers in AGN

The maser in NGC 4258 was discovered as part of a series of early surveys seeking
water maser systems in late-type galaxies, where accumulations of bright Orion- and
W49-type masers would be expected to make them detectable in the nearest galaxies.
The brightest and most luminous extragalactic masers discovered in these surveys,
however, were found to be associated with the nuclear gas near AGN [10], and so
search strategies later refocused on AGN in the 1990 s. Several large surveys at that
time [2, 3] with the Effelsberg 100-m and Parkes 64-m telescopes, the most sensitive
single dish telescopes at K-band in the northern and southern hemispheres at that
time, soon identified Seyfert 2 galaxies (and some LINERs) as the most likely hosts
of detectable maser emission, with the implication being that the maser amplification
benefits from the edge-on orientation of the putative obscuring toroidal structures in
these AGN, providing long amplification gain paths. Such early surveys revealed a
diversity of maser spectra, including the prototypical jet maser in NGC 1052 [2] and
the remarkable luminous gigamaser in TXFS 2226–184 [30].
Two important analogs to the maser disk in NGC 4258 were discovered as part
of the early AGN surveys. Like in NGC 4258, the initial discovery of the mega-
maser in IC 2560 [3] revealed only maser components near the systemic recession
velocity, and subsequent, more sensitive wideband observations with Nobeyama [29]
identified high velocity emission, both red- and blue-shifted. Despite the relatively
strong emission from this maser disk galaxy, its southern location at .δ = -34.◦ has
hampered the detailed study needed to get a precise maser distance. Second, the rich
64 J. A. Braatz and D. W. Pesce

maser system in Mrk 1419 [24] shows the characteristic triple-peaked spectral profile
indicative of a disk maser. Despite its maser system having the signature disk profile
with many bright spectral components, subsequent studies have shown it is difficult
to model, possibly because of an apparent thickness in its observed maser disk as
well as incompletely-traced substructure.
Progress with maser surveys focusing on AGN flourished in subsequent years,
initially with programs that used the 70-m NASA Deep Space Stations [18, 31] and
the Effelsberg 100-m telescope [20]. It was the significant step up in sensitivity at
K-band offered by the GBT, however, that opened the field for studies of distant disk
masers. The most productive GBT surveys in terms of total number of megamaser
discoveries focused on narrow-line active galaxies [4, 19, 32]. However the signifi-
cant discovery of the megamaser disk in the galaxy UGC 3789 resulted, surprisingly,
from a nonselective GBT snapshot survey of nearby, luminous galaxies [5]. This
galaxy was confirmed to host a Seyfert 2 nucleus by optical spectroscopy only after
the disk maser discovery.

4.4.2 The Megamaser Cosmology Project

The Megamaser Cosmology Project [6, 47] (MCP) is an international effort formed to
measure the Hubble Constant using the megamaser method. It started by measuring
the distance to the maser disk in UGC 3789, and then followed up by conducting a
large survey with the GBT to find and measure distances to additional megamaser
disk galaxies in the Hubble Flow. The MCP and its precursor surveys with the GBT
eventually surveyed .> 2500 galaxies and discovered about half of the currently-
known 160 megamaser galaxies. The MCP surveys focused on nearby (z .< 0.05),
narrow-line AGN selected from the Sloan Digital Sky Survey (SDSS), the 6dF galaxy
survey, and the 2MASS Redshift Survey (2MRS), and included a smaller number
of X-ray selected AGN and “apparently normal” galaxies. The survey also included
candidates selected from infrared properties observed by the Wide-Field Infrared
Survey Explorer (WISE).
Figure 4.2 shows the progress made by extragalactic maser surveys over time.
The first detection of an extragalactic maser in a star-forming cloud in M33 [8] was
made at Effelsberg and published in 1977. GBT surveys were particularly productive
starting in 2004, and then surveys slowed in 2016 as the MCP refocused efforts on
VLBI and monitoring the known disk masers to measure their distances.
The overall detection rate of.< 3% for MCP surveys reflects the difficulty in identi-
fying optimal samples of AGN candidates. The large maser surveys have mostly sam-
pled nearby (z .< 0.05) Seyfert 2 galaxies indiscriminately. Efforts to find improved
selection criteria based on a multivariate parameter analysis of the optical and mid-
infrared photometric properties show promise of increasing overall detection rates
[36]. In particular, compared to a control sample of AGN, megamaser galaxies have
been found to have stronger emission at mid-IR wavelengths observed with WISE,
with the strongest enhancement at 22 .μm. In addition, megamaser galaxies tend to
4 Measuring . H0 with H.2 O Megamasers 65

Fig. 4.2 The number of extragalactic water masers discovered each year, labeled by the telescope
that made the discovery

have redder IR colors and larger integrated mid-IR luminosities [36]. The correlation
of megamasers with X-ray bright galaxies and galaxies with X-ray column densi-
ties N. H .> 10.23 cm.−2 is well known [41, 54], and recent work demonstrates that
the X-ray and mid-IR properties could be examined together to enhance megamaser
detection rates significantly [37]. Surveys of nearby narrow-line Seyfert 1 (NLS1)
galaxies have also identified an enhanced detection rate [52], but the overall sample
sizes are small and the newly identified galaxies from NLS1 samples have so far not
been identified with disk masers.
The combined total of extragalactic maser surveys have so far identified .∼ 160
megamaser galaxies in AGN. About 40 show high velocity masers consistent with a
disk origin, making them candidates for precise black hole mass measurements (e.g.,
[33, 55]), and .∼ 35 of those show the triple-peaked spectra characteristic of “clean”
disk masers [42, 44], making them further candidates for distance measurements. A
clean disk megamaser in this context is a maser demonstrating a clear, triple-peaked
spectral profile with little or no evident maser emission at intermediate velocities
from non-disk components such as jets or outflows.
Because of the significant observing time requirements for VLBI and monitoring
to measure each megamaser distance, only 1 or 2 maser galaxies have been observed
as part of a distance campaign using the HSA in any one year. The MCP therefore
selected and began measuring the best candidates from the set of clean disk masers.
Those candidates are identified based on (1) having a declination.δ .> -20.◦ suitable for
sensitive and efficient observing with the (northern) HSA; (2) having bright and rich
maser spectra in all of the systemic, red-shifted, and blue-shifted parts of the spectral
66 J. A. Braatz and D. W. Pesce

Fig. 4.3 The Hubble diagram for H.2 O megamaser galaxies reproduced from [45]. Each galaxy is
plotted with 1.σ uncertainties in distance and 250 km s.−1 in velocity set by the unknown peculiar
velocities. The solid black line shows the best fit Hubble Constant and the shaded gray regions show
1.σ and 2.σ confidence intervals

profile, with many components in each to better constrain the disk model; and (3)
having distinct and clearly measured velocity drifts among the systemic components
in preliminary monitoring observations.
The best current megamaser measurement of H.0 by the MCP incorporates mea-
surements for five systems in the Hubble Flow (UGC 3789 [48], NGC 6264 [34],
NGC 6323 [35], NGC 5765b [13], and CGCG 074–064 [44]) as well as NGC 4258
[49]. Distances to each galaxy are calculated using a flat .ɅCDM cosmology with
.Ωm = 0.315, but the measurement is not strongly dependent on the cosmological
model. The galaxies are all within 150 Mpc and the distance calculations vary by
.< 1% for any .Ωm .< 0.5. The recession velocities are corrected to the frame of the
Cosmic Microwave Background. With the number of galaxies involved in the overall
measurement being small, the unknown peculiar velocities of the galaxies represents
a significant source of uncertainty. When applying a conservatively-estimated uni-
form uncertainty of 250 km s.−1 on the recession velocity of each galaxy, added in
quadrature with the (much smaller) uncertainty on the fit velocity of the dynamical
center, the MCP determines . H0 = 73.9 .± 3.0 km s.−1 Mpc.−1 [45] (“best result”).
The Hubble diagram for this determination is presented in Fig. 4.3. When applying a
group recession velocity for each measured megamaser galaxy rather than the maser-
determined recession velocity, the result becomes . H0 = 73.3 .± 2.8 km s.−1 Mpc.−1 ,
and when applying different galaxy flow models the result varies between 71.8 and
76.9 km s.−1 Mpc.−1 , with similar uncertainties.
4 Measuring . H0 with H.2 O Megamasers 67

4.5 Prospects for Improving the Measurement of . H0

with Megamasers

Improvements in measuring H.0 with megamasers will progress on a number of fronts.

Several known, clean disk megamaser galaxies remain to be observed and analyzed
with sufficient precision to contribute to the combined MCP result. These include
J0437+2456, ESO 558-G009, and Mrk 1419 [28]. In addition, the megamaser in
IC 2560, which shows the promising characteristics of a high quality distance target
[53] despite its southern (.δ .≈ -34.◦ ) location, could contribute meaningfully to the
measurement. Improvements in survey strategies and the availability of new catalogs
of Seyfert galaxies will lead to more efficient discovery of new megamaser systems.
Meanwhile, improvements in disk modeling that relax the thin disk constraint could
further reduce the statistical errors in current fits.

4.5.1 Sub-millimeter Water Megamasers and Other

Molecules

Water is not the only molecule known to exhibit megamaser emission. The hydroxyl
(OH) molecule is also widely observed as a megamaser, primarily associated with
gas-rich major galaxy mergers where it traces densities .n(H.2 ) . 10.4 cm.−3 [12], and
other molecules have been detected as megamasers as well (e.g. [14]), tracing similar
density gas. The 6.16 - 5.23 rotational transition of ortho-water at 22 GHz is, however,
particularly useful for these cosmological studies because it is optimally excited at
warm temperatures (T. K .∼ 1000 K) and high densities.n(H.2 ).∼ 10.9 cm.−3 [15], making
it a clear tracer of the Keplerian accretion disks in AGN. Indeed, water remains the
only molecule detected as a maser directly in the AGN accretion disks. Moreover,
the relatively high frequency of the well-studied 22 GHz water megamasers enables
a smaller VLBI beam and precise mapping of the pc-scale disks.
Other rotational transitions of the water molecule at even higher frequencies have
also been observed to emit as masers, with several sub-millimeter transitions detected
as megamasers in AGN [27]. The development and expansion of ALMA has opened
the opportunity to study these sub-mm transitions expansively. The 183 GHz maser
line has particularly attracted interest because, in early ALMA studies, it has been
found to accompany 22 GHz megamasers [46] widely in accretion disk systems. The
321 GHz transition has also been detected in low-z active galaxies [22, 23, 43]. The
promising 380 GHz transition, which is anticipated to be more luminous even than
the 22 GHz line when the gas temperature exceeds 1000 K, falls in a particularly
strong atmospheric absorption band for low-z systems, but can be detected in high-z
[38] QSOs, where it gets red-shifted into a frequency range where the atmosphere
is more transparent. Other maser transitions from the water molecule that have been
observed in Galactic sources or predicted in molecular models are yet to be explored
in AGN.
68 J. A. Braatz and D. W. Pesce

Compared to the 22 GHz masers, the potential factor of 8 to 17 improvement in

VLBI angular resolution from these higher frequency masers, combined with the
unexplored incidence rate of sub-mm masers in AGN, has elevated the community
interest in finding and exploring such masers as potential future distance indicators.
Spectral line VLBI at sub-mm wavelengths is a developing field, and as VLBI facil-
ities expand and mature at these wavelengths, newly discovered maser disk systems
may generate even more precise results than the 22 GHz transition.

4.5.2 The ngVLA Era

Megamaser surveys are sensitivity limited. The advent of the sensitive GBT opened
the field of direct H.0 measurement with megamasers, and the next leap forward in
K-band sensitivity will come in the mid-2030s with the Next Generation Very Large
Array (ngVLA). This telescope will include 244 antennas each 18-m in diameter
as well as 19 more antennas 6-m in diameter and will offer better than an order of
magnitude improvement in 22 GHz sensitivity over the VLA while incorporating the
long baselines of the VLBA.
The sensitivity improvement will mean that, for maser systems of a given isotropic
luminosity, the ngVLA will be able to sample a volume .∼ 30X larger than the VLA
or GBT (which have similar K-band sensitivities to each other). The ngVLA will
also be an efficient maser survey instrument, and every observation will include
both spectral information and VLBI imaging. Combined spectral monitoring with
ngVLA mapping will eliminate the heterogeneous aspect of data used in current
measurements and enable an improved analysis scheme to model the disk parameters
and dynamics in one global fit, rather than separate modeling of accelerations and
disk structure.
By discovering and measuring new megamaser systems, and improving the mea-
surements of known systems, the ngVLA will enable a.∼ 1% geometric measurement
of the Hubble Constant using the megamaser method [7]. Since individual maser dis-
tances
√ are uncorrelated, the cumulative improvement in the uncertainty improves as
.1/ N and so the telescope could attain such a 1% result by measuring, e.g., 7% dis-
tances to 50 galaxies. In practice, the uncertainty in each individual maser galaxy’s
distance determination will vary, being dependent on the quality of its disk maser
system. If the galaxies are widely spaced on the sky, the larger number of measure-
ments will average out the uncertainty associated with peculiar velocities and flow
models.
4 Measuring . H0 with H.2 O Megamasers 69

4.6 Summary

Circumnuclear water vapor megamasers at 22 GHz in AGN provide a 1.5% anchor for
extragalactic distance ladders and also enable a one-step, geometric measurement of
the Hubble Constant. The best current measurement of H.0 using only the megamaser
method by the MCP is 73.9 .± 3.0 km s.−1 Mpc.−1 . Incremental improvements to this
measurement will become available as more galaxies are measured and as megamaser
methods extend into sub-mm maser transitions. The ngVLA, offering an order of
magnitude better sensitivity at 22 GHz, will enable a leap forward in measurements
and will permit a 1% measurement of H.0 .

References

1. A.L. Argon, L.J. Greenhill, M.J. Reid, J.M. Moran, E.M.L. Humphreys, Astrophys. J. 659,
1040 (2007). https://doi.org/10.1086/512718
2. J.A. Braatz, A.S. Wilson, C. Henkel, Astrophys. J. 437, L99 (1994). https://doi.org/10.1086/
187692
3. J.A. Braatz, A.S. Wilson, C. Henkel, Astrophys. J. Supp. Ser. 106, 51 (1996). https://doi.org/
10.1086/192328
4. J.A. Braatz, C. Henkel, L.J. Greenhill, J.M. Moran, A.S. Wilson, Astrophys. J. Lett. 617, L29
(2004). https://doi.org/10.1086/427185. (arXiv:astro-ph/0412352 [astro-ph].)
5. J.A. Braatz, N.E. Gugliucci, Astrophys. J. 678, 96–101 (2008). https://doi.org/10.1086/529538
(arXiv:0804.4136 [astro-ph])
6. J.A. Braatz, M.J. Reid, E.M.L. Humphreys, C. Henkel, J.J. Condon, K.Y. Lo, Astrophys. J.
718, 657–665 (2010). https://doi.org/10.1088/0004-637X/718/2/657 (arXiv:1005.1955 [astro-
ph.CO])
7. J. Braatz, D. Pesce, J. Condon, M. Reid, A.S.P. Conf. Ser. 517, 821 (2018). (arXiv:1810.06686
[astro-ph.CO])
8. E. Churchwell, A. Witzel, W. Huchtmeier, I. Pauliny-Toth, J. Roland, W. Sieber, Astron. Astro-
phys. 54, 969 (1977)
9. M.J. Claussen, G.M. Heiligman, K.Y. Lo, Nature 310, 298 (1984). https://doi.org/10.1038/
310298a0
10. M.J. Claussen, K.Y. Lo, Astrophys. J. 308, 592 (1986). https://doi.org/10.1086/164529
11. M.J. Claussen, P.J. Diamond, J.A. Braatz, A.S. Wilson, C. Henkel, Astrophys. J. 500, L129
(1998). https://doi.org/10.1086/311405
12. J. Darling, Astrophys. J. Lett. 669, L9 (2007). https://doi.org/10.1086/523756.
(arXiv:0710.1080 [astro-ph].)
13. F. Gao, J.A. Braatz, M.J. Reid, K.Y. Lo, J.J. Condon, C. Henkel, C.Y. Kuo, C.M.V. Impellizzeri,
D.W. Pesce, W. Zhao, Astrophys. J. 817(2), 128 (2016). https://doi.org/10.3847/0004-637X/
817/2/128 (arXiv:1511.08311 [astro-ph.GA])
14. M.D. Gorski, S. Aalto, J. Mangum, E. Momjian, J.H. Black, N. Falstad, B. Gullberg, S. König,
K. Onishi, M. Sato, F. Stanley, Astron. Astrophys. 654, A110 (2021). https://doi.org/10.1051/
0004-6361/202141633 ([arXiv:2108.06331])
15. M.D. Gray, A. Baudry, A.M.S. Richards, E.M.L. Humphreys, A.M. Sobolev, J.A.
Yates, Mon. Not. R. Astron. Soc. 456, 374 (2016). https://doi.org/10.1093/mnras/stv2437.
([arXiv:1510.06182])
16. L.J. Greenhill, C. Henkel, R. Becker, T.L. Wilson, J.G.A. Wouterloot, Astron. Astrophys. 304,
21 (1995)
70 J. A. Braatz and D. W. Pesce

17. L.J. Greenhill, S.P. Ellingsen, R.P. Norris, R.G. Gough, M.W. Sinclair, J.M. Moran,
R. Mushotzky, Astrophys. J. Lett. 474, L103 (1997). https://doi.org/10.1086/310434.
(arXiv:astro-ph/9611127 [astro-ph].)
18. L.J. Greenhill, P.T. Kondratko, J.E.J. Lovell, T.B.H. Kuiper, J.M. Moran, D.L. Jauncey, G.P.
Baines, Astrophys. J. 582, L11 (2003). https://doi.org/10.1086/367602
19. L.J. Greenhill, P.T. Kondratko, J.M. Moran, A. Tilak, Astrophys. J. 707, 787–799 (2009).
https://doi.org/10.1088/0004-637X/707/1/787 (arXiv:0911.0382 [astro-ph.CO])
20. Y. Hagiwara, P.J. Diamond, M. Miyoshi, E. Rovilos, W. Baan, Mon. Not. R. Astron. Soc. 344,
L53 (2003). https://doi.org/10.1046/j.1365-8711.2003.07005.x
21. A.D. Haschick, W.A. Baan, E.W. Peng, Astrophys. J. Lett. 437, L35 (1994). https://doi.org/
10.1086/187676
22. Y. Hagiwara, M. Miyoshi, A. Doi, S. Horiuchi, Astrophys. J. Lett. 768, L38 (2013). https://
doi.org/10.1088/2041-8205/768/2/L38. (arXiv:1305.0076 [astro-ph.GA].)
23. Y. Hagiwara, S. Horiuchi, A. Doi, M. Miyoshi, P.G. Edwards, Astrophys. J. 827, 69 (2016).
https://doi.org/10.3847/0004-637X/827/1/69. ([arXiv:1604.07937])
24. C. Henkel, J.A. Braatz, L.J. Greenhill, A.S. Wilson, Astron. Astrophys. 394, L23 (2002). https://
doi.org/10.1051/0004-6361:20021297. (arXiv:astro-ph/0211226 [astro-ph].)
25. J.R. Herrnstein, J.M. Moran, L.J. Greenhill, P.J. Diamond, M. Inoue, N. Nakai, M.
Miyoshi, C. Henkel, A. Riess, Nature 400, 539–541 (1999). https://doi.org/10.1038/22972.
(arXiv:astro-ph/9907013 [astro-ph])
26. E.M.L. Humphreys, M.J. Reid, J.M. Moran, L.J. Greenhill, A.L. Argon, Astrophys. J. 775, 13
(2013). https://doi.org/10.1088/0004-637X/775/1/13
27. E.M.L. Humphreys, L.J. Greenhill, M.J. Reid, H. Beuther, J.M. Moran, M. Gurwell, D.J.
Wilner, P.T. Kondratko, Astrophys. J. Lett. 634, L133–L136 (2005). https://doi.org/10.1086/
498890 (arXiv:astro-ph/0511014 [astro-ph])
28. C.M.V. Impellizzeri, J.A. Braatz, C.Y. Kuo, M.J. Reid, K.Y. Lo, C. Henkel, J.J. Condon,
IAU Symp. 287, 311 (2012). https://doi.org/10.1017/S1743921312007235. (arXiv:1205.4758
[astro-ph.CO].)
29. Y. Ishihara, N. Nakai, N. Iyomoto, K. Makishima, P. Diamond, P. Hall, Publ. Astron. Soc. Jap.
53, 215 (2001). https://doi.org/10.1093/pasj/53.2.215
30. A.M. Koekemoer, C. Henkel, L.J. Greenhill, A. Dey, W. van Breugel, C. Codella, R. Antonucci,
Nature 378, 697 (1995). https://doi.org/10.1038/378697a0
31. P.T. Kondratko, L.J. Greenhill, J.M. Moran, J.E.J. Lovell, T.B.H. Kuiper, D.L. Jauncey, L.B.
Cameron, J.F. Gomez, C. Garcia-Miro, E. Moll et al., Astrophys. J. 638, 100–105 (2006).
https://doi.org/10.1086/498641. (arXiv:astro-ph/0510851 [astro-ph].)
32. P.T. Kondratko, L.J. Greenhill, J.M. Moran, Astrophys. J. 652, 136–145 (2006). https://doi.
org/10.1086/507885 (arXiv:astro-ph/0610060 [astro-ph].)
33. C.Y. Kuo, J.A. Braatz, J.J. Condon, C.M.V. Impellizzeri, K.Y. Lo, I. Zaw, M. Schenker, C.
Henkel, M.J. Reid, J.E. Greene, Astrophys. J. 727, 20 (2011). https://doi.org/10.1088/0004-
637X/727/1/20. (arXiv:1008.2146 [astro-ph.CO].)
34. C. Kuo, J.A. Braatz, M.J. Reid, F.K.Y. Lo, J.J. Condon, C.M.V. Impellizzeri, C. Henkel, Astro-
phys. J. 767, 155 (2013). https://doi.org/10.1088/0004-637X/767/2/155. (arXiv:1207.7273
[astro-ph.CO].)
35. C.Y. Kuo, J.A. Braatz, K.Y. Lo, M.J. Reid, S.H. Suyu, D.W. Pesce, J.J. Condon, C. Henkel,
C.M.V. Impellizzeri, Astrophys. J. 800(1), 26 (2015). https://doi.org/10.1088/0004-637X/800/
1/26. (arXiv:1411.5106 [astro-ph.GA].)
36. C.Y. Kuo, A. Constantin, J.A. Braatz, H.H. Chung, C.A. Witherspoon, D. Pesce, C.M.V. Impel-
lizzeri, F. Gao, L. Hao, J.-H. Woo, I. Zaw, Astrophys. J. 860, 169 (2018). https://doi.org/10.
3847/1538-4357/aac498
37. C.Y. Kuo, J.Y. Hsiang, H.H. Chung, A. Constantin, Y.-Y. Chang, E. da Cunha, D. Pesce, W.T.
Chien, B.Y. Chen, J.A. Braatz, I. Zaw, S. Matsushita, J.C. Lin, Astrophys. J. 892, 18 (2020).
https://doi.org/10.3847/1538-4357/ab781d
38. C.Y. Kuo, S.H. Suyu, V. Impellizzeri, J.A. Braatz, https://doi.org/10.1093/pasj/psz032.
(arXiv:1901.04745 [astro-ph.GA])
4 Measuring . H0 with H.2 O Megamasers 71

39. M. Miyoshi, J. Moran, J. Hernstein, L. Greenhill, N. Nakai, P. Diamond, M. Inoue, Nature 373,
127–129 (1995). https://doi.org/10.1038/373127a0
40. N. Nakai, M. Inoue, M. Miyoshi, Nature 361, 45 (1999)
41. F. Panessa, P. Castangia, A. Malizia, L. Bassani, A. Tarchi, A. Bazzano, P. Uber-
tini, Astron. Astrophys. 641, A162 (2020). https://doi.org/10.1051/0004-6361/201937407.
(arXiv:2006.08280 [astro-ph.GA].)
42. D.W. Pesce, J.A. Braatz, J.J. Condon, F. Gao, C. Henkel, E. Litzinger, K.Y. Lo, M.J. Reid, Astro-
phys. J. 810(1), 65 (2015). https://doi.org/10.1088/0004-637X/810/1/65. arXiv:1507.07904
[astro-ph.GA].)
43. D.W. Pesce, J.A. Braatz, C.M.V. Impellizzeri, Astrophys. J. 827, 68 (2016). https://doi.org/10.
3847/0004-637X/827/1/68. ([arXiv:1604.03789])
44. D.W. Pesce, J.A. Braatz, M.J. Reid, J.J. Condon, F. Gao, C. Henkel, C.Y. Kuo, K.Y. Lo, W.
Zhao, Astrophys. J. 890, 118 (2020). https://doi.org/10.3847/1538-4357/ab6bcd
45. D.W. Pesce, J.A. Braatz, M.J. Reid, A.G. Riess, D. Scolnic, J.J. Condon, F. Gao, C. Henkel,
C.M.V. Impellizzeri, C.Y. Kuo et al., Astrophys. J. Lett. 891(1), L1 (2020). https://doi.org/10.
3847/2041-8213/ab75f0. (arXiv:2001.09213 [astro-ph.CO])
46. D.W. Pesce, J.A. Braatz, C. Henkel, E.M.L. Humphreys, C.M.V. Impellizzeri, C.-Y. Kuo, Astro-
phys. J. 948, 134 (2023). https://doi.org/10.3847/1538-4357/acc57a. ([arXiv:2302.02572])
47. M.J. Reid, J.A. Braatz, J.J. Condon, L.J. Greenhill, C. Henkel, K.Y. Lo, Astrophys. J. 695, 287–
291 (2009). https://doi.org/10.1088/0004-637X/695/1/287. (arXiv:0811.4345 [astro-ph].)
48. M.J. Reid, J.A. Braatz, J.J. Condon, K.Y. Lo, C.Y. Kuo, C.M.V. Impellizzeri, C. Henkel, Astro-
phys. J. 767, 154 (2013). https://doi.org/10.1088/0004-637X/767/2/154. (arXiv:1207.7292
[astro-ph.CO].)
49. M.J. Reid, D.W. Pesce, A.G. Riess, Astrophys. J. Lett. 886(2), L27 (2019). https://doi.org/10.
3847/2041-8213/ab552d. (arXiv:1908.05625 [astro-ph.GA].)
50. A.G. Riess, W. Yuan, L.M. Macri, D. Scolnic, D. Brout, S. Casertano, D.O. Jones, Y. Murakami,
L. Breuval, T.G. Brink et al., Astrophys. J. Lett. 934(1), L7 (2022). https://doi.org/10.3847/
2041-8213/ac5c5b. (arXiv:2112.04510 [astro-ph.CO].)
51. S. Sawada-Satoh, S. Kameno, K. Nakamura, D. Namikawa, K.M. Shibata, M. Inoue, Astrophys.
J. 680, 191 (2008). https://doi.org/10.1086/587886. (arXiv:0802.4149 [astro-ph].)
52. A. Tarchi, P. Castangia, A. Columbano, F. Panessa, J.A. Braatz, Astron. Astrophys. 532, A125
(2011). https://doi.org/10.1051/0004-6361/201117213. (arXiv:1107.5155 [astro-ph.CO].)
53. A. Yamauchi, N. Nakai, Y. Ishihara, P. Diamond, N. Sato, Publ. Astron. Soc. Jap. 64, 103
(2012). https://doi.org/10.1093/pasj/64.5.103. (arXiv:1207.6820 [astro-ph.GA].)
54. J.S. Zhang, C. Henkel, M. Kadler, L.J. Greenhill, N. Nagar, A.S. Wilson, J.A.
Braatz, Astron. Astrophys. 450, 933 (2006). https://doi.org/10.1051/0004-6361:20054138.
(arXiv:astro-ph/0512459 [astro-ph].)
55. W. Zhao, J.A. Braatz, J.J. Condon, K.Y. Lo, M.J. Reid, C. Henkel, D.W. Pesce, J.E. Greene,
F. Gao, C.Y. Kuo et al., Astrophys. J. 854, 124 (2018). https://doi.org/10.3847/1538-4357/
aaa95c. (arXiv:1801.06332 [astro-ph.GA].)
Chapter 5
Eclipsing Binary Stars as Precise
and Accurate Distance Indicators

Grzegorz Pietrzyński and Dariusz Graczyk

5.1 Introduction

Eclipsing binary stars offer a unique opportunity to measure very precisely and
accurately stellar parameters like mass, radius, etc. [21, 66]. The first precise mea-
surements were done more than hundred years ago for famous Algol [59], see also
[31] for a short historical review. Nowadays with advent of very high quality photo-
metric data from space missions and radial velocities measurement from very stable
spectrographs an amazing progress in precision of measurement of stellar parameters
has been achieved. Indeed several authors reported mass and radius measurements
with a precision close to 0.1% [12, 24, 38]. Eclipsing binary stars can be also used
to measure geometrical distances even to nearby galaxies. This makes them a per-
fect tool to calibrate the best standard candles like Cepheids, TRGB, JAGB and as a
consequence the SN Ia to determine . H0 .
In this section we will briefly describe the progress on distance determinations
with eclipsing binary stars. The detailed description of the technique and our future
plans will follow in next sections.

5.1.1 Astrometric Binaries

Before proceeding to the main subject, let’s discuss a bit the astrometric binary stars
in context of distance measurements. In general both groups of stars overlap, and

G. Pietrzyński (B)
Nicolaus Copernicus Astronomical Centre, Bartycka 18, Warszawa, Poland
e-mail: pietrzyn@camk.edu.pl
D. Graczyk
Nicolaus Copernicus Astronomical Centre, Rabiańska 8, Toruń, Poland
e-mail: darek@ncac.torun.pl

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 73
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_5
74 G. Pietrzyński and D. Graczyk

techniques used to analyse them are also similar. When the spectral lines from both
components can be detected the radial velocity of the orbital motion around the centre
of mass of both components can be measured providing the spectroscopic orbit of
the system. Together with the apparent orbit from astrometry (done with speckle or
long-base interferometry) we can measure very precisely stellar masses and distance
.d using the following equations [66]:

1.036149 × 10−7 (K 1 + K 2 )2 K 2 P(1 − e2 )3/2

. M1 =
sin3 i
−7
1.036149 × 10 (K 1 + K 2 )2 K 1 P(1 − e2 )3/2
M2 =
sin3 i
−5
9.191940 × 10 (K 1 + K 2 )P(1 − e2 )1/2
aAU =
sin i
aAU
d = '' (5.1)
a
where the masses . M1 , . M2 are expressed in solar units, the distance .d in parsecs,
semi amplitudes of radial velocities . K 1 and . K 2 in km s.−1 , the period . P in days,
and the apparent semi-major axis .a '' in arcseconds. The parameter .aAU is the linear
semi-major axis expressed in astronomical units.
Indeed from the fit of the spectroscopic orbit we get the linear size of the semi
major axis while astrometric orbit gives its projection on the sky.
This very powerful technique has been applied several times to measure geometric
distances [7, 13, 40, 65, 67]. A 1% precision in distance measurement level was
obtained with this method more than a decade ago. The limiting factor was the
calibration of the interferometric data. Recently Gallenne et al. [12] published ultra-
precise distances to ten binary systems. Observations obtained with ESO instruments
GRAVITY allowed to reach 0.02% accuracy in wavelength calibration and together
with spectra obtained with UVES spectrograph resulted in distances precise to 0.2%
on average.
However this method is limited to the systems that can be spatially resolved.
Even when using the very high resolution observations from the state-of-the-art
interferometric instruments, and studying systems with periods that allow to complete
data in a reasonable amount of time (e.g. 1 year or so) the range of this technique is
limited to about 200 pc (Fig. 5.1).

5.1.2 Eclipsing Binaries

In order to reach nearby galaxies and use the full potential of binary systems to
calibrate the extragalactic distance scale we have to use eclipsing binary systems.
The huge potential of eclipsing binaries as precise and accurate distance indicators
5 Eclipsing Binary Stars as Precise and Accurate Distance Indicators 75

Fig. 5.1 Spectroscopic and astrometric orbits of a binary system HD 188088. Based on these data
the distance of 14.1129.±0.0038 pc was measured

was recognized a long time ago [25, 32, 43]. However as usual the devil is in the
details and several different problems prevented astronomers to measure precise
geometrical distances with detached eclipsing binaries.
First of all it is not possible to measure distances with eclipsing binaries based on
solely results of analysis of the photometric light curves and spectroscopic orbit. An
additional information is needed. The most common approach applied to measure
extragalactic distances was to determine the intrinsic flux of each eclipsing binary
component using bolometric correction and effective temperature determinations
and then calculate a distance:

d [pc] = 3.360 Ri Ti2 10 0.2(BCi +Vi )

. i (5.2)

where .Vi is an observed, extinction corrected .V -band magnitude of .i-th component

and . BC is the bolometric correction. This way of measuring distances makes the
whole method sensitive to significant errors on effective temperature determination.
One way of dealing with this difficulty is to fit spectrophotometry using stellar atmo-
spheric models [26], however that makes the procedure strongly model-dependent.
Another complication to apply eclipsing binary method to measure distances to
nearby galaxies is a limited number of good systems. The most common systems
identified there are bright, massive close EBs of late O-type or early B-type. Very
often they are multiple systems and they show complication in light curves due to a
presence of strong stellar winds, circumstellar matter, thin accretion disc, pulsations
or spot activity. In spite of this situations substantial progress was achieved and
several projects reported distance determinations to nearby galaxies using early type
EBs in the LMC [4, 8, 54], M33 [55, 74], and M31 [3]. All of them are subject of
76 G. Pietrzyński and D. Graczyk

rather unknown systematic errors related to adopted stellar atmospheric models and
strong degeneracy between effective temperature and reddening.
Another way of using eclipsing binary systems to measure distances is to use a
very well established relationship between the surface brightness and color (SBCR).
In this approach the angular diameters of the stars, components of the binary system
comes from the surface - brightness (see more details in the next paragraph) while
their linear dimensions come from modelling of the light curve and radial velocity
curve.
The big complication in applying this method to measure extragalactic distances
is that the SBCR is well calibrated for late type stars only (with spectral types later
than middle A-type). Late typed dwarfs are intrinsically to faint to secure high-
resolution spectra for such system even in the LMC. Fortunately the Magellanic
Clouds have been monitored for more than 25 years by the Optical Gravitational
Lensing Experiment. These unique dataset allowed to discover long period eclipsing
binaries composed of helium burning giants in both LMC and SMC [45]. Such
objects are perfectly suited for modeling. In the course of the Araucaria project
([49]) we started to observe these targets in order to provide high-resolution spectra,
near-infrared photometry and improve the optical light curves.
The first results published for one of these systems [46] fully confirmed big
prospects of this method. The first precise distance determinations with our late-typed
systems to the LMC [47] and SMC [19], with precision of 2% and 3%, respectively
provided the best zero points for calibration of the standard candles. These determi-
nations were limited mostly (60% of the total error) by the systematic errors related
to the SBCR. Therefore we dedicated a lot of efforts to improve on that. With new
precision near-infrared photometry [33] and interferometric angular diameter deter-
minations [14] obtained for a sample of nearby red clump stars (e.g. stars very similar
to the components of our eclipsing binaries in the Magellanic Clouds) we established
much better the SBCR. Based on it and studying more systems we finally provided
the best so far distances to the LMC (.49.50 ± 0.09 (statistical) .±0.54 (systematic)
kpc [48] and SMC (.62.44 ± 0.47 (st.) .±0.81 (sys.) kpc [22]. These results clearly
show that the eclipsing binary systems can provide distances precise and accurate
to about 1%. However our results are limited again by the precision of the SBCR
(about 0.8% systematic contribution to the total error on distance).

5.2 Methods Details

The eclipsing binary method is combining information derived from spectroscopy

and photometry. In this regard it is similar to Baade-Wesselink method of distance
determination to radially pulsating stars [2, 76]. The method utillises a concept of the
stellar surface brightness [77] and empirical calibrations of the surface brightness—
color relations (SBCRs) in order to avoid systematic uncertainties arising from a use
of bolometric corrections or absolute fluxes predicted by stellar atmospheric models
[32]. Currently the most precise SBCRs are calibrated against optical-infrared colors
5 Eclipsing Binary Stars as Precise and Accurate Distance Indicators 77

(.V − K ) and (. B − K ) and they allow to calculate angular diameters of stars with a
precision of 1% or even slightly better e.g. [23, 30, 48, 58].
The simplicity of the method is actually its largest advantage: to derive a distance
.d to a star it is used a basic geometrical relation between a physical radius . R and an

angular diameter .φ:

R
.d [pc] = 9.301 (5.3)
φ

where . R is expressed in the solar nominal radii [52], .φ in miilliarcseconds and the
conversion factor is equal .2RʘN
/1 AU.
Determination of radii of eclipsing binary components is the most critical and
time consuming part of the method. It consists of several steps:
1. Selection of suitable eclipsing binary candidates, usually based on light curves
from large photometric surveys.
2. Spectroscopic follow up of the best candidates.
3. Securing near-infrared photometry in . J and . K band.
4. Combined analysis of photometry and spectroscopy using a well-tested software.

5.2.1 Candidate EBs

For the distance determination to the Magellanic Clouds we searched published

catalogs of eclipsing binary stars [70, 80], based on the OGLE project photometry
[69], to find good candidates. Moreover we developed our own method to identify
detached eclipsing binary stars from ground-based non continuous photometry [17]
and we employed it to data from the third phase of OGLE project [68] in order to
find more candidates [18, 44].
In order to retain a geometrical simplicity and to avoid additional corrections an
eclipsing binary should have a number of specific properties. Consequently only a
small fraction of eclipsing binaries are actually suited for the method. The near-
ideal system is a double-lined, well detached eclipsing binary (DEB), thus both
components are seen in observed spectra and stars can be safely treated as spheres.
Components should be photometrically quite avoiding this way complications intro-
duced by stellar spots, winds, flares, etc. The orbital inclination .i should be close
to .90◦ resulting in total eclipses or deep both eclipses (.∆m > 0.4 mag) making the
determination of radii more unequivocal. Triple or multiple systems are disfavoured
unless a light contribution of additional components to the total system’s light is neg-
ligible, eventually it is small (of order of several per cents) and very well constrained.
Finally, candidate DEBs with smallest interstellar extinction are preferred.
For the Magellanic Clouds even more constrains were needed. The systems should
have sufficient brightness for the spectroscopic follow up with 4m-8m class tele-
scopes. However, at the same time, they should consist of components of late spec-
78 G. Pietrzyński and D. Graczyk

tral type in order to use high precision SBCRs. The only choice, up to now, is to use
eclipsing binary stars consisting of late-type giants of F,G and K type. Such eclipsing
binaries fulfilling all described above criteria are extremely rare because their orbital
periods are of order of 100–1000 days. We found only 40 suitable systems in the
LMC and 18 in the SMC, while total number of known eclipsing binaries in both
galaxies are .∼40000 and .∼8400 in the LMC and the SMC, respectively [45].

5.2.2 Spectroscopic Follow-Up

The spectroscopic follow up have a number of objectives. They are:

• Obtaining radial velocity curves of both components.

• Search for eventual multiplicity of a DEB.
• Determination of spectroscopic light ratio.
• The atmospheric analysis of disentangled spectra of components.
• Determination of the interstellar extinction.

To determine precise radial velocities using spectra of binary stars a number of

methods were developed. The cross correlation technique [64] based on comparison
of an observed spectrum with a template spectrum is widely used, as well as its
extension to two-dimensions – TODCOR [81]. Very good results, especially in cases
when stellar lines are blended or S/N of spectra is small, gives another method which
is based on a determination of the spectral line broadening function (BF) [56]. We
used this method successfully to measure radial velocities of DEBs in the Magellanic
Clouds with a dedicated software RaVeSpAn [50]. For targets as faint as 17–18 mag
in V band we could extract radial velocities with a precision better than 400 m s.−1
for spectra with S/N .< 10 [20, 22].
When a third set of lines was observed in a spectrum or the systemic radial
velocity showed clear and large radial velocity trend a DEB was discarded from
the sample. However we found only two such cases in our full sample of systems.
From observed spectra we determined the light ratio of the components in a specific
photometric band, usually .V -band, by measuring the strength of absorption lines of
both components. Of course one needs to correct it for the temperature difference
of components, and that information comes from light curve modelling. The light
ratio helps constrain the ratio of components’ radii in case of partial and relatively
shallow eclipses eclipses (.∆m ∼ 0.25–0.5 mag).
The whole set of observed spectra for a particular system was separated into
single spectra of individual components using the iterative method [15]. Resulting
components’ spectra were analysed in order to determine atmospheric parameters
like the surface temperature and the metallicity. The former is important during
light curve modeling to constrain the limb darkening coefficients. The latter allows
for more precise comparison of components with the stellar evolution models and
determination of the age of a system.
5 Eclipsing Binary Stars as Precise and Accurate Distance Indicators 79

Precise determination of the interstellar extinction in direction of each eclipsing

binary reduces significantly errors on distances. We employed three methods to
calculate the reddening. First was based on spectra and used a calibration between
reddening and equivalent width of interstellar sodium absorption lines NaI (5890.0,
5895.9Å)[42, 51]. Second method used color - effective temperature calibrations like
e.g. [53, 79] to calculate intrinsic colors of components. The effective temperature
are derived from atmospheric analysis of decomposed spectra. Simply comparison of
observed and intrinsic colors allow for calculation of the interstellar reddening. And
finally we used the reddening maps in the Magellanic Clouds [16, 27]. In order to
diminish systematic uncertainties of each method we calculated an average reddening
from all above methods.

5.2.3 The Infrared Photometry

Securing high quality near-infrared photometry is vital for the method, because most
precise SBCR are calibrated for (.V − K ) color. For every system we collected multi-
epoch . J ,. K photometry with SOFI camera installed on 3.5m NNT telescope on La
Silla. Resulting precision of . J and . K magnitudes were about 0.01 mag or better.
Finally we converted magnitudes into 2MASS system [60] in order to have photom-
etry which is consistent with that obtained by Laney et al. [33].

5.2.4 Combined Analysis of Photometry and Velocimetry

Details of modelling of eclipsing binary giant stars in the Magellanic Clouds are
given in [20, 22]. Here we give only a brief description. Light curves and radial
velocity curves were analysed simultaneously with the Wilson-Devinney program
[78] and also with JKTEBOP code [61]. An extensive modelling was performed for
every system. To validate consistency of a model we compared the spectroscopic light
and temperature ratios with model predictions. In most cases consistency was fully
satisfactory. We put a lot of attention to constrain the third light which can origin
from a blend or an unresolved physical companion star. In some cases we could
strengthen the analysis by inclusion of additional . B (blue) and . R (red) light curves
from MACHO project [1]. Modelling multi-band photometry has an advantage to
better constrain the temperature ratio of components and the third light.
We used two different modelling codes in order to find the best model and to
investigate eventual discrepancies in the model parameters. JKTEBOP served as
auxiliary code to confirm results obtained with the WD code as well as to calculate
statistical errors of model parameters with the Monte Carlo method. These errors
were compared with formal errors returned by the WD code, and always the larger
one was chosen as the final uncertainty on a parameter. Significant discrepancies
between parameters returned by the WD and JKTEBOP were sporadic and limited
80 G. Pietrzyński and D. Graczyk

to cases where components were significantly distorted due to tidal forces; then
parameters and their errors were adopted after the WD code.

5.2.5 Error Analysis

We followed [9, 71] in dividing the distance error into so called the statistical and the
systematic parts. We used this division in most of our papers reporting distance mea-
surements. That was strongly motivated by a need to underline a correlation between
derived individual eclipsing binary distances and also the zero-point uncertainty of
the SBCR used. In other words there is a “systematic” part of the total error which
does not reduce while enlarging a sample of eclipsing binary stars used to measure
a distance to a particular object. This part of the error is an immanent feature of the
method and only a modification/improvement of the method can change/decrease it.
This nomenclature is a bit unfortunate, because the phrase “systematic error” has
in general a different meaning in physical sciences. It is a “consistent, repeatable error
(bias) associated with faulty equipment or a flawed experiment design” leading to a
permanent shift of a measured quantity (e.g. a distance) from the true one. Once such
error is identified an instrument/method can be corrected for it and one can pretty
much forget about it. However, we had something else in mind: not that the method
we used is somehow “flawed”, but rather that its accuracy is limited because of a
number of slight but still not fully recognised effects. In our opinion a good measure
of this “unknown” systematic is a standard deviation (a spread) of the SBCR fit used
by us [48] because the uncertainty on the SBCR dominates the total error budget of
our eclipsing binary distances.
The statistical part of the LMC distance error was found simply by the Monte Carlo
simulations. We draw 10 000 times from individual distance measurements within
errors assuming their Gaussian distribution. For each draw we computed the mean
distance and geometrical orientation of the disc (the inclination to the line-of-sight,
the longitude of nodes) and we made a histogram of the results.
In case of the SMC situation was more difficult because of the large extension of
this galaxy in a direction of view and a significant distortion due to gravitational inter-
action with the LMC and the Milky Way. In that case statistical error was determined
by analysing distance trends of individual targets in two perpendicular directions
within the galaxy [22].
5 Eclipsing Binary Stars as Precise and Accurate Distance Indicators 81

5.3 Toward Sub-percent Distances Within the Local Group

5.3.1 New Calibrations of SBCR

5.3.1.1 SPICA

CHARA/SPICA project [41] aims at precise measuring angular diameters of about

1000 stars with long-base optical interferometer on Mount Wilson observatory.
Non-variable and non-multiple stars brighter than 10 mag in V-band were carefully
selected to cover all spectral types. The main advantage of this project is a homoge-
nous way all angular diameters will determined minimising this way systematic
effects when angular diameters obtained with different instruments are combined.
The angular diameters will be used to extract modern and accurate SBCRs over very
wide range of colors, metallicities and the surface gravities.
The novelty of project is a use of a near-infrared fringe tracker allowing for longer
exposures of the fringe signals in the visible. This way SPICA has higher sensitivity
and will allow to measure angular diameters as small as 0.1–0.2 mas even for faintest
stars in the sample. The resulting multi-parameter SBCRs are expected to reach sub-
percent precision.

5.3.1.2 Nearby DEBs

Another, independent route to improve significantly SBCRs provide eclipsing binary

stars. By reversing the eclipsing binary method one can derive angular diameters of an
eclipsing binary components and calibrate the SBCR, provided a distance is known.
We selected over 100 suitable, well detached eclipsing binary stars with spectral types
from mid B-type to early K-type. In the analysis we use superb quality light curves
from space mission Kepler K2 [28] and TESS [57] to find photometric solutions
with the WD code. The spectroscopic follow up was performed with the HARPS
spectrograph [39] well known for its high precision performance. As a result we can
derive absolute radii of components usually with a precision much better 1%.
The Gaia satellite mission [10, 11] resulted in a tremendous improvement in a
number of stars with precisely measured trigonometric parallaxes. Although there is
still a debate on how accurate are Gaia parallaxes from latest DR3 (what is called a
parallax bias) after correcting them using a general scheme proposed by Lindegren
et al. [36] they have only very small zero-point parallax offset of order of several .μas
[6, 75]. For nearby eclipsing binary stars (d .< 200 pc) such systematic uncertainty
is practically insignificant and one can almost safely use Gaia parallaxes. We write
“almost” because Gaia see photocenter movement of nearby EBs. For overwhelming
majority of them Gaia DR3 provide only five- or six-parameter solutions what may,
in principle, lead to some bias when an amplitude of photocenter movements is
comparable or larger than a parallax. However, full 9-parameter solutions accounting
for photocenter movements will likely be published by Gaia DR4.
82 G. Pietrzyński and D. Graczyk

5.3.1.3 Early Type DEBs in the LMC

There are practically no nearby well detached eclipsing binary stars containing com-
ponents of O-type or early B-type. Those which are at larger distances of about
0.5–1.5 kpc lie close to the galactic plane and suffer from significant reddening. As
a consequence we can’t follow the same route like in case of EBs of later spectral
types.
However, there is another possibility to calibrate precise SBCRs for hottest stars
using EBs. The distance to the LMC was firmly established with late-type eclipsing
binary stars. On the other hand there is a large number of early type EBs identified
in both Magellanic Clouds [45] from which we choose about 20 best candidates.
The systems are bright, well detached, little reddened and have good quality light
curves. High precision spectroscopic follow-up of them with UVES spectrograph
was performed in order to derive their fundamental parameters. First results for two
eclipsing binaries were already published [62, 63]. We expect to obtain this way
precise radii for about 40 early type stars in the LMC. Individual distances of stars
follow from their positions within the disc of the LMC, the geometry of which is
known [48, 72]. The resulting angular diameters would define precise SBCRs for
this important type of stars.

5.3.2 Distances to More Local Group Galaxies

In order to precisely calibrate standard candles like Cepheids, TRGB etc., it is crucial
to provide geometrical distances to several galaxies having a wide range of metal-
licities. Unfortunately, the eclipsing binaries containing late-type giant stars which
were used successfully to determine distances to the Magellanic Clouds are rather
too faint. The brightest such systems containing bright giants had combined absolute
magnitude . MV in a range from .−3.0 up to .−3.5 mag. That translates into, neglecting
the reddening, observed .V magnitudes of 21.4–20.9 mag in M31 [35] and 21.6–21.1
mag in M33 [5]. Note that individual components will be even .∼0.7 mag fainter!
Given their long orbital period it will be very challenging to detect such systems at a
distance of M33, and then to perform follow up observations of such a faint targets.
The precise SBCRs calibrated for massive and hot stars, see Sects. 5.3.1.1 and
5.3.1.3, could be used to determine direct near-geometric distances to more distant
galaxies within the Local Group like M31, M33, NGC 6822, IC 1613 etc. A number
of candidate detached EBs comprising very early type components were identified
in M31 and M33 by DIRECT project [29, 37], in M31 by [73] and in M31 by
Pan-STARRS [34]. Three systems with eccentric orbits were selected in M31 for
spectroscopic follow-up with 8-m class telescopes, based on high brightness, quality
of light curves and low reddening. More effort is required to provide fully empirical
distanced based on DEBs for M31 and M33. While in other galaxies dedicated
surveys should be performed to select good systems first (Fig. 5.2).
5 Eclipsing Binary Stars as Precise and Accurate Distance Indicators 83

Fig. 5.2 Range and precision of the geometrical methods. The range of distances that can be
measured with astrometric binaries, Gaia parallaxes, eclipsing binaries together with the surface
brightness—colour relation (SBCR), the Cepheid P-L relation and type Ia supernovae are marked
with the corresponding bars. Astrometric binaries, Gaia parallaxes, and eclipsing binaries with the
SBCR can be used to calibrate the Cepheid P-L relation, TRGB, JAGB, and as a result the supernova
Ia brightness. As can be appreciated, eclipsing binaries together with our new relation offer us the
opportunity to determine distances competitive in precision to Gaia already at about 1 kpc from the
Sun, while retaining their high precision for distances up to 1 Mpc. The lower bar divided into three
sections shows the range of distances within our Galaxy, Local Group, and more distant galaxies.
The vertical grey line marks the distance to the LMC

Acknowledgements The research leading to these results has received funding from the Euro-
pean Research Council (ERC) under the European Unions Horizon 2020 research and innova-
tion program (grant agreement No. 951549). We also acknowledge support from the National
Science Center, Poland grants MAESTRO UMO-2017/26/A/ ST9/00446, BEETHOVEN UMO-
2018/31/G/ST9/03050, and DIR/WK/2018/09 and 2024/WK/02 grants of the Polish Ministry of
Education and Science and Polish Ministry of Science and Higher Education.

References

1. C. Alcock et al., [MACHO], Astrophys. J. 486, 697–726 (1997). https://doi.org/10.1086/

304535. (arXiv:astro-ph/9606165 [astro-ph])
2. W. Baade, Astron. Nachr. 228, 359–362 (1926)
3. A.Z. Bonanos, K.Z. Stanek, R.P. Kudritzki, L.M. Macri, D.D. Sasselov, J. Kaluzny, P.B. Stetson,
D. Bersier, F. Bresolin, T. Matheson et al., Astrophys. J. 652, 313–322 (2006). https://doi.org/
10.1086/508140. (arXiv:astro-ph/0606279 [astro-ph])
84 G. Pietrzyński and D. Graczyk

4. A.Z. Bonanos, N. Castro, L.M. Macri, R.P. Kudritzki, Astrophys. J. Lett. 729, L9–L14 (2011).
https://doi.org/10.1088/2041-8205/729/1/L9. (arXiv:1101.1953 [astro-ph.SR])
5. L. Breuval, A.G. Riess, L.M. Macri, S. Li, W. Yuan, S. Casertano, T. Konchady, B. Trahin,
M.J. Durbin, B.F. Williams, Astrophys. J. 951, 118 (2023). https://doi.org/10.3847/1538-4357/
acd3f4. ([arXiv:2304.00037])
6. M. Cruz Reyes, R.I. Anderson, Astron. Astrophys. 672, A85 (2023). https://doi.org/10.1051/
0004-6361/202244775. ([arXiv:2208.09403])
7. J.A. Docobo, M. Andrade, Mon. Not. R. Astron. Soc. 428, 321–339 (2013). https://doi.org/10.
1093/mnras/sts045
8. E.L. Fitzpatrick, I. Ribas, E.F. Guinan, F.P. Maloney, A. Claret, Astrophys. J. 587, 685–700
(2003). https://doi.org/10.1086/368309. (arXiv:astro-ph/0301296 [astro-ph])
9. W.L. Freedman et al., [HST], Astrophys. J. 553, 47–72 (2001). https://doi.org/10.1086/320638.
(arXiv:astro-ph/0012376 [astro-ph])
10. Gaia Collaboration, A.G.A. Brown, A. Vallenari, T. Prusti, J.H.J. de Bruijne, F. Mignard, R.
Drimmel, C. Babusiaux, C.A.L. Bailer-Jones, U. Bastian et al., Astron. Astrophys. 595, A2
(2016). https://doi.org/10.1051/0004-6361/201629512. ([arXiv:1609.04172])
11. Gaia Collaboration, A. Vallenari, A. G. A. Brown, T. Prusti, J. H. J. de Bruijne, F. Arenou, C.
Babusiaux, M. Biermann, O. L. Creevey, C. Ducourant, et al., Astronomy and Astrophysics
674, A1 (2023) https://doi.org/10.1051/0004-6361/202243940 [arXiv:2208.00211]
12. A. Gallenne, A. Mérand, P. Kervella, D. Graczyk, G. Pietrzyński, W. Gieren, B.
Pilecki, Astron. Astrophys. 672, A119 (2023). https://doi.org/10.1051/0004-6361/202245712.
([arXiv:2302.12960])
13. A. Gallenne, G. Pietrzyński, D. Graczyk, P. Konorski, P. Kervella, A. Mérand, W. Gieren, R.I.
Anderson, S. Villanova, Astron. Astrophys. 586, A35 (2016). https://doi.org/10.1051/0004-
6361/201526764. ([arXiv:1511.07971])
14. A. Gallenne, G. Pietrzyński, D. Graczyk, N. Nardetto, A. Mérand, P. Kervella, W. Gieren, S.
Villanova, R.E. Mennickent, B. Pilecki, Astron. Astrophys. 616, A68 (2018). https://doi.org/
10.1051/0004-6361/201833341. ([arXiv:1806.09572])
15. J.F. González, H. Levato, Astron. Astrophys. 448, 283–292 (2006). https://doi.org/10.1051/
0004-6361:20053177
16. M. Górski, B. Zgirski, G. Pietrzyński, W. Gieren, P. Wielgórski, D. Graczyk, R.-P. Kudritzki,
B. Pilecki, W. Narloch, P. Karczmarek et al., Astrophys. J. 889, 179 (2020). https://doi.org/10.
3847/1538-4357/ab65ed. ([arXiv:2001.08242])
17. D. Graczyk, L. Eyer, Acta Astron. 60, 109–119 (2010)
18. D. Graczyk, I. Soszynski, R. Poleski, G. Pietrzynski, A. Udalski, M.K. Szymanski, M. Kubiak,
L. Wyrzykowski, K. Ulaczyk, Acta Astron. 61, 103 (2011). (arXiv:1108.0446 [astro-ph.SR])
19. D. Graczyk, G. Pietrzynski, I.B. Thompson, W. Gieren, B. Pilecki, P. Konorski, A. Udalski,
I. Soszynski, S. Villanova, M. Górski et al., Astrophys. J. 780, 59 (2014). https://doi.org/10.
1088/0004-637X/780/1/59. (arXiv:1311.2340 [astro-ph.CO])
20. D. Graczyk, G. Pietrzyński, I.B. Thompson, W. Gieren, B. Pilecki, P. Konorski, S. Villanova,
M. Górski, K. Suchomska, P. Karczmarek et al., Astrophys. J. 860, 1 (2018). https://doi.org/
10.3847/1538-4357/aac2bf. ([arXiv:1805.04952])
21. D. Graczyk, G. Pietrzyński, W. Gieren, J. Storm, N. Nardetto, A. Gallenne, P.F.L. Maxted, P.
Kervella, Z. Kołaczkowski, P. Konorski et al., Astrophys. J. 872, 85 (2019). https://doi.org/10.
3847/1538-4357/aafbed. ([arXiv:1902.00589])
22. D. Graczyk, G. Pietrzyński, I.B. Thompson, W. Gieren, B. Zgirski, S. Villanova, M. Górski, P.
Wielgórski, P. Karczmarek, W. Narloch et al., Astrophys. J. 904, 13 (2020). https://doi.org/10.
3847/1538-4357/abbb2b. ([arXiv:2010.08754])
23. D. Graczyk, G. Pietrzyński, C. Galan, W. Gieren, A. Tkachenko, R.I. Anderson, A. Gallenne,
M. Górski, G. Hajdu, M. Kałuszyński et al., Astron. Astrophys. 649, A109 (2021). https://doi.
org/10.1051/0004-6361/202140571. ([arXiv:2103.02077])
24. D. Graczyk, G. Pietrzyński, C. Galan, J. Southworth, W. Gieren, M. Kałuszyński, B. Zgirski,
A. Gallenne, M. Górski, G. Hajdu et al., Astron. Astrophys. 666, A128 (2022). https://doi.org/
10.1051/0004-6361/202244122. ([arXiv:2208.07257])
5 Eclipsing Binary Stars as Precise and Accurate Distance Indicators 85

25. E.F. Guinan, D.H. Bradstreet, L.E. Dewarf, The Origins, Evolution, and Destinies of Binary
Stars in Clusters, vol. 90, p. 196 (1996)
26. E.F. Guinan, E.L. Fitzpatrick, L.E. DeWarf, F.P. Maloney, P.A. Maurone, I. Ribas, A. Gimenz,
J.D. Pritchard, D.H. Bradstreet, Astrophys. J. Lett. 509, L21–L24 (1998). https://doi.org/10.
1086/311760. (arXiv:astro-ph/9809132 [astro-ph])
27. R. Haschke, E.K. Grebel, S. Duffau, Astron. J. 141, 158 (2011). https://doi.org/10.1088/0004-
6256/141/5/158. (arXiv:1104.2325 [astro-ph.GA])
28. S.B. Howell, C. Sobeck, M. Haas, M. Still, T. Barclay, F. Mullally, J. Troeltzsch, S. Aigrain,
S.T. Bryson, D. Caldwell et al., Publ. Astron. Soc. Pacific 126, 398 (2014). https://doi.org/10.
1086/676406. ([arXiv:1402.5163])
29. J. Kaluzny, K.Z. Stanek, M. Krockenberger, D.D. Sasselov, J.L. Tonry, M. Mateo, Astron. J.
115, 1016 (1998). https://doi.org/10.1086/300235. (arXiv:astro-ph/9703124 [astro-ph])
30. P. Kervella, F. Thevenin, E. Di Folco, D. Segransan, Astron. Astrophys. 426, 297–307 (2004).
https://doi.org/10.1051/0004-6361:20035930. (arXiv:astro-ph/0404180 [astro-ph])
31. A. Kruszewski, I. Semeniuk, Acta Astron. 49, 561 (1999). (arXiv:astro-ph/9912270 [astro-ph])
32. C.H. Lacy, Astrophys. J. 213, 458–463 (1977). https://doi.org/10.1086/155176
33. C.D. Laney, M.D. Joner, G. Pietrzynski, Mon. Not. R. Astron. Soc. 419, 1637 (2012). https://
doi.org/10.1111/j.1365-2966.2011.19826.x. (arXiv:1109.4800 [astro-ph.SR])
34. C.H. Lee, J. Koppenhoefer, S. Seitz, R. Bender, A. Riffeser, M. Kodric, U. Hopp, J. Snigula, C.
Gossl, R.P. Kudritzki et al., Astrophys. J. 797, 22 (2014). https://doi.org/10.1088/0004-637X/
797/1/22. ([arXiv:1411.1115])
35. S. Li, A.G. Riess, M.P. Busch, S. Casertano, L.M. Macri, W. Yuan, Astrophys. J. 920, 84 (2021).
https://doi.org/10.3847/1538-4357/ac1597. ([arXiv:2107.08029])
36. L. Lindegren, U. Bastian, M. Biermann, A. Bombrun, A. de Torres, E. Gerlach, R. Geyer, J.
Hernández, T. Hilger, D. Hobbs et al., Astron. Astrophys. 649, A4 (2021). https://doi.org/10.
1051/0004-6361/202039653. ([arXiv:2012.01742])
37. L.M. Macri, K.Z. Stanek, D.D. Sasselov, M. Krockenberger, J. Kaluzny, Astron. J. 121, 870
(2001). https://doi.org/10.1086/318773. (arXiv:astro-ph/0102123 [astro-ph])
38. P.F.L. Maxted, P. Gaulme, D. Graczyk, K.G. Hełminiak, C. Johnston, J.A. Orosz, A. Prša, J.
Southworth, G. Torres, G.R. Davies et al., Mon. Not. R. Astron. Soc. 498, 332–343 (2020).
https://doi.org/10.1093/mnras/staa1662. ([arXiv:2003.09295])
39. M. Mayor, F. Pepe, D. Queloz, F. Bouchy, G. Rupprecht, G. Lo Curto, G. Avila, W. Benz, J.-L.
Bertaux, X. Bonfils et al., The Messenger 114, 20–24 (2003)
40. H.A. McAlister, Publ. Astron. Soc. Pac. 88, 957–959 (1976). https://doi.org/10.1086/130053
41. D. Mourard, P. Berio, C. Pannetier, N. Nardetto, F. Allouche, C. Bailet, J. Dejonghe, P. Geneslay,
E. Jacqmart, S. Lagarde et al., Optical and infrared interferometry and imaging VIII, vol. 12183,
1218308 (2022). https://doi.org/10.1117/12.2628881. ([arXiv:2210.09096])
42. U. Munari, T. Zwitter, Astron. Astrophys. 318, 269–274 (1997)
43. B. Paczynski, The Extragalactic Distance Scale 273 (1997)
44. M. Pawlak, D. Graczyk, I. Soszyński, P. Pietrukowicz, R. Poleski, A. Udalski, M.K. Szymański,
M. Kubiak, G. Pietrzyński, Ł. Wyrzykowski et al., Acta Astronomica 63, 323–338 (2013).
https://doi.org/10.48550/arXiv.1310.3272. ([arXiv:1310.3272])
45. M. Pawlak, I. Soszyński, A. Udalski, M.K. Szymański, Ł. Wyrzykowski, K. Ulaczyk, R.
Poleski, P. Pietrukowicz, S. Kozłowski, D.M. Skowron et al., Acta Astronomica 66, 421–432
(2016). https://doi.org/10.48550/arXiv.1612.06394. ([arXiv:1612.06394])
46. G. Pietrzynski, I. Thompson, D. Graczyk, W. Gieren, A. Udalski, O. Szewczyk, D. Minniti, Z.
Kolaczkowski, F. Bresolin, R.P. Kudritzki, Astrophys. J. 697, 862 (2009). https://doi.org/10.
1088/0004-637X/697/1/862. (arXiv:0903.0855 [astro-ph.CO])
47. G. Pietrzyński, D. Graczyk, W. Gieren, I.B. Thompson, B. Pilecki, A. Udalski, I. Soszyński,
S. Kozłowski, P. Konorski, K. Suchomska et al., Nature 495, 76–79 (2013). https://doi.org/10.
1038/nature11878. (arXiv:1303.2063 [astro-ph.GA])
48. G. Pietrzyński, D. Graczyk, A. Gallenne, W. Gieren, I.B. Thompson, B. Pilecki, P. Karczmarek,
M. Górski, K. Suchomska, M. Taormina et al., Nature 567, 200–203 (2019). https://doi.org/
10.1038/s41586-019-0999-4. ([arXiv:1903.08096])
86 G. Pietrzyński and D. Graczyk

49. G. Pietrzyński, W. Gieren, P. Karczmarek et al., Book 1, 1–123 (2023). https://doi.org/10.

48550/arXiv.2305.17247. ([arXiv:2305.17247])
50. B. Pilecki, P. Konorski, M. Gorski, From interacting binaries to exoplanets: essential modeling
tools 282, 301–302 (2012). https://doi.org/10.1017/S174392131102761X
51. D. Poznanski, J.X. Prochaska, J.S. Bloom, Mon. Not. Roy. Astron. Soc. 426, 1465 (2012).
https://doi.org/10.1111/j.1365-2966.2012.21796.x. (arXiv:1206.6107 [astro-ph.IM])
52. A. Prsa, P. Harmanec, G. Torres, E. Mamajek, M. Asplund, N. Capitaine, J. Christensen-
Dalsgaard, É. Depagne, M. Haberreiter, S. Hekker et al., Astron. J. 152, 41 (2016). https://doi.
org/10.3847/0004-6256/152/2/41. ([arXiv:1605.09788])
53. I. Ramirez, J. Melendez, Astrophys. J. 626, 465–485 (2005). https://doi.org/10.1086/430102.
(arXiv:astro-ph/0503110 [astro-ph])
54. I. Ribas, E.L. Fitzpatrick, F.P. Maloney, E.F. Guinan, A. Udalski, Astrophys. J. 574, 771–782
(2002). https://doi.org/10.1086/341003 (arXiv:astro-ph/0204061 [astro-ph])
55. I. Ribas, C. Jordi, F. Vilardell, E.L. Fitzpatrick, R.W. Hilditch, E.F. Guinan, Astrophys. J. Lett.
635, L37–L40 (2005). https://doi.org/10.1086/499161. (arXiv:astro-ph/0511045 [astro-ph])
56. S.M. Rucinski, Astron. J. 104, 1968 (1992). https://doi.org/10.1086/116372
57. G.R. Ricker, J.N. Winn, R. Vanderspek, D.W. Latham, G.Á. Bakos, J.L. Bean, Z.K. Berta-
Thompson, T.M. Brown, L. Buchhave, N.R. Butler et al., J. Astron. Telesc. Instrum. Syst. 1,
014003 (2015). https://doi.org/10.1117/1.JATIS.1.1.014003
58. A. Salsi, N. Nardetto, D. Mourard, D. Graczyk, M. Taormina, O. Creevey, V. Hocdé, F. Morand,
K. Perraut, G. Pietrzynski et al., Astron. Astrophys. 652, A26 (2021). https://doi.org/10.1051/
0004-6361/202140763. ([arXiv:2106.01073])
59. F. Schlesinger, R.H. Curtiss, Publ. Allegheny Obs. Univ. Pittsburgh 1, 25–32 (1910)
60. M.F. Skrutskie et al., 2MASS. Astron. J. 131, 1163–1183 (2006). https://doi.org/10.1086/
498708
61. J. Southworth, Mon. Not. R. Astron. Soc. 386, 1644 (2008). https://doi.org/10.1111/j.1365-
2966.2008.13145.x. (arXiv:0802.3764 [astro-ph].)
62. M. Taormina, R.-P. Kudritzki, J. Puls, B. Pilecki, E. Sextl, G. Pietrzyński, M.A. Urbaneja,
W. Gieren, Astrophys. J. 890, 137 (2020). https://doi.org/10.3847/1538-4357/ab6bd0.
([arXiv:2001.04762].)
63. M. Taormina, G. Pietrzyński, B. Pilecki, R.-P. Kudritzki, I.B. Thompson, D. Graczyk, W.
Gieren, N. Nardetto, M. Górski, K. Suchomska et al., Astrophys. J. 886, 111 (2019). https://
doi.org/10.3847/1538-4357/ab4b57. ([arXiv:1910.08111])
64. J. Tonry, M. Davis, Astron. J. 84, 1511 (1979). https://doi.org/10.1086/112569
65. G. Torres, A.S.P. Conf. Ser. 318, 123 (2004). (arXiv:astro-ph/0312147 [astro-ph])
66. G. Torres, J. Andersen, A. Gimenez, Astron. Astrophys. Rev. 18, 67–126 (2010). https://doi.
org/10.1007/s00159-009-0025-1. (arXiv:0908.2624 [astro-ph.SR])
67. G. Torres, A. Claret, K. Pavlovski, A. Dotter, Astrophys. J. 807, 26 (2015). https://doi.org/10.
1088/0004-637X/807/1/26. ([arXiv:1505.07461].)
68. A. Udalski, Acta Astron. 53, 291 (2003). (arXiv:astro-ph/0401123 [astro-ph])
69. A. Udalski, M. Kubiak, M. Szymanski, Acta Astron. 47, 319–344 (1997).
(arXiv:astro-ph/9710091 [astro-ph])
70. A. Udalski, I. Soszynski, M. Szymanski, M. Kubiak, G. Pietrzynski, P. Wozniak, K. Zebrun,
Acta Astron. 48, 563 (1998). (arXiv:astro-ph/9812348 [astro-ph])
71. A. Udalski, M. Szymanski, M. Kubiak, G. Pietrzynski, P. Wozniak, K. Zebrun, Acta Astron.
48, 1 (1998). (arXiv:astro-ph/9803035 [astro-ph])
72. R.P. van der Marel, D.R. Alves, E. Hardy, N.B. Suntzeff, Astron. J. 124, 2639–2663 (2002).
https://doi.org/10.1086/343775. (arXiv:astro-ph/0205161 [astro-ph])
73. F. Vilardell, I. Ribas, C. Jordi, Astron. Astrophys. 459(1), 321–331 (2006). https://doi.org/10.
1051/0004-6361:20065667. (arXiv:astro-ph/0607236 [astro-ph])
74. F. Vilardell, I. Ribas, C. Jordi, E.L. Fitzpatrick, E.F. Guinan, Astron. Astrophys. 509, A70
(2010). https://doi.org/10.1051/0004-6361/200913299. (arXiv:0911.3391 [astro-ph.CO])
75. C. Wang, H. Yuan, Y. Huang, Astron. J. 163, 149 (2022). https://doi.org/10.3847/1538-3881/
ac4dec. ([arXiv:2201.09438])
5 Eclipsing Binary Stars as Precise and Accurate Distance Indicators 87

76. A.J. Wesselink, Bull. Astron. Inst. Netherlands 10, 91 (1946)

77. A.J. Wesselink, Mon. Not. R. Astron. Soc. 144, 297 (1969)
78. R.E. Wilson, E.J. Devinney, Astrophys. J. 166, 605 (1971). https://doi.org/10.1086/150986
79. G. Worthey, H.C. Lee, Astrophys. J. Suppl. 193, 1 (2011). https://doi.org/10.1088/0067-0049/
193/1/1. (arXiv:astro-ph/0604590 [astro-ph])
80. L. Wyrzykowski, A. Udalski, M. Kubiak, M. Szymanski, K. Zebrun, I. Soszynski, P.R. Woz-
niak, G. Pietrzynski, O. Szewczyk, Acta Astron. 53, 1 (2003). https://doi.org/10.1016/S0094-
5765(02)00195-9. (arXiv:astro-ph/0304458 [astro-ph])
81. S. Zucker, T. Mazeh, Astrophys. J. 420, 806 (1994). https://doi.org/10.1086/173605
Chapter 6
On Cepheid Distances in the H0 .

Measurement

Richard I. Anderson

6.1 Cepheids: From Variable Stars to Cosmic Yardsticks

The variability of classical Cepheids (henceforth for simplicity: Cepheids, cf.

Sect. 6.2) was first discovered by visual obserations in the 18th century by John
Goodricke and Edward Piggott [1]. Friedrich Argelander began systematically sur-
veying stars on both hemispheres in the 1850s, and Charles Pickering pushed for
systematic observations of variable stars towards the end of the 19th century. Pick-
ering hired female assistants (the so-called Harvard “Computers”) to process the
observations collected [2]. The photographic plates were astrometrically aligned
and variable stars identified by blinking observations of the same sky area observed
at different epochs. Henrietta Leavitt thus conducted an unprecedented census of
1777 variable stars in the Small Magellanic Cloud [3].
In 1912, Pickering reported on Leavitt’s groundbreaking discovery based on 25
classical Cepheids, that the apparent magnitude .m of 25 SMC Cepheids correlated
well with the logarithm of their period of variability . P, both near minimum and
maximum light [4]. Since all SMC stars were assumed to reside at a common distance,
it was clear that the observed relation between brightness and period must be an
intrinsic property of the stars. This period-luminosity relation (PLR) is nowadays
referred to as Leavitt’s law (LL).
Just one year later, Hertzsprung [5] made the first attempt to measure distance
using Leavitt’s discovery. He estimated the absolute magnitude of Cepheids at a
period near .6.6 d using secular parallax and, in turn, determined the distance modu-
lus of the SMC Cepheids by comparing their apparent magnitude at the same period
to that of the MW Cepheids. Hertzsprung’s work laid the foundation for the distance
ladder in use today. Despite concluding on an incorrect distance (.2000 lightyears), he
convincingly demonstrated that the SMC is external to the MW based on a compar-

R. I. Anderson (B)
Institute of Physics, École Polytechnique Fédérale de Lausanne (EPFL), Observatoire de
Sauverny, Chemin Pegasi 51, 1290 Versoix, Switzerland
e-mail: richard.anderson@epfl.ch
© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 89
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_6
90 R. I. Anderson

ison of the scale height of stars in the Solar Neighborhood. Thus, Cepheid distances
already informed the (extragalactic) distance scale before their variability mecha-
nism was understood [6–8], and well before the Great Debate of 1920, Hubble’s
works on M33 [9], as well as the subsequent discovery of the Hubble-Lemaître law
[10, 11].

6.1.1 The Hubble Constant Measured Using the Distance

Ladder

The relation between Hubble’s constant, H0 , and . D L is given by the Friedmann

equation, developed here to second order using the Robertson-Walker metric:
[ ]
cz 1
. DL = 1 + (1 − q0 )z − O(z ) + . . . ,
2
(6.1)
H0 2

where .q0 ≈ −0.55 is the deceleration parameter, whose observed negative value
implies the Universe’s accelerated expansion [12, 13]. Clearly, the wish to measure
H0 accruately is intimately tied to measuring accurate distances.
Equation 6.1 requires the measured redshifts to be purely cosmological, that is, any
motions due to graviational attraction (large scale flows, peculiar motions, etc.) must
be subtracted or negligible, so that H0 is best measured in the Hubble flow at distances
of 90–600 Mpc (.0.01 < z < 0.0233) [14]. Such distances are accessible thanks to
the extremely luminous type-Ia supernovae (SNeIa, cf. Vincenzi this volume) that
enable .∼5% relative distance estimates [15]. However, since SNeIa are very rare, no
SNIa has as yet been observed in a galaxy whose distance is known by geometrical
means. SNeIa thus require external calibration by other means, such as Cepheids,
via the distance ladder (DL).
The modern DL consists of three rungs as illustrated in Fig. 6.1. On the first
rung, Cepheids are calibrated as standard candles (SCs) using distances measured by
geometric methods. On the second rung, Cepheids are used to determine distances
to galaxies (SN-hosts) where at least one SNIa has been observed, out about .80 Mpc
[16]. On the last rung, SNeIa map cosmic expansion in the Hubble flow.
Seen the other way around, the DL connects the SNeIa distance-redshift relation
to an angular scale provided by geometry using Cepheids as the intermediary. By
further connecting the Hubble diagram to the early Universe’s angular scale (the
cosmological sound horizon) via the cosmic expansion history (. H (z)) using baryon
acoustic oscillations, the DL enables a crucial cosmological end-to-end test involving
two angular scales at opposite ends of the cosmos—trigonometric parallaxes and the
quantum fluctuations of the early Universe [17].
H0 measurements using a DL have a rich history (cf. Tully, this volume), involving
much controversy and disagreements. The HST Key Project to measure the Hubble
6 On Cepheid Distances in the . H0 Measurement 91

Fig. 6.1 The extragalactic distance ladder uses a two-rung approach for measuring H0 . Geometric
distances calibrate stellar standard candles for determining luminosity distances, which in turn
calibrate the luminosity of SNeIa that trace the Hubble flow. Credit: NASA, ESA, A. Feild (STScI),
and A. Riess (STScI/JHU)

constant was designed to resolved the ongoing disputes and for the first time achieved
a .10% measurement of H0 [18].
In the late 2000s, the SH0 ES (Supernova H0 for the Equation of State) project
was launched to measure H0 in order to determine the nature of Dark Energy using a
differential DL built on the principle of data homogeneity [19, 20]. In its most recent
form, the SH0 ES DL is implemented as a least squares problem that minimizes [16]:

χ 2 = (y − Lq)T C −1 (y − Lq) ,
. (6.2)

with the data vector . y, the design matrix . L that encodes the relevant equations, such
as the LL, distance moduli, etc., the covariance matrix .C, and the best-fit parameter
vector .q. Covariance among observables can be directly considered by quantifying
the corresponding elements of.C, and very significant amount of covariance informa-
tion is already being considered, notably among background corrections of Cepheids
in SN host galaxies [15, 16, 21]. Using 3445 degrees of freedom, 5 best fit parameters
are determined by the SH0 ES DL (see Sect. 6.1.2 for definitions): the LL slope .β
(.bW ), metallicity effect .γ (. Z W ), the fiducial absolute magnitude of a 10-day Cepheid
W
in the HST NIR Wesenheit magnitude . M H,1 , the absolute . B−band magnitude of
0
SNeIa, . M B , and Hubble’s constant .5 log10 H0 . Alternative symbols used in literature
are listed in parenthesis.
The matrix formalism is both simple and accurate, and allows straightforward
investigation of analysis variants, notably to assess the systematic effects of mod-
eling choices or dataset selection. Nevertheless, the results from the matrix inver-
sion are cross-checked with computationally intensive Markov chain Monte Carlo
(MCMC) simulations that allow inspection of marginalized posterior distributions
92 R. I. Anderson

and correlations among fit parameters. TRGB distances have also been included in
this procedure and can help to further improve precision. Including the latest results
based on cluster Cepheids, this DL yields . H0 = 73.15 ± 0.97 km s−1 Mpc−1 [22].
The specific construction of the SH0 ES DL, with the simultaneous fit to all observ-
ables resembles a tightly controlled experiment that seeks to maximize the benefits
of purely differential flux comparisons to leverage statistical precision while clearly
quantifying the systematic uncertainties involved in calibrating the absolute lumi-
nosity scale. In this spirit, the SH0 ES DL emphasized the use of homogeneous
and very well calibrated Hubble Space Telescope (HST ) photometry. The greatly
improved spatial resolution and enhanced sensitivity of the James Webb Space Tele-
cope (JWST ) is increasingly replacing the near-infrared (NIR) photometry observed
using HST WFC3/IR [23]. Careful cross-calibration between HST and JWST NIR
photometry is thus required to ensure direct compatibility. As a result of its spe-
cific construction, any modifications to the experimental setup must consider conse-
quences across the entire DL in order to ensure apples-to-apples comparisons.

6.1.2 On Standard Candles, Calibration, and Standardization

Astronomical standard candles (SCs) are objects that fulfill at least two conditions:
(a) their intrinsic brightness (absolute magnitude . M) can be calibrated in a given
photometric band, or combinations of several bands, and (b) they can be faithfully
recognized in different observational contexts, notably in different galaxies. SCs
are crucial to astronomy because they allow to determine luminosity distances, . D L
(cf. Eq. 6.2), by exploiting the fact that the observed flux . F decreases with distance
squared, that is, . F ∝ L/D L2 . In magnitude space, this becomes the distance modulus
equation .μ = m − M − A = 5 log10 D L + 25, with . D L in Mpc and extinction . A. As
noted by Hertzsprung, using standard candles for distance determination presumes
that the absolute magnitude established in one context, e.g., in the Milky Way, . Mcal ,
is representative of the absolute magnitude of sources seen in other contexts, such
as SN host galaxies, . M0 . Assuming this correspondence, distance is determined by
comparing the observed (apparent) magnitude to the calibrated absolute magnitude,
.μ = m obs − Mcal − A.
Calibrating standard candles to establish . Mcal requires observing them at known
distance, . Mcal = m − μ − A. Galaxies where this is possible are frequently referred
to as anchors, and three anchors are of particular importance for Cepheids: The
Milky Way (MW), the Large Magellanic Cloud (LMC), and the water megamaser-
host galaxy NGC 4258. Despite an available geometric distance estimate of the SMC,
complications related to its elongated line-of-sight orientation have thus limited its
usefulness for measuring H0 at the current levels of precision. Interestingly, the
distance measurements used for all three anchors are systematically very different. In
the MW, trigonometric parallaxes of SCs, notably from the European Space Agency
Gaia mission (cf. Lindegren, this volume, and references therein) [24–26], set the
current gold standard. In particular, parallaxes of open star clusters that host Cepheids
6 On Cepheid Distances in the . H0 Measurement 93

have provided the most accurate Cepheid calibration to date (.0.9% in distance) [27]
thanks to better understood systematics of the fainter, non-variable cluster stars, and
the statistical gain of averaging over many cluster members. The typical uncertainty
of such cluster parallaxes is .7 μarcsec, three times better than for individual field
Cepheids, so that one cluster Cepheid carries as much weight as 9 field Cepheids
[22].
The geometric distance measurements of the LMC (detached eclipsing binaries)
and NGC4258 (Keplerian disk model of observed megamaser features observed
using VLBI) rely on very different systematics, including instrumentation used and
modeling assumptions, rendering them truly independent checks of the calibration
based on parallax. The close agreement among the Cepheid calibrations based on
these anchors is therefore a strong attestation to the accuracy of the anchor distances as
well as the consistency among Cepheid observations spanning an enormous contrast
range from approximately .8th to .28th magnitude (8 dex in flux).
Standardization refers to the process of mitigating differences that disturb the
equivalence between . Mcal and . M0 . All stellar standard candles (including Cepheids,
SNeIa, and the TRGB method) require standardization to provide distances to better
than a few percent.
As Leavitt showed more than a century ago, Cepheids differ very significantly
in luminosity depending on their pulsation period (by about 3 mag/dex of period),
which ranges from a few days to more than 100 days. Cepheid distances are thus
obtained by comparing fiducial magnitudes. The most commonly adopted form of
the Leavitt law of Cepheids is

. M(P) = α + β · log P/P0 + γ · [Fe/H] , (6.3)

with .α the fiducial absolute magnitude of a Solar metallicity Cepheid at the pivot
period . P0 (also often referred to as the zero-point), .β the LL slope, and .γ the metal-
licity term that allows to correct the impact of chemical composition using iron (also:
oxygen) abundances, cf. below. Importantly, .α, .β, and .γ generally depend on the
photometric passband (or combinations thereof) used.
W
The fiducial absolute Cepheid magnitude in the SH0 ES DL, . M H,1 , represents
a .10 d Cepheid at the center of the instability strip in a specific combination of
photometric passbands (cf. Sect. 6.2.2.2). The commonly used pivot period of .10 d
is motivated by literature reports of LL non-linearity [28] and the period-amplitude
diagram of Cepheids [29], and corresponds to an intermediate period among Cepheids
in anchor galaxies, although Cepheids at greater distance tend to have increasingly
longer periods. .β and .γ are determined in the global DL fit using all Cepheids in
anchor and SN host galaxies simultaneously.
At present, there is no clearly demonstrated need to consider LL non-linearity
when using Cepheids to measure distances, and allowing for two different slopes
near the pivot period does not significantly improve the fit. Moreover, the astrophys-
ical motivation for using non-linear LLs is limited, since stellar models of single
chemical composition yield log-linear LLs across a large range of periods [e.g. [30,
31]]. However, stellar evolution models do predict a small dependence of chemi-
94 R. I. Anderson

cal composition on LL slope [30]. Since the effect is rather small, it has not yet
been confidently detected [32]. LL slope changes due to metallicity are furthermore
particularly small in the infrared, rendering the SH0 ES Wesenheit formalism very
insensitive to them.
It seems that observed LL non-linearity arises in a context-specific way from an
uneven sampling of the period-color relation, which is caused by the finite width of
the classical instability strip. Multi-phase PL relations [33] compare LLs at different
pulsation periods and have demonstrated broken PL relations. However, Cepheid
magnitudes used for measuring distances correspond to the average flux of Cepheids,
and cannot be directly compared to PL relations observed at specific phases. The
coincidence of the reported LL breaks near .10 d strongly suggest an origin in the
Hertzsprung progression [34].
Two further considerations are in order concerning LL non-linearity. Firstly, H0
bias due to LL non-linearity arises only in case of systematic differences among the
distributions of Cepheid periods in anchor and SN-host galaxies, and this effect is of
order .≾ 0.15% (the author & Bastian Lengen, priv. comm.). Secondly, variants of the
SH0 ES DL that include a break period of .10 d are presented alongside the baseline
analysis. These variants also indicate a very insignificant effect on H0 , while not
being preferred according to the change in .χ 2 .
Aside from pulsation period, the most common parameter used for standardization
considers metallicity differences [35–38]. For Cepheids, most studies assume a global
offset to the LL using a linear dependence on chemical composition usually quantified
via oxygen or iron abundances relative to Solar composition, that is, .ΔMmet = M0 −
Mcal = γ · Δ[O/H ]. Determining .γ is challenging because the effect is rather small
(.∼ −0.25 mag/dex) [16, 35], and metallicity differences among most Cepheids used
in the DL are typically less than 0.5–0.6 dex. Oxygen abundances estimated in SN-
host galaxies fall within a rather narrow range that is fully contained within the
abundance ranges of Cepheids in the Milky Way, LMC, and the SMC [16]. As
a result, the uncertainty in the metallicity effect does not significantly affect H0 .
However, metallicity can be a limiting factor for determining distances to individual
Cepheids, e.g., to map Galactic structure, or for determining Gaia’s parallax offsets
[39].
Standardization is best done using quantities than can be directly observed in the
specific sources of interest, such as individual elemental abundances derived from
Cepheid spectra. Such comparisons are presently limited to MW and LMC Cepheids,
although future 30m class telescopes will significantly increase the ability to observe
individual stellar spectra in other nearby galaxies. However, the majority of SN-host
galaxies will remain off-limits to individual star spectroscopy. At such distances,
equivalent information based on the closest proxys is typically used, such as oxygen
abundances of H II regions derived by the strong line method [40, 41], or abundances
of blue supergiants [42, 43]. Ensuring the compatibility of proxy information used
for standardization is thus crucial to obtaining unbiased distance estimates. For the
realm where this has been feasible, such agreement has been shown, for example
among HII regions in the Milky Way and Cepheid abundances determined by optical
spectra [16].
6 On Cepheid Distances in the . H0 Measurement 95

6.1.2.1 A Sidenote on TRGB Standardization

As all SCs, the TRGB method [44] is also subject to standardization. However,
metallicity corrections are complicated by the statistical nature of the measurement
(.m TRGB is measured based on a sample, not for each star individually) and the diverse
population of giant stars composed of different ages and chemical composition. As
a result, there is less agreement over how population differences should be treated,
with some arguing for color-based metallicity corrections [38] and others assuming
that differences among the bluest (interpreted to be the oldest) stars will be negligible
[45].
Ubiquitous small amplitude pulsations observed in luminous red gaints near the
tip [46, 47] provide useful addition information that may resolve this deadlock.
[48] recently showed that all stars near the TRGB feature are variable, and that the
variabilty of red giants provides crucial information about the composition of the
diverse red giant population. Comparison between the Large and Small Magellanic
Clouds (Koblischke & Anderson, in prep.) reveals that the PL sequences of small
amplitude red giants in the OGLE (Optical Gravitational Lensing Experiment) cat-
alog of variable stars (OSARGs) are systematically shifted to shorter period in the
SMC. This is indicative of a metallicity effect that causes lower-metallicity stars to
be more compact, increasing the average density of stars, and thus reducing their
√
periods, since . P ∝ 1/ ρ. Thus, OSARG-like variability can be used to standardize
metallicity differences for the TRGB method, provided such low levels of variability
(amplitudes of order 0.01–0.04 mag) can be detected.
Another form of standardization is required for TRGB distances due to method-
ological choices. The CATs (Comparative Analysis of TRGBs) team [49–51] demon-
strated a significant tip-contrast relation (TCR) for TRGB measurements, whereby
fields observed in the same (very distant) galaxy differed in TRGB magnitude accord-
ing to the ratio of the number of stars that are fainter and brighter (. R = N+ /N− ) than
the TRGB feature. Standardization to a fiducial value of . R = 4 was thus proposed
to ensure consistency. Since TCR standardization does not depend on distance, the
effect mainly increases the scatter of TRGB distances to SN host galaxies, rather
than biasing H0 . Meanwhile, [48] showed that the TCR is a methodological arti-
fact introduced by weighting the edge detection response (Sobel filter output) in an
effort to assign statistical significance to it [52] and that unweighted edge detec-
tion responses do not exhibit a significant TCR. Moreover, simulations showed that
unweighted EDRs yield less biased TRGB magnitudes than weighted EDRs [48].
Since . R is defined after determining .m TRGB , TCR corrections are slightly circular,
and unweighted EDRs should be the preferred choice of measuring .m TRGB . Addi-
tional details on the TRGB method are presented in Beaton (this volume), and [53]
provides a detailed account of other methodological aspects of TRGB measurements
that must be considered to obtain consistent results.
96 R. I. Anderson

6.2 Cepheids and Where to Find Them—Observational

Considerations

6.2.1 On the Usefulness of Cepheids in the DL

Classical Cepheids afford the best precision for measuring distances thanks to the
fortuitous combination of (a) high luminosity, (b) highly regular variability, (c) com-
paratively straightforward identification thanks to large (optical) amplitudes, (d) and
relatively high prevalence in star forming galaxies. Of course, Cepheids also come
with their own observational challenges and are not suitable for every context. For
example, Cepheids in highly inclined galaxies suffer from enhanced crowding and
reddening issues. In such cases, the Tip of the Red Giant Branch (TRGB) method
can offer a suitable alternative [44].
An important advantage of Cepheids is that they are observed specifically and
individually. This differs notably from the detection of the TRGB feature, which
is ascribed to an astrophysically extremely short-lived phenomenon and which is
always measured using an inhomogeneous and contaminated (e.g., by AGB stars)
sample of stars. Galaxies where multiple types of SCs can be observed thus provide
cross-checks essential for understanding systematic uncertainties while allowing to
cross-calibrate multiple SC types, e.g., to increase the number of SN host galaxies
in the DL. The following reviews observational aspects of Cepheids and any known
issues involving their calibration or use as SCs.

6.2.2 Cepheid Measurements Used in the DL

6.2.2.1 Cepheid Types, Variability, and Classification

This chapter uses the term “Cepheid” synonymous with classical (type-I) Cepheids
that belong to the young population I, and whose prototype is.δ Cephei. However, it is
important to note that other Cepheid types exist, notably the type-II (subgroups W Vir,
BL Her, and RV Tau) and anomalous Cepheids that belong to the old population II
and exhibit relatively similar light variability, at least in terms of timescales and
amplitudes. Distinguishing these groups is crucial to ensure distance accuracy and
indeed led to a major revision of the distance scale by Walter Baade [54].
Classical Cepheids are, first and foremost, identified by their variability. Addi-
tional context, such as Galactic latitude for MW stars, or the position within a galaxy,
usefully separates between young and old stars. However, misclassification between
type-I and type-II Cepheids is relatively common if parallax information is not reli-
able or available. For stars at a common distance, PL sequences are extremely power-
ful at distinguishing among different types, notably in the Magellanic Clouds, where
many PL sequences have been identified [55]. Since classical Cepheids are easily
6 On Cepheid Distances in the . H0 Measurement 97

identified by their high luminosity, large number, and period range, there is little
room for confusion among Cepheid types in SN host galaxies.
Where available, multi-band time-series data are helpful to establish tempera-
ture variations that distinguish chromatic pulations from usually achromatic eclipse-
related phenomena. Spectroscopy can provide certainty by allowing to measure
effective temperature variations, asymmetric line shapes (a telltale of pulsations),
RV variations, as well as surface gravity and abundances.
Gaia’s magnitude limited survey has provided the largest homogeneous sample
of Cepheids classified solely based on (multi-band) light curves, containing 15 006
Cepheids of all sub-types, including approximately 12 554 classical Cepheids. These
Cepheids belong to galaxies within the Local Group (.< 1 Mpc), including the MW,
the Magellanic Clouds, M31, and M33, and nearby dwarfs. In the MW, there are
approximately .3500 classical Cepheids [56, 57], whereas nearly .10 000 Cepheids
are nowadays known in the Magellanic System [55, 57].
The prime targets for H0 are Cepheids that pulsate in the fundamental radial
mode (FM Cepheids). These are the most numerous, most regular in terms of their
variability, and the longest-period Cepheids. Their detection is further aided by saw-
tooth shaped light curves that exhibit a characteristic dependence on pulsation period.
This so-called Hertzsprung progression [34] provides important evidence of the sim-
ilarity of nearby and distant Cepheids. FM Cepheids have their highest amplitudes
between approximately 5–7 and 15–30 d, and amplitudes decrease for the longest-
period Cepheids [29].
Shorter-period Cepheids that pulsate in the first overtone (FO Cepheids) feature
sinusoidal light curves of lower amplitudes, and higher overtones feature progres-
sively lower amplitudes that can become challenging to detect and distinguish from
other types of variability. Approximate relations between FO and FM pulsation peri-
ods [58] are occasionally used to combine samples of FO and FM Cepheids. In these
cases, FO Cepheid periods are “fundamentalized” using period ratios of FO and
FM Cepheids in well observed samples, such as the LMC. However, the metallicity
dependence of . PFO /PFM [59] complicates this process and adds systematic uncer-
tainty that is difficult to quantify [e.g. [60]]. As a result, FO Cepheids are not usually
considered in the modern DL.
Besides being more luminous, FM Cepheids also have the advantage of being more
regular in their variability than FO Cepheids [61–65]. However, the very long period
Cepheids (. P ≿ 25 d) exhibit period fluctuations that can be stochastic or (semi-
)periodic in nature [66], notably causing issues with de-phasing to mean magnitudes
due to variable pulsation ephemerides [67]. However, even very fast period changes of
−5
.100 s/yr (. Ṗ/P ∼ 4 · 10 at .30 d, .Δ log P ∼ 6 · 10−7 ) are irrelevant for determining
distances using LLs.

6.2.2.2 Reddening-Free Wesenheit Magnitudes

Near-infrared (NIR) photometry significantly improves the use of Cepheids as stan-

dard candles due to the intrinsically reduced sensitivity to extinction, lower ampli-
98 R. I. Anderson

tudes, and an intrinsically reduced width of the instability strip, which results in lower
dispersion. Until recently, the main drawback of NIR photometry for the DL was the
reduced spatial resolution of HST ’s WFC3/IR channel compared to WFC3/UVIS.
However, NIRCAM on board the JWST now provides .4× better spatial resolution
than WFC3/IR, with a larger aperture, resulting in much improved Cepheid photom-
etry at large distances [23]. Unfortunately, the reduced amplitudes of Cepheids in the
NIR compared to the optical render Cepheid discovery using JWST not applicable
to large samples.
Issues related to dust extinction are commonly mitigated by computing Wesen-
heit1 magnitudes [68, 69], .m W , which are reddening-free by construction for a given
reddening law [SH0 ES uses . RV = 3.3 from [70, 71]]. The SH0 ES near-IR Wesen-
heit formalism combines . H −band (F160W) magnitudes with optical colors, m W H .=
F160W − R W · (F555W − F814W), where . R W = A H /(A V − A I ) = 0.386 set by
the reddening law. For Cepheids observed by Gaia, . RGW = A G /(A G B P − A G R P ) ≈
1.91. Importantly, . R W depends on intrinsic source color when using wide-band pho-
tometry due to the shape of the spectral energy distribution incident on the photomet-
ric passband [72]. The SH0 ES approach is thus particularly robust against reddening
issues thanks to the combination of intrinsically extinction-insensitive NIR photome-
try with a rather small color-dependent term. For example, a.10% change in reddening
law (from . RV = 3.3 to .3.0) would result in a .0.01 − 0.05 change in . R H (depending
on reddening law), which multiplied by a typical color excess of . E(V − I ) ≈ 0.6
would merely change m W H by .∼ 0.03 · 0.6 = 0.02 mag, that is, much less than the
dispersion in most SN host galaxies, cf. Fig. 22 in [16].
Wesenheit magnitudes have the desirable side-effect of narrowing LL dispersion
due to the finite width of the instability strip, since. R W happens to be close to the slope
of the period-color relation. Of course, Wesenheit magnitudes require knowledge of
the reddening law, which can vary among sightlines, and this can be an issue for
specific sightlines, such as Cepheids in the SMC, which has been reported to have
an abnormal reddening law [73]. However, given the thousands of sightlines among
Cepheids in anchor galaxies (MW, LMC, NGC4258) and the thousands of sightlines
to Cepheids in SN-host galaxies it is highly unlikely for reddening-law related bias
to significantly affect H0 , as was also found empirically [16, 74].

6.2.2.3 Cepheids in the Milky Way

MW Cepheids are particularly important for the DL due to the ability to measure
trigonometric parallax [75]. However, MW Cepheids are subject to several challenges
that arise from their location inside the MW’s disk, where dust extinction is high and
a strong function of distance, and where the LL cannot be recovered using apparent
magnitudes alone. Observationally, the brightness of Cepheids can be an asset in
that high-resolution spectra can be observed [76–78]. However, the same brightness

1The German word “Wesenheit” relates to the abstract innate nature (or essence) of an entity, cf.
https://www.dwds.de/wb/Wesenheit.
6 On Cepheid Distances in the . H0 Measurement 99

can easily saturate photometric observations of Cepheids, notably from space, using
large apertures, and in the infrared. Another practical issue is the distribution of
MW Cepheids across a very large sky area, requiring MW Cepheids to be targeted
individually and causing large observational overheads.
Drift scanning observations with HST have allowed the collection of photometry
in the SH0 ES Wesenheit system for a legacy sample of 75 MW FM Cepheids that
feature low extinction and are sufficiently nearby (.∼ 1 − 5 kpc) for high-quality Gaia
parallaxes [79, 80]. HST drift scan observations have further provided parallaxes of
8 Cepheids, with a typical accuarcy of .40 − 50 μas [81–83].
Hundreds of high-quality parallaxes of MW Cepheids have been published by
Gaia [24–26, 84, 85], an order of magnitude increase over ESA’s predecessor Hip-
parcos [86, 87]. Clearly, Gaia has been a game changer for DL calibration and
has ushered in a new era for SCs. However, Gaia (E)DR3 parallaxes are subject to
unfortuante systematics at the level of 10–20.µas (a 3–6% error at.3 kpc) that correlate
strongly with sky position (ecliptic latitude), magnitude, and color, cf. Lindegren, this
volume. [84] (L21) determined parallax offset corrections based on these parameters
using bright physical pairs, LMC stars, and quasars. However, several follow-up stud-
ies [16, 39, 88] have shown that stars brighter than .G ≾ 11 mag require additional
correction at levels similar to the L21 corrections [89]. Very recently, [90] showed
that the L21 corrections require significant further corrections at bright (.G ≾ 11 mag)
magnitudes and that these residual corrections depend significantly on sky position,
likely because the bright physical pairs insufficiently sample the full sky. As a result,
all SC calibrations in this magnitude range must simultaneously solve for a sample
parallax offset, which requires careful analysis and generally comes at the cost of
precision, and this notably includes Cepheids that are typically brighter than .11 mag.
The best parallax accuracy for MW Cepheids is afforded by the fainter non-
variable member (main sequence) stars of open star clusters that host Cepheids [22,
27]. The numerous member stars improve statistical precision while Gaia’s parallax
sysetmatics are best understood in the magnitude range probed by the members stars
[84, 91]. This fortuitous combination yields a typical accuracy of.7 μarcsec for cluster
Cepheids, affording them the same weight as .∼ 9 field Cepheids for determining
W W
. M H,1 . Combining cluster and field Cepheids thus has allowed to calibrate . M H,1 , and
an analogous magnitude based on Gaia photometry, to better than .0.020 mag (.0.9%
in distance), while simultaneously solving the sample’s overall parallax offset to
.−13 ± 5 μas (.−19 ± 3 μas for the larger sample based on Gaia photometry) [27].

6.2.2.4 Cepheids in the Magellanic Clouds

The Magellanic Clouds host very large Cepheid populations that have provided the
most important proving grounds for testing stellar models thanks to the ability to
observe the Cepheids in great detail at a virtually constant distance.
LMC Cepheids are particularly valuable due to the extremely accurate distance
determined based on 20 detached eclipsing binaries (cf. Pietrzynski & Graczyk, this
volume; [92]). The intrinsic depth of the SMC combined with the relatively few (15)
100 R. I. Anderson

detached eclipsing binaries that define its distance [93] has thus far limited the use
of the SMC as an anchor galaxy however, see Breuval et al. [94]. Photometric obser-
vations using HST ’s DASH mode have enabled efficient photometric observations
of a significant sample of long-period Cepheids, despite the large angular separa-
tion among LMC Cepheids [95]. While ground-based photometry of LMC Cepheids
abounds from several surveys [58, 96–98], HST ’s photometry provides unmatched
photometric consistency and avoids issues related to Earth’s atmosphere, notably in
the NIR.
Cepheids in the Magellanic Clouds usefully constrain the slope of the metallicity
term, .γ [99], since they are less metallic than most Cepheids in the Solar neighbor-
hood, are not heavily extincted, and not subject to Gaia parallax offsets. However,
the internal depth of both galaxies must be considered for best results, and correcting
for the LMC plane geometry [97] reduces the scatter of the observed Cepheid LL.
This effect is even more pronounced in the SMC [100], which also allows to extend
the metallicity lever to the lowest values [35].

6.2.2.5 Cepheids in NGC 4258

The spiral galaxy NGC 4258 (Messier 106) is a crucial DL anchor thanks to a water
megamaser that orbits its supermassive black hole [101], cf. Braatz & Pesce this
volume. Modeling the observed maser emission as a warped Keplerian disk has
allowed to determine the distance to this galaxy to within .1.5% [102]. Importantly
for the DL, this measurement shares virtually no common systematics with Gaia
parallaxes, or the modeling of binary star orbits (apart from Kepler’s laws) that
underlies the LMC distance. The largest population of .669 Cepheids in NGC 4258
has been observed in the SH0 ES Wesenheit system by [103]. However, long-period
Cepheids in NGC 4258 are sufficiently close to detect them using ground-based,
8m-class telescopes [104, 105].
At a distance of .7.6 Mpc, Cepheids in NGC 4258 are observed under conditions
that resemble the conditions of other galaxies much more than Cepheids in the
MW, or the LMC, notably with regards to the process by which the photometry is
collected (cf. Sect. 6.2.3), the typical flux range, the blending of sources, and the
metallicity corrections based on HII regions. This makes NGC 4258 the only anchor
galaxy where Cepheids, as well as other SCs, such as the TRGB and J-region AGB
methods [106, 107], can be calibrated in the same way as they are used for deter-
mining distances. Thus, NGC 4258 allows very direct apples-to-apples comparisons
for determining distances to individual galaxies, such as M51 [60]. Additionally, the
consistent absolute LL calibrations determined using MW parallaxes, the LMC dis-
tance, and NGC 4258 underline the accuracy of the measurements in all three anchor
galaxies.
6 On Cepheid Distances in the . H0 Measurement 101

6.2.2.6 Cepheids in SN-Host Galaxies

The SH0 ES DL uses 2150 Cepheids in 37 SN-host galaxies to calibrate the SNIa
luminosity zero-point, and these galaxies are located across a wide range of distances,
from .6.7 (M101) to .80 Mpc (NGC 105), with an average distance of approximately
.27 Mpc. This dataset represents a major time investment, totaling approximately
.1050 orbits, or the equivalent of .∼ 66 full days of continuous HST observations. In

the nearest galaxies, several fields can be monitored to map Cepheids across galaxies,
whereas a single frame (WFC3/NIR footprint is .2.7 × 2.7 arcmin.2 ) centered on the
galactic bulge captures the majority of the Cepheids in distant galaxies.
Cepheids in SN-hosts are discovered at optical wavelengths where their ampli-
tudes are largest, .∼ 1 mag. .V − (F555W) and . I −band (F814W) observations are
de-phased to estimate mean magnitudes, whereas . H −band (F160W) observations
can usually be used at random phase thanks to the reduced amplitudes. The combina-
tion of. H −band magnitudes, and.V − I colors is then used to compute the Wesenheit
magnitudes. Further information on Cepheid photometry beyond the Local Group
and Wesenheit magnitudes is provided in Sect. 6.2.3.
Bona fide Cepheids used for the DL are discovered based on their light variations
and subsequently vetted using several quality criteria, cf. [103]. These cuts include
amplitudes, amplitude ratios, dispersion of Cepheid observations around template
light curve fits, minimum period of .5 d, and average color to ensure a matching
temperature range. Finally, the sample is cleaned by .3σ clipping in all three bands.
Impressively, binning the light curves of many Cepheids in period ranges allows to
illustrate the Hertzsprung progression, which adds further confidence concerning the
equivalence of the distant and nearby Cepheids.
The constant angular resolution of the photometric system corresponds to increas-
ingly larger physical scales, the more distant the SN host galaxy. This creates a fairly
clean correlation between the degree to which Cepheids are blended with other
sources and the distance at which they are observed. Variations in surface source
density, e.g., due to inclination or the region of the galaxy observed, further impact
the amount of blending. At large source densities, the term crowding refers to heavy
blending in fields with many overlapping sources.
Crowding corrections are used to remove background light contributions due to
unresolved sources [108]. The background due to blended sources is measured using
artificial star tests that locally insert and remeasure the flux of sources in the vicinity
of the target star, and with similar brightness [14, 109]. This is a significant effect
in many SN host galaxies (mean of .∼ 0.4 mag, up to nearly .1 mag at the largest
distances), although corrections as small as a few hundredths of a magnitude are
also estimated in low density regions of nearby galaxies. The accuracy of crowding
corrections has been extensively tested [16] and recently been validated using JWST ’s
higher spatial resolution [23].
Stellar association bias arises because wide binaries and open star clusters add
flux to Cepheids in NGC 4258 and SN-host galaxies, while being resolved (and
hence not contributing) to Cepheid flux measurements used to calibrate the LL in the
LMC and MW [110]. Given a typical physical scale of clusters being approximately
102 R. I. Anderson

4 pc, the artificial star tests used for crowding corrections cannot correct for this
.

effect, leading to a small, albeit one-sided bias for the DL because the effect occurs
primarily in the SN-host galaxies. Using M31 as a SN-host analog, [110] estimated
an effect of .∼ 0.007 mag per SN-host Cepheid, and a correction of this size has been
included in the SH0 ES analysis. A similar estimate has been found in M33 [111],
and additional work based on M101 and NGC 4258 is ongoing (Spetsieri et al., in
prep.). While a better quantification of this bias will further improve H0 accuracy, it
is important to note that stellar association bias cannot resolve the H0 tension, since
both the requirements of Cepheid discovery (notably sufficiently large amplitudes and
characteristic light curve shapes) and the quality cuts applied to Cepheid candidates
(notably based on color and LL .σ clipping) provide very strong protection against
significant bias. In essence, Cepheids blended by very luminous clusters are either
not discovered due to diminished amplitudes, or rejected due to their blue color.
Metallicity corrections in SN-host galaxies can unfortunately not (yet) be mea-
sured directly from Cepheid spectra. Instead, oxygen abundances of Cepheids in
SN-host galaxies are estimated based on their galactocentric location and metal-
licity gradients estimated using HII regions [40]. Care must be taken to ensure the
consistency of the derived oxygen abundances, notably with regards to the Solar oxy-
gen abundance problem. In that case, a high degree of agreement between Cepheid
oxygen abudances and HII region abundances has been shown [16][Appendix C].
Relativistic effects can bias H0 measurements based on the DL because SN-host
galaxies are systematically redshifted relative to the anchor galaxies due to cos-
mic expansion. For example, time dilation leads to systematically overestimated
Cepheid periods at greater distances, and thus, to overestimated luminosities and
underestimated distances. Correcting the dilated observed Cepheid periods therefore
has the effect of increasing H0 , by approximately .0.2% [112]. This effect is already
accounted for in the SH0 ES DL. Two additional relativistic effects have been con-
sidered [72]: . K −corrections and the redshifting of the reddening law on Wesenheit
magnitudes. While both effects are very small when working with NIR Wesenheit
magnitudes, single-band IR observations (e.g., of TRGB stars observed with JWST )
can exceed .1% distance bias at .100 Mpc and must therefore be considered as the use
of stellar SCs is pushed to ever greater distances.

6.2.3 Cepheid Photometry in SN-Hosts and NGC 4258

The quality of the observed photometry is paramount for accurate luminosity esti-
mates and H0 . As such, a few words on the measurement process are in order.
Point spread function (PSF) photometry remains the state-of-the-art method for
measuring Cepheid light curves, mostly using DAOPHOT and DOLPHOT [113,
114]. PSF photometry treats each star as a point source that can be distinguished
from neighboring sources and background, and it is used for busy and inhomogeneous
scenes, where the simpler aperture photometry approach is not suitable. An important
benefit of PSF photometry over aperature photometry is its ability to distinguish
6 On Cepheid Distances in the . H0 Measurement 103

resolved and unresolved sources, and PSF fitting produces quantitative information
that can be used to study the shape of the light’s distribution over the detector.
However, the key limitation of PSF photometry is the need to know how light is
dispersed by the optical system either a priori, or by measuring the PSF directly
from isolated sources. The details of how temporal variations in the PSF, e.g., due
to telescope breathing or focus changes, affect photometry are challenges where
observer experience frequently outperforms automated procedures.
In many ways, measuring photometry well is an art that requires a significant
time investment to master. Given the high hurdle for participating and the prevalence
of two methods commonly adopted as black boxes, there is certaintly room for
more independent cross-checks. A notable recent study presented new photometric
measurements based on the raw HST frames of NGC 5584 [115]. Comparison with
the photometry by the SH0 ES team revealed very good agreement. While further
cross-checks would be welcome, it is worth noting that H0 bias would primarily
arise due to inconsistency among photometric measurements between near and far
galaxies, rather than due to errors involving the absolute flux scale of all galaxies.

6.2.4 Baade-Wesselink Distances

The large scale radial pulsations of Cepheids enable distance determination using
the ratio of the observed angular (.ΔΘ) and linear radius variations (.ΔR), since . D ∝
ΔR/ΔΘ. Such distances are commonly referred to as Baade-Wesselink
distances [8, 116], and their development was intricately linked with establishing
pulsations as the cause of Cepheid pulsations [6, 117, 118]. .ΔΘ is observable for
a relatively small number of MW Cepheids using very long baseline optical inter-
ferometry [119–121] as well as surface brightness color relations (SBCRs) [122],
which require only photometric measurements and are thus applicable to longer dis-
tances. The SpectroPhoto-Interferometry of Pulsating Stars (SPIPS) algorithm per-
forms simultaneous modeling of interferometry, multi-band photometry, and radial
velocity observations
{ to determine distance using the same concepts [123, 124].
.ΔR = p · vr dφ is the product of the integral of the line of sight (radial) velocity
variations measured using optical spectra and the projection factor . p, which trans-
lates the observed line of sight velocity to the pulsational velocity as measured from
the center of the star. The requirement for optical high resolution spectra limits the
use of BW distances to relatively nearby objects, such as Cepheids in the MW and the
Magellanic Clouds [125–127]. However, the projection factor . p has been a limiting
factor to the precision attainable for BW distances due to its complex task of simul-
taneously correcting for geometric effects, limb darkening, and dynamical aspects
of stellar atmospheres [61, 128–131]. For example, despite the availability of Gaia
parallaxes, it has not been possible to determine whether or how . p may depend on
Cepheid properties, such as pulsation period . P [124, 132, 133]. While BW distances
thus provide a potentially interesting alternative for measuring individual distances
to Cepheids, they do not currently afford the accuracy required to inform H0 .
104 R. I. Anderson

6.2.5 On the Impact of Cepheid Multiplicity on H0

The majority of Cepheids reside in multiple systems [134–137], with estimates of the
multiplicity fraction ranging from .60% to more than .100% (many Cepheids occur in
triple or even quadruple systems). Cepheid companions can affect distances either
astrometrically or photometrically.
Astrometric issues falsify the parallax measurements of Cepheids and can arise
either due to real or apparent orbital motion. The former is caused by the move-
ment around the common center of gravity, whereas apparent orbital motion is a
photometric effect in disguise and caused by the shifting photocenter in a binary sys-
tem containing stars of similar magnitude. Timescales can be used to separate these
two effects, with true orbital motion generally being much longer than the pulsation
period, and apparent motion occuring in phase with the Cepheid’s variability. Poor
astrometry thus obviously can affect LL calibration, albeit on a star-by-star basis,
increasing scatter rather than biasing LL calibration. Thankfully, radial velocity mon-
itoring can very sensitively estimate whether orbital motion can significantly affect
parallax measurements [138]. Gaia’s astrometric quality flags, such as RUWE, are
also commonly used to identify high quality astrometric solutions.
Photometric distance bias can in principle arise if the calibrated LL does not
correspond to the observed LL due to differences in the nature of the companion
stars, or the prevalence of multiple star systems. However, these concerns are essen-
tially limited to individual Cepheids [139], which may stand out significantly from
observed LLs. For example, abnormally luminous and red LL outliers in the LMC
have been successfully identified as high-value double-lined spectroscopic binaries
(SB2s) [140, 141]. A similar population of SB2 Cepheids in the MW has not yet been
detected, which prompts the interesting question of how much the companion star
population of Cepheids may differ from galaxy to galaxy. Reasonable assumptions
on the distribution of wide binary separations and contrasts have allowed to constrain
a possible DL bias to .< 0.004% [110].

6.2.6 Cepheids in (Open) Star Clusters

In the Milky Way, approximately 10% of Cepheids are associated with (gravitation-
ally bound to) open clusters [27]. This estimate is based on Gaia (E)DR3 data and
considers the 2 kpc surrounding the Sun, where a rather complete census of both
Cepheids and star clusters is currently feasible. Different galaxies present different
clustered Cepheid fractions, with the LMC having the highest known fraction. In
M31, the fraction of Cepheids in star clusters is much less, approximately 2.5%
[110, 142], and M33 has a clustered Cepheid fraction of .3.3% [111]. HST UV pho-
tometry of M101 shows that Cepheid clustering is highly variable across the two
fields in which Cepheids have been observed and that it correlates strongly with star
formation (Spetsieri et al. in prep.).
6 On Cepheid Distances in the . H0 Measurement 105

Dynamical simulations [143] explain the low fraction of Cepheids that occur
in clusters as a consequence of cluster dissociation over time. Cluster dispersal is
driven by the violent gas expulsion due to massive star winds at very early times and
the dynamical interactions over longer timescales, and mostly occurs well before
intermediate-mass stars reach the Cepheid stage. Despite sensitivity to initial con-
ditions, the observed trend that younger Cepheids cluster more frequently strongly
supports the paradigm that a majority, if not all, stars are born in clusters [144].
Cepheid clustering is thus both dependent on global and local properties of galaxies,
as is observed.

6.2.7 Circumstellar Environments

Circumstellar environments (CSEs) of Cepheids have been detected using mid-IR

photometry [145, 146] and HI imaging [147], near- and mid-IR interferometry [148–
150], as IR excess among spectral energy distributions [151–153], and using Spitzer
spectroscopy [154]. Based on the IR excesses, it appears that . H −band is just short
of the region where IR excess becomes most noticeable, and that only a fraction of
Cepheids exhibit significant mid-IR excess. At present, it is unclear whether and how
CSEs can bias Cepheid distances. However, the following preliminary conclusions
appear reasonably secure: (a) CSEs can lead to DL bias only if the nature of CSEs
differs systematically among anchor and SN-host galaxies — in this regard, CSE
distance bias behaves like photometric bias due to companions; (b) CSEs may lead
individual Cepheids to be outliers from LLs; (c) long-wavelength observations, such
as JWST NIRCAM/F277W and F444W are likely to be more sensitive to CSEs
than . H −band and shorter wavelength observations; (d) if chemical composition
significantly affects the presence of CSEs, then metallicity corrections in use today
likely already partially correct the effect induced by CSEs.

6.2.8 Stellar Evolution Models

Cepheids pose many interesting challenges to stellar evolution theory, starting from
the occurrence of blue loops during core He burning, to observed rates of period
change, to predicting the correct instability strip boundaries and the right luminosity
(mass) range where Cepheids occur, to the mass-luminosity relation and its depen-
dence on metallicity, convection, and rotation, and the ability to predict Cepheid
variability, among others. Stellar models thus benefit from a large number of observa-
tional constraints that allow to understand the physical nature of these important stars
and to rigorously test model predictions along several axes at once. With thousands
of Cepheids in dozens of galaxies, the DL offers unprecedented means for compar-
isons, albeit at a reduced level of detail at greater distances. However, given their
106 R. I. Anderson

systematic uncertainties, stellar models have limited predictive power for Cepheid
distances.
Areas where stellar models can usefully inform the DL include, among others, (a)
the linearity of the LL (cf. above); (b) the dispersion of the LL observed in different
passbands; (c) differential comparisons of Cepheid luminosity under otherwise iden-
tical conditions as required, e.g., to determine the effects of chemical composition
on the LL. For example, predictions based on Geneva stellar evolution models [30]
predict a negative sign for the metallicity term, .γ , both in the individual .V − and
. H −band PL-relations and in the optical . WVI and near-IR Wesenheit systems . WH,VI
used in the SH0 ES analysis. Conversely, a widely used set of Cepheid models [31,
155–157] predicts a positive sign for PL-relations in the individual passbands (.γλ > 0
in those cases) and negatively signed period-Wesenheit relations, that is, when com-

Fig. 6.2 The slope of the metallicity effect on Wesenheit magnitudes determined by different
methods (bottom x-axis) and the impact of metallicity on H0 (top x-axis). Circles show results based
on different galaxies, squares results based on galactic abundance gradients, diamonds results based
on MW Cepheids observed with Gaia, downward triangles results based on BW-type distances,
and asterisks results based on stellar evolution models. Figure courtesy of Louise Breuval
6 On Cepheid Distances in the . H0 Measurement 107

bining a single photometric passband with a color (.γW < 0). Recent empirical work
by different teams has shown the metallicity effect to have negative sign across several
individual photometric bands as well as in the Wesenheit formulations [35, 77, 125,
158–160], cf. Fig. 6.2. Hence, it appears that the disagreement between the two sets
of models is due to different predicted locations of the instability strip boundaries,
as well as their dependence on metallicity.

6.3 Conclusions

Classical Cepheids provide the base calibration for the H0 measurement using the
extragalactic DL. They benefit from more than .110 years of study for measuring
distances. Recent progress in quantifying systematics has centered on sub-percent
effects for H0 , which ended up sharpening the accuracy of the measurement and
the significance of the Hubble tension. In the process, corrections that increase and
decrease the value of H0 have been identified. Gaia has revolutionized the cali-
bration of the LL in the MW based on trigonometric parallaxes, and JWST stands
to significantly improve H0 precision by better sensitivity and reduced LL disper-
sion in SN-host galaxies due to lower background (enabled by the .4× better spatial
resolution than HST ’s WFC3/IR).
Further improvements of the distance ladder and its experimental setup will yield
significant improvements in H0 accuracy, hopefully to deliver the desired .1% direct
measurement. Further improvements to Gaia’s parallax systematics are paramount to
achieving this level of accuracy. However, despite very significant efforts to identify
systematics that could bias H0 sufficiently to explain the H0 constant tension, no
such biases have been identified. For this reason, the author is cautiously optimistic
that the problem with the H0 constant will eventually enable new interesting insights
into the physics governing our Universe.

Acknowledgements RIA is funded by the SNSF through an Eccellenza Professorial Fellowship,

grant number PCEFP2_194638 and acknowledges support from the European Research Coun-
cil (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant
agreement No 947660).

References

1. J. Goodricke, A series of observations on, and a discovery of, the period of the variation of the
light of the star marked δ by Bayer, near the head of Cepheus, in a letter from John Goodricke,
Esq. to Nevil Maskelyne, D. D. F. R. S. and astronomer royal. Philos. Trans. R. Soc. London
Ser. I 76, 48–61 (1786)
2. D. Sobel, The Glass Universe: The Hidden History of the Women Who Took the Measure of
the Stars (HarperCollins Publishers, 2016)
3. H.S. Leavitt, 1777 variables in the magellanic clouds. Ann. Harvard College Obs. 60, 87–108.3
(1908)
108 R. I. Anderson

4. H.S. Leavitt, E.C. Pickering, Periods of 25 variable stars in the small magellanic cloud.
Harvard College Obs. Circular 173, 1–3 (1912)
5. E. Hertzsprung, Über die räumliche Verteilung der Veränderlichen vom δ Cephei-Typus.
Astron. Nachr. 196, 201 (1913)
6. A.S. Eddington, The pulsation theory of Cepheid variables. The Observatory 40, 290–293
(1917)
7. H. Shapley, On the nature and cause of cepheid variation. ApJ 40, 448 (1914)
8. A.J. Wesselink, The observations of brightness, colour and radial velocity of δ Cephei and
the pulsation hypothesis (Errata: 10 258, 310). Bull. Astron. Inst. Netherlands 10, 91 (1946)
9. E.P. Hubble, A spiral nebula as a stellar system: messier 33. ApJ 63, 236–274 (1926)
10. E. Hubble, A relation between distance and radial velocity among extra-galactic nebulae.
Proc. Natl. Acad. Sci. 15(3), 168–173 (1929)
11. G. Lemaître, Un Univers homogène de masse constante et de rayon croissant rendant compte
de la vitesse radiale des nébuleuses extra-galactiques. Annales de la Société Scientifique de
Bruxelles 47, 49–59 (1927)
12. S. Perlmutter, G. Aldering, G. Goldhaber, R.A. Knop, P. Nugent, P.G. Castro, S. Deustua,
S. Fabbro, A. Goobar, D.E. Groom, I.M. Hook, A.G. Kim, M.Y. Kim, J.C. Lee, N.J. Nunes,
R. Pain, C.R. Pennypacker, R. Quimby, C. Lidman, R.S. Ellis, M. Irwin, R. G. McMahon,
P. Ruiz-Lapuente, N. Walton, B. Schaefer, B. J. Boyle, A.V. Filippenko, T. Matheson, A.S.
Fruchter, N. Panagia, H.J.M. Newberg, W.J. Couch, T.S.C. Project, Measurements of Ω and
Ʌ from 42 high-redshift supernovae. ApJ 517(2), 565–586 (1999)
13. A.G. Riess, A.V. Filippenko, P. Challis, A. Clocchiatti, A. Diercks, P.M. Garnavich, R.L.
Gilliland, C.J. Hogan, S. Jha, R.P. Kirshner, B. Leibundgut, M.M. Phillips, D. Reiss, B.P.
Schmidt, R.A. Schommer, R.C. Smith, J. Spyromilio, C. Stubbs, N.B. Suntzeff, J. Tonry,
Observational evidence from supernovae for an accelerating universe and a cosmological
constant. AJ 116(3), 1009–1038 (1998)
14. A.G. Riess, L.M. Macri, S.L. Hoffmann, D. Scolnic, S. Casertano, A.V. Filippenko, B.E.
Tucker, M.J. Reid, D.O. Jones, J.M. Silverman, R. Chornock, P. Challis, W. Yuan, P.J. Brown,
R.J. Foley, A 2.4% determination of the local value of the hubble constant. ApJ 826(1), 56
(2016)
15. D. Brout, D. Scolnic, B. Popovic, A.G. Riess, A. Carr, J. Zuntz, R. Kessler, T.M. Davis, S.
Hinton, D. Jones, W.D. Kenworthy, E.R. Peterson, K. Said, G. Taylor, N. Ali, P. Armstrong, P.
Charvu, A. Dwomoh, C. Meldorf, A. Palmese, H. Qu, B.M. Rose, B. Sanchez, C.W. Stubbs,
M. Vincenzi, C.M. Wood, P.J. Brown, R. Chen, K. Chambers, D.A. Coulter, M. Dai, G.
Dimitriadis, A.V. Filippenko, R.J. Foley, S.W. Jha, L. Kelsey, R.P. Kirshner, A. Möller, J.
Muir, S. Nadathur, Y.-C. Pan, A. Rest, C. Rojas-Bravo, M. Sako, M.R. Siebert, M. Smith,
B.E. Stahl, P. Wiseman, The Pantheon+ analysis: cosmological constraints. ApJ 938(2), 110
(2022). https://doi.org/10.3847/1538-4357/ac8e04
16. A.G. Riess, W. Yuan, L.M. Macri, D. Scolnic, D. Brout, S. Casertano, D.O. Jones,
Y. Murakami, G.S. Anand, L. Breuval, T.G. Brink, A.V. Filippenko, S. Hoffmann, S.W. Jha,
W. D’arcy Kenworthy, J. Mackenty, B.E. Stahl, W. Zheng, A comprehensive measurement
of the local value of the hubble constant with 1 km s−1 Mpc−1 uncertainty from the hubble
space telescope and the SH0 ES team. ApJ 934(1), L7 (2022)
17. A.G. Riess, The expansion of the universe is faster than expected. Nat. Rev. Phys. 2(1), 10–12
(2020)
18. W.L. Freedman, B.F. Madore, B.K. Gibson, L. Ferrarese, D.D. Kelson, S. Sakai, J.R. Mould,
J. Kennicutt, C. Robert, H.C. Ford, J.A. Graham, J.P. Huchra, S.M.G. Hughes, G.D. Illing-
worth, L.M. Macri, P.B. Stetson, Final results from the hubble space telescope key project to
measure the hubble constant. ApJ 553(1), 47–72 (2001)
19. A.G. Riess, L. Macri, S. Casertano, M. Sosey, H. Lampeitl, H.C. Ferguson, A.V. Filippenko,
S.W. Jha, W. Li, R. Chornock, D. Sarkar, A redetermination of the hubble constant with the
hubble space telescope from a differential distance ladder. ApJ 699(1), 539–563 (2009)
20. A.G. Riess, L. Macri, W. Li, H. Lampeitl, S. Casertano, H.C. Ferguson, A.V. Filippenko,
S.W. Jha, R. Chornock, L. Greenhill, M. Mutchler, M. Ganeshalingham, M. Hicken, Cepheid
6 On Cepheid Distances in the . H0 Measurement 109

calibrations of modern type Ia supernovae: implications for the hubble constant. ApJS 183(1),
109–141 (2009)
21. D. Scolnic, D. Brout, A. Carr, A.G. Riess, T.M. Davis, A. Dwomoh, D.O. Jones, N. Ali,
P. Charvu, R. Chen, E.R. Peterson, B. Popovic, B.M. Rose, C.M. Wood, P.J. Brown, K.
Chambers, D.A. Coulter, K.G. Dettman, G. Dimitriadis, A.V. Filippenko, R.J. Foley, S.W.
Jha, C.D. Kilpatrick, R.P. Kirshner, Y.-C. Pan, A. Rest, C. Rojas-Bravo, M.R. Siebert, B.E.
Stahl, W. Zheng, The Pantheon+ analysis: the full data set and light-curve release. ApJ 938(2),
113 (2022)
22. A.G. Riess, L. Breuval, W. Yuan, S. Casertano, L.M. Macri, J.B. Bowers, D. Scolnic, T. Cantat-
Gaudin, R.I. Anderson, M. Cruz Reyes, Cluster cepheids with high precision gaia parallaxes,
low zero-point uncertainties, and hubble space telescope photometry. ApJ 938(1), 36 (2022)
23. A.G. Riess, G.S. Anand, W. Yuan, S. Casertano, A. Dolphin, L.M. Macri, L. Breuval,
D. Scolnic, M. Perrin, R.I. Anderson, Crowded no more: the accuracy of the hubble con-
stant tested with high resolution observations of cepheids by JWST (2023). arXiv e-prints:
arXiv:2307.15806
24. Gaia Collaboration, A.G.A. Brown, A. Vallenari, T. Prusti, J.H.J. de Bruijne, C. Babusiaux,
M. Biermann, O.L. Creevey, D.W. Evans, L. Eyer, A. Hutton, F. Jansen, C. Jordi, S.A. Klioner,
U. Lammers, L. Lindegren, X. Luri, F. Mignard, C. Panem, D. Pourbaix, S. Randich, P. Sar-
toretti, C. Soubiran, N.A. Walton, F. Arenou, C.A.L. Bailer-Jones, U. Bastian, M. Cropper,
R. Drimmel, D. Katz, M.G. Lattanzi, F. van Leeuwen, J. Bakker, C. Cacciari, J. Castañeda,
F. De Angeli, C. Ducourant, C. Fabricius, M. Fouesneau, Y. Frémat, R. Guerra, A. Guer-
rier, J. Guiraud, A. Jean-Antoine Piccolo, E. Masana, R. Messineo, N. Mowlavi, C. Nicolas,
K. Nienartowicz, F. Pailler, P. Panuzzo, F. Riclet, W. Roux, G. M. Seabroke, R. Sordo, P. Tanga,
F. Thévenin, G. Gracia-Abril, J. Portell, D. Teyssier, M. Altmann, R. Andrae, I. Bellas-
Velidis, K. Benson, J. Berthier, R. Blomme, E. Brugaletta, P.W. Burgess, G. Busso, B. Carry,
A. Cellino, N. Cheek, G. Clementini, Y. Damerdji, M. Davidson, L. Delchambre, A. Dell’Oro,
J. Fernández-Hernández, L. Galluccio, P. García-Lario, M. Garcia-Reinaldos, J. González-
Núñez, E. Gosset, R. Haigron, J.L. Halbwachs, N. C. Hambly, D. L. Harrison, D. Hatzidim-
itriou, U. Heiter, J. Hernández, D. Hestroffer, S.T. Hodgkin, B. Holl, K. Janßen, G. Jevardat de
Fombelle, S. Jordan, A. Krone-Martins, A. C. Lanzafame, W. Löffler, A. Lorca, M. Manteiga,
O. Marchal, P.M. Marrese, A. Moitinho, A. Mora, K. Muinonen, P. Osborne, E. Pancino,
T. Pauwels, J. M. Petit, A. Recio-Blanco, P.J. Richards, M. Riello, L. Rimoldini, A. C. Robin,
T. Roegiers, J. Rybizki, L. M. Sarro, C. Siopis, M. Smith, A. Sozzetti, A. Ulla, E. Utrilla,
M. van Leeuwen, W. van Reeven, U. Abbas, A. Abreu Aramburu, S. Accart, C. Aerts, J. J.
Aguado, M. Ajaj, G. Altavilla, M.A. Álvarez, J. Álvarez Cid-Fuentes, J. Alves, R. I. Ander-
son, E. Anglada Varela, T. Antoja, M. Audard, D. Baines, S.G. Baker, L. Balaguer-Núñez,
E. Balbinot, Z. Balog, C. Barache, D. Barbato, M. Barros, M.A. Barstow, S. Bartolomé,
J. L. Bassilana, N. Bauchet, A. Baudesson-Stella, U. Becciani, M. Bellazzini, M. Bernet,
S. Bertone, L. Bianchi, S. Blanco-Cuaresma, T. Boch, A. Bombrun, D. Bossini, S. Bouquil-
lon, A. Bragaglia, L. Bramante, E. Breedt, A. Bressan, N. Brouillet, B. Bucciarelli, A. Burlacu,
D. Busonero, A. G. Butkevich, R. Buzzi, E. Caffau, R. Cancelliere, H. Cánovas, T. Cantat-
Gaudin, R. Carballo, T. Carlucci, M. I. Carnerero, J. M. Carrasco, L. Casamiquela, M. Castel-
lani, A. Castro-Ginard, P. Castro Sampol, L. Chaoul, P. Charlot, L. Chemin, A. Chiavassa,
M. R. L. Cioni, G. Comoretto, W. J. Cooper, T. Cornez, S. Cowell, F. Crifo, M. Crosta,
C. Crowley, C. Dafonte, A. Dapergolas, M. David, P. David, P. de Laverny, F. De Luise, R. De
March, J. De Ridder, R. de Souza, P. de Teodoro, A. de Torres, E. F. del Peloso, E. del Pozo,
M. Delbo, A. Delgado, H. E. Delgado, J. B. Delisle, P. Di Matteo, S. Diakite, C. Diener, E. Dis-
tefano, C. Dolding, D. Eappachen, B. Edvardsson, H. Enke, P. Esquej, C. Fabre, M. Fabrizio,
S. Faigler, G. Fedorets, P. Fernique, A. Fienga, F. Figueras, C. Fouron, F. Fragkoudi, E. Fraile,
F. Franke, M. Gai, D. Garabato, A. Garcia-Gutierrez, M. García-Torres, A. Garofalo, P. Gavras,
E. Gerlach, R. Geyer, P. Giacobbe, G. Gilmore, S. Girona, G. Giuffrida, R. Gomel, A. Gomez,
I. Gonzalez-Santamaria, J. J. González-Vidal, M. Granvik, R. Gutiérrez-Sánchez, L. P. Guy,
M. Hauser, M. Haywood, A. Helmi, S. L. Hidalgo, T. Hilger, N. Hładczuk, D. Hobbs, G. Hol-
land, H. E. Huckle, G. Jasniewicz, P. G. Jonker, J. Juaristi Campillo, F. Julbe, L. Karbevska,
110 R. I. Anderson

P. Kervella, S. Khanna, A. Kochoska, M. Kontizas, G. Kordopatis, A.J. Korn, Z. Kostrzewa-

Rutkowska, K. Kruszyńska, S. Lambert, A.F. Lanza, Y. Lasne, J.F. Le Campion, Y. Le Fustec,
Y. Lebreton, T. Lebzelter, S. Leccia, N. Leclerc, I. Lecoeur-Taibi, S. Liao, E. Licata, E. P.
Lindstrøm, T. A. Lister, E. Livanou, A. Lobel, P. Madrero Pardo, S. Managau, R. G. Mann,
J. M. Marchant, M. Marconi, M. M. S. Marcos Santos, S. Marinoni, F. Marocco, D. J. Mar-
shall, L. Martin Polo, J. M. Martín-Fleitas, A. Masip, D. Massari, A. Mastrobuono-Battisti,
T. Mazeh, P. J. McMillan, S. Messina, D. Michalik, N. R. Millar, A. Mints, D. Molina,
R. Molinaro, L. Molnár, P. Montegriffo, R. Mor, R. Morbidelli, T. Morel, D. Morris, A. F.
Mulone, D. Munoz, T. Muraveva, C.P. Murphy, I. Musella, L. Noval, C. Ordénovic, G. Orrù,
J. Osinde, C. Pagani, I. Pagano, L. Palaversa, P.A. Palicio, A. Panahi, M. Pawlak, X. Peñalosa
Esteller, A. Penttilä, A. M. Piersimoni, F.X. Pineau, E. Plachy, G. Plum, E. Poggio, E. Poretti,
E. Poujoulet, A. Prša, L. Pulone, E. Racero, S. Ragaini, M. Rainer, C.M. Raiteri, N. Ram-
baux, P. Ramos, M. Ramos-Lerate, P. Re Fiorentin, S. Regibo, C. Reylé, V. Ripepi, A. Riva,
G. Rixon, N. Robichon, C. Robin, M. Roelens, L. Rohrbasser, M. Romero-Gómez, N. Rowell,
F. Royer, K. A. Rybicki, G. Sadowski, A. Sagristà Sellés, J. Sahlmann, J. Salgado, E. Salguero,
N. Samaras, V. Sanchez Gimenez, N. Sanna, R. Santoveña, M. Sarasso, M. Schultheis, E. Sci-
acca, M. Segol, J.C. Segovia, D. Ségransan, D. Semeux, S. Shahaf, H. I. Siddiqui, A. Siebert,
L. Siltala, E. Slezak, R.L. Smart, E. Solano, F. Solitro, D. Souami, J. Souchay, A. Spagna,
F. Spoto, I.A. Steele, H. Steidelmüller, C.A. Stephenson, M. Süveges, L. Szabados, E. Szegedi-
Elek, F. Taris, G. Tauran, M.B. Taylor, R. Teixeira, W. Thuillot, N. Tonello, F. Torra, J. Torra,
C. Turon, N. Unger, M. Vaillant, E. van Dillen, O. Vanel, A. Vecchiato, Y. Viala, D. Vicente,
S. Voutsinas, M. Weiler, T. Wevers, Ł. Wyrzykowski, A. Yoldas, P. Yvard, H. Zhao, J. Zorec,
S. Zucker, C. Zurbach, T. Zwitter, Gaia Early Data Release 3 summary of the contents and
survey properties. A&A 649, A1 (2021)
25. Gaia collaboration, T. Prusti, J. De Bruijne, A.G. Brown, A. Vallenari, C. Babusiaux, C. Bailer-
Jones, U. Bastian, M. Biermann, D.W. Evans, L. Eyer et al., The Gaia mission. A&A 595,
A1 (2016)
26. Gaia Collaboration, A. Vallenari, A.G.A. Brown, T. Prusti, J.H.J. de Bruijne, F. Arenou,
C. Babusiaux, M. Biermann, O.L. Creevey, C. Ducourant, D.W. Evans, L. Eyer, R. Guerra,
A. Hutton, C. Jordi, S.A. Klioner, U.L. Lammers, L. Lindegren, X. Luri, F. Mignard, C. Panem,
D. Pourbaix, S. Randich, P. Sartoretti, C. Soubiran, P. Tanga, N.A. Walton, C.A.L. Bailer-
Jones, U. Bastian, R. Drimmel, F. Jansen, D. Katz, M.G. Lattanzi, F. van Leeuwen, J. Bakker,
C. Cacciari, J. Castañeda, F. De Angeli, C. Fabricius, M. Fouesneau, Y. Frémat, L. Galluc-
cio, A. Guerrier, U. Heiter, E. Masana, R. Messineo, N. Mowlavi, C. Nicolas, K. Nienar-
towicz, F. Pailler, P. Panuzzo, F. Riclet, W. Roux, G.M. Seabroke, R. Sordo, F. Thévenin,
G. Gracia-Abril, J. Portell, D. Teyssier, M. Altmann, R. Andrae, M. Audard, I. Bellas-Velidis,
K. Benson, J. Berthier, R. Blomme, P. W. Burgess, D. Busonero, G. Busso, H. Cánovas,
B. Carry, A. Cellino, N. Cheek, G. Clementini, Y. Damerdji, M. Davidson, P. de Teodoro,
M. Nuñez Campos, L. Delchambre, A. Dell’Oro, P. Esquej, J. Fernández-Hernández, E. Fraile,
D. Garabato, P. García-Lario, E. Gosset, R. Haigron, J. L. Halbwachs, N. C. Hambly, D. L.
Harrison, J. Hernández, D. Hestroffer, S. T. Hodgkin, B. Holl, K. Janßen, G. Jevardat de
Fombelle, S. Jordan, A. Krone-Martins, A. C. Lanzafame, W. Löffler, O. Marchal, P. M.
Marrese, A. Moitinho, K. Muinonen, P. Osborne, E. Pancino, T. Pauwels, A. Recio-Blanco,
C. Reylé, M. Riello, L. Rimoldini, T. Roegiers, J. Rybizki, L. M. Sarro, C. Siopis, M. Smith,
A. Sozzetti, E. Utrilla, M. van Leeuwen, U. Abbas, P. Ábrahám, A. Abreu Aramburu, C. Aerts,
J. J. Aguado, M. Ajaj, F. Aldea-Montero, G. Altavilla, M. A. Álvarez, J. Alves, F. Anders,
R. I. Anderson, E. Anglada Varela, T. Antoja, D. Baines, S. G. Baker, L. Balaguer-Núñez,
E. Balbinot, Z. Balog, C. Barache, D. Barbato, M. Barros, M.A. Barstow, S. Bartolomé,
J. L. Bassilana, N. Bauchet, U. Becciani, M. Bellazzini, A. Berihuete, M. Bernet, S. Bertone,
L. Bianchi, A. Binnenfeld, S. Blanco-Cuaresma, A. Blazere, T. Boch, A. Bombrun, D. Bossini,
S. Bouquillon, A. Bragaglia, L. Bramante, E. Breedt, A. Bressan, N. Brouillet, E. Bru-
galetta, B. Bucciarelli, A. Burlacu, A. G. Butkevich, R. Buzzi, E. Caffau, R. Cancelliere,
T. Cantat-Gaudin, R. Carballo, T. Carlucci, M. I. Carnerero, J. M. Carrasco, L. Casamiquela,
M. Castellani, A. Castro-Ginard, L. Chaoul, P. Charlot, L. Chemin, V. Chiaramida, A. Chi-
6 On Cepheid Distances in the . H0 Measurement 111

avassa, N. Chornay, G. Comoretto, G. Contursi, W. J. Cooper, T. Cornez, S. Cowell, F. Crifo,

M. Cropper, M. Crosta, C. Crowley, C. Dafonte, A. Dapergolas, M. David, P. David, P. de
Laverny, F. De Luise, R. De March, J. De Ridder, R. de Souza, A. de Torres, E. F. del
Peloso, E. del Pozo, M. Delbo, A. Delgado, J.B. Delisle, C. Demouchy, T.E. Dharmawar-
dena, P. Di Matteo, S. Diakite, C. Diener, E. Distefano, C. Dolding, B. Edvardsson, H. Enke,
C. Fabre, M. Fabrizio, S. Faigler, G. Fedorets, P. Fernique, A. Fienga, F. Figueras, Y. Fournier,
C. Fouron, F. Fragkoudi, M. Gai, A. Garcia-Gutierrez, M. Garcia-Reinaldos, M. García-
Torres, A. Garofalo, A. Gavel, P. Gavras, E. Gerlach, R. Geyer, P. Giacobbe, G. Gilmore,
S. Girona, G. Giuffrida, R. Gomel, A. Gomez, J. González-Núñez, I. González-Santamaría,
J. J. González-Vidal, M. Granvik, P. Guillout, J. Guiraud, R. Gutiérrez-Sánchez, L. P. Guy,
D. Hatzidimitriou, M. Hauser, M. Haywood, A. Helmer, A. Helmi, M. H. Sarmiento, S. L.
Hidalgo, T. Hilger, N. Hładczuk, D. Hobbs, G. Holland, H. E. Huckle, K. Jardine, G. Jas-
niewicz, A. Jean-Antoine Piccolo, Ó. Jiménez-Arranz, A. Jorissen, J. Juaristi Campillo,
F. Julbe, L. Karbevska, P. Kervella, S. Khanna, M. Kontizas, G. Kordopatis, A. J. Korn,
Á. Kóspál, Z. Kostrzewa-Rutkowska, K. Kruszyńska, M. Kun, P. Laizeau, S. Lambert, A.F.
Lanza, Y. Lasne, J.F. Le Campion, Y. Lebreton, T. Lebzelter, S. Leccia, N. Leclerc, I. Lecoeur-
Taibi, S. Liao, E.L. Licata, H.E.P. Lindstrøm, T. A. Lister, E. Livanou, A. Lobel, A. Lorca,
C. Loup, P. Madrero Pardo, A. Magdaleno Romeo, S. Managau, R. G. Mann, M. Manteiga,
J.M. Marchant, M. Marconi, J. Marcos, M.M.S. Marcos Santos, D. Marín Pina, S. Marinoni,
F. Marocco, D. J. Marshall, L. Martin Polo, J.M. Martín-Fleitas, G. Marton, N. Mary, A. Masip,
D. Massari, A. Mastrobuono-Battisti, T. Mazeh, P. J. McMillan, S. Messina, D. Michalik, N. R.
Millar, A. Mints, D. Molina, R. Molinaro, L. Molnár, G. Monari, M. Monguió, P. Montegriffo,
A. Montero, R. Mor, A. Mora, R. Morbidelli, T. Morel, D. Morris, T. Muraveva, C. P. Murphy,
I. Musella, Z. Nagy, L. Noval, F. Ocaña, A. Ogden, C. Ordenovic, J. O. Osinde, C. Pagani,
I. Pagano, L. Palaversa, P. A. Palicio, L. Pallas-Quintela, A. Panahi, S. Payne-Wardenaar,
X. Peñalosa Esteller, A. Penttilä, B. Pichon, A. M. Piersimoni, F. X. Pineau, E. Plachy,
G. Plum, E. Poggio, A. Prša, L. Pulone, E. Racero, S. Ragaini, M. Rainer, C.M. Raiteri,
N. Rambaux, P. Ramos, M. Ramos-Lerate, P. Re Fiorentin, S. Regibo, P.J. Richards, C. Rios
Diaz, V. Ripepi, A. Riva, H.W. Rix, G. Rixon, N. Robichon, A.C. Robin, C. Robin, M. Roe-
lens, H.R.O. Rogues, L. Rohrbasser, M. Romero-Gómez, N. Rowell, F. Royer, D. Ruz Mieres,
K.A. Rybicki, G. Sadowski, A. Sáez Núñez, A. Sagristà Sellés, J. Sahlmann, E. Salguero,
N. Samaras, V. Sanchez Gimenez, N. Sanna, R. Santoveña, M. Sarasso, M. Schultheis, E. Sci-
acca, M. Segol, J.C. Segovia, D.Ségransan, D. Semeux, S. Shahaf, H.I. Siddiqui, A. Siebert,
L. Siltala, A. Silvelo, E. Slezak, I. Slezak, R.L. Smart, O. N. Snaith, E. Solano, F. Solitro,
D. Souami, J. Souchay, A. Spagna, L. Spina, F. Spoto, I.A. Steele, H. Steidelmüller, C. A.
Stephenson, M. Süveges, J. Surdej, L. Szabados, E. Szegedi-Elek, F. Taris, M.B. Taylor,
R. Teixeira, L. Tolomei, N. Tonello, F. Torra, J. Torra, G. Torralba Elipe, M. Trabucchi, A. T.
Tsounis, C. Turon, A. Ulla, N. Unger, M.V. Vaillant, E. van Dillen, W. van Reeven, O. Vanel,
A. Vecchiato, Y. Viala, D. Vicente, S. Voutsinas, M. Weiler, T. Wevers, Ł. Wyrzykowski,
A. Yoldas, P. Yvard, H. Zhao, J. Zorec, S. Zucker, T. Zwitter, Gaia Data Release 3 summary
of the content and survey properties. A&A 674, A1 (2023)
27. M. Cruz Reyes, R.I. Anderson, A 0.9% calibration of the Galactic Cepheid luminosity scale
based on Gaia DR3 data of open clusters and Cepheids. A&A 672, A85 (2023)
28. A. Bhardwaj, S.M. Kanbur, L.M. Macri, H.P. Singh, C.C. Ngeow, E.E.O. Ishida, Large magel-
lanic cloud near-infrared synoptic survey—III a statistical study of non-linearity in the Leavitt
Laws. MNRAS 457(2), 1644–1665 (2016)
29. P. Klagyivik, L. Szabados, Observational studies of cepheid amplitudes I period-amplitude
relationships for galactic cepheids and interrelation of amplitudes. A&A 504(3), 959–972
(2009)
30. R.I. Anderson, H. Saio, S. Ekström, C. Georgy, G. Meynet, On the effect of rotation on
populations of classical cepheids II pulsation analysis for metallicities 0.014, 0.006, and
0.002. A&A 591, A8 (2016)
31. G. De Somma, M. Marconi, S. Cassisi, V. Ripepi, A. Pietrinferni, R. Molinaro, S. Leccia,
I. Musella, Period-age-metallicity and period-age-colour-metallicity relations for classical
112 R. I. Anderson

Cepheids: an application to the Gaia EDR3 sample. MNRAS 508(1), 1473–1488 (2021)
32. V. Ripepi, G. Catanzaro, R. Molinaro, M. Marconi, G. Clementini, F. Cusano, G. De Somma,
S. Leccia, I. Musella, V. Testa, Period-luminosity-metallicity relation of classical cepheids.
A&A 642, A230 (2020)
33. C.-C. Ngeow, S.M. Kanbur, E.P. Bellinger, M. Marconi, I. Musella, M. Cignoni, Y.-H. Lin,
Period-luminosity relations for cepheid variables: from mid-infrared to multi-phase. Ap& SS
341(1), 105–113 (2012)
34. E. Hertzsprung, On the relation between period and form of the light-curve of variable stars
of the δ Cephei type. Bull. Astron. Inst. Netherlands 3, 115 (1926)
35. L. Breuval, A.G. Riess, P. Kervella, R.I. Anderson, M. Romaniello, An improved calibration
of the wavelength dependence of metallicity on the cepheid Leavitt law. ApJ 939(2), 89 (2022)
36. W.L. Freedman, B.F. Madore, Two new tests of the metallicity sensitivity of the cepheid
period-luminosity relation (the Leavitt Law). ApJ 734(1), 46 (2011)
37. J. Kennicutt, C. Robert., P.B. Stetson, A. Saha, D. Kelson, D.M. Rawson, S. Sakai, B.F.
Madore, J.R. Mould, W.L. Freedman, F. Bresolin, L. Ferrarese, H. Ford, B.K. Gibson, J.A.
Graham, M. Han, P. Harding, J.G. Hoessel, J.P. Huchra, S.M.G. Hughes, G.D. Illingworth,
L.M. Macri, R.L. Phelps, N.A. Silbermann, A.M. Turner, P.R. Wood, The hubble space tele-
scope key project on the extragalactic distance scale XIII the metallicity dependence of the
cepheid distance scale. ApJ 498(1), 181–194 (1998)
38. L. Rizzi, R.B. Tully, D. Makarov, L. Makarova, A.E. Dolphin, S. Sakai, E.J. Shaya, Tip of
the red giant branch distances II zero-point calibration. ApJ 661(2), 815–829 (2007)
39. R. Molinaro, V. Ripepi, M. Marconi, M. Romaniello, G. Catanzaro, F. Cusano, G. De Somma,
I. Musella, J. Storm, E. Trentin, Cepheid metallicity in the Leavitt law (C-MetaLL) survey—III
simultaneous derivation of the Gaia parallax offset and period-luminosity-metallicity coeffi-
cients. MNRAS 520(3), 4154–4166 (2023)
40. F. Bresolin, Revisiting the abundance gradient in the maser host galaxy NGC 4258. ApJ
729(1), 56 (2011)
41. D. Zaritsky, J. Kennicutt, C. Robert, J.P. Huchra, HII regions and the abundance properties of
spiral galaxies. ApJ 420, 87 (1994)
42. F. Bresolin, R.-P. Kudritzki, M.A. Urbaneja, The metallicity and distance of NGC 2403 from
blue supergiants. ApJ 940(1), 32 (2022)
43. R.-P. Kudritzki, M.A. Urbaneja, F. Bresolin, N. Przybilla, W. Gieren, G. Pietrzyński, Quan-
titative spectroscopy of 24 a supergiants in the sculptor galaxy NGC 300: flux-weighted
gravity-luminosity relationship, metallicity, and metallicity gradient. ApJ 681(1), 269–289
(2008)
44. M.G. Lee, W.L. Freedman, B.F. Madore, The tip of the red giant branch as a distance indicator
for resolved galaxies. ApJ 417, 553 (1993)
45. W.L. Freedman, Measurements of the hubble constant: tensions in perspective. ApJ 919(1),
16 (2021)
46. L. Eyer, Search for QSO candidates in OGLE-II data. Acta Astron. 52, 241–262 (2002)
47. Y. Ita, T. Tanabé, N. Matsunaga, Y. Nakajima, C. Nagashima, T. Nagayama, D. Kato, M.
Kurita, T. Nagata, S. Sato, M. Tamura, H. Nakaya, Y. Nakada, Pulsation at the tip of the first
giant branch? MNRAS 337(3), L31–L34 (2002)
48. R.I. Anderson, N.W. Koblischke, L. Eyer, Reconciling astronomical distance scales with
variable red giant stars (2023). arXiv e-prints: arXiv:2303.04790
49. S. Li, A.G. Riess, D. Scolnic, G.S. Anand, J. Wu, S. Casertano, W. Yuan, R. Beaton, R.I.
Anderson, Standardized luminosity of the tip of the red giant branch utilizing multiple fields
in NGC 4258 and the CATs algorithm (2023). arXiv e-prints: arXiv:2306.10103
50. D. Scolnic, A.G. Riess, J. Wu, S. Li, G.S. Anand, R. Beaton, S. Casertano, R. Anderson,
S. Dhawan, X. Ke, CATS: the hubble constant from standardized TRGB and type Ia supernova
measurements (2023). arXiv e-prints: arXiv:2304.06693
51. J. Wu, D. Scolnic, A.G. Riess, G.S. Anand, R. Beaton, S. Casertano, X. Ke, S. Li, Comparative
analysis of TRGBs (CATs) from unsupervised multi-halo-field measurements: contrast is key.
ApJ 954(1), 87 (2023)
6 On Cepheid Distances in the . H0 Measurement 113

52. B.F. Madore, V. Mager, W.L. Freedman, Sharpening the tip of the red giant branch. ApJ
690(1), 389–393 (2009)
53. R.L. Beaton, G. Bono, V.F. Braga, M. Dall’Ora, G. Fiorentino, I.S. Jang, C.E. Martínez-
Vázquez, N. Matsunaga, M. Monelli, J.R. Neeley, M. Salaris, Old-aged primary distance
indicators. Space Sci. Rev. 214(8), 113 (2018)
54. W. Baade, The period-luminosity relation of the cepheids. PASP 68(400), 5 (1956)
55. I. Soszyński, A. Udalski, M. K. Szymański, D. Skowron, G. Pietrzyński, R. Poleski,
P. Pietrukowicz, J. Skowron, P. Mróz, S. Kozłowski, Ł. Wyrzykowski, K. Ulaczyk, M. Pawlak,
The OGLE collection of variable stars: classical cepheids in the magellanic system. Acta
Astron. 65(4), 297–312 (2015)
56. P. Pietrukowicz, I. Soszyński, A. Udalski, Classical cepheids in the milky way. Acta Astron.
71(3), 205–222 (2021)
57. V. Ripepi, G. Clementini, R. Molinaro, S. Leccia, E. Plachy, L. Molnár, L. Rimoldini,
I. Musella, M. Marconi, A. Garofalo, M. Audard, B. Holl, D.W. Evans, G. Jevardat de
Fombelle, I. Lecoeur-Taibi, O. Marchal, N. Mowlavi, T. Muraveva, K. Nienartowicz, P. Sar-
toretti, L. Szabados, L. Eyer, Gaia DR3: specific processing and validation of all-sky RR
Lyrae and Cepheid stars—The Cepheid sample (2022). arXiv e-prints: arXiv:2206.06212
58. C. Alcock, R.A. Allsman, D.R. Alves, T.S. Axelrod, A.C. Becker, D.P. Bennett, D.F. Bersier,
K.H. Cook, K.C. Freeman, K. Griest, J.A. Guern, M. Lehner, S.L. Marshall, D. Minniti, B.A.
Peterson, M.R. Pratt, P.J. Quinn, A.W. Rodgers, C.W. Stubbs, W. Sutherland, A. Tomaney,
T. Vandehei, D.L. Welch, The MACHO project LMC variable star inventory VIII the recent
star formation history of the large magellanic cloud from the cepheid period distribution. AJ
117(2), 920–926 (1999)
59. V. Kovtyukh, B. Lemasle, F. Chekhonadskikh, G. Bono, N. Matsunaga, A. Yushchenko, R.I.
Anderson, S. Belik, R. da Silva, L. Inno, The chemical composition of galactic beat cepheids.
MNRAS 460(2), 2077–2086 (2016)
60. G. Csörnyei, R.I. Anderson, C. Vogl, S. Taubenberger, S. Blondin, B. Leibundgut, W. Hille-
brandt, Reeling in the Whirlpool: the distance to M 51 clarified by cepheids and the type IIP
SN 2005cs (2023). arXiv e-prints: arXiv:2305.13943
61. R.I. Anderson, S. Ekström, C. Georgy, G. Meynet, N. Mowlavi, L. Eyer, On the effect of
rotation on populations of classical cepheids I predictions at solar metallicity. A&A 564,
A100 (2014)
62. N.R. Evans, R. Szabó, A. Derekas, L. Szabados, C. Cameron, J.M. Matthews, D. Sasselov, R.
Kuschnig, J.F. Rowe, D.B. Guenther, A.F.J. Moffat, S.M. Rucinski, W.W. Weiss, Observations
of Cepheids with the MOST satellite: contrast between pulsation modes. MNRAS 446(4),
4008–4018 (2015)
63. R. Smolec, M. Śniegowska, Non-radial pulsation in first overtone cepheids of the small mag-
ellanic cloud. MNRAS 458(4), 3561–3577 (2016)
64. M. Süveges, R.I. Anderson, Investigating light-curve modulation via kernel smoothing—II
new additional modes in single-mode OGLE classical Cepheids. MNRAS 478(2), 1425–1441
(2018)
65. M. Süveges, R.I. Anderson, Investigating light curve modulation via kernel smoothing I
application to 53 fundamental mode and first-overtone cepheids in the LMC. A&A 610, A86
(2018)
66. D. Turner, M. Abdel-Sabour Abdel-Latif, L.N. Berdnikov, Using Cepheid period changes to
test stellar evolutionary models. JRASC 99(4), 144 (2005)
67. G. Csörnyei, L. Szabados, L. Molnár, B. Cseh, N. Egei, C. Kalup, V. Kecskeméthy, R.
Könyves-Tóth, K. Sárneczky, R. Szakáts, Study of changes in the pulsation period of 148
galactic cepheid variables. MNRAS 511(2), 2125–2146 (2022)
68. B.F. Madore, The period-luminosity relation IV intrinsic relations and reddenings for the large
magellanic cloud cepheids. ApJ 253, 575–579 (1982)
69. S. van den Bergh, The extragalactic distance scale, in Galaxies and the Universe, ed. by
A. Sandage, M. Sandage, J. Kristian, p. 509 (1975)
114 R. I. Anderson

70. E.L. Fitzpatrick, Correcting for the effects of interstellar extinction. PASP 111(755), 63–75
(1999)
71. E.F. Schlafly, D.P. Finkbeiner, Measuring reddening with sloan digital sky survey stellar
spectra and recalibrating SFD. ApJ 737(2), 103 (2011)
72. R.I. Anderson, Relativistic corrections for measuring Hubble’s constant to 1% using stellar
standard candles. A&A 658, A148 (2022)
73. K.D. Gordon, G.C. Clayton, K.A. Misselt, A.U. Landolt, M.J. Wolff, A quantitative com-
parison of the small magellanic cloud, large magellanic cloud, and milky way ultraviolet to
near-infrared extinction curves. ApJ 594(1), 279–293 (2003)
74. E. Mörtsell, A. Goobar, J. Johansson, S. Dhawan, The hubble tension revisited: additional
local distance ladder uncertainties. ApJ 935(1), 58 (2022)
75. M.W. Feast, R.M. Catchpole, The Cepheid period-luminosity zero-point from HIPPARCOS
trigonometrical parallaxes. MNRAS 286(1), L1–L5 (1997)
76. R. da Silva, J. Crestani, G. Bono, V.F. Braga, V. D’Orazi, B. Lemasle, M. Bergemann,
M. Dall’Ora, G. Fiorentino, P. François, M.A.T. Groenewegen, L. Inno, V. Kovtyukh, R.P.
Kudritzki, N. Matsunaga, M. Monelli, A. Pietrinferni, L. Porcelli, J. Storm, M. Tantalo,
F. Thévénin, A new and homogeneous metallicity scale for galactic classical cepheids II
abundance of iron and α elements. A&A 661, A104 (2022)
77. V. Ripepi, G. Catanzaro, G. Clementini, G. De Somma, R. Drimmel, S. Leccia, M. Marconi,
R. Molinaro, I. Musella, E. Poggio, Classical Cepheid period-Wesenheit-metallicity relation
in the Gaia bands. A&A 659, A167 (2022)
78. E. Trentin, V. Ripepi, G. Catanzaro, J. Storm, M. Marconi, G. De Somma, V. Testa, I. Musella,
Cepheid metallicity in the leavitt law (C-MetaLL) survey-II high-resolution spectroscopy of
the most metal poor galactic cepheids. MNRAS 519(2), 2331–2348 (2023)
79. A.G. Riess, S. Casertano, W. Yuan, J.B. Bowers, L. Macri, J.C. Zinn, D. Scolnic, Cosmic
distances calibrated to 1% precision with Gaia EDR3 parallaxes and hubble space telescope
photometry of 75 milky way cepheids confirm tension with ɅCDM. ApJ 908(1), L6 (2021)
80. A.G. Riess, S. Casertano, W. Yuan, L. Macri, B. Bucciarelli, M.G. Lattanzi, J.W. MacK-
enty, J.B. Bowers, W. Zheng, A.V. Filippenko, C. Huang, R.I. Anderson, Milky way cepheid
standards for measuring cosmic distances and application to Gaia DR2: implications for the
hubble constant. ApJ 861(2), 126 (2018)
81. S. Casertano, A.G. Riess, J. Anderson, R.I. Anderson, J.B. Bowers, K.I. Clubb, A.R. Cukier-
man, A.V. Filippenko, M.L. Graham, J.W. MacKenty, C. Melis, B.E. Tucker, G. Upadhya,
Parallax of galactic cepheids from spatially scanning the wide field camera 3 on the hubble
space telescope: the case of SS Canis Majoris. ApJ 825(1), 11 (2016)
82. A.G. Riess, S. Casertano, J. Anderson, J. MacKenty, A.V. Filippenko, Parallax beyond a
kiloparsec from spatially scanning the wide field camera 3 on the hubble space telescope. ApJ
785(2), 161 (2014)
83. A.G. Riess, S. Casertano, W. Yuan, L. Macri, J. Anderson, J.W. MacKenty, J.B. Bowers, K.I.
Clubb, A.V. Filippenko, D.O. Jones et al., New parallaxes of galactic cepheids from spatially
scanning the hubble space telescope: implications for the hubble constant. ApJ 855(2), 136
(2018)
84. L. Lindegren, U. Bastian, M. Biermann, A. Bombrun, A. de Torres, E. Gerlach, R. Geyer,
J. Hernández, T. Hilger, D. Hobbs, S.A. Klioner, U. Lammers, P.J. McMillan, M. Ramos-
Lerate, H. Steidelmüller, C.A. Stephenson, F. van Leeuwen, Gaia Early Data Release 3 parallax
bias versus magnitude, colour, and position. A&A 649, A4 (2021)
85. L. Lindegren, S.A. Klioner, J. Hernández, A. Bombrun, M. Ramos-Lerate, H. Steidelmüller,
U. Bastian, M. Biermann, A. de Torres, E. Gerlach, R. Geyer, T. Hilger, D. Hobbs, U. Lammers,
P.J. McMillan, C.A. Stephenson, J. Castañeda, M. Davidson, C. Fabricius, G. Gracia-Abril,
J. Portell, N. Rowell, D. Teyssier, F. Torra, S. Bartolomé, M. Clotet, N. Garralda, J.J. González-
Vidal, J. Torra, U. Abbas, M. Altmann, E. Anglada Varela, L. Balaguer-Núñez, Z. Balog,
C. Barache, U. Becciani, M. Bernet, S. Bertone, L. Bianchi, S. Bouquillon, A.G.A. Brown,
B. Bucciarelli, D. Busonero, A.G. Butkevich, R. Buzzi, R. Cancelliere, T. Carlucci, P. Charlot,
M.R.L. Cioni, M. Crosta, C. Crowley, E. F. del Peloso, E. del Pozo, R. Drimmel, P. Esquej,
6 On Cepheid Distances in the . H0 Measurement 115

A. Fienga, E. Fraile, M. Gai, M. Garcia-Reinaldos, R. Guerra, N.C. Hambly, M. Hauser,

K. Janßen, S. Jordan, Z. Kostrzewa-Rutkowska, M.G. Lattanzi, S. Liao, E. Licata, T.A. Lister,
W. Löffler, J.M. Marchant, A. Masip, F. Mignard, A. Mints, D. Molina, A. Mora, R. Morbidelli,
C.P. Murphy, C. Pagani, P. Panuzzo, X. Peñalosa Esteller, E. Poggio, P. Re Fiorentin, A. Riva,
A. Sagristà Sellés, V. Sanchez Gimenez, M. Sarasso, E. Sciacca, H.I. Siddiqui, R.L. Smart,
D. Souami, A. Spagna, I. A. Steele, F. Taris, E. Utrilla, W. van Reeven, A. Vecchiato, Gaia
Early Data Release 3 the astrometric solution. A&A 649, A2 (2021)
86. The HIPPARCOS and TYCHO catalogues. Astrometric and photometric star catalogues
derived from the ESA HIPPARCOS Space Astrometry Mission, vol. 1200. (ESA Special Pub-
lication, 1997)
87. F. van Leeuwen, M.W. Feast, P.A. Whitelock, C.D. Laney, Cepheid parallaxes and the hubble
constant. MNRAS 379(2), 723–737 (2007)
88. J.C. Zinn, Validation of the Gaia Early Data Release 3 parallax zero-point model with aster-
oseismology. AJ 161(5), 214 (2021)
89. M. Groenewegen, Primary period-luminosity-relation calibrators in the milky way: cepheids
and RR Lyrae physical basis, calibration, and applications (2023). arXiv e-prints:
arXiv:2307.03033
90. S. Khan, R.I. Anderson, A. Miglio, B. Mosser, Y.P. Elsworth, Investigating gaia edr3 par-
allax systematics using asteroseismology of cool giant stars observed by kepler, k2, and
tess ii. deciphering gaia parallax systematics using red clump stars (2023). arXiv e-prints:
arXiv:2310.03654
91. J. Maíz Apellániz, An estimation of the Gaia EDR3 parallax bias from stellar clusters and
magellanic clouds data. A&A 657, A130 (2022)
92. G. Pietrzyński, D. Graczyk, A. Gallenne, W. Gieren, I.B. Thompson, B. Pilecki, P. Karcz-
marek, M. Górski, K. Suchomska, M. Taormina, B. Zgirski, P. Wielgórski, Z. Kołaczkowski,
P. Konorski, S. Villanova, N. Nardetto, P. Kervella, F. Bresolin, R.P. Kudritzki, J. Storm, R.
Smolec, W. Narloch, A distance to the large magellanic cloud that is precise to one per cent.
Nature 567(7747), 200–203 (2019)
93. D. Graczyk, G. Pietrzyński, I.B. Thompson, W. Gieren, B. Zgirski, S. Villanova, M. Górski, P.
Wielgórski, P. Karczmarek, W. Narloch, B. Pilecki, M. Taormina, R. Smolec, K. Suchomska,
A. Gallenne, N. Nardetto, J. Storm, R.-P. Kudritzki, M. Kałuszyński, W. Pych, A distance
determination to the small magellanic cloud with an accuracy of better than two percent based
on late-type eclipsing binary stars. ApJ 904(1), 13 (2020)
94. L. Breuval, A.G. Riess, S. Casertano, W. Yuan, L.M. Macri, M. Romaniello, Y.S. Murakami,
D. Scolnic, G.S. Anand, I. Soszyński, Small Magellanic Cloud Cepheids Observed with the
Hubble Space Telescope Provide a New Anchor for the SH0 ES Distance Ladder (2024).
https://doi.org/10.48550/arXiv.2404.08038. ArXiv: 2404.08038
95. A.G. Riess, S. Casertano, W. Yuan, L.M. Macri, D. Scolnic, Large magellanic cloud cepheid
standards provide a 1% foundation for the determination of the hubble constant and stronger
evidence for physics beyond ɅCDM. ApJ 876(1), 85 (2019)
96. L.M. Macri, C.C. Ngeow, S.M. Kanbur, S. Mahzooni, M.T. Smitka, Large magellanic cloud
near-infrared synoptic survey I cepheid variables and the calibration of the Leavitt Law. AJ
149(4), 117 (2015)
97. V. Ripepi, L. Chemin, R. Molinaro, M.R.L. Cioni, K. Bekki, G. Clementini, R. de Grijs,
G. De Somma, D. El Youssoufi, L. Girardi, M.A.T. Groenewegen, V. Ivanov, M. Marconi,
P.J. McMillan, J.T. van Loon, The VMC survey—XLVIII classical cepheids unveil the 3D
geometry of the LMC. MNRAS 512(1), 563–582 (2022)
98. I. Soszyński, A. Udalski, M.K. Szymański, P. Pietrukowicz, J. Skowron, D.M. Skowron, R.
Poleski, S. Kozłowski, P. Mróz, K. Ulaczyk, K. Rybicki, P. Iwanek, M. Wrona, Final release
of the OGLE collection of cepheids and RR Lyrae stars in the magellanic system the outer
regions. Acta Astron. 69(2), 87–99 (2019)
99. M. Romaniello, A. Riess, S. Mancino, R.I. Anderson, W. Freudling, R.P. Kudritzki, L. Macrì,
A. Mucciarelli, W. Yuan, The iron and oxygen content of LMC classical cepheids and its
implications for the extragalactic distance scale and hubble constant: equivalent width analysis
with Kurucz stellar atmosphere models. A&A 658, A29 (2022)
116 R. I. Anderson

100. V. Scowcroft, W.L. Freedman, B.F. Madore, A. Monson, S.E. Persson, J. Rich, M. Seibert,
J.R. Rigby, The carnegie hubble program: the distance and structure of the SMC as revealed
by mid-infrared observations of cepheids. ApJ 816(2), 49 (2016)
101. E.M.L. Humphreys, M.J. Reid, J.M. Moran, L.J. Greenhill, A.L. Argon, Toward a new geo-
metric distance to the active galaxy NGC 4258 III final results and the hubble constant. ApJ
775(1), 13 (2013)
102. M.J. Reid, D.W. Pesce, A.G. Riess, An improved distance to NGC 4258 and its implications
for the hubble constant. ApJ 886(2), L27 (2019)
103. W. Yuan, L.M. Macri, A.G. Riess, T.G. Brink, S. Casertano, A.V. Filippenko, S.L. Hoffmann,
C.D. Huang, D. Scolnic, Absolute calibration of cepheid period-luminosity relations in NGC
4258. ApJ 940(1), 64 (2022)
104. M.M. Fausnaugh, C.S. Kochanek, J.R. Gerke, L.M. Macri, A.G. Riess, K.Z. Stanek, The
Cepheid distance to the maser-host galaxy NGC 4258: studying systematics with the large
binocular telescope. MNRAS 450(4), 3597–3619 (2015)
105. B.J. Shappee, K.Z. Stanek, A new cepheid distance to the giant spiral M101 based on image
subtraction of hubble space telescope/advanced camera for surveys observations. ApJ 733(2),
124 (2011)
106. B.F. Madore, W.L. Freedman, A.J. Lee, Astrophysical distance scale IV preliminary zero-
point calibration of the JAGB method in the HST/WFC3-IR broad J-band (F110W) filter. ApJ
926(2), 153 (2022)
107. B. Zgirski, G. Pietrzyński, W. Gieren, M. Górski, P. Wielgórski, P. Karczmarek, F. Bresolin,
P. Kervella, R.P. Kudritzki, J. Storm, D. Graczyk, G. Hajdu, W. Narloch, B. Pilecki,
K. Suchomska, M. Taormina, The Araucaria project: distances to nine galaxies based on
a statistical analysis of their carbon stars (JAGB Method). ApJ 916(1), 19 (2021)
108. L. Ferrarese, N.A. Silbermann, J.R. Mould, P.B. Stetson, A. Saha, W.L. Freedman, J. Kenni-
cutt, C. Robert, Photometric recovery of crowded stellar fields observed with HST/WFPC2
and the effects of confusion noise on the extragalactic distance scale. PASP 112(768), 177–201
(2000)
109. A.G. Riess, L. Macri, S. Casertano, H. Lampeitl, H.C. Ferguson, A.V. Filippenko, S.W. Jha,
W. Li, R. Chornock, A 3% solution: determination of the hubble constant with the hubble
space telescope and wide field camera 3. ApJ 730(2), 119 (2011)
110. R.I. Anderson, A.G. Riess, On cepheid distance scale bias due to stellar companions and
cluster populations. ApJ 861(1), 36 (2018)
111. L. Breuval, A.G. Riess, L.M. Macri, S. Li, W. Yuan, S. Casertano, T. Konchady, B. Trahin,
M.J. Durbin, B.F. Williams, A 1.3% distance to M33 from hubble space telescope cepheid
photometry. ApJ 951(2), 118 (2023)
112. R.I. Anderson, Towards a 1% measurement of the Hubble constant: accounting for time
dilation in variable-star light curves. A&A 631, A165 (2019)
113. A. Dolphin, DOLPHOT: stellar photometry. Astrophysics Source Code Library, record
ascl:1608.013 (2016)
114. P.B. Stetson, DAOPHOT: crowded-field stellar photometry package. Astrophysics Source
Code Library, record ascl:1104.011 (2011)
115. B. Javanmardi, A. Mérand, P. Kervella, L. Breuval, A. Gallenne, N. Nardetto, W. Gieren, G.
Pietrzyński, V. Hocdé, S. Borgniet, Inspecting the cepheid distance ladder: the hubble space
telescope distance to the SN Ia host galaxy NGC 5584. ApJ 911(1), 12 (2021)
116. W. Baade, Über eine Möglichkeit, die Pulsationstheorie der δ Cephei-Veränderlichen zu
prüfen. Astron. Nachr. 228(20), 359 (1926)
117. F.A. Lindemann, Note on the pulsation theory of cepheid variables. MNRAS 78, 639 (1918)
118. G. Tiercy, G. Abetti, Recherches sur SU Cassiopeiae et les Cepheides a courtes periodes.
Osservazioni e memorie dell’Osservatorio astrofisico di Arcetri 44, 1 (1927)
119. P. Kervella, V. Coudé du Foresto, W.A. Traub, M.G. Lacasse, Interferometric observations
of the Cepheid ζ Geminorum with FLUOR/IOTA, in Working on the Fringe: Optical and IR
Interferometry from Ground and Space, ed. by S. Unwin, R. Stachnik. Astronomical Society
of the Pacific Conference Series, vol. 194, p. 22 (1999)
6 On Cepheid Distances in the . H0 Measurement 117

120. B.F. Lane, M.J. Kuchner, A.F. Boden, M. Creech-Eakman, S.R. Kulkarni, Direct detection of
pulsations of the cepheid star ζ Gem and an independent calibration of the period-luminosity
relation. Nature 407(6803), 485–487 (2000)
121. D. Mourard, D. Bonneau, L. Koechlin, A. Labeyrie, F. Morand, P. Stee, I. Tallon-Bosc, F. Vak-
ili, The mean angular diameter of δ cephei measured by optical long-baseline interferometry.
A&A 317, 789–792 (1997)
122. P. Kervella, D. Bersier, D. Mourard, N. Nardetto, P. Fouqué, V. Coudé du Foresto, Cepheid
distances from infrared long-baseline interferometry. III. Calibration of the surface brightness-
color relations. A&A 428, 587–593 (2004)
123. A. Mérand, P. Kervella, J. Breitfelder, A. Gallenne, V. Coudé du Foresto, T.A. ten Brummelaar,
H.A. McAlister, S. Ridgway, L. Sturmann, J. Sturmann, N.H. Turner, Cepheid distances from
the spectrophoto-interferometry of pulsating stars (SPIPS): application to the prototypes δ
cephei and η aquilae. A&A 584, A80 (2015)
124. B. Trahin, L. Breuval, P. Kervella, A. Mérand, N. Nardetto, A. Gallenne, V. Hocdé, W. Gieren,
Inspecting the cepheid parallax of pulsation using Gaia EDR3 parallaxes: projection factor
and period-luminosity and period-radius relations. A&A 656, A102 (2021)
125. W. Gieren, J. Storm, P. Konorski, M. Górski, B. Pilecki, I. Thompson, G. Pietrzyński, D.
Graczyk, T.G. Barnes, P. Fouqué, N. Nardetto, A. Gallenne, P. Karczmarek, K. Suchomska, P.
Wielgórski, M. Taormina, B. Zgirski, The effect of metallicity on cepheid period-luminosity
relations from a Baade-Wesselink analysis of cepheids in the milky way and magellanic clouds
★. A&A 620, A99 (2018)
126. J. Storm, W. Gieren, P. Fouqué, T. G. Barnes, G. Pietrzyński, N. Nardetto, M. Weber,
T. Granzer, K.G. Strassmeier, Calibrating the Cepheid period-luminosity relation from the
infrared surface brightness technique I the p-factor, the Milky Way relations, and a universal
K-band relation. A&A 534, A94 (2011)
127. J. Storm, W. Gieren, P. Fouqué, T. G. Barnes, I. Soszyński, G. Pietrzyński, N. Nardetto,
D. Queloz, Calibrating the Cepheid period-luminosity relation from the infrared surface
brightness technique II the effect of metallicity and the distance to the LMC. A&A 534,
A95 (2011)
128. R.I. Anderson, Discovery of cycle-to-cycle modulated spectral line variability and velocity
gradients in long-period cepheids. MNRAS 463(2), 1707–1723 (2016)
129. R.I. Anderson, A. Mérand, P. Kervella, J. Breitfelder, J.B. LeBouquin, L. Eyer, A. Gallenne, L.
Palaversa, T. Semaan, S. Saesen, N. Mowlavi, Investigating cepheid l carinae’s cycle-to-cycle
variations via contemporaneous velocimetry and interferometry. MNRAS 455(4), 4231–4248
(2016)
130. N. Nardetto, A. Fokin, D. Mourard, P. Mathias, P. Kervella, D. Bersier, Self consistent mod-
elling of the projection factor for interferometric distance determination. A&A 428, 131–137
(2004)
131. N. Nardetto, D. Mourard, P. Mathias, A. Fokin, D. Gillet, High-resolution spectroscopy for
Cepheids distance determination II: a period-projection factor relation. A&A 471(2), 661–669
(2007)
132. J. Breitfelder, A. Mérand, P. Kervella, A. Gallenne, L. Szabados, R.I. Anderson, J.B. Le
Bouquin, Observational calibration of the projection factor of cepheids II application to nine
cepheids with HST/FGS parallax measurements. A&A 587, A117 (2016)
133. R.I. Anderson, G. Viviani, S. Shreeya Shetye, N. Mowlavi, L. Eyer, L. Palaversa, B. Holl,
S. Blanco-Cuaresma, K. Kravchenko, M. Pawlak, M. Cruz Reyes, S. Khan, H.E. Netzel, L.
Löbling, P.I. Pápics, A. Postel, M. Roelens, Z.T. Spetsieri, A. Thoul, J. Zák, V. Bonvin, D.V.
Martin, M. Millon, S. Saesen, A. Wyttenbach, P. Figueira, M. Marmier, S. Prins, G. Raskin,
H. van Winckel, VELOcities of CEpheids (VELOCE) I. High-precision Radial Velocities of
Cepheids (2024). https://doi.org/10.48550/arXiv.2404.12280. ArXiv: 2404.12280
134. N.R. Evans, L. Berdnikov, J. Lauer, D. Morgan, J. Nichols, H.M. Günther, N. Gorynya,
A. Rastorguev, P. Moskalik, Binary properties from cepheid radial velocities (CRaV). AJ
150(1), 13 (2015)
118 R. I. Anderson

135. P. Kervella, A. Gallenne, N.R. Evans, L. Szabados, F. Arenou, A. Mérand, N. Nardetto,

W. Gieren, G. Pietrzynski, Multiplicity of galactic cepheids and RR Lyrae stars from Gaia
DR2 II resolved common proper motion pairs. A&A 623, A117 (2019)
136. P. Kervella, A. Gallenne, N. Remage Evans, L. Szabados, F. Arenou, A. Mérand, Y. Proto,
P. Karczmarek, N. Nardetto, W. Gieren, G. Pietrzynski, Multiplicity of galactic cepheids and
RR Lyrae stars from Gaia DR2 I binarity from proper motion anomaly. A&A 623, A116
(2019)
137. M. Moe, R. Di Stefano, Mind your Ps and Qs: the interrelation between period (P) and
mass-ratio (Q) distributions of binary stars. ApJS 230(2), 15 (2017)
138. R.I. Anderson, S. Casertano, A.G. Riess, C. Melis, B. Holl, T. Semaan, P.I. Papics, S. Blanco-
Cuaresma, L. Eyer, N. Mowlavi, L. Palaversa, M. Roelens, Vetting galactic leavitt law cal-
ibrators using radial velocities: on the variability, binarity, and possible parallax error of 19
long-period cepheids. ApJS 226(2), 18 (2016)
139. A. Gallenne, P. Kervella, N.R. Evans, C.R. Proffitt, J.D. Monnier, A. Mérand, E. Nelan,
E. Winston, G. Pietrzyński, G. Schaefer, W. Gieren, R.I. Anderson, S. Borgniet, S. Kraus,
R.M. Roettenbacher, F. Baron, B. Pilecki, M. Taormina, D. Graczyk, N. Mowlavi, L. Eyer, A
geometrical 1% distance to the short-period binary cepheid V1334 Cygni. ApJ 867(2), 121
(2018)
140. B. Pilecki, G. Pietrzyński, R.I. Anderson, W. Gieren, M. Taormina, W. Narloch, N.R. Evans,
J. Storm, Cepheids with giant companions I revealing a numerous population of double-lined
binary cepheids. ApJ 910(2), 118 (2021)
141. B. Pilecki, I.B. Thompson, F. Espinoza-Arancibia, R.I. Anderson, W. Gieren, W. Narloch,
J. Minniti, G. Pietrzyński, M. Taormina, G. Bono, G. Hajdu, Discovery of a binary-origin
classical cepheid in a binary system with a 59 day orbital period. ApJ 940(2), L48 (2022)
142. P. Senchyna, L.C. Johnson, J.J. Dalcanton, L.C. Beerman, M. Fouesneau, A. Dolphin, B.F.
Williams, P. Rosenfield, S.S. Larsen, Panchromatic hubble andromeda treasury XIV the
period-age relationship of cepheid variables in M31 star clusters. ApJ 813(1), 31 (2015)
143. F. Dinnbier, R.I. Anderson, P. Kroupa, On the dynamical evolution of Cepheids in star clusters.
A&A 659, A169 (2022)
144. F. Dinnbier, P. Kroupa, R.I. Anderson, Do the majority of stars form as gravitationally
unbound? A&A 660, A61 (2022)
145. P. Barmby, M. Marengo, N.R. Evans, G. Bono, D. Huelsman, K.Y.L. Su, D.L. Welch, G.G.
Fazio, Galactic cepheids with spitzer II search for extended infrared emission. AJ 141(2), 42
(2011)
146. M. Marengo, N.R. Evans, P. Barmby, L.D. Matthews, G. Bono, D.L. Welch, M. Romaniello,
D. Huelsman, K.Y.L. Su, G.G. Fazio, An infrared nebula associated with δ cephei: evidence
of mass loss? ApJ 725(2), 2392–2400 (2010)
147. L.D. Matthews, M. Marengo, N.R. Evans, G. Bono, New evidence for mass loss from δ Cephei
from HI 21 cm line observations. ApJ 744(1), 53 (2012)
148. A. Gallenne, A. Mérand, P. Kervella, O. Chesneau, J. Breitfelder, W. Gieren, Extended
envelopes around galactic cepheids IV T Monocerotis and X Sagittarii from mid-infrared
interferometry with VLTI/MIDI. A&A 558, A140 (2013)
149. P. Kervella, A. Mérand, G. Perrin, V. Coudé du Foresto, Extended envelopes around galactic
cepheids I. l carinae from near and mid-infrared interferometry with the VLTI. A&A 448(2),
623–631 (2006)
150. A. Mérand, P. Kervella, V. Coudé du Foresto, G. Perrin, S.T. Ridgway, J.P. Aufdenberg,
T.A. Ten Brummelaar, H.A. McAlister, L. Sturmann, J. Sturmann, N.H. Turner, D.H. Berger,
Extended envelopes around galactic cepheids II polaris and δ Cephei from near-infrared
interferometry with CHARA/FLUOR. A&A 453(1), 155–162 (2006)
151. M.A.T. Groenewegen, Analysing the spectral energy distributions of galactic classical
cepheids. A&A 635, A33 (2020)
152. M.A.T. Groenewegen, J. Lub, Spectral energy distributions of classical cepheids in the mag-
ellanic clouds. A&A 676, A136 (2023)
6 On Cepheid Distances in the . H0 Measurement 119

153. E.G. Schmidt, Excess mid-infrared flux: an indicator of mass loss in cepheids? ApJ 813(1),
29 (2015)
154. V. Hocdé, N. Nardetto, E. Lagadec, G. Niccolini, A. Domiciano de Souza, A. Mérand, P.
Kervella, A. Gallenne, M. Marengo, B. Trahin, W. Gieren, G. Pietrzyński, S. Borgniet, L.
Breuval, B. Javanmardi, A thin shell of ionized gas as the explanation for infrared excess
among classical Cepheids. A&A 633, A47 (2020)
155. G. De Somma, M. Marconi, R. Molinaro, V. Ripepi, S. Leccia, I. Musella, An updated metal-
dependent theoretical scenario for classical cepheids. ApJS 262(1), 25 (2022)
156. G. Fiorentino, I. Musella, M. Marconi, Cepheid theoretical models and observations in
HST/WFC3 filters: the effect on the Hubble constant H0 . MNRAS 434(4), 2866–2876 (2013)
157. M. Marconi, I. Musella, G. Fiorentino, Cepheid pulsation models at varying metallicity and
ΔY/ΔZ. ApJ 632(1), 590–610 (2005)
158. A. Bhardwaj, A.G. Riess, G. Catanzaro, E. Trentin, V. Ripepi, M. Rejkuba, M. Marconi, C.C.
Ngeow, L.M. Macri, M. Romaniello, R. Molinaro, H.P. Singh, S.M. Kanbur, High-resolution
spectroscopic metallicities of milky way cepheid standards and their impact on the Leavitt
Law and the hubble constant (2023). arXiv e-prints: arXiv:2309.03263
159. E. Trentin, V. Ripepi, G. Catanzaro, J. Storm, M. Marconi, G. De Somma, V. Testa, I. Musella,
Cepheid metallicity in the leavitt law (C-MetaLL) survey: II High-resolution spectroscopy of
the most metal poor galactic cepheids. arXiv e-prints: arXiv:2209.03792 (2022)
160. P. Wielgórski, G. Pietrzyński, W. Gieren, M. Górski, R.-P. Kudritzki, B. Zgirski, F. Bresolin,
J. Storm, N. Matsunaga, D. Graczyk, I. Soszyński, A precision determination of the effect of
metallicity on cepheid absolute magnitudes in VIJHK bands from magellanic cloud cepheids.
ApJ 842(2), 116 (2017)
Chapter 7
The Role of Type Ia Supernovae
in Constraining the Hubble Constant

Dan Scolnic and Maria Vincenzi

7.1 Introduction

When Edwin Hubble first measured the expansion of the universe, no supernovae
(SNe) were used in this measurement [1]. Instead, Hubble measured distances to
galaxies by observing the Cepheids in the galaxy and Cepheids’ properties discovered
by Henrietta Leavitt [2]. Today, both Cepheids and SNe are often used together to
infer what is now called the Hubble constant . H0 , which parameterizes the relation
between expansion velocity and distance such that .v = H0 d. The Cepheids allow for
a projection of a geometric scale, like that from parallax or megamasers, out onto
the Megaparsec scale. However, the range that Cepheids can be found and measured
with the Hubble Space Telescope is limited to .∼ 40 Mpc, where the measurements
of motions of galaxies are still strongly affected by gravitational pulls of nearby
galaxies (motions called peculiar velocities). Therefore, SNe are needed as an extra
rung along the distance ladder to reach out further into what is now deemed the
‘Hubble flow’ (.∼100 to 600 Mpc), and precisely measure the expansion rate of the
universe.
In the typical three rung distance ladder, SNe are used in the second and third rung.
In the second rung, their luminosities are calibrated with stellar distance indicators
in the same galaxy. In the third rung, their brightnesses calibrate the Hubble relation.
Ideally, one could remove the intermediate step, like Cepheids or Tip of the Red Giant
Branch (TRGB), and go straight from geometric calibration to SNe. Unfortunately,
the rate of SNe in the local universe is not nearly frequent enough to provide multiple
SNe in the few galaxies used for geometric anchors (the Milky Way, the Large
Magellanic Cloud, NGC 4258). Even with Cepheids or TRGB as a go-between, the

D. Scolnic (B) · M. Vincenzi (B)

Department of Physics, Duke University, Durham, NC 27708, USA
e-mail: daniel.scolnic@duke.edu
M. Vincenzi
e-mail: maria.vincenzi@duke.edu

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 121
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_7
122 D. Scolnic and M. Vincenzi

low rate of SNe in the nearby universe (roughly one per galaxy per 100 years) is
the limiting component of the precision of . H0 measurements [3]. At this point, the
S. H0 ES team [3] for instance, utilizes every SN Ia that pass cosmological quality
requirements and Cepheid suitability within .40 Mpc.
SNe Ia are used in the most cited local distance ladder analyses due to their high
precision, large luminosity, and number of discovered sources. However, as we will
discuss in this review, studies have attempted to replace SNe Ia with hydrogen-
rich core collapse SNe [4], which while being less-precise standardizable candles,
still provide an effective crosscheck. Along the same lines, the role of SNe can
be replaced with measurements of Surface Brightness Fluctuations [5] or the Tully
Fisher relation [6], a different type of standard candle that allows reach into the
Hubble Flow. With a similar purpose, there have been a tremendous number of re-
analyses and studies of potential systematics of analyses of SNe Ia. These SNe have
been used in cosmological measurements for decades, so there is no shortage of
checks on their accuracy.
A related benefit of SNe Ia in particular is their usage for measurements of the
expansion history, beyond just the current expansion rate. There are multiple large
surveys that have been charged over the last two decades to discover and measure
SNe Ia in order to constrain properties of dark energy and the dark matter content
of the Universe [7–11]. Leverage in understanding these properties comes from
modelling the change of brightnesses of SNe Ia over a range of redshift (like within
.0 < z < 2). As such, low redshift SNe are critically important, and the SNe that
make up the third rung of the distance ladder are the same that provide the anchor for
the relative distance constraints for measurements of expansion history [8, 9, 11].
Thereby, progress made in growing and understanding samples of low redshift SNe
has been directly beneficial to both . H0 measurements of dark energy / dark matter
measurements.
For this review, we first discuss in Sect. 7.2 the common three-rung path to con-
straining . H0 using SNe Ia. We then discuss in Sect. 7.3 the top systematic uncer-
tainties in this measurement on the SN side, and how they have been quantified.
In Sects. 7.4 and 7.5, we present an accounting of the different crosschecks, using
replacements or substitutions of the SNe sample. In Sect. 7.6, we discuss the impli-
cations of the inverse distance ladder approach. Finally, we present our discussions
and conclusions in Sects. 7.7 and 7.8.

7.2 The Canonical Path to . H0 with Type Ia Supernovae

We review here the formalism for deriving the Hubble constant with SNe Ia in the
local distance ladder, as measured in [11, 12] and used in [3]. We follow the typical
three-rung ladder, as shown on the left hand side of Fig. 7.1. As also shown in Fig. 7.1,
the third rung of the distance ladder is the low-redshift (.z < 0.15) part of the Hubble
diagram, used to measure cosmological parameters like the equation-of-state of dark
energy .w.
7 The Role of Type Ia Supernovae … 123

Fig. 7.1 Left panel: The ‘distance ladder’ from [3]. Right panel: The Pantheon+ “Hubble diagram”
from [11] showing the distance modulus .μ versus redshift .z. The SNe from the second rung make
up the majority of the .z < 0.01 SNe in the Pantheon+ Hubble diagram, and the third rung makes
up the rest of the .z < 0.15 SNe. The uncertainties for .z < 0.01 shown on the Hubble diagram are
much larger than those in the second rung, due to propagation of redshift uncertainties

We assume that a set of Cepheid or TRGB distances are calibrated with geometric
measurements like parallax or megamasers. This set of Cepheid or TRGB distances,
expressed as .μ0 here, can then be compared to SNe Ia brightnesses .m X in order to
find a single offset . M B , which describes the absolute magnitude of a SN Ia.
In most recent cosmological analyses with SNe Ia, the standardized brightness is
measured with the Tripp formula such that

m X = m B + αx1 βc − δ Bias + δ H ost ,

. (7.1)

where .m B , .x1 and .c are all independent properties of each light curve, .α and .β
are correlation coefficients that help standardize the brightness, .δ Bias is a correction
due to selection effects and other biases as predicted by simulations, and .δ H ost is a
final correction due to residual correlations with host galaxy properties. We call the
standardized brightness .m X instead of .m B as in [3] to be clear that the brightness
is standardized. We note that typically SN Ia analyses like that in [11] will subtract
off an absolute luminosity . M B of SNe Ia to derive a distance modulus, but this must
assume a fiducial . H0 value, which is what we are trying to derive.
124 D. Scolnic and M. Vincenzi

For a SN Ia in the .i-th Cepheid host,

.m X,i = μ0,i + M B , (7.2)

where . M B is the fiducial SN Ia luminosity, and .μ0,i is the distance derived from
Cepheid measurements for each galaxy.
The ladder is completed with a set of SNe Ia that measure the expansion rate
quantified as the intercept, .a B , of the distance (or magnitude)–redshift relation. For
an arbitrary expansion history and for .z > 0 as
{ }
1 1[ ]
a = log cz 1 + [1 − q0 ] z −
. B 1 − q0 − 3q02 + j0 z 2 + O(z 3 ) − 0.2m 0X ,
2 6
(7.3)
measured from a set of SNe Ia (.z, m 0X ), where .z is the redshift due to expansion, .q0
is the deceleration parameter, and . j0 is the jerk parameter. Typically, for .ɅCDM, . j0
is set to 1. The determination of . H0 follows from

. log H0 = 0.2M B0 +a B +5. (7.4)

The approach of [13, 14] is now often used to account for covariance between
rungs, as measurements of SNe Ia used in the second and third rung are correlated.
Importantly,.q0 (and. j0 ) can only be constrained from SN Ia data, without the require-
ment of any additional information.

7.3 Top Systematics on the Path to . H0 with SN Ia

Cosmological measurements using SNe Ia are affected by various source of system-

atic uncertainties. However, measurements of .w using SNe are significantly more
sensitive than measurements of . H0 with SNe. A simple explanation of the sensitivity
is that . H0 is constrained by comparing SNe at .z ∼ 0.005 to SNe in the Hubble flow at
. z ∼ 0.05, not too different in redshift and therefore similar in terms of SN properties

and survey methods and telescopes used for SN observations (see Fig. 7.2). On the
other hand, .w is constrained by comparing SNe at .z ∼ 0.05 to .z ∼ 0.5, in which evo-
lution of SN properties is possible, and the surveys and telescopes used to find and
measure the SNe will be quite different. While new surveys like Foundation [10, 15],
ZTF [16] and DEBASS are attempting to measure more low-redshift SNe with the
same telescopes as high-redshift SNe have been measured, it will take on the order
of 10–20 years for these surveys to measure an equivalent amount of second-rung
SNe (.∼ 40) as those in the current second rung. One route to overcome this hurdle
is by extending the local volume for which host galaxy distances can be obtained,
e.g., [17].
There have been a number of papers that have focused on systematics impact on
. H0 and we list these in the top of Table 7.1. [11] give a comprehensive overview of
7 The Role of Type Ia Supernovae … 125

Fig. 7.2 Adapted from [3], a comparison of SN Ia light-curve shape (.x1 ), colour (.c) and the log host
galaxy stellar mass (log . Mʘ ) for calibrator and Hubble flow samples (with different host galaxy
type and redshift cuts)

many of these systematics and how they may affect measurements of .w or . H0 . In

Fig. 7.3, from [11], we show the impact in units of . H0 for applying .1σ shift of each
systematic. We group the systematics into various categories: Redshifts, SN Physics,
Selection, Calibration, and Milky Way Dust.
Before discussing in more details each category of systematics, it is important to
understand which are the systematics . H0 is more sensitive to and, on the contrary,
which are generally mitigated and have little impact on . H0 . The distance ladder is
constructed such that the same probe is used in two of the three rungs (e.g. Cepheids
in 1 and 2, SN Ia in 2 and 3). If a systematic is disproportionately affecting SN Ia
in the second and third rung, or if it is introducing some substantial differences in
the second rung and third rung SN Ia population, this will have a large impact on
. H0 . On the contrary, if a source of systematic is introducing some consistent offset

that affects all SN Ia coherently, the impact of that systematic will be null due to the
formalism described above. In other words, the offset will cancel and the effect of
the systematic will be mitigated.
Calibration: While calibration dominates the systematic error budget for .w, its
impact is small for . H0 because of the systematic mitigation discussed above. An
illustration of this is given in [20, 30]. In particular, [20] check the possibility of
‘gray’ calibration offsets per survey and finds a total uncertainty due to this effect no
larger than 0.2 km/s/Mpc. They show that since many of the same surveys are used
to measure SNe in the second and third rung of the distance ladder, the impact of
potential biases in photometric calibration of the surveys will cancel. This is shown
to not be the case when different surveys are used for the second and third rung. This
approach is done in [22, 31], which while emphasizing the role of the CSP SNe,
increases the sensitivity to calibration errors.
126 D. Scolnic and M. Vincenzi

Table 7.1 A summary of the various crosschecks and systematics on the supernova component of
the distance ladder. If two uncertainties are given, the first one is the statistical uncertainty and the
second one is the systematic uncertainty
References Notes Result (km/s/Mpc)
Specific systematic check for optical SNIa:
Murakami et al. [18] Uses spectral feature twinning process to improve standardization; .73.01 ± 0.92
checks dust modelling, intrinsic scatter modelling;
Peterson et al. [19] Checks different models of peculiar velocities/bulk flows .σ H < 0.2
0
Brownsberger et al. [20] Checks SNIa Calibration by allowing individual SN survey offsets .σ H0 < 0.2
Jones et al. [21] Checks impact of mass step, global vs local correlations .σ H < 0.15
0
Burns et al. [22] Checks light-curve fitting method; also does NIR fits .73 ± 2
Garnavich et al. [23] Uses 4 rung distance ladder, checks SNIa host demographic .74.6 ± 0.9 ± 2.7
systematic
Dhawan et al. [16] Uses ZTF data alone, check on SNIa calibration .76.94 ± 6.4
NIR SN Ia:
Dhawan et al. [24] Uses literature NIR SN (restframe . J ) and peak fitting .72.8 ± 2.8
Galbany et al. [25] Uses literature NIR SN (restframe . J and . H band) .72.3 ± 1.4 ± 1.4
Jones et al. [26] Uses RAISIN.\+literature NIR SN (restframe .Y band) and .75.9 ± 2.2
SNooPy fitting, check on dust
Dhawan et al. [27] Uses literature NIR SN and BayesN fitting .74.82 ± 0.97 ± 0.84
Removing SNIa:
Kenworthy et al. [28] Eliminates SNIa rung entirely with 2 rung distance ladder .72.9+2.4
−2.2
De Jaeger et al. [4] Uses SN II instead of SNIa .77.6+5.2
−4.8
Blakeslee et al. [5] Uses SBF standard candles instead of SNIa .73.3 ± 0.7 ± 2.4
Kourkchi et al. [29] Uses TF relation instead of SNIa .76.0 ± 2.3 ± 1.5
Schombert et al. [6] Uses TF relation .75.1 ± 1.1 ± 2.3

Redshifts: . H0 measurements are more sensitive to redshift related systematic

because redshift information is only used in one rung of the distance ladder (the
third), so there is no systematic mitigation. [19] explore a large number of models
of bulk flows in the nearby universe, given various models that best improve the
Hubble residual scatter from supernova measurements, and find that there could be
changes in. H0 up to 0.2 km/s/Mpc. They also find that the impact of including peculiar
velocity corrections of the nominal method versus not including them is .ΔH0 ∼ 0.5.
Peculiar velocities are discussed further in chapters of this book. [32] check biases
due in redshift measurements and find uncertainties on the level of 0.1 km/s/Mpc.
A predecessor study was done by [33], which similarly found small changes to . H0 ,
though saw possible discrepancies in .Ω M H02 plane between subsamples of SNe for
measured redshifts with different levels of precision. No evidence for subsample
differences was found by [32].
SN Physics and Selection: SN intrinsic astrophysics and the role of SN dust remain
one of the most poorly understood aspects of SN Ia cosmology. Recent analyses
have investigated how systematics due to uncertainties in the physics of the supernova
explosions and extragalactic dust extinction impact . H0 measurements [18, 34]. Most
7 The Role of Type Ia Supernovae … 127

Fig. 7.3 Adapted from [11], the impact on recovery of . H0 of the various systematic uncertainties
tabulated. The units of these measurements are km/s/Mpc. The dashed lines are given at .ΔH0 of
0.7, which is the entire contribution of the uncertainty in [3] from SN measurements. We add on
labels explaining categories of systematic uncertainties

analyses have found that these effects have a small impact on . H0 because the SN sub-
populations selected in the second and third rung are not expected to be significantly
different.
A good example of a potential systematic related to SN physics was discussed in
[35] following earlier S. H0 ES analyses (e.g., [36]). [35] showed evidence for a cor-
relation between standardized brightness and the age of the host galaxy (quantified
estimating the specific star-formation at the SN location). In earlier measurements
like [36], the third rung of SNe had no galaxy-based selection applied, but the sec-
ond rung, by tying to Cepheid discovery, favored star-forming galaxies. This would
potentially lead to a bias in the recovery of . H0 . The size of the bias would depend
on the relative differential fraction of host-galaxy demographics between the second
and third rung multiplied by the size of the correlation. Subsequent analyses [21]
showed that this effect would likely be insufficient to explain the Hubble tension.
Still, in the most recent S. H0 ES analysis [3], the selection of SNe in the third
rung of the distance ladder was done to be as similar as possible as the second
rung. The S. H0 ES team only selected SNe found in star-forming galaxies which
thereby removed the sensitivity to this systematic; the impact from this change was
less than the statistical uncertainty from the supernova component of the distance
ladder. Similarly, significant differences in dust extinction and/or color-related effects
128 D. Scolnic and M. Vincenzi

between SNe in the second and third rung could potentially bias . H0 measurements
[37]. SN dust extinction and color-dependent corrections are encapsulated in the
nuisance parameter .β (see Eq. 7.1). As a crosscheck, the .β parameter has been fitted
separately in SN used in the second and third rung, and it has been found to be
consistent [11]. This test, together with various NIR SN . H0 measurements [24–
27], see Table 7.1, suggests that dust or color-dependent effects are not expected to
significantly bias . H0 , or to be a significantly underestimated systematic in current
. H0 measurements.
In general, it is relatively simple to modify the SN sample selection in the second
and third rung to ensure better consistency in terms of SN physics/dust and sample
selection between rungs, and this approach has been implemented in the latest S. H0 ES
analysis.
Other systematics: Finally, there have been various analyses (e.g., [18, 22]) that
implemented additional systematic tests, e.g., changing the light-curve fitter or adding
spectroscopic information, but overall these appear to give very consistent (within
.∼ 0.3 km/s/Mpc) results. Similarly, isolating only to a single survey, like done in
[27], gives consistent results, though with much larger uncertainties due to the small
number of calibrator SNe.

7.4 Variants on the Path to . H0 with Supernovae Ia

Some of the main crosschecks on the SNe Ia used for these analyses is varying the
wavelength regime in which light curves are measured (i.e., optical to NIR) or the
dataset used (i.e., the survey used to measure light curves). As . H0 constraints are
limited by the number of SNe found within 40 Mpc, there is considerable overlap in
the data between these various studies. The most popular path to check/improve the
distance ladder with SN Ia is with measuring brightnesses from NIR light curves. We
list these papers in Table 7.1. Overall, even though the rest-frame band in which light
curves are measured varies between these analyses, and the fitting method varies
between these methods, there is generally very good agreement in recovered values
of . H0 . One challenge multiple of these studies have found (e.g. [24, 26]) is larger
calibration offsets between samples than those found for optical studies. A benefit of
this type of study is the possibility of improved precision of distance measurements
from NIR data, but the quality of older light curves has not typically been good
enough to evaluate this possibility [38].
An additional path is creating a “4 rung distance ladder”, as done in [23] and shown
in Fig. 7.4. SNe Ia used in the S. H0 ES distance ladder are those found in late-type
galaxies. To avoid this specific subsample, one can add another rung in the distance
ladder between TRGB/Cepheids and SNe - that from Surface Brightness Fluctuations
(SBF). The analysis of [23] improve on that of [39], which follow a similar method,
but uses an inhomogeneous set of SBF measurements, which significantly increases
the scatter of the tie between SBF and supernova measurements, and appears to
7 The Role of Type Ia Supernovae … 129

Fig. 7.4 A summary of papers and . H0 results given different uses and non-uses of SNe Ia. All . H0
measurements are in km/s/Mpc

bias . H0 to lower values. [23] find a value of . H0 = 74.6 ± 2.8 km/s/Mpc, in good
agreement with the S. H0 ES value.

7.5 Removing Type Ia Supernovae on the Path to . H0

An alternative path is to remove SNe Ia altogether. First, one can attempt to measure
H0 with a S. H0 ES-like distance ladder, but without the third rung; this is called a “two-
.
rung” distance ladder (see Fig. 7.4). This relies on knowing the redshifts to the nearby
galaxies in which Cepheids are discovered. As these galaxies are mostly .z < 0.01,
uncertainties in redshifts on the order of 300 km/s can contribute more than 10%
of the uncertainty per galaxy. Given the low signal-to-noise, any systematic biases
from selection effects or peculiar velocity corrections may have a large effect on the
measured value of . H0 . [28] attempt this with well-calibrated HST measurements of
Cepheids from the S. H0 ES team, and measures . H0 = 73.1 ± 2.5 km/s/Mpc. A key
component of this type of analysis is accounting for the inherent volumetric bias that
there are more galaxies further away than nearby.
A separate path, instead of removing the third rung entirely, is to replace SN
Ia with another type of standard candle. [4] attempt to replace SN Ia with SN II;
while SN II are not typically considered standardizable candles, there has been good
progress to improve their precision as distance indicators. The scatter in the Hubble
flow is .0.27 mag, compared to the .∼ 0.17 mag found for SNe Ia. The precision is
enabled by leveraging a correlation between the luminosity and photospheric velocity
130 D. Scolnic and M. Vincenzi

and a color correction [40, 41]. The study finds weaker constraints on . H0 , but good
agreement with the S. H0 ES analysis: . H0 = 77.6+5.2
−4.8 km/s/Mpc.
Another is using Surface Brightness Fluctuations, but in the same role as SNe
Ia [5]. With those, they find . H0 = 73.3 ± 0.7 ± 2.4 km/s/Mpc. This combines first-
rung calibration from Cepheids and TRGB, though they yield similar results (. H0 =
73.44 for Cepheids and . H0 = 73.20 km/s/Mpc for TRGB, each with .∼ 5% total
uncertainty). The number of calibrators is just 7, compared to the 42 that S. H0 ES uses,
thereby limiting the overall precision. Additionally, the Tully-Fisher (TF) relation
can be done in this third-rung role. References [6, 29] calibrate the TF relation with
Cepheids and TRGB; after doing so, [6] recovers .75.1 ± 2.3 ± 1.5 km/s/Mpc and
[29] finds .76.0 ± 1.1 ± 2.3 km/s/Mpc.

7.6 Inverse Distance Ladder to . H0 with Supernova

The same set of SNe Ia that are calibrated to a physical distance scale in the S. H0 ES
distance ladder, can also be used as un-calibrated relative distance indicators to
constrain . H0 in combination with other probes. Here we describe the process of
combining the un-calibrated SNe Ia with constraints from Baryon Acoustic Oscilla-
tion (BAO). BAO constraints on the expansion history . H (z) and that extrapolate . H0
have been shown to achieve low values of . H0 [42, 43, 45, 46]. We note that such
constraints (1) assume the sound horizon from CMB constraints and (2) are model
dependent because they assume .ɅCDM to infer . H (z = 0) from BAO datasets at
. z ∼ 0.5 and do not allow for late-time physical solutions. However, the SNe Ia can
be used to solve the latter of these two.
Instead of calibrating the SN Ia intrinsic magnitude to the distance ladder, they
can be calibrated at the typical BAO redshift to the distance scale set by BAO and for
which there is substantial overlap in .z of many SNe Ia in current datasets. Having
calibrated the SNe Ia to the BAO scale at moderate redshift, one can use the SNe Ia to
constrain the expansion history without assuming .ɅCDM to infer . H0 (see smoothed
blue curve of Fig. 7.5). In this case, one finds a very similar result (. H0 = 68.57 ± 0.9)
to that of the CMB under the assumption of .ɅCDM (grey curve of Fig. 7.5). We note
here that using SNe Ia calibrated to the BAO, which itself obtains the physical scale
from the sound speed in the early universe, to recover a low value of . H0 has spurred
many discussion to understand the impact of the value of the sound horizon.

7.7 Discussion

Improving measurements of . H0 in the future:

While constraints on .w should easily improve with upcoming large SN samples,
the road to improving constraints on . H0 is more challenging. There are a limited
number of SNe Ia that will explode in the near future within a .∼ 40 Mpc radius, a
7 The Role of Type Ia Supernovae … 131

Fig. 7.5 From [43], history for BOSS BAO, Pantheon SNe and Planck .rd assuming smooth expan-
sion and early-time physics only (blue), and for Planck assuming .ɅCDM (grey). BAO redshifts are
shown as short-dashed lines. Left panel: corresponding . H0 posteriors and Cepheid distance ladder
measurement (orange). Top panel: redshift distribution of Pantheon SNe

constraint due to HST discovery limits of Cepheids. At roughly one to three SN Ia per
year, it will likely take more than one decade to double the current sample of .42 SNe
calibrated by S. H0 ES Cepheid hosts. An exciting possibility is that the distance range
for precise measurements of Cepheids or TRGB can be extended. As the volume of
discovered SNe roughly increases with distance cubed, then even a .25% increase in
the range of Cepheids or TRGB measurements would allow a doubling in the number
of usable SN Ia in the second rung of the distance ladder. This may be a possibility
with new telescopes such as the James Webb Space Telescope, or further into the
future, the Nancy Grace Roman Space Telescope, or ground-based Extremely Large
Telescope (ELT).
The other path towards improving the constraint from SN Ia is to improve the
precision of the measurements. This is the path followed by papers like [18] which
tried using spectral features to improve the standardization, or the large number of
papers that measure SNe Ia in the NIR. [18] are able to improve the scatter of SN Ia
from 0.14 to 0.12 mag, which is similar to increasing the number of SN Ia by .∼ 35%.
The main challenge of this approach is that the benefit of the new idea depends on
how much data is available for its application in past literature measurements. New
types of measurements can be made for future nearby supernovae, but we can not
re-measure past SN Ia.
The role of SNe Ia when comparing local probes:
Within the larger discussion of tension between early and late universe probes, there
has been various discussions of smaller tensions between late-universe probes. One of
note has been potential discrepancies when using Cepheids or TRGB in the distance
132 D. Scolnic and M. Vincenzi

Table 7.2 Sources of differences in . H0 Between TRGB analysis by CATs [44] and CCHP [31] (in
units of . H0 )
Term .ΔCCHP

(km/s/Mpc)
SN related
1. Include SN 2021pit, 2021rhu, 2007on 0.6
2. No TRGB detected in N5584, N3021, 0.0
N1309, N3370
3. Peculiar flows (Pantheon.+) 0.4
4. Hubble flow surveys (Pantheon.+) 1.1
SN subtotal 2.0
TRGB related
5. Fiducial TRGB calibration/Tip-contrast .1.4
relation
Total 3.4
Comments: Adapted from [44], this shows how differences that appear to be from anchor mea-
surements of Cepheids versus TRGB are more due to differences in the supernova analysis. Here,
.ΔCCHP= differences between [31]. Descriptions of individual entries: (1) CCHP did not include
the two SNe from 2021 and excluded SN 2007on, (2) CATs did not detect the TRGB in these four
most-distant SN host galaxies; (3) Pantheon.+ accounts for peculiar motions which produce a highly
significant improvement in the dispersion of the Hubble diagram (see [19]) (4) CCHP measures the
Hubble flow from a single SN survey which has an offset with respect to the mean of many surveys
in Pantheon.+ as shown in [20]; (5) This term is difference in the calibration of TRGB from NGC
4258 as applied to SN hosts and is discussed by [44] (see Sect. 7.4)

ladder. While a lot of this focus has been on Cepheids versus TRGB, we note that
often the SNe that are used to project out the distance ladder can cause discrepancies.
An interesting example of that is shown from [44], which breaks down changes in . H0
from the CATS and CCHP teams [31]. As shown in Table 7.2, the CATS team explains
that SN choices account for more than 60% of the difference with other analyses,
whereas the TRGB side is less than 40%. A number of these choices have been
discussed in this review, including the application of peculiar velocity corrections
and consistency amongst surveys used for the different rungs of the distance ladder.

7.8 Conclusion

We discuss the various systematic uncertainties of as well as the crosschecks on the

SN Ia component of the distance ladder, which occupies two of the three rungs. We
find generally that systematics in the measurement of. H0 from the SNe are at the scale
of 0.3 km/s/Mpc. The range of . H0 values recovered from crosschecks is from .72 <
H0 < 78 km/s/Mpc. Therefore, it appears unlikely that SN Ia are artificially causing a
high value of . H0 to be found with the local distance ladder. Still, additional studies of
SN Ia are important in order to compare other parts of the distance ladder, like TRGB
7 The Role of Type Ia Supernovae … 133

versus Cepheids, as the sample overlap does not always yield high significance from
a direct comparison.

Acknowledgements D. S. thanks the John Templeton Foundation for their support of grant #62314.
D. S. is supported by DOE grant DE-SC0010007, the David and Lucile Packard Foundation. M. V. is
supported by NASA through the NASA Hubble Fellowship grant HST-HF2-51546.001-A awarded
by the Space Telescope Science Institute, which is operated by the Association of Universities for
Research in Astronomy, Incorporated, under NASA contract NAS5-26555.

References

1. E. Hubble, Proc. Natl. Acad. Sci. 15, 168 (1929). https://doi.org/10.1073/pnas.15.3.168

2. H.S. Leavitt, E.C. Pickering, Harvard College Observatory Circular, vol. 173 (1912), pp. 1–3,
https://ui.adsabs.harvard.edu/abs/1912HarCi.173....1L
3. A.G. Riess, W. Yuan, L.M. Macri et al., ApJL 934, L7 (2022). https://doi.org/10.3847/2041-
8213/ac5c5b
4. T. de Jaeger, L. Galbany, A.G. Riess et al., MNRAS 514(3), 4620–4628 (2022). https://doi.
org/10.1093/mnras/stac1661
5. J.P. Blakeslee, J.B. Jensen, C.-P. Ma et al., ApJ 911(1), 65 (2021). https://doi.org/10.3847/
1538-4357/abe86a
6. J. Schombert, S. McGaugh, F. Lelli, AJ 160, 71 (2020). https://doi.org/10.3847/1538-3881/
ab9d88
7. M. Sullivan, J. Guy, A. Conley et al., ApJ 737(2), 102 (2011). https://doi.org/10.1088/0004-
637X/737/2/102
8. M. Betoule, R. Kessler, J. Guy et al., A&A 568, A22 (2014). https://doi.org/10.1051/0004-
6361/201423413
9. D.M. Scolnic, D.O. Jones, A. Rest et al., ApJ 859, 101 (2018). https://doi.org/10.3847/1538-
4357/aab9bb
10. D.O. Jones, D.M. Scolnic, R.J. Foley et al., ApJ 881, 19 (2019). https://doi.org/10.3847/1538-
4357/ab2bec
11. D. Brout, D. Scolnic, B. Popovic et al., ApJ 938, 110 (2022). https://doi.org/10.3847/1538-
4357/ac8e04
12. D. Scolnic, D. Brout, A. Carr et al., ApJ 938, 113 (2022). https://doi.org/10.3847/1538-4357/
ac8b7a
13. A. Conley, J. Guy, M. Sullivan et al., ApJS 192, 1 (2011). https://doi.org/10.1088/0067-0049/
192/1/1
14. S. Dhawan, D. Brout, D. Scolnic et al., ApJ 894, 54 (2020). https://doi.org/10.3847/1538-
4357/ab7fb0
15. R.J. Foley, D. Scolnic, A. Rest et al., MNRAS 475, 193 (2018). https://doi.org/10.1093/mnras/
stx3136
16. S. Dhawan, A. Goobar, J. Johansson et al., ApJ 934(2), 185 (2022). https://doi.org/10.3847/
1538-4357/ac7ceb
17. D. Jones, S. Casertano, A.V. Filippenko et al., HST Proposal Cycle 28 ID #16269 (2020).
https://ui.adsabs.harvard.edu/abs/2020hst..prop16269J
18. Y.S. Murakami, A.G. Riess, B.E. Stahl et al. (2023). arXiv e-prints: arXiv:2306.00070 [astro-
ph.CO]
19. E.R. Peterson, W.D. Kenworthy, D. Scolnic et al., ApJ 938(2), 112 (2022). https://doi.org/10.
3847/1538-4357/ac4698
20. S. Brownsberger, D. Brout, D. Scolnic et al. (2021). arXiv e-prints arXiv:2110.03486, [astro-
ph.CO]
134 D. Scolnic and M. Vincenzi

21. D.O. Jones, A.G. Riess, D.M. Scolnic et al., ApJ 867(2), 108 (2018). https://doi.org/10.3847/
1538-4357/aae2b9
22. C.R. Burns, E. Parent, M. Phillips et al., ApJ 869(1), 56 (2018). https://doi.org/10.3847/1538-
4357/aae51c
23. P. Garnavich, C.M. Wood, P. Milne et al. ApJ 953(1), 35 (2023). https://doi.org/10.3847/1538-
4357/ace04b
24. S. Dhawan, S.W. Jha, B. Leibundgut, A&A 609, A72 (2018). https://doi.org/10.1051/0004-
6361/201731501
25. L. Galbany, T. de Jaeger, A.G. Riess et al. (2022). arXiv e-prints: arXiv:2209.02546, [astro-
ph.CO]
26. D.O. Jones, K.S. Mandel, R.P. Kirshner et al., ApJ 933, 172 (2022). https://doi.org/10.3847/
1538-4357/ac755b
27. S. Dhawan, S. Thorp, K.S. Mandel et al., MNRAS 524(1), 235–244 (2023). https://doi.org/10.
1093/mnras/stad1590
28. W.D. Kenworthy, A.G. Riess, D. Scolnic et al., ApJ 935(2), 83 (2022). https://doi.org/10.3847/
1538-4357/ac80bd
29. E. Kourkchi, R.B. Tully, G.S. Anand et al., ApJ 896(1), 3 (2020). https://doi.org/10.3847/1538-
4357/ab901c
30. D. Brout, G. Taylor, D. Scolnic et al., ApJ 938, 111 (2022). https://doi.org/10.3847/1538-4357/
ac8bcc
31. W.L. Freedman, B.F. Madore, D. Hatt et al., ApJ 882, 34 (2019). https://doi.org/10.3847/1538-
4357/ab2f73
32. A. Carr, T.M. Davis, D. Scolnic et al., PASA 39, e046 (2022). https://doi.org/10.1017/pasa.
2022.41
33. C.L. Steinhardt, A. Sneppen, B. Sen, VizieR Online Data Catalog. J. ApJ. 902 (14) (2022).
DOIurlhttps://doi.org/10.26093/cds/vizier.19020014
34. B. Popovic, D. Brout, R. Kessler, D. Scolnic (2021) arXiv e-prints: arxiv:2112.04456
35. M. Rigault, G. Aldering, M. Kowalski et al., ApJ 802, 20 (2015). https://doi.org/10.1088/0004-
637X/802/1/20
36. A.G. Riess, L. Macri, S. Casertano et al., ApJ 730, 119 (2011). https://doi.org/10.1088/0004-
637X/730/2/119
37. R. Wojtak, J. Hjorth, MNRAS 515(2), 2790–2799 (2022). https://doi.org/10.1093/mnras/
stac1878
38. E.R. Peterson, D.O. Jones, D. Scolnic et al., MNRAS 522, 2478 (2023). https://doi.org/10.
1093/mnras/stad1077
39. N. Khetan, L. Izzo et al., A&A 647, A72 (2021). https://doi.org/10.1051/0004-6361/
202039196
40. M. Hamuy, P.A. Pinto, ApJL 566, L63 (2002). https://doi.org/10.1086/339676
41. T. de Jaeger, B.E. Stahl, W. Zheng et al., MNRAS 496, 3402 (2020). https://doi.org/10.1093/
mnras/staa1801
42. E. Macaulay, R.C. Nichol, D. Bacon et al., MNRAS 486, 2184–2196 (2019). https://doi.org/
10.1093/mnras/stz978
43. S.M. Feeney, H.V. Peiris, A.R. Williamson et al., Phys. Rev. Lett. 122(6), 061105 (2019).
https://doi.org/10.1103/PhysRevLett.122.061105
44. D. Scolnic, A.G. Riess, J. Wu et al., ApJL 954, L31 (2023). https://doi.org/10.3847/2041-
8213/ace978
45. D. Camarena, V. Marra, MNRAS 495, 2630 (2020). https://doi.org/10.1093/mnras/staa770
46. S. Taubenberger, S.H. Suyu, E. Komatsu et al., A&A 628, L7 (2019). https://doi.org/10.1051/
0004-6361/201935980
Chapter 8
Tip of the Red Giant Branch

Siyang Li and Rachael L. Beaton

8.1 Introduction

8.1.1 Physical Basis of the TRGB

The tip of the red giant branch (TRGB) marks the evolutionary transition of stars
from the red giant branch (RGB) to the horizontal branch and can be used as a stan-
dardizable candle to construct an extragalactic distance ladder to measure . H0 . After
a low mass red giant star (.∼ .≤ 2 . Mʘ ) finishes burning through the hydrogen in its
core, it leaves a degenerate Helium core. The star then continues to burn hydrogen
in a shell surrounding the Helium core, which increases the core temperature and
stellar luminosity. Once the core temperature reaches a critical temperature, helium
burning begins. This results in runaway reaction called the Helium Flash [1]. Here,
core degeneracy is lifted, which prevents the star from further increasing in lumi-
nosity. After the Helium Flash, the star decreases in luminosity and transitions onto
the horizontal branch. The maximum luminosity reached during this evolutionary
sequence is a function of initial Helium core mass, which varies very little (.∼ 0.001
. Mʘ ) for stars from roughly 1.5 to 3 Gyr to .∼13 Gyr [2]. This small range of core
masses in turn results in a small range of maximum luminosities that allows the
TRGB to be used as a standardizable candle.

S. Li (B)
Department of Physics and Astronomy, Johns Hopkins University, Baltimore,
MD 21218, USA
e-mail: sli185@jh.edu
R. L. Beaton
Space Telescope Science Institute, Baltimore, MD 21218, USA
e-mail: rbeaton@stsci.edu

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 135
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_8
136 S. Li and R. L. Beaton

8.1.2 Observational Basis of the TRGB

The increase, then decrease, of a star’s luminosity before and after the Helium flash
creates a distinct ‘tip’ of the red giant branch that can be visualized in a color mag-
nitude diagram (CMD). To show this, we plot in Fig. 8.1 a CMD and luminosity
function of stars simulated with the Python Artpop package [3] and the noise model
from Li et al. [4] in the Hubble Space Telescope (HST) . F814W and . F606W filters.
We mark the location of the TRGB with a dashed red line, which corresponds to a dis-
continuity in the corresponding . F814W luminosity function and separates the RGB
population below the TRGB from the asymptotic giant branch (AGB) population
above the TRGB.
The TRGB is measured statistically with a population of stars, as the TRGB
feature can only be reliably identified and measured when using stars at multiple
locations along their evolution leading up to the TRGB. A variety of approaches
have been developed to measure the TRGB; we refer the reader to Beaton et al.
[2] for a detailed historical review. Generally, these approaches first apply selection
cuts to remove possible contaminants and isolate the red giant branch. These can
involve color cuts, which are often based on color calibrations such as from Rizzi
et al. [5] and Jang et al. [6] to isolate metal-poor stars, and spatial cuts involving
young star clipping [7], B mag 25th isophotal radius [8] or other elliptical cuts [9].
We describe algorithms used to measure the TRGB after such cuts in the context of
. H0 in Sects. 8.3.1 and 8.3.1.3.

Observationally, the TRGB is typically measured in the relatively metal-poor and

old halos of galaxies. This is to primarily reduce the effects of contamination from
younger stars [9] and to avoid internal extinction and crowding. In addition, the
TRGB exhibits a metallicity dependence, which can be traced with color due to the

Fig. 8.1 CMD and LF simulated using Artpop [3] in the . H ST . F814W and . F606W filters and 10
Gyr stars with [Fe/H] = .−1 dex. We add noise using the noise model from Li et al. [4]. The location
of the TRGB discontinuity is marked with the red dashed line. The RGB and AGB lie below and
above the TRGB, respectively
8 Tip of the Red Giant Branch 137

line-blanketing effect. The TRGB becomes fainter and brighter at redder colors in
the optical and infrared, respectively. Empirically, the TRGB shows the least color
dependency in the . I or HST . F814W filters [5, 6]; for this reason, these filters are
most commonly used to calibrate and measure . H0 . There is, however, an increasing
interest in measuring the TRGB in redder colors to extend TRGB measurements in
the Hubble flow and take advantage of the IR capabilities of the James Webb Space
Telescope (. J W ST ). We discuss this further in Sect. 8.4.1.

8.2 The TRGB Path to . H0

In this section, we briefly describe how the TRGB can be used to construct a distance
ladder to measure . H0 . We emphasize that this description is not comprehensive and
encourage the reader to delve into the literature to learn more about the various
methodologies and considerations needed to measure a robust TRGB-based . H0 .

8.2.1 Absolute Calibrations

Measuring extragalactic distances with the TRGB to measure. H0 requires knowledge

of its apparent magnitude, absolute magnitude, and extinction:

μ0 = m host − M − A
. (8.1)

where .μ0 is the distance modulus to the host galaxy, .m host is the apparent magnitude
of the TRGB in the host galaxy, . M is the absolute magnitude of the TRGB, and . A is
the extinction (where we combine foreground and internal extinctions). .m host can be
directly measured using the methods described later in this chapter. The extinction,
. A, can be obtained from dust maps, for instance from Schlafly & Finkbeiner 2011
[10], among others, the discussion of which is outside the scope of this chapter.
With measurements of .m host and . A, we are left with two unknowns, .μ0 and . M.
To obtain . M, one can invert Eq. 8.1 such that

. M = m − μ0 − A (8.2)

.M is intrinsic to the TRGB and varies very little galaxy to galaxy, with variations that
can be standardized via measurements of the color and contrast ratio, for instance.
This means Eq. 8.2 is general and we can, in principle, exchange .m host with a general
.m and use any galaxy to obtain . M provided we also have a direct measurement of its

distance, .μ0 , that is obtained independently of the TRGB. Examples of galaxies with
distances measured geometrically and independently from the TRGB that are most
commonly used to calibrate, or ‘anchor’, the TRGB are the Milky Way, Large Mag-
ellanic Cloud (LMC), Small Magellanic Cloud (SMC), and NGC 4258. Distances to
138 S. Li and R. L. Beaton

Milky Way red giants can be determined via parallaxes, the most precise of which
have been measured by the .Gaia mission [11]. The LMC, SMC, and NGC 4258
have geometric distances measured via eclipsing binaries [12, 13] and water masers
[14], respectively. For derivations of the LMC, SMC, and NGC 4258 distances using
these systems, we refer the reader to their respective references.
With measurements of the geometric distance to and apparent TRGB magnitude
in the anchor galaxy (as well as an estimate of the extinction, . A), one can solve for
the absolute magnitude, . M, of the TRGB using Eq. 8.2. With . M as well as .m host and
. A for the host galaxy, one can then return to Eq. 8.1 and solve for .μ0 .
We briefly pause to highlight differences between two distinct approaches of
calibrating the TRGB and how they affect . H0 measurements. The first approach
is to minimize uncertainties in the TRGB by applying particular selections. For
instance, Anderson et al. [15] utilize the variability of Optical Gravitational Lensing
Experiment (OGLE) Small Amplitude Red Giants (OSARGS) [16–19] in the LMC to
achieve a 1.39.% calibration of the TRGB. Another instance can be seen in Hoyt [20],
who selects 5 out of 20 LMC fields that have more symmetric Sobel responses and
narrower bootstrap widths to achieve a 1.5.% calibration. While both these methods
are effective at lowering uncertainties in the TRGB calibration to the 1.% level, it
is important to note that these calibrations are distinct from calibrations that can be
used (at the moment) to measure . H0 , as these selections are not currently feasible in
SN Ia host galaxies with the current resolution of space telescopes such as . H ST and
. J W ST . For instance, the amplitudes of red giants are the same order magnitude of
the photometric uncertainties in many SN Ia host galaxies, which makes it difficult
to differentiate intrinsic and extrinsic variability. Li et al. 2023 demonstrate in their
Appendix D that the same bootstrap width criterion of 0.025 mag used to reject 15
out of 20 LMC fields used in Hoyt [20] would eliminate all TRGBs measured in SN
Ia host galaxies, precluding its use to measure . H0 [4]. The key concept here is that
similar selections should be applied to both the anchor and host galaxies. Inconsistent
selections can introduce biases, such as mismatch of the calibration due to different
stellar populations.
The second approach applies the same or similar measurement and selection cri-
terion to both the anchor and host galaxies, even if this results in a higher uncertainty
in the calibration. Here, the calibration and selection of the TRGB remain consistent
across the distance ladder such that the calibration more closely reflects the mea-
sured luminosity of the TRGB in the host galaxy such as in Scolnic et al. [21]. We
emphasize that both these approaches are very useful for advancing understanding
of the TRGB. One should be careful, however, to be aware of these two conceptually
different approaches when choosing a calibration to anchor distance and . H0 mea-
surements. We look forward to future instrumentation that can provide the resolution
that erases this distinction.
8 Tip of the Red Giant Branch 139

8.2.2 Calibrating SN Ia

Now with a calibration of the TRGB, one can measure distances to galaxies using
Eq. 8.1. To measure . H0 , one in particular would want to measure TRGB distances
to galaxies that contain both a well-observed TRGB as well as a luminous standard
candle that can reach further into the Hubble flow, such as SN Ia, to obtain a calibration
of the more luminous standard candle. The Tully-Fisher relation is another way one
can reach far into the Hubble flow; we refer the reader to the corresponding section in
this book for more information. For simplicity, we will refer to SN Ia when describing
the distance ladder for the remainder of this chapter.
A TRGB distance measurement to a galaxy, combined with a measurement of
the apparent magnitude of SN Ia (such as from Pantheon.+ [22] or the Carnegie
Supernovae Project [23, 24]), allows one to calibrate the SN Ia luminosity using the
same principles described in Sect. 8.2.1. A luminosity calibration of SN Ia can then
be combined with apparent magnitude SN Ia measurements to measure distances to
galaxies further in the Hubble flow where the TRGB is too faint to be observed.
One illustrative example of how to calibrate SN Ia is described in Scolnic et al.
[21]. They first find the difference in magnitude between the apparent TRGB and
mean SN Ia apparent magnitude in a given galaxy:

. ΔS = m I,T RG B − m 0B (8.3)

where .ΔS is the difference between the apparent TRGB and SN Ia magnitudes,
m I,T RG B is the apparent TRGB magnitude, and .m 0b is the apparent SN Ia magnitude.
.
They then obtain the difference between the absolute magnitudes of the TRGB and
SN Ia by taking the mean .ΔS from all SN Ia host galaxies and subtracting it from
the TRGB luminosity:

. M B0 = M I,T RG B − Δ S̄ (8.4)

to obtain the luminosity of SN Ia, . M B0 . We note that Eq. 8.4 is essentially an algebraic
rearrangement of Eq. 8.1 for two standard candles to the same galaxy. .μ0 = m − M
with TRGB should yield the same distance when using .μ0 = m − M with SN Ia
when both are in the same galaxy. By plugging in .ΔS from Eqs. 8.3 into 8.4, one can
rearrange to find .(m − M)T RG B = (m − M) S N I a .

8.2.3 Measuring . H0

Measuring . H0 requires knowledge of both redshifts and distances to galaxies. The

derivations of the equations used to measure . H0 are outside the scope of this chapter;
we instead briefly mention that one can conveniently obtain . H0 using the Hubble
140 S. Li and R. L. Beaton

diagram intercepts from previous studies. With a calibration of SN Ia obtained with

the method from the previous section, one can use the equation:

.log H0 = 0.2M B0 + a B + 5 (8.5)

where . M B0 is the SN Ia luminosity, and .a B is the intercept of the Hubble diagram

at low redshift and can be found from various studies, such as Riess et al. [25].
Equation 8.5 can also be used to measure . H0 without .a B by comparing distances
relative to another . H0 measurement.

8.2.4 Measurements of . H0 with the TRGB

In this section, we highlight a few . H0 measurements from the past couple of years.
We note that this discussion is not exhaustive and encourage the reader to read the
references within these studies for more examples of TRGB-based. H0 measurements.
1. The most recent . H0 measurement from the Carnegie-Chicago Hubble Program
(CCHP) results in . H0 = 69.8 ± 0.6 (stat) .± 1.6 (sys) km s.−1 Mpc.−1 [26]. This
work updates the Freedman et al. [27] measurement, which used a Monte Carlo
Markov Chain approach and 18 SN Ia calibrators to measure . H0 . In this update,
Freedman 2021 [26] replaces the Freedman et al. [27] LMC only calibration
with a combination of calibrations from NGC 4258 [9], Milky Way globular
clusters [28], LMC [20], and SMC [20]. They use a Sobel-filter based TRGB
measurement methodology which will be described in Sect. 8.3.1.
2. Anand et al. [8] independently reduced the data used by CCHP to find . H0 =
71.5 ± 1.8 km s.−1 Mpc.−1 . They incorporate . H ST NGC 4258 observations that
were not used in Freedman 2021 to anchor their distance ladder and notably found
that they could not reliable identify the TRGB in 4 out of the 15 hosts used by
Freedman et al. 2019. Anand et al. [8] also used a model based least-squares fit
instead of a Sobel method, a different photometry pipeline, and apply color cor-
rections from Rizzi et al. [5]. They find a .−0.028 mag difference (Extragalactic
Distance Database - CCHP) distance scales and did not find a significant differ-
ence in their . H0 measurement when using SN Ia from the Carnegie Supernovae
Project [23, 24] or Pantheon [29].
3. Scolnic et al. [22] recently measured . H0 = 73.22 ± 2.06 km s.−1 Mpc.−1 . In
their measurement, Scolnic et al. 2023 [22] correct their measured TRGBs to
a fiducial contrast ratio to improve standardization across rungs in the distance
ladder. They calibrate their TRGB to NGC 4258 [4] and include three new SN Ia
in their analysis that were not previously used in Freedman 2021 or Anand
et al. 2022 [8, 26], in addition to applying peculiar flow corrections and the
Pantheon.+ sample. They find that applying contrast ratio corrections to their
TRGB measurements increases the Freedman [26] and Anand et al. [8] . H0 by
8 Tip of the Red Giant Branch 141

1.4 and.−0.3 km s.−1 Mpc.−1 , respectively. The largest differences, however, came
from different treatments of SN Ia, which would change the Freeman [26] and
Anand et al. [8] . H0 by 2 and 1.3 km s.−1 Mpc.−1 , respectively. Their algorithm
and data are publicly available via the link provided in Sect. 8.3.2.
4. Dhawan et al. [30] used a hierarchical Bayesian SED model, BayeSN, to infer
SN Ia distances calibrated with Cepheids and TRGB to find . H0 = 70.92 .± 1.14
(stat) .± 1.49 (sys) km s.−1 Mpc.−1 with the TRGB calibration. Their hierarchical
Bayesian SED model allowed them to simultaneously model optical and NIR SN
Ia light curves, decreasing . H0 uncertainty by .∼ 15% compared to their optical
only case.
5. Dhawan et al. [31] measured . H0 = 76.94 ± 6.4 km s.−1 Mpc.−1 using only SN Ia
from the from the Zwicky Transient Facility [32] to minimize systematics from
host-galaxy bias and different treatments of photometry. They use a single host
galaxy NGC 7814 containing SN Ia ZTF SN Ia SN 2021rhu and measure the
TRGB with the same pipeline from CCHP. The large uncertainty is primarily
due to their single SN Ia calibrator; they note that future Zwicky Transient
Facility observations of SN Ia combined with . J W ST observations of TRGB
will significantly improve this measurement.
6. Blakeslee et al. [33] tied Cepheid and TRGB calibrations to surface brightness
fluctuation measurements of 64 galaxies ranging from 19 to 99 Mpc to find
−1
. H0 = 73.3 ± 0.7 ± 2.4 km s. Mpc.−1 . They obtain their sample from a variety
of programs using the. H ST WFC3/IR. F110W filter and four different treatments
of galaxy velocities.

8.3 Challenges for . H0 Measurements

8.3.1 Different Measurement Algorithms

The TRGB can be measured using a variety of different methodologies, each of

which can introduce challenges for ensuring consistency along the distance ladder
to measure . H0 . Here, we focus on describing the Sobel filter and least-squares fit
methods, the two of which are most commonly used to measure TRGB . H0 .

8.3.1.1 Sobel Filter

The Sobel filter was first used to measure the TRGB by Lee et al. [34] and has since
been adopted in several different forms. For a detailed review of these variations,
we refer the reader to Beaton et al. [2]. Here, we focus on describing the primary
method adopted by the Carnegie-Chicago Hubble Program (CCHP) [35], which has
been used extensively to measure . H0 . An unsupervised variant of this method has
142 S. Li and R. L. Beaton

been developed by the Comparative Analysis of TRGBs (CATs) team [4, 7, 21]; we
describe this method separately in Sect. 8.3.1.3.
The Sobel filter concept originates from the field of computer vision and is used to
identify edges by evaluating the first derivative of intensities. In computer vision, this
may involve finding the edges of an object in a 2D image to isolate the object from its
surroundings. For a 1D luminosity function around the TRGB, the Sobel filter can be
used to locate the TRGB discontinuity in the luminosity function with the maximum
first derivative. To use this method, one first finely bins the luminosity function around
the TRGB in bin widths of order 0.005 mag. Then, the binned luminosity function
is smoothed using Gaussian-windowed, Locally Weighted Scatterplot Smoothing
(GLOESS) [35–37] to suppress false detections caused by noise. Poisson weights
are also applied to de-weight sparsely populated bins. Finally, a Sobel kernel of
the form [-1, 0, 1] is evaluated across the luminosity function to compute the first
derivatives of the smoothed and weighted luminosity function. The TRGB is taken
to be at the location of the maximum Sobel response.
This method has its advantages and disadvantages. GLOESS is effective at reduc-
ing false TRGB detections caused by noise and can help suppress smaller local
maxima in the Sobel response to isolate the strongest discontinuities in a luminos-
ity function. In addition, the method is computationally inexpensive compared to
maximum likelihood estimation, for instance, and can be used to detect multiple
TRGBs in a sample containing mixed stellar populations. However, the smoothing
filter, by nature, reduces the TRGB discontinuity by moving red giant branch (RGB)
stars fainter than the TRGB discontinuity into the region brighter than the TRGB
discontinuity containing asymptotic giant branch (AGB) stars, as RGB stars are
more abundant than AGB stars. This can cause a smoothing bias at certain levels of
smoothing, where changing the strength of the smoothing changes the location of the
measured TRGB. The amount of smoothing bias can also depend on the heights of
the TRGB discontinuities. This effect was noted by Hatt et al. [35] and investigated
and documented by Anderson et al. [15] and Li et al. [4]. Hatt et al. [35] describes an
artificial star luminosity function test, which is used to choose a smoothing level by
minimizing the quadrature sum of statistical and systematic errors. As the smoothing
bias can be a function of the contrast ratio [4, 15], it is essential that the artificial star
luminosity function accurately reflects the observed sample for which is being used
to estimate uncertainties. One should keep in mind that using different smoothing
levels for different TRGB measurements in different galaxies could potentially lead
to inconsistencies along the distance ladder. In addition, there can often be several
local maximum in the Sobel response that are close together in height and magni-
tude even after applying GLOESS and Poisson weighting. Some studies suggest to
increase the smoothing interval in these cases until there is a single, well defined
peak (for instance, in Hatt et al. [35]). However, doing so runs the risk of combining
Sobel peaks corresponding to multiple true TRGBs from different stellar popula-
tions or merging a local maximum corresponding to the true TRGB with another
local maximum caused by noise.
8 Tip of the Red Giant Branch 143

8.3.1.2 Least Squares Fitting

The least-squares fitting method is frequently used by the Extragalactic Distance

Database team and has been used to measure . H0 [8]. This method assumes that
the RGB/AGB luminosity function around the TRGB is well described by a broken
power law model of the form:
{
10a(m−m T RG B )+b m ≥ m T RG B
.ψ(m) = (8.6)
c(m−m T RG B )
10 m < m T RG B .

where .m T RG B is the magnitude of the TRGB, .m is the magnitude of a star, .a and .c

are the logarithmic slopes of the luminosity function corresponding to the RGB and
AGB, respectively, and .b represents the strength or height of the TRGB break. This
form of the RGB/AGB luminosity function was first described by Zoccali and Piotto
[38] and subsequently used for TRGB measurements by Mendez et al. 2002, Makarov
et al. 2006, Wu et al. 2014, Li et al. 2022, and Li et al. 2023, and references therein
[39–43]. We note that Eq. 8.6 is also used for the maximum likelihood approach to
measuring the TRGB, which is also used by the EDD team [27] and has been used
to measure the TRGB using Milky Way field giants [29, 30]. Although both the
least-squares and maximum likelihood methods are model-based, we focus here on
the the least-squares fitting method here due to its use to measure . H0 .
The TRGB least-squares fitting procedure as described in Wu et al. [41] is as
follows: The luminosity function model described in Eq. 8.6 is first convolved with
the completeness, uncertainty, and bias of the observations. Then, an initial guess
for the location of the TRGB is obtained by taking first derivatives of the luminosity
function and locating their maximum. The luminosity function is binned in 0.05 mag
intervals across .± 1 mag from the anticipated TRGB magnitude. Finally, a least-
squares fit using the Levenberg-Marquardt algorithm is applied to the luminosity
function to determine .m T RG B , .a, .b, and .c. Uncertainties of the fitted parameters are
estimated with the square root of the variance.
This method has the advantage of also being computationally inexpensive and
is easily implemented in Python, with least-squares fitting routines such as from
scipy curve_fit [44] readily available. However, this method assumes a sin-
gle, true TRGB in its model and cannot separate out two TRGB breaks that may arise
from the mixture of stellar populations. In principle, one could modify the model to
include multiple TRGB discontinuities, however, this would still lack the flexibility
of the Sobel filter at detecting multiple Sobel peaks. In addition, one must select
an appropriate binning width; for well-populated stellar populations, the measured
TRGB may be insensitive to binning, however, the measurements using sparsely pop-
ulated populations may be susceptible to noise and sensitive to the choice of binning.
The least-squares fit can also potentially converge onto unrealistic model parameters
depending on the initial guess. For instance, in some cases the fit may identify the
faint magnitude cutoff as the primary discontinuity if the TRGB discontinuity itself
144 S. Li and R. L. Beaton

is not well defined or identify the tip of the AGB as the tip of the RGB (for examples,
see Li et al. [42] and Anand et al. [45]).

FAQ: Which method should I use to measure the TRGB?

Response: Each TRGB measurement method has its advantages and disadvan-
tages and should be selected based on characteristic of the RGB/AGB sample
and end goal. For instance, a Sobel filter paired with GLOESS can be more
effective at reducing noise and false detections in noisy data compared to the
binned power law least-squares fit, but the latter can also provide information
about the RGB and AGB slopes, which may be useful for comparing stel-
lar populations in other fields or galaxies. In many cases when the sample
is well defined and populated, different methodologies agree well. However,
we emphasize that regardless of the method used, it is important to closely
examine potential sources of measurement biases to ensure that the methods
are consistently across a distance ladder such that any biases do not propagate
into the final distance or . H0 measurement.

8.3.1.3 Combining Algorithms and Stellar Populations

The TRGB measurement methods described in Sect. 8.3.1 have been used exten-
sively to measure extragalactic distances and . H0 . However, significant variations in
calibrated TRGB luminosities (see compilations in Blakeslee et al. 2021, Freedman
2021, and Li et al. 2022 [26, 33, 42]) and . H0 (for instance, see Freedman 2021 and
Anand et al. 2021 [8, 26]) suggest there may be additional considerations that need
to be taken into account to improve the consistency of TRGB measurements. In par-
ticular, one must ensure that the TRGB calibrated in an anchor galaxy is standarized
such that it accurately reflects the true luminosity of the TRGB in a host galaxy.
Mismatches between the TRGBs in anchor and host galaxies can occur with incon-
sistent measurement or selection choices in the two galaxies. The stellar population
used to calibrate the TRGB in a host galaxy should have similar characteristics or be
corrected to a fiducial value. One must also be careful that the measurement method
also does not introduce biases, or that these biases cancel out along the distance
ladder (Fig. 8.2).
Some of these issues can manifest when using Sobel filter and least-squares fit
TRGB measurement methods. Model based methods, such as the least-squares fit
described in Sect. 8.3.1.2, have the disadvantage of assuming only a single TRGB
break. While such model based methods are effective at measuring populations that
show a single, well defined TRGB discontinuity, a population consisting of stars of
several characteristics may exhibit multiple TRGB discontinuities [4, 7, 21]. Multi-
ple TRGB discontinuities cannot be separated with a broken power law model that
assumes a single TRGB discontinuity. On the other hand, it is difficult to differentiate
8 Tip of the Red Giant Branch 145

Fig. 8.2 Examples of a

Sobel filter response (top;
blue) and broken power law
fit (bottom; blue) TRGB
measurements on a simulated
luminosity function (black).
The red vertical line marks
the location of the simulated
TRGB

between local maxima in a Sobel response that correspond to noise and true TRGBs.
An approach adopted by Hatt et al. [35] is to increase the smoothing until a single
local maximum in the Sobel response remains. However, this can potentially com-
bine multiple true TRGBs or a true TRGB with nearby local maxima in the Sobel
response corresponding to noise, and different choices of smoothing in the calibra-
tion and host can lead to inconsistencies along the distance ladder. The selection of
different smoothing parameters across the distance ladder, or one local maximum
in the Sobel response over another of similar height, introduces subjective choices
in the measurement method that can easily lead to confirmation bias, if one is not
careful.
To address these issues, the CATs team developed an unsupervised Sobel-filter
based algorithm that can be used to measure TRGBs using objectively determined
measurement parameters and improve the consistency of TRGB measurements
across the distance ladder to measure . H0 [7]. The algorithm first applies a step
called spatial clipping, which removes regions in an observed field that contain high
ratios of blue to red stars, as defined on a CMD, that might contaminate the red giant
sample. Then, the algorithm applies a color band with a fixed width that maximizes
the number of stars in the sample and applies GLOESS and Poisson weights to a
luminosity function. A Sobel filter of the form [-1, 0, 1] is then evaluated across the
luminosity function to identify the TRGB. The key differences in the CATs algorithm
compared to previous Sobel filter algorithms are that the measurement parameters,
such as the width of the color bands, choice of smoothing, and quality cuts, are
first optimized to minimize the scatter in the observed field-to-field variation in the
measured TRGBs in multiple fields from the Galaxy Halos, Outer disks, Substruc-
ture, Thick disks, and Star clusters (GHOSTS) galaxies [46]. These measurement
parameters are subsequently fixed when measuring extragalactic distances and . H0 .
The algorithm also defines a contrast ratio, taken to be the ratio between number of
stars in the 0.5 mag interval fainter and brighter than the measured TRGB break. In
addition to being a useful for estimating ages [7, 47], the contrast ratio provides a
metric for differentiating high and low signal TRGB discontinuities. A high contrast
ratio corresponds to a well defined break, while a low contrast ratio corresponds to
146 S. Li and R. L. Beaton

an ill-defined break that may be sensitive to noise. The CATs algorithm filters out
contrast ratios less than 3 to avoid false detections [7]. This algorithm is publicly
available via the GitHub repository link provided in Sect. 8.3.2.
In addition to developing an unsupervised Sobel-based TRGB measurement algo-
rithm, the CATs team discovered a 5 .σ relationship between the measured TRGB
and contrast ratio (tip-contrast ratio, or TCR). This finding suggests that the TRGB
can be further standardized via the contrast ratio so that it is used more consistently
across the distance ladder. The CATs team used their unsupervised algorithm and
TCR to calibrate the TRGB luminosity using NGC 4258 [4] and measure . H0 [21]
and find a . H0 of 73.22 .± 2.06 km s.−1 Mpc.−1 . There are four key differences in the
CATs . H0 compared to previous Sobel-filter based . H0 measurements. First, the CATs
algorithm uses measurement parameters such as smoothing and spatial clipping con-
sistently across the distance ladder to facilitate first order cancellation of potential
biases. Second, the CATs algorithm allows for the possibility of multiple true TRGB
breaks in a luminosity function due to mixtures of stellar populations. Third, the
CATs team corrects measured TRGBs to a fiducial contrast ratio to facilitate closer
alignment between the anchor and host TRGB and to account for variations in the
measured TRGB as seen in the TCR. Fourth, the CATs measurement follows a dif-
ferent treatment of SN Ia, such as applying peculiar velocity corrections (see Scolnic
et al. [21] for more details).
The CATs team hypothesize that the TCR is caused by a combination of astro-
physical properties as well as the measurement process such as smoothing [4, 7, 21].
However, the exact nature of the TCR is inconsequential to the validity of the trend;
it is important to remember that the TCR is an empirical relationship, such that, at
the fundamental level, there is a variation in the measured TRGB that correlates with
another parameter. The TRGB is corrected to account for this variation irrespective
of the exact origin of the variation, whether due to astrophysical differences, mea-
surement artifacts, or both. There are still improvements that can be made to this
approach, and more studies can be conducted to better understand the nature of the
TCR and whether the TCR exists when using other TRGB measurement methods.
While there may be other avenues to further standardize the TRGB, these efforts
show the beginning of a paradigm shift in TRGB measurements with the recognition
that there are still small variations in measured TRGBs using the Sobel-filter method
that were not previously accounted for, and that additional standardization should be
applied to the measured TRGB as used to measure . H0 .

FAQ: How much does correcting for the contrast ratio change . H0 ?
Response: Scolnic et al. [21] compared the CATs . H0 result with previous
. H0 measurements in their Table 5 and found that applying contrast ratio stan-

dardization increased . H0 from Freedman et al. [26] by 1.4 km s.−1 Mpc.−1 and
decreased . H0 from Anand et al. [8] by 0.3 km s.−1 Mpc.−1 . It’s worth noting
that some of the largest differences between the different . H0 values also come
8 Tip of the Red Giant Branch 147

from different treatments of SNe Ia. We refer the reader to Table 5 and its
corresponding text in Scolnic et al. [21] for more details.

8.3.2 Open Datasets

Open datasets and algorithms help improve transparency and community-wide dis-
cussions in the field. We list here currently public TRGB datasets and algorithms
available through websites that we encourage the reader to explore in addition to
datasets made directly available via journal publications. We encourage future stud-
ies to make their datasets and algorithms available to the public.
• Photometry from the Extragalactic Distance Database [48, 49]:
https://edd.ifa.hawaii.edu/.
• Unsupervised Sobel-filter algorithm and results from the CATs team: [4, 7,
21]: https://github.com/JiaxiWu1018/Unsupervised-TRGB, https://github.com/
JiaxiWu1018/CATS-\inlinealttextupperH0$H_0$.
• 2D Maximum Likelihood algorithm used to calibrate TRGB using field stars and
. Gaia: [43]:
https://github.com/siyangliastro/Gaia-DR3-MilkyWay-TRGB.

8.3.3 TRGB at the Crossroads

We pause here to highlight two distinct ideologies in constructing a TRGB distance

ladder to measure . H0 that arise as a natural consequence of developments in obser-
vational abilities. The first we refer to as artisanal measurements, which involve
adjusting measurement and selection parameters to optimize TRGB measurements.
For instance, one may increase the smoothing parameter when using a Sobel filter
method until there is a single and well defined TRGB detection, as suggested by Hatt
et al. [35], with the optimal smoothing parameter being different for each galaxy.
Advocates for this approach may argue that using different measurement and selec-
tion choices is necessary when using heterogeneous datasets having different levels
of photometric uncertainties and sample size, for instance, which requires different
treatments to isolate the TRGB. We note that when doing so, it is especially impor-
tant to understand and minimize potential biases of the measurement method when
working with data of varying qualities.
On the other hand, applying different measurement and selection criteria runs the
risk of introducing inconsistencies along the distance ladder. For instance, Ander-
son et al. 2023 and Li et al. 2023 [4, 15] demonstrate that different smoothing,
as applied to the same luminosity function, introduces different biases. Advocates
148 S. Li and R. L. Beaton

for this approach argue that measurements must be applied consistently across the
distance ladder for all galaxies to ensure first order cancellation of any biases and
minimize inconsistencies between calibrator and host measurements. This approach
is more feasible when working with datasets with taken with the same telescope,
such as . H ST observations of NGC 4258 and GHOSTS galaxies.
These two ideologies can be seen in the context of a natural shift that occurs as
more homogeneous observations are made possible with improved telescopes such
as . H ST and . J W ST . We leave the question of which approach is best open to the
reader. Both have their advantages and disadvantages. While there is not a single,
accepted answer, it is important to be aware of this distinction, especially when
interpreting the results and methodologies of different . H0 results.

8.4 Future Developments

8.4.1 New Prospects with JWST

The successful launch of. J W ST opened up the possibly of probing the TRGB further
in the IR in greater resolutions and distances than. H ST . In addition, the configuration
of detector modules on the . J W ST Near Infrared Camera (NIRCam) opens up the
possibility for observational programs to simultaneously observe multiple standard
candles at once. NIRCam contains two cameras each consisting of four chips that cre-
ates a rectangular array. This setup allows . J W ST to simultaneously image both the
halos and disks of galaxies to observe red giants for TRGB, Cepheid variables, Mira
variables, and the Carbon asymptotic giant branch stars for the J-region Asymptotic
Giant Branch (JAGB), potentially enabling four . H0 measurements with one set of
anchor and host observations once the JAGB is better understood and standardized.
This setup also will allow for a powerful cross-calibration of these standard candles.
In addition, NIRCam is a two-channel instrument which allows observations to be
taken simultaneously in the short and long wavelength IR, further increasing the
breadth of observations possible within a single observing program. However, as
the TRGB exhibits a greater metallicity dependence in the IR compared to the . I or
. F814W bands [50], further calibration and investigation will be needed before the IR

TRGB can be used to measure robust extragalactic distances. We provide an example

of the metallicity dependence of the NIRCam . F150W TRGB in Fig. 8.3, where we
use the Python ArtPop package [3] to generate 10 Gyr isochrones for the red giant
branch with metallicities ranging from [Fe/H] of .−2 to .−1 in 0.1 increments starting
from the left most isochrone. . J W ST programs designed with these features in mind
have already been completed; we highlight some of these in Sect. 8.4.1.1.
8 Tip of the Red Giant Branch 149

Fig. 8.3 Isochrones

generated using the Python
ArtPop package [3] in
. J W ST NIRCam . F150W
and . F277W using stars of 10
Gyr and metallicities ranging
from [Fe/H] of .−2 to .−1 dex

8.4.1.1 Early . J W ST Observational Programs

As of writing this chapter, . J W ST has undergone two general observer (GO) pro-
posal cycles in addition to the Director’s Discretionary (DD) Early Release Science
Programs (ERS), all of which contain accepted proposals containing observations
that can be used to measure the TRGB. We briefly highlight some of these programs
that contain observations that can be used for TRGB measurements and encourage
the reader to explore images from these proposals on MAST.1 Proposal abstracts can
be obtained from STScI.2.,3 We anticipate this list to grow as . J W ST undergoes more
proposal cycles in the future.

• ERS DD-1334 (PI: Daniel Weisz): The Resolved Stellar Populations Early
Release Science Program
• Cycle 1 GO-1638 (PI: Kristen McQuinn): Securing the TRGB Distance Indica-
tor: A Pre-Requisite for a JWST Measurement of . H0
• Cycle 1 GO-1685 (PI: Adam Riess): Uncrowding the Cepheids for an Improved
Determination of the Hubble Constant
• Cycle 1 GO-1995 (PI: Wendy Freedman & Barry Madore): Answering the
Most Important Problem in Cosmology Today: Is the Tension in the Hubble Con-
stant Real?
• Cycle 2 GO-3055 (PI: Richard Brent Tully): A TRGB calibration of Surface
Brightness Fluctuations

1 https://mast.stsci.edu/portal/Mashup/Clients/Mast/Portal.html.
2 https://www.stsci.edu/jwst/science-execution/approved-ers-programs.
3 https://www.stsci.edu/jwst/science-execution/approved-programs.
150 S. Li and R. L. Beaton

8.4.2 Future Gaia Data Releases

The .Gaia mission [11] has observed over a billion stars in the Milky Way and signif-
icantly augmented the understanding of our home galaxy. Among the measurements
of Milky Way field stars made by .Gaia include low-resolution spectra that can be
used to construct synthetic photometry and parallaxes. These measurements can be
retrieved via the .Gaia archive4 and .Gaia Synthetic Photometry Catalogue [51] and
have been used to calibrate the TRGB. Dixon et al. [52] used .Gaia Early Data
Release (EDR) 3 parallaxes and SkyMapper Data Release (DR) 3, American Asso-
ciation of Variable Star Observers (AAVSO) Photometric All-Sky Survey (APASS)
DR9, Asteroid Terrestrial-impact Last Alert System (ATLAS) Refcat2, and .Gaia
DR3 synthetic photometry to calibrate the TRGB using Milky Way field stars. They
directly convert apparent magnitudes into absolute magnitudes using distances con-
structed from .Gaia parallaxes and apply a Sobel filter to find a weighted average
calibration of. M IT RG B =.−4.042.± 0.041 (stat).± 0.031 (sys). Caution should be used,
however, when directly converting apparent magnitudes to absolute magnitudes via
parallaxes due to asymmetric probability distributions (see Bailer-Jones et al. [53]).
Li et al. developed a two-dimensional maximum likelihood estimation algorithm to
ensure simultaneous treatment of uncertainties in parallaxes and magnitudes. The
algorithm was first introduced in Li et al. [42] and optimizes model parameters for
a broken power law luminosity function and exponential density distribution to find
the set of parameters that maximizes the likelihood of observing the sample of field
giants. This first study used .Gaia EDR3 parallaxes and APASS DR9 photometry
and was subsequently improved using .Gaia DR3 synthetic photometry in Li et al.
2023 [43] to find a calibration of . M IT RG B = .−3.970 .+0.042
−0.024 (sys) .± 0.062 (stat). The
algorithm used for these two studies by Li et al. is publicly available via the GitHub
repository link provided in Sect. 8.3.2. The differences between the Dixon et al. [52]
and Li et al. [43] field star results can be due to several factors, such as a different
treatment of the .Gaia parallax zero-point offset (differences of 10 .μas), which is
still not well understood, and differences in measurement methodologies, with one
being in magnitude space and the other in parallax space.
. Gaia parallaxes can also be used to calibrate the TRGB in globular clusters
such as .ω Centauri. Soltis et al. [54] used .Gaia EDR3 parallaxes to calibrate the
TRGB measured by Bellazzini et al. [55] and find . M I,T RG B = .−3.97 .± 0.06 mag.
Li et al. [43] investigated the possibility of using .Gaia DR3 synthetic photometry
to measure the TRGB in .ω Centauri but found evidence of blending that prevents it
from being used to obtain an accurate TRGB calibration. They instead update the
TRGB measurement in .ω Centauri using the database from Stetson et al. [56] to find
. M I,T RG B = .−3.97 .± 0.04 (stat) .± 0.10 (sys) mag.

Higher precision parallaxes and synthetic photometry in future.Gaia data releases

can further improve these TRGB calibrations. We note that a better understanding
of the .Gaia parallax zero-point offset will be essential before .Gaia data can be
used to obtain a TRGB calibration that is competitive in precision with calibrations

4 https://gea.esac.esa.int/archive/.
8 Tip of the Red Giant Branch 151

using other anchors such as NGC 4258. At the moment, the parallax zero-point
offset is not well understood; the standard offset is described in Lindegren et al. [57],
but other estimates can differ [58–62]. Regardless, .Gaia has the potential to help
increase leverage in TRGB . H0 calibrations through observations independent from
those used for other anchors.

Acknowledgements The authors would like to thank Dr. Gagandeep Anand for helpful comments.
S. L. is supported by the National Science Foundation Graduate Research Fellowship Program under
grant number DGE2139757. R. L. B. is supported by the National Science Foundation through grant
number AST-2108616.

References

1. I. Iben, A. Renzini, Ann. Rev. Astron. Astrophys. 21, 271–342 (1983). https://doi.org/10.1146/
annurev.aa.21.090183.001415
2. R.L. Beaton, G. Bono, V.F. Braga, M. Dall’Ora, G. Fiorentino, I.S. Jang, C.E. Martínez-
Vázquez, N. Matsunaga, M. Monelli, J.R. Neeley et al. https://doi.org/10.1007/s11214-018-
0542-1. ([arXiv:1808.09191 [astro-ph.GA]].)
3. J.P. Greco, S. Danieli, Astrophys. J. 941(1), 26 (2022). https://doi.org/10.3847/1538-4357/
ac75b7. ([arXiv:2109.13943 [astro-ph.GA]].)
4. S. Li, A.G. Riess, D. Scolnic, G.S. Anand, J. Wu, S. Casertano, W. Yuan, R. Beaton,
R.I. Anderson, Astrophys. J. 956(1), 32 (2023). https://doi.org/10.3847/1538-4357/acf4fb.
([arXiv:2306.10103 [astro-ph.GA]].)
5. L. Rizzi, R.B. Tully, D. Makarov, L. Makarova, A.E. Dolphin, S. Sakai, E.J. Shaya, Astrophys. J.
661, 815–829 (2007). https://doi.org/10.1086/516566. ([arXiv:astro-ph/0701518 [astro-ph]].)
6. I.S Jang, M.G. Lee, Astrophys. J. 835(1), 28 (2017). https://doi.org/10.3847/1538-4357/
acdd7b. ([arXiv:2211.06354 [astro-ph.CO]].)
7. J. Wu, D. Scolnic, A.G. Riess, G.S. Anand, R. Beaton, S. Casertano, X. Ke, S. Li, Astrophys.
J. 954(1), 87 (2023). https://doi.org/10.3847/1538-4357/acdd7b. ([arXiv:2211.06354 [astro-
ph.CO]].)
8. G.S. Anand, R.B. Tully, L. Rizzi, A.G. Riess, W. Yuan, Astrophys. J. 932(1), 15 (2022). https://
doi.org/10.3847/1538-4357/ac68df. ([arXiv:2108.00007 [astro-ph.CO]].)
9. I.S Jang, T.J. Hoyt, R.L. Beaton, W.L. Freedman, B.F. Madore, M.G. Lee, J.R. Neeley, A.J. Mon-
son, J.A. Rich, M. Seibert, Astrophys. J. 906(2), 125 (2017). https://doi.org/10.3847/1538-
4357/abc8e9. ([arXiv:2008.04181 [astro-ph.GA]].)
10. E.F. Schlafly, D.P. Finkbeiner, Astrophys. J. 737, 103 (2011). https://doi.org/10.1088/0004-
637X/737/2/103. ([arXiv:1012.4804 [astro-ph.GA]].)
11. T. Prusti et al., [Gaia], Astron. Astrophys. 595(1), A1 (2016). https://doi.org/10.1051/0004-
6361/201629272. ([arXiv:1609.04153 [astro-ph.IM]].) (Gaia Data Release)
12. G. Pietrzyński, D. Graczyk, W. Gieren, I.B. Thompson, B. Pilecki, A. Udalski, I. Soszyński,
S. Kozłowski, P. Konorski, K. Suchomska et al., Nature 495, 76–79 (2013). https://doi.org/10.
1038/nature11878. ([arXiv:1303.2063 [astro-ph.GA]].)
13. D. Gracyzk, G. Pietrzyński, I.B. Thompson, W. Gieren, B. Zgirski, S. Villanova, M. Górski,
P. Wielgórski, P. Karczmarek, W. Narloch et al., Astrophys. J. 904(1), 13 (2020). https://doi.
org/10.3847/1538-4357/abbb2b. ([arXiv:2010.08754 [astro-ph.GA]].)
14. M.J. Reid, D.W. Pesce, A.G. Riess, Astrophys. J. Lett. 886(2), L27 (2019). https://doi.org/10.
3847/2041-8213/ab552d. ([arXiv:1908.05625 [astro-ph.GA]].)
15. R.I. Anderson, N.W. Koblischke, L. Eyer, [arXiv:2303.04790 [astro-ph.CO]]
152 S. Li and R. L. Beaton

16. L.L. Kiss, T. Bedding, Mon. Not. R. Astron. Soc. 343, L79 (2003). https://doi.org/10.1046/j.
1365-8711.2003.06931.x. ([arXiv:astro-ph/0306426 [astro-ph]].)
17. J.J. Wray, L. Eyer, B. Paczynski, Mon. Not. R. Astron. Soc. 349, 1059 (2004). https://doi.org/
10.1111/j.1365-2966.2004.07587.x. ([arXiv:astro-ph/0310578 [astro-ph]].)
18. I. Soszynski, A. Udalski, M. Kubiak, M. Szymanski, G. Pietrzynski, K. Zebrun, O. Szewczyk,
L. Wyrzykowski, Acta Astron. 54, 129 (2004). ([arXiv:astro-ph/0407057 [astro-ph]])
19. I. Soszynski, A. Udalski, M.K. Szymanski, M. Kubiak, G. Pietrzynski, L. Wyrzykowski, O.
Szewczyk, K. Ulaczyk, R. Poleski, Acta Astron. 59, 239 (2009). ([arXiv:0910.1354 [astro-
ph.SR]])
20. T.J. Hoyt, [arXiv:2106.13337 [astro-ph.GA]]
21. D. Scolnic, A.G. Riess, J. Wu, S. Li, G.S. Anand, R. Beaton, S. Casertano, R.I. Anderson,
S. Dhawan, X. Ke, Astrophys. J. Lett. 954(1), L31 (2023). https://doi.org/10.3847/2041-8213/
ace978. ([arXiv:2304.06693 [astro-ph.CO]].)
22. D. Scolnic, D. Brout, A. Carr, A.G. Riess, T.M. Davis, A. Dwomoh, D.O. Jones, N. Ali,
P. Charvu, R. Chen et al., Astrophys. J. 938(2), 113 (2022). https://doi.org/10.3847/1538-
4357/ac8b7a. ([arXiv:2112.03863 [astro-ph.CO]].)
23. K. Krisciunas, C. Contreras, C.R. Burns, M.M. Phillips, M.D. Stritzinger, N. Morrell,
M. Hamuy, J. Anais, L. Boldt, L. Busta et al., Astron. J. 154(5), 211 (2017). https://doi.org/10.
3847/1538-3881/aa8df0. ([arXiv:1709.05146 [astro-ph.IM]].)
24. M. Hamuy, G. Folatelli, N.I. Morrell, M.M. Phillips, N.B. Suntzeff, S.E. Persson, M. Roth,
S. Gonzalez, W. Krzeminski, C. Contreras et al., Publ. Astron. Soc. Pac. 118, 2–20 (2006).
https://doi.org/10.1086/500228. ([arXiv:astro-ph/0512039 [astro-ph]].)
25. A.G. Riess, W. Yuan, L.M. Macri, D. Scolnic, D. Brout, S. Casertano, D.O. Jones, Y. Murakami,
L. Breuval, T.G. Brink et al., Astrophys. J. Lett. 934(1), L7 (2022).https://doi.org/10.3847/
2041-8213/ac5c5b. ([arXiv:2112.04510 [astro-ph.CO]].)
26. W.L. Freedman, Astrophys. J. 919(1), 16 (2021). https://doi.org/10.3847/1538-4357/ac0e95.
([arXiv:2106.15656 [astro-ph.CO]].)
27. W.L. Freedman, B.F. Madore, D. Hatt, T.J. Hoyt, I.S. Jang, R.L. Beaton, C.R. Burns, M.G.
Lee, A.J. Monson, J.R. Neeley et al., Astrophys. J. 882, 34 (2019). https://doi.org/10.3847/
1538-4357/ab2f73. ([arXiv:1907.05922 [astro-ph.CO]].)
28. W. Cerny, W.L. Freedman, B.F. Madore, F. Ashmead, T. Hoyt, E. Oakes, N.Q.H. Tran, B. Moss
(arXiv:2012.09701 [astro-ph.GA]].)
29. D.M. Scolnic et al., [Pan-STARRS1], Astrophys. J. 859(2), 101 (2018). https://doi.org/10.
3847/1538-4357/aab9bb. ([arXiv:1710.00845 [astro-ph.CO]].)
30. S. Dhawan, S. Thorp, K.S. Mandel, S.M. Ward, G. Narayan, S.W. Jha, T. Chant, Mon.
Not. R. Astron. Soc. 524(1), 235–244 (2023). https://doi.org/10.1093/mnras/stad1590.
([arXiv:2211.07657 [astro-ph.CO]].)
31. S. Dhawan, A. Goobar, J. Johansson, I.S. Jang, M. Rigault, L. Harvey, K. Maguire, W.L. Freed-
man, B.F. Madore, M. Smith et al., Astrophys. J. 934(2), 185 (2022). https://doi.org/10.3847/
1538-4357/ac7ceb. ([arXiv:2203.04241 [astro-ph.CO]].)
32. M.J. Graham, S.R. Kulkarni, E.C. Bellm, S.M. Adams, C. Barbarino, N. Blagorodnova,
D. Bodewits, B. Bolin, P.R. Brady, S.B. Cenko et al., Publ. Astron. Soc. Pac. 131(1001), 078001
(2019). https://doi.org/10.1088/1538-3873/ab006c. ([arXiv:1902.01945 [astro-ph.IM]].)
33. J.P. Blakeslee, J.B. Jensen, C.P. Ma, P.A. Milne, J.E. Greene, Astrophys. J. 911(1), 65 (2021).
https://doi.org/10.3847/1538-4357/abe86a. ([arXiv:2101.02221 [astro-ph.CO]].)
34. M.G. Lee, W.L. Freedman, B.F. Madore, Astrophys. J. 417, 553 (1993). https://doi.org/10.
1086/173334
35. D. Hatt, R.L. Beaton, W.L. Freedman, B.F. Madore, I.S. Jang, T.J. Hoyt, M.G. Lee, A.J. Monson,
J.A. Rich, V. Scowcroft et al., Astrophys. J. 845(2), 146 (2017). https://doi.org/10.3847/1538-
4357/aa7f73. ([arXiv:1703.06468 [astro-ph.CO]].)
36. S.E. Persson, B.F. Madore, W. Krzemiński, W.L. Freedman, M. Roth, D.C. Murphy, Astron.
J. 128, 5 (2004). https://doi.org/10.1086/424934
37. A.J. Monson, R.L. Beaton, V. Skowcroft, W.L. Freedman, B.F. Madore, J.A. Rich, M. Seib-
ert, J.A. Kollmeier, G. Clementini, Astron. J. 153, 3 (2017). https://doi.org/10.1086/424934.
([arXiv:1703.01520 [astro-ph.SR]].)
8 Tip of the Red Giant Branch 153

38. M. Zoccali, G. Piotto, Astron. Astrophys. 358, 943 (2000). ([arXiv:astro-ph/0005034 [astro-
ph]])
39. B. Mendez, M. Davis, J. Moustakas, J. Newman, B.F. Madore, W.L. Freedman, Astron. J. 124,
213 (2002). https://doi.org/10.1086/341168. ([arXiv:astro-ph/0204192 [astro-ph]].)
40. D. Makarov, L. Makarova, L. Rizzi, R.B. Tully, A.E. Dolphin, S. Sakai, E.J. Shaya, Astron.
J. 132, 2729–2742 (2006). https://doi.org/10.1086/508925. ([arXiv:astro-ph/0603073 [astro-
ph]].)
41. P.F. Wu, R.B. Tully, L. Rizzi, A.E. Dolphin, B.A. Jacobs, I.D. Karachentsev, Astron. J. 148, 7
(2014). https://doi.org/10.1088/0004-6256/148/1/7. ([arXiv:1404.2987 [astro-ph.GA]].)
42. S. Li, S. Casertano, A.G. Riess, Astrophys. J. 939(2), 96 (2022). https://doi.org/10.3847/1538-
4357/ac7559. ([arXiv:2202.11110 [astro-ph.GA]].)
43. S. Li, S. Casertano, A.G. Riess, Astrophys. J. 950(2), 83 (2023. https://doi.org/10.3847/1538-
4357/accd69. ([arXiv:2304.06695 [astro-ph.GA]].)
44. P. Virtanen, R. Gommers, T.E. Oliphant, M. Haberland, T. Reddy, D. Cournapeau, E. Burovski,
P. Peterson, W. Weckesser, J. Bright et al., Nature Meth. 17, 261 (2020). https://doi.org/10.
1038/s41592-019-0686-2. ([arXiv:1907.10121 [cs.MS]].)
45. G.S. Anand, R.B. Tully, L. Rizzi, I.D. Karachentsev, Astrophys. J. Lett. 872(1), L4 (2019).
https://doi.org/10.3847/2041-8213/aafee6. ([arXiv:1901.05981 [astro-ph.GA]].)
46. D.J. Radburn-Smith, R.S. de Jong, A.C. Seth, J. Bailin, E.F. Bell, T.M. Brown, J.S Bullock,
S. Courteau, J.J. Dalcanton, H.C. Ferguson et al., Astrophys. J. Suppl. 195, 2 (2011). https://
doi.org/10.1088/0067-0049/195/2/18
47. B. Harmsen, E.F. Bell, R. D’Souza, A. Monachesi, R.S. de Jong, A. Smercina, I. Jang, B.W. Hol-
werda, Mon. Not. R. Astron. Soc. 525(3), 4497–4514 (2023). https://doi.org/10.1093/mnras/
stad2480. ([arXiv:2308.11499 [astro-ph.GA]].)
48. R.B. Tully, L. Rizzi, E.J. Shaya, H.M. Courtois, D.I. Makarov, B.A. Jacobs, Astron. J. 138,
323–331 (2009). https://doi.org/10.1088/0004-6256/138/2/323. ([arXiv:0902.3668 [astro-
ph.CO]].)
49. G.S. Anand, L. Rizzi, R.B. Tully, E.J. Shaya, I.D. Karachentsev, D.I. Makarov, L. Makarova,
P. Wu, A.E. Dolphin, E. Kourkchi, Astron. J. 162(2), 80 (2021). https://doi.org/10.3847/1538-
3881/ac0440. ([arXiv:2104.02649 [astro-ph.GA]].)
50. K.B.W. McQuinn, M. Boyer, E.D. Skillman, A.E. Dolphin, Astrophys. J. 880(1), 63 (2019).
https://doi.org/10.3847/1538-4357/ab2627. ([arXiv:1904.01571 [astro-ph.GA]].)
51. P. Montegriffo et al., [Gaia], Astron. Astrophys. 674(3), A33 (2023). https://doi.org/10.1051/
0004-6361/202243709 (Gaia Data Release 3)
52. M. Dixon, J. Mould, C. Flynn, E.N. Taylor, C. Lidman, A.R. Duffy, Mon. Not. R. Astron.
Soc. 523(2), 2283–2295 (2023). https://doi.org/10.1093/mnras/stad1500. ([arXiv:2305.09215
[astro-ph.GA]].)
53. C.A.L. Bailer-Jones, J. Rybizki, M. Fouesneau, M. Demleitner, R. Andrae, Astron. J. 161(3),
147 (2021). https://doi.org/10.3847/1538-3881/abd806. ([arXiv:2012.05220 [astro-ph.SR]].)
54. J. Soltis, S. Casertano, A.G. Riess, Astrophys. J. Lett. 908(1), L5 (2021). https://doi.org/10.
3847/2041-8213/abdbad. ([arXiv:2012.09196 [astro-ph.GA]].)
55. M. Bellazzini, F.R. Ferraro, E. Pancino, Astrophys. J. 556, 635–640 (2001). https://doi.org/10.
1086/321613. ([arXiv:astro-ph/0104114 [astro-ph]].)
56. P.B. Stetson, E. Pancino, A. Zocchi, N. Sanna, M. Monelli, Mon. Not. R. Astron. Soc.
485(3), 3042–3063 (2019). https://doi.org/10.1093/mnras/stz585. ([arXiv:1902.09925 [astro-
ph.SR]].)
57. L. Lindegren, S.A. Klioner, J. Hernández, A. Bombrun, M. Ramos-Lerate, H. Steidelmüller,
U. Bastian, M. Biermann, A. de Torres, E. Gerlach et al., Astron. Astrophys. 649 Gaia Data
Release 3, A2 (2021). https://doi.org/10.1051/0004-6361/202039709. ([arXiv:2012.03380
[astro-ph.IM]].)
58. Y. Huang, H. Yuan, T.C. Beers, H. Zhang, Astrophys. J. Lett. 910(1), L5 (2021). https://doi.
org/10.3847/2041-8213/abe69a.Huang:2021. ([arXiv:2101.09691 [astro-ph.GA]].)
59. M.A.T. Groenewegen, Astron. Astrophys. 654, A20 (2021). https://doi.org/10.1051/0004-
6361/202140862. ([arXiv:2106.08128 [astro-ph.GA]].)
154 S. Li and R. L. Beaton

60. F. Ren, X. Chen, H. Zhang, R. de Grijs, L. Deng, Y. Huang, Astrophys. J. Lett. 911(2), L20
(2021). https://doi.org/10.3847/2041-8213/abf359. ([arXiv:2103.16096 [astro-ph.SR]].)
61. J.C. Zinn, Astron. J. 161(5), 214 (2021). https://doi.org/10.3847/1538-3881/abe936.
([arXiv:2101.07252 [astro-ph.GA]].)
62. J. Maíz Apellániz, Astron. Astrophys. 657, A130 (2022). https://doi.org/10.1051/0004-6361/
202142365. ([arXiv:2110.01475 [astro-ph.IM]].)
63. I. Ribas, C. Jordi, F. Vilardell, E.L. Fitzpatrick, R.W. Hilditch, E.F. Guinan, Astrophys. J. Lett.
635, L37–L40 (2005). https://doi.org/10.1086/499161. ([arXiv:astro-ph/0511045 [astro-ph]].)
64. F. Vilardell, I. Ribas, C. Jordi, E.L. Fitzpatrick, E.F. Guinan, Astron. Astrophys. 509, A70
(2010). https://doi.org/10.1051/0004-6361/200913299. ([arXiv:0911.3391 [astro-ph.CO]].)
65. L. Lindegren, J. Hernandez, A. Bombrun, S. Klioner, U. Bastian, M. Ramos-Lerate, A. de
Torres, H. Steidelmuller, C. Stephenson, D. Hobbs et al., Astron. Astrophys. 616, A2 (2018).
https://doi.org/10.1051/0004-6361/201832727. ([arXiv:1804.09366 [astro-ph.IM]].)
Chapter 9
Surface Brightness Fluctuations

Michele Cantiello and John P. Blakeslee

9.1 Introduction

Stellar counts have long been used to study unresolved or partially-resolved stellar
populations [1]. However, it was not until the development of CCD astronomy that
the statistical characteristics of the discrete nature of stellar counts allowed for the
formulation of the surface brightness fluctuations (SBF) method [50, 51], one of the
most reliable and stable extragalactic distance indicators used within the range of a
few to about 150 Mpc.
As the distance to a stellar population increases, individual stars blend together
and the brightness profile becomes smoother, but there remain statistical fluctua-
tions on small scales because of the discrete nature of the stars. The SBF method
leverages the random nature of stellar counts and luminosities to measure a quantity
that corresponds to the luminosity-weighted mean brightness of the stars in the stel-
lar population. For evolved stellar systems, this is approximately equal to the mean
brightness of the stars on the red giant branch (RGB).
The SBF method was initially developed to determine distances of relatively
nearby elliptical galaxies observed from the ground. Over the past few decades,
the technique has undergone improvements in the analysis tools and benefited from
increasingly advanced astronomical facilities. Consequently, SBF distances are now
derived for a wider class of objects beyond ellipticals: spiral bulges, dwarf galaxies,
irregulars, and others. Moreover, the distance limit has increased by a factor of .∼7
from the original predictions, and will increase further in the coming years.

M. Cantiello (B)
INAF-Astronomical Observatory of Abruzzo, Via Maggini Snc, 64020 Teramo, Italy
e-mail: michele.cantiello@inaf.it
J. P. Blakeslee (B)
NSF’s NOIRLab, Tucson, AZ 85719, USA
e-mail: john.blakeslee@noirlab.edu

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 155
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_9
156 M. Cantiello and J. P. Blakeslee

Quantitatively, what is measured in the SBF analysis is the ratio between the
intrinsic brightness variance (in the absence of blurring) and the average surface
brightness over a specific region of a galaxy. This ratio has the units of flux and it is
equivalent to the ratio between the second and first moments of the stellar luminosity
function within the region analyzed. Although the definition is simple, there are a
number of intricacies in practice. These involve how to measure the SBF amplitudes
in astronomical images, the optimal targets for the measurement, and how to calibrate
SBF magnitudes, either empirically or using stellar population synthesis.
In the following sections, we provide a detailed description of the method and
summarize the recent results, especially regarding the “Hubble tension.” We conclude
with a discussion of the future prospects for the method.

9.2 Characterization of the SBF Signal

The main conceptual difficulty related to SBF does not pertain to the measurement
of fluctuations or their calibration, which we discuss below. Rather, a cause of con-
fusion, for some, concerns the nature of the indicator itself. Unlike other commonly
used distance indicators like parallaxes, Cepheid variables, and Type Ia supernovae,
the SBF signal is not easily “visualized”, and it is not immediately apparent how
fluctuations in star counts can be converted into a reliable estimate of distance.
To grasp how the SBF signal is produced and used for distance estimation, we
can consider a simplified model of a CCD image of a stellar population unblurred by
a point spread function (PSF). Let all the stars have the same (unknown) luminosity
. L ∗ , with .n ∗ being the average number of stars per pixel and .d the distance. Then,
. f ∗ = L ∗ /(4π d ) is the flux contributed per star. We cannot directly measure . f ∗ , but
2

we can measure the surface brightness . F∗ as the mean flux per pixel, . F∗ = n ∗ × f ∗ .
Since .n ∗ ∝ d 2 and . f ∗ ∝ 1/d 2 , . F∗ is independent of the distance of the stellar system.
We can also measure the standard deviation in the flux per pixel. Because of
√
Poisson fluctuations in the numbers of stars, the standard deviation is .σ F = n ∗ f ∗ .
Now, we define the SBF flux as the ratio of the variance .σ F2 to the mean:

. f¯ = σ F2 /F∗ = n ∗ f ∗2 /(n ∗ f ∗ ) = f ∗ = L ∗ /(4π d 2 ). (9.1)

Thus, in this trivial case where the stars are all identical, the SBF flux reduces to the
mean flux per star, which depends inversely on the square of the distance.
In reality, stars follow a luminosity function, and a more realistic treatment [39]
shows that the ratio of the variance to the mean surface brightness reduces to:
∑
n i f i2 L̄
. f¯ = ∑i ≡ , (9.2)
n
i i if 4π d2
9 Surface Brightness Fluctuations 157

where .n i is the number of stars of apparent brightness . f i , and . L̄ is the ratio of the
second and first moments of the stellar luminosity function, corresponding to the
luminosity-weighted mean stellar luminosity of the population. For a more detailed
derivation of . f¯, see [18] or Sect. 3.9.1 of [39]. . L̄ is easily calculated from stellar
population models; because of the luminosity weighting, it is sensitive to the brightest
stars in the population. Thus, the SBF flux f¯, or the equivalent SBF magnitude m,
is solely determined by the stellar luminosity function and the galaxy distance.

9.3 From SBF Magnitudes to Distances

Similar to other standard candle methods, deriving SBF distances requires measuring
an apparent magnitude, m, and then adopting an absolute M based on a calibration
relation to obtain the distance modulus, .m−M. The following sections describe the
process for measuring m and estimating M, as well as the characteristics of suitable
target galaxies and the preferred passbands for reliable SBF measurements.

9.3.1 SBF Measurements Step by Step

Accurate SBF measurements require correcting and accounting for the different
sources of noise or contamination that afflict the signal. In a sense, the fluctuation
signal itself is a sort of noise in the star counts per pixel; in the absence of detector
noise, photon shot noise, contamination from background sources, and blurring by
the PSF, the SBF would simply be the Poisson variance among pixels due to the
varying number and luminosity of stars in each pixel, normalized to the local mean
flux of the galaxy. However, all of these effects are present and must be dealt with.
Since the SBF signal is convolved with the PSF, it can be distinguished from
photon and detector noise, which have white-noise power spectra, at least prior to
distortion correction. For this reason, the analysis is done in the Fourier domain,
where the SBF signal is measured by determining the amplitude of the variance
component in the image power spectrum on the scale of the PSF. Numerous papers
have described the steps involved in SBF measurements [6, 7, 14, 17, 28, 30, 38,
39, 51]. In general, the procedure can be summarized as follows.
After combining individual dithered exposures to make a cleaned, stacked image
free from cosmic rays, satellite trails, and detector defects, we estimate and subtract
the sky background, mask any bright stars and neighboring galaxies, then derive a
smooth isophotal model of the galaxy surface brightness. This model is then sub-
tracted from the image, the remaining point-like and extended sources are masked,
and a low-order fit to the background is derived and subtracted to eliminate large-scale
residuals from the galaxy model subtraction. This masking, modeling and subtrac-
tion procedure is typically iterative. We then detect and mask all compact objects
(foreground stars, globular clusters in the galaxy, faint background galaxies) down to
158 M. Cantiello and J. P. Blakeslee

a signal-to-noise threshold of .∼ 4.5 using an automatic photometry program. These

objects, along with any other sources of non-SBF variance (visible dust, brighter
satellite galaxies, tidal features, regions of poor model residuals, etc.) are masked
after the smooth model and large-scale residuals are removed.
The residual masked frame is then normalized by the square root of the model
and contains both the variance from the stellar fluctuations of interest as well as
fluctuations from unexcised sources fainter than the detection limit, along with the
unconvolved photon and detector noise. The next step is to analyze the power spec-
trum of the normalized residual masked frame. The fluctuations from the stellar
counts and unexcised sources are convolved with the instrumental PSF in the spatial
domain; hence, they are multiplied by the Fourier transform of the PSF in the Fourier
domain. For accurate SBF measurements, it is crucial to have an accurate character-
ization of the local PSF. This can be achieved using individual point sources in the
field or by utilizing PSF modeling and reconstruction methods.
After azimuthally averaging the power spectrum,1 the amplitude . P0 of the astro-
physical component of the normalized and masked residual frame’s power spectrum
. P(k) is obtained as the constant term that multiples the “expectation power spectrum”

. E(k), which is calculated as:. E(k) = ~ PSF(k) ⊗ M ~ (k) is the power spec-
~ (k), where.M
~
trum of the normalized PSF, and .PSF(k) is the power spectrum of the mask window
function. There is also the white noise (approximately independent of wavenumber
.k) component . P1 ; thus the power spectrum is modeled as: . P(k) = P0 × E(k) + P1 .

If the large-scale light distribution is well-subtracted, the power spectrum will be

well represented by these two components over most the .k range. Additional power
may exist at low .k (largest spatial scales) due to imperfect galaxy subtraction. The
power may also be suppressed at some values of.k due to correlation of the pixel noise
resulting from geometric distortion correction; this is especially problematic when
using linear interpolation of the pixel values. However, for sinc-like interpolation
kernels [14, 38], the “damage” to the power spectrum is confined to high .k (the
smallest scales). In this case, the highest and lowest wavenumbers can simply be
omitted from the fit range to ensure an accurate measurement.
It is crucial to note that the fitted . P0 value includes contamination from unexcised
sources fainter than the detection limit. Any astrophysical source of fluctuation that is
not generated from the stellar counts must be estimated and removed by subtracting
the contribution from . P0 . The presence of dust in some galaxies may also affect the
power spectrum, and hence the fitted . P0 , but in ways that depend on the distribution
and are not easy to quantify. Optical images and color maps can be used to identify
and mask any dusty regions. Typically the SBF measurement is confined to “clean
regions,” and we assume all dust patches are identified and masked; thus, no further
contribution is considered in . P0 .
In most cases, the globular clusters (GCs) within the target galaxy are the main
source of contamination for . P0 , with faint background galaxies being a lesser con-
taminant. We estimate the extra power from these sources by fitting, down to the

1 The power spectrum is by definition the squared modulus of the Fourier transform; this explains
the normalization of the residual frame by the square root of the galaxy model.
9 Surface Brightness Fluctuations 159

detection limit, a combined luminosity function that includes a radially-dependent

Gaussian GC luminosity function (GCLF; e.g., [26]) and a power-law LF for the
galaxies. We then estimate the “residual power” . Pr from sources fainter than this
limit (which varies with radius) by extrapolating the fitted combined LF, multiplied
by the square of the source flux, from the radially-dependent detection limit down
to zero flux (see [10]). This residual power is then subtracted from . P0 to obtain the
intrinsic stellar fluctuations, typically denoted . P f = P0 − Pr .
Because the GCs have a radial density distribution that increases toward the galaxy
center, while the detection limit also gets brighter near the center, the fluctuations
from GCs in the central region can overwhelm the signal from the stellar fluctuations,
especially for massive galaxies with rich GC systems. Space-based observations (or
high-resolution ground-based data obtained with adaptive optics [AO]) can greatly
improve the distance limit for SBF by reducing the effect of crowding and pushing
the detection limit fainter for the GCs. The surface density of background galaxies is
typically lower and relatively uniform over the area, which means that their impact
is less dramatic and is easier to robustly constrain.
Finally, the SBF magnitude is then calculated as .m = −2.5 log(P0 − Pr ) + m ZP ,
where .m ZP is the zero point magnitude. For a schematic illustration of the intricate
SBF analysis procedure, see Fig. 9.1.

9.3.2 SBF Calibration

To convert m into a distance, knowledge of the absolute SBF magnitude M of the pop-
ulation is required. The most common method for estimating M is through empirical
relationships. Studies of large samples of galaxies at known distances or in compact
groups have shown that M can be described as a one-parameter family with integrated
color (e.g., .V −I or .g−z) as a useful parameter. The ground-based SBF survey of
Tonry and collaborators [7, 52–54] found that a linear relation describes accurately
the color dependence of .m I in nearby groups and clusters. They also showed that M
in the . I band is a universal function of color, with an intrinsic scatter in the range of
0.05 to 0.10 mag. Later studies confirmed a similar intrinsic scatter in the calibration
relation for red galaxies in suitable passbands [9, 28], although for high-quality SBF
measurements over a large color range, the M-vs-color relation is more accurately
described by a higher-order polynomial curve [9].
Since the introduction of SBF, many empirical calibrations of M as a function of
various optical and near-IR colors have been presented [6, 11, 28, 37]. However, a
major drawback of empirical approaches is the need to establish a new calibration
whenever the reference passband for SBF or the reference color of the observing
campaign changes. An alternative approach is to use SBF magnitude versus color
relations derived from stellar population synthesis models [11, 16, 56]. The advan-
tage of theoretical calibrations is that M can easily be computed for any passband and
reference color. Furthermore, fully theoretical calibrations are independent of other
distance indicators, avoiding systematic uncertainties associated with the primary
160 M. Cantiello and J. P. Blakeslee

Fig. 9.1 Illustration of SBF observations and measurements..(a) Simulation of the stellar population
in a spheroidal galaxy at the distance of the Virgo cluster (. DV irgo ≃ 16.5 Mpc [9]) as observed with
the E-ELT in .∼1 h. .(b) Same as in panel .(a), but for a galaxy ten times more distant. .(c) Same as in
panel.(a), but for a galaxy fifty times more distant. Stars, which appear marginally resolved in panel
.(a), blend together into a smooth brightness profile at larger distances. .(d) Near-infrared image of
NGC 1399 from the HST WFC3 camera. .(e) Model of NGC 1399s surface brightness distribution
derived from the WFC3/IR image. .( f ) Residual frame, obtained from the galaxy image .(d) minus
the model .(e). .(g) Typical luminosity function analysis for estimating the “residual variance” . Pr
due to contaminating sources: green squares show the data, the blue curve and red line show the
fits to the globular cluster and background galaxy luminosity functions, respectively, and the solid
black line is the combined model luminosity function (data and fits are from [12]). The vertical
gray dashed line indicates the GCLF turnover magnitude and the shaded area shows the magnitude
interval where the detection is incomplete..(h) Color-magnitude diagram of an old stellar population
(data for the MW globular cluster NGC 1851 [43]); the RGB/AGB population is highlighted with
red dots..( f ) A schematic illustration of the SBF power spectrum analysis. Images reproduced with
permission from [39, 43]
9 Surface Brightness Fluctuations 161

indicator. Consequently, theoretically calibrated SBF can be regarded as a primary

(though not geometric) distance indicator [3].
There is generally good agreement in SBF predictions from independent stellar
population synthesis models in red optical bands. Discrepancies arise when com-
paring SBF magnitudes predicted by different models in bluer passbands or in the
near-IR, where the models are not yet as robust [28]. These discrepancies can exceed
.∼0.2 mag. Such a difference is considerably larger than the empirically derived

scatter, rendering stellar population model predictions unreliable as an alternative to

empirical . M. Therefore, except for a few notable cases [2], numerically derived M
vs color relations serve mainly as validation tools for empirical calibrations.
Before concluding this subsection, it is important to emphasize that accurate
distance estimation using . M calibrations requires precise galaxy colors. Hence, a
reliable photometric calibration is crucial to ensure consistency and precision in the
distances derived.

9.3.3 Reference Targets and Passbands

Owing to the definition of SBF and the measurement procedures outlined above,
the ideal target galaxy for fluctuation measurements should have minimal dust, a
brightness profile that is sufficiently regular, and a relatively homogeneous population
of stars with a constant or at least not highly variable stellar luminosity function across
its isophotes. This ensures a relatively constant SBF signal across the regions of the
galaxy where the measurement is carried out, reducing the intrinsic variation of m
and resulting in a more accurate distance estimate with lower errors.
In short, the ideal SBF targets are passively evolving, massive red ellipticals that
are free from dust contamination and have a well-mixed population of stars with no
recent star formation. Empirical calibrations have shown that such galaxies exhibit
a tight correlation between absolute SBF magnitude and optical or near-IR colors,
enabling SBF distances with a contribution from intrinsic scatter as low as 0.06 mag
in certain passbands2 [6, 9, 17]. Stellar population models also support the narrow
scatter of M for colors typical of “red and dead” galaxies with stellar populations
dominated by an old, metal-rich component [11, 16].
While the SBF method was originally developed for bright, morphologically reg-
ular early-type galaxies, it has more recently been used for a wider variety of stellar
systems, including bulges of spirals and the low mass ultra-diffuse galaxies [4, 20].
In contrast to massive red ellipticals, bluer galaxies with intermediate masses have
properties consistent with younger ages, lower metallicities, or both. Galaxies in this
color regime have more uncertainties from the age-metallicity degeneracy,3 which

2The final distance errors are larger due to measurement uncertainties, as discussed in Sec. 9.4.
3The age-metallicity degeneracy refers to the effect of stellar populations of a specific age and
metallicity displaying similar colors, SBF magnitudes, and certain spectral indices as populations
with a younger age and higher metallicity, or vice versa [57].
162 M. Cantiello and J. P. Blakeslee

can result in targets with different stellar populations having the same integrated
color. However, having the same integrated colors (which are mostly dominated by
main sequence stars) does not always mean the same SBF absolute magnitude. This
effect implies a larger scatter in the . M-vs-color relations as compared to red massive
galaxies, leading to larger scatter associated with the derived distances.
Additionally, galaxies with intermediate colors often host an intermediate-age
stellar population (3–7 Gyr), which includes many asymptotic giant branch (AGB)
stars. These stars are in a bright and rapidly evolving phase, hence their stochastic
appearance can make the SBF signal less stable, especially in near-IR bands [45].
Although AGBs have little impact on integrated colors they represent an additional
source of scatter to the SBF in these galaxies. As a result of degeneracy effects and
AGB stars, in some bands intermediate-mass targets suffer a larger scatter in SBF
amplitude at fixed color, which makes it challenging to measure individual distances
with an accuracy similar to the red-massive galaxies.
The accuracy of the SBF technique further drops for bluer, low-mass, low surface
brightness galaxies. This is partly because of increased scatter from stellar population
effects, including stochastic variations in the AGB stars. Additionally, measuring SBF
in diffuse blue galaxies is challenging because low stellar densities make it harder
to achieve an adequate signal-to-noise for the stellar-count fluctuations. As a result,
a precision of approximately 15% should be considered the norm for such galaxies
under optimal conditions [25].
Despite these challenges, the development of large-format imagers and wide-
field surveys on 4m-class telescopes has increased the interest in SBF as a tool
for determining distances to smaller galaxies. Recent studies have applied the SBF
technique as a distance indicator for galaxies with surface brightness levels as low
as .μr ∼ 24 mag/arcsec.2 [4, 17, 25, 33]. In these cases, the technique is typically
not exploited for its accuracy but rather because these faint and diffuse galaxies
do not offer many reliable alternatives for measuring distances. Despite the larger
uncertainty, the SBF technique represents a unique opportunity to confirm galaxy
associations and group or cluster membership for candidate satellite galaxies.
Concerning the preferred passbands, SBF distances have been measured in var-
ious filters, ranging from visible to near-infrared wavelengths. Shorter wavelength
bands are generally avoided because the SBF magnitudes are much fainter and more
susceptible to contamination from GCs and dust. Nevertheless, SBF data in bluer
bands can be valuable tracers for studying unresolved stellar populations due to their
sensitivity to stellar population properties [11, 16, 47, 56]. Redder optical (. I and .z)
and nearer IR (.Y J H ) bands are instead preferred for distance measurements.
Empirical studies have shown that the intrinsic scatter of M at a fixed galaxy
color is around .∼ 0.06 mag [6, 9] for the most massive ellipticals in optical bands.
In the near-IR bands, the intrinsic dispersion of M is closer to 0.10 mag due to the
larger variance arising mainly from bright AGB stars [15, 28]. Despite the somewhat
larger instrinsic scatter, near-IR bands such as . J and . H have several advantages. The
SBF is intrinsically brighter, up to three magnitudes brighter in . K than . I , although
the intrinsic scatter in . K is significant and not well characterized. Near-IR bands
also offer a much more favorable contrast with respect to GCs. Reliable m mea-
9 Surface Brightness Fluctuations 163

surements need to mask GCs to at least .∼ 0.5 mag fainter than the GCLF peak in
the . I band, while contamination can be reduced to the same level in the . K band by
reaching .∼ 2 mag brighter than the GCLF peak. Finally, the impact of residual dust
contamination in the near-IR is negligible in most cases [36].
The region around .∼ 1. μm (.Y band) is particularly interesting for SBF. According
to independent stellar population models, and as expected from the inversion of the
slope of the empirical SBF vs color relation between optical and near-IR bands for
intermediate-color evolved galaxies, this wavelength should exhibit nearly complete
degeneracy with respect to stellar population: m remains constant with color, making
it an ideal distance indicator [16, 56]. In practice, this does not hold for the reddest,
most massive ellipticals, for the which the SBF continues to get fainter with color,
but these galaxies are comparatively rare. For future wide-area surveys the .∼ Y band
may be the best choice for SBF measurements, although further study is needed.
In summary, there is not a definitive spectral “sweet spot” for measuring SBF, and
any passband from.∼ 0.8 to.∼ 2.μm has its pros and cons. With the rapid development
of near-IR technologies and the significant investment in future astronomical facilities
in these passbands (e.g., Roman, the ELTs, and the recently launched JWST), along
with the brighter intrinsic luminosity and negligible effect of dust contamination,
it seems likely that the future of the SBF method, in terms of distance/depth and
number of targets reached, lies primarily in the near-IR.

9.4 Uncertainties

9.4.1 Statistical Uncertainties

Statistical uncertainties affecting SBF can be grouped into three main categories:
m measurement errors, uncertainty in M due to photometric errors in the galaxy
color and intrinsic stellar-population scatter in the calibration relation, and uncer-
tainties from data processing and calibration. The last category includes effects like
flat fielding and uncertainty in the photometric calibration. While not unique to SBF,
it is important to note their presence since they can impact various parts of the
SBF measurement process in non-trivial ways. The uncertainty in the flat-fielding
multiplicative factor affects the m measurement by affecting both the variance and
average term in Eq. 9.2 and the color used in the calibration equation. Any sources
of uncertainty in the reduction stage impact in a relatively intertwined way the final
uncertainty on the distances from SBF. However, these uncertainties are typically
well-controlled with standard data reduction and calibration procedures, with a typ-
ical level of .∼ 1%. Regarding SBF measurement uncertainties, these are directly
tied to the specific steps required for the measurement. Usually, each error term
introduced during any of the various analysis stages described in Sect. 9.3.1, can be
independently estimated and then combined in quadrature with the others. The most
relevant error terms are the following.
164 M. Cantiello and J. P. Blakeslee

• Sky background. The process of sky subtraction has an associated error due to
spatial and/or temporal variations in the sky emission (especially in the near-IR,
where the sky background varies on the time scale of the observations), scattered
light, and the extended nature of the galaxy, especially if the field of view is com-
parable to the galaxy size. The effect of the sky uncertainty increases significantly
in the outer parts of the galaxy and for galaxies of lower surface brightness. To
determine the final error associated to the sky uncertainty, the entire SBF mea-
surement may be repeated by offsetting the background by the uncertainty level;
this alters the final m, and the size of the change is then taken into account for
the final uncertainty. The sky uncertainty typically contributes .∼ 0.02 mag of the
total error on m in the near-IR, and less than that in optical bands.
• Point Spread Function. Accurate characterization of the PSF is essential for a reli-
able SBF measurement, and this depends on the number of bright, unsaturated stars
in the field. Thus, for a given galaxy, the quality of the PSF template depends on
the size of the field, stability of the PSF across the field, signal-to-noise, saturation
level, etc. Once a reliable set of reference stars is identified, the SBF measurement
is repeated using each available PSF and possibly a composite PSF model. The
standard deviation of these measurements, typically less than .∼ 0.03 mag, is then
used as the PSF-fitting uncertainty.
• Power Spectrum Fitting. The accuracy of the power spectrum fit is affected by
several factors, such as the number of good pixels in the annulus, the spatial
structure of the mask, and the presence of patterns in the residual image caused by
bars, shells, tidal interaction features or spiral arms. Low- and high-wavenumber
power excesses or deficits can be filtered out during the power spectrum fit, yet the
range of useful .k-numbers needs to be large enough for reliable fitting. The fitting
of the PSF power spectrum to the galaxy-subtracted data involves a statistical fit
uncertainty of .∼0.02 mag.
• Residual Variance . Pr . Various approaches have been used to determine the uncer-
tainty in correcting for variance from GCs and background galaxies that are too
faint for direct detection. One approach is to examine the change in . Pr for differ-
ent LF parameters and source detection thresholds. Another involves using differ-
ent, independent detection and photometry tools (SExtractor, DAOphot, DoPHOT,
etc.). Based on a variety of tests, a typical uncertainty of about 25% is adopted for
the . Pr estimate; thus, to ensure the contribution to the total error on m from the
. Pr correction is .< 0.05 mag requires that . Pr /P f ∼ 0.2. To keep this error term
at an acceptable level, it is essential for the images to have sufficient depth; this
is vastly easier to achieve with the superior resolution and reduced background of
space-based observations.
• Extinction. Another component in the error budget is the uncertainty in the extinc-
tion correction, typically about 10% of the extinction correction itself. This is
added in quadrature to the other sources of statistical error.

As for uncertainty in the adopted value of M for a given galaxy, apart from the
systematic zero-point error from the distance calibration discussed in the next section,
the statistical error in . M includes terms arising from the error in the integrated color
9 Surface Brightness Fluctuations 165

of the galaxy (needed to estimate M from the calibration relation) and from the
intrinsic, irreducible scatter in the calibration equation. The latter effect results from
stellar population variations and varies with the bandpass.
Overall, SBF measurement errors for well designed observations can achieve
statistical uncertainties of .∼ 0.05 mag in m. For a red galaxy observed in a filter
close to 1 .μm, the intrinsic scatter in M is .∼ 0.06 mag. Hence, when combined,
these errors result in a purely statistical uncertainty that can be as low as 0.08 mag,
or 4% on the distance. Nonetheless, fainter, blue galaxies and sub-optimal observing
conditions will have larger statistical uncertainties.

9.4.2 Systematic Uncertainties

The main systematic uncertainty in SBF distances is in the M zero point4 . Empirical
calibrations of M are typically based on the zero point from the SBF ground-based
survey by Tonry and colleagues, which obtained SBF for bulges of spiral galaxies
with distance estimates from Cepheids [54]. A recent reassessment of this calibration
reports a total systematic uncertainty of 0.09 mag in the Cepheid-based SBF distance
zero point [8]. This includes contributions from the tie between near-IR and optical
SBF distances, the tie between SBF and Cepheids, and the current uncertainty in
the Cepheid zero point, taking into account the improved distance to the Large
Magellanic Cloud from detached eclipsing binaries [42].
The largest source of systematic uncertainty in this calibration is the tie between
SBF and Cepheids. This can be improved to a limited degree with further refinements
in Cepheid luminosities with Gaia parallaxes (e.g., [19]) and by using more modern
Cepheid distances, including a recently revised distance to M31 (one of the SBF cal-
ibrators) with a precision better than 2% [35]. New space-based SBF measurements
for all the calibrator galaxies would also help. However, it is important to note that
the ideal galaxies for measuring SBF (red ellipticals, Sect. 9.3.3), will never have
Cepheid distances because they have no recent star formation.
An alternative to calibrating SBF from Cepheids is to use the tip of the red giant
branch (TRGB) method. The TRGB method is ideal for measuring distances to early-
type galaxies and can be calibrated using geometric distances from Gaia (e.g., [21,
34]). Unlike the case for Cepheids, the stellar populations underlying both the SBF
and TRGB methods are the same, namely old low-mass stars. Given this, TRGB
distances are the natural choice for calibrating SBF. In an early effort to calibrate
SBF using TRGB, [40] examined a set of 16 galaxies within 10 Mpc. These were
predominantly blue, low-mass early-type galaxies, and thus the scatter was large,
although the zero point was consistent with the Cepheid calibration.

4 As future facilities become capable of reaching distances well beyond 100 Mpc, SBF.k-corrections
that take the galaxy spectrum redshift properly into account may be necessary to avoid introducing
further systematic effects [29]. As of now, the 100 Mpc limit is usually not surpassed, so the
correction does not pose any problems when addressing systematic uncertainties.
166 M. Cantiello and J. P. Blakeslee

A high-quality TRGB-based SBF calibration must be based on massive red early-

type galaxies. Recently, [8] compared SBF and TRGB distances to the few such
galaxies for which both types of measurment exist, including the bright ellipticals
M 60 and M 87 in Virgo, and the dusty merger remnant NGC 1316 in Fornax. The
authors found that the mean offset between the TRGB and Cepheid-calibrated SBF
distances of the two galaxies in Virgo was only .−0.01 mag. NGC 1316, a far less
ideal galaxy, had a larger offset of 0.14 mag. Although preliminary, this result shows
no significant difference between Cepheid- and TRGB-based SBF calibrations.
With the launch of JWST, the prospects have never been better for improving the
SBF zero-point calibration using only modest amounts of observing time. A recently
approved Cycle 2 program will measure TRGB distances and SBF magnitudes for
14 luminous, red early-type galaxies reaching out to 20 Mpc. The sample is designed
to produce an absolute calibration for the SBF method of better than 2%. Further
reduction in the zero-point uncertainty can be achieved by combining the TRGB-
and Cepheid-based calibrations.
In summary, the current systematic uncertainty on SBF distances using the
Cepheid-based calibration is 0.09 mag or about 4.2%. This value could be improved
by obtaining new SBF measurements for the bulges of spiral galaxies that host
Cepheids, but this approach will never be ideal because the best targets for the SBF
method are massive red ellipticals. The preferred route for substantially reducing the
systematic uncertainty in M is via the the TRGB method, which relies on the same
underlying stellar population as the SBF method. A new JWST program, measuring
both the TRGB and SBF for a well-selected set of early-type calibrators, should
soon decrease the systematic uncertainty in SBF distances to below 2%. Further
improvements can be achieved with additional TRGB calibrators and/or combining
the TRGB and Cepheid calibrations.

9.5 Hubble-Lemaitre Constant

9.5.1 Current Constraints from SBF

Since its introduction for measuring extragalactic distances, the SBF method has
been applied to about 400 galaxies. The random uncertainty associated with these
distances varies greatly, from .∼ 4% for space-based observations of massive, red
galaxies, to .∼ 20% for dwarf galaxies observed from the ground. The distances
obtained have been used for a variety of scientific cases. For example, SBF has been
used to measure distances to:

• NGC 4993, the lenticular galaxy host of the gravitational wave event GW 170817
which had the first observed electromagnetic counterpart [15];
• M 87, the cD host galaxy of the first supermassive black hole to have its “shadow”
imaged using the EHT [23];
9 Surface Brightness Fluctuations 167

• hundreds of other member galaxies in the Virgo and Fornax clusters to explore the
substructure in these environments [9, 13, 22, 37];
• numerous ultra-diffuse galaxies [20], including NGC 1052-DF2, the galaxy pro-
posed to be devoid of dark matter [4, 55].
In addition, as the method reaches into the Hubble flow, many authors have used this
technique for the estimation of the Hubble constant.
The first study to estimate. H0 directly from SBF distances decisively in the Hubble
flow used a sample of 15 galaxies from 40 to 130 Mpc observed with NICMOS on
HST and obtained . H0 ≈ 76 km s−1 Mpc−1 (dropping to .∼ 72 when restricted to the
six most distant targets).5 Another early direct measurement of H0 from SBF used
archival HST/ACS observations combined with a theoretical calibration based on
population models and likewise found . H0 = 76 km s−1 Mpc−1 [2].
Among recent results, we mention three studies that are closely related to each
other. The first study, by [32] (hereafter K21), presented a re-calibration of the
peak magnitude for a sample of 24 local Type Ia supernovae (SNe Ia) using SBF
distances and a hierarchical Bayesian approach. The authors then extended the
calibration to a sample of 96 SNeIa at redshifts .0.02 < z < 0.08 and obtained
−1
. H0 = 71.2 ± 2.4 (stat) ± 3.4 (sys) km s Mpc−1 , where we have updated the value
to use the improved LMC distance from [42] for consistency with the other recent . H0
values discussed here. K21 used a diverse sample of SBF distances, which were col-
lected from observations over a span of twenty years, along with an equally diverse
sample of SNeIa. While normalizing the SBF distance sample to a common reference
calibration was difficult, the heterogeneity of the sample provided some benefits, as
it should in principle be less prone to systematic effects, such as those related to
specific instruments or data analysis procedures. However, this also resulted in large
final statistical uncertainties. The value of . H0 reported by K21 was roughly midway
between those found directly from Cepheid-calibrated SNe Ia [46] and predicted
from analysis of the cosmic microwave background (CMB) [44].
In another recent study, Blakeslee et al. [8] (hereafter B21) used the catalog of SBF
distances from Jensen et al. [29] to estimate . H0 . This sample comprises 63 bright
early-type galaxies observed with the WFC3/IR instrument on the HST, reaching
distances up to 100 Mpc. The resulting value of . H0 = 73.3 ± 0.7 (stat) ± 2.4 (sys)
km s−1 Mpc−1 agrees well with most other direct measurements in the local universe,
but deviates significantly (by about 2.5.σ , including the systematic uncertainty) from
the CMB prediction, assuming the standard .ΛCDM model. It is noteworthy that the
random uncertainty is about one third of that reported in K21.
To ensure the reliability of their findings, B21 conducted an analysis of . H0 involv-
ing four different treatments of galaxy velocities, including group-averaged CMB
frame velocities, individual CMB velocities, and flow-corrected velocities from two
different models. Additionally, their final . H0 combines the mutually consistent SBF
calibrations from Cepheids and the TRGB in order to reduce the final systematic

5These numbers should be updated to the most recent Cepheids calibration, which has a minor
impact on the final . H0 value.
168 M. Cantiello and J. P. Blakeslee

error down to the .∼ 3.3% level. As noted previously, the number of TRGB calibra-
tors in this analysis was very small, and future studies incorporating a larger sample of
giant ellipticals with TRGB distances measured by JWST and tied to Gaia parallaxes
should vastly reduce this systematic error.
The third recent work [24] related to . H0 combines an approach like that of K21
with the WFC3/IR SBF distances from [29]. The authors of [24] used SBF dis-
tances for 25 host galaxies of SNe Ia to provide an absolute calibration for the Pan-
theon+ compilation [49] of 1500 SNe Ia extending far out into the Hubble flow. Using
the original Pantheon+ “SALT2” parameters for the dependence of peak luminos-
ity on decline rate and color, these authors find . H0 = 74.6 ± 0.9 (stat) ± 2.7 (syst)
km s−1 Mpc−1 , which agrees to within .∼ 1σ with the value obtained by B21 (the
systematic error is common between the two studies, so the comparison must be
done with respect to the statistical error only).
However, it is important to note that the SNe Ia whose luminosities can be cali-
brated with SBF distances are generally fast decliners in massive host galaxies. When
[24] limit their analysis to fast-declining SNe Ia in massive hosts and rederive the
SALT2 parameters, the SBF-calibrated. H0 from these SNe becomes.73.3 ± 1.0 ± 2.7
km s−1 Mpc−1 , which is identical to the value derived by B21 from SBF distances
in the Hubble flow. This .∼ 1.σ change in . H0 suggests a slight difference between
the best-fit SALT2 parameters for the relatively small set of fast-declining SNe Ia in
massive galaxies as compared to the full Pantheon+ sample.

9.5.2 Forecasts for New and Future Facilities

As outlined in the previous section, recent SBF-based estimates of . H0 align more

closely with the results found by Cepheid-calibrated SNe Ia than with the predictions
from the CMB. The potential use of SBF, particularly with JWST and upcoming
astronomical facilities, offers a promising route to derive a more precise and accurate
value of . H0 that is fully independent of SNe Ia and Cepheids. Here, we present the
expected results of the SBF method for estimating . H0 using various existing and
forthcoming facilities: JWST, Rubin Observatory, ESA’s Euclid mission, the Roman
Space Telescope, and the Extremely Large Telescopes (ELTs).
To evaluate the potential for constraining. H0 from SBF with each new or forthcom-
ing facility, we proceed as follows. First, we estimate the maximum distance.d Max that
can be reached with the specific facility, as explained below. Then, we count the num-
ber of massive elliptical galaxies that can be observed by the telescope/instrument
within a distance range of .dmin ≤ D ≤ d Max . Here, we assume .dmin = 40 Mpc to
exclude galaxies in the local universe where flow motions and peculiar velocities
are large compared to distance errors. To estimate .d Max , we assume (as is generally
the case) that contamination by GCs is the limiting factor for well-designed SBF
measurements. Thus, .d Max is the distance to which the GCs can be detected and
removed to a faint enough limit, and the residual contamination reliably estimated,
so that the error in the correction drops below the intrinsic scatter in the method. For
9 Surface Brightness Fluctuations 169

instance, in the . I band, where the peak of the GC luminosity function (GCLF) is
at . M I ≈ −8.0 AB mag, accurate SBF measurements require detecting and remov-
ing sources to +0.5 mag fainter than the GCLF peak, or . M I ≈ −7.5 AB mag. In
the . H -band (which has slightly larger intrinsic scatter), the same relative amount of
contamination from GCs can be reached for a detection limit about 1.5 mag brighter
than the GCLF peak; [28]. We estimate the near-IR GCLF peak magnitudes from
[41] and convert to the AB system.
The expected depth is determined based on available information for the specific
facilities. For instance, in the case of Rubin Observatory’s LSST, we use the expected
10-year .5σ point-source depth for the .i band, with a GCLF peak of .∼ −8.0 AB mag
and sky coordinates with .Dec(J2000) ≤ 15◦ .
Another important factor is the selection of an appropriate list of targets. We
use the 2MASS Redshift Survey (2MRS), which comprises approximately 43,000
galaxies brighter than . K s = 11.75 mag (Vega) [27]. This survey is nearly complete,
with the exception of regions near the Galactic plane, which account for only 9%
of the sky, and an . L ∗ galaxy is included if it is within .∼135 Mpc. We first focus on
“optimal” targets, i.e., bright galaxies with a morphological .T -type.≤ −4 (indicating
morphologically regular ellipticals) and an absolute . K -band Vega magnitude . M K ≤
−25 mag (similar to M87), where distance is estimated from the 2MRS redshift
assuming . H0 = 75 km s−1 Mpc−1 . Under these assumptions, Table 9.1 (column 7)
reports the numbers of optimal targets reachable by each of the new and upcoming
facilities mentioned above.
Now, if we relax the selection somewhat on the galaxy morphology and luminosity,
accepting .T ≤ −1 and . M K ≤ −22.5 mag as “good” candidates, then the numbers
of SBF targets increase dramatically, as again shown in column 7 of the table. It
is important to note that these numbers represent conservative lower limits. For
reference, the NGVS survey on the CFHT, with a median .i-band seeing of .0.'' 7 and
5.σ depth three magnitudes fainter than the GCLF peak, enabled SBF measurements
for about 300 galaxies in the Virgo cluster [13] down to . M K ∼ −16 Vega mag. It was
also possible to measure SBF in many galaxies with positive T-type, such as bulges
of spirals and galaxies with some dust contamination, thanks to the availability of
.u-band data. Thus, the numbers of candidates measurable with the facilities specified
in Table 9.1 could increase by a factor of a few, or even orders of magnitude, when
fainter and more irregular targets are considered.
For estimating the constraints on. H0 , we also need to adopt values for the expected
distance errors. Thus, for ground-based observations in natural seeing (Rubin/LSST),
we conservatively assume a total error in the SBF distance of 8% for the optimal
target galaxies, and 16% for the “good” targets. For high-resolution space-based and
AO-assisted observations, we assume total errors of 4% and 8% for the optimal and
“good” targets, respectively.
All the necessary parameters required for estimating the number of reachable
targets with the selected facilities are listed in Table 9.1, along with the estimated
maximum distances and sample sizes for each facility in selected bands. Taking Rubin
as an example, it is expected that approximately 50 optimal targets within .d Max ≈ 70
Mpc will be measurable in the Baseline survey area, resulting in a constraint on . H0
170 M. Cantiello and J. P. Blakeslee

of .∼ 1.2% (stat). For “good” targets, the estimated number increases to 1500, and
the statistical error drops well below the systematic uncertainties from the distance
calibration and possible large-scale flow motions.
The Roman Space Telescope is especially promising for future constraints on . H0
using SBF because of its combination of HST-like resolution, excellent PSF stability,
near-IR coverage, and planned large-area surveys [5]. While awaiting the Roman
telescope, however, the JWST has unsurpassed capabilities for targeting galaxies
one at a time out to .∼ 300 Mpc, an unprecedented distance for SBF that exceeds
the limit of claimed “Hubble bubbles” (e.g., [31, 48]). In the NIRCam passbands
reported in Table 9.1, we find that around a thousand candidates from the 2MRS
catalog will be optimal targets for JWST. Of course, only a few dozen are needed
to bring the statistical error comfortably below 1%, so there will be no shortage of
targets from which to choose.
To demonstrate the feasibility of SBF measurements with JWST/NIRCam, we
conducted a preliminary analysis on NGC 7317, the elliptical galaxy in Stephan’s
Quintet at a distance of about 100 Mpc, using the public Early Release Data.
The characteristic SBF signal is clearly observed across all passbands, including
mid-infrared bands that have not previously been used in SBF studies. The left
panels of Fig. 9.2 display the measured power spectra in the F150W band of the

Fig. 9.2 Power spectrum analyses of JWST data on NGC 7317, an elliptical galaxy at .∼ 100 Mpc,
in the F150W (top row) and F277W (bottom row) passbands. Left panels: power spectra of the
central masked region after galaxy subtraction (blue solid curve), the scaled PSF (red curve), and
an empty region away from the galaxy (orange curve); dotted blue lines show the range of the
starting waveneumbers .k for the fits. Middle panels: ratio of the power spectra on- and off-galaxy;
the structure in this ratio for F150W may indicate a variable PSF in the combined image. Right:
Uncorrected power spectrum normalization . P0 derived from the fits in the left panels as a function
of the starting .k value (the fitted . P0 values here are insensitive to the starting .k)
9 Surface Brightness Fluctuations 171

short-wavelength channel and the F277W band of the long-wavelength channel. The
other panels show the ratio of the fluctuation amplitudes on- and off- the galaxy center
and the . P0 coefficient of the PSF term in the power spectrum, using a star within the
image as the reference PSF. Analysis of the SBF signal for this target is in progress,
and further work is needed to improve the agreement with the PSF templates, but the
signal is very strong, given the short exposure times for this distance.
With the forthcoming generation of wide-area imaging surveys such as
Rubin/LSST, Euclid, and the Roman Community Surveys providing hundreds or
even thousands of galaxies suitable for SBF measurements, it is essential that the
SBF analysis should be carried out in a robust, streamlined, and automatic way.
Historically, the limited number of available SBF measurement pipelines have been
poorly automated and required extensive user intervention. This is especially true for
the galaxy fitting and residual masking, as these steps must be optimized to ensure
clean power spectra. However, it is clear that the method needs to be adapted to run
in “production mode” to deliver robust results with minimal human intervention.
Multiple parallel efforts are ongoing in this regard.
The prospects for SBF measurements with AO-assisted observations on 30–40m
telescopes are also exciting. Table 9.1 reports the expectations for E-ELT/MICADO
with the MORFEO AO module. With their extremely large apertures and diffraction-
limited resolution .< 0.'' 02, such facilities should be able to measure SBF distances
out to redshift .z ∼ 0.1. However, measuring SBF with AO-assisted instruments may

Table 9.1 Maximum distances and numbers of candidates for SBF analysis
Telescope Band (2) Depth.(a) GCLF GCLF .dMax (6) Optimal/good .δ H0 /H0
(constraint) (mag) peak.(b) (4) offset.(c) (5) candidates (7) (%) opti-
(1) mal/good
(8)
Rubin .i 26.8 .−8.5 .+0.5 70 50/1500 1.2/.<0.5
(Dec.≤ 15◦ )
Euclid . H Euclid 24.0 .−8.6 .−1.5 55 20/600 1.0/.<0.5
(.|b| ≥ 23◦ )
Roman.(e) . J F129 28.0 .−8.3 .−1.0 230 1500/.>5000 0.7/0.5.(g)
Roman.(e) . H F158 28.0 .−8.6 .−1.5 300 1600/.>5000 0.7/0.5.(g)
JWST.(e) . J F115W 27.5 .−8.3 .−1.0 180 900/.>5000 0.9/0.5.(g)
JWST.(e) . H F150W 27.7 .−8.6 .−1.5 290 1600/.>5000 0.7/0.5.(g)
JWST.(e) . K F200W 27.8 .−8.1 .−2.0 310 1600/.>5000 0.7/0.5.(g)
ELT.( f ) .K 28.5 .−8.1 .−2.0 400 750/.>4000 1.0/0.7.(g)
(Dec.≤ 0◦ )
(a) Nominal .5σ point-source depth accounting for host galaxy background
(b) Peak, or turnover, magnitude (AB) of the GCLF in each band
(c) Offset of the required detection limit with respect to the GCLF peak (see text)
(d) Estimated number of optimal/suitable SBF targets based on 2MRS [27]
(e) .5σ depth in one hour integration time
(f) Depth derived from the preliminary ELT/MICADO exposure time calculator, assuming a. S/N =
5 in one hour in multi-conjugate adaptive optics mode
(g) Assuming that a fraction of .∼ 2% of possible candidates are indeed observed
172 M. Cantiello and J. P. Blakeslee

be challenging because of the spatially and temporally varying PSF, significant over-
heads, and the requirement for accurate .k-corrections at these distances [29]. Never-
theless, once the technique is optimized for 30m-class telescopes, a carefully selected
set of targets could have a significant impact, potentially providing an independent
direct detection of cosmic acceleration.

9.6 Conclusions

Over the last 30 years, the SBF method has become one of the most powerful meth-
ods for measuring galaxy distances from the Local Group to the Hubble flow and
constraining the value of . H0 . Unlike other precision distance indicators (Cepheids,
TRGB, SN Ia, masers), SBF measurements require only a modest investment of tele-
scope time, especially when using high-resolution space-based data. SBF distances
to large numbers of galaxies observed with HST using WFC3/IR have already con-
strained . H0 to 3.4%, where the main source of uncertainty is the zero-point calibra-
tion. Now, with JWST/NIRCam, it becomes possible to directly calibrate SBF using
TRGB distances to the same galaxies, opening a path to reduce the uncertainty on
. H0 from SBF to below 2%. The large aperture, superior resolution, PSF stability,
and near-infrared capability of JWST also make it possible to measure robust SBF
distances out to .∼ 300 Mpc. This will reduce the uncertainty from possible large-
scale bulk motions and vastly increase the number of optimal targets in range of the
method. Other forthcoming facilities such as Euclid, Roman, Rubin, and the ELTs
also hold great potential for the method. With an improved TRGB-based calibration
and samples of hundreds, or even thousands, of SBF distances reaching out to hun-
dreds of Mpc, a fully independent measurement of . H0 to a precision approaching
1% will be in reach by the end of the current decade.

References

1. W.A. Baum, M. Schwarzschild, A comparison of stellar populations in the Andromeda galaxy

and its elliptical companion. AJ 60, 247 (1955)
2. I. Biscardi, G. Raimondo, M. Cantiello, E. Brocato, Optical surface brightness fluctuations of
shell galaxies toward 100 Mpc. Astrophys. J. 678(1), 168–178 (2008)
3. J.P. Blakeslee, Surface brightness fluctuations as primary and secondary distance indicators.
Ap&SS 341, 179–186 (2012)
4. J.P. Blakeslee, M. Cantiello, Independent analysis of the distance to NGC 1052-DF2. Res.
Notes Am. Astron. Soc. 2(3), 146 (2018)
5. J.P. Blakeslee, M. Cantiello, M.J. Hudson, L. Ferrarese, N. Hazra, J.B. Jensen, E.W. Peng, G.
Raimondo, Gathering galaxy distances in abundance with roman wide-area data (2023). arXiv
e-prints: arXiv:2306.15170
6. J.P. Blakeslee, M. Cantiello, S. Mei, P. Côté, R. Barber DeGraaff, L. Ferrarese, A. Jordán, E.W.
Peng, J.L. Tonry, G. Worthey, Surface brightness fluctuations in the hubble space telescope
9 Surface Brightness Fluctuations 173

ACS/WFC F814W bandpass and an update on galaxy distances. Astrophys. J. 724(1), 657–
668 (2010)
7. J.P. Blakeslee, M. Davis, J.L. Tonry, A. Dressler, E.A. Ajhar, A first comparison of the surface
brightness fluctuation survey distances with the galaxy density field: implications for H.0 and
.Ω. Astrophys. J. 527(2), L73–L76 (1999)
8. J.P. Blakeslee, J.B. Jensen, C.-P. Ma, P.A. Milne, J.E. Greene, The hubble constant from infrared
surface brightness fluctuation distances. Astrophys. J. 911(1), 65 (2021)
9. J.P. Blakeslee, A. Jordán, S. Mei, P. Côté, L. Ferrarese, L. Infante, E.W. Peng, J.L. Tonry, M.J.
West, The ACS fornax cluster survey V measurement and recalibration of surface brightness
fluctuations and a precise value of the Fornax-Virgo relative distance. Astrophys. J. 694, 556–
572 (2009)
10. J.P. Blakeslee, J.L. Tonry, Measurements of globular cluster specific frequencies and luminosity
function widths in coma. Astrophys. J. 442, 579 (1995)
11. J.P. Blakeslee, A. Vazdekis, E.A. Ajhar, Stellar populations and surface brightness fluctuations:
new observations and models. MNRAS 320, 193 (2001)
12. M. Cantiello, I. Biscardi, E. Brocato, G. Raimondo, VLT optical BVR observations of two
bright supernova Ia hosts in the Virgo cluster-surface brightness fluctuation analysis. A&A
532, A154 (2011)
13. M. Cantiello, J.P. Blakeslee, L. Ferrarese, P. Côté, J.C. Roediger, G. Raimondo, E.W. Peng,
S. Gwyn, P.R. Durrell, J.C. Cuillandre, The next generation virgo cluster survey (NGVS) XVIII
measurement and calibration of surface brightness fluctuation distances for bright galaxies in
virgo (and beyond). Astrophys. J. 856(2), 126 (2018)
14. M. Cantiello, J.P. Blakeslee, G. Raimondo, S. Mei, E. Brocato, M. Capaccioli, Detection of
radial surface brightness fluctuations and color gradients in elliptical galaxies with the advanced
camera for surveys. Astrophys. J. 634, 239 (2005)
15. M. Cantiello, J.B. Jensen, J.P. Blakeslee, E. Berger, A.J. Levan, N.R. Tanvir, G. Raimondo, E.
Brocato, K.D. Alexander, P.K. Blanchard, M. Branchesi, Z. Cano, R. Chornock, S. Covino,
P.S. Cowperthwaite, P. D’Avanzo, T. Eftekhari, W. Fong, A.S. Fruchter, A. Grado, J. Hjorth,
D.E. Holz, J.D. Lyman, I. Mandel, R. Margutti, M. Nicholl, V.A. Villar, P.K.G. Williams, A
precise distance to the host galaxy of the binary neutron star merger GW170817 using surface
brightness fluctuations. Astrophys. J. 854(2), L31 (2018)
16. M. Cantiello, G. Raimondo, E. Brocato, M. Capaccioli, New optical and near-infrared surface
brightness fluctuation models: a primary distance indicator ranging from globular clusters to
distant galaxies? Astrophys. J. 125, 2783 (2003)
17. S.G. Carlsten, R.L. Beaton, J.P. Greco, J.E. Greene, Using surface brightness fluctuations to
study nearby satellite galaxy systems: calibration and methodology. Astrophys. J. 879(1), 13
(2019)
18. M. Cerviño, V. Luridiana, L. Jamet, On surface brightness fluctuations: probabilistic and sta-
tistical bases I. Stellar population and theoretical surface brightness fluctuations. A&A 491(3),
693–701 (2008)
19. G. Clementini, V. Ripepi, R. Molinaro, A. Garofalo, T. Muraveva, L. Rimoldini, L.P. Guy,
G.J. de Fombelle, K. Nienartowicz, O. Marchal, M. Audard, B. Holl, S. Leccia, M. Mar-
coni, I. Musella, N. Mowlavi, I. Lecoeur-Taibi, L. Eyer, J. De Ridder, S. Regibo, L.M. Sarro,
L. Szabados, D.W. Evans, M. Riello, Gaia data release 2. specific characterisation and validation
of all-sky Cepheids and RR Lyrae stars. A&A 622, A60 (2019)
20. Y. Cohen, P. van Dokkum, S. Danieli, A.J. Romanowsky, R. Abraham, A. Merritt, J. Zhang,
L. Mowla, J.M.D. Kruijssen, C. Conroy, A. Wasserman, The Dragonfly nearby galaxies survey
V HST/ACS observations of 23 low surface brightness objects in the fields of NGC 1052, NGC
1084, M96, and NGC 4258. Astrophys. J. 868(2), 96 (2018)
21. M. Dixon, J. Mould, C. Flynn, E.N. Taylor, C. Lidman, A.R. Duffy, A geometric calibration of
the tip of the red giant branch in the Milky Way using Gaia DR3. MNRAS 523(2), 2283–2295
(2023)
22. L.P. Dunn, H. Jerjen, First results from SAPAC: toward a three-dimensional picture of the
Fornax cluster core. AJ 132(3), 1384–1395 (2006)
174 M. Cantiello and J. P. Blakeslee

23. Event Horizon Telescope Collaboration, K. Akiyama, A. Alberdi, W. Alef, K. Asada, R. Azulay,
A.K. Baczko, D. Ball, M. Baloković, J. Barrett, D. Bintley, L. Blackburn, W. Boland, K.L.
Bouman, G.C. Bower et al., First M87 event horizon telescope results VI the shadow and mass
of the central black hole. Astrophys. J. 875(1), L6 (2019)
24. P. Garnavich, C.M. Wood, P. Milne, J.B. Jensen, J.P. Blakeslee, P.J. Brown, D. Scolnic, B. Rose,
D. Brout, Connecting infrared surface brightness fluctuation distances to type Ia supernova
hosts: testing the top rung of the distance ladder. Astrophys. J. 953(1), 35(2023). https://doi.
org/10.3847/1538-4357/ace04b
25. J.P. Greco, P. van Dokkum, S. Danieli, S.G. Carlsten, C. Conroy, Measuring distances to low-
luminosity galaxies using surface brightness fluctuations. Astrophys. J. 908(1), 24 (2021)
26. W.E. Harris, Globular cluster systems, in Saas-Fee Advanced Course 28: Star Clusters (2001)
27. J.P. Huchra, L.M. Macri, K.L. Masters, T.H. Jarrett, P. Berlind, M. Calkins, A.C. Crook, R.
Cutri, P. Erdoǧdu, E. Falco, T. George, C.M. Hutcheson, O. Lahav, J. Mader, J.D. Mink, N.
Martimbeau, S. Schneider, M. Skrutskie, S. Tokarz, M. Westover, The 2MASS redshift survey-
description and data release. Astrophys. J. Supp. Ser. 199(2), 26 (2012)
28. J.B. Jensen, J.P. Blakeslee, Z. Gibson, H.-C. Lee, M. Cantiello, G. Raimondo, N. Boyer, H.
Cho, Measuring infrared surface brightness fluctuation distances with HST WFC3: calibration
and advice. Astrophys. J. 808, 91 (2015)
29. J.B. Jensen, J.P. Blakeslee, C.-P. Ma, P.A. Milne, P.J. Brown, M. Cantiello, P.M. Garnavich, J.E.
Greene, J.R. Lucey, A. Phan, R.B. Tully, C.M. Wood, Infrared surface brightness fluctuation
distances for MASSIVE and type Ia supernova host galaxies. Astrophys. J. Supp. Ser. 255(2),
21 (2021)
30. J.B. Jensen, J.L. Tonry, G.A. Luppino, Measuring distances using infrared surface brightness
fluctuations. Astrophys. J. 505(1), 111–128 (1998)
31. R.C. Keenan, A.J. Barger, L.L. Cowie, Evidence for a .∼ 300 Megaparsec scale under-density
in the local galaxy distribution. Astrophys. J. 775(1), 62 (2013)
32. N. Khetan, L. Izzo, M. Branchesi, R. Wojtak, M. Cantiello, C. Murugeshan, A. Agnello,
E. Cappellaro, M. Della Valle, C. Gall, J. Hjorth, S. Benetti, E. Brocato, J. Burke, D. Hiramatsu,
D.A. Howell, L. Tomasella, S. Valenti, A new measurement of the Hubble constant using Type
Ia supernovae calibrated with surface brightness fluctuations. A&A 647, A72 (2021)
33. Y.J. Kim, M.G. Lee, Calibration of surface brightness fluctuations for Dwarf galaxies in the
hyper suprime-cam gi filter system. Astrophys. J. 923(2), 152 (2021)
34. S. Li, S. Casertano, A.G. Riess, A Gaia data release 3 view on the tip of the red giant branch
luminosity. Astrophys. J. 950(2), 83 (2023)
35. S. Li, A.G. Riess, M.P. Busch, S. Casertano, L.M. Macri, W. Yuan, A sub-2% distance to
M31 from photometrically homogeneous near-infrared Cepheid period-luminosity relations
measured with the Hubble space telescope. Astrophys. J. 920(2), 84 (2021)
36. G.A. Luppino, J.L. Tonry, Infrared surface brightness fluctuations: K-band observations of
M31, M32, and the distance to Maffei 1. Astrophys. J. 410, 81 (1993)
37. S. Mei, J.P. Blakeslee, P. Côté, J.L. Tonry, M.J. West, L. Ferrarese, A. Jordán, E.W. Peng,
A. Anthony, D. Merritt, The ACS Virgo cluster survey XIII. SBF distance catalog and the
three-dimensional structure of the Virgo cluster. Astrophys. J. 655(1), 144–162 (2007)
38. S. Mei, J.P. Blakeslee, J.L. Tonry, A. Jordán, E.W. Peng, P. Côté, L. Ferrarese, D. Merritt,
M. Milosavljević, M.J. West, The ACS Virgo cluster survey IV. Data reduction procedures for
surface brightness fluctuation measurements with the advanced camera for surveys. Astrophys.
J. Supp. Ser. 156, 113–125 (2005)
39. M. Moresco, L. Amati, L. Amendola, S. Birrer, J.P. Blakeslee, M. Cantiello, A. Cimatti, J. Dar-
ling, M. Della Valle, M. Fishbach, C. Grillo, N. Hamaus, D. Holz, L. Izzo, R. Jimenez, E. Lusso,
M. Meneghetti, E. Piedipalumbo, A. Pisani, A. Pourtsidou, L. Pozzetti, M. Quartin, G. Risaliti,
P. Rosati, L. Verde, Unveiling the universe with emerging cosmological probes. Living Rev.
Relativ. 25(1), 6 (2022)
40. J. Mould, S. Sakai, The extragalactic distance scale without Cepheids II surface brightness
fluctuations. Astrophys. J. 694(2), 1331–1334 (2009)
9 Surface Brightness Fluctuations 175

41. J.B. Nantais, J.P. Huchra, P. Barmby, K.A.G. Olsen, T.H. Jarrett, Nearby spiral globular cluster
systems I: luminosity functions. AJ 131(3), 1416–1425 (2006)
42. G. Pietrzyński, D. Graczyk, A. Gallenne, W. Gieren, I.B. Thompson, B. Pilecki, P. Karcz-
marek, M. Górski, K. Suchomska, M. Taormina, B. Zgirski, P. Wielgórski, Z. Kołaczkowski,
P. Konorski, S. Villanova, N. Nardetto, P. Kervella, F. Bresolin, R.P. Kudritzki, J. Storm, R.
Smolec, W. Narloch, A distance to the large Magellanic cloud that is precise to one per cent.
Nature 567(7747), 200–203 (2019)
43. G. Piotto, I.R. King, S.G. Djorgovski, C. Sosin, M. Zoccali, I. Saviane, F. De Angeli, M. Riello,
A. Recio-Blanco, R.M. Rich, G. Meylan, A. Renzini, HST color-magnitude diagrams of 74
galactic globular clusters in the HST F439W and F555W bands. A&A 391, 945–965 (2002)
44. Planck Collaboration, N. Aghanim, Y. Akrami, M. Ashdown, J. Aumont, C. Baccigalupi,
M. Ballardini, A.J. Banday, R.B. Barreiro, N. Bartolo, S. Basak, R. Battye, K. Benabed, J.P.
Bernard, M. Bersanelli, P. Bielewicz, J.J. Bock, J.R. Bond et al., Planck 2018 results VI cos-
mological parameters. A&A 641, A6 (2020)
45. G. Raimondo, Joint analysis of near-infrared properties and surface brightness fluctuations of
large Magellanic cloud star clusters. Astrophys. J. 700(2), 1247–1261 (2009)
46. A.G. Riess, S. Casertano, W. Yuan, L.M. Macri, D. Scolnic, Large magellanic cloud Cepheid
standards provide a 1% foundation for the determination of the Hubble constant and stronger
evidence for physics beyond .ΛCDM. Astrophys. J. 876(1), 85 (2019)
47. P. Rodríguez-Beltrán, A. Vazdekis, M. Cerviño, M.A. Beasley, Surface brightness fluctua-
tions to constrain secondary stellar populations: revealing very low-metallicity stars in massive
galaxies. MNRAS 507(2), 3005–3029 (2021)
48. A.E. Romano, Hubble trouble or Hubble bubble? Int. J. Mod. Phys. D 27(9), 1850102 (2018)
49. D. Scolnic, D. Brout, A. Carr, A.G. Riess, T.M. Davis, A. Dwomoh, D.O. Jones, N. Ali, P.
Charvu, R. Chen, E.R. Peterson, B. Popovic, B.M. Rose, C.M. Wood, P.J. Brown, K. Cham-
bers, D.A. Coulter, K.G. Dettman, G. Dimitriadis, A.V. Filippenko, R.J. Foley, S.W. Jha, C.D.
Kilpatrick, R.P. Kirshner, Y.C. Pan, A. Rest, C. Rojas-Bravo, M.R. Siebert, B.E. Stahl, W.K.
Zheng, The Pantheon + analysis: the full data set and light-curve release. Astrophys. J. 938(2),
113 (2022)
50. J. Tonry, D.P. Schneider, A new technique for measuring extragalactic distances. Astrophys. J.
96, 807–815 (1988)
51. J.L. Tonry, E.A. Ajhar, G.A. Luppino, Observations of surface-brightness fluctuations in Virgo.
Astrophys. J. 100, 1416 (1990)
52. J.L. Tonry, J.P. Blakeslee, E.A. Ajhar, A. Dressler, The SBF survey of galaxy distances I sample
selection, photometric calibration, and the hubble constant. Astrophys. J. 475, 399 (1997)
53. J.L. Tonry, J.P. Blakeslee, E.A. Ajhar, A. Dressler, The surface brightness fluctuation survey
of galaxy distances II local and large-scale flows. Astrophys. J. 530(2), 625–651 (2000)
54. J.L. Tonry, A. Dressler, J.P. Blakeslee, E.A. Ajhar, A.B. Fletcher, G.A. Luppino, M.R. Metzger,
C.B. Moore, The SBF survey of galaxy distances IV SBF magnitudes, colors, and distances.
Astrophys. J. 546, 681 (2001)
55. P. van Dokkum, S. Danieli, Y. Cohen, A.J. Romanowsky, C. Conroy, The distance of the dark
matter deficient galaxy NGC 1052-DF2. Astrophys. J. 864(1), L18 (2018)
56. G. Worthey, The dependence of the brightness fluctuation distance indicator on stellar popula-
tion age and metallicity. Astrophys. J. 409, 530–536 (1993)
57. G. Worthey, Comprehensive stellar population models and the disentanglement of age and
metallicity effects. Astrophys. J. Supp. Ser. 95, 107 (1994)
Chapter 10
The Pursuit of the Hubble Constant
Using Type II Supernovae

Thomas de Jaeger and Lluís Galbany

10.1 Introduction

Thermonuclear explosions of a carbon-oxygen white dwarf in multi-star systems,

known as Type Ia supernovae (SNe Ia), are among the best distance indicators [30,
45, 49]. For more than three decades, they have been used to measure extragalactic
distances with an accuracy of 5–10% [7, 9] and have been fundamental to identify
a tension between local [50] and distant [46] measurements of the Hubble constant
(. H0 ). There are two complementary ways to solve this tension: first, increasing
the level of precision of the different methods, and second, developing as many
independent methods as possible with their own systematic uncertainties. In this
chapter, as an independent approach, we describe the use of Type II supernovae
(SNe II) to estimate . H0 .
After being overshadowed by the better standardized SNe Ia, recent studies have
demonstrated that SNe II can be used to measure extragalactic distances and have
a role to play in the “. H0 tension.” Observationally, their optical spectra are char-
acterized by the presence of strong Balmer lines exhibiting broad P-Cygni profiles
(e.g., [24, 26]), and their light curves display a plateau of varying steepness and
length ([1, 3, 16, 27]) owing to hydrogen recombination. Among all the supernova
types, SN II is the one for which the physics and the nature of their progenitors
are the most well-understood. Unlike SNe Ia, for which no clear progenitors have

T. de Jaeger (B)
LPNHE, CNRS/IN2P3 & Sorbonne Université, 4 place Jussieu, 75005 Paris, France
e-mail: thomas.dejaeger@lpnhe.in2p3.fr
L. Galbany
Institute of Space Sciences (ICE-CSIC), Campus UAB, Can Magrans S/N, 08193 Barcelona,
Spain
e-mail: lg@csic.es
Institut d’Estudis Espacials de Catalunya (IEEC), 08034 Barcelona, Spain

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 177
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_10
178 T. de Jaeger and L. Galbany

been yet identified, SNe II have dozens of direct progenitor detections [54]. It is now
accepted that their progenitors have retained a significant fraction of their Hydrogen
envelopes and arise from the explosion of red supergiants in late-type galaxies via
core collapse. The understanding of the explosion mechanisms of SN II itself has
also made remarkable progress in the past few decades [36, 60].
Although SNe II are not perfect standard candles and display a large range of
peak luminosities ([1]), they have the potential to be excellent distance indicators
with different systematics than SNe Ia. Their extrinsic differences can be calibrated
using theoretical or empirical distance measurement methods such as the expand-
ing photosphere method (EPM; [40]), the spectral expanding atmosphere method
(SEAM; [6, 21]), the tailored EPM ([58]), the standard-candle method (SCM; [33]),
the photospheric-magnitude method (PMM; [51, 52]), and the only method based
solely on photometry, the photometric-color method (PCM; [12, 13, 17]).
Among this wide variety of methods to measure distances from SN II, here, we
review only those used to measure . H0 . For each method, we will describe the under-
lying theory behind them and present the most updated Hubble diagram together
with a nearly exhaustive list of their . H0 measurements.

10.2 EPM: Expanding Photosphere Method

The Expanding Photosphere Method is a geometric-based method based on the

approach used by [2] to measure distances for pulsating stars. The distance (. D) is
directly obtained by measuring the physical photosphere radius (. R phot ) and the angu-
lar size (.θ ) of the SN II. Unfortunately, extragalactic SNe are unresolved and those
parameters cannot be measured directly, they need to be derived by making some
assumptions. First, supported by the fact that SNe II show a low polarization degree
during the plateau phase ([41]), we assume a spherically symmetric photosphere.
Second, the ejecta is freely expanding, so we can determine the photosphere radius
from its velocity (.v) and time since explosion (.t − t0 ): . R phot (t) = v(t − t0 ) + R0 .
Due to a fast expansion, it is safe to assume that the initial radius (. R0 at.t = t0 ) is neg-
ligible after a few days. The velocities of the photosphere are determined through the
minimum flux of the absorption component of the P-Cygni line profile of weak lines
such as those of Fe II (.λ5018,5169). Finally, to derive the theoretical angular size .θ ,
we assume that the photosphere radiates as a blackbody (BB), i.e., the thermalization
layer from which the thermal photons originated (R.ther m ) and the photosphere have
−0.4 Aλ
the same radius. Therefore, we have .4π D 2 f ν = 4π Rther 2
m π Bν (Tc )10 , where
B.ν (Tc ) is the Planck function at color temperature .Tc of the BB radiation estimated
by fitting Planck curves to observed broadband magnitudes, A.λ the total dust extinc-
tion (host galaxy and Milky Way), and . f ν is the flux received at Earth. Finally, for
. z << 1, .θ can be written as,
10 The Pursuit of the Hubble Constant Using Type II Supernovae 179

Fig. 10.1 Left Top: Figure updated from figure of [29] representing the Type II supernova Hubble
diagram using the EPM. The EPM was applied to the data taken from [23] (blue triangles; E96),
[37] (red squares; E96), [8] (green diamonds; B14), and [29] (cyan dots; G18). Left Bottom: Type
II supernova Hubble diagram of six objects using the distance luminosities derived by [58] (brown
dots; V20) when applying the tailored EPM. Right Top and Right bottom show the probability
distributions of . H0 for the individual SNe in grey and the combined estimate in black. We include
the value from local measurement [50] (pink filled region) and from the high-redshift results [46]
(orange filled region) for comparison
/
R phot BB Rther m fν
.θ = = = . (10.1)
D D π Bν (Tc )10−0.4 Aλ

Combining . R(t) = v(t − t0 ) and the above equation implies that .t = D(θ/v) + t0 ,
i.e., the distance . D corresponds to the slope of a straight line made in the .t versus
(.θ/v) plot and the explosion date to the y-intercept.
The EPM was first used by [40] to calculate distances of two SNe II (69L and
70g). They also obtained the first . H0 from SN II and found a value of . H0 = 65 ±
15 km s.−1 Mpc.−1 . However, a few years later, [59] noticed that the photosphere
radiation deviates from that of a BB. The reason is that the thermalization layer from
which the BB radiation is generated is deeper than the photosphere. The result is
that the photosphere radiates less strongly than a BB of the same color temperature
and therefore, appears diluted. To take into account this effect, the EPM was refined
by incorporating a dilution factor (.ξ ) in Eq. 10.4 which parameterized the relative
position of the two surfaces (. R phot = Rther m /ξ ). In principle, the dilution factor
should account for all deviations of the spectrum from BB emission, including the
chemical composition or density structure of the progenitor star. However, theoretical
dilution factors computed from non-LTE models with a wide range of properties have
shown that they depend only on the color temperature ([20, 23]).
180 T. de Jaeger and L. Galbany

[53] were the first to apply this refined method to a large sample of SN II and
derive a. H0 . Using 16 objects and preliminary values of the dilution factors computed
by [23], they obtained a value of 73 .± 6 (stat) ± 7 (sys) km s.−1 Mpc.−1 . Then, using 9
well-observed SNe II, with a slightly modified version of [23] dilution factors, and dif-
ferent filter combinations, [32] obtained an average value . H0 =71 .± 9 (stat) km s.−1
Mpc.−1 . Based on [32] work, [42] derived a value of . H0 =57 .± 7 (stat)± 13 (sys)
km s.−1 Mpc.−1 . Two other works have used the EPM to measure the Universe expan-
sion rate. Jones et al. [37] constructed a Hubble diagram with 12 well-observed
SNe II using dilutions factors from [20, 23]. They obtained a value of . H0 =100.5
−1
.± 8.4 (stat) km s. Mpc.−1 and . H0 =56.9 .± 4.2 (stat) km s.−1 Mpc.−1 , respectively.
The dilution factor choice generates a large systematic difference in the . H0 values.
The factors calculated by [20] are systematically larger than [23] which implies that
the EPM distances are also .∼ 10–20% larger. The difference between the two sets
of dilution factors is unclear but it is an important source of concern for using the
EPM to measure distances. Gall [28] also found different values for . H0 depending
on the dilution factors. For one object (13eq) in the Hubble flow (.z = 0.041), they
obtained a value of. H0 = 83.± 10 (stat) km s.−1 Mpc.−1 and. H0 = 76.± 9 (stat) km s.−1
Mpc.−1 using dilution factors from [20, 32] respectively. Finally, [29] have success-
fully extended the EPM to high-z and measured the distances to 10 SNe II with a
. z > 0.04. Using these distance measurements , together with those of seven other
1

SNe II in the Hubble flow (.z > 0.01) from the literature, we construct a combined
probability distribution from the individual . H0 posteriors of each SN II using Gaus-
sian kernel-density estimates (see Fig. 10.1)2 . We obtained . H0 = 75.53.+2.07−1.97 (stat)
and . H0 = 84.70.+2.28
−2.21 (stat) using dilution factors from [ 20, 32], respectively.

Pros: Absolute distances, i.e., do not require any external calibration. Less sensitive
to uncertainties in the dust extinction.
Cons: Dilution factors: different sets of dilution factors yield differences of .∼ 20%
in the inferred distance.
Requirements: Optical photometry (at least . BV I ), at least two optical spectra, and
dilution factors.
Future: New dilution factors based on independent numerical methods (e.g. [56]).

10.3 Tailored EPM

To avoid systematic biases introduced by the use of the dilution factors with the EPM,
the Spectral Expanding Atmosphere Method (SEAM; [4–6] and the tailored EPM
([20, 58]) have been developed. Both methods take advantage of the full information
contained in the observed spectra through detailed spectral fitting to infer the chem-

1 We have not considered LSQ13cuw, which is an outlier in the EPM and SCM Hubble diagrams
(see Fig. 7 from [29]).
2 Using . H = cz/Dl[1 + 0.5(1 − q )z] with q. .= .−0.55.
0 0 0
10 The Pursuit of the Hubble Constant Using Type II Supernovae 181

ical composition, the density profile, the temperature, the ejecta velocity, and other
parameters of the SN. Afterwards, because the spectral energy distribution is known,
one can determine the absolute magnitude in any bands from the calculated synthetic
spectra. Then a comparison with to the observed apparent magnitudes allows us to
obtain the distance modulus.
Mainly three radiative-transfer codes have been used to fit observed spectra:
PHOENIX ([6, 35]), CMFGEN ([20], and a modified version of Tardis ([39, 56]). All
the codes take into account the departure from the local thermodynamic equilibrium
(LTE) seen in low-density, scattering-dominated SN II atmospheres via the use of
a non-LTE approximation for the ionization and excitation treatment of the species.
However, radiative-transfer simulations are complicated and time-consuming, rang-
ing from minutes to days on large supercomputers depending on the number of free
parameters. Additionally, it requires high signal to noise spectrum sequence but only
recently such data have been made publicly accessible. For these reasons, only three
well-observed SNe II in the local Universe (99em, 05cs, 06bp) have been studied
using detailed radiative-transfer models ([6, 21]). The authors provided excellent fits
between the observed and synthetic spectra, with a level of accuracy on the distances
.∼ 10%. However, none of those SNe II are enough distant to derive a reliable . H0 .

Recently, radiative-transfer simulations have regained popularity thanks to the

development of a fitting method using a spectral emulator ([57]) to predict the SN
II spectra and magnitudes. The emulator has been trained on a large grid of spectra
calculated with Tardis, and can produce synthetic spectra in only 10.−2 s instead of
10.5 s. It has been built in two steps: (1) data dimensionality reduction by Principal
Component Analysis (PCA), (2) Gaussian processes to interpolate the spectra within
the PCA space and to predict preprocessed, dimensionality-reduced spectra for new
sets of input parameters including: photospheric temperature and velocity, metallic-
ity, time since explosion, and the exponent of the density profile. Finally, a synthetic
spectrum is obtained by reversing the preprocessing procedure and then compare to
the observations. Vogl et al. [57] demonstrated the emulator predictions match the
Tardis simulations with a precision of better than 99%. Vogl [58] applied their new
method to six low-redshift SNe II (03bn, 06it, 10id, 12ck, 13fs, and iPTF13bjx) and
constructed the Hubble diagram displayed in Fig. 10.1 (bottom left). From those six
distances, they were able to derive a value of . H0 = 75.28.+2.80
−2.85 (stat). This result is
currently limited by a small sample size and low redshift range (.z < 0.04), and it
is affected by the systematic uncertainties of atmosphere models. However, this is a
promising method which is independent of local calibrations. Also, it has recently
been checked using sibling SNe II, i.e., SNe II in the same host galaxy. With four
different host galaxies (8 SNe II), [10] found consistent distances within the errors for
each of the SN sibling pairs and with a precision better than 5% for two hosts. Finally,
the adH0 cc collaboration (accurate determination of . H0 with core-collapse super-
novae3 ) has now gathered observations for 26 SNe II in the Hubble flow (.z > 0.03).
This new dataset will provide the basis for a highly-competitive . H0 measurement.

3 https://adh0cc.github.io/.
182 T. de Jaeger and L. Galbany

Fig. 10.2 Figure from [19] representing the last two rungs of the distance ladder with the SCM. For
each rung, distances on the abscissa serve to calibrate a relative distance on the ordinate. Blue dots
represent geometric, Cepheid, or TRGB distances used to calibrate SN II through the determination
of the .i-band absolute magnitudes (Bottom left). Red dots are the SNe II in the Hubble flow used
to constrain . H0 (Top right)

Pros: Absolute distances, i.e., do not require any external calibration. Unlike the
EPM, dilution factors are not needed.
Cons: Small sample size (only 6 objects) and low redshift range (.z < 0.04). Assume
fractional uncertainties for .θ/v ph .
Requirements: A well-defined explosion date, optical photometry (at least . BV I )
interpolated to the spectral epochs, multiple well-calibrated spectra in the first month
after the explosion, and spectral emulator to predict synthetic spectra.
Future: Increase the sample of SN II in the Hubble flow. Better treatment of the
uncertainties.
10 The Pursuit of the Hubble Constant Using Type II Supernovae 183

10.4 SCM: Standard Candle Method

Unlike the EPM and its tailored version, the Standard Candle Method (SCM) is an
empirical method. First developed by [31], it has been later theoretically explained
by [38]. SCM is based on the correlation between the SN absolute magnitude and
the photospheric expansion velocity determined from the blueshift of the Fe II lines
(.λ5018). Such a correlation is expected when the SN is on the plateau phase, i.e.,
between 30–80 d after the explosion ([1]). This phase is physically well-understood
and is only due to a change in opacity and density in the outermost layers of the SN. A
few weeks after the explosion, the ionized hydrogen present in the outermost layers
of the progenitor starts to recombine when the temperature drops and levels of the
hydrogen recombination temperature (.∼5,000 K). During this phase, the light curve
is powered by the sudden release of energy caused by the recession of the photosphere
in mass that corresponds to a fixed radius in space. Therefore, as the temperature and
the radius are nearly constant, the luminosity also appears constant. Consequently,
in more luminous SNe II, the hydrogen recombination front will be fixed at a larger
radius, and therefore, assuming a homologous expansion, the photosphere will be
maintained at higher velocities.
Hamuy [34] applied this technique to 24 SNe II in the Hubble flow and reduced
the Hubble diagram scatter from 0.83 mag to only 0.32 mag, i.e., 15% in distances.
Distances derived using the SCM are not absolute and must be calibrated using
primary distance indicators like Cepheids or Tip Red Giant Branch (TRGB). Using
four SNe II (68L, 70G, 73R, 99em) with precise Cepheid distances, [34] derived a
−1
. H0 of 75 .± 7 km s. Mpc.−1 . With the same technique but with only one SN (99em)
calibrated via Cepheid, [42] obtained a smaller . H0 of 59 .± 3 (stat) ± 11 (sys).
Recently, many others studies have refined the SCM to simply its use and to extend
it to higher redshifts. First, [43] added an extinction correction (intrinsically brighter
SNe II are bluer) like for SN Ia cosmology (.β .× color) rather than relying on a method
to measure the host-galaxy extinction. Second, he adjusted the velocity of the Fe II
lines with a power-law decline, reducing the need of obtaining several spectra around
the middle of the plateau phase. Third, [13, 43, 47, 55] have proposed to use the
H.β .λ4861 absorption line which has the advantage of being stronger than Fe II lines
(.λ5018), and therefore easier to measure. Finally, because the measurement of the
velocity can be difficult for noisy even for the strong H.β .λ4861 absorption line, [14,
47] have developed a method based on the cross-correlation between the observed
spectra and a library of high S/N SN II spectra. Velocities from direct measurement
or cross-correlation techniques have shown a good agreement with a dispersion of
only 400 km s.−1 ([14]). Therefore, finally, using the SCM, the corrected magnitude
is written as,

m corr = m + α log10 (vHβ ) − βcolor,

. (10.2)

where .m is the apparent magnitude in a given passband around 50 d after the explo-
sion, .vHβ is the velocity measured using H.β absorption from an optical spectrum, .α
184 T. de Jaeger and L. Galbany

and .β are nuisance parameters. The fine-tuned version of the SCM have been used
by several studies ([11–14, 17, 29, 44, 47, 48]) to measure SN II distances in the
Hubble flow with a precision of 12–15% in distance. However, only three works
have applied this updated version to measure . H0 . First, [44] used two SNe II with
Cepheid distances (99em, 04dj) and obtained a value of 71 .± 12 km s.−1 Mpc.−1 , 69
−1
.± 16 km s. Mpc.−1 , and 70 .± 16 km s.−1 Mpc.−1 depending on the passband used
(. B, .V , and . I , respectively). More recently, using seven objects (99em, 99gi, 05ay,
05cs, 09ib, 12aw, and 13ej) with Cepheid or TRGB independent host-galaxy distance
measurements, [18] found a . H0 value of .75.8+5.2 −4.9 (stat) .± 2.8 (sys) km s.
−1
Mpc.−1 ,
which differs by .1.4σ from the high-redshift result ([46]). Thanks to six addi-
tional calibrators (04et, 08bk, 14bc, 17eaw, 18aoq, 20yyz) and a better analysis
of the systematic errors, [19] were able to obtain a . H0 with an uncertainty of 5%,
+3.8 (stat) −1
.75.4−3.7 (stat) ± 1.5 (sys) km s. Mpc.−1 . In Fig. 10.2, the last two rungs of the dis-
tance ladder from their latest work are displayed. Their value is consistent with the
local measurement [50] but differs by .2.0σ from the high-redshift results [46]. Addi-
tionally, using seven Cepheids or five TRGB, they derived values which differ by
−1
.< 1.0σ (i.e., 4.5 kms. Mpc.−1 ), suggesting that neither Cepheids nor TRGB are the
source of the “. H0 tension.”. It is important to note that one SN II (14bc) was dis-
covered in NGC 4258 for which we have geometric distance. Therefore, we could
use 14bc to directly calibrate the SN II in the Hubble flow, without the need for any
calibrators, i.e., it will be just a two-rungs distance ladder. However, the uncertainties
on its SCM distance and mostly the uncertainty on the ejecta expansion velocity are
too large to constrain . H0 .
Pros: Straightforward simple method and has been applied at high redshift.
Cons: Needs to be calibrated. Common reddening law for all SNe.
Requirements: Optical photometry (at least two bands) up to 50 d after the explo-
sion, one optical spectrum during the plateau phase, a well-defined explosion date
(.σTex p <10 d), calibrators (Cepheids, TRGB) in the same host.
Future: Increase the number of calibrators, and find new correlations to decrease
the scatter in the Hubble diagram (e.g. host galaxy environment).

10.5 PMM: Photospheric Magnitude Method

Rodríguez [51] developed another method to measure SN II distances and build

Hubble diagram with a scatter of 0.15 mag. This method is a generalization of the
SCM (Sect. 10.4) at any epoch during the photospheric phase. For a given bandpass
. X , the absolute magnitude (. M X ) at an epoch .ti (since the explosion .t0 ) can be written

as,

.MX,ti = a X,ti − 5log10 (v ph ), (10.3)

where .a X,ti is a function that can be calibrated empirically and .v ph the photospheric
velocity derived from the Fe II.λ5169 line absorption minima. Finally, the SN distance
modulus is obtained using,
10 The Pursuit of the Hubble Constant Using Type II Supernovae 185

Fig. 10.3 Figure updated from figure of [52] representing SNe II distance moduli derived with
the PMM for the . B (blue triangles) and . H (red squares) bandpass versus the redshift corrected for
peculiar velocities. Red and blue lines correspond to the Hubble law fits to only the SNe II at .cz .>
2000 km s.−1 (shaded region)

.μ X,ti = m corr
X,ti − a X,ti + 5log10 (v ph ), (10.4)

where.m corr
X,ti is the apparent magnitude corrected for Galactic, host galaxy extinctions
and K-correction. Unlike the SCM method for which a color term is added in the
SN standardisation, [51] measured the host galaxy extinction using the colours and
assuming that all SNe II have the same intrinsic colour. However, this assumption
has recently been challenged by [15] who demonstrated that intrinsic SN II colors
are most probably dominated progenitor properties like its radius.
To calibrate the PMM zero-point (encapsulated in .a X,ti ), [52] used four SNe II
with known distances from TRGB (SN 03hn, 05cs, 12aw, and 13ej). With only 9
SNe II with .cz .> 2000 km s.−1 , they were able to derive . H0 value between 67.1 and
74.9 km s.−1 Mpc.−1 . The large range of . H0 values is a source of concern for using
the PMM as the . H0 values strongly depend on the filter chosen. For example, using
the . B band, they obtained 74.9.+10.6
−9.2 (stat) km s.
−1
Mpc.−1 while using redder filter,
+5.4 −1 −1
this value decrease to 67.1.−5.0 (stat) km s. Mpc. in . H band (see Fig. 10.3).
Pros: Straightforward simple method. Smaller scatter in the Hubble diagram than
the SCM.
Cons: Small sample size and affected by peculiar velocities.. H0 depends on the filters
used. Host galaxy extinction estimated with colours. Needs to be calibrated. Never
applied at high redshifts.
Requirements: Photometry (at least two bands) between 35–75 d since explo-
sion, one optical spectrum during the plateau phase, a well-defined explosion date
(.σTex p <5 d), calibrators (Cepheids, TRGB) in the same host.
Future: Increase the number of calibrators and SN in the Hubble flow.
186 T. de Jaeger and L. Galbany

Table 10.1 . H0 values from Type II supernovae

. H0 Method.+ Year References
km s.−1 Mpc.−1 dilution factors or N.cal
60 .± 15 (stat) EPM 1974 [40]
73 .± 6 (stat) ± 7 (sys) EPM.+ [22] 1994 [53]
71 .± 9 (stat) EPM.+ [32] 2001 [32]
57 EPM.+ [32] 2003 [42]
.± 7 (stat) ± 13 (sys)

100.5.± 8.4 (stat) EPM.+ [23] 2009 [37]

59.6.± 4.2 (stat) EPM.+ [20] 2009 [37]
83 .± 10 (stat) EPM.+ [32] 2016 [28]
76 .± 9 (stat) EPM.+ [20] 2016 [28]
84.70.+2.28
−2.21 (stat) ± EPM.+ [32] 2023 This work
10.47 (sys)
75.57.+2.04
−1.89 (stat) ± EPM.+ [20] 2023 This work
13.46 (sys)
72.28.+2.80
−2.85 (stat) ± Tailored EPM 2020 [58]
3.79 (sys)
74.9.+10.6
−9.2 (stat) PMM.+4.+B 2019 [52]
+5.4
67.1.−5.0 (stat) PMM.+4.+H 2019 [52]
59 SCM.+1 2003 [42]
.± 3 (stat) ± 11 (sys)

75 .± 7 (stat) SCM .+4 2003 [34]

71 .± 12 (stat) SCM.+2 2010 [44]
75.8.+5.2
−4.9 (stat) ± SCM.+7 2020 [18]
2.8 (sys)
75.4.+3.8
−3.7 (stat) ± SCM.+13 2022 [19]
1.5 (sys)

10.6 Conclusions

All . H0 measurements from SNe II are summarized in Table 10.1 and presented in
Fig. 10.4 in comparison to late-Universe Cepheid ([50]) and TRGB ([25]) estimates,
and the early-Universe ([46]) value from Planck. For the EPM values measured in
this work and the tailored EPM by Vogl et al. ([58]) we considered the dispersion of
all individual . H0 measurements as an estimate of the systematic uncertainty.
Although not yet as mature as those using SNe Ia, different methods have been
developed to standardize SNe II to make them useful extragalactic distance indicators
and, in turn, provide independent estimates of. H0 . Their main advantages over SNe Ia
are: (i) the extended knowledge on their progenitor systems and the physics behind
their explosion mechanism; (ii) their larger volumetric rates; and (iii) their more
direct connection with the local environment, which could potentially be used to
10 The Pursuit of the Hubble Constant Using Type II Supernovae 187

Fig. 10.4 Comparison of . H0 derived using SNe II, SNe Ia calibrated with Cepheids ([50]) and
TRGB ([25]), and from the CMB anisotropies and lensing ([46]). For the Type II supernovae
measurements, we represent in red those using the SCM, in brown those from the PMM method,
and in blue those using the EPM, including the tailored version. Light errors bars represent statistical
and sytematic uncertainties in quadrature, while solid error bars are only statistical uncertainties

improve their standardization. In addition, their constraints on . H0 are independent

to those from SNe Ia since they are affected by different systematic uncertainties.
Advantageously, some stand-alone methods exist that do not need calibrators as the
distance ladder does. The different methods summarized here, some of which do
not depend on external calibrators, are more likely to be improved in the following
years thus reducing the impact of systematic uncertainties in the determination of
. H0 . Moreover, with more well-observed objects to be included that will reduce the
statistical uncertainties, SNe II will certainly contribute to understanding the reason
behind the Hubble tension.

Acknowledgements T. dJ acknowledges financial support from the French Centre National de

la Recherche Scientifique (CNRS/IN2P3) and from the French Agence National de la Recherche
project 21-CE31-0016. L.G. acknowledges financial support from the Spanish Ministerio de Ciencia
e Innovación (MCIN), the Agencia Estatal de Investigación (AEI) 10.13039/501100011033, and
the European Social Fund (ESF) “Investing in your future” under the 2019 Ramón y Cajal program
RYC2019-027683-I and the PID2020-115253GA-I00 HOSTFLOWS project, from Centro Superior
de Investigaciones Científicas (CSIC) under the PIE project 20215AT016, and the program Unidad
de Excelencia María de Maeztu CEX2020-001058-M.
188 T. de Jaeger and L. Galbany

References

1. J.P. Anderson, S. González-Gaitán, M. Hamuy, C.P. Gutiérrez, M.D. Stritzinger, F.E. Olivares.,
M.M. Phillips, S. Schulze et al. Astrophys. J. 786, 67 (2014). https://doi.org/10.1088/0004-
637X/786/1/67[arXiv:1403.7091 [astro-ph.HE]]
2. W. Baade, Astron. Nachr. 228, 359–362 (1926)
3. R. Barbon, F. Ciatti, L. Rosino, Astron. Astrophys. 72, 287–292 (1979)
4. E. Baron, P.H. Hauschildt, P. Nugent, D. Branch, Mon. Not. R. Astron. Soc. 283(1), 297–315
(1996). https://doi.org/10.1093/mnras/283.1.297
5. E. Baron, D. Branch, P.H. Hauschildt, A.V. Filippenko, R.P. Kirshner, P.M. Challis,
S. Jha, R. Chevalier et al. Astrophys. J. 545, 444–448 (2000). https://doi.org/10.1086/
317795[arXiv:astro-ph/0010614 [astro-ph]]
6. E.A. Baron, P.E. Nugent, D. Branch, P.H. Hauschildt, Astrophys. J. Lett. 616, L91–L94 (2004).
https://doi.org/10.1086/426506 [arXiv:astro-ph/0410153 [astro-ph]]
7. M. Betoule et al. [SDSS] Astron. Astrophys. 568, A22 (2014). https://doi.org/10.1051/0004-
6361/201423413 ([arXiv:1401.4064 [astro-ph.CO]])
8. S. Bose, B. Kumar, Astrophys. J. 782, 98 (2014). https://doi.org/10.1088/0004-637X/782/2/
98. ([arXiv:1401.5115 [astro-ph.CO]].)
9. D. Brout, D. Scolnic, B. Popovic, A.G. Riess, J. Zuntz, R. Kessler, A. Carr,
T.M. Davis et al. Astrophys. J. 938(2), 110 (2022). https://doi.org/10.3847/1538-4357/
ac8e04([arXiv:2202.04077 [astro-ph.CO]])
10. G. Csörnyei, C. Vogl, S. Taubenberger, A. Flörs, S. Blondin, M.G. Cudmani, A. Holas,
S. Kressierer et al. ([arXiv:2302.03112 [astro-ph.SR]])
11. C.B. D’Andrea, M. Sako, B. Dilday, J.A. Frieman, J. Holtzman, R. Kessler, K. Konishi,
D.P. Schneider et al. Astrophys. J. 708, 661–674 (2010). https://doi.org/10.1088/0004-637X/
708/1/661 ([arXiv:0910.5597 [astro-ph.CO]])
12. T. de Jaeger, S. González-Gaitán, J.P. Anderson, M. Hamuy, L. Galbany, M.M. Phillips, M.D.
Stritzinger, C.P. Gutiérrez et al., Astrophys. J. 815, 121 (2015). https://doi.org/10.1088/0004-
637X/815/2/121. ([arXiv:1511.05145 [astro-ph.HE]].)
13. T. de Jaeger, S. González-Gaitán, M. Hamuy, L. Galbany, J.P. Anderson, M.M. Phillips,
M.D. Stritzinger, R.G. Carlberg et al. Astrophys. J. 835(2), 166 (2017). https://doi.org/10.
3847/1538-4357/835/2/166 ([arXiv:1612.05636 [astro-ph.CO]])
14. T. de Jaeger, L. Galbany, A.V. Filippenko, S. González-Gaitán, N. Yasuda, K. Maeda,
M. Tanaka, T. Morokuma et al. Mon. Not. R. Astron. Soc. 472(4), 4233–4243 (2017). https://
doi.org/10.1093/mnras/stx2300 ([arXiv:1709.01513 [astro-ph.HE]])
15. T. de Jaeger, J.P. Anderson, L. Galbany, S. González-Gaitán, M. Hamuy, M.M. Phillips,
M.D. Stritzinger, C. Contreras et al. Mon. Not. R. Astron. Soc. 476(4), 4592–4616 (2018).
https://doi.org/10.1093/mnras/sty508 ([arXiv:1802.07254 [astro-ph.HE]])
16. T. de Jaeger, W. Zheng, B. E. Stahl, A. V. Filippenko, T. G. Brink, A. Bigley, K. Blanchard,
P. K. Blanchard, et al. Mon. Not. Roy. Astron. Soc. 490, no.2, 2799-2821 (2019) doi:https://
doi.org/10.1093/mnras/stz2714[arXiv:1909.13813 [astro-ph.HE]]
17. T. de Jaeger et al. [DES], https://doi.org/10.1093/mnras/staa1402[arXiv:2005.09757 [astro-
ph.HE]]
18. T. de Jaeger, B.E. Stahl, W. Zheng, A.V. Filippenko, A.G. Riess, L. Galbany, Mon.
Not. R. Astron. Soc. 496(3), 3402–3411 (2020). https://doi.org/10.1093/mnras/staa1801
([arXiv:2006.03412 [astro-ph.CO]])
19. T. de Jaeger, L. Galbany, A.G. Riess, B.E. Stahl, B.J. Shappee, A.V. Filippenko, W. Zheng,
Mon. Not. R. Astron. Soc. 514(3), 4620–4628 (2022). https://doi.org/10.1093/mnras/stac1661
([arXiv:2203.08974 [astro-ph.CO]])
20. L. Dessart, D.J. Hillier, Astron. Astrophys. 439, 671 (2005). https://doi.org/10.1051/0004-
6361:20053217. ([arXiv:astro-ph/0505465 [astro-ph]].)
21. L. Dessart, S. Blondin, P.J. Brown, M. Hicken, D.J. Hillier, S.T. Holland, S. Immler, R.P. Kir-
shner et al., Astrophys. J. 675, 644 (2008). https://doi.org/10.1086/526451. ([arXiv:0711.1815
[astro-ph]].)
10 The Pursuit of the Hubble Constant Using Type II Supernovae 189

22. R.G. Eastman, P.A. Pinto, Astrophys. J. 412, 731 (1993). https://doi.org/10.1086/172957
23. R.G. Eastman, B.P. Schmidt, R.P. Kirshner, Astrophys. J. 466, 911 (1996). https://doi.org/10.
1086/177563
24. A.V. Filippenko, Ann. Rev. Astron. Astrophys. 35, 309–355 (1997). https://doi.org/10.1146/
annurev.astro.35.1.309
25. W.L. Freedman, Astrophys. J. 919(1), 16 (2021). https://doi.org/10.3847/1538-4357/ac0e95
([arXiv:2106.15656 [astro-ph.CO]])
26. A. Gal-Yam, https://doi.org/10.1007/978-3-319-20794-0_35-1([arXiv:1611.09353 [astro-
ph.HE]])
27. L. Galbany, M. Hamuy, M.M. Phillips, N.B. Suntzeff, J. Maza, T. de Jaeger, T. Moraga,
S. González-Gaitán et al. https://doi.org/10.3847/0004-6256/151/2/33 ([arXiv:1511.08402
[astro-ph.SR]])
28. E.E.E. Gall, R. Kotak, B. Leibundgut, S. Taubenberger, W. Hillebrandt, M. Kromer,
Astron. Astrophys. 592, A129 (2016). https://doi.org/10.1051/0004-6361/201628333.
([arXiv:1603.04730 [astro-ph.CO]].)
29. E.E.E. Gall, R. Kotak, B. Leibundgut, S. Taubenberger, W. Hillebrandt, M. Kromer, W.S.
Burgett, K. Chambers et al., Astron. Astrophys. 611, A25 (2018). https://doi.org/10.1051/
0004-6361/201731271. ([arXiv:1705.10806 [astro-ph.CO]].)
30. M. Hamuy, M.M. Phillips, N.B. Suntzeff, R.A. Schommer, J. Maza, R. Aviles, Astron. J. 112,
2398 (1996). https://doi.org/10.1086/118191. ([arXiv:astro-ph/9609062 [astro-ph]].)
31. M. Hamuy, Type II supernovae as distance indicators. Ph.D. Thesis, The University of Arizona
(2001)
32. M. Hamuy, P.A. Pinto, J. Maza, N.B. Suntzeff, M.M. Phillips, R.G. Eastman, R.C.
Smith, C.J. Corbally et al., Astrophys. J. 558, 615 (2001). https://doi.org/10.1086/322450.
([arXiv:astro-ph/0105006 [astro-ph]].)
33. M. Hamu, P.A. Pinto, Astrophys. J. Lett. 566, L63–L65 (2002). https://doi.org/10.1086/339676
([arXiv:astro-ph/0201279 [astro-ph]])
34. M. Hamuy, Springer Proc. Phys. 99, 535–541 (2005). https://doi.org/10.1007/3-540-26633-
X_71 ([arXiv:astro-ph/0309122 [astro-ph]])
35. P.H. Hauschildt, E. Baron, J. Comput. Appl. Math. 109, 41–63 (1999). https://doi.org/10.1016/
S0377-0427(99)00153-3 ([arXiv:astro-ph/9808182 [astro-ph]])
36. H.T. Janka, Ann. Rev. Nucl. Part. Sci. 62, 407–451 (2012). https://doi.org/10.1146/annurev-
nucl-102711-094901 ([arXiv:1206.2503 [astro-ph.SR]])
37. M.I. Jones, M. Hamuy, P. Lira, J. Maza, A. Clocchiatti, M. Phillips, N. Morrell, M. Roth
et al. Astrophys. J. 696, 1176–1194 (2009). https://doi.org/10.1088/0004-637X/696/2/1176
([arXiv:0903.1460 [astro-ph.CO]])
38. D. Kasen, S.E. Woosley, Astrophys. J. 703, 2205 (2009). https://doi.org/10.1088/0004-637X/
703/2/2205. ([arXiv:0910.1590 [astro-ph.CO]].)
39. W.E. Kerzendorf, S.A. Sim, Mon. Not. R. Astron. Soc. 440(1), 387–404 (2014). https://doi.
org/10.1093/mnras/stu055 ([arXiv:1401.5469 [astro-ph.CO]])
40. R.P. Kirshner, J. Kwan, Astrophys. J. 193, 27 (1974). https://doi.org/10.1086/153123
41. D.C. Leonard, A.V. Filippenko, Publ. Astron. Soc. Pac. 113, 920 (2001). https://doi.org/10.
1086/322151. ([arXiv:astro-ph/0105295 [astro-ph]].)
42. D.C. Leonard, S.M. Kanbur, C.C. Ngeow, N.R. Tanvir, Astrophys. J. 594, 247–278 (2003).
https://doi.org/10.1086/376831 ([arXiv:astro-ph/0305259 [astro-ph]])
43. P. Nugent et al. [SNLS], Astrophys. J. 645, 841–850 (2006). https://doi.org/10.1086/504413
([arXiv:astro-ph/0603535 [astro-ph]])
44. F.E. Olivares, M. Hamuy, G. Pignata, J. Maza, M. Bersten, M.M. Phillips, N.B. Suntzeff,
A.V. Filippenko et al., Astrophys. J. 715, 833–853 (2010). https://doi.org/10.1088/0004-637X/
715/2/833 ([arXiv:1004.2534 [astro-ph.CO]])
45. M.M. Phillips, Astrophys. J. Lett. 413, L105–L108 (1993). https://doi.org/10.1086/186970
46. N. Aghanim et al., [Planck], Astron. Astrophys. 641, A6 (2020) [erratum: Astron. Astrophys.
652, C4 (2021)]. https://doi.org/10.1051/0004-6361/201833910 ([arXiv:1807.06209 [astro-
ph.CO]])
190 T. de Jaeger and L. Galbany

47. D. Poznanski, N. Butler, A.V. Filippenko, M. Ganeshalingam, W. Li, J.S. Bloom, R. Chornock,
R.J. Foley et al., Astrophys. J. 694, 1067–1079 (2009). https://doi.org/10.1088/0004-637X/
694/2/1067 ([arXiv:0810.4923 [astro-ph]])
48. D. Poznanski, P.E. Nugent, A.V. Filippenko, Astrophys. J. 721, 956–959 (2010). https://doi.
org/10.1088/0004-637X/721/2/956 ([arXiv:1008.0877 [astro-ph.CO]])
49. A.G. Riess, W.H. Press, R.P. Kirshner, Astrophys. J. 473, 88 (1996). https://doi.org/10.1086/
178129. ([arXiv:astro-ph/9604143 [astro-ph]].)
50. A.G. Riess, W. Yuan, L.M. Macri, D. Scolnic, D. Brout, S. Casertano, D.O. Jones, Y. Murakami
et al., Astrophys. J. Lett. 934(1), L7 (2022). https://doi.org/10.3847/2041-8213/ac5c5b
([arXiv:2112.04510 [astro-ph.CO]])
51. Ó. Rodríguez, A. Clocchiatti, M. Hamuy, Astron. J. 148, 107 (2014). https://doi.org/10.1088/
0004-6256/148/6/107 ([arXiv:1409.3198 [astro-ph.CO]])
52. Ó. Rodríguez, G. Pignata, M. Hamuy, A. Clocchiatti, M.M. Phillips, K. Krisciunas, N.I. Morrell,
G. Folatelli et al., Mon. Not. R. Astron. Soc. 483(4), 5459–5479 (2019). https://doi.org/10.
1093/mnras/sty3396 ([arXiv:1812.04982 [astro-ph.CO]])
53. B.P. Schmidt, R.P. Kirshner, R.G. Eastman, M.M. Phillips, N.B. Suntzeff, M. Hamuy, J. Maza,
R. Aviles, Astrophys. J. 432, 42 (1994). https://doi.org/10.1086/174546
54. S.J. Smartt, Publ. Astron. Soc. Aust. 32, e016 (2015). https://doi.org/10.1017/pasa.2015.17.
([arXiv:1504.02635 [astro-ph.SR]].)
55. K. Takats, J. Vinko, Mon. Not. R. Astron. Soc. 419, 2783 (2012). https://doi.org/10.1111/j.
1365-2966.2011.19921.x. ([arXiv:1109.5873 [astro-ph.SR]].)
56. C. Vogl, S.A. Sim, U.M. Noebauer, W.E. Kerzendorf, W. Hillebrandt, Astron. Astrophys.
621, A29 (2019). https://doi.org/10.1051/0004-6361/201833701. ([arXiv:1811.02543 [astro-
ph.HE]].)
57. C. Vogl, W.E. Kerzendorf, S.A. Sim, U.M. Noebauer, S. Lietzau, W. Hillebrandt, Astron. Astro-
phys. 633, A88 (2020). https://doi.org/10.1051/0004-6361/201936137. ([arXiv:1911.04444
[astro-ph.HE]].)
58. C. Vogl, Cosmological distances of type II supernovae from radiative transfer modeling, PhD
thesis, Munich Technical University, 2020
59. R.V. Wagoner, Astrophys. J. Lett. 250, L65–L69 (1981). https://doi.org/10.1086/183675
60. S.E. Woosley, T.A. Weaver, Astrophys. J. Suppl. 101, 181–235 (1995). https://doi.org/10.1086/
192237
Chapter 11
The Mira Distance Ladder

Caroline D. Huang

11.1 Introduction

Mira variables are fundamental-mode, radially-pulsating tip of the Asymptotic Giant

Branch (AGB) stars with periods ranging from.∼100–1000 days or longer. They were
the first class of periodic variables to be identified, starting with the discovery of Mira
(Omicron Ceti) by David Fabricius on August 3, 1596. With their high luminosities
and the largest amplitudes of any regular pulsator (.>2.5 .m V [1]), they remain a
popular and easy target for amateur astronomers to this day.
As highly-evolved AGB stars, they contribute to the chemical enrichment of
the interstellar medium (ISM) through stellar mass loss [2] and are also extremely
luminous—generally the brightest stars in an intermediate-to-old stellar population.
They also also ubiquitous, as nearly all stars (.0.8MΘ < M < 8MΘ ) will experience
a Mira phase in evolution.
Period-luminosity relations (P–L relations) for Miras have been known for several
decades and have been employed to measure distances to many Local Group galaxies
([3] and references therein). However, the persistent Hubble tension has resulted in
renewed interest in Mira distances as an alternative and cross-check to the more
commonly-used Cepheid variables and Tip of the Red Giant Branch (TRGB) thanks
to their age (intermediate-age Miras are.∼3 Gyr) and luminosity (up to a few.104 L Θ ).
With their low effective temperatures (.Teff < 3500 K) they are particularly attractive
in light of the upcoming decade’s focus on NIR and IR wavelengths since the spectral
intensity of Miras with thin circumstellar shells (such that those currently used in
distance measurements) peaks between 1–2.µm.
Miras may be classified into two spectral types—Carbon-(C-) and Oxygen-(O-)
rich. Current distance measurements using Miras employ exclusively short-period
(. P < 400 d), O-rich variables, which have been shown to have a tight (.σ ∼ 0.12

C. D. Huang (B)
Center for Astrophysics | Harvard & Smithsonian, 60 Garden St., Cambridge, MA 02138, USA
e-mail: caroline.huang@cfa.harvard.edu
© The Smithsonian Institution 2024 191
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_11
192 C. D. Huang

mag) P–L relation in the near-infrared. C-rich Miras have larger scatter at shorter
wavelengths, and longer-period (. P > 400 days) Miras may follow a different P–L
relation due to hot-bottom burning [4].
In this review, we will briefly discuss the evolutionary status of Miras in Sect. 11.2.
Next, we will give an overview of the history of various Mira Period-Luminosity
Relations and some local measurements and studies with Miras in Sect. 11.3. In
Sect. 11.4 we will give outline the previous efforts related to measuring H.0 with
Miras. In Sect. 11.5 we will briefly discuss prospects and previous observations of
long-period Miras. In Sect. 11.6 we compare Miras with other stellar primary distance
indicators and conclude in Sect. 11.7.

11.2 Evolutionary Status

Miras are the brightest of the AGB stars and the AGB phase is the final stage of
nuclear-burning life for vast majority of stars. It is often stated that stars with masses
.0.8MΘ < M < 8MΘ will experience an AGB phase of evolution, though the exact

upper limit depends on how convection is handled. The AGB stage is also brief,
lasting .< 1% of the main sequence lifetime. On the Hertzsprung-Russell diagram,
these stars are typically identified as the bright stars above the tip of the Red Giant
Branch (TRGB). Figure 11.1 shows theoretical evolutionary tracks for stars from 0.4
. MΘ to 16 . MΘ and the location of the AGB, where the Miras reside.
AGB stars have evolved past core hydrogen and helium fusion. Instead, they
have a degenerate CO core and sustain themselves against gravitational collapse pri-
marily through hydrogen-shell burning and quasi-periodic helium-shell flashes that

Fig. 11.1 Hertzsprung-

Russell Diagram created
using MIST evolutionary
tracks [5–7]. The zero-age
main sequence is shown in
black, individual
evolutionary tracks in gray,
and the thermally-pulsating
AGB evolutionary stage in
red. Miras are located at the
uppermost region of the
thermally-pulsating AGB
branch
11 The Mira Distance Ladder 193

ignite helium-shell burning. Stars actively undergoing this process are also known
as thermally-pulsating (TP-AGB) stars.
The cycle of thermal pulses is also responsible for dredging-up carbon from the
core and potentially changing the surface chemistry of a star from O-rich to C-
rich. AGB stars exhibit enrichments of elements produced through the slow neutron
capture process (s-process) and are also often experiencing rapid mass loss with
rates of .10−8 MΘ < Ṁ < 10−4 MΘ year.−1 [8] and typical wind velocities of .∼ 10
km/second.
Stars ascending the AGB will are likely to experience multiple phases of pul-
sational instability, first as low-amplitude, overtone pulsators, and then finally as
high-amplitude, fundamentally-pulsating Miras at the tip of the AGB. During the
Mira phase, stars are at their most luminous and also have the highest mass-loss
rates. This stage is also quite brief, lasting .∼ 105 years for low-mass stars. Depend-
ing on their mass and abundance, due to the He-shell flashes, it is also possible that
during their evolution, stars may also enter and leave the Mira phase more than once.
The post-AGB evolution of a star is one of the least well-understood stages of
stellar evolution. Low-to-intermediate mass stars are thought to become post-AGB
stars and eventually blow off the outer layers to form planetary nebulae. The final
life stage of more massive stars is even more uncertain, although they may go on to
become super-AGB stars.
There have been several excellent previous reviews of the AGB evolution, includ-
ing [8–14]. Here we will provide only a brief summary of the evolutionary status of
Miras with a focus on the aspects most relevant to distance measurements.

11.2.1 The TP-AGB and Third Dredge up

A dredge-up is the process that mixes the products of nucleosynthesis from the
nuclear burning zone of the star to the its upper layers via convection. The first
dredge-up occurs after core hydrogen burning but before helium burning starts, as
the star is ascending the red giant branch. The second dredge up occurs for stars of 4–
8 . MΘ after core helium fusion comes to an end. Lower mass stars will not experience
a second dredge-up at all. Therefore, the names of the dredge up processes do not
necessarily indicate the order in which they occur. We will primarily focus on the
third dredge-up, which takes place when a star has entered the asymptotic giant
branch, after a helium flash, and brings helium, carbon, and s-process elements to
the surface.
During the AGB phase, the hydrogen-burning shell is the primary source of energy.
As the star evolves up the AGB, the helium shell becomes thinner and thermally
unstable, resulting in explosive, quasi-periodic ignitions of the helium-burning shell
[15]. These are known as shell flashes or thermal pulses (TPs). Prior to the first TP,
a star is in the early AGB phase and after the first TP, it is on the TP-AGB phase.
194 C. D. Huang

Each of these cycles of AGB evolution can be broken down into four phases
[14, 16]:
1. Thermal pulse: Helium-burning is ignited explosively at the base of the inter-
shell region. High luminosities of .≥ 8L Θ are produced for a short timescale of
.∼ 10 years. The energy from the flash creates a convective zone that extends
2

from the He-burning shell to the H-burning shell and mixes the abundances in
the intershell region.
2. Power Down: During power-down, energy produced from the thermal pulse
expands the star, which cools the material outwards of the He-shell and halts
fusion the H-shell. The surface luminosity will dip by .∼ 10% during this expan-
sion. The inner layers will cool, leading to a increase in stellar opacity.
3. Third dredge-up: The increase in stellar opacity results in the base of outer
convective envelope moving inwards, past the previous H-burning shell and into
the intershell region. This mixes the elements from He-burning to the surface,
where they can be observed.
4. Interpulse: After the third dredge-up, the star will contract once again, reigniting
the H-burning shell. The star then enters a long quiescent phase for .∼.104 years.
During this time, the H-shell provides most of the surface luminosity before the
cycle repeats.
Figure 11.2 shows the interior of an AGB star and the relative positions of the regions
where these processes take place.
Stars enter the AGB with a photospheric ratio of C/O .< 1 (the exact ratio depends
on the initial oxygen abundance) and are classified as O-rich (M-) stars. On the AGB,
this can change through the third dredge up process, which brings carbon synthesized
through the triple-alpha reactions at the base of the helium shell to the surface via a

Fig. 11.2 The interior of an

AGB star—not to
scale—based on Fig. 1 from
[17]. The degenerate CO
core is surrounded by a
He-burning shell. The
He-burning shell is separated
from the H-burning shell by
an intershell region
consisting mostly of helium
and carbon. The H-burning
shell lies directly below the
deep, outer convective
envelope. The ratio of radial
thickness of the core
compared to the envelope is
.∼ 1 × 10
−5
11 The Mira Distance Ladder 195

convective zone. If the surface C/O ratio becomes .>1, then the star is classified as
C-rich (C-type) star. Environments with lower oxygen abundance are host to more
C-rich stars, since less carbon is required to change the surface chemistry from O-
to C-rich.
The two subtypes can be readily separated in the basis of the molecular features in
their spectra. Carbon and oxygen will form stable CO molecules, which then leaves
only the element with greater abundance for other chemical reactions.

11.2.2 Hot-Bottom Burning

Another process that affects the surface chemistry of some AGB stars is hot-bottom
burning. Hot-bottom burning occurs in relatively massive AGB stars .∼ 4 − 7MΘ
and happens alongside the third dredge-up process. In stars experiencing hot-bottom
burning, the surface convective envelope is deep enough to that proton-capture nucle-
osynthesis can occur at the base [18–21].
When the temperature at the base of the convective envelope exceeds.5 × 107 K the
CNO cycle is activated, which burns .12 C to .14 N. Furthermore, the entire convective
envelope has a turnover time of 1 year, meaning that the entire envelope is well-
mixed and will be exposed to this hot region within a year. This process prevents a
C-rich atmosphere from forming and is responsible for limiting the number of C-rich
variables at the longest periods and highest masses.
The exact mass at which this process activates is unknown [22]. Models indicate
that AGB stars with lower metallicity experience hot-bottom burning at lower initial
mass compared to stars with higher metallicity. Therefore, the masses at which hot-
bottom burning occurs—and consequently, the masses of the stars that will become
C-rich is not known and will almost certainly depend on the environment.

11.3 Period-Luminosity Relations

The Period-Luminosity Relation itself is typically defined as a linear relationship

between mean magnitudes of variables and the logarithm of their periods. The mag-
nitudes can be determined through averaging the flux or magnitude mean or by the
mean of a (generally sinusoidal for Miras) fit to fluxes or magnitudes. Fit means are
typically more consistent than an average, and therefore, preferred.
While mean magnitudes are by far the most commonly-used for P–L relations,
Mira P–L relations also occasionally use bolometric magnitude or peak magnitude.
Bolometric luminosities are more fundamental, easier to compare with models, and
independent of circumstellar reddening, but they require more extensive observations
over a longer period of time and in a large wavelength range. In addition, the results
can be dependent on the color-dependent bolometric corrections used [23]. Peak
magnitudes, on the other hand, are sometimes used with optical observations of
196 C. D. Huang

Miras [24, 25]. Due to the large amplitudes of these stars at shorter wavelengths,
mean magnitudes can be difficult to measure and therefore peak magnitude can result
in lower-dispersion P–L relations. In addition, the luminosity is inversely dependent
on period in the shorter optical wavelengths, and the P–L relation becomes roughly
flat around the I-band. Figure 11.3 from [26] shows the single-epoch P–L relations
for Large Magellanic Cloud Miras at a range of wavelengths, from optical to the far
infrared. With mean magnitudes the scatter in the P–L relations would be significantly
smaller.

11.3.1 Basics of Mira Pulsation

Pulsating stars exist in various parts of the Hertzsprung-Russell Diagram, grouped

in narrow (in color/temperature), nearly-vertical regions known as instability strips.
While the relationship between luminosity and period for variable stars (specifically
Cepheids) was determined empirically [27], an intuitive explanation for the existence
of the P–L relation can be derived by examining the relationship between pulsational
period and radius.
The boundary conditions for a pulsating star are similar to those in an organ pipe,
where one end (the center) is closed, and the other (the surface) is open. In the
fundamental, lowest-frequency mode, the entire star will contract and expand at the
same time. Thus we expect that,

R
.∏∼4 (11.1)
v̄s

where .∏ is the period, . R is the radius, and .v̄s is the mean sound speed. Writing sound
speed in terms of density gives us,
√
γP
v =
. s (11.2)
ρ

where .γ is the specific heat ratio of the gas, . P is the pressure, and .ρ is the density.
Then, taking the mean density.ρ̄ = M/( 43 π R 3 ) (an approximation) and plugging this
back into Eq. 11.1 we quickly find

∏ ∝ R 3/2 M −1/2
. (11.3)

which indicates that pulsational period is strongly dependent on radius, but only
weakly dependent on mass.
To connect this more directly to the Period-Luminosity relation, we can rewrite
. R in terms of . L, the luminosity. We know that luminosity of a spherical blackbody
is given by . L = 4π R 2 σ Teff
4
. Assuming .Teff is fixed, then . L ∝ R 2 . Now we can write
Eq. 11.3 as
11 The Mira Distance Ladder 197

∏ ∝ L 3/4 M −1/2
. (11.4)

Luminosity is strongly dependent on mass, however mass is relatively insensitive

to luminosity. Thus, we would find .∏ ∝ L x , where .x ≤ 3/4. In practice, however,
. Teff is not constant and thus . x can be .>3/4. Nonetheless, this indicates that there
is a power law relationship between .∏ and . L. As a result of this, P–L relations are
typically shown and fit in log-log space, as a function of .log period and magnitude.
In order to compare theory with observations and determine the pulsational mode
of a star, pulsation is sometimes described using the pulsational constant . Q, which
is defined as:
( )1/2
M
. Q=∏ (11.5)
R3

Here we solved for the constant term in Eq. 11.3 which we previously ignored.
In practice, we typically cannot measure . R directly, so we substitute . R with . L
(luminosity) and .Teff (effective temperature) and get

3
M 1/2 Teff
. Q=∏ × 5.13 × 10−12 (11.6)
L 3/4
for .∏ and . Q in days and . M and . L in solar units. However, . Q values can be difficult
to calculate, because . L is the bolometric luminosity. For a Mira-like star with .Teff ∼
3000 K, . Q ∼ 0.1 days for fundamental pulsators and . Q ∼ 0.05 days for overtone
pulsators. In practice, some measured . Q values have been more than a factor of 2
off from the theoretical values. Wood [28] argued that the differences in theoretical
and measured . Q could potentially be attributed to metallicity (discussed further in
Sect. 11.4.4.1). However, the magnitude of metallicity dependence needed to account
for this difference would be quite significant: a difference of .0.4 mag between a solar
metallicity Mira one with metallicity comparable to the LMC. So far, such a strong
dependence has not been observationally confirmed.
This discrepancy likely lies in the pulsational models used. Recently, [29, 30] have
shown that linear calculations of overtone long-period variable sequences match well
with observations. However, nonlinear pulsational models of O-rich AGB stars are
required in order to reproduce the large-amplitude fundamental mode pulsations of
Miras. These nonlinear models provide significantly better agreement with obser-
vations, by reproducing the earlier onset of dominant mode pulsation and shorter
periods at larger radii that are present in observed P–L relations.
198 C. D. Huang

11.3.2 Ages and Initial Masses of Miras

Miras are the brightest stars in old or intermediate age populations. Similar to
other fundamentally-pulsating variable stars, younger (and more massive) Miras are
expected to have longer periods than older (and less massive) ones. This is known
as the period-age (P-A) relation, and has been empirically motivated by stellar kine-
matics [31]. Theoretical investigations of the P-A relation for LPVs have generally
been sparse due to the difficulty of modeling evolved red stars, but recent models
have shown agreement with observations [32]. There is also evidence to suggest that
C-rich Miras—which are confined to a smaller period range than O-rich Miras—
may be slightly more massive than O-rich Miras at the same period and are generally
younger stars.
Observational evidence comes from a range of studies in the Galaxy and Local
Group. O-rich Mira variables with log. P ∼ 2.0 − 2.5 days have been found in Galac-
tic globular clusters [33], which suggests that they have low initial mass. Comparisons
of O-rich Mira velocity dispersion with the age and velocity dispersion relations of
stars in the solar neighborhood indicates that stars with log . P < 2.3 days have initial
masses of .< 1MΘ . Most Galactic O-rich Miras have log. P ∼ 2.5 days and age .∼ 7
Gyr. More massive Miras with log . P ∼ 2.65 days are .∼ 3 Gyr.
C-Miras on the other hand, have mean log . P = 2.717, age of .∼ 1.8 Gyr, and
initial mass of .1.8 ∼ MΘ [34]. Feast [35] provides a more detailed analysis of the
relationship between ages, periods, and kinematics.

11.3.3 Variability Classification

Miras are a part of a broader group of variables known as long-period variables

(LPVs), which have periods ranging from .∼ 100 − 1000 days. Their P–L relations
were first studied by [36], nearly a hundred years ago. He compared the periods and
mean visual magnitudes of 10 LPVs with trignometric parallaxes. LPVs are very
common since nearly all stars on the red giant branch or AGB will exhibit some
form of variability and many of these will fall under the classification of LPV. As
noted by [37], more luminous variables are also typically more variable, therefore
the degree to which we can identify which stars are variable is dependent on the
photometric precision of our measurements.
LPVs can typically be grouped into one of three classes: Miras, semi-regular vari-
ables (SRVs), and slow irregular variables. As the names suggest, Miras are the most
regular pulsators amongst LPVs while semi-regular variables and irregular variables
often exhibit variations in their periodicity. Miras are typically distinguished from
SRVs by their larger amplitudes. A visual-band peak-to-trough amplitude larger than
2.5 mag within a single pulsational cycle and spectroscopic identification tradition-
ally served as the cutoff between Miras and SRVs [38]. Recently, however, it has
become more common to use only amplitude (often .ΔI > 0.8 mag, .ΔK > 0.4 mag)
11 The Mira Distance Ladder 199

for classification since spectroscopic followup is limited and expensive [39]. While
pulsational periods for Miras are generally very stable over many decades, with only
an estimated 1% of them showing variations [40], it is more common to see stochastic
variations in the cycle-to-cycle mean flux [41, 42]. Huang et al. [43] estimated this
to be .∼ 0.07 mag in . H -band.

11.3.4 LMC

The Mira variables in the Large Magellanic Cloud (LMC) are the best studied and
have informed our knowledge of Miras in environments where the data is sparser.
The LMC has been a critical galaxy both in the history and development of Mira
P–L relations and also in the extragalactic distance ladder, as an anchor galaxy (see
Sect. 11.4.1).
With a small number of Miras from the Large Magellanic Cloud (LMC), [44]
created the first P–L relation for Miras using bolometric magnitude. Subsequently,
much of the history and study of Miras was a direct result of the availability of
vast amounts of time-series photometry from microlensing surveys including the
Optical Gravitational Lensing Experiment (OGLE, [45]) and MACHO [46]. Using
MACHO data, [47] identified five parallel sequences (labeled . A − E) on the Period-
Luminosity plane, including Miras and SRVs. Miras were found to exist exclusively
on sequence .C and were confirmed to be radial, fundamental mode pulsators, now
in good agreement with theoretical models [30].
Later, [48] determined that the. K -band (2.2.µm) Mira P–L relation exhibits a small
scatter of .σ ∼ 0.13 mag. Using OGLE II and OGLE III time-series [45] and 2MASS
photometry, [49] found that sequence .C could further be separated into two – .C and
'
.C in the period-Wesenheit index plane. With spectroscopically-confirmed variables,
they were able to attribute this separation to O- and C-rich surface chemistries.
Whitelock et al. [50] examined the . K -band P–L relations for the O- and C-rich
subtypes separately and found a similar level of dispersion (.σ = 0.14 and 0.15 mag,
respectively). This dispersion is also roughly comparable to the scatter in the Cepheid
. K -band P–L relation (.σ = 0.09 mag, [51]).

Ita and Matsunaga [26] showed that P–L relations for O-rich Miras in the LMC
exist in a wide range of wavelengths, from . I -band to [24] while C-rich Miras typi-
cally only exhibit a P–L relation at longer wavelengths (see Fig. 11.3). Since then,
numerous other studies have examined P–L relations in the from optical ([24], using
maximum light) to mid-infrared [52]. . K -band is typically the bandpass of choice
for ground-based observations [53] thanks to the low intrinsic scatter of the P–L and
the Mira spectral intensity peaking around 1–2.µm. There is also less circumstellar
absorption in . K than at shorter wavelengths, and less circumstellar emission than at
longer wavelengths. Thus, NIR P–L relations are the most widely studied, including
[3, 26, 47, 50, 54].
200 C. D. Huang

Fig. 11.3 Figure 2 from [26]

showing the
period-magnitude relations
for OGLE III Miras in the
LMC. The red points
represent C-rich stars, while
black points represent O-rich
stars. The dashed and solid
blue lines represent
least-squares fits to the data.
Mean magnitudes for the
NIR and IR bands are
derived from single-epoch
observations, and thus show
larger scatter

11.3.5 M33

M33 is another Local Group galaxy that has been extensively analyzed in variability
studies and thus offers interesting comparisons between independent distance cali-
brators. There have been numerous Cepheid and TRGB distances to the galaxy, in
addition to two JAGB distance determinations. Yuan et al. [55], which focused on
the classification of the Miras in M33, the first to fit the Mira P–L using a quadratic
relation and [56] was the first to derive a distance to M33 using Miras with both
linear and quadratic P–L relations. The quadratric form of the P–L was empirically
motivated by the steeper slope and excess of above-PL stars at longer periods due to
hot-bottom burning. A quadratic functional form also allows Miras of all periods to
be fit simultaneously. However, this form has not been physically motivated. HBB
models generally still predict a linear relation at long periods and previous studies
had addressed this discontinuity by using two separate P–L relations.
In [55], Miras were classified using a Random Forest classifier and novel semi-
parameteric periodgram method developed and described in detail in [57]. . I −band
11 The Mira Distance Ladder 201

light curves obtained from the M33 Synoptic Stellar Survey were combined with near
and mid-infrared wavelength observations from the 3.8m UK InfraRed Telescope
(UKIRT), the 4-m May-all telescope at Kitt Peak National Observatory (KPNO)
and the 8-m Gemini North telescope. From this they obtained .μ M33 = 24.80 ± 0.06
mag.
Most recently, [58] used photometry from several surveys, including
PANSTARRS, PTF, and ZTF, to create well-sampled optical light curves and derive
a distance modulus of .24.67 ± 0.06 mag to the galaxy using O-rich Mira variables.
They fit both maximum and mean light P–L relations. Miras were classified using
their separation along the . J − K s period-color plane. The majority (96%) of periods
were in agreement with those determined in the previous study by [55].

11.3.6 NGC 5128

NGC 5128 (Centaurus A) is the nearest giant elliptical galaxy and the first galaxy
outside of the local group to have a Mira distance. It remains the only giant elliptical
galaxy with a Mira distance.
Rejkuba [59] determined the distance to NGC 5128 using both Mira variables
and infrared tip of the red giant branch (IR-TRGB) and found good agreement
between the two methods, with .μ Mira = 27.92 ± 0.19, .μT RG B K = 27.87 ± 0.16,
and .μT RG B H = 27.9 ± 0.2. Data used in the study taken using the ISAAC near-IR
array on ESO VLT and absolute calibration for the Miras came from Hipparcos par-
allaxes and the K-band Mira P–L relation in the LMC. Observations consisted of 20
epochs of . K s -band observations spanning 1197 d and 1 epoch each of the . Js and . H
bands. Periods, amplitudes, and mean . K s magnitudes were them calculated for the
variables with 10 or more measurements.
Only variables with regular periods, a sine-curve fit of.χ 2 < 5, and color bluer than
. Js − K s ≤ 1.4 were used in the determination of distance. In addition, stars that were

suspected to be aliased were not included in the fit. The mean log . P for the remaining
sample was 2.5 (.∼ 300 days). The color cut removed LPVs with more circumstellar
reddening and also matched the colors of the Miras used for the calibration of the Mira
P–L relation in the LMC (. J − K < 1.5 [3]). Rejkuba [59] considered the possibility
of hot-bottom burning in variables with . P > 400 days. However, in practice, they
also found that the distances determined with the full period range (up to just under
800 d at the longest end) and the shorter period variables only (. P < 400 days) were
nearly the same (differing by a few hundredths of a magnitude and well within the
uncertainties). This could indicate that many stars, even at longer periods, are not
actively undergoing hot-bottom burning.
202 C. D. Huang

11.3.7 Dwarf Galaxies

P–L relations also exist for some local group dwarf irregular galaxies, including
NGC 6822 [60] and IC 1613 [61], which have distances determined using C-rich
Miras. Perhaps most importantly for distance measurements, dwarf galaxies span a
range of metallicities and star formation histories and thus are ideal environments for
studying the relationship between stellar populations and galaxy evolution. While
the C/O ratio is known to vary strongly with metallicity [62], studies of the AGB P–L
relations in dwarf galaxies with a wide range of metallicities have been instrumental
in showing that there is no measurable metallicity dependence for the P–L relation
[63]. A more extensive review can be found in [64].

11.4 Measuring the Hubble Constant

Hubble’s law states that:

v = H0 × D
. r (11.7)

where .vr is the recessional velocity and . D is distance. Hubble’s Constant can be
thought of as the constant of proportionality of .vr and . D. At large distances (.z > 1),
this will relationship will begin to deviate from linearity depending on the cosmology
of the universe. For nearby galaxies, local gravitational interactions can contribute
significantly to or dominate the observed velocity.
Galaxies used to measure H.0 must be within the “Hubble flow". In other words,
they must be in the regime where their recessional velocities are due almost entirely
to the expansion of the Universe. In practice, this often means galaxies distances. D >
100 Mpc, or redshifts .z > 0.02. Peculiar velocity corrections are typically applied
to account for any remaining effects due to local gravitational interactions.
However, geometric methods generally cannot reach galaxies in the Hubble flow.
Even using Miras as secondary distance indicators, only the most luminous and
massive stars would be able to reach these distances directly and their P–L relations
are less well-understood and studied (see Sect. 11.5). Thus, a typical path towards
measuring H.0 consists of a three-rung ladder. The first rung uses nearby (in the case of
Miras, . D < 7.5 Mpc) geometric anchors which determine the absolute magnitude of
the Mira P–L relation. Short-period Miras (typically hundreds can be discovered per
galaxy) are then used to reach .< 40 Mpc distances, enough to measure the distances
to and calibrate the peak luminosities of the nearest SN Ia, which are relatively rare
(occurring at a rate of approximately 1 per galaxy per century). Thus, Miras serve as
a bridge between the distances reachable by most geometric methods and the SN Ia
in the Hubble flow that are capable of measuring H.0 directly.
In the low-redshift (.z ∼ 0) limit, the intercept of the Hubble diagram .a B is given
by,
11 The Mira Distance Ladder 203

a = log cz − 0.2m 0B
. B (11.8)

where .m 0B is the standardized (corrected for variations from fiducial color, luminos-
ity, or host dependence) maximum-light apparent magnitude of the SNe. From this
equation, we can see that .a B , which is determined solely from the third rung of the
distance ladder, does not depend on any Mira distance measurements.
In order to solve for H.0 , we combine the absolute SN Ia peak magnitude along
with the intercept for the SN Ia Hubble diagram to write,

. log H0 = (M B0 + 5a B + 25)/5 (11.9)

where . M B0 is the fiducial SN Ia absolute magnitude. We determine . M B0 by measuring

the distance to local SN Ia host galaxies using Miras (or other primary distance
indicators), and using the distance to convert the observed apparent SN Ia magnitudes
(.m 0B ) to absolute magnitudes (. M B0 ).

11.4.1 Anchors

The first rung of a cosmic distance ladder is often referred to as an anchor because it
sets absolute scale for the later rungs using geometry. The most obvious way to obtain
a geometric distance is through parallax. However, in the case of highly-extended
and evolved stars like Miras the current consensus is that parallax measurements
of AGB using Gaia have large, asymmetric, and underestimated uncertainties and
should be treated with caution (see Sect. 11.4.4.2 for a more in-depth discussion).
Thus, we currently do not recommend using parallax measurements for calibrating
the Mira H.0 measurement.
Instead, since Miras are a ubiquitous, intermediate-age population, they can be
discovered, and subsequently calibrated in almost any galaxy with a geometric dis-
tance. Mira-based H.0 measurements [65] rely primarily on geometric measurements
to two anchor galaxies—the LMC and NGC 4258 [43].

11.4.1.1 NGC 4258

NGC 4258 is a relatively nearby water megamaser host galaxy which has a geometric
distance of . D = 7.576 ± 0.082 (stat.).±0.076 (sys.) Mpc with 1.4% precision [66].
Angular diameter distances to megamasers are obtained by combining spectral mon-
itoring of the Keplerian motion of the masing clouds orbiting the black hole, along
with VLBI mapping and a model of the maser disk geometry. Given that the galaxy
is quite close, NGC 4258s peculiar motion is likely to contribute significantly to its
recessional velocity. Thus, while it is unsuitable for direct H.0 measurements, it has
served as an important anchor galaxy for Cepheids, TRGB, and Miras.
204 C. D. Huang

[43] first calibrated the distance to this galaxy using a quadratic P–L relation based
on [54] and the 2.6% distance to the galaxy from [67, 68]. For P–L relation,

m F160W = a0 − 3.59(log P − 2.3) − 3.40(log P − 2.3)2

. (11.10)

where .m F160W is the magnitude in F160W, .a0 is the zeropoint, and . P is period, in
days. [43] obtained .a0 = −6.15 ± 0.09 mag as the absolute mean magnitude of a
200-day O-rich Mira using the quadratic relation. Huang et al. [65] later updated this
with the improved distance from [66] and used two linear P–L relations,

.m F160W = a0 − 3.64(log P − 2.3) (11.11)

.m F160W = a0 − 3.35(log P − 2.3) (11.12)

Eq. 11.11 used the linear slope derived for the . H −band from [54] while Eq. 11.12
used a color transformation from . H to F160W derived using O-rich Mira spectra to
solve for a new slope. This resulted in .a0 = −6.21 ± 0.042 mag for Eq. 11.11 and
.a0 = −6.25 ± 0.042 mag for Eq. 11.12.
The data for the Mira observations were taken using the HST WFC3/IR channel
in two filters F125W (HST wide . J ) and F160W (HST wide . H ) in GO-13445 (PI:
Bloom). Twelve epochs, spaced approximately monthly, were obtained of a single
field over the course of a year. While the colors were originally intended to aid in
separating the Miras into their O- and C-rich spectral subtypes, the color separation
was found to be poor. Therefore, additional measures were taken to identify the two
classes without using color information. Given that C-rich variables are rare at the
shortest periods and also tend to have larger amplitudes, the sample was limited to
. P < 300 days and .0.4 mag .< ΔF160W < 0.8 mag to reduce C-rich contamination.
Of the 438 Mira candidates, 139 remained in the most stringent “gold" sample.
For the gold sample, optical data from a previous Cepheid observing campaign [69]
was used to verify that the objects were variable in the F814W (HST wide . I ) time-
series. In addition, detection in F814W was also taken as a proxy for color since
the redder C-rich Miras were expected to be difficult to detect in the Cepheid data.
However, despite these additional requirements, the zeropoints obtained for the full
sample and gold sample were within 0.02 magnitudes, well within the uncertainties.
Finally, the relative distance between NGC 4258 and the LMC was calculated after
applying a Mira-specific color correction to ground-based . H band P–L relation. This
was found to be in agreement with the geoemtric distances for the two galaxies.

11.4.1.2 LMC (Again)

Huang et al. [65] used the 1.2% [70] distance to the LMC of .μLMC = 18.477 ±
0.004(stat) ± (sys) mag to calculate an absolute calibration of .a0 = −6.27 ± 0.03
mag for Eq. 11.12, consistent to within the uncertainties with the zeropoint of the
Mira PLR from NGC 4258.
11 The Mira Distance Ladder 205

While LMC Miras were not directly observed with HST WFC3/IR, using the
OGLE-III sample of O-rich Mira variables [39, 43] determined the relative distance
modulus between the LMC and NGC 4258,

μN4258 − μLMC = 10.95 ± 0.01(stat) ± 0.06(sys) mag.

. (11.13)

which is consistent with the relative Cepheid distance between these two galaxies.
In order to directly compare the two samples without observations in the same
bandpasses, [43] derived an . H to . F160W transformation of

. F160W = H + 0.39(J − H ) (11.14)

using O-rich Mira spectra from [71] as input into PySynPhot [72].
In addition, the sample of Mira variables from OGLE-III also included variables
with a wider range of amplitudes and periods than those used in the NGC 4258
sample. Thus, the OGLE-III sample was restricted to only O-rich Miras with .240 <
P < 400 days and .0.4 < ΔH < 0.8 mag to provide a closer match to the HST
sample for the calibration.

11.4.2 SN Ia Calibrators

The second rung and largest contributer to the overall H.0 error budget is the calibration
of the SN Ia peak luminosity with Miras in local SN Ia host galaxies. To date, only
two SN Ia have had luminosities calibrated with Mira distances.

11.4.2.1 NGC 1559

Huang et al. [65] determined the distance modulus to NGC 1559, the host of SN
2005df to be .μ N 1559 = 31.41 ± 0.050(stat) ± 0.060(sys) using a sample of 115
Miras. At .19.1 ± 1.1 Mpc, NGC 1559 is the most distant galaxy in which Miras
have been studied. The distance for this galaxy was used by [65] to determine a fidu-
cial SN Ia luminosity of . M B0 = −19.27 ± 0.13 mag, and in agreement with the [73]
Cepheid measurement of . M B0 = −19.360 ± 0.106 mag for this galaxy. Using both
the NGC 4258 and LMC P–L relations as anchors, this results in H.0 = 73.3 ± 4.0
km s.−1 Mpc .−1 . The methodology laid out in this paper is the blueprint for measuring
Miras in other local SN Ia hosts. However, because this H.0 measurement uses only
one SN Ia calibrator, the 5% uncertainty is dominated by the statistical uncertainty
in the peak magnitude of the single SN Ia and the result is within 2-.σ agreement
with both Planck and the most precise Cepheid measurement.
Data used for the NGC 1559 study consisted of 10 epochs of WFC3/IR channel
F160W observations (GO-15145; PI: Riess) with roughly monthly spacing and a
baseline slightly longer than 1 year. Unlike most Mira studies, the NGC 1559 Mira
206 C. D. Huang

sample did not use color as a selection criterion since [43, 74] showed that HST
wide-band colors are an ineffective discriminant of both O- and C-rich AGB stars in
general and Miras in particular. Instead, amplitude and period cuts were used to limit
the C-rich contamination. The remaining C-rich contamination (estimated to have
an bias of .∼ −0.07 mag in the worst-case) was estimated by using the LMC Mira
sample to model a pure O-rich sample and a C-rich contaminated sample. Because
C-rich Miras are expected to be fainter in F160W than O-rich Miras and have a
different period distribution than O-rich Miras, the shape of the zeropoint-vs-period
curve (. Z (P)) is different for a pure O-rich sample and a sample contaminated with
C-rich Miras. The contaminated sample is then scaled to fit the . Z (P) relation of
NGC 1559, yielding a correction of .−0.057 ± 0.024 mag and no correction when
applied to NGC 4258. This method was tested on the M33 Mira sample from [56] by
creating a contaminated population and then correcting the zeropoint. The correction
of .−0.04 ± 0.01 mag was found to be in agreement with the true value of .−0.034
mag.

11.4.2.2 M101

M101 is the most recent SN Ia host galaxy to have a distance measurement using
Miras (Huang+2023). 288.Ω-rich Mira candidates were discovered in the field of SN
2011fe using a long baseline of HST observations, spanning .∼ 2900 days. Many of
the observations were originally taken to the study the late-time lightcurve of the SN.
From the sample of Miras with periods between 260 and 400 d, the distance modulus
to M101, using the LMC and NGC 4258 as anchors, was determined to be .μM101 =
29.10 ± 0.06 mag. In combination with the SN Ia calibration from NGC 1559, this
yielded . M B0 = −19.27 ± 0.09 mag. The resulting H.0 measurement combining both
SN Ia calibrators is .73.14 ± 2.98 km s.−1 Mpc.−1 . H.0 = 72.37 ± 2.97 kms.−1 Mpc.−1 ,
and confirmed previous indications that the local universe value of . H0 is higher
than the early-universe value at .∼ 95% confidence. This 4% measurement is still
dominated by the statistical uncertainty in the peak luminosity of the two SN Ia
calibrators.
As the nearest SN Ia host galaxy with modern (CCD) photometry, the distance
to this galaxy has also been measured numerous times using Cepheids and TRGB,
offering a opportunity for comparison with the two most commonly-used primary
distance indicators. The recent Mira measurement falls within 1-.σ of most other
recent measurements with Cepheids and TRGB.

11.4.3 Methodology

The methodology we describe here applies primarily to the targeted Mira searches
in galaxies calibrated using sparse HST data. Ground-based observations have used
diverse observing strategies and typically prefer . K -band for P–L relations. How-
11 The Mira Distance Ladder 207

ever, Mira H.0 measurements have necessitated the use of HST to resolve stellar
populations in SN Ia host galaxies and therefore use a slightly shorter bandpass
– WCF3/IR’s F160W (.∼ 1.6μm). This is the longest wavelength wide-band filter
available with HST and also coincides well with the peak of a typical Mira’s spectral
energy distribution.

11.4.3.1 Observations

Observational baselines for Miras span around one year or slightly longer and typi-
cally consist of .∼ 10 epochs, with roughly monthly spacing. To reduce aliasing, it is
generally preferable to space the observations such that the Miras at the target period
ranges (usually .∼ 100 − 400 days) will be sampled a minimum of three phase points
in one cycle.
The initial steps of the variability search are similar to the process for Cepheids
described in several S. H0 ES papers including [75–77]. Once the data is collected, a
variability search is performed to cull as many nonvariable objects as possible while
retaining the variables. Crowded-field photometry is used to create lightcurves for the
candidates, which are then cut down to a sample of O-rich Mira candidate variables
using selection criteria.
The major difference in the analysis process between Cepheids and Miras lies in
the wavelengths used for determining the positions for the variables. Miras are very
red stars and therefore typically not observed in optical wavelengths. With Cepheids,
the finer-resolution optical observations are generally used to classify the stars and
pinpoint their locations. However, given that they are an older population, Miras are
generally found in less dusty areas than the younger Cepheids, and thus this entire
search is performed only in the NIR.

11.4.3.2 Selection Criteria

Selection criteria are used to separate the Miras from SRVs and other potential
contaminants. Unlike Cepheids, which have a distinctive sawtooth shape in optical
wavelengths and are often identified on the basis of their fit to templates, there
currently are no templates for Miras and their light curve shapes of Miras in optical
wavelengths. However, even in the optical, most (.∼ 80%) short-period Miras do not
show substantial deviations from a symmetric light curve [78], and .∼70% of them
do not significantly deviate from a purely sinusoidal shape [79].
Similarly to Cepheids, the lightcurves of Miras in NIR are even more sinusoidal
(see Fig. 11.4 showing composite light curves at a range of periods). Therefore, the
period, amplitude, and mean magnitude of Miras are typically fit using only a sine
function. When [43] applied a Bayesian information criteria before adding additional
harmonics, they found that .∼ 99% of Miras favored a first order fit.
Amplitude cuts are used to remove SRVs or potential C-rich Miras (which often
have larger amplitudes). An initial period cut is also typically applied to remove
208 C. D. Huang

Fig. 11.4 Figure 3 from [80] showing the stacked lightcurves of .∼ 300 Miras in the SN Ia host
galaxy M101. Individual observations are shown with their fit phase, mean magnitude is subtracted,
but amplitudes are not scaled

Table 11.1 Mira selection criteria

M101 NGC 1559 NGC 4258 (gold)
Period cut 200 d .< P < 500 d .240 d .< P < 400 d .P < 300 d
Amplitude cut .0.4 mag < .0.4 mag < .0.4 mag <
ΔF160W < 0.8 mag ΔF160W < 0.8 mag ΔF160W < 0.8 mag
Surface brightness cut – 421 counts/second –
Color Cut .m F110W − m F160W < – .m F125W − m F160W <
1.3 mag 1.3 mag
F814W Detection – – Slope-fit to F814W
data .> 3σ
F814W Amplitude – – .Δ F814W .> 0.3 mag

.χs
. F-statistic
: 2 /χ 2 < 0.5 .χs
2 /χ 2 < 0.5 –
l l
A comparison of the criteria for the final Mira sample in M101 with those used for NGC 1559 and
NGC 4258. Owing to differences in available data, we were unable to match the criteria exactly
between all of the observations

objects with periods that are either too short (.<100 days) or more likely to be C-rich
or hot-bottom burning (.>300–400 days). A summary and comparison of the cuts
used in a few recent studies is shown in Table 11.1.
At this stage, the remaining objects are likely to be Miras, but further corrections
may be applied to remove objects that are found to be particularly crowded (either
through surface brightness cuts or cuts on the crowding corrections) or below the
completeness limit and thus may bias the zeropoint of P–L relation. C-rich contam-
ination corrections are also typically applied, with [65] estimating the level to be a
few hundredths of a magnitude.
11 The Mira Distance Ladder 209

11.4.4 Other Considerations

11.4.4.1 Metallicity

As touched upon in Sect. 11.3.1, theory suggests that the Mira P–L relation may
have a dependence on metallicity. Wood [28] noted that if two stars with the same
mass but different metallicities begin to pulsate at the same radius (and thus, period),
they will exhibit different luminosities. However, this assumes that these stars would
become unstable to Mira-like pulsations at the same radius. It is also possible that
Miras with the same mass but different metallicity would begin to pulsate at different
radii, which may have the effect of mitigating the change in luminosity. Thus far, the
agreement between theory and observations of Miras in a wide range of metallicities
suggests that the dependence of the zeropoint of the P–L on metallicity is weak. For
a more detailed explanation, see [81].
As discussed in Sect. 11.3.7, dwarf galaxies are the ideal testing grounds to under-
stand the connection between galaxy evolution and stellar populations. Most recently,
[63] examined the IR P–L relations of AGB stars in environments with metallicity
ranging from 1.4% solar metallicity ([Fe/H] = -1.85) to 50% solar metallicity ([Fe/H]
= -0.38) and found no discernible difference in the pulsational properties of stars in
metal-poor and metal-rich environments. Similarly, [24] also found that the mean
. K s band O-rich Mira measurement between the LMC and SMC showed no sign of

a metallicity effect.
Other works have put tentative upper limits on the metallicity effect by showing
that agreement in Mira P–L relations without metallicity corrections is within a few
hundredths of a magnitude in the NIR. A study of the . K -band P–L relation of Miras
in the LMC, Galactic globular clusters, and Galaxy [82] showed agreement to within
the measurement uncertainties. Whitelock et al. [50] found an upper limit of .∼ 0.1
mag in the zeropoint of LMC and Galactic O-rich Miras with Hipparcos and VLBI
parallaxes and those in globular clusters. However, when [43] used the same data,
but an updated LMC distance from [70] this discrepancy dropped to 0.01 mag. For
HST F160W, the wide-band filter used for H.0 measurements, [43] found .∼ 0.03
mag agreement (within errors) in the relative distance modulus between the LMC
and NGC 4258 determined using Miras that from Cepheids and from geometry.
Together, this suggests that the AGB P–L relations are a reliable tool for determine
distances regardless of metallicity, at least down to the .∼ 0.03 mag level in the NIR.

11.4.4.2 Parallaxes

While other distance indicators such as Cepheids or RR Lyrae have been calibrated
using Gaia parallaxes, parallax measurements are difficult to obtain for large, red,
evolved stars such as Miras using an optical telescope like Gaia. Typical radii for
Miras range from .250RΘ − 800RΘ [83]. Xu et al. [84] found that the redder the
star, the larger parallax uncertainties in Gaia DR2. Red stars, including AGB stars,
210 C. D. Huang

typically have more complex surface dynamics, larger physical size, and more dust
[85]. Empirically, the Galactic Mira PLR using Gaia shows large scatter.
Gaia parallax uncertainties for Miras and AGB stars in general are both underesti-
mated (up to a factor of 5.44 for sources with .G < 8 mag) and asymmetrical. Andri-
antsaralaza et al. [86] caution against using parallaxes when uncertainties are larger
than 20% and employing the Gaia catalogue parameter RUWE (re-normalised unit
weight error) as the sole measure of quality of Gaia astrometric data for AGB stars.
They were able to obtain distance measurements to AGB stars using maser obser-
vations with very long baseline interferometry (VLBI). Unfortunately, this method
can only measure distances to Miras with substantial circumstellar envelopes, and is
typically not applicable to the Miras currently used for the distance ladder. However,
it is appropriate as a cross-check to the AGB distances obtained by Gaia.
Difficulties in parallax measurements with Miras were not unexpected and sev-
eral factors—both astrophysical and instrumental, contribute. Physically, Mira radii
are larger than 1 AU—in other words, their angular diameters are larger than their
parallaxes. Both angular size and parallax scale as .1/d, where .d is distance to the
star. Thus, the physical size of a star is a problem whether the Mira is located nearby
or far away. On its face, this alone does not appear to be particularly problematic
for parallax measurements since an astrometric solution can be found using only
on the motion of the photocenter of the extended source. However, Miras are not
uniformly-illuminated discs. Moreover, they have complex stellar surface dynamics
caused by large convective cells that can result in the motion of their stellar photo-
centers [87]. Bright sources, including Galactic Miras, can also reach instrumental
saturation which results in less accurate astrometry [88].
Hipparcos had more success obtaining parallaxes for Miras. Whitelock and Feast
[89] examined subsets of the data and found that the best solution excluded short-
period red group stars and also low-amplitude variables. Whitelock et al. [50] used
the revised Hipparcos parallaxes from [90] data to obtain a calibration for Galactic,
O-rich Mira variables in the . K band of .δ = −7.25 ± 0.07 mag.

11.5 Long-Period Miras

The long-period end of the Mira distribution is less studied than the short-period
objects discussed throughout the this review for a number of reasons. In addition
to more complicated physics (including the effects of hot-bottom burning, heavier
mass-loss, and larger circumstellar envelopes compared to short-period Miras), they
require extended observational baselines to accurately determine their periods and
are formed from stars of higher initial mass and are therefore considerably rarer,
younger, and more likely to be located closer to dusty regions where they were born
than short-period Miras. Additionally, OH/IR stars, which are bright infrared sources
with Type II OH masing emission and intermediate initial mass, fall into this category
[91] and some of these extreme objects have periods as long as 1800 d. Long-period
Miras also often follow a different P–L relation than their shorter-period counterparts.
11 The Mira Distance Ladder 211

Despite these difficulties, there remains great interest in studying long-period

Miras both for distance ladder applications and as tracers of structure [92]. Assuming
the same P–L slope as the shorter-end, their . F160W luminosities will start at .−7.3
mag on the low (.400 days) end. As seen in Fig. 11.3, the slope for these longer-period
objects can often be steeper than at the short-period end, likely due to hot-bottom
burning [4]. Thus, the brightest objects can be up to 2.5 magnitudes more luminous
and reach distances.>50 Mpc with Hubble Space Telescope. Long-period objects also
appear to have tighter P–L relations in the mid-infrared which may be particularly
advantageous with James Webb Space Telescope observations.
Very long period (. P > 1000 day) Miras have been discovered locally in the SMC,
LMC, and Galactic Bulge by OGLE [37, 39, 93] and in the nearby galaxies Sgr dIC
and NGC 3109 [94, 95] using the . J H K s bands from the InfraRed Survey Facility
(IRSF). The SPIRITS [96, 97] and DUSTiNGS [2, 63, 98] collaborations have also
found hundreds of long-period variables using 3.6.µm and 4.5.µm that are likely
Miras in a wide range of host environments using Spitzer Space Telescope some of
them with . P ∼ 1000 days.

11.6 Comparison with Other Indicators

While there are numerous excellent stellar distance indicators, we will limit this
comparison to those that have currently been shown to be capable of bridging the
gap between nearby, geometric measures and the local host galaxies of SN Ia. This
includes Cepheids, the Tip of the Red Giant Branch (TRGB), J-Region Asymptotic
Giant Branch (JAGB) stars [99], and short-period Miras and typically excludes fainter
stellar indicators such as RR Lyrae and Type II Cepheids. Long-period Miras are quite
different, so for brevity when we refer to Miras we refer exclusively to short-period
O-rich Miras (Table 11.2).
Luminosity: Cepheids and short-period Miras are roughly comparable in luminosity
the NIR. Both are .> 104 L Θ in H-band. JAGB (which includes C-rich Miras) are
brighter than the shortest-period Miras (. P < 200 days), but typically fainter than
the brightest Miras. As a result, despite requiring time-series photometry, Miras
and Cepheids can be more efficient than TRGB and JAGB larger distances (e.g. . D >
10 − 15 Mpc). It is necessary to resolve stars.0.5 − 1.0 mag below the tip/main clump
for TRGB and JAGB measurements. On the other hand, peak-to-trough amplitudes
of Cepheids and Miras generally span .< 0.3 mag and .0.4 − 0.8 mag respectively,
and must also be detected in order to determine their mean magnitudes.
Age: Cepheids are young, massive stars, typically found clustered near dusty spiral
arms and star-forming regions and are on average more crowded than the older
indicators. They are generally found in the disk of spiral galaxies. JAGB are C-
rich stars and trend younger than O-rich Miras, but are still found in many galaxy
morphologies. TRGB are a ubiquitious older population and are best observed in the
halos of galaxies [102, 103].
212 C. D. Huang

Table 11.2 A comparison of the stellar distance indicators that have been shown to be capable of
reaching SN Ia host galaxy distances. For the sake of brevity, Miras refers to O-rich, short period
(. P < 400 days) Miras
Indicator Age Time-series mag. H Optical data
Miras: .2 − 8 Gyr Yes.a .< −6.25b No
Cepheids: .< 100 Myr Yes.c .−6.1d Yes
TRGB: .> 4 − 5 Gyr No – Yes
IR-TRGB: .> 4 − 5 Gyr No .[−5.5 to − 6.5]e No
JAGB: 300 Myr–1 Gyr No .−6.2[J ], −6.8[H ] f No
A comparison of the stellar distance indicators that have either been shown (Cepheids, TRGB,
Miras) or are thought to be capable of reaching SN Ia host galaxy distances (JAGB). For the sake
of brevity, Miras refers to O-rich, short period (. P < 400 days) Miras.a timescale .>1 year; .b for
. P > 200 d; . timescale .> 60 days; . for . P > 15 d; . IR-TRGB luminosity depends strongly on
c d e

metallicity; . f JAGB is most-commonly studied in the . J −band (calibration from [100]) but a recent
work by [101] provided a number of . H -band absolute magnitude calibrations

Time-Series Requirements: A major advantage that TRGB and JAGB have over
Cepheids or Miras is that they do not require time-series observations since both are
identified using their positions on a color-magnitude diagram (CMD). This makes
TRGB and JAGB more efficient for nearby galaxies . D < 10 − 15 Mpc, since only
a single epoch of observation is necessary.

11.7 Conclusions and Future Prospects

As the most luminous of the intermediate-to-old stellar distance indicators, Miras

are in a unique position to expand the sample of local SN Ia with calibrated peak
magnitudes and are also the only distance indicator capable of cross-checking all
of the others. They will be particularly exciting in the upcoming decade with the
increase focus on NIR and IR observations.
Presently, due to the small number of SN Ia calibrators, the uncertainty in the
Mira-based H.0 is dominated by the statistical uncertainty in the peak magnitude of
SN Ia. The best path forward to improving H.0 measurements with Miras is to increase
the number of SN Ia that have been calibrated with Miras. This can be done either
by revisiting the local host galaxies that already have Cepheid distances—allowing
for a direct comparison of Mira and Cepheid distances—and by searching for them
in the galaxies of SN Ia that have not yet been calibrated with other indicators due
to their distance, or lack of star formation.
James Webb Space Telescope (JWST), while not optimally suited for obtain-
ing time-series, is able to photometrically separate O- and C-rich Miras in SN Ia
host galaxies with medium-band filters and will also be able to un-crowd the back-
grounds of previously-observed variables, similar to what has currently been done
for Cepheids [104]. JWST also has the sensitivity to measure a two-rung H.0 using
11 The Mira Distance Ladder 213

geometry and Miras only—without the inclusion of SNe Ia. However, it suffers from
large overheads that make multi-epoch visits undesirable.
The upcoming Vera C. Rubin Observatory’s Legacy Survey of Space and Time,
will cover nearly half of the sky in .u, g, r, i, z, and . y filters. Rubin has the potential
to be particularly useful in developing the first rung of the Mira distance ladder, since
it is expected to discover variables in galaxies out .< 15 Mpc [105]. While this is
too small of a volume to substantially increase the number of SN Ia hosts that can
be calibrated with Miras, it is possible that some of these galaxies can eventually
become additional geometric anchors. Miras will be brightest in the .r, i, and .z bands,
and most will be discovered in the .i band. As previous optical studies have found,
maximum-light Period-Luminosity Relations may be more useful and have smaller
scatter than mean-light relations for these shorter wavelengths.
The Nancy Grace Roman Observatory will have a similar sensitivity and resolution
to Hubble, but will cover a much larger field-of-view. This will allow it to continue
to study Miras at the distances of SN Ia hosts. In addition, it has a redder, . K -band
equivalent that may produce tighter P–L relations for Miras than the HST F160W
(H) band currently used in the distance ladder. Given its the large field of view, many
Local Group Miras are also likely to fall into survey regions, potentially providing
NIR data for previously-identified stars.

Acknowledgements I would like to thank Eleonora di Valentino and Dillon Brout for inviting me
to write this review chapter and for their hard work in assembling this textbook. I also thank Morgan
MacLeod, John Menzies, Patricia A. Whitelock, Louise Breuval, and Andrea Sacchi for their helpful
comments and discussions regarding this review. Additionally, I thank Morgan MacLeod for his
help in the creation of Fig. 11.1.

References

1. N.N. Samus’, E.V. Kazarovets, O.V. Durlevich, N.N. Kireeva, E.N. Pastukhova, General
catalogue of variable stars: version GCVS 5.1. Astron. Rep. 61, 80–88 (2017)
2. M.L. Boyer, K.B.W. McQuinn, P. Barmby, A.Z. Bonanos, R.D. Gehrz, K.D. Gordon, M.A.T.
Groenewegen, E. Lagadec, D. Lennon, M. Marengo, M. Meixner, E. Skillman, G.C. Sloan,
G. Sonneborn, J.T. van Loon, A. Zijlstra, An infrared census of dust in nearby galaxies with
spitzer (dustings) I overview. ApJS 216, 10 (2015)
3. M.W. Feast, I.S. Glass, P.A. Whitelock, R.M. Catchpole, A period-luminosity-colour relation
for Mira variables. MNRAS 241, 375–392 (1989)
4. P.A. Whitelock, M.W. Feast, J.T. van Loon, A.A. Zijlstra, Obscured asymptotic giant branch
variables in the Large Magellanic Cloud and the period-luminosity relation. MNRAS 342,
86–104 (2003)
5. A. Dotter, MESA isochrones and stellar tracks (MIST) 0: methods for the construction of
stellar isochrones. ApJS 222, 8 (2016)
6. J. Choi, A. Dotter, C. Conroy, M. Cantiello, B. Paxton, B.D. Johnson, Mesa isochrones and
stellar tracks (MIST) I. Solar-scaled Models. ApJ 823, 102 (2016)
7. B. Paxton, P. Marchant, J. Schwab, E.B. Bauer, L. Bildsten, M. Cantiello, L. Dessart, R.
Farmer, H. Hu, N. Langer, R.H.D. Townsend, D.M. Townsley, F.X. Timmes, Modules for
experiments in stellar astrophysics (MESA): binaries, pulsations, and explosions. ApJS 220,
15 (2015)
214 C. D. Huang

8. S. Höfner, H. Olofsson, Mass loss of stars on the asymptotic giant branch: mechanisms,
models and measurements. A & A Rev 26, 1 (2018)
9. J. Iben, Asymptotic giant branch stars: thermal pulses, carbon production, and dredge up;
neutron sources and s-process nucleosynthesis, in Evolution of Stars: The Photospheric Abun-
dance Connection, ed. by G. Michaud, A.V. Tutukov, vol. 145, p. 257 (1991)
10. C.A. Frost, J.C. Lattanzio, On the numerical treatment and dependence of the third dredge-up
phenomenon. ApJ 473, 383 (1996)
11. P.R. Wood, Final stages of AGB evolution (Invited Review), in Planetary Nebulae, ed. by
H.J. Habing, H.J.G.L.M. Lamers, vol. 180, p. 297 (1997)
12. M. Busso, R. Gallino, G.J. Wasserburg, Nucleosynthesis in asymptotic giant branch stars:
relevance for galactic enrichment and solar system formation. ARA&A 37, 239–309 (1999)
13. F. Herwig, Evolution of Asymptotic Giant Branch Stars. ARA&A 43, 435–479 (2005)
14. A.I. Karakas, J.C. Lattanzio, The Dawes review 2: nucleosynthesis and stellar yields of low-
and intermediate-mass single stars. PASA 31, e030 (2014)
15. M. Schwarzschild, R. Härm, Thermal instability in non-degenerate stars. ApJ 142, 855 (1965)
16. J. Iben, On the interior properties of red giants, in Physical Processes in Red Giants, ed. by
J. Iben, A. Renzini. Astrophysics and Space Science Library, vol. 88 (1981), pp. 3–24
17. A.I. Karakas, J.C. Lattanzio, O.R. Pols, Parameterising the third dredge-up in asymptotic
giant branch stars. PASA 19, 515–526 (2002)
18. T. Bloecker, D. Schoenberner, A 7M AGB model sequence not complying with the core
mass-luminosity relation. A&A 244, L43–L46 (1991)
19. J.C. Lattanzio, Hot bottom burning in a 5 solar mass model. PASA 10, 120 (1992)
20. A.I. Boothroyd, I.J. Sackmann, G.J. Wasserburg, Hot bottom burning in asymptotic giant
branch stars and its effect on oxygen isotopic abundances. ApJL 442, L21 (1995)
21. P. Marigo, A. Bressan, A. Nanni, L. Girardi, M.L. Pumo, Evolution of thermally pulsing
asymptotic giant branch stars-I the COLIBRI code. MNRAS 434, 488–526 (2013)
22. A.I. Boothroyd, I.J. Sackmann, Breakdown of the core mass-luminosity relation at high lumi-
nosities on the asymptotic giant branch. ApJL 393, L21 (1992)
23. P.A. Whitelock, J.W. Menzies, M.W. Feast, N. Matsunaga, T. Tanabé, Y. Ita, Asymptotic giant
branch stars in the Fornax dwarf spheroidal galaxy. MNRAS 394, 795–809 (2009)
24. A. Bhardwaj, S. Kanbur, S. He, M. Rejkuba, N. Matsunaga, R. de Grijs, K. Sharma, H.P. Singh,
T. Baug, C.-C. Ngeow, J.-Y. Ou, Multiwavelength period-luminosity and period-luminosity-
color relations at maximum light for mira variables in the magellanic clouds. ApJ 884, 20
(2019)
25. C.C. Ngeow, J.Y. Ou, A. Bhardwaj, J. Purdum, B. Rusholme, A. Wold, Zwicky transient facil-
ity and globular clusters: the gr-band period-luminosity relations for mira variables at maxi-
mum light and their applications to local galaxies (2023). arXiv e-prints: arXiv:2307.06749
26. Y. Ita, N. Matsunaga, Period-magnitude relation of Mira-like variables in the Large magellanic
cloud as a tool to understanding circumstellar extinction. MNRAS 412, 2345–2352 (2011)
27. H.S. Leavitt, E.C. Pickering, Periods of 25 variable stars in the small magellanic cloud.
Harvard College Obs. Circ. 173, 1–3 (1912)
28. P.R. Wood, Mira variables: pulsation, mass loss and evolution, in Confrontation Between
Stellar Pulsation and Evolution, ed. by C. Cacciari, G. Clementini. Astronomical Society of
the Pacific Conference Series, , vol. 11 (1990), pp. 355–363
29. M. Trabucchi, P.R. Wood, J. Montalbán, P. Marigo, G. Pastorelli, L. Girardi, A new interpre-
tation of the period-luminosity sequences of long-period variables. ApJ 847, 139 (2017)
30. M. Trabucchi, P.R. Wood, N. Mowlavi, G. Pastorelli, P. Marigo, L. Girardi, T. Lebzelter,
Modelling long-period variables–II: fundamental mode pulsation in the non-linear regime.
MNRAS 500, 1575–1591 (2021)
31. M.W. Feast, The long period variables. MNRAS 125, 367 (1963)
32. M. Trabucchi, N. Mowlavi, The period-age relation of long-period variables. A&A 658, L1
(2022)
33. M. Feast, P. Whitelock, J. Menzies, Globular clusters and the Mira period-luminosity relation.
MNRAS 329, L7–L12 (2002)
11 The Mira Distance Ladder 215

34. M.W. Feast, P.A. Whitelock, J.W. Menzies, Carbon-rich Mira variables: kinematics and abso-
lute magnitudes. MNRAS 369, 791–797 (2006)
35. M.W. Feast, The ages, masses, evolution and kinematics of mira variables, in AGB Stars and
Related Phenomena, ed. by T. Ueta, N. Matsunaga, Y. Ita (2009), p. 48
36. B.P. Gerasimovic, The absolute magnitudes of long period variable stars. Proc. Natl. Acad.
Sci. 14, 963–968 (1928)
37. I. Soszyński, A. Udalski, M.K. Szymański, M. Kubiak, G. Pietrzyński, Ł Wyrzykowski, K.
Ulaczyk, R. Poleski, S. Kozłowski, P. Pietrukowicz, J. Skowron, The optical gravitational
lensing experiment. The OGLE-III catalog of variable stars XV. Long-period variables in the
galactic bulge. AcA 63, 21–36 (2013)
38. P.N. Kholopov, N.N. Samus, E.V. Kazarovets, N.B. Perova, The 67th name-list of variable
stars. Inf. Bull. Variable Stars 2681, 1 (1985)
39. I. Soszyński, A. Udalski, M.K. Szymański, M. Kubiak, G. Pietrzyński, Ł Wyrzykowski, O.
Szewczyk, K. Ulaczyk, R. Poleski, The optical gravitational lensing experiment. The OGLE-
III catalog of variable stars IV. Long-period variables in the large magellanic cloud. AcA 59,
239–253 (2009)
40. A.A. Zijlstra, T.R. Bedding, Period evolution in Mira variables. JAAVSO 31, 2–10 (2002)
41. J.A. Mattei, Introducing Mira variables. JAAVSO 25, 57–62 (1997)
42. P.A. Whitelock, F. van Leeuwen, M.W. Feast, The luminosities and diameters of Mira vari-
ables from HIPPARCOS parallaxes, in Hipparcos–Venice 1997, ed. by R.M. Bonnet, E. Høg,
P.L. Bernacca, L. Emiliani, A. Blaauw, C. Turon, J. Kovalevsky, L. Lindegren, H. Hassan,
M. Bouffard, B. Strim, D. Heger, M.A.C. Perryman, L. Woltjer. ESA Special Publication,
vol. 402 (1997), pp. 213–218
43. C.D. Huang, A.G. Riess, S.L. Hoffmann, C. Klein, J. Bloom, W. Yuan, L.M. Macri, D.O. Jones,
P.A. Whitelock, S. Casertano, R.I. Anderson, A near-infrared period-luminosity relation for
Miras in NGC 4258, an anchor for a new distance ladder. ApJ 857, 67 (2018)
44. I.S. Glass, T.L. Evans, A period-luminosity relation for Mira variables in the large magellanic
cloud. Nature 291, 303–304 (1981)
45. A. Udalski, M.K. Szymanski, I. Soszynski, R. Poleski, The optical gravitational lensing exper-
iment. final reductions of the OGLE-III data. AcA 58, 69–87 (2008)
46. C. Alcock, R.A. Allsman, T.S. Axelrod, D.P. Bennett, K.H. Cook, K.C. Freeman, K. Griest,
J.A. Guern, M.J. Lehner, S.L. Marshall, H.S. Park, S. Perlmutter, B.A. Peterson, M.R. Pratt,
P.J. Quinn, A.W. Rodgers, C.W. Stubbs, W. Sutherland, The MACHO project first-year large
magellanic cloud results: the microlensing rate and the nature of the galactic dark halo. ApJ
461, 84 (1996)
47. P.R. Wood, C. Alcock, R.A. Allsman, D. Alves, T.S. Axelrod, A.C. Becker, D.P. Bennett,
K.H. Cook, A.J. Drake, K.C. Freeman, K. Griest, L.J. King, M.J. Lehner, S.L. Marshall,
D. Minniti, B.A. Peterson, M.R. Pratt, P.J. Quinn, C.W. Stubbs, W. Sutherland, A. Tomaney,
T. Vandehei, D.L. Welch, MACHO observations of LMC red giants: Mira and semi-regular
pulsators, and contact and semi-detached binaries, Asymptotic Giant Branch Stars, ed. by
T. Le Bertre, A. Lebre, C. Waelkens, vol. 191 (1999), p. 151
48. I.S. Glass, T. Lloyd Evans, The calibrating stars of the Mira P-L relation. MNRAS 343, 67–74
(2003)
49. I. Soszynski, A. Udalski, M. Kubiak, M.K. Szymanski, G. Pietrzynski, K. Zebrun, O.
Szewczyk, L. Wyrzykowski, K. Ulaczyk, The optical gravitational lensing experiment. Miras
and semiregular variables in the large magellanic cloud. AcA 55, 331–348 (2005)
50. P.A. Whitelock, M.W. Feast, F. Van Leeuwen, AGB variables and the Mira period-luminosity
relation. MNRAS 386, 313–323 (2008). (May)
51. L.M. Macri, C.C. Ngeow, S.M. Kanbur, S. Mahzooni, M.T. Smitka, Large magellanic cloud
near-infrared synoptic survey. I cepheid variables and the calibration of the Leavitt law. AJ
149, 117 (2015)
52. P. Iwanek, I. Soszyński, S. Kozłowski, Mid-infrared period-luminosity relations for Miras in
the large magellanic cloud. ApJ 919, 99 (2021)
216 C. D. Huang

53. P.A. Whitelock, Asymptotic giant branch variables in the galaxy and the local group. ApSS
341, 123–129 (2012)
54. W. Yuan, L.M. Macri, S. He, J.Z. Huang, S.M. Kanbur, C.C. Ngeow, Large magellanic cloud
near-infrared synoptic survey V. period-luminosity relations of Miras. AJ 154, 149 (2017)
55. W. Yuan, S. He, L.M. Macri, J. Long, J.Z. Huang, The M33 synoptic stellar survey II Mira
variables. AJ 153, 170 (2017)
56. W. Yuan, L.M. Macri, A. Javadi, Z. Lin, J.Z. Huang, Near-infrared Mira period-luminosity
relations in M33. AJ 156, 112 (2018)
57. S. He, W. Yuan, J.Z. Huang, J. Long, L.M. Macri, Period estimation for sparsely-sampled
quasi-periodic light curves applied to Miras. AJ 152, 164 (2016)
58. J.Y. Ou, C.C. Ngeow, A. Bhardwaj, M.J. Graham, R.R. Laher, F.J. Masci, R. Riddle, A distance
measurement to M33 using optical photometry of Mira variables. AJ 165, 137 (2023)
59. M. Rejkuba, The distance to the giant elliptical galaxy NGC 5128. A&A 413, 903–912 (2004)
60. P.A. Whitelock, J.W. Menzies, M.W. Feast, F. Nsengiyumva, N. Matsunaga, The local group
galaxy NGC 6822 and its asymptotic giant branch stars. MNRAS 428, 2216–2231 (2013)
61. J.W. Menzies, P.A. Whitelock, M.W. Feast, The local group galaxy IC 1613 and its asymptotic
giant branch variables. MNRAS 452, 910–923 (2015)
62. P. Battinelli, S. Demers, The calibration of the metallicity versus C/M relation. A&A 434,
657–663 (2005). (May)
63. S.R. Goldman, M.L. Boyer, K.B.W. McQuinn, P.A. Whitelock, I. McDonald, J.T. van Loon,
E.D. Skillman, R.D. Gehrz, A. Javadi, G.C. Sloan, O.C. Jones, M.A.T. Groenewegen, J.W.
Menzies, An infrared census of DUST in nearby galaxies with spitzer (DUSTiNGS). V. The
period-luminosity relation for dusty metal-poor AGB stars. ApJ 877, 49 (2019). (May)
64. P.A. Whitelock, AGB variables in local group dwarf irregular galaxies, in Stellar Populations
and the Distance Scale, ed. by J. Jensen, R.M. Rich, R. de Grijs. Astronomical Society of the
Pacific Conference Series, , vol. 514 (2018), p. 49
65. C.D. Huang, A.G. Riess, W. Yuan, L.M. Macri, N.L. Zakamska, S. Casertano, P.A. Whitelock,
S.L. Hoffmann, A.V. Filippenko, D. Scolnic, Hubble space telescope observations of Mira
variables in the SN Ia host NGC 1559: an alternative candle to measure the Hubble constant.
ApJ 889, 5 (2020)
66. M.J. Reid, D.W. Pesce, A.G. Riess, An improved distance to NGC 4258 and its implications
for the Hubble constant. ApJL 886, L27 (2019)
67. E.M.L. Humphreys, M.J. Reid, J.M. Moran, L.J. Greenhill, A.L. Argon, Toward a new geo-
metric distance to the active galaxy NGC 4258. III. Final results and the Hubble constant.
ApJ 775, 13 (2013)
68. A.G. Riess, L.M. Macri, S.L. Hoffmann, D. Scolnic, S. Casertano, A.V. Filippenko, B.E.
Tucker, M.J. Reid, D.O. Jones, J.M. Silverman, R. Chornock, P. Challis, W. Yuan, P.J. Brown,
R.J. Foley, A 2.4% determination of the local value of the Hubble constant. ApJ 826, 56
(2016)
69. L.M. Macri, K.Z. Stanek, D. Bersier, L.J. Greenhill, M.J. Reid, A new Cepheid distance to
the Maser-Host galaxy NGC 4258 and its implications for the Hubble constant. ApJ 652,
1133–1149 (2006)
70. ...G. Pietrzyński, D. Graczyk, A. Gallenne, W. Gieren, I.B. Thompson, B. Pilecki, P. Karcz-
marek, M. Górski, K. Suchomska, M. Taormina, B. Zgirski, P. Wielgórski, Z. Kołaczkowski,
P. Konorski, S. Villanova, N. Nardetto, P. Kervella, F. Bresolin, R.P. Kudritzki, J. Storm, R.
Smolec, W. Narloch, A distance to the large magellanic cloud that is precise to one percent.
Nature 567, 200–203 (2019)
71. A. Lançon, P.R. Wood, A library of 0.5 to 2.5 mu m spectra of luminous cool stars. A&AS
146, 217–249 (2000)
72. STScI Development Team, pysynphot: synthetic photometry software package. Astrophysics
Source Code Library, record ascl:1303.023 (2013)
73. A.G. Riess, W. Yuan, L.M. Macri, D. Scolnic, D. Brout, S. Casertano, D.O. Jones,
Y. Murakami, G.S. Anand, L. Breuval, T.G. Brink, A.V. Filippenko, S. Hoffmann, S.W. Jha,
W. D’arcy Kenworthy, J. Mackenty, B.E. Stahl, W. Zheng, A comprehensive measurement
11 The Mira Distance Ladder 217

of the local value of the Hubble constant with 1 km s.−1 Mpc.−1 uncertainty from the Hubble
space telescope and the S. H0 ES team. ApJL 934, L7 (2022)
74. J.J. Dalcanton, B.F. Williams, J.L. Melbourne, L. Girardi, A. Dolphin, P.A. Rosenfield, M.L.
Boyer, R.S. de Jong, K. Gilbert, P. Marigo, K. Olsen, A.C. Seth, E. Skillman, Resolved
near-infrared stellar populations in nearby galaxies. ApJS 198, 6 (2012)
75. S.L. Hoffmann, L.M. Macri, A.G. Riess, W. Yuan, S. Casertano, R.J. Foley, A.V. Filippenko,
B.E. Tucker, R. Chornock, J.M. Silverman, D.L. Welch, A. Goobar, R. Amanullah, Optical
identification of Cepheids in 19 host galaxies of type Ia supernovae and NGC 4258 with the
Hubble space telescope. ApJ 830, 10 (2016)
76. W. Yuan, L.M. Macri, B.M. Peterson, A.G. Riess, M.M. Fausnaugh, S.L. Hoffmann, G.S.
Anand, M.C. Bentz, E. Dalla Bontà, R.I. Davies, G. De Rosa, L. Ferrarese, C.J. Grier, E.K.S.
Hicks, C.A. Onken, R.W. Pogge, T. Storchi-Bergmann, M. Vestergaard, The Cepheid distance
to the narrow-line Seyfert 1 galaxy NGC 4051. ApJ 913, 3 (2021)
77. W. Yuan, L.M. Macri, A.G. Riess, T.G. Brink, S. Casertano, A.V. Filippenko, S.L. Hoffmann,
C.D. Huang, D. Scolnic, Absolute calibration of Cepheid period-luminosity relations in NGC
4258. ApJ 940, 64 (2022)
78. M.S. Vardya, Classification of Mira variables based on visual light curve shape. A&AS 73,
181–194 (1988)
79. T. Lebzelter, The shapes of light curves of Mira-type variables. Astronomische Nachrichten
332, 140 (2011)
80. C.D. Huang, W. Yuan, A.G. Riess, W. Hack, P.A. Whitelock, N.L. Zakamska, S. Casertano,
L.M. Macri, M. Marengo, J.W. Menzies, R.K. Smith, The Mira distance to M101 and a
4% measurement of H 0. Astrophys. J. 963(2), 83 (2024). https://ui.adsabs.harvard.edu/abs/
2024ApJ...963...83H/abstract
81. M.W. Feast, The pulsation, temperatures and metallicities of Mira and semiregular variables
in different stellar systems. MNRAS 278, 11–21 (1996)
82. P. Whitelock, J. Menzies, M. Feast, F. Marang, B. Carter, G. Roberts, R. Catchpole, J. Chap-
man, High-mass-loss AGB stars in the South Galactic Cap. MNRAS 267, 711–742 (1994)
83. G.T. van Belle, R.R. Thompson, M.J. Creech-Eakman, Angular size measurements of mira
variable stars at 2.2 microns II. AJ 124, 1706–1715 (2002)
84. S. Xu, B. Zhang, M.J. Reid, X. Zheng, G. Wang, Comparison of Gaia DR2 parallaxes of stars
with VLBI astrometry. ApJ 875, 114 (2019)
85. J.L. Sanders, The period–luminosity relation for Mira variables in the Milky Way using Gaia
DR3: a further distance anchor for H 0. Monthly Not. R. Astron. Soc. 523(2), 2369–2398
(2023). https://ui.adsabs.harvard.edu/abs/2023MNRAS.523.2369S/abstract
86. M. Andriantsaralaza, S. Ramstedt, W.H.T. Vlemmings, E. De Beck, Distance estimates for
AGB stars from parallax measurements. A&A 667, A74 (2022)
87. A. Chiavassa, B. Freytag, M. Schultheis, Heading Gaia to measure atmospheric dynamics in
AGB stars. A&A 617, L1 (2018)
88. K. El-Badry, H.-W. Rix, T.M. Heintz, A million binaries from Gaia eDR3: sample selection
and validation of Gaia parallax uncertainties. MNRAS 506, 2269–2295 (2021)
89. P. Whitelock, M. Feast, Hipparcos parallaxes for Mira-like long-period variables. MNRAS
319, 759–770 (2000)
90. F. van Leeuwen, Hipparcos, the New Reduction of the Raw Data, vol. 350 (2007)
91. D. Engels, E. Kreysa, G.V. Schultz, W.A. Sherwood, The nature of OH/IR stars I. Infrared
Mira variables. A&A 124, 123–138 (1983)
92. R. Urago, T. Omodaka, T. Nagayama, Y. Watabe, R. Miyanosita, N. Matsunaga, R.A. Burns,
The 3D distribution of long-period Mira variables in the galactic disk. ApJ 891, 50 (2020)
93. I. Soszyński, A. Udalski, M.K. Szymański, M. Kubiak, G. Pietrzyński, Ł Wyrzykowski,
K. Ulaczyk, R. Poleski, S. Kozłowski, P. Pietrukowicz, The optical gravitational lensing
experiment. The OGLE-III catalog of variable stars XIII long-period variables in the small
magellanic cloud. AcA 61, 217–230 (2011)
94. P.A. Whitelock, J.W. Menzies, M.W. Feast, P. Marigo, A remarkable oxygen-rich asymptotic
giant branch variable in the Sagittarius Dwarf Irregular Galaxy. MNRAS 473, 173–184 (2018)
218 C. D. Huang

95. J.W. Menzies, P.A. Whitelock, M.W. Feast, N. Matsunaga, Luminous AGB variables in the
dwarf irregular galaxy, NGC 3109. MNRAS 483, 5150–5165 (2019)
96. M.M. Kasliwal, J. Bally, F. Masci, A.M. Cody, H.E. Bond, J.E. Jencson, S. Tinyanont, Y. Cao,
C. Contreras, D.A. Dykhoff, S. Amodeo, L. Armus, M. Boyer, M. Cantiello, R.L. Carlon,
A.C. Cass, D. Cook, D.T. Corgan, J. Faella, O.D. Fox, W. Green, R.D. Gehrz, G. Helou,
E. Hsiao, J. Johansson, R.M. Khan, R.M. Lau, N. Langer, E. Levesque, P. Milne, S. Mohamed,
N. Morrell, A. Monson, A. Moore, E.O. Ofek, D. O’ Sullivan, M. Parthasarathy, A. Perez,
D.A. Perley, M. Phillips, T.A. Prince, D. Shenoy, N. Smith, J. Surace, S.D. Van Dyk, P.A.
Whitelock, R. Williams, SPIRITS: uncovering unusual infrared transients with spitzer. ApJ
839, 88 (2017)
97. V.R. Karambelkar, S.M. Adams, P.A. Whitelock, M.M. Kasliwal, J.E. Jencson, M.L. Boyer,
S.R. Goldman, F. Masci, A.M. Cody, J. Bally, H.E. Bond, R.D. Gehrz, M. Parthasarathy,
R.M. Lau, SPIRITS Collaboration, SPIRITS catalog of infrared variables: identification of
extremely luminous long period variables. ApJ 877, 110 (2019)
98. M.L. Boyer, K.B.W. McQuinn, P. Barmby, A.Z. Bonanos, R.D. Gehrz, K.D. Gordon, M.A.T.
Groenewegen, E. Lagadec, D. Lennon, M. Marengo, I. McDonald, M. Meixner, E. Skillman,
G.C. Sloan, G. Sonneborn, J.T. van Loon, A. Zijlstra, An infrared census of DUST in nearby
galaxies with spitzer (DUSTiNGS) II. Discovery of metal-poor dusty AGB stars. ApJ 800,
51 (2015)
99. A.J. Lee, W.L. Freedman, B.F. Madore, K.A. Owens, I. Sung Jang, A preliminary calibration
of the JAGB method using Gaia EDR3. ApJ 923, 157 (2021)
100. B.F. Madore, W.L. Freedman, Astrophysical distance scale: the AGB J-band method. I. Cali-
bration and a first application. Astrophys. J. 899(1), 66 (2020). https://ui.adsabs.harvard.edu/
abs/2020ApJ...899...66M/abstract
101. S. Li, A.G. Riess, S. Casertano, G.S. Anand, D.M. Scolnic, W. Yuan, L. Breuval, C.D.
Huang, Reconnaissance with JWST of the J-region asymptotic giant branch in distance
ladder galaxies: from irregular luminosity functions to approximation of the Hubble con-
stant. arXiv preprint arXiv:2401.04777 (2024, Jan 9). https://ui.adsabs.harvard.edu/abs/
2024arXiv240104777L/abstract
102. R.L. Beaton, G. Bono, V.F. Braga, M. Dall’Ora, G. Fiorentino, I.S. Jang, C.E. Martínez-
Vázquez, N. Matsunaga, M. Monelli, J.R. Neeley, M. Salaris, Old-aged primary distance
indicators. SSRv 214, 113 (2018)
103. K.B.W. McQuinn, M. Boyer, E.D. Skillman, A.E. Dolphin, Using the tip of the red giant
branch as a distance indicator in the near infrared. ApJ 880, 63 (2019)
104. A.G. Riess, G.S. Anand, W. Yuan, S. Casertano, A. Dolphin, L.M. Macri, L. Breuval,
D. Scolnic, M. Perrin, R.I. Anderson, Crowded no more: the accuracy of the Hubble con-
stant tested with high resolution observations of Cepheids by JWST (2023). arXiv e-prints:
arXiv:2307.15806
105. W. Yuan, Period-luminosity relations of Cepheid and Mira variables and their application to
the extragalactic distance scale. PhD thesis, Texas A&M University System (2017)
Chapter 12
Tully–Fisher Relation

Khaled Said

12.1 Introduction

12.1.1 Description of the Tully–Fisher Relation

The Tully–Fisher (TF) relation is an empirical relation that correlates the intrinsic
brightness of a spiral galaxy, measured by its total luminosity, and its dynamical prop-
erties, measured by its rotational velocity. In 1977, Brent Tully and Richard Fisher
proposed the use of this relation as a distance indicator based on their observations
of spiral galaxies in the Virgo cluster [1]. The main idea behind distance indicators is
to use distance-independent observables to predict a distance-dependent parameter,
which can subsequently be compared with the corresponding observable to derive a
distance estimate. In the case of the Tully–Fisher relation, the rotational velocity of a
spiral galaxy serves as the distance-independent observable that is used to predict the
total absolute luminosity, which is distance-dependent. The total luminosity can then
be compared with apparent magnitude to measure distance via the distance modulus.
The Tully–Fisher relation has since been extensively studied and refined [2–9]. It
has become a workhorse tool for measuring the distances and properties of galaxies,
particularly in the context of large-scale structure surveys and cosmological studies
[10–16].

12.1.2 Historical Background

The historical background of the Tully–Fisher relation can be traced back to the Great
Debate in the 1920s, when Ernst Öpik used an expression between the observed rota-
tional velocity and the absolute magnitude to measure the true distance to Andromeda

K. Said (B)
School of Mathematics and Physics, University of Queensland, Brisbane, QLD 4072, Australia
e-mail: k.saidahmedsoliman@uq.edu.au

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 219
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_12
220 K. Said

[17]. This measurement helped to prove that Andromeda is an independent galaxy,

not part of our own Milky Way Galaxy, as Shapley had thought.
More than 50 years later, Balkowski et al. used the Nançay radio telescope in
France to measure the line widths of a sample of 13 irregular and spiral galaxies
[18]. They found a correlation between the line width and the luminosity but they
did not apply this correlation as a distance indicator. This laid the foundation for the
ground-breaking paper by Brent Tully and Richard Fisher in 1977 [1]. In their paper,
Tully and Fisher used only inclined spiral galaxies and proposed the usage of the
linear relation between H I profile and absolute magnitude as a distance indicator.
The publication of the Tully–Fisher relation in 1977 and the proposal to use it
as a distance indicator was significant in many different ways. Firstly, it provided
a robust new tool for measuring distance at redshifts that other methods such as
Cepheid variable stars cannot reach. This opened up a whole range of large scale
structure and cosmological studies. Notably, similar relations like the Fundamental
Plane relation [19, 20] only emerged a decade later. Secondly, Tully and Fisher
measured the Hubble constant . H0 to be 80 .km s−1 Mpc−1 from the Virgo cluster and
Ursa Major. This value was the first to deviate from the two mainstream values at
that time, low value of . H0 = 50 .km s−1 Mpc−1 which was promoted by Sandage
and Tammann [21] and a higher value of . H0 = 100 .km s−1 Mpc−1 claimed by de
Vaucouleurs [22]. This value is much closer to the value of Hubble constant that we
know today.

12.1.3 Tully–Fisher Relation in Cosmology

The Tully–Fisher relation has been heavily used in measuring the Hubble constant
H0 . Additionally, it has also been used as a valuable tool in many other cosmological
.
studies1 :
• Cosmography: The Tully–Fisher relation uses the maximum rotational velocity
of a galaxy to predict its luminosity. This can be used to estimate the distance
to the galaxy, and to measure its peculiar velocity. The peculiar velocity is the
deviation from the smooth Hubble flow and can be used to reconstruct 3D maps of
the density and velocity fields of the universe. In 2014, a team led by Brent Tully
used the Wiener filter reconstruction [23] to recover the 3D density and velocity
maps. They used these maps to set the borders of our home supercluster which they
called Laniakea [24]. More recently, Alexandra Dupuy and Hélène Courtois used
the full CosmicFlows-4 [25] catalog2 to reconstruct the 3D density and velocity
fields [26] and then redefine the boundaries of the Laniakea supercluster along
with other structures such as Perseus-Pisces and Shapley [27].

1 I do not intend to give a comprehensive review of all cosmological studies that used TF relation,
but I will discuss a few examples before focusing on the role of TF relation in the . H0 tension which
is the main focus of this chapter.
2 This catalog contains TF distances along with other distance indicators.
12 Tully–Fisher Relation 221

• Bulk Flow: The bulk flow is the average peculiar velocity over a sphere of a
given radius R. The theoretical expected value of this measurement relies on a
model of a specific set of cosmological parameters, initial conditions and a chosen
gravitational law [28]. In 2009, Watkins, Feldman and Hudson used the largest TF
survey available at that time, the SFI++ survey [10], to measure the bulk flow of
galaxies. They used a new method, the minimal variance estimator, and found a
large bulk flow amplitude of .431 ± 102 km s.−1 within a scale of .50 Mpc .h −1 [29].
The amplitude of this bulk flow is larger than expected from the .ΛCDM model.
More recently, several teams have used the state-of-the-art CosmicFlows-4 catalog
[25], which contains more than 10,000 newly measured TF distances, to measure
the bulk flow of galaxies. They have also found large bulk flows, arising on a much
larger scale, that are higher than the expected value from the .ΛCDM model [26,
30, 31]
• Growth rate of structure: Over the last three decades, many collaborations have
started to test Einstein’s general theory of relativity over a range of scales. One way
to do that is by measuring the growth rate of cosmic structure, or . fσ8 . This method
is capable of measuring. fσ8 at low redshifts (.z < 0.1), a range inaccessible to other
methods such as the redshift space distortion. The growth rate can be constrained by
comparing the density field from redshift surveys and peculiar velocity fields from
TF surveys. The growth rate can then be parameterized as a function of the mass
density parameter .Ωm and the growth index .γ , which is determined by the theory
of gravity. In 2011, Marc Davis and others used the TF survey SFI++ to find a . fσ8
value of .0.31 ± 0.06 [32]. Other teams also used the TF SFI++ catalog and found
a slightly higher value of .0.401 ± 0.024 [33]. More recently, the CosmicFlows-4
catalog also was used to measure . fσ8 and found a similar value of .0.40 ± 0.07
[16].
• Galaxy formation and evolution: As an empirical relation between fundamental
properties of galaxies, TF relation has been used to refine semi-analytical models
and numerical simulations of galaxy formation and evolution [34–36].
• Dark matter in galaxies: The Tully–Fisher relation have also been used to probe
the distribution and properties of dark matter in galaxies. Several studies put con-
straints on galaxy halo profiles by comparing the halo profile required by the
Tully–Fisher relation to the density profile that is well-described by the Navarro-
Frenk-White (NFW) profile [37–39].

12.2 The Tully–Fisher Relation

12.2.1 How Does It Work?

The Tully–Fisher relation is defined and its usages in cosmology are discussed above.
But how does it work in practice?
222 K. Said

Building the Tully–Fisher relation requires two data sets: photometry and spec-
troscopy. Photometry provides a measurement of the galaxy’s luminosity, which is the
distance-dependent parameter in the TF relation. Spectroscopy provides a measure-
ment of the galaxy’s rotational velocity, which is the distance-independent parame-
ter in the relation. The Tully–Fisher relation establishes the connection between the
luminosity and rotational velocity of spiral galaxies. The TF relation is a secondary
distance indicator and as the name suggests it requires calibration from a sample of
galaxies with known distances usually coming from a primary distance indicator or
a sample of cluster galaxies. Figure 12.1a represents the initial step in utilizing the
TF relation for distance measurements. It shows a sample of galaxies with known
rotational velocities and absolute magnitudes. The TF relation of the form:

. M = a + b log W (12.1)

where . M is the absolute magnitude, .W represents the rotation, .a and .b are the
TF slope and intercept, respectively, can be fit to that sample. This relation, which
we often call a template relation (Fig. 12.1b), can be used to predict the absolute
magnitude of a galaxy with unknown distance given a measurement of its rotational
velocity as shown in Fig. 12.1c. The predicted absolute magnitude can be compared
to the galaxy’s apparent magnitude to yield the distance via the distance modulus as:

. μ = m − M(W ). (12.2)

where .μ is the distance modulus, .m is the apparent magnitude, and . M(W ) is the
predicted absolute magnitude of a galaxy from the TF relation given its rotational
velocity.
It is important to note that the derived template relation is not a universal relation.
Instead, the data, parameters, and corrections used in the measurements of the TF
distances should be consistent with those used in building the TF template relation.
This has encouraged many teams to derive their own template relations instead of
using an existing one. In the past few decades, different measurements, wavelengths,
and methodologies have been used to do this. For example, shortly after the inception
of the TF relation, several studies began to use near-infrared (NIR) bands because
they suffer less than optical bands from dust extinction (both internal and external)
and are more sensitive to stellar mass [9, 11, 40]. Other groups used isophotal
magnitudes instead of total magnitudes to extend the Tully–Fisher studies into the
Zone of Avoidance around the plane of the Milky Way [41, 42].
In terms of spectroscopy, many parameters have been used in an attempt to min-
imise the intrinsic scatter of the TF relation. Many teams have used the 20% H I line
width (the width measured at 20% of the peak flux of the H I profile) including the
original TF relation [1, 43, 44], while others suggested using the width measured at
50% instead [45–48]. Furthermore, other teams suggest that the flat part of the optical
rotation curve is the best proxy for the rotational velocity, and they have shown that
it does indeed yield the tightest Tully–Fisher relation [49, 50].
12 Tully–Fisher Relation 223

Fig. 12.1 Illustrating how the Tully–Fisher relation combines photometric and spectroscopic data
to enable the estimation of galaxy distance. In order to use the Tully–Fisher relation to measure
distances, we need to build something called template relation. a Shows a sample of galaxies
with known rotation velocity, apparent magnitude, and distances which allows the determination of
luminosity. b By fitting the TF relation to this sample, a template relation is derived. This establishes
the correlation between the absolute magnitude and the rotational velocity of galaxies. c Applying
the Tully–Fisher template to a galaxy of unknown distance, the absolute magnitude can be estimated
based on its measured rotational velocity

Different methodologies have been used to extract the best-fitting parameters for
the Tully–Fisher relation by applying different fitting procedures. In an attempt to
overcome the Malmquist bias, Renée C. Kraan-Korteweg and others proposed the
use of the inverse TF relation, which only takes into account the errors on the x-axis
(rotational velocity) [4]. Since then, many teams have used both direct and inverse
fitting procedures in their analyses [51, 52]. Other teams have used bivariate forms
that take into account errors in both the x- and y-axes [9, 11, 45, 48]. Given the
scatter in TF relation, a suggestion has been made to use a Bayesian mixture model,
which is less biased by outliers [53].
In summary, the Tully–Fisher relation has been observed to hold over a wide range
of applications, including different parameters, methodologies, and wavelengths.
This suggests that the TF relation is one of the most robust and powerful tools for
measuring galaxy distances.

12.2.2 Theoretical Basis

The physical origin of the Tully–Fisher relation is still not fully understood, but it
is certainly rooted in the physics of gravity and the dynamics of galactic rotation.
The TF relation links the mass of a galaxy (characterized by the luminosity) to its
dynamics (characterized by the rotational velocity). Specifically, it is based on the
idea that the rotation velocity of a galaxy is related to the mass contained within a
certain radius from the galaxy’s center, known as the circular velocity.
224 K. Said

Michael Strauss and Jeffrey Willick provided a widely accepted theoretical expla-
nation3 for the Tully–Fisher relation in their 1995 review [54]. They wrote that by
equating the centrifugal force of an object moving in a circle of radius r to the gravi-
tational attraction on the same object due to the mass inside a sphere of radius r, one
can write

M
.
2
Vrot ∝ (12.3)
r
where .Vrot is the rotational velocity, . M is the mass within a sphere of radius .r , and
r is the distance from the galaxy’s center. Then, assuming a constant value for both
.
the mass-to-light ratio (M/L) and mean surface brightness for spirals, one can obtain
that

. L ∝ Vrot
4
(12.4)

However, most Tully–Fisher studies have derived a power-law exponent that devi-
ates from this theoretical explanation. This is usually attributed to the complication
of the dark matter halo that surrounds a galaxy [55]. Early near-infrared Tully–Fisher
analyses showed that the complications of halo mass could be avoided, and indeed
derived a power-law exponent that was closer to the theoretical expectation [40].
However, their conclusion has been contradicted by many larger subsequent surveys
[11, 45].
In 2000, Stacy McGaugh and his team investigated the deviation of dwarf faint
galaxies from the linear Tully–Fisher relation [6]. They found that the linear TF
relation could be restored if they used the sum of both stellar and gas mass components
instead of luminosity. This relation takes the form

. Md ∝ Vrot
4
(12.5)

where . Md is the sum of stellar and gas masses,

. Md = M∗ + Mgas . (12.6)

Not only did they establish the baryonic TF relation as the fundamental relation,
but they also found that the intrinsic scatter was smaller than the original TF relation.
This result can be simply explained if we suppose that this estimate of the baryonic
mass is a better proxy for the true total mass than the luminosity or the stellar mass
alone.

3 This was not the first such explanation, but it was particularly clear.
12 Tully–Fisher Relation 225

This result, at first glance, may seem to support the Modified Newtonian Dynamics
(MOND: [56]) as an alternative to dark matter. MOND is a theory of gravity that
modifies Newtonian dynamics. In MOND, the baryonic mass is actually the total
mass, which means that the baryonic TF relation is the fundamental TF relation.
However, it is important to note that Modified Newtonian Dynamics (MOND) has
its own critics [57, 58]. We refer the reader to the literature for more information
about the debate between MOND and Cold Dark Matter.
One important theoretical development in recent years has been the use of hydro-
dynamic simulations to study the formation and evolution of galaxies in a cosmo-
logical context [35, 59]. These simulations incorporate the effects of gravity, gas
dynamics, star formation, and feedback, and have been shown to predict the Tully–
Fisher relation in different environments and at different redshifts better than semi-
analytical models [60].

12.2.3 Advantages and Limitations

The Tully–Fisher relation has several advantages as a tool for measuring distances of
galaxies. The first and main advantage is its simplicity and robustness. The TF relation
provides a linear correlation between only two easily measurable galaxy properties,
luminosity and rotational velocity. Several studies have shown that, unlike other
distance indicators, the scatter of the TF relation is relatively insensitive to other
galaxy properties [8, 61, 62]. This makes the TF relation a fundamental, robust and
reliable tool for measuring distances.
Another advantage of the Tully–Fisher relation is that it uses spiral galaxies, which
are the most numerous type of galaxies at low redshifts (.z < 0.07). This makes it
a convenient tool for measuring peculiar velocities at these redshifts, where the
true distances can be determined. Beyond this redshift, it is more difficult to obtain
accurate peculiar velocities as the errors in measuring these velocities from distance
indicators scale with redshift.
In Fig. 12.2, I show the redshift distribution of spiral and elliptical galaxies from
the SDSS survey DR14 [63]. I separated spirals from ellipticals using the galaxy zoo
catalog [64]. I also over plot the redshift distribution of the state-of-the-art Pantheon+
sample [65, 66]. I used only the redshift below .z < 0.15, as this is the range used for
Hubble constant (. H0 ) measurements [67]. It is clear that spiral galaxies (Tully–Fisher
galaxies) are the most numerous type at low redshift .(z < 0.07), with a mean redshift
of .z = 0.077. This makes TF relation one of the most precise tools for measuring
. H0 at this redshift.
However, there are also some limitations to the Tully–Fisher relation as a tool for
distance measurements. One of the main limitations of the TF relation is its intrinsic
scatter. This scatter can lead to errors of around .∼20% in distance measurements.
The baryonic Tully–Fisher relation is often shown to be a tighter relation than the
TF relation, with a much smaller intrinsic scatter [68, 69]. However, measuring the
sum of the stellar and gas components is much more difficult than measuring just the
226 K. Said

Fig. 12.2 The redshift distribution of spiral (black) and elliptical (red) galaxies in the SDSS survey,
as well as the redshift distribution of the state-of-the-art SN Ia from Pantheon+ (blue), up to redshift
.z < 0.15. The number of galaxies per 0.0025 redshift bin is shown

luminosity of a galaxy. This adds a lot of observational uncertainties to the baryonic

TF relation.
Another limitation of the Tully–Fisher relation and a significant source of uncer-
tainty is the inclination of galaxies. The inclination of a spiral galaxy affects both TF
parameters. It affects the derived maximum rotational velocity during the process of
correcting the observed line width for projection effects (i.e. to edge-on orientation).
Additionally, to correct the magnitude for the internal extinction, one needs to use
the inclination to account for the path-length of the light through the galaxy. There-
fore, inclination errors lead to correlated errors in the two TF parameters that have
the same sense (qualitatively) as errors in distance, so an error in inclination is (at
least partly) degenerate with an error in distance. Most Tully–Fisher surveys select
only edge-on galaxies by applying a cut to inclinations in an attempt to minimize
the line width correction. Although this seems like the right thing to do, inaccurate
inclinations will lead to a selection bias in the TF distance. This is because galaxies
with inaccurate inclinations may be excluded or included in the survey, depending on
whether their inclinations are overestimated or underestimated. Furthermore, edge-
on galaxies are subject to higher internal extinction, which means that more light is
blocked by dust and gas inside the galaxy. This results in a dimmer magnitude, and
thus a larger correction is needed to correct for the extinction. Some studies have
even suggested that when the inclination measurements have errors larger than 10.◦ C,
it may be more prudent to omit inclinations entirely rather than assuming them to be
exact [70].
12 Tully–Fisher Relation 227

Selection effects and Malmquist bias are other limitations of the TF relation.
However, these limitations are not specific to the TF relation, as they affect all
distance indicators.

12.3 The Role of the Tully–Fisher Relation in . H0

Measurements

The methodology for using the Tully–Fisher relation in . H0 measurements typically

involves the following steps. First, a sample of galaxies with well-measured rotational
velocities and luminosities (for TFR) or masses (for the BTFR) is selected (see
Fig. 12.1a). These galaxies are typically chosen to be representative of the population
of galaxies being studied. Next, the Tully–Fisher relation is calibrated for this sample
of galaxies. This involves measuring the slope, intercept, and scatter of the relation
for the sample, and correcting for any systematic biases or selection effects (see
Fig. 12.1b). Once the Tully–Fisher relation has been calibrated for the sample, it can
be used to infer the distances to other galaxies with similar properties (see Fig. 12.1c).
The final step is measuring. H0 , which can be done by plotting the derived TF distances
against redshifts.
This methodology has been used in a number of studies to estimate the value of
the Hubble constant (. H0 ). In 1977, Tully and Fisher applied their newly discovered
Tully–Fisher relation to derive distances to the Virgo and Ursa Major clusters. They
measured a Hubble constant of . H0 = 84 km s.−1 Mpc.−1 for Virgo and . H0 = 75
km s.−1 Mpc.−1 for Ursa Major. They conclude a preliminary value for the Hubble
constant of . H0 = 80 km s.−1 Mpc.−1 . Although they called it preliminary and did not
propagate their distance error to the Hubble constant value, we know today that it
was the most accurate result at that time.
In 1983, Visvanathan used the Tully–Fisher relation to measure distances to 52
cluster spiral galaxies. They found that the Hubble constant derived from nearby
clusters varied from 58.5 to 83.5, but the range was much smaller for distant clusters,
between 76.3 and 78.9 km s.−1 Mpc.−1 . Visvanathan concluded that the best fit value
for the Hubble constant from all clusters in their study was . H0 = 74.4 ± 11 km s.−1
Mpc.−1 [71]. That result is very close to the value of . H0 we have today.
Shortly after, Sandage and Tammann used the Tully–Fisher relation in the infrared
(IR) and blue bands to estimate distances to galaxies in the Virgo and Coma clusters,
respectively. Their analysis gave a lower value of the Hubble constant . H0 = 55 ± 9
km s.−1 Mpc.−1 [72]. This result aligned with their previous measurement of . H0 =
50 ± 7 km s.−1 Mpc.−1 obtained through Type I supernovae [73].
Aaronson led a team in 1986 to derive distances to 10 galaxy clusters using the
infrared Tully–Fisher relation. They stated that the current evidence favors a large
value for . H0 of 90 km s.−1 Mpc.−1 [74]. In two subsequent studies, Bottinelli led a
team to use the infrared Tully–Fisher relation to derive distances to galaxy clusters.
They concluded that the . H0 lies between 70 and 75 km s.−1 Mpc.−1 . During the same
228 K. Said

time, Renée C. Kraan-Korteweg and others derived distances to a larger sample of 82

cluster galaxies using both the infrared Tully–Fisher relation and the optical Tully–
Fisher relation. They found that their data is consistent with a Hubble constant of
−1
. H0 = 56.6 ± 0.9 km s. Mpc.−1 [75]. However, this low value was contradicted by
three studies. One study used the TF relation to measure distances to Virgo and Ursa
Major clusters, and derived a value of . H0 = 85 ± 10 km s.−1 Mpc.−1 [43]. The other
two studies derived distances to the Coma cluster, and also found a higher value
of . H0 ≈ 90 km s.−1 Mpc.−1 [76, 77]. Bureau, Mould, and Staveley-Smith [78] used
the Tully–Fisher relation to measure distances to galaxies in the Fornax cluster, and
derived a value of . H0 = 74 ± 11 km s.−1 Mpc.−1 .
This significant discrepancy in these early measurements can be attributed to
various factors, including the challenge of sample incompleteness correction and the
absence of reliable distances to nearby objects for use as a calibration sample.
The Hubble Space Telescope Key Project on the Extragalactic Distance Scale used
Cepheid distances to re-calibrate the Tully–Fisher relation. In a series of papers, they
found results consistent with a value of . H0 of .71 ± 4 km s.−1 Mpc.−1 [79–81].
The Cosmicflows project combines distances measured through various meth-
ods, including the Tully–Fisher relation, surveys, and collaborations. This project
has made significant contributions to measuring the Hubble constant, consistently
obtaining values in the range of .74 ± 4 to .76 ± 1 km s.−1 Mpc.−1 over an extended
period [26, 82–88].
Figure 12.3 shows a graphical representation of the published efforts to measure
the Hubble constant (. H0 ) using the Tully–Fisher relation. The red band shows the
the most recent result of Hubble constant and its associated errors measured using
Type Ia supernovae by the S. H0 ES team in 2022 [89]. Additionally, the blue band
represent the Hubble constant value and its associated error from the CMB measured
by the Planck Collaboration in 2020 [90]. The figure shows a lot of scatter in the
TF results prior to 2000 with some unrealistically optimistic uncertainties. However,
after 2000, the results seem consistent and in more agreement with the value derived
by the S. H0 ES team in 2022 [89] than the CMB value by the Planck Collaboration
in 2020 [90]. Table 12.1 provides a summary of the results from these studies.
In summary, the Tully–Fisher relation has been used for decades to estimate dis-
tances to galaxies and to measure the Hubble constant. Although Tully–Fisher mea-
surements of . H0 have shown a lot of scatter prior to 2000, they have been consistent
since then with a value of . H0 that agrees with the other low-redshift probes.

12.4 Future Prospects

The largest homogeneous Tully–Fisher sample currently exists is the CosmicFlows-

4 survey [87]. This survey includes 10,000 TF distances. Imminent surveys such
as DESI [98] and WALLABY [15] will provide 50,000 and 200,000 TF distances,
respectively, over the coming few years. With approximately 20% TF distance errors
for about 250,000 galaxies covering a significant portion of the sky up to redshift
12 Tully–Fisher Relation 229

Fig. 12.3 A sample of published Hubble constant (. H0 ) measurements using Tully–Fisher relation
spanning from 1977 to 2023. Most of these measurements are obtained by different teams using
different data given in Table 12.1. The red band highlights the most recent Hubble constant value
and its associated uncertainty from the baseline results of the Cepheid-SN Ia sample by the S. H0 ES
team [89]. The blue band represents the Hubble constant value reported by Planck Collaboration
2020 [90]. The majority of the data points align closely with the red band, particularly from 2000
onward. It is important to note that this sample of data is just a subset from the literature and
additional studies may already exist beyond those represented here

Table 12.1 Evolution of published Hubble constant values from the Tully–Fisher relation (1977–
2023)
Date . H0 Refs. Date . H0 Refs.
1977 80.0 [1] 2001 71.0 .± 3.0 [81]
1983 74.3 .± 11.0 [71] 2001 73.0 .± 4.0 [91]
1984 55.0 .± 9.0 [72] 2006 74.0 .± 2.0 [46]
1986 90.0 [74] 2009 84.2 .± 6.0 [92]
1987 72.0 .± 5.0 [93] 2011 79.0 .± 2.0 [94]
1988 85.0 .± 10.0 [43] 2012 75.0 [82]
1988 72.5 [95] 2012 75.2 .± 3.0 [83]
1988 56.6 .± 0.9 [75] 2013 74.0 .± 4.0 [84]
1991 92.0 .± 21.0 [76] 2014 74.4 .± 1.0 [85]
1993 90.0 [77] 2014 75.2 .± 3.3 [86]
1996 74.0 .± 11.0 [78] 2020 75.1 .± 2.3 [96]
1999 76.0 .± 3.0 [79] 2020 76.0 .± 1.1 [87]
2000 77.0 .± 8.0 [97] 2022 75.5 .± 2.5 [88]
2000 71.0 .± 4.0 [80] 2023 74.5 .± 0.1 [26]
a This is just a subset of the published Hubble constant values using TF relation, and additional
.
studies may already exist beyond those represented here
230 K. Said

. z ≈ 0.1, both random and systematic errors in . H0 from large-scale structure will be
considerably smaller compared to the most optimistic uncertainties associated with
other low-redshift probes.

Acknowledgements KS acknowledges support from the Australian Government through the Aus-
tralian Research Council’s Laureate Fellowship funding scheme (project FL180100168). I would
like to express my gratitude to Tamara Davis and Matthew Colless for their valuable feedback on
the initial draft of this review.

References

1. R.B. Tully, J.R. Fisher, A new method of determining distances to galaxies. A&A 54, 661–673
(1977)
2. J. Silk, Feedback, disk self-regulation, and galaxy formation. ApJ 481, 703–709 (1997)
3. V. Avila-Reese et al., On the formation and evolution of disk galaxies: cosmological initial
conditions and the gravitational collapse. ApJ 505, 37–49 (1998)
4. R.C. Kraan-Korteweg et al., 21 CM line widths and distances of spiral galaxies, in NATO ASIC
Proceedings of the 180 Galaxy Distances and Deviations from Universal Expansion, ed. by
B.F. Madore, R.B. Tully, pp. 65–72 (1986)
5. K. O’Neil et al., Red, gas-rich low surface brightness galaxies and enigmatic deviations from
the Tully–Fisher relation. AJ 119, 136–152 (2000)
6. S.S. McGaugh et al., The Baryonic Tully–Fisher relation. ApJ 533, L99–L102 (2000)
7. S.S. McGaugh, The Baryonic Tully–Fisher relation of galaxies with extended rotation curves
and the stellar mass of rotating galaxies. ApJ 632, 859–871 (2005)
8. S. Courteau et al., Scaling relations of spiral galaxies. ApJ 671, 203–225 (2007)
9. K. Said et al., On how to extend the NIR Tully-Fisher relation to be truly all-sky. MNRAS
447(2), 1618–1629 (2015)
10. C.M. Springob et al., SFI++. II a new i-band Tully–Fisher catalog, derivation of peculiar
velocities, and data set properties. ApJS 172, 599–614 (2007)
11. K.L. Masters et al., 2MTFI the Tully–Fisher relation in the two micron all sky survey J, H, and
K bands. AJ 135, 1738–1748 (2008)
12. R.B. Tully et al., The extragalactic distance database. AJ 138(2), 323–331 (2009)
13. H.M. Courtois et al., Cosmography of the local universe. AJ 146, 69 (2013)
14. T. Hong et al., 2MTF-VII, 2MASS Tully–Fisher survey final data release: distances for 2062
nearby spiral galaxies. MNRAS 487(2), 2061–2069 (2019)
15. H.M. Courtois et al., WALLABY pre-pilot and pilot survey: the Tully Fisher relation in Eri-
danus, Hydra, Norma, and NGC4636 fields. MNRAS 519(3), 4589–4607 (2023)
16. P. Boubel et al., Cosmic growth rate measurements from Tully–Fisher peculiar velocities (2023).
arXiv e-prints: arXiv:2301.12648
17. E. Opik, An estimate of the distance of the Andromeda Nebula. ApJ 55, 406–410 (1922)
18. C. Balkowski et al., Neutral hydrogen study of spiral and irregular dwarf galaxies. A&A 34,
43–52 (1974)
19. S. Djorgovski, M. Davis, Fundamental properties of elliptical galaxies. ApJ 313, 59–68 (1987)
20. A. Dressler et al., Spectroscopy and photometry of elliptical galaxies—a large-scale streaming
motion in the local universe. ApJ 313, L37–L42 (1987)
21. A. Sandage, G.A. Tammann, Steps toward the Hubble constant V the Hubble constant from
nearby galaxies and the regularity of the local velocity field. ApJ 196, 313–328 (1975)
22. G. de Vaucouleurs, G. Bollinger, The extragalactic distance scale VII the velocity-distance
relations in different directions and the Hubble ratio within and without the local supercluster.
ApJ 233, 433–452 (1979)
12 Tully–Fisher Relation 231

23. S. Zaroubi et al., Wiener reconstruction of the large-scale structure. ApJ 449, 446 (1995)
24. R.B. Tully et al., The Laniakea supercluster of galaxies. Nature 513(7516), 71–73 (2014)
25. R.B. Tully et al., Cosmicflows-4. ApJ 944(1), 94 (2023)
26. H.M. Courtois et al., Gravity in the local universe: density and velocity fields using
CosmicFlows-4. A&A 670, L15 (2023)
27. A. Dupuy, H.M. Courtois, Watersheds of the Universe: Laniakea and five newcomers in the
neighborhood (2023). arXiv e-prints: arXiv:2305.02339
28. P. Andersen et al., Cosmology with peculiar velocities: observational effects. MNRAS 463(4),
4083–4092 (2016)
29. R. Watkins et al., Consistently large cosmic flows on scales of 100h−1 Mpc: a challenge for the
standard ΛCDM cosmology. MNRAS 392(2), 743–756 (2009)
30. R. Watkins et al., Analyzing the large-scale bulk flow using CosmicFlows4: increasing tension
with the standard cosmological model (2023). arXiv e-prints: arXiv:2302.02028
31. A.M. Whitford et al., Evaluating bulk flow estimators for CosmicFlows-4 measurements (2023).
arXiv e-prints: arXiv:2306.11269
32. M. Davis et al., Local gravity versus local velocity: solutions for β and non-linear bias. MNRAS
413(4), 2906–2922 (2011)
33. J. Carrick et al., Cosmological parameters from the comparison of peculiar velocities with
predictions from the 2M++ density field. MNRAS 450(1), 317–332 (2015)
34. J.F. Navarro, M. Steinmetz, Dark halo and disk galaxy scaling laws in hierarchical universes.
ApJ 538(2), 477–488 (2000)
35. M. Vogelsberger et al., Properties of galaxies reproduced by a hydrodynamic simulation. Nature
509(7499), 177–182 (2014)
36. A.A. Ponomareva et al., From light to baryonic mass: the effect of the stellar mass-to-light
ratio on the Baryonic Tully–Fisher relation. MNRAS 474(4), 4366–4384 (2018)
37. J.F. Navarro et al., The structure of cold dark matter halos. ApJ 462, 563 (1996)
38. H.J. Mo, S. Mao, The Tully–Fisher relation and its implications for the halo density profile and
self-interacting dark matter. MNRAS 318(1), 163–172 (2000)
39. U. Seljak, Constraints on galaxy halo profiles from galaxy-galaxy lensing and Tully-
Fisher/fundamental plane relations. MNRAS 334(4), 797–804 (2002)
40. M. Aaronson et al., The infrared luminosity/H I velocity-width relation and its application to
the distance scale. ApJ 229, 1–13 (1979)
41. N. Bouché, S.E. Schneider, IR-TF relation in the zone of avoidance with 2MASS, in Mapping
the Hidden Universe: The Universe behind the Mily Way—The Universe in HI, ed. by R.C.
Kraan-Korteweg et al. Astronomical Society of the Pacific Conference Series, , vol. 218, pp.
111 (2000)
42. K. Said et al., NIR Tully-Fisher in the zone of avoidance—III deep NIR catalogue of the HIZOA
galaxies. MNRAS 462(3), 3386–3400 (2016)
43. M.J. Pierce, R.B. Tully, Distances to the Virgo and Ursa major clusters and a determination of
H0 . ApJ 330, 579 (1988)
44. H.M. Courtois et al., The extragalactic distance database: all digital HI profile catalog. AJ
138(6), 1938–1956 (2009)
45. R. Giovanelli et al., The I band Tully–Fisher relation for cluster galaxies: a template relation,
its scatter and bias corrections. AJ 113, 53–79 (1997)
46. K.L. Masters et al., SFI++ I: a new I-band Tully-Fisher Template, the cluster peculiar velocity
dispersion, and H0 . ApJ 653(2), 861–880 (2006)
47. K. Said et al., NIR Tully–Fisher in the zone of avoidance—II 21 cm H I-line spectra of southern
ZOA galaxies. MNRAS 457(3), 2366–2376 (2016)
48. R. Bell et al., Calibration of the Tully–Fisher relation in the WISE W1 (3.4 µm) and W2 (4.6
μm) bands. MNRAS 519(1), 102–120 (2023)
49. M.A.W. Verheijen, The Ursa major cluster of galaxies V HI rotation curve shapes and the
Tully–Fisher relations. ApJ 563(2), 694–715 (2001)
50. F. Lelli et al., The baryonic Tully-Fisher relation for different velocity definitions and implica-
tions for galaxy angular momentum. MNRAS 484(3), 3267–3278 (2019)
232 K. Said

51. J.A. Willick et al., Homogeneous velocity-distance data for peculiar velocity analysis I cali-
bration of cluster samples. ApJ 446, 12 (1995)
52. J.A. Willick et al., Homogeneous velocity-distance data for peculiar velocity analysis II cali-
bration of field samples. ApJ 457, 460 (1996)
53. D.W. Hogg et al., Data analysis recipes: fitting a model to data (2010). arXiv e-prints:
arXiv:1008.4686
54. M.A. Strauss, J.A. Willick, The density and peculiar velocity fields of nearby galaxies. Phys.
Rep. 261, 271–431 (1995)
55. J. Mould, Understanding the fundamental plane and the Tully Fisher relation. Front. Astron.
Space Sci. 7, 21 (2020)
56. M. Milgrom, A modification of the Newtonian dynamics as a possible alternative to the hidden
mass hypothesis. ApJ 270, 365–370 (1983)
57. F.C. van den Bosch, J.J. Dalcanton, Semianalytical models for the formation of disk galaxies
II dark matter versus modified Newtonian dynamics. ApJ 534(1), 146–164 (2000)
58. F.C. van den Bosch, J.J. Dalcanton, Disk galaxies as cosmological benchmarks: cold dark
matter versus modified Newtonian dynamics, in Galaxy Disks and Disk Galaxies, ed. by J.G.
Funes, E.M. Corsini. Astronomical Society of the Pacific Conference Series, vol. 230, pp.
549–552 (2001)
59. J. Schaye et al., The EAGLE project: simulating the evolution and assembly of galaxies and
their environments. MNRAS 446(1), 521–554 (2015)
60. E. Papastergis et al., An accurate measurement of the baryonic Tully-Fisher relation with
heavily gas-dominated ALFALFA galaxies. A&A 593, A39 (2016)
61. S. Courteau, H.-W. Rix, Maximal disks and the Tully–Fisher relation. ApJ 513(2), 561–571
(1999)
62. N.N.Q. Ouellette et al., The spectroscopy and H-band imaging of Virgo cluster galaxies
(SHIVir) survey: scaling relations and the stellar-to-total mass relation. ApJ 843(1), 74 (2017)
63. B. Abolfathi et al., The Fourteenth data release of the sloan digital sky survey: first spectroscopic
data from the extended baryon oscillation spectroscopic survey and from the second phase of
the apache point observatory galactic evolution experiment. ApJS 235(2), 42 (2018)
64. C. Lintott et al., Galaxy Zoo 1: data release of morphological classifications for nearly 900000
galaxies. MNRAS 410(1), 166–178 (2011)
65. A. Carr et al., The Pantheon+ analysis: improving the redshifts and peculiar velocities of Type
Ia supernovae used in cosmological analyses. PASA 39, e046 (2022)
66. D. Brout et al., The Pantheon+ analysis: cosmological constraints. ApJ 938(2), 110 (2022)
67. A.G. Riess et al., A 2.4% determination of the local value of the Hubble constant. ApJ 826(1),
56 (2016)
68. S.S. McGaugh, The Baryonic Tully–Fisher relation of gas-rich galaxies as a test of ΛCDM
and MOND. AJ 143(2), 40 (2012)
69. F. Lelli et al., The small scatter of the Baryonic Tully–Fisher relation. ApJ 816(1), L14 (2016)
70. D. Obreschkow, M. Meyer, Precise Tully–Fisher relations without galaxy inclinations. ApJ
777(2), 140 (2013)
71. N. Visvanathan, A global value of the Hubble constant. ApJ 275, 430–444 (1983)
72. A. Sandage, G.A. Tammann, The Hubble constant as derived from 21 cm linewidths. Nature
307(5949), 326–329 (1984)
73. A. Sandage, G.A. Tammann, Steps toward the Hubble constant VIII—the global value. ApJ
256, 339–345 (1982)
74. M. Aaronson et al., A distance scale from the infrared magnitude/Hi velocity-width relation
V distance moduli to 10 galaxy clusters, and positive detection of bulk supercluster motion
toward the microwave anisotropy. ApJ 302, 536 (1986)
75. R.C. Kraan-Korteweg et al., 21 Centimeter line width distances of cluster galaxies and the
value of H0 . ApJ 331, 620 (1988)
76. M. Fukugita et al., The distance to the Coma cluster using the B-Band Tully–Fisher relation.
ApJ 376, 8 (1991)
12 Tully–Fisher Relation 233

77. H.J. Rood, B.A. Williams, Tully–Fisher distances to M31-like galaxies in the Coma cluster.
MNRAS 263, 211–228 (1993)
78. M. Bureau et al., A new i-band Tully–Fisher relation for the Fornax Cluster: implication for
the Fornax distance and local supercluster velocity field. ApJ 463, 60 (1996)
79. B.F. Madore et al., The Hubble space telescope key project on the extragalactic distance scale.
XV. a Cepheid distance to the Fornax cluster and its implications. ApJ 515(1), 29–41 (1999)
80. S. Sakai et al., The Hubble space telescope key project on the extragalactic distance scale
XXIV the calibration of Tully–Fisher relations and the value of the Hubble constant. ApJ
529(2), 698–722 (2000)
81. W.L. Freedman et al., Final results from the Hubble Space Telescope key project to measure
the Hubble constant. ApJ 553(1), 47–72 (2001)
82. R.B. Tully, H.M. Courtois, Cosmicflows-2: i-band luminosity-HI linewidth calibration. ApJ
749(1), 78 (2012)
83. J.G. Sorce et al., The Mid-infrared Tully–Fisher relation: calibration of the type Ia supernova
scale and H0 . ApJ 758(1), L12 (2012)
84. J.G. Sorce et al., Calibration of the mid-infrared Tully–Fisher relation. ApJ 765(2), 94 (2013)
85. J.D. Neill et al., The calibration of the WISE W1 and W2 Tully–Fisher relation. ApJ 792(2),
129 (2014)
86. J.G. Sorce et al., From spitzer galaxy photometry to Tully–Fisher distances. MNRAS 444(1),
527–541 (2014)
87. E. Kourkchi et al., Cosmicflows-4: the calibration of optical and infrared Tully–Fisher relations.
ApJ 896(1), 3 (2020)
88. E. Kourkchi et al., Cosmicflows-4: the Baryonic Tully-Fisher relation providing 10000 dis-
tances. MNRAS 511(4), 6160–6178 (2022)
89. A.G. Riess et al., A comprehensive measurement of the local value of the Hubble constant
with 1 km s−1 Mpc−1 uncertainty from the Hubble space telescope and the SH0 ES team. ApJ
934(1), L7 (2022)
90. Planck Collaboration et al., Planck 2018 results VI cosmological parameters. A&A 641, A6
(2020)
91. M. Watanabe et al., Surface photometric calibration of the infrared Tully–Fisher relation using
Cepheid-based distances of galaxies. ApJ 555(1), 215–231 (2001)
92. D.G. Russell, The Ks-band Tully–Fisher relation—a determination of the Hubble parameter
from 218 ScI galaxies and 16 galaxy clusters. J. Astrophys. Astron. 30, 93–118 (2009)
93. L. Bottinelli et al., Cluster population incompleteness bias and the value of HO from the
Tully-Fisher B relation. A&A 181, 1–13 (1987)
94. L. Hislop et al., The extragalactic distance scale without Cepheids IV. ApJ 733(2), 75 (2011)
95. L. Bottinelli et al., The value of HO from the infrared Tully–Fisher relation. A&A 196, 17–25
(1988)
96. J. Schombert et al., Using the Baryonic Tully–Fisher relation to measure Ho . AJ 160(2), 71
(2020)
97. R.B. Tully, M.J. Pierce, Distances to galaxies from the correlation between luminosities and
line widths III cluster template and global measurement of H0 . ApJ 533(2), 744–780 (2000)
98. C. Saulder et al., Target selection for the DESI peculiar velocity survey (2023). arXiv e-prints
arXiv:2302.13760
Chapter 13
Determination of the Local Hubble
Constant Using Giant Extragalactic H ii
Regions and H ii Galaxies

David Fernández-Arenas and Ricardo Chávez

13.1 Introduction

The search for standard candles or rulers to measure cosmic distances is one of the
fundamental efforts in the last decades in the interface of observational astronomy
and cosmology. However, it has been a difficult task for a long time. In this chapter,
we present a historical review of the use of H ii galaxies (HIIGs) and giant H ii
regions (HIIRs) to carry out this task of measuring distances and its application in
the determination of the Hubble constant, from its beginnings to recent years. We
discuss its applicability, its potential physical origin and the associated errors which
can be improved in the future.

“Peculiar” objects from Zwicky list of compact galaxies were identified by [1,
2] and further studied by [3]. Those that have spectral characteristics indistin-
guishable from giant extragalactic H ii regions (like 30 Dor in the LMC) were
designated by them as “Isolated Extragalactic H ii Regions”. Later on, these
compact, dwarf, star-forming galaxies, were dubbed H ii Galaxies [4–6, 42,
43].

D. Fernández-Arenas (B)
Canada France Hawaii Telescope, 65-1238 Mamalahoa Hwy Kamuela, Waimea, HI 96743, USA
e-mail: arenas@cfht.hawaii.edu
R. Chávez
Instituto de Radioastronomía Y Astrofísica, UNAM, Campus Morelia, C.P., 58089 Morelia,
Mexico
e-mail: r.chavez@irya.unam.mx
Consejo Nacional de Ciencia y Tecnología, Av. Insurgentes Sur 1582, 03940 Mexico City, Mexico

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 235
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_13
236 D. Fernández-Arenas and R. Chávez

In the early 1960s, Sersic discovered that the diameters of the largest HIIRs in
spiral galaxies increase with the galaxy luminosity. The use of the HIIRs for the
determination of distances goes back to the mid-70s, with the studies of Sandage
and Tammann [7], mainly with the “classical” application, an empirical correlations
between the diameters of the HIIRs and the parent galaxies luminosity. However,
the correlation was not useful beyond 20 Mpc where the angular diameters become
comparable with the seeing even for the largest HIIRs.
Later, Melnick in 1977 [8] studying the structures of HIIRs, noticed that their
linear diameters also correlated well with the width of their integrated Hα line.
This correlation between the velocity dispersion (line-width) in HIIRs and its size is
analyzed and established in members of the Local Group and M81 and NGC 2403
group allowing a new determination of the distance to M101. In 1978, Melnick [9]
explored the close correlation between the linear diameters of HIIRs in late-type
galaxies and the width of their global Hα emission and used it to establish a close
correlation between this width and the absolute magnitude of the parent galaxies;
this correlation was calibrated in members of the local and NGC 2403 groups, and
then used again to estimate the distance to NGC 300 and M101. The values obtained
were consistent with those reported in the literature at that epoch.
In 1979, the relation between the absolute Hβ fluxes from HIIRs and the corre-
sponding linear diameters and turbulent velocities (width of their integrated Hα line)
is examined and the results presented in [10]. In this study, a strong correlation was
found between the linear diameters and the absolute Hβ luminosities, no longer of
the galaxies but of the HIIRs themselves. However, the large scatter precludes its
application to distance determination with the properties of the dust being largely
responsible for the observed scatter in the relation.

In 1981, Terlevich and Melnick [5] analyzed the correlations between Hβ

luminosities, linear diameters and the widths of the global emission line profiles
of 19 giant extragalactic HIIRs in 6 galaxies and 4 HII galaxies. They found
that the relations: luminosity.∝(linewidth).4 and size.∝(linewidth).2 which are
valid for pressure-supported stellar systems (elliptical galaxies, bulges of spiral
galaxies and globular clusters) are also valid for giant HIIRs.

From this and other considerations, they proposed that giant extragalactic HII
regions are self-gravitating systems in which the observed emission-line profile
widths represent the velocity dispersion of discrete gas clouds in the gravitational
potential of these massive gas-young star complexes.
The analysis also showed that the scatter observed in the Hβ luminosity-linewidth
relation for their sample is largely due to a metallicity effect and after correction for
this effect, the scatter in the relation is reduced with strong implications as a new
method for distance calibrations and although their results are based on a very small
sample, they highlight the need for more and better observations of HIIRs. They
proposed that these correlations can be calibrated using HIIRs in nearby galaxies
13 Determination of the Local Hubble Constant … 237

for which the distances are known to obtain distances to field spirals and irregular
galaxies beyond the Virgo cluster.
From that moment on, the empirical correlations among integrated parameters of
HIIRs are considered to have fundamental importance for the understanding of the
nature and origin of these objects and represent an independent indicator of distance.
In 1987, with a new homogeneous sample of data for 22 HIIRs in 10 galaxies with
known distances, Melnick et al. [11] studied the different correlations among inte-
grated parameters of HIIRs. They confirmed the existence of significant correlations
between core radius, emission line width, Hβluminosity, and oxygen abundance,
providing a powerful distance indicator for galaxies with HIIRs. Since, the distance
indicator of HIIRs relies on the correlations between Hβ luminosity, the emission
line-profile widths and the oxygen abundance of giant HIIRs, they proposed a pos-
sible physical scenario where the stellar winds and gravity are the most promising
mechanisms for explaining the observations, although not explaining completely the
presence of supersonic motions in the ionized gas.
In 1987, Melnick et al. [12] proposed that the most useful application of HIIRs
to the determination of distances, is the correlation find by [5], which, we will call
the L(Hβ) − σ relation. Using this extragalactic distance indicator tied to galactic
scale via cepheid variables, 41 HIIGs from the group’s spectrophotometric catalogue
[13] and the zero point calibrated with the giant HIIRs in nearby spiral galaxies, for
the first time a value of H0 =.95 ± 10 km s−1 Mpc−1 was obtained for the Hubble
parameter, with the distance calibrator given by:

σ5
. log L(Hβ) = log + 41.32 (13.1)
(O/H)

A year later, Melnick et al. [14] calibrated the correlation between luminos-
ity, line width and oxygen abundance. They reported a value for H0 of .89.0 ± 10
km s−1 Mpc−1 , after correcting the flows for Malmquist bias and the radial velocities
for the movement of the Local Group. In this study, they discuss the initial potential
problems related to the application of the method: the zero point errors and the curva-
ture of the L(Hβ) − σ relation. An underestimate of the zero point may arise if many
HIIGs are composed of more than one HIIR superimposed along the line of sight.
On the other hand, the curvature related to the broadening of the emission lines may
be affected by the gravity of the overall galaxy or systems with rotation. If present,
this effect will saturate the L(Hβ) − σ relation at high luminosities where the line
broadening will be dominated by the gravity of the old stellar component. Excluding
systems with a linewidth .> 60 km s−1 the correlation between Hβ luminosity, the
width of the line and oxygen abundance, HIIGs show the same functional form and
the same scatter as that for local HIIRs, suggesting that any broadening of the lines
by a source not related to the young stellar population must be small. Additionally,
by imposing the condition that the equivalent width of Hβ.> 30Å, effectively the
number of galaxies where the light of the young component is severely affected by
the underlying older stellar population is reduced.
238 D. Fernández-Arenas and R. Chávez

In the last two decades, with the advance of technology in new detectors, the
new generation of big telescopes and given the potential use of HIIRs and HIIGs as
distance indicators, a strategy to measure the Hubble constant and to constrain cos-
mological parameters was devised. It is a long-term program using HIIGs, as standard
candles that can, in principle, be applied out to .z ∼ 3.5 [15–17] with present-day
ground-based instrumentation and up to .z ∼ 9 with the Near Infrared Spectrograph
(NIRSpec) component of the James Webb Space Telescope (JWST) ([18, 19] and
Chavez et al. in preparation). Let’s remember that the potential of HIIGs as distance
indicators stems from the existence of a correlation between the luminosity of a
hydrogen recombination line, e.g. L(Hβ) (proportional to the number of ionising
photons), and the line width of the emission line, .σ .
In what follows, we show the latest advances in the description of the L(Hβ) − σ
relation, its calibrations with new data sets and the application to the determination
of the local value of the Hubble constant.

13.2 The L-.σ Relation

Two of the more recent works related to the L(Hβ) − σ relation for HIIGs have been
presented in the works of Bordallo and Telles [20] and Chávez et al. [21] in 2011
and 2014, respectively. They presented new observational datasets of high- and low-
spectral resolution. After multi-parametric fits, they found that the Hβ equivalent
width is more important for determining the scatter in the L(Hβ) − σ relation than
the dependence on metallicity, contrary to the suggestion by Terlevich et al. [5].
Assuming that the EW(Hβ) is an indicator of the burst’s age, the interpretation is
that the age of the underlying older stellar population could be contaminating the
luminosity of a recent dominant young burst more than its velocity dispersion. Chávez
et al. [21] show that the majority of the scatter in the L(Hβ) − σ relation is reduced
using the size of the burst as a second parameter, which allows us to think of a
fundamental plane defined by the luminosity, velocity dispersion and size in HIIGs
and HIIRs.
Both studies conclude that, in addition, part of the scatter in the L(Hβ) − σ
relation can be reduced by using the objects with the most Gaussian profiles in their
emission lines. However, the cost of reducing the samples is of the order of 50% using
this criterion, which can be parameterized in different ways, (see Sect. 4.2 in [20]
and sec 6.1 in [21]). The main parametric relations presented in the above papers are
reproduced in Table 13.1. According to different settings, constraints are placed on
the slope between 3 and 5, which could be consistent with a fundamental plane similar
to that followed by elliptical galaxies and old star systems supported by gravity. A
study on the evolution of these young systems and their scale relationships compared
to old systems can be found in Terlevich et al. [22]. However, the existence of the
L − σ relation for HIIRs and HIIGs poses an intriguing question about the origin
of the supersonic velocity widths of the emission line profiles. A simple explanation
would be that metal rich galaxies produce more massive starbursts, and this should
13 Determination of the Local Hubble Constant … 239

Table 13.1 Fits for the L − σ relation in local HIIGs

.log L = α + β × log σ

Linear Intercept (.α) Slope (.α) rms N

Regression
(BT11, Here the luminosity of Hα has been used instead of Hβ luminosity)
OLS(Y.|X) 36.21 .± 0.32 3.01 .± 0.23 0.37 81.a
OLS(X.|Y) 34.52 .± 0.38 4.18 .± 0.27
OLS Bisector 35.49 .± 0.32 3.51 .± 0.23
OLS(Y.|X) 35.29 .± 0.42 3.72 .± 0.31 0.31 53.b
OLS(X.|Y) 33.73 .± 0.47 4.85 .± 0.34
OLS Bisector 34.61 .± 0.41 4.22 .± 0.30
OLS(Y.|X) 34.80 .± 0.41 4.14 .± 0.29 0.29 37.c
OLS(X.|Y) 33.45 .± 0.53 5.13 .± 0.38
OLS Bisector 34.19 .± 0.43 4.58 .± 0.30
(C14)
Error in both axes 33.71 .± 0.21 4.65 .± 0.14 0.332 107.d
Error in both axes 33.22 .± 0.27 4.97 .± 0.17 0.332 69.e
a
. All galaxies with homogeneous spectrophotometry in. .b The sample showing regular Gaussian
profiles. .c The sample showing regular Gaussian profiles and homogeneous data. .d All galaxies with
homogeneous data in. .e The sample showing regular Gaussian profiles

be verified by observations. Another hint about the origin of velocity widths found in
HIIGs is provided by the systematic difference found between the width of Balmer
lines and [O iii] lines, the most recent studies relating to “The L − σ relation for
HIIGs in green” can be found in Melnick et al. [23]. A possible explanation would
be the existence of a well behaved ionization structure, in which more excited ions
are concentrated closer to the ionizing sources and densest regions.

13.3 The Data

The use of the L − σ relation as a distance indicator and as a tool to derive the Hubble
constant requires accurate determination of both the luminosity and the FWHM or
velocity dispersion of the emission lines in giant HIIRs and HIIGs. Observationally
HIIGs and HIIRs represent the youngest Super Star Clusters that can be observed in
any detail. The measurements that are required for the calibration of the L(Hβ) − σ
relation are the following:
1. Narrow-band imaging or low-resolution-wide slit spectroscopy to obtain the
integrated Hβ flux.
2. High-resolution spectroscopy to measure the velocity dispersion from the Hβ
(or [O iii] for comparison) line profile.
240 D. Fernández-Arenas and R. Chávez

Fig. 13.1 Hα image obtained from NASA/IPAC Extragalactic Database (NED), high-resolution
profile for the giant HIIR (IC10: 101) and low-resolution spectrum. Taken from [24]

In Fig. 13.1, we show as an example the data for one HIIR in a nearby galaxy used
in the calibration of the zero-point of the L − σ relation.
To guarantee the best-integrated spectrophotometry and the youngest objects, the
following selection criteria are applied:

• A lower limit for the equivalent width, EW(Hβ).> 50Å(HIIGs and giant HIIRs).
• An upper limit to the Balmer line widths, .log σ < 1.8 km s−1 (HIIGs and giant
HIIRs).
• Redshift range .0.01 < z < 0.2 (HIIGs).
• Petrosian diameter less than 6 arcsecs (HIIGs).

Figure 13.2 shows typical HIIG in the local universe and their high-resolution
spectrum dominated by strong emission lines.
The above criteria allow us to exclude highly evolved regions, to diminish con-
tamination by an underlying older stellar population and to avoid objects with a high
rate of escape of ionizing photons. They also minimize the possibility of including
systems supported by rotation or with multiple young ionizing clusters as it was dis-
cussed in [14]. The lower redshift limit in HIIGs is selected to avoid nearby objects
that are more affected by local peculiar motions relative to the Hubble flow and the
upper limit was set to minimize cosmological non-linearity effects.
13 Determination of the Local Hubble Constant … 241

Fig. 13.2 A selection of colour images of HIIGs from the sample of [21]. The SDSS name and the
index number are indicated in the stamps. The changes in colour are related to the redshift of the
object. In the right panel: Examples of the high dispersion spectra obtained for the same object with
Subaru HDS (top) and VLT UVES (bottom), showing the region covering Hβ and the [O iii] lines
at .λλ 4959,5007 Å. The instrumental profile is shown in red at the left of each spectrum. Taken
from [21]

In Fig. 13.3, we can see the evolution of the Hβ equivalent width for a simple
instantaneous burst with metallicity . Z = 0.004. In the right panel, we show the
location of the HIIGs in the BPT diagrams [25] which are widely used to differentiate
the excitation mechanism for star-forming galaxies from AGN, the EW criterion
clearly can separate between other similar star-forming ionizing mechanism as Blue
Compact Dwarfs and also show green peas galaxies as a class of HIIGs discovered
using different selection criteria. Interesting, in recent years, making evolution studies
of these objects the interpretation has been that these young massive stellar clusters,
HIIGs and GHIIRs can evolve to form globular clusters and ultracompact dwarf
ellipticals in about 12 Gyr so that present-day globular clusters and ultracompact
dwarf ellipticals could have originated in conditions similar to those observed in today
GHIIR and HIIG [22]. However, this is a topic for another discussion.

13.3.1 Giant HII Regions and HII Galaxies

HIIG are compact and massive systems experiencing luminous bursts of star forma-
tion generated by the formation of young super clusters (SSCs) with a high luminosity
per unit mass and with properties similar, if not identical, to giant HIIRs. The poten-
tial of GHIIRs as distance indicators was originally realized from the existence of
a correlation between the giant HIIR diameter and the luminosity [28, 29] see also
[30].
242 D. Fernández-Arenas and R. Chávez

Fig. 13.3 The evolution of the Hβequivalent width for an instantaneous burst with metallicity. Z =
0.004 and a Salpeter IMF with an upper limit of 100 Mʘ [26]. The blue area marks the Hβ equivalent
width of 50Å, corresponding to an age of .∼ 5 Myrs. In the right plot: BPT diagram showing the
high excitation level of a sample of HIIGs selected mainly as having high equivalent width in their
Balmer emission lines. The solid line represents the upper limit for stellar photoionization, from
[27]. Taken from Fernandez PhD Thesis, 2018

Most HIIG were discovered in objective-prism surveys thanks to their strong

narrow emission lines. Currently, in spectroscopic surveys like Sloan Digital Sky
Survey (SDSS), they are selected by very large equivalent widths in the Balmer
lines. Since the luminosity of HIIG is dominated by the starburst component they
can be observed even at large redshifts becoming interesting standard candles.
So far, we have analyzed a sample of HIIG and GHIIRs containing 217 objects,
which can be split into: the anchor sample (36 GHIIRs in 13 local galaxies with
distances from primary indicators, [24, 44]), the local sample of HIIG (107 with
. z < 0.16, [21]) and a high-. z sample based on 29 KMOS, 15 MOSFIRE, 6 XShooter

and 24 literature objects (see [31, 32] for more details). We used these data to
constrain cosmological parameters and to determine the local value of the Hubble
constant by means of the distance indicator defined by the L − σ relation.

13.4 Methodology

The Hubble constant is determined as follows: first we fix the slope of the L −
σ relation using the velocity dispersion and luminosity of the HIIGs. The slope is
independent of the actual value of H0 .
To estimate the Hubble constant we use the slope (.α) of the L − σ relation of the
HIIG and new GHIIR data (anchor sample) to calibrate the zero point (. Z p ) of the
distance indicator (see Fig. 13.4) as follows:
13 Determination of the Local Hubble Constant … 243

Fig. 13.4 L − σ relation for the GHIIRs. The adopted distances to the parent galaxies come from
Cepheids. The green solid line is the fit to the data given in the inset and the dashed line is the fit.
Taken from [24]

∑36
Wi (log L GHR,i − α × log σGHR,i )
. Zp = i=1
∑36 (13.2)
i=1 Wi

where . L G H R,i is the H.β luminosity of each GHIIR and .σGHR,i is the corresponding
velocity dispersion. The statistical weights .Wi are calculated as:
( δL GHR,i )2 ( δσGHR,i )2
. Wi−1 = 0.4343 + 0.4343α + (δα)2 (σGHR,i − < σHIIG >)2
L GHR,i σGHR,i
(13.3)
where .< σHIIG > is the average velocity dispersion of the HIIG that define the
slope of the relation. Thus, the calibrated L − σ relation or distance estimator is:
.log L(Hβ) = α log σ + Z p . To calculate the Hubble constant we minimise the func-
tion,
∑N
.χ (H0 ) = [Wi (μi − μ H0 ,i )2 − ln(Wi )]
2
(13.4)
i=1

where .μi is the logarithmic distance modulus to each HIIG calculated using the
distance indicator and the H.β flux F(H.β) as
244 D. Fernández-Arenas and R. Chávez

μi = 2.5[Z p + α × log σi − log Fi (Hβ) − log 4π ]

. (13.5)

and .μ H0 ,i is the distance modulus calculated from the redshift using either the linear
relation . D L = zc/H0 or the full cosmological prescription with .ΩΛ = 0.71.
The best value of H0 is then obtained minimising .χ 2 with statistical weights
−1
. Wi = δμi2 + δμ2H0 ,i calculated as,
( δ Fi )2 ( δσi )2
. Wi−1 = 6.25[(δ Z p )2 + 0.4343 + 0.4343α +(δα)2 (σi − < σ >)2 ]
Fi σi
(13.6)

13.5 Systematics

Genuine systematic errors are difficult to estimate. However, we have explored and
listed a range of parameters to quantify at least part of the systematic error component.
In particular, we present different systematic errors that can be taken into account in
order to determine the value of H0 using giant HIIRs and HIIGs, these effects are
described in the following.

13.5.1 Age Effects

Even in our sample of HIIGs and giant HIIRs, chosen to be the youngest systems,
it is crucial to verify that the rapid luminosity evolution of the stellar cluster does
not introduce a systematic bias in the distance indicator. This would happen, for
example, if the average age of the giant HIIRs is different from that of the HIIGs
or if luminous and faint HIIGs have different average ages. In general, if velocity
dispersion measures mass, younger clusters will be more luminous than older ones
for a given velocity dispersion. The evolution effect can be scrambled by the super-
position of bursts of different ages along the line of sight. Nevertheless, even small
systematics can have a sizable effect on the value of H0 so it is important to remove
the evolution effect from the data in a similar fashion as for the dust extinction.
The equivalent width of Hβ [EW(Hβ)] is a useful age estimator [33–35] or at least
it provides an upper limit of the age of the burst [36]. Indeed there is some empirical
evidence for this as discussed by [16] and [20]. [21] explored the possibility that
the age of the burst is a second parameter, using the EW(Hβ) as an age estimator in
the L(Hβ) − σ relation for HIIGs, and found a rather weak dependence. Different
evolution corrections have been explored in recent years, e.g. [24] and [37]. This
correction must be taken into account in order to improve the estimates of the value
of H0 but it must still be explored in more detail with better observations especially
to characterize the underlying older population.
13 Determination of the Local Hubble Constant … 245

13.5.2 Extinction Effects

One of the major contributions to the systematic errors is related to extinction cor-
rection, in particular the use of different extinction laws. In particular,[38] law yields
larger values of H0 than [39] typically by about 1.5.∼km s−1 Mpc−1 . As the [38]
law was derived from a sample of eight heterogeneous starburst galaxies where
only two, Tol 1924–416 and UGCS410 are bonafide HIIGs and the rest are evolved
high metallicity starburst galaxies, we prefer to use the [39] extinction curve which
corresponds to the LMC2 supershell near the 30 Dorado star-forming region, the
prototypical giant HIIR, in the Large Magellanic Cloud.

13.5.3 Metallicity Effects

In 1981, Terlevich et al. [5] proposed that oxygen abundance is a good indicator of
the long-term evolution of the system. They proposed a simple ‘closed box’ chemical
evolution model with many successive cycles of star formation. In it, for each cycle,
evolution is traced by the EW(Hβ) whereas the long-term evolution of the system,
spanning two or more cycles, could be traced by the oxygen abundance, which then
becomes a plausible second parameter in the L(Hβ) − σ correlation.

13.5.4 Aperture Effects

The sensitivity to Hβ photometry could be crucial for the determination of the lumi-
nosity especially when multiple bursts of star-forming regions and the nearest HIIGs
are present. This effect was explored by [24] by comparing SDSS and [21] photo-
metric measurements. The results reveal that the values of H0 are on average sys-
tematically lower by about 4.9.∼km s−1 Mpc−1 for SDSS fluxes and no evolutionary
corrections. This systematic difference is reduced to 1.7.∼km s−1 Mpc−1 when evo-
lutionary corrections are included. The smaller value of H0 for the SDSS photometry
is related to the systematically steeper slope of the L(Hβ) − σ relation compared
with that obtained when using [21] photometry, both with and without evolution
correction. Since the lower luminosity HIIGs are also the closer ones, the steepening
is a consequence of the smaller SDSS aperture, compared with that of [21], under-
estimating the line fluxes of the nearest galaxies. This effect may also introduce a
bias in luminosity or velocity dispersion due to the multiplicity seen in some objects
(e.g. J084220+115000) as presented by [40].
246 D. Fernández-Arenas and R. Chávez

Fig. 13.5 Left: The L − σ relation for the [21] sample using the velocity dispersions in the original
paper; the fluxes have been corrected using [39] extinction law. The solid line is the fit to the HIIGs.
The inset equation is the distance indicator where the slope is obtained from the fit to the HIIG and the
zero point is determined following the procedure described in the text. In the right panel: Our main
result incorporating the evolution correction, is H0 =.71.0 ± 3.5 km s−1 Mpc−1 (random+systematic)
a value that is between the most recent determination from Planck and SNIa. Taken and updated
from [24]. We also plot the most recent results derived from the cosmological constrain using HIIGs
from Gónzalez-Morán et al. [32]

13.5.5 Redshift Effects

An upper redshift cutoff of .z = 0.1 was set (S2) instead of .z = 0.16 (S1) for the
HIIGs. Comparing S1 (in which the distances were computed using a flat cosmol-
ogy with .Ωm = 0.29) with sample S2, using the linear Hubble relation for the dis-
tances, we found that using .z < 0.1 reduces the value of H0 typically by about 2
−1
.∼km s Mpc−1 with a range from 1.1 to 3.3 km s−1 Mpc−1 . The sensitivity of H0
to the actual value of .Ωm is low, amounting in our case to an uncertainty of about
0.1% in H0 for an uncertainty in .Ωm of 0.02 [41]. This, together with the larger
uncertainty when using S2 drives us to use S1 with the distance determined using a
flat cosmology with .Ωm = 0.29. This reinforces the point that we are not biasing the
results by using additional cosmological parameters.

13.6 Cosmological Constraints

To calculate the parameters of the L − σ relation in a unified way including HIIGs

and GHIIRs, we define the following likelihood function:
13 Determination of the Local Hubble Constant … 247

1
L ∝ exp(− χ H2 I I )
. (13.7)
2
where:
∑ (μ0 (log f, log σ |α, β) − μθ (z|θ ))2
χ H2 I I =
. (13.8)
n
∈2

where .μ0 is the distance modulus calculated from a set of observables as:

μ0 = 2.5(α + β log σ − log f − 40.08)

. (13.9)

α and .β are the L − σ relation’s intercept and slope, respectively, .log σ is the
.
logarithm of the measured velocity dispersion and log f is the logarithm of the
measured flux. For HIIG the theoretical distance modulus, .μθ , is given as:

μθ = 5 log d L (z, θ ) + 25
. (13.10)

where .z is the redshift, .d L is the luminosity distance in Mpc and .θ is a given set
of cosmological parameters. For GHIIRs, the value of .μθ is inferred from primary
indicators and finally .∈ 2 are the weights in the likelihood function.
The luminosity distance .d L of the sources tracing the Hubble expansion is
employed to calculate the theoretical distance moduli. We define, for convenience,
an extra parameter independent of the Hubble constant as:
{ z
dz '
. D L (z, θ ) = (1 + z) (13.11)
0 E(z ' , θ )

i.e., .d L = cD L /H0 . . E(z.θ ) for a flat Universe is given by:

( )
−3wa z
. E (z, θ ) = Ωr (1 + z) + Ωm (1 + z) + Ωw (1 + z) exp
2 4 3 3y
(13.12)
1+z

with. y = (1 + w0 + wa ). The parameters.w0 and.wa refer to the dark energy equation

of state (DE EoS) the general form of which is:

. pw = w(z)ρw (13.13)

with.ρw the pressure and.ρw the density of the postulated Dark Energy fluid. Different
DE models have been proposed and many are parametrized using a Taylor expansion
around the present epoch:
z
w(a) = w0 + wa (1 − a) =⇒ w(z) = w0 + wa
. (13.14)
1+z

The cosmological constant is just a special case of DE, given for .(w0 , wa ) = (−1, 0),
while the so called wCDM models are such that.wa = 0 but.w0 can take values./= −1.
248 D. Fernández-Arenas and R. Chávez

Fig. 13.6 Hubble diagram connecting our local and high redshift samples up to .z ∼ 2.6. Red
circles represent averages of the distance moduli in redshift bins. The continuous line corresponds
to .Ωm = 0.249 and .w0 = −1.18 (our best cosmological model using only HIIG). The insets show
the distribution of the residuals of the fit that are plotted in the bottom panels. Taken from [32]

13.7 Conclusions

The accurate determination of the Hubble constant, H0 , is considered one of the

most fundamental tasks in the interface between astronomy and cosmology and
the L(Hβ) − σ relation is a powerful independent distance indicator that has to be
explored with more precise data in the future. Here we summarise the main results
of the last year in the use of the HIIRs and HIIGs, the Hubble constant and the
cosmological constraints.
This chapter presents the determination of the local value of the Hubble constant
and observational constraints of the cosmological parameters making use of the
GHIIRs in nearby galaxies and HIIG local and at high-.z, the most recent results are
summarized in the Figs. 13.4, 13.5 and 13.6, where we describe the local L(Hβ) − σ
relation followed by GHIIRs, the determination of the local value of the Hubble
constant, the local L(Hβ) − σ relation, the Hubble diagram tracing local GHIIRs
up to high-.z HIIG, the space of solutions in the cosmological parameters and the
joint analysis of the other tracers of the Hubble expansion.
From the analysis of different systematic effects, we can infer that they cause
changes in the L(Hβ) − σ relation that translate into r.m.s. variations in the value of
H0 of 1.5 to 2.5 km s−1 Mpc−1 .
On the other hand, combining HIIGs (the L − σ relation as distance indi-
cator), CMB and BAO yields: .Ωm = 0.298 ± 0.012 and .w0 = −1.005 ± 0.051,
fully consistent with the .ΛCDM model. It is clear that the solution space of
HIIG/CMB/BAO, although less constrained, is certainly compatible with the solution
space of SNIa/CMB/BAO.
13 Determination of the Local Hubble Constant … 249

Acknowledgements D.F.A acknowledges the support from the National Science Foundation under
grant 2109124 for SIGNALS: Unveiling Star-Forming Regions in Nearby Galaxies. R.C and D.F.A
gratefully acknowledge support from the CONAHCyT of Mexico under the grant CF2022-320152.
D.F.A and R.C. would like to thank professors Roberto Terlevich and Elena Terlevich for their
constant support throughout these years and for their valuable suggestions, which have allowed us
to improve the chapter that we present here.

References

1. W.L.W. Sargent, Astrophys. J. 159, 765 (1970). https://doi.org/10.1086/150353

2. W. L. W. Sargent, The Astrophysical Journal, 160, no. 405 (1970) https://doi.org/10.1086/
150443
3. W.L.W. Sargent et al., Astrophys. J. 162, L155 (1970). https://doi.org/10.1086/180644
4. A. Campbell et al., Mon. Not. R. Astron. Soc. 223, 811–825 (1986). https://doi.org/10.1093/
mnras/223.4.811
5. R. Terlevich, J. Melnick et al., Mon. Not. R. Astron. Soc. 195, 839–851 (1981). https://doi.org/
10.1093/mnras/195.4.839
6. J. Maza et al., Astron. Astrophys. Suppl. Ser. 89, 389 (1991)
7. A. Sandage, G.A. Tammann, Astrophys. J. 190, 525–538 (1974). https://doi.org/10.1086/
152906
8. J. Melnick, Astrophys. J. 213, 15–17 (1977). https://doi.org/10.1086/155122
9. J. Melnick, Astron. Astrophys. 70, 157 (1978)
10. J. Melnick, Astrophys. J. 228, 112–117 (1979). https://doi.org/10.1086/156827
11. J. Melnick et al., Mon. Not. R. Astron. Soc. 226, 849–866 (1987). https://doi.org/10.1093/
mnras/226.4.849
12. J. Melnick et al., Revista Mexicana de Astronomia y Astrofisica 14, 158–164 (1987)
13. R. Terlevich et al., Astron. Astrophys. Suppl. Ser. 91, 285 (1991)
14. J. Melnick et al., Mon. Not. R. Astron. Soc. 235, 297–313 (1988). https://doi.org/10.1093/
mnras/235.1.297
15. R. Terlevich et al., Mon. Not. R. Astron. Soc. 451(3), 3001–3010 (2015). https://doi.org/10.
1093/mnras/stv1128
16. J. Melnick et al., Mon. Not. R. Astron. Soc. 311(3), 629–635 (2000). https://doi.org/10.1046/
j.1365-8711.2000.03112.x
17. M. Plionis et al., Mon. Not. R. Astron. Soc. 416(4), 2981–2996 (2011). https://doi.org/10.1111/
j.1365-2966.2011.19247.x
18. de Graaff et al. (2023), arXiv e-prints: arXiv: 2308.09742, https://doi.org/10.48550/arXiv.2308.
09742
19. Bunker et al. (2023), arXiv e-prints: arXiv:2306.02467, https://doi.org/10.48550/arXiv.2306.
02467
20. V. Bordalo, E. Telles et al., Astrophys. J. 735(1), 52 (2011). https://doi.org/10.1088/0004-
637X/735/1/52
21. R. Chávez et al., Mon. Not. R. Astron. Soc. 442(4), 3565–3597 (2014). https://doi.org/10.1093/
mnras/stu987
22. E. Terlevich et al., Mon. Not. R. Astron. Soc. 481, 268–276 (2018). https://doi.org/10.1093/
mnras/sty2325
23. J. Melnick et al., Astron. Astrophys. 599(A76) (2017).https://doi.org/10.1051/0004-6361/
201629728
24. D. Fernández Arenas et al., Mon. Not. R. Astron. Soc. 474(1), 1250–1276 (2018). https://doi.
org/10.1093/mnras/stx2710
25. J.A. Baldwin et al., Publ. Astron. Soc. Pac. 93, 5–19 (1981). https://doi.org/10.1086/130766
250 D. Fernández-Arenas and R. Chávez

26. C. Leitherer et al., Astrophys. J. Suppl. Ser. 123(1), 3–40 (1999). https://doi.org/10.1086/
313233
27. L.J. Kewley et al., Astrophys. J. 556(1), 121–140 (2001). https://doi.org/10.1086/321545
28. A. Sandage, Problems of extra-galactic research 15, 359 (1962)
29. J.L. Sérsic, Zeitschrift fur Astrophysik 50, 168 (1960)
30. R.C. Kennicutt, Astrophys. J. 228, 394–404 (1979). https://doi.org/10.1086/156858
31. A.L. González-Morán et al., Mon. Not. R. Astron. Soc. 487(4), 4669–4694 (2019). https://doi.
org/10.1093/mnras/stz1577
32. A.L. González-Morán et al., Mon. Not. R. Astron. Soc. 505(1), 1441–1457 (2021). https://doi.
org/10.1093/mnras/stab1385
33. H.A. Dottori, E.L.D. Bica et al., Astron. Astrophys. 102(2), 245–249 (1981)
34. G. Stasińska, C. Leitherer et al., Astrophys. J. Suppl. Ser. 107, 661 (1996). https://doi.org/10.
1086/192377
35. M.L. Martín-Manjón et al., Mon. Not. R. Astron. Soc. 385(2), 854–866 (2008). https://doi.org/
10.1111/j.1365-2966.2008.12875.x
36. R. Terlevich et al., Mon. Not. R. Astron. Soc. 348(4), 1191–1196 (2004). https://doi.org/10.
1111/j.1365-2966.2004.07432.x
37. E. Telles, J. Melnick et al., Astron. Astrophys. 615 (A55) (2018). https://doi.org/10.1051/0004-
6361/201732275
38. D. Calzetti et al., Astrophys. J. 533(2), 682–695 (2000). https://doi.org/10.1086/308692
39. K.D. Gordon et al., Astrophys. J. 594(1), 279–293 (2003). https://doi.org/10.1086/376774
40. D. Fernández Arenas et al., Mon. Not. R. Astron. Soc. 519(3), 4221–4240 (2023). https://doi.
org/10.1093/mnras/stac3309
41. M. Betoule et al., Astron. Astrophys. 568 (A22) (2014). https://doi.org/10.1051/0004-6361/
201423413
42. R.J. Terlevich, Star-forming dwarf galaxies and related objects, 395–402 (1986)
43. M.L. Martín-Manjón et al., Mon. Not. R. Astron. Soc. 385(2), 854–866 (2008). https://doi.org/
10.1111/j.1365-2966.2008.12875.x
44. R. Chávez et al., Mon. Not. R. Astron. Soc. 425(1), L56–L60 (2012). https://doi.org/10.1111/
j.1745-3933.2012.01299.x
Chapter 14
Strong Lensing and H0

Tommaso Treu and Anowar J. Shajib

14.1 Introduction

Following Fermat’s Principle, multiple images of the lensed source form at stationary
points of the arrival time surface. The difference in arrival time between multiple
images arises from the combination of the geometric delay, which makes some light
paths longer than others, and the Shapiro delay [94], arising from the difference in
gravitational potential along different paths.
Thus, if one can measure the difference in arrival time and reconstruct the gravita-
tional potential, one obtains an absolute measurement of the difference in the length
of the light paths. It is then a simple matter to convert this measurement into a com-
bination of angular diameter distances, known as the “time-delay distance” [98], and
thus obtain a direct measurement of the Hubble constant, H0 [76].
The principle is elegant and straightforward, and Refsdal [76] recognized its
potential well before strong gravitational lenses were discovered in the late 70 s
[117]. Since then, breakthroughs in observations, methodology, and theory have
finally made Resfdal’s dream a reality. Time-delay cosmography, as it has come to
be called, has been demonstrated to yield measurements of H0 at the level of precision
and accuracy of a few percent for a single strong lens system [88, 98]. As strong
lens systems are being discovered and followed up at an increasing pace, the overall
precision achievable by combining multiple systems is improving rapidly. Progress
has also been achieved in understanding systematic errors [7, 28, 38, 39, 64, 114]

T. Treu (B)
Physics and Astronomy Department, University of California, Los Angeles, CA 90095, USA
e-mail: tt@astro.ucla.edu
A. J. Shajib
NHFP Einstein Fellow, Department of Astronomy and Astrophysics, University of Chicago,
Chicago, IL 60637, USA
e-mail: ajshajib@uchicago.edu
Kavli Institute for Cosmological Physics, University of Chicago, Chicago, IL 60637, USA

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 251
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_14
252 T. Treu and A. J. Shajib

and how to control them so as to achieve unbiased sample averages at the level of
1–2% in accuracy and precision required to help settle the Hubble tension [9].
Time-delay cosmography has several advantages as a probe of H0 . This method
provides a one-step measurement of H0 that is completely independent of other
H0 probes. The measured cosmological distances are angular diameter distances.
Thus, the method is not susceptible to uncertain dust extinction laws that luminosity
distance indicators could be susceptible to [31].
In this Chapter, we first provide some background in the theory and history of
the method in Sect. 14.2. We then review the current state of the art in Sect. 14.3,
by covering two recent case studies in some detail. First, in 14.3.1 to Sect. 14.3.4,
we describe the case of RXJ1131–1231, a galaxy deflector lensing a quasar, which
has been modeled based on excellent data, including ground-based spatially resolved
kinematics. Second, in Sect. 14.3.5 we described the case of supernova (SN) Refsdal,
multiply imaged by a foreground cluster, the first lensed SN that has been used
to measure . H0 . We conclude with our future outlook in Sect. 14.4. Due to space
limitations, we only focus on the main points currently relevant to the measurement
of H0 and refer to previous reviews for more extensive treatments of the theory and
history of the method [8, 102, 107, 110].

14.2 Background

14.2.1 Theory

This section briefly reviews the theory of time-delay cosmography. See [85] for a
detailed treatment of the strong lensing formalism.

14.2.1.1 Strong Lensing Formalism

The lensing phenomenon is described by the lens equation

β = θ − α(θ)
. (14.1)

that maps a source plane coordinate vector .β to an image plane coordinate vector .θ,
where .α(θ ) is the deflection angle vector. The lensing deflection is produced by the
mass distribution between the source and the observer. In the thin lens approximation,
the lensing mass distribution is described by the surface mass density .∑(θ ) projected
on the lens plane. The dimensionless lensing convergence is defined as .κ ≡ ∑/∑cr ,
where the critical density .∑cr is given by

c2 Ds
.∑cr ≡ . (14.2)
4πG Dd Dds
14 Strong Lensing and H0 253

Here, . Dd is the angular diameter distance between the observer and the deflector,
Ds is the angular diameter distance between the observer and the source, and . Dds
.
is the angular diameter distance between the deflector and the source. The angular
diameter distance between two redshifts .z 1 and .z 2 is given by
c
. DA (z 1 , z 2 ) = f k (z 1 , z 2 , Θ), (14.3)
H0 (1 + z 2 )

where .Θ is the set of cosmological parameters excluding H0 in a given cosmology

and . f k (z 1 , z 2 , Θ) is a function whose form depends on the sign of the curvature
density .Ωk [73, 118].
The deflection angle .α is related to the convergence .κ as

1
κ=
. ∇ · α, (14.4)
2
and to the lensing potential .ψ as

α(θ) = ∇ψ(θ ).
. (14.5)

Thus, the lensing potential is connected to the surface mass distribution by the
two-dimensional Poisson equation:

.∇ 2 ψ(θ ) = 2κ(θ). (14.6)

We can define the Fermat potential [10, 84] as

(θ − β)2
. τ (θ , β) = + ψ(θ ). (14.7)
2
The time delay between the arrival times of photons for images A and B is given by

DΔt
.Δt = [τ (θ A ) − τ (θ B )] , (14.8)
c
where . DΔt is the “time-delay distance” [76, 98] defined as

Dd Ds
. DΔt ≡ (1 + z d ) . (14.9)
Dds

The time delay .Δt can be measured for transient sources such as SNe or quasars. The
time-delay distance is inversely proportional to the Hubble constant . H0 . Therefore,
if we can measure the time delay .Δt and the lensing potential .ψ, we can measure
. H0 . The two unknowns .Δψ and .β in Eq. 14.8 are obtained by modeling the data. The

practical aspects of lens modeling with imaging data are described in Sect. 14.3.2.
254 T. Treu and A. J. Shajib

The next section discusses the well-known “mass-sheet degeneracy” [MSD 32, 85].
Limiting the MSD is crucial to achieving precise and accurate H.0 measurements.

14.2.1.2 Mass-Sheet Degeneracy

The mass-sheet transformation (MST) of the convergence .κ given by

κ → κ ' = λκ + (1 − λ)
. (14.10)

leaves all the imaging observables invariant with a simultaneous rescaling of the
unknown source position

β → β ' = λβ,
. (14.11)

where .λ is a constant. This degeneracy in the convergence .κ and thus the potential
ψ is called the MSD.
.
The time delay .Δt transforms under the mass-sheet transformation as

Δt → Δt ' = λΔt.
. (14.12)

Thus, the time-delay distance. DΔt and Hubble constant. H0 inferred from the observed
time-delay .Δt will change as
'
. DΔt = DΔt /λ, (14.13)
'
. H0 = λH0 . (14.14)

To gain an understanding of the MST, it is useful to describe the “true” physical

convergence .κtrue in terms of two components as

κ
. true (θ ) = κint (θ) + κext , (14.15)

where .κint (θ ) is the convergence produced by the mass distribution of the galaxy or
group or cluster acting as the main deflector, while.κext is the convergence produced by
the mass distribution not physically associated with the main deflector, e.g., along the
line of sight (LOS). Since for a physical deflector.limθ →∞ κint (θ) = 0, we can see that
.lim θ →∞ κtrue (θ ) = κext . Due to the MSD, this external convergence term cannot be
constrained from the imaging of the lensing system. However, it can be estimated by
comparing the statistics of LOS mass distribution between cosmological simulations
and that observed using photometric and spectroscopic surveys (Sect. 14.3.3). If
the external convergence .κext is ignored during lens modeling, then the modeled
'
convergence.κmodel will be an MST of the true convergence.κtrue with.λ = 1/(1 − κext )
as
14 Strong Lensing and H0 255

κtrue (θ ) 1 κint (θ )
κ'
. model (θ ) = +1− = . (14.16)
1 − κext 1 − κext 1 − κext
' '
Here, the condition .limθ →∞ κmodel = 0 is satisfied, since .κmodel is only attributed to
the central galaxy’s or galaxies’ mass distribution that ought to vanish at infinity.
In practice, lens models are often described with simply parameterized profiles,
which implicitly break the MSD. Therefore, the best fit simply-parametrized model
'
.κmodel could be thought to be an approximate MST of .κmodel as

κ'
. model (θ ) = λint (θ )κmodel (θ) + 1 − λint (θ ). (14.17)

'
Here, the internal MST parameter.λint (θ ) cannot be a constant to satisfy.lim θ→∞ κmodel =
0. A .λint (θ ) can be designed, which acts as a constant sheet of mass near the central
region of the lensing system (.θ ≲ O(10θE )), but vanishes at.θ ≫ θE [7, 11], where.θE
is the Einstein radius. As a result, the true internal mass distribution can be expressed
as

κ (θ) = (1 − κext )[λint (θ)κmodel (θ ) + 1 − λint (θ )]

. int (14.18)

Thus, the true Fermat potential difference relates to the modeled Fermat potential
difference as

. Δτtrue = Δτmodel λint (1 − κext ). (14.19)

Here, we have expressed .λint as a constant since it is approximately a constant in the

region where lensed images appear.

Combining Eqs. 14.8, 14.9, and 14.19, the measured Hubble constant from
time-delay cosmography can be expressed as

Δτmodel (1 − κext )λint f k (0, z d , Θ) f k (0, z s , Θ)

. H0 = . (14.20)
Δt f k (z d , z s , Θ)

In this formulation, the time delay .Δt is directly measured from light curves of
the images (Sect. 14.3.1), .Δτmodel is obtained from lens modeling of high-resolution
imaging data of the lens system (Sect. 14.3.2), .κext is estimated from photometric and
spectroscopic surveys of the lens environment (Sect. 14.3.3), and .λint is constrained
from the stellar kinematics of the lens galaxy (Sect. 14.3.4).
This formulation also illustrates that . H0 measured by time-delay cosmography
weakly depends on the expansion history of the Universe from redshift .z s , through
redshift .z d , up to redshift .z = 0. Usually, a cosmological model is assumed to com-
pute the . f k functions, and . H0 can be slightly degenerate with other cosmological
parameters, such as the dark energy’s equation of state parameter .w [12, 120]. How-
256 T. Treu and A. J. Shajib

ever, it is also possible to infer . H0 by constraining the Universe’s expansion history

empirically, using relative distances of the Type Ia SNe. This inverse distance ladder
method [103] allows one to obtain H0 without assuming specific values for the other
cosmological parameters.

14.2.2 A Brief History

After Refsdal’s [76] suggestion, time-delay cosmography stayed dormant until the
first lensed quasars were discovered some 15 years later [117]. A period of excitement
followed in the 1980s when astronomers tried to measure time delays for the first
time [32]. However, determining time delays proved to be a significant challenge.
First, the stochastic nature of quasar light curves makes it more difficult than using
the well-behaved SN light curves, as originally suggested by Refsdal. Second, light
curves at optical wavelengths (corresponding usually to rest frame UV or blue) are
severely affected by microlensing, which further confounds the signal. Third, the
typical image separation of galaxy-scale lenses is similar to the image quality of
seeing-limited ground-based optical telescopes.
Overcoming these challenges required monitoring campaigns with significantly
higher precision per epoch, higher cadence, and longer duration than initially thought.
By the end of the 1990s and early 2000s, such campaigns started to produce reliable
time delays with few percent precision, both in the optical [57, 71, 83] and in the
radio [33].
Once the ability to obtain time delays was demonstrated, attention turned to
constraining the lensing potential. A two-pronged approach proved successful. On
the one hand, improvements in data quality and lens modeling techniques allowed
astronomers to capture the information content of multiply imaged extended sources,
e.g., quasar host galaxies and radio jets and emissions. This step increases by several
orders of magnitude the constraints on the mass distribution of the deflector with
respect to the positions of the quasars in the image plane used in previous work.
On the other hand, it became possible to measure the stellar kinematics of the
deflector, a dynamical mass tracer [112] that is crucially insensitive to the limiting
factors of lensing, chiefly the MSD. Stellar kinematics also provides an independent
handle on the angular diameter distance to the deflector, further enhancing the cosmo-
logical constraints [49, 50]. Conversely, lensing is insensitive to the mass-anisotropy
degeneracy affecting dynamical estimates of mass. As shown in Sect. 14.3.4, the
combination of lensing and stellar dynamics is crucial to the success of this method-
ology [87].
The final piece of the puzzle for precision time-delay cosmography at the galaxy
scale was accounting for the effects of the LOS and the local environment [98]. Under
or overdense LOSs, nearby perturbers, multiple-plane lensing, and group/cluster
scale halos can affect the inferred distances for galaxy-scale lenses as much as 5–
10% if not properly accounted for.
14 Strong Lensing and H0 257

Fittingly, 50 years after Refsdal’s paper, the first multiply imaged SN was dis-
covered in 2014 [54]. The SN, properly named “Refsdal”, was multiply imaged by
a cluster of galaxies, making the geometry and arrival time sequence significantly
more complex than those observed for quasars lensed by galaxy-scale potentials. The
discovered “Einstein-cross” configuration is primarily due to a cluster galaxy and has
little value for cosmography since the time delays are short [78]. Additional images,
however, were predicted upon discovery based on the cluster mass model [54, 69,
95]. One was in the past and thus sadly lost, but the other was predicted to re-appear
approximately a year later [113], and it indeed appeared when and where the model
predicted [53]. Measurements of H0 based on SN Refsdal have been reported [55,
56].
Additional multiply imaged SNe have been discovered since SN Refsdal [35],
and this is an area where much growth is expected in the coming decade.

14.3 Current State of the Art

We now review the current state of the art by means of two case studies. First,
in Sects. 14.3.1, 14.3.2, 14.3.3 and Sect. 14.3.4, we describe the galaxy-scale lens
RXJ1131–1231. In this system, the variable source is a quasar (Fig. 14.2), and
it is arguably the galaxy-scale system with the highest quality dataset to date.
Then, in Sect. 14.3.5, we examine the case of the SN Refsdal, lensed by cluster
MACS1149.5+2223. This is the first example of H0 measured via a multiply imaged
SN and presents a number of interesting differences with respect to the more estab-
lished method of using quasars lensed by galaxies.

14.3.1 Time Delay Measurements

Time delays (i.e., .Δt in Eq. 14.20) between lensed quasar images are measured
using the quasar’s intrinsic variability. The light curves of the individual images are
measured with 1-m or 2-m class telescopes. The COSMOGRAIL collaboration [22]
has monitored several lensed quasar systems, spanning up to more than a decade,
providing precise measurements to a few percent [23, 63, 104]. Figure 14.1 shows
light curves from COSMOGRAIL for RXJ1131–1231 based on 16 years (2003–
2019) of monitoring. Courbin et al. [24] have demonstrated that robust time delays
can be measured within a single season with almost a daily cadence and millimag
photometric precision to detect small amplitude variations of the order of 10–50
millimag. Given the numerous discoveries of new lensed quasar systems from recent,
ongoing, and future surveys, such fast turnaround for time delay measurements will
be essential to rapidly produce cosmological constraints.
The main challenge in measuring the time delay is the extrinsic variability caused
by the microlensing of foreground stars. This extrinsic variability pattern is unique
258 T. Treu and A. J. Shajib

Fig. 14.1 Light curve of RXJ1131–1231 in the . R-band from COSMOGRAIL. The bottom panels
show the difference curves between pairs of images after shifting the curves corresponding to
the measured delays. These difference curves illustrate the long-term extrinsic variability due to
microlensing. Figure from [63]

to each lensed image. The microlensing variability can be of two types. The first
one is a fast rise and fall, giving a sharp peak in the light curve. This fast variability
happens when the lines of formally infinite magnification (i.e., caustics [85]) from
a single foreground star cross over the quasar accretion disk. The second one is a
long-term variability owing to overlapping caustics resulting from crowded stars
creating a smooth change in the microlensing magnification as they move in front of
the background quasar, due to internal and peculiar transverse motions.
Techniques developed to extract the time delays from the light curves while
accounting for the microlensing variability include: cross-correlation that does not
require explicit modeling of the microlensing variability [74]; explicit modeling of
the intrinsic and microlensing variabilities with spline fitting or Gaussian processes
[46, 104]. The “Time Delay Challenge” (TDC) [29, 59] validated these techniques
with simulated data. The PyCS software used for the COSMOGRAIL data analysis
was among the techniques achieving the target precision and accuracy in the TDC.
14 Strong Lensing and H0 259

14.3.2 Lens Modeling

The main objective in lens modeling is to constrain the lensing potential that gives
rise to the observed lensed images and distorted arcs from the background quasar and
its host. This lensing potential provides.Δτmodel in Eq. 14.20. Usually, high-resolution
imaging from the Hubble Space Telescope (HST ) is used in lens modeling due to
the superb stability in the point spread function (PSF) and diffraction-limited nature
of the PSF to resolve lensed quasars that are separated by .∼ 1'' (e.g., Fig. 14.2).
However, imaging from large ground-based telescopes assisted with adaptive optics
(AO), such as the NIRC2 imager at the Keck Observatory, has also been successfully
modeled for cosmographic analysis after meticulous reconstruction of the PSF [19].
The lens model has four main components: (i) the lensing potential or the deflector
mass distribution, (ii) the flux distribution in the deflector galaxy, (ii) the flux distri-
bution in the quasar host galaxy, and (iv) the point spread function (PSF). Thus, the
model for the imaging data can be reconstructed from which the likelihood function
. p(d imaging | ξmass , ξlight , ξsource , P) can be computed, where . d imaging is the imaging

data, .ξmass is the mass model parameters, .ξlight is the deflector galaxy’s light model
parameters .ξsource is the quasar host galaxy’s light model parameters and .P is the
PSF model. The PSF.P can be initially estimated from a few stars within the imaging
data. However, due to color mismatch between the stellar type and the quasar, spatial
variations, and undersampling, this initial PSF model needs to be improved during
the lens modeling to fit the pixels around the lensed quasar images to the noise level
[4, 18]. The PSF also usually needs to be reconstructed at a higher pixel resolution
than the original image [91, 92]. The lens model parameters .ξmass , .ξlight , and .ξsource
are then constrained by sampling from the likelihood function.
The most common choices for the deflector galaxy’s mass or potential model are
“simply parametrized” models, where the mass distribution of the main deflector
galaxy and other nearby galaxies are described with functions depending on a few
free parameters. The nearby galaxies along the LOS are added to the parametric lens
model when their higher order lensing effect (i.e., flexion) is non-negligible [6, 62, 89,
119]. The lensing contribution from all the other LOS structures can be accounted for
by the independently estimated external convergence term (Sec. 14.3.3) and an exter-
nal shear component added to the lens model’s parametric description. The simplest
choice that yields good residuals for the simply parametrized model of galaxy-scale
lenses is the elliptical power-law mass distribution with the 3D radial density profile
−γ
.ρ(r ) ∝ r [79]. The dark matter and baryonic distributions in massive galaxies are
individually not power-laws. Yet, a simple power-law radial mass profile has been
sufficient to describe both lensing [1, 36] and non-lensing observables such as stel-
lar dynamics [17, 27] and X-ray intensities [47]. The total density profile is well
approximated by a power-law close to the isothermal .r − 2, a phenomenon known as
the “bulge–halo conspiracy” [30, 109]. Suyu et al. [98] allowed departure from the
power-law model using pixelated perturbations in the potential and found that large
deviations (.> 2%) from the power-law form were not required.
260 T. Treu and A. J. Shajib

An alternative choice of simply parametrized mass profile is a composite of the

baryonic mass distribution that follows the observed light distribution with a spatially
uniform mass-to-light ratio and the dark matter distribution described with a Navarro–
Frenk–White (NFW) profile [66, 68]. The H0 inferred using this composite mass
model is reassuringly consistent with that using the simpler power-law model [64].
The composite model with the dark matter distribution described by an NFW profile
has also been consistent for a sample of non-time-delay galaxy–galaxy lenses [30,
90]. These simply parametrized mass models are sufficient to fit the lens imaging
data to the noise level and are internally consistent within a few percent for each
individual system and within 1% for the combined H0 values from 7 systems [64,
120]. However, the true mass distribution can potentially differ from that inferred
from lensing data only, owing to the MSD [90]. A physical interpretation for such
possible deviation—small and not required by the data—is shown in Fig. 14.3.
To model the deflector light profile, one, two, or three Sérsic profiles [86] are
used [92, 98, 119]. The quasar host galaxy’s light profile can be described on a
pixellated grid with a regularization condition [99, 116]. For example, such a pixel-
lated scheme is adopted to reconstruct the source of RXJ1131–1231 illustrated in
Fig. 14.2. An alternative parametric approach is to adopt a basis set of a Sérsic pro-
file and shapelets (i.e., 2D Gauss–Hermite polynomials [75]), whose amplitudes are
determined through linear inversion of the observed image [5].
It is often necessary to make choices in the model settings, for example, the pixel
resolution of the reconstructed host galaxy. To avoid any systematic bias and account
for modeling uncertainty, different models are constructed by taking a combination of
the plausible choices, and then this source of error is marginalized over by combining
the posteriors from all these models, generally weighting by goodness of fit to avoid
overcounting relatively poor models. Figure 14.4 illustrates this marginalization over
multiple lens models with differing resolutions in the host galaxy reconstruction.

14.3.3 External Convergence from the LOS

The mass structures between the background source and the observer (i.e., galaxies,
groups, and clusters) contribute additional lensing effect that can modify the esti-
mated Fermat potential difference and thus the inferred H0 by a few percent. This
shift is estimated with the external convergence term .κext in Eq. 14.20. The contribut-
ing LOS structures can be grouped into two categories: (i) for lensing effects falling
outside the tidal regime1 , the non-linear lensing contribution needs to be taken into
account by directly including their mass distributions in the lens model (described in
Sect. 14.3.2), and (ii) for lensing effect falling within the tidal regime, the combined
lensing effect can be accounted for with the external convergence term .κext and an
external shear term that is already included in the lens modeling. The LOS struc-

1Tidal regime is when the perturber’s gravitational field is smaller than the gradient in the deflection
angle field created by the main deflector(s).
14 Strong Lensing and H0 261

Fig. 14.2 Illustration of lens model of RXJ1131–1231 from [98]. Top left: Observed HST
ACS/F814W image. Top middle: Reconstructed lensed arcs of the quasar host galaxy using the
best-fit lens model. Top right: Reconstructed quasar images and the lens galaxy light. Bottom
right: Total reconstructed image obtained by summing the top-middle and top-right panels. Bot-
tom middle: Normalized residuals for the reconstructed image. Bottom right: The reconstructed
quasar host galaxy morphology on the source plane

tures of the first category are typically within .∼10.'' from the central deflector. The
commonly used criterion to select the perturbers in this category is a “flexion shift”
threshold [62]. Estimating flexion shift requires photometric redshifts and mass mea-
surements of the LOS galaxies. Spectroscopic redshifts are used over photometric
redshifts whenever available. If the LOS galaxies form a group (or cluster), then the
group-scale (or cluster-scale) halo also needs to be explicitly modeled if it satisfies
the flexion-shift criterion [52, 65, 81]. The velocity dispersions of the LOS galaxies
are additionally used to infer group or cluster memberships of those galaxies [14,
96].
The external convergence for the second category of perturbers can be estimated
statistically from the number count of galaxies [34] around the lens system (usually
within .120'' ) or using weak lensing effect created by these LOS perturbers on the
shapes of background galaxies. The galaxy number counts, either directly or weighted
by quantities that correlate with lensing strength, such as projected distance from the
central deflector, the perturber’s mass, and redshift [40, 80], are used as summary
statistics for the lens environment. LOS cones with matching summary statistics
are then selected from cosmological simulations [44, 45]. The external convergence
values computed for these simulated cones then provide a probability distribution of
the external convergence of the real target (e.g., as obtained by [98] for RXJ1131–
262 T. Treu and A. J. Shajib

Fig. 14.3 Physical interpretation of residual uncertainty allowed by MSD on mass density profile.
[89] modeled a set of non-time-delay lens galaxies with exquisite HST images and unresolved stellar
velocity dispersion of the deflector fully accounting for the MSD and expressed the results as devia-
tions from standard “composite” (top row) and power-law mass profiles (bottom row). The standard
composite model comprising an NFW [67] dark matter halo and a stellar component with constant
mass-to-light ratio is consistent with the data, although a small amount of contraction/expansion
of the halo (top left panel) or a small gradient in the mass-to-light ratio (top right panel) cannot be
ruled out. Similarly, a power-law mass density profile is consistent with the data, although the data
cannot rule out small deviations (purple bands in the bottom panels). See [89] for more description.
When available, additional information—such as spatially resolved stellar kinematics—reduces the
residual freedom and thus tightens the bounds on H0 when applied to time-delay lenses

1231). The number counts for the observed lens environment are normalized with a
large number of control fields for which the photometric data is available at the same
quality, and the same is done for the simulated fields. Taking these relative number
counts as the summary statistics minimizes the impact of the chosen cosmological
parameters in the cosmological simulation.
From high-quality and wide-field imaging, the distortion in the shapes of back-
ground galaxies can be used to constrain the weak lensing shear created by the LOS
structures. In the linear regime, this shear uniquely maps to external convergence [51].
Tihhonova et al. [105, 106] applied this technique to provide alternative estimates of
the external convergence .κext for two systems, finding estimates in agreement with
those based on galaxy number counts.
14 Strong Lensing and H0 263

Fig. 14.4 Combined posterior of the lens model parameters power-law exponent.γ , Einstein radius
.θE , and external shear magnitude.γext . The illustrated model-predicted time-delay distance. DΔt
model is

obtained using a fiducial cosmology from the predicted Fermat potential difference.Δτmodel . There-
model is simply a reformulation of .Δτ
fore, the illustrated . DΔt model , and the measured time delays are
still necessary to measure H0 using the lens model posteriors. The combined posterior marginalizes
over the choice of pixel resolution in the host galaxy reconstruction, where the individual posterior
for each choice is illustrated with colored distributions. Figure from [101]
264 T. Treu and A. J. Shajib

14.3.4 Stellar Kinematics

Stellar kinematics traces the 3D potential of the deflector, whereas lensing traces the
2D projected potential. Thus, in combination, dynamics and lensing can break the
degeneracies inherent to each method: the MSD in lensing and the mass-anisotropy
degeneracy in dynamics. Therefore, stellar kinematics provide the .λint term in
Eq. 14.20.
Owing to the angular size and faintness of the deflectors, the most common type of
measurement for stellar kinematics is an unresolved (i.e., aperture-integrated) LOS
velocity dispersion from long-slit spectra. These spectra are usually taken using large
ground-based telescopes, for example, the Keck Observatory and the Very Large
Telescope (VLT), in seeing-limited cases in the optical.
The stellar velocity dispersion is usually modeled by solving the Jeans equation
[16, 48], which is derived from the collisionless Boltzmann equation [3]. The LOS
velocity dispersion can then be expressed, in the spherical case, as
ʃ ∞ ( r ) l(r )M(r )
2G
σ 2 (R) =
. los Kβ dr, (14.21)
I (R) R R r

where . I (R) is the surface brightness distribution of the kinematic tracer in the deflec-
tor, .l(r ) is the 3D luminosity density of the same tracer, . M(r ) is the 3D enclosed
mass, .Kβ is a function that depends on the anisotropy profile .β(r ) [61].
The anisotropy parameter .β(r ) is defined as

σt2 (r )
β(r ) ≡ 1 −
. , (14.22)
σr2 (r )

where .σt is the tangential component of the velocity dispersion and .σr is the radial
component. In Eq. 14.21, the terms related to the light distribution, that is, . I (R)
and .l(r ), are well-constrained from the observed light, albeit with assumptions to
deproject from the 2D . I (R) to the 3D .l(r ). However, there is a degeneracy between
the mass distribution giving . M(r ) and the anisotropy profile giving .Kβ , namely the
mass–anisotropy degeneracy [25]. Although combining constraints from lensing and
dynamics breaks the mass-sheet and mass–anisotropy degeneracies, the breaking
power from unresolved velocity dispersion is limited. In the past, the mass-sheet
degeneracy was broken chiefly through the mass profile assumption (i.e., the power-
law or composite form), and then the stellar kinematics provided tighter constraint
on the mass profile distribution with the mass-anisotropy degeneracy marginalized
over [101]. In such cases, the internal mass-sheet transformation parameter .λint (in
Eq. 14.20) was effectively set to .λint = 1. With these assumptions, for RXJ1131–
1231 H0 =.78.3+3.4
−3.3 km s
−1
Mpc−1 based on joint lens modeling constraints from
HST.+AO imaging [19, 97].
Spatially resolved velocity dispersion constrains anisotropy more tightly than the
unresolved case. Thus, spatially resolved velocity dispersion is much more effective
14 Strong Lensing and H0 265

Fig. 14.5 Keck/KCWI integral field spectra of RXJ1131–1231. Left: 2D representation of the
datacube obtained by summing across the wavelength dimension. The yellow contour shows the
region within which spectra were extracted to measure the resolved velocity dispersion of the
deflector. The grey box marks the pixel for which the observed spectra and model fitting are shown
in the right-hand panel. Right: The observed spectra (grey line) and the estimated deflector spectra
(orange line) after removing the modeled quasar contribution (blue line). The model for the observed
spectra using X-shooter Spectral Library (XSL) and having fitted the velocity dispersion is shown
with the red line. The spatially resolved velocity dispersion map is thus extracted in 41 bins within
the yellow contour (Fig. 14.6). Figure from [93]

in simultaneously breaking the above-mentioned degeneracies [87]. Furthermore,

an axisymmetric mass model can be constrained based on the spatial information
in the resolved kinematics, allowing one to go beyond simple spherical symmetry
in the mass models [15]. The oblateness or prolateness of the 3D mass shape can
potentially be constrained by the misalignment between the kinematic and photo-
metric major axes [58]. The only time-delay lens system with spatially resolved
kinematics published so far is RXJ1131–1231, based on data from the Keck Cosmic
Web Imager (KCWI) on the Keck Telescope (Figs. 14.5 and 14.6). Obtaining inte-
gral field spectra for kinematic measurement has been challenging from the ground
since the quasar contribution will contaminate the central deflector’s spectra in the
seeing-limited case without AO as the typical separation between the two is .∼ 1'' .
The kinematic measurement made by [93] accounted for quasar light contamina-
tion in extracting the velocity dispersion of the deflector by forward modeling it.
Combining this resolved kinematics with the HST imaging, measured time delays,
and estimated external convergence yields . H0 = 77.3+7.1
−7.3 km s
−1
Mpc−1 [93]. This
value, obtained by combining resolved kinematics and mass modeling with relaxed
assumptions, agrees very well with that obtained from simple parametric mass mod-
els, thus corroborating the standard assumptions made in time-delay cosmography
[93, 120] (Fig. 14.7).
266 T. Treu and A. J. Shajib

Fig. 14.6 Spatially resolved kinematic maps in 41 bins for RXJ1131–1231 from Keck/KCWI
integral field spectra. The top row corresponds to the LOS velocity dispersion, and the bottom row
corresponds to the LOS mean velocity. The left column shows the mean values, and the right column
shows the uncertainties. Figure from [93]

Fig. 14.7 Comparison of H0 values with various ways to break the MSD. For the seven systems
analyzed by the TDCOSMO collaboration, the MSD broken by simple parametric assumption on
+1.6
the mass profile (with .λint = 1 fixed) gives .74.2−1.6 km s−1 Mpc−1 (grey) [64], and the MSD
+5.8
broken by unresolved kinematics gives .73.3−5.8 km s Mpc−1 (emerald) [7]. For RXJ1131–1231,
−1

the MSD broken by simple parametric mass profile assumption gives .78.3+3.4 −3.3 km s
−1 Mpc−1
+7.3
(blue) [19], and the MSD broken by spatially resolved kinematics gives .77.1−7.1 km s−1 Mpc−1
(red) [93]. Using spatially resolved kinematics for one system gives similar uncertainty on H0 from
seven systems with unresolved kinematics, illustrating the superior power of resolved kinematics
in breaking the MSD. Figure from [93]
14 Strong Lensing and H0 267

14.3.5 SN Time Delay Cosmography: The “Refsdal” Case

Study

SN Refsdal (Fig. 14.8; [54]) is the first multiply imaged SN with measured time
delays [56, 78]. It is worth studying this case in some detail as it provides some
important lessons for SN time delays, which we expect to be a major contributor to
cosmography in the next decade.
Before going through all the steps leading to the recent measurements of H0 [55],
it is useful to summarize the main differences between this case and the one examined
in the previous sections, representing quasars lensed by galaxy-scale deflectors:
1. The intrinsic light curve of an SN (Fig. 14.8) is usually well described by a
template or a low-order polynomial. This regular nature simplifies the task of
obtaining a high-precision time delay with respect to the stochastic light curves of
lensed quasars (Fig. 14.1). Extrinsic effects from the foreground, e.g., microlens-
ing, can affect both cases.
2. The main deflector of SN Refsdal is a cluster of galaxies. Clusters of galaxies
are not dynamically relaxed, in contrast to the inner regions of massive ellipti-
cal galaxies—the typical deflectors of galaxy-scale lenses. Thus, cluster lenses
are generally significantly more complex to model, both from a lensing and
dynamical point of view, with respect to galaxy-scale ones.
3. The caustics of clusters of galaxies cover a much larger solid angle on the sky
than those of galaxies. Therefore, tens or even hundreds [2] of multiply imaged
sources can be used to constrain the lens model in a cluster. There is generally
one multiply imaged source for galaxy-scale lenses, although systems with up to
a handful of them have been found [37, 88]. Note that having several families of

Fig. 14.8 Left: HST images of SN Refsdal, summarizing the time evolution of the phenomenon,
including the original discovery of the Einstein Cross in 2014 and the appearance of SX in
2015/2016. Right: light curve of SN Refsdal compared to SN1987A-like light curves and polyno-
mial fits. Figures from [56]
268 T. Treu and A. J. Shajib

Fig. 14.9 Left: unblinded posterior distribution function of . H0 based on SN Refsdal. The lens
models have been weighted according to their agreement with the . H0 -independent observables,
such as image positions and magnification ratios. Right: error budget on . H0 measurement. Figures
from [55]

multiply imaged sources at different redshifts helps mitigate the MSD [13, 41,
42] (Fig. 14.9).
SN Refsdal was first discovered [54] as an Einstein cross in images taken as part
of the Grism Lens-Amplified Survey from Space (GLASS) program [111]. Although
the time delays between the cross images were expected to be of order days/week
(and thus not very useful for cosmography) because of the symmetry and separation
of the configuration, it was immediately recognized [54, 69, 95] that one additional
and more distant image (SX) would re-appear some time in the near future, with
great potential for cosmography.
Considerable effort went into predicting SX’s timing, position, and brightness
using updated lens models based on extensive spectroscopy [43, 113] before it
appeared in the sky. Indeed, SX appeared as predicted [54], helping to build confi-
dence in the models. The good agreement was not a foregone conclusion, considering
the complexity of the cluster lens models.
The first estimate of the magnification ratio and of the long time delay (SX to
the cross; the one with the most cosmographic constraining power) was used by one
team to produce an estimate of H0 [115]. That study highlighted the spread between
the different cluster lens models. The spread between all the possible models is not
surprising, considering that many of the lensing codes employed were not designed
for precision cosmology and, therefore, did not have the necessary numerical preci-
sion and resolution. Pixelated models and models based on heavily regularized basis
sets do not have by design the angular resolution needed for time-delay cosmogra-
phy. The resolution requirement can be understood in terms of astrometric precision.
Measuring the Hubble constant to a few percent precision usually requires knowing
how the Fermat potential (and image positions) vary over tens of milliarcseconds
[4]. Furthermore, many of the early models did not make full use of all the available
information, e.g., cluster membership, stellar velocity dispersions of cluster mem-
14 Strong Lensing and H0 269

bers, and spectroscopic redshift of multiple images. A thorough discussion of the

sources of uncertainty is given by [42].
After the re-appearance, a major effort was undertaken to obtain a blind, high-
precision (1.5%) measurement of the time delay [56] and fold it in with existing
lens models to obtain a blind measurement of H0 [55]. It is crucial to stress that all
the analysis choices were made blindly with respect to H0 (the time delay was kept
blind throughout the analysis, for example) and that the analysis was not modified
after unblinding (with the exception of the correction of a mistake in sign convention
on magnifications). This is a crucial step to prevent “experimenter bias,” and we
advocate that every cosmological measurement should be carried out blindly.
As highlighted by [115], the spread between lens models is clearly significant. It
should be noted, however, that they include in their analysis models that were later
discovered to be affected by substantial numerical errors [55], and therefore their
spread was overestimated. Two options were considered by [55] to account for the
spread between models. First, it was decided to consider only the two models that are
based on codes designed to do high-precision cosmography, glee [100] and glafic
[70], weighted by their agreement with observables independent of H0 , yielding
+4.1 −1
. H0 = 66.6−3.3 km s Mpc−1 . If one had preferred to do the straight average of the
two models selected prior to unblinding to be the most accurate, the result would
have changed negligibly: combining the glee and glafic models with equal weight
+4.1
yields . H0 = 67.2−3.7 km s−1 Mpc−1 , consistent within the errors. Alternatively,
an analysis considering all models was run, again weighted by the same scheme,
finding . H0 = 64.8+4.3
−4.3 km s
−1
Mpc−1 . The results are very similar between the two
schemes because the glee and glafic models provide, by far, the best match to the
observables. The good agreement between the two options is partly due to the fact
that the glee and glafic models are by far the ones that best match the observables.
This is not surprising, considering that some of the other models were not designed
for precision cosmography, as discussed by [108] and [56]. Figure 14.2 in [56] shows
that the models with the least weights (Zitrin and Diego) have very broad tails to
significantly lower H0 , while Jauzac and Sharon yield results comparable to Grillo
and Oguri.
The measurement from SN Refsdal [56] is not precise enough to help solve the
“Hubble tension.” Given the uncertainties [42, 56], it is consistent with both Planck
and the local distance ladder method [77]. The latter is only 1.5.σ away from the best
fit, thus consistent with the SN Refsdal analysis.
The important lessons from SN Refsdal are two. First, the blind efforts succeeded
to the level that only the most optimistic practitioners of cluster lens modeling would
have expected. SX appeared exactly where and when it was predicted to be, and
the inferred value of H0 is perfectly consistent with those of other methods. It
did not have to be this way, considering the complexity of the mass distribution
in MACS1149.5+2223. SX could have appeared elsewhere or at a different time,
and H0 could have turned out to be 30 or 100 km s−1 Mpc−1 . This blind prediction
helps build confidence in the models. Although it is clearly impossible to prove that
a better model does not exist, the best models appear to be sufficiently precise and
accurate, to the extent that we can probe them empirically. Second, 6% is a very
270 T. Treu and A. J. Shajib

small uncertainty for an absolute distance measurement from a single system, and
compelling arguments show that it is not significantly underestimated [42, 55] when
the data quality is as high as in this case. We stress that high-quality data is vital to
obtain this kind of precision. The model from [42] is constrained by 89 spectroscop-
ically confirmed multiple images arising from 28 distinct sources. Most clusters to
date do not have the same level of empirical
√ constraints [60], although the situation
is rapidly improving [2]. A simple . N scaling suggests that ten systems similar
to MACS1149.5+2223 will suffice to reach .∼ 2% precision and thus contribute to
solving the Hubble tension if no yet-to-be-discovered systematic floor arises.

14.4 Conclusions and Future Outlook

The two recent case studies we chose to highlight represent genuine breakthroughs.
We will now discuss them in the context of other measurements to give a sense of
the landscape (a selection of measurements is summarized in Fig. 14.10).
RXJ1131–1231 is the first galaxy-scale lens with time delay and spatially resolved
kinematics. The exquisite data for this system enabled [93] to reach 9% precision
from a single lens, using what we call “conservative” assumptions about the mass
distribution, i.e., allowing the MSD to be only constrained by kinematic data, and
thus allowing maximum uncertainty on H0 . This is remarkable, as the precision is
comparable to what was obtained by [7] with seven lenses for which only unresolved
kinematics were available. With the higher angular resolution attainable with the
JWST or with future instruments with AO, we can expect the precision per system to
reach .∼ 4% [121], better than that was achieved in the past for individual systems by
using the more “assertive” approach of breaking the MSD by imposing a functional
form for the mass density profile [89]. With this kind of precision, reaching the
2% precision, which was previously achieved under “assertive” assumptions by [64,
120], seems within reach under “conservative” assumptions with the systems known
today [9].“Free form” models, that is, the ones in which the surface mass density or
potential of the main deflector is rendered as pixels (see [20, 26, 72, 82]), should be
able to obtain similar precision if they can be constrained by the full, high-information
content from data with state-of-the-art quality. This has not been achieved yet but
should be attainable. Along the way, work will need to continue in order to uncover
and mitigate new sources of systematic errors, including those arising from selection
effects [21].
The success of MACS1149.5+2223 paves the way for lensed SNe to become a
major contributor to time-delay cosmography. A single system with excellent data
quality, modeled in an “assertive” way, obtains a 6% precision on H0 , comparable
to the average time-delay quasars of those analyzed by [120]. With major synoptic
surveys such as Euclid and the Vera C. Rubin Observatory’s Legacy Survey of Space
and Time (LSST) about to begin, we are confident that many such systems will be
discovered in the coming decade [70]. Hopefully, samples of lensed SN will soon
reach comparable precision (or better!) to lensed quasars, thus providing a vital sanity
14 Strong Lensing and H0 271

Fig. 14.10 Comparison of H0 measurements based on time-delay cosmography, in.ɅCDM cosmol-

ogy. The measurements are grouped by: (i) the lensing configuration (galaxy+QSO vs. cluster+SN).
(ii) Assumptions on the mass distribution of the main deflector, “assertive” and “conservative” for
simply parametrized models or “free form” for pixellated models. (iii) Amount of information used
per lens; in the case of a galaxy-scale lens, “low info” utilizes quasar positions and time delays,
“high info” adds the extended surface brightness distribution of the multiple images of the quasar
host galaxy, stellar kinematics of the main deflector, and number counts or weak lensing to estimate
the LOS convergence. For the cluster+SN case, we define it as high info due to the large number
of spectroscopically confirmed multiply-images and cluster members, and we define it as assertive
because the glee and glafic models used for this measurement are based on simply parametrized
forms. We give the reference and the number of time delay lenses for each measurement. The
measurements shown in red have been blinded to prevent experimenter bias. The figure is updated
from [110]

check and increased overall precision with respect to quasar-only forecasts such as
those presented by [110], thus getting us closer to solving the “Hubble Tension.”

Acknowledgements TT and AJS thank their many colleagues and friends working in the field of
time delay cosmography, without whom the progress described here would have been impossible.
We thank Fred Courbin, Pat Kelly, Veronica Motta, and Paul Schechter for their constructive com-
ments on an early draft of this manuscript. TT gratefully acknowledges support by the National
Science Foundation, the National Aeronautics and Space Administration, the Packard Foundation,
and the Moore Foundation. Support for this work was provided by NASA through the NASA Hub-
ble Fellowship grant HST-HF2-51492 awarded to AJS by the Space Telescope Science Institute,
which is operated by the Association of Universities for Research in Astronomy, Inc., for NASA,
under contract NAS5-26555.
272 T. Treu and A. J. Shajib

References

1. M.W. Auger et al., Astrophys. J. 724, 511–525 (2010). https://doi.org/10.1088/0004-637X/

724/1/511
2. P. Bergamini et al., Astrophys. J., in press (2023). arXiv:2303.10210. [astro-ph.GA]
3. J. Binney et al. (1987). arXiv: 1987gady . Accessed 7 Nov 2022
4. S. Birrer et al., Mon. Not. R. Astron. Soc. 489(2), 2097–2103 (2019). https://doi.org/10.1093/
mnras/stz2254. https://ui.adsabs.harvard.edu/abs/2019MNRAS.tmp.2172B
5. S. Birrer et al., Astrophys. J. 813, 102 (2015). https://doi.org/10.1088/0004-637X/813/2/102
6. S. Birrer et al., Mon. Not. R. Astron. Soc. 484(4), 4726–4753 (2019). https://doi.org/10.1093/
mnras/stz200. http://adsabs.harvard.edu/abs/2019MNRAS.484.4726B
7. S. Birrer et al., Astron. Astrophys. 643, A165 (2020). ISSN: 0004-6361, 1432-
0746. https://doi.org/10.1051/0004-6361/202038861, https://www.aanda.org/articles/aa/
abs/2020/11/aa38861-20/aa38861-20.html. Accessed 1 Apr 2021
8. S. Birrer et al. (2022). arXiv e-prints, arXiv:2210.10833 [astro-ph.CO]
9. S. Birrer, ApJ 919(1), 38 (2021). ISSN: 0004-637X. https://doi.org/10.3847/1538-4357/
ac1108. Accessed 21 Feb 2022
10. R. Blandford et al., Astrophys. J. 310, 568 (1986). https://doi.org/10.1086/164709
11. K. Blum et al. (2020). arXiv e-prints, arXiv:2001.07182
12. V. Bonvin et al., Astron. Astrophys. 616, A183 (2018). https://doi.org/10.1051/0004-6361/
201833287
13. M. Bradač et al., Astron. Astrophys. 424, 13–22 (2004). https://doi.org/10.1051/0004-6361:
20035744. arXiv:astro-ph/0405357 [astro-ph]
14. E.J. Buckley-Geer et al., Mon. Not. R. Astron. Soc. 498, 3241–3274 (2020). ISSN: 0035-8711.
https://doi.org/10.1093/mnras/staa2563. arXiv:2020MNRAS.498.3241B. https://ui.adsabs.
harvard.edu/abs/2020MNRAS.498.3241B. Accessed 21 Feb 2022
15. M. Cappellari, Mon. Not. R. Astron. Soc. 390, 71–86 (2008). https://doi.org/10.1111/j.1365-
2966.2008.13754.x
16. M. Cappellari, Mon. Not. R. Astron. Soc. 494(4), 4819–4837 (2020). ISSN: 0035-8711.
https://doi.org/10.1093/mnras/staa959. Accessed 3 June 2022
17. M. Cappellari, Ann. Rev. Astron. Astrophys. 54, 597–665 (2016). https://doi.org/10.1146/
annurev-astro-082214-122432
18. G.C.F. Chen et al., Mon. Not. R. Astron. Soc. 462(4), 3457–3475 (2016). https://doi.org/10.
1093/mnras/stw991
19. G.C. Chen et al., Mon. Not. R. Astron. Soc. 490, 1743–1773 (2019). ISSN: 0035-8711. https://
doi.org/10.1093/mnras/stz2547. arXiv: 2019MNRAS.490.1743C. https://ui.adsabs.harvard.
edu/abs/2019MNRAS.490.1743C. Accessed 21 Feb 2022
20. J. Coles, Astrophys. J. 679(1), 17–24 (2008). https://doi.org/10.1086/587635.
(arXiv:0802.3219 [astro-ph])
21. T.E. Collett et al., Mon. Not. R. Astron. Soc. 462(3), 3255–3264 (2016). https://doi.org/10.
1093/mnras/stw1856
22. F. Courbin et al., 225, 297–303 (2005). https://doi.org/10.1017/S1743921305002097. arXiv:
2005IAUS.225.297C. https://ui.adsabs.harvard.edu/abs/2005IAUS.225.297C. Accessed 20
Dec 2022
23. F. Courbin et al., Astron. Astrophys. 536, A53 (2011). https://doi.org/10.1051/0004-6361/
201015709. (arXiv:1009.1473 [astro-ph.CO])
24. F. Courbin et al., Astron. Astrophys. 609, A71 (2018). https://doi.org/10.1051/0004-6361/
201731461. https://adsabs.harvard.edu/abs/2018A%26A...609A..71C
25. S. Courteau et al., Rev. Mod. Phys. 86, 47–119 (2014). https://doi.org/10.1103/RevModPhys.
86.47
26. P. Denzel et al., Mon. Not. R. Astron. Soc. 501(1), 784–801 (2021). https://doi.org/10.1093/
mnras/staa3603. (arXiv:2007.14398 [astro-ph.CO])
27. C. Derkenne et al., Mon. Not. R. Astron. Soc. 506(3), 3691–3716 (2021). https://doi.org/10.
1093/mnras/stab1996. (arXiv:2107.05206 [astro-ph.GA])
14 Strong Lensing and H0 273

28. X. Ding et al., Mon. Not. R. Astron. Soc. 503, 1096–1123 (2021). ISSN: 0035-8711. https://
doi.org/10.1093/mnras/stab484. arXiv: 2021MNRAS.503.1096D. https://ui.adsabs.harvard.
edu/abs/2021MNRAS.503.1096D. Accessed 21 Feb 2022
29. G. Dobler et al., Astrophys. J. 799(2), 168 (2015). https://doi.org/10.1088/0004-637X/799/
2/168
30. A.A. Dutton et al., Mon. Not. R. Astron. Soc. 438, 3594–3602 (2014). https://doi.org/10.
1093/mnras/stt2489
31. G. Efstathiou, Mon. Not. R. Astron. Soc. 505(3), 3866–3872 (2021). https://doi.org/10.1093/
mnras/stab1588. (arXiv:2103.08723 [astro-ph.CO])
32. E.E. Falco et al., Astrophys. J. Lett. 289, L1–L4 (1985). https://doi.org/10.1086/184422
33. C.D. Fassnacht et al., Astrophys. J. 581, 823–835 (2002). https://doi.org/10.1086/344368
34. C.D. Fassnacht et al., Mon. Not. R. Astron. Soc. 410(4), 2167–2179 (2011). https://doi.org/
10.1111/j.1365-2966.2010.17591.x. arXiv: 0909.4301 [astro-ph.CO]
35. B. Frye et al., Transient Name Server AstroNote 96, 1 (2023)
36. R. Gavazzi et al., Astrophys. J. 667, 176–190 (2007). https://doi.org/10.1086/519237
37. R. Gavazzi et al., Astrophys. J. 677, 1046–1059 (2008). https://doi.org/10.1086/529541
38. D. Gilman et al., Astron. Astrophys. 642, A194 (2020). https://doi.org/10.1051/0004-6361/
202038829
39. M.R. Gomer et al., A&A 667, A86 (2022). ISSN: 004–6361, 1432–0746 (2022). https://doi.
org/10.1051/0004-6361/202244324. Accessed 28 Dec 2022
40. Z.S. Greene et al., Astrophys. J. 768, 39 (2013). https://doi.org/10.1088/0004-637X/768/1/
39
41. C. Grillo et al., Astrophys. J. 860(2), 94 (2018). https://doi.org/10.3847/1538-4357/aac2c9.
(arXiv:1800.1584 [astro-ph.CO])
42. C. Grillo et al., Astrophys. J. 898(1), 87 (2020). https://doi.org/10.3847/1538-4357/ab9a4c.
(arXiv:2001.02232 [astro-ph.CO])
43. C. Grillo et al., Astrophys. J. 822(2), 78 (2016). https://doi.org/10.3847/0004-637X/822/2/
78. (arXiv:1511.04093 [astro-ph.GA])
44. S. Hilbert et al., Astron. Astrophys. 499, 31–43 (2009). https://doi.org/10.1051/0004-6361/
200811054
45. S Hilbert et al., Mon. Not. R. Astron. Soc. 382(1), 121–132 (2007). https://doi.org/10.1111/
j.1365-2966.2007.12391.x. arXiv:astro-ph/0703803 [astro-ph]
46. A. Hojjati et al., Phys. Rev. D 87(12), 123512 (2013). https://doi.org/10.1103/PhysRevD.87.
123512. (arXiv:1304.0309 [astro-ph.CO])
47. P.J. Humphrey et al., Mon. Not. R. Astron. Soc. 403(4), 2143–2151 (2010). https://doi.org/
10.1111/j.1365-2966.2010.16257.x
48. J.H. Jeans, Mon. Not. R. Astron. Soc. 82, 122–132 (1922). ISSN: 0035-8711. https://doi.org/
10.1093/mnras/82.3.122. arXiv:1922MNRAS.82.122J. https://ui.adsabs.harvard.edu/abs/
1922MNRAS.82.122J. Accessed 07 Nov 2022
49. I. Jee et al., J. Cosmol. Astroparticle Phsy. 11, 033 (2015). https://doi.org/10.1088/1475-
7516/2015/11/033
50. I. Jee et al., J. Cosmol. Astroparticle Phys. 4, 031 (2016). https://doi.org/10.1088/1475-7516/
2016/04/031
51. N. Kaiser et al., Astrophys. J. 404, 441 (1993). https://doi.org/10.1086/172297
52. C.R. Keeton et al., Astrophys. J. 612(2), 660–678 (2004). https://doi.org/10.1086/422745.
arXiv:astro-ph/0406060 [astro-ph]
53. P.L. Kelly et al., Astrophys. J. Lett. 819(1), L8 (2016). https://doi.org/10.3847/2041-8205/
819/1/L8. (arXiv:1512.04654 [astro-ph.CO])
54. P.L. Kelly et al., Science 347, 1123–1126 (2015). https://doi.org/10.1126/science.aaa3350
55. P.L. Kelly et al., Science, in press (2023a). arXiv:2305.06367 [astro-ph.CO]
56. P.L. Kelly et al., Astrophys. J. 948(2), 93 (2023). https://doi.org/10.3847/1538-4357/ac4ccb.
(arXiv:2305.06377 [astro-ph.CO])
57. T. Kundic et al., Astrophys. J. 482, 75 (1997). https://doi.org/10.1086/304147.
(arXiv:astro-ph/9610162)
274 T. Treu and A. J. Shajib

58. H. Li et al., Astrophys. J. 863, L19 (2015). ISSN: 0004-637X. https://doi.org/10.3847/

2041-8213/aad54b. arXiv:2018ApJ.863L.19L. https://ui.adsabs.harvard.edu/abs/2018ApJ.
863L.19L. Accessed 17 Nov 2022
59. K. Liao et al., Astrophys. J. 800(1), 11 (2015). https://doi.org/10.1088/0004-637X/800/1/11
60. Y. Liu et al. (2023). arXiv:2307.14833 [astro-ph.CO]
61. G.A. Mamon et al., Mon. Not. R. Astron. Soc. 363, 705–722 (2005). https://doi.org/10.1111/
j.1365-2966.2005.09400.x
62. C. McCully et al., Astrophys. J. 836, 141 (2017). https://doi.org/10.3847/1538-4357/836/1/
141
63. M. Millon et al., Astron. Astrophys. 640, A105 (2020). https://doi.org/10.1051/0004-6361/
202037740. (arXiv:2002.05736 [astro-ph.CO])
64. M. Millon et al., Astron. Astrophys. 639, A101 (2020). https://doi.org/10.1051/0004-6361/
201937351
65. I.G. Momcheva et al., Astrophys. J. Suppl. 219(2), 29 (2015). https://doi.org/10.1088/0067-
0049/219/2/29. (arXiv:1503.02074 [astro-ph.CO])
66. J.F. Navarro et al., Astrophys. J. 490, 493–508 (1997). https://doi.org/10.1086/304888
67. J.F. Navarro et al., Astrophys. J. 490, 493 (1997). https://doi.org/10.1086/304888.
(arXiv:astro-ph/9611107)
68. J.F. Navarro et al., Astrophys. J. 462, 563 (1996). https://doi.org/10.1086/177173
69. M. Oguri, Mon. Not. R. Astron. Soc. 449, L86–L89 (2015). https://doi.org/10.1093/mnrasl/
slv025. (arXiv:1411.6443 [astro-ph.CO])
70. M. Oguri et al., Mon. Not. R. Astron. Soc. 405, 2579–2593 (2010). https://doi.org/10.1111/
j.1365-2966.2010.16639.x
71. A. Oscoz et al., Astrophys. J. Lett. 479(2), L89–L92 (1997). https://doi.org/10.1086/310599
72. D. Paraficz et al., Astrophys. J. 712(2), 1378–1384 (2010). https://doi.org/10.1088/0004-
637X/712/2/1378. (arXiv:1002.2570 [astro-ph.CO])
73. P.J.E. Peebles (1993). https://doi.org/10.1515/9780691206721. arXiv:1993ppc.book.P.
https://ui.adsabs.harvard.edu/abs/1993ppc.book.P. Accessed 28 Apr 2023
74. J. Pelt et al., Astron. Astrophys. 305, 97 (1996). https://doi.org/10.48550/arXiv.astro-ph/
9501036. arXiv:astro-ph/9501036 [astro-ph]
75. A. Refregier, Mon. Not. R. Astron. Soc. 338, 35–47 (2003). https://doi.org/10.1046/j.1365-
8711.2003.05901.x
76. S. Refsdal, Mon. Not. R. Astron. Soc. 128, 307 (1964). https://doi.org/10.1093/mnras/128.4.
307
77. A.G. Riess et al., Astrophys. J. Lett. 934(1), L7 (2022). https://doi.org/10.3847/2041-8213/
ac5c5b. (arXiv:2112.04510 [astro-ph.CO])
78. S.A. Rodney et al., Astrophys. J. 820(1), 50 (2016). https://doi.org/10.3847/0004-637X/820/
1/50. (arXiv:1512.05734 [astro-ph.CO])
79. D. Rusin et al., Astrophys. J. 587, 143–159 (2003). https://doi.org/10.1086/346206.eprint:
astro-ph/0211229
80. C.E. Rusu et al., Mon. Not. R. Astron. Soc. 467, 4220–4242 (2017). https://doi.org/10.1093/
mnras/stx285
81. C.E. Rusu et al., Mon. Not. R. Astron. Soc. 498, 1440–1468 (2020). https://doi.org/10.1093/
mnras/stz3451
82. P. Saha et al., Astrophys. J. Lett. 650(1), L17–L20 (2006). https://doi.org/10.1086/507583.
arXiv:astro-ph/0607240 [astro-ph]
83. P.L. Schechter et al., Astrophys. J. Lett. 475, L85+ (1997). https://doi.org/10.1086/310478.
arXiv:astro-ph/9611051
84. P. Schneider, Astron. Astrophys. 143, 413–420 (1985). ISSN: 0004-6361. arXiv:1985A and
arXiv:A.143.413S. Accessed 27 Apr 2023
85. P. Schneider et al. (1992). https://doi.org/10.1007/978-3-662-03758-4
86. J.L. Sérsic (1968), http://adsabs.harvard.edu/abs/1968adga.book.S
87. A.J. Shajib et al., Mon. Not. R. Astron. Soc. 473, 210–226 (2018). https://doi.org/10.1093/
mnras/stx2302
14 Strong Lensing and H0 275

88. A.J. Shajib et al., Mon. Not. R. Astron. Soc. 483, 5649–5671 (2019). https://doi.org/10.1093/
mnras/sty3397. http://adsabs.harvard.edu/abs/2019MNRAS.483.5649S
89. A.J. Shajib et al., Mon. Not. R. Astron. Soc. 494(4), 6072–6102 (2020). https://doi.org/10.
1093/mnras/staa828
90. A.J. Shajib et al., MNRAS 503(2), 2380–2405 (2021). ISSN: 0035-8711. https://doi.org/10.
1093/mnras/stab536. Accessed 31 Mar 2021
91. A.J. Shajib et al., Astron. Astrophys. 667, A123 (2022a). ISSN: 0004-6361. https://doi.org/10.
1051/0004-6361/202243401. https://ui.adsabs.harvard.edu/abs/2022A-abstract. Accessed
01 Feb 2023
92. A.J. Shajib et al. (2022b), arXiv: 2022arXiv220305170S. Accessed 08 Aug 2022
93. A.J. Shajib et al., Astron. Astrophys. 673, A9 (2023). https://doi.org/10.1051/0004-6361/
202345878. (arXiv:2301.02656 [astro-ph.CO])
94. I.I. Shapiro, Phys. Rev. Lett. 13(26), 789–791 (1964). https://doi.org/10.1103/PhysRevLett.
13.789
95. K. Sharon et al., Astrophys. J. Lett. 800(2), L26 (2015). https://doi.org/10.1088/2041-8205/
800/2/L26. (arXiv:1411.6933 [astro-ph.CO])
96. D. Sluse et al., Mon. Not. R. Astron. Soc. (2019). https://doi.org/10.1093/mnras/stz2483.
https://ui.adsabs.harvard.edu/abs/2019MNRAS.490..613S/abstract
97. S.H. Suyu et al., Astrophys. J. Lett. 788, L35 (2014). https://doi.org/10.1088/2041-8205/788/
2/L35
98. S.H. Suyu et al., Astrophys. J. 711, 201–221 (2010). https://doi.org/10.1088/0004-637X/711/
1/201
99. S.H. Suyu et al., Astron. Astrophys. 524, A94 (2010). https://doi.org/10.1051/0004-6361/
201015481
100. S.H. Suyu et al., Astron. Astrophys. 524, A94 (2010). https://doi.org/10.1051/0004-6361/
201015481. (arXiv:1007.4815 [astro-ph.CO])
101. S.H. Suyu et al., Astrophys. J. 766, 70 (2013). https://doi.org/10.1088/0004-637X/766/2/70
102. S.H. Suyu et al., Space Sci. Rev. 214(5), 91 (2018). https://doi.org/10.1007/s11214-018-
0524-3. (arXiv:1801.07262 [astro-ph.CO])
103. S. Taubenberger et al., Astron. Astrophys. 628, L7 (2019). ISSN: 0004-6361, 1432-0746.
https://doi.org/10.1051/0004-6361/201935980. Accessed 5 Jan 2023
104. M. Tewes et al., Astron. Astrophys. 556, A22 (2013). https://doi.org/10.1051/0004-6361/
201220352
105. O. Tihhonova et al., Mon. Not. R. Astron. Soc. 498(1), 1406–1419. https://doi.org/10.1093/
mnras/staa1436. arXiv:2005.12295 [astro-ph.GA]
106. O. Tihhonova et al., Mon. Not. R. Astron. Soc. 477(4), 5657–5669 (2018). https://doi.org/10.
1093/mnras/sty1040
107. T. Treu et al., Astrophys. J. 817, 60 (2016). https://doi.org/10.3847/0004-637X/817/1/60
108. T. Treu et al., Astrophys. J. 817(1), 60 (2016). https://doi.org/10.3847/0004-637X/817/1/60.
(arXiv:1510.05750 [astro-ph.CO])
109. T. Treu et al., Astrophys. J. 611, 739–760 (2004). https://doi.org/10.1086/422245.
eprint:astro-ph/0401373
110. T. Treu et al. (2022). arXiv: 2022arXiv221015794T. Accessed 3 Nov 2022
111. T. Treu et al., Astrophys. J. 812(2), 114 (2015). https://doi.org/10.1088/0004-637X/812/2/
114. (arxiv:1509.00475 [astro-ph.GA])
112. T. Treu et al., Mon. Not. R. Astron. Soc. 337, L6–L10 (2002). https://doi.org/10.1046/j.1365-
8711.2002.06107.x. eprint: astro-ph/0210002
113. T. Treu et al., Astron. Astrophys. Rev. 24(1), 11 (2016). https://doi.org/10.1007/s00159-016-
0096-8. https://adsabs.harvard.edu/abs/2016A%26ARv..24...11T
114. L. Van de Vyvere et al., Astron. Astrophys. 659, A127 (2022). ISSN: 0004-6361. https://
doi.org/10.1051/0004-6361/202141551. https://ui.adsabs.harvard.edu/abs/2022A&A.659A.
127V/abstract. Accessed 22 Aug 2022
115. J. Vega-Ferrero et al., Astrophys. J. Lett. 853(2), L31 (2018). https://doi.org/10.3847/2041-
8213/aaa95f. arXiv preprint: arxiv:1712.05800 [astro-ph.CO]
276 T. Treu and A. J. Shajib

116. S. Vegetti et al., Mon. Not. R. Astron. Soc. 392, 945–963 (2009). https://doi.org/10.1111/j.
1365-2966.2008.14005.x
117. D. Walsh et al., Nature 279, 381–384 (1979). https://doi.org/10.1038/279381a0
118. S. Weinberg (1972). arXiv: 1972gcpa.book.W. Accessed 28 Apr 202
119. K.C. Wong et al., Mon. Not. R. Astron. Soc. 465, 4895–4913 (2017). https://doi.org/10.1093/
mnras/stw3077
120. K.C. Wong et al., Mon. Not. R. Astron. Soc. 498, 1420–1439 (2020). https://doi.org/10.1093/
mnras/stz3094
121. A. Yıldırım et al. (2021). arXiv:2109.14615 [astro-ph]. Accessed 21 Feb 2022
Chapter 15
Addressing the Hubble Tension with
Cosmic Chronometers

Michele Moresco

15.1 Introduction

The discovery of the accelerated expansion of the Universe [64, 67] has been a major
revolution in modern cosmology, introducing a shift in the paradigm of the Universe
considered until the end of the XX century. This discovery required scientists to
introduce the presence of an unknown form of energy driving this acceleration (the
“dark energy”), or to postulate that the laws of General Relativity as known would
break at some large cosmological scales.
In this context, finding diverse independent ways to measure the expansion history
of the Universe became one of the main goals in astrophysics, since it could provide
fundamental inside into understanding the nature of dark energy, or if modified
gravity models are a better explanation. The quantity that measures how the Universe
is expanding as a function of time (or equivalently redshift) is the so-called Hubble
parameter, . H (z). It is typically useful to link this quantity to the scale factor .a(t)
(providing the relation between physical and comoving distances .rphys = a(t)rcom )
with the equation:
ȧ
. H (z) = . (15.1)
a
Several cosmological probes have been introduced during these years to address
these open questions, measuring how the Universe is expanding and modeling it
through cosmological parameters to characterize the behavior of dark energy and
dark matter. Due to methodological, technical, and theoretical advances, several
probes have reached nowadays a maturity that allowed them to be considered “stan-
dard” probes, among which e.g. the Cosmic Microwave Background (CMB), Type Ia
Supernovae (SNeIa), Weak Lensing (WL), and Baryon Acoustic Oscillations (BAO)

M. Moresco (B)
Dipartimento di Fisica e Astronomia “Augusto Righi”, Università di Bologna, Viale Berti Pichat
6/2, 40127 Bologna, Italy
e-mail: michele.moresco@unibo.it

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 277
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_15
278 M. Moresco

(for a detailed review, see, e.g. [33]). Cosmological missions have been devised and
carried out constantly improving the accuracy and precision of the measurements,
pushing all these probes almost to their limits (e.g. Planck [2] for the CMB, Pan-
theon+ [8] for SNeIa, and BOSS and eBOSS [69] for BAO). It is worth mentioning
among these the just launched ESA mission Euclid [45], which will provide the most
advanced results for BAO and WL.
Recently, significant attention has been given also to the determination of the local
value of the expansion rate, the Hubble constant (. H0 = H (z = 0)), that since first
Hubble’s measurement [32] in 1929 has improved its accuracy by orders of magni-
tude. In particular, with the observational and theoretical of recent years both direct
measurements from the local Universe considering Cepheids and SNeIa (S. H0 ES,
[68]) and indirect measurements from the Early Universe based on the analysis of
CMB (Planck, [2]) have reached the golden goal to constrain H0 with an uncertainty
of 1 km/s/Mpc and less. However, the results obtained from the two probes presented
a tension increasing with the decreasing errors of the two analyses, reaching now
a 4–5.σ disagreement called the “Hubble tension” (for a comprehensive review, see
[1, 76]). While the origin of this tension is still under investigation, as done in the
past it is fundamental now to go beyond the standard probes, which are reaching
their intrinsic limits, and explore new solutions to constrain H0 , providing new and
independent measurements that could help to shed light in this debate between early-
and late-Universe analyses.
A typical pattern of the various cosmological probes previously discussed is the
search for a standard property of objects (luminosity, length) that allows us to decou-
ple in the observable the effect due to the intrinsic evolution and to the expansion of
the Universe, making it possible to measure the latter. Historically, standard candles
(SNeIa) and standard rulers (BAO) have been amongst the most successful ones, but
in this context it is possible to exploit also the time. Using the oldest objects in the
Universe to set constraints on its age and on cosmological parameters is not new in
the literature (e.g., see [14, 38, 41, 43], and for a review [52]). Moreover, given how
the age of the Universe scales as a function of redshift:
ʃ z
dz'
t(z) =
. . (15.2)
0 H(z' )(1 + z' )

finding the oldest objects at each redshift it is possible to use them to constrain
the Hubble parameter inside the integral; for some applications of this method, see,
e.g., Refs. [6, 74, 75]. In this way, we would be searching for standard clocks,
objects whose absolute ages can be used to constrain cosmological models. This
approach has, however, some downsides. First of all, robustly determining absolute
ages is quite difficult, requiring an absolute calibration that increases the total error
and being prone to not negligible systematic effects. Moreover, as can be evinced
from Eq. 15.2, with this method we measure an integral of . H (z)1 , and typically
one has to assume a cosmological model, substitute it in Eq. 15.2, and constrain its

1 The same holds for SNeIa, that measure the luminosity distance .d L (z).
15 Addressing the Hubble Tension with Cosmic Chronometers 279

parameters. The method is, therefore, not purely cosmology-independent, since we

have to assume each time a particular cosmological model.
With the cosmic chronometers method, we can improve on this approach, bypass-
ing all these issues.

15.2 The Cosmic Chronometers Method

The idea behind the cosmic chronometer method has been introduced by Ref. [39].
Under minimal assumptions, namely just a Friedmann-Lemaître-Robertson-Walker
(FLRW) metric, given the relationship between the scale factor and redshift .a(t) =
1/(1 + z), the Hubble parameter can be derived directly from Eq. 15.1 as:

1 dz
. H (z) = − (15.3)
1 + z dt

This means that by measuring the differential age of the Universe (how much
the Universe has aged between two redshifts) it is possible to obtain a direct and
model-independent determination of the expansion rate . H (z). The main difference
here is that instead of looking for some standard clocks, we will be looking for stan-
dard chronometers, a homogeneous population of objects with a synchronized star
formation, i.e. whose clocks started “ticking” at the same time and that are there-
fore optimal tracers of the differential age evolution of the Universe. There are many
advantages in this technique with respect to using absolute ages: (i) the measurement
of differential ages has been proven to be more accurate and robust than absolute
ages, since many systematic effects in age determination are minimized when esti-
mating the differences .dt; (ii) the differential approach allows us to minimize many
other systematic effects that might affect this type of analysis (e.g. progenitor bias,
rejuvenation, ...) since for the application of the method we are only interested in
the homogeneity between the two redshift bins where the quantity .dt is estimated;
and (iii) it provides a direct determination of the Hubble parameter without relying
on any cosmological assumptions (apart the cosmological principal and a metric). A
detailed review of cosmic chronometers is provided in Ref. [52].

The pillars of the cosmic chronometers method are the following:

• Selection of a population of optimal cosmic chronometers. As can be
evinced from Eq. 15.3, the first point to address is finding a population
of objects able to trace homogeneously how much the Universe has aged
between two redshifts. We also need to ensure a way to maximize the homo-
280 M. Moresco

geneity and purity of the sample, devising criteria to minimize the potential
contamination from outliers. We describe the selection process in Sect. 15.3.
• Robust measurement of the differential age.dt. While in Eq. 15.3 there are
formally two observables (.dz and .dt) the advances in spectroscopic surveys
makes the measurement of .dz remarkably accurate when a spectroscopic
redshift is available (typically .δz/(1 + z) ≲ 10−3 , see e.g. [53]). Therefore,
for a proper application of the method, we need to find ways to obtain a
robust and unbiased determination of .dt, as we will describe in Sect. 15.4.
• Assessment of the statistical and systematic effects. For a proper appli-
cation of the method, we need to consider carefully and assess the various
sources of errors—statistical and systematic—and include those in the total
error budget with a proper covariance matrix. We will discuss this at the end
of Sect. 15.4.

15.3 Selecting an Optimal Sample Cosmic Chronometers

The most elementary objects in the Universe that we can date and that can be found
from the local Universe up to high redshifts are, of course, galaxies. To apply the
cosmic chronometer method, the idea is therefore to find at each redshift the oldest
population of galaxies available. These will, of course, have an offset with respect to
the age of the Universe given the fact that their redshift of formation is not infinite,
but as long as our selection is homogeneous this will not impact the analysis, since
the differential age .dt will not be affected by this issue.
The first and more straightforward solution that can be explored is to avoid any
selection, collect at each redshift all available galaxies, and determine the upper
envelope of the distribution, selecting the eldest crust of objects. This approach
(applied, e.g., in [39, 40, 42, 53, 71]) while having the advantage of relying on the
easiest possible selection, has the problem that to be applied we need a complete
sample with a very high statistic (typically . O(103 ) galaxies, see [40]) to robustly
detect the upper envelope.
A viable and more feasible alternative suggested in the literature is instead to select
the most massive (.log(M/Mʘ )>10.5–11) and passively evolving galaxies. There is a
large literature confirming that these objects have formed at extremely high redshifts
(.z > 2 − 3, [11–13, 21, 22, 47]) over a very short time-scale (.τ < 0.3 Gyr, [12,
16, 37, 47, 73, 74]). These galaxies have been found up to .z ∼ 2.5 [9, 20, 25, 49,
61], and with the advent of JWST this boundary has been pushed up to .z ∼ 5 [10].
These systems, having formed most of their mass at very high redshifts, in a very
quick episode of star formation, and having mostly exhausted their gas reservoir are
expected to evolve passively as a function of cosmic time; therefore, they represent the
best cosmic chronometers. This is confirmed by the fact that their stellar metallicity
is found to be very stable as a function of redshift, around solar to slightly over-solar
15 Addressing the Hubble Tension with Cosmic Chronometers 281

values at .0 ≲ z ≲ 2 [19, 22, 26, 27, 44, 47, 62, 63]. In addition, it has also been
found that more massive galaxies undergo a faster process of formation with respect
to less massive ones, forming also at earlier cosmic times. This scenario, referred to
as mass downsizing, suggests that most massive passive galaxies are therefore ideal
cosmic chronometers, being at each redshift the oldest objects and having a much
more synchronized time of formation [15, 31, 73].
Different definitions have been proposed in the literature to select passive galaxies,
based on photometry, spectroscopy, physical properties, and/or morphology. How-
ever, it has been also proven that a single selection criterion is not sufficient to obtain
a sample with high purity and that depending on the adopted criterion the contam-
ination by young/star-forming outlier can be as high as 30–50% [59] (see, e.g., the
lower-right panel of Fig. 15.1). This effect can be significantly mitigated by combin-
ing different selection criteria, following the workflow proposed by Ref. [57] and
shown in the upper panel of Fig. 15.1.
In summary, for an optimal selection of cosmic chronometers, it has been demon-
strated that the following multiple selection criteria should be combined.

• A photometric selection. It has been shown that the wider the photometric cover-
age, the better passive and star-forming galaxies can be separated. From this point
of view, the near-UV and bluer optical bands are crucial, in particular to detect
possible contamination by a younger population (0.1–1 Gyr). So far, the NUV-r-J
color-color diagram [35] has been identified as the criterion that allows the best
identification of passive galaxies (being minimally sensitive to contamination due
to dust), but valid alternatives are also the U-V-J diagram [78] and the NUV-r-K
[4]; as another option, it is also possible to assess the galaxy type modeling the
full spectral energy distribution (as done e.g. in [34, 83]).
• A spectroscopic selection. Different spectroscopic features have been identified
as ideal tracers of ongoing star formation, or of potential contamination by a
younger population; among those, the most used are the [OII].λ3727, H.β (.λ =
4861 Å), [OIII].λ5007, and H.α (.λ = 6563 Å) emission lines. Different approaches
can be used to reject objects with significant emission, either adopting a cut on the
equivalent width of the line (a typical threshold is EW.<5Å[7, 50, 53]) or on the
signal to noise ratio [58, 77]). These cuts, while increasing the purity of the sample,
are not optimized to detect the presence of potential sub-dominant contamination
by a young population, which might affect the oldest galaxies e.g. via episodes
of mergers. Recently, Ref. [57] proposed a novel spectroscopic indicator able
to precisely assess this level of contamination. The ratio between the CaII H
(.λ = 3969 Å) & K (.λ3934 Å) lines was found to be maximally sensitive to this
ratio, since the presence of a younger population, characterized by a significant
H.ϵ (.λ = 3970 Å) absorption lines, affects the CaII H/K inverting its typical ratio.
For this reason, it is a powerful diagnostic for the presence of contamination, and
it has been shown to correlate extremely well with other indicators such as star
formation rate, NUV, and optical colors [7]. As an example, in the bottom-left
panel of Fig. 15.1 it is shown how the CaII H/K ratio perfectly map the region of
passive galaxies. Depending on the data availability, it is also advisable to combine
282 M. Moresco

Fig. 15.1 Cosmic chronometer selection process. In the upper panel is shown the optimal workflow
to select a pure sample of cosmic chronometers (reproduced from [52]). The lower left plot show a
NUV-r-J color-color diagram, colored as a function of the CaII H/K ratio The dashed line displays
the theoretical lines separating passive galaxies (upper part) from star-forming galaxies (lower part),
showing the very good correlation with the CaII H/K ratio, demonstrating how it is an excellent
diagnostic to select passive galaxies. The right panel shows the effect of different selection criteria
on the spectra of passive galaxies. The sample has been obtained from Ref. [74], showing, from top
to bottom, the stacked spectra of passive galaxies selecting only on photometry (parent), adding a
U-V-J color-color cut (.UVJcut ), adding a spectroscopic cut on [OII] emission lines (.[OII]cut ), and
finally a cut on CaII H/K ratio (CC). It is evident how combining more and more restrictive selection
criteria, the emission lines progressively disappear from the stacked spectra, displaying at the end
the characteristic features of a purely passive population

several of these indicators to maximize the purity of the sample. In the end, in the
spectra of the galaxies ideally should not be visible any residual emission lines, as
shown by the bottom-right plot of Fig. 15.1.
• A cut in stellar mass or stellar velocity dispersion. As previously discussed,
different studies have shown that the most massive galaxies are the ones that formed
before and over the shortest timescale[15, 31, 73]; a cut in stellar mass (typically
.log(M/Mʘ ) >10.6–11) ensures the synchronicity of the sample selected, and, if
15 Addressing the Hubble Tension with Cosmic Chronometers 283

not available, it can equivalently be done on the stellar velocity dispersion of the
galaxy, since it presents a good correlation with the stellar mass.
The combination of these criteria has been demonstrated to be effective to maximize
the purity of the sample, minimizing possible residual contamination of a young
and/or star-forming subdominant population. However, it might not be always feasi-
ble to combine multiple criteria, due to data quality or availability. In that case, it is
crucial to try to characterize the purity of the sample and properly take into account
potential issues in the total covariance matrix as discussed in Sect. 15.4.

15.4 Measuring the Differential Ages

Measuring the age of a stellar population is not a trivial task. While encoded in the
integrated spectrum of the galaxy analyzed, this information presents some degener-
acy with other physical properties, the most known being the age-metallicity degen-
eracy [24, 79] (other noticeable degeneracies are the ones with the star formation
history [28] or with dust [65]). However, the information encoded in the photometry
and spectrum of a galaxy allows us to break (or mitigate in some cases) these degen-
eracies, and different methods have been introduced and exploited to provide a robust
measurement. First of all, we need to underline here that in the cosmic chronometers
approach we are interested in determining the relative age .dt, and not the absolute
age .t; this means that we will not be dependent on any offset in the determination
of the age, provided that the differential measurement is unbiased, and it has been
shown that the relative ages are more robust and can achieve higher accuracy than
absolute ones, removing many systematic effects when taking the differences [46].
We summarize here in the following the main methods that have been used to
derive .dt, leaving to Ref. [52] a detailed discussion, and report in Table 15.1 and
Fig. 15.2 the Hubble parameters determined with the various approaches.
Full-spectral fitting. The first approach that can be adopted is to exploit the full
information available. Different codes are publicly available performing a fit to the
photometric and spectroscopic data with theoretical spectra obtained combining the
different physical ingredients of the stellar population of a galaxy (age, metallicity,
mass, star formation history). This allows us to determine the ages of our chronome-
ters, from which it is possible to derive the differential age by binning the data at
different redshifts and estimating the quantity .dt. Successful cosmic chronometers
analyses have been performed with this method by several groups [66, 71, 72, 81],
that have also demonstrated how with a good photometric and spectroscopic cover-
age the .dt measurement is very robust and independent of the specific assumptions
chosen to model the spectra [37, 74].
Analysis of absorption lines (Lick indices). Instead of considering the full spec-
tral information as in the previous method, an alternative is to study instead the
absorption lines that characterize the spectra of passive galaxies, since they are known
284 M. Moresco

Table 15.1 Measurements of the Hubble parameter derived with the cosmic chronometers method
obtained with the CC method (in units of km/s/Mpc). The error on. H (z) reported here accounts only
for the diagonal part of the covariance matrix, while for a proper analysis also the full covariance
should be considered as discussed in Sect. 15.4. The method (M) used to derive the differential age
.dt is also provided (full-spectrum fitting F, Lick indices L, . D4000 D, Machine Learning ML), and
the corresponding reference
.z . H (z) .σ H (z) M References .z . H (z) .σ H (z) M References
0.07 69.0 19.6 F [81] 0.593 104 13 D [53]
0.09 69 12 F [71] 0.68 92 8 D [53]
0.12 68.6 26.2 F [81] 0.75 98.8 33.6 L [6].∗
0.17 83 8 F [71] 0.75 105 10.76 ML [40].∗
0.179 75 4 D [53] 0.781 105 12 D [53]
0.199 75 5 D [53] 0.8 113.1 25.22 F [37].∗
0.20 72.9 29.6 F [81] 0.875 125 17 D [53]
0.27 77 14 F [71] 0.88 90 40 F [72]
0.28 88.8 36.6 F [81] 0.9 117 23 F [71]
0.352 83 14 D [53] 1.037 154 20 D [53]
0.38 83 13.5 D [58] 1.26 135 65 F [74]
0.4 95 17 F [71] 1.3 168 17 F [71]
0.4004 77 10.2 D [58] 1.363 160 33.6 D [51]
0.425 87.1 11.2 D [58] 1.43 177 18 F [71]
0.445 92.8 12.9 D [58] 1.53 140 14 F [71]
0.47 89.0 49.6 F [66] 1.75 202 40 F [71]
0.4783 80.9 9 D [58] 1.965 186.5 50.4 D [51]
0.48 97 62 F [72]
.∗ These data have been obtained from the same sample, and should not be used together in an analysis

Fig. 15.2 Latest compilation of Hubble parameter measurements as a function of redshift obtained
with cosmic chronometers. Different symbols indicate a measurement performed with a different
method, as discussed in Sect. 15.4 and presented in Table 15.1
15 Addressing the Hubble Tension with Cosmic Chronometers 285

to be particularly sensitive to the age (e.g., Balmer lines), metal content (e.g., iron
lines), and enhancement in .α elements (e.g., magnesium lines) of the population.
Similarly as before, it is possible to build theoretical models to interpret the strength
of these features, known as Lick indices, as a function of the age, metallicity, and.α/Fe
of the population, and therefore extract this information by analyzing a combination
of indices. Also with this method, it has been recently obtained a determination of
the Hubble parameter with the cosmic chronometer approach [6].
Analysis of specific spectral features (.D4000). A slightly different approach
consists in still considering spectroscopic features, but focusing on a specific spectral
feature known to correlate particularly well with the age of the population instead
of analyzing a joint combination of those. This approach, introduced by Ref. [54],
suggested using the break at 4000 Å restframe (known as . D4000, see Fig. 15.1) as
an age indicator since it has been found that for cosmic chronometers it correlates
linearly with the stellar age (at fixed metallicity and star formation history). If there is
a linear relation between the. D4000 and the age modeled as. D4000 = A(Z , S F H ) ×
age + B, Ref. [53] proposed to rewrite Eq. 15.3 in the form:

1 dz
. H (z) = − A(S F R, Z /Z ʘ ) , (15.4)
1+z d D4000

where . A(S F R, Z /Z ʘ ) is a calibration parameter depending on the metallicity and

star formation history obtained from stellar population synthesis models. We illustra-
tively show how this method works in Fig. 15.3. This equation improves on Eq. 15.3
since it allows to decouple statistical and systematic effects because the differential
evolution will be measured on purely observational data (.d D4000, purely including
statistical errors), while the systematic errors will be all contained in the calibration
parameter. To derive this properly, the metallicity of the sample has to be derived
with independent measurements, or assumed in a realistic range; here we underline
that given the cosmic chronometers selection described in Sect. 15.3, we expect the
metal content to be rather homogeneous as a function of redshift. However, it is
fundamental to keep into account this effect, as well as the dependence of the results
on the model to derive the calibration parameter, in the total covariance matrix, to
properly include in the error budget all systematic effects as discussed below. With
this approach, the most accurate . H (z) determinations of . H (z) have been derived to
date, with accuracy going from .∼5% at .z ∼ 0.15, to .∼10% at .z ∼ 0.8, to .∼25% at
. z ∼ 2 [51, 53, 58].

Other approaches. Recently, Ref. [40] demonstrated that it is possible to apply

the cosmic chronometer approach also considering purely photometric data, using a
Machine Learning algorithm trained on well-determined aged (obtained from Lick
indices analysis). This method, while needing to rely on larger samples (. O(104 )) to
determine the upper envelope of the age-redshift relation, is extremely promising in
view of all the surveys that will provide photometry for almost the full sky in the
near future (e.g. Euclid [45] and Rubin [36]).
286 M. Moresco

Fig. 15.3 Illustrative representation of how to apply the . D4000 method in the cosmic chronome-
ters approach (reproduced from [52]. The left panel shows an example of median . D4000-z val-
ues obtained from an analysis of SDSS BOSS survey, representing how to derive the quantity
.dz/d D4000. The right panel show for a specific model the . D4000-age relations, where it is high-
lighted their dependence on metallicity (different colors) and star formation history (different lines).
It is also possible to see how these relations can be approximated with piecewise linear relation
with good accuracy

Cosmic chronometers full covariance matrix. The full covariance matrix for
cosmic chronometers has been formalized in Ref. [55], and can be divided into
different contributions:

Covitotj = Covistat
. j +
young
+Covimet
j + Covi j +
+CoviSFH
j + CoviIMF
j + Covist.lib.
j + CoviSPS
j . (15.5)

The first row of Eq. 15.5 reports the component related to the statistical errors (e.g. the
one related to the measurement of .d D4000 if using the . D4000 method to measure
.dt). The second row reports the systematic contribution to the covariance due to
uncertainty in the determination of the metallicity of the sample, or to a non-perfect
selection of the sample when residual contamination by a young population is not
completely removed; this second effect has been quantified in Ref. [57], where a
recipe to derive it directly from observable is provided. These two effects are expected
to provide diagonal contributions to the covariance since they depend on uncorrelated
properties of the galaxies selected. The last row describes instead the systematic
component of the covariance due to the ingredients of the model assumed. It has
been divided into various parts, i.e. the one related to the uncertainty on the stellar
population synthesis model considered (SPS), the one related to the stellar library
that the model adopts (st.lib.), the one related to the initial mass function considered
in the model (IMF), and the one depending on the assumed star formation history
(SFH). All these parts have been thoroughly explored in Ref. [55], where also a
15 Addressing the Hubble Tension with Cosmic Chronometers 287

Jupyter notebook is provided as an example of how to derive the total covariance

matrix for cosmic chronometers.
To conclude, we note here that several other effects have been considered that
might in principle affect the cosmological measurements, such as the progenitor
bias or a dependence on the stellar mass of the sample. These effects, discussed in
Ref. [52], are expected however not to have a significant impact on the analysis due
to the strict selection process described in Sect. 15.3 and since in the differential
approach relative ages are estimated in homogeneous samples really close in cosmic
time, minimizing therefore by definition these effects.

15.5 Cosmic Chronometers and the Hubble Constant

As shown in Fig. 15.2, cosmic chronometers do not directly constrain the Hubble
constant, but rather measure the Hubble parameter independently of any cosmologi-
cal assumptions. For this reason, they represent the ideal testbed to compare different
cosmological models. While not measuring H0 directly, however, it can be derived
from these data in different ways.
A possible solution is to estimate the Hubble constant as the extrapolated value
of . H (z) at .z = 0. To exploit at most the cosmology-independent nature of these
measurements, different methods have been suggested, either based on Gaussian
Process regression (e.g. [29, 30, 70], but see also [18] on the assumptions of the
method and potential dependence of the measured errors on the assumed kernel)
or on Machine Learning (e.g. [3, 48]). The most recent constraint obtained with
these methods give . H0 = 70.7 ± 6.7 km/s/Mpcwhen considering the full covariance
matrix [23].
The most direct way to derive H0 is to assume a cosmological model and fit the data
presented in Table 15.1, properly considering their covariance matrix as discussed
above. We assumed here two cosmological models, a flat .ɅCDM model(the one with
the least free parameters, H0 and Ωm ) and the open wCDM model, where also the
curvature of the Universe and the dark energy equation of state parameter are left
free. This can give us a range of the expected constraints while varying the number
of parameters in the fit. Moreover, we also extend our analysis by adding to the
current . H (z) measurements the forecasts discussed by Ref. [52], where two sets of
simulated data are provided as expected by the analysis of the just launched Euclid
mission and from the exploitation of newly derived data (such as the latest SDSS
BOSS and eBOSS surveys). In this case, two different simulations are explored, a
standard one where current data are combined with these simulated data (combined)
and one where a more optimistic impact of systematic error is considered for the
simulated data, minimizing the contribution due to different models in the analysis.
The constraints are shown in Fig. 15.4, and the results are reported in Table 15.2. We
find that current data are able to measure H0 with an accuracy of 8% in a flat .λCDM
cosmology, and of 11% in an open wCDM cosmology, with values . H0 = 66.7 ± 5.3
+8.8
and .68.0−7.1 km/s/Mpc, respectively. As can be seen, the constraints at the moment
288 M. Moresco

Fig. 15.4 Constraints in the H0 -Ωm plane obtained from the analysis of cosmic chronometers for
two cosmological models, a flat.ɅCDM (left panel) and an open wCDM (right panel). The different
colors refer to the analysis of the current dataset presented in Table 15.1 (in blue), and with the
forecasts for future measurements in two configurations (a standard one in red, and an optimistic
one in gold, as discussed in Sect. 15.5. Dashed lines report as a reference the cosmological parameters
obtained by Planck2018 [2]

are preferring a lower value of . H0 , confirming the results obtained with Gaussian
Processes, but with an errorbar that does not allows us to significantly discriminate
between S. H0 ES and CMB values. However, it is interesting to notice how future
data could significantly improve, reaching an accuracy of around 3–1% depending
on the cosmological model assumed and on the simulation considered. This sheds
a very promising light on this method since it suggests that future analysis and a
more thorough study of the models impacting on the systematic errors might provide
competitive H0 measurements, that can help in giving important independent pieces
of evidence to address the H0 tension.
Finally, another possibility is to explore how much the constraints would improve
by combining cosmic chronometers with other cosmological probes. This has been
done by several authors (e.g. [5, 56, 60, 80, 82]), and the most recent measurement
to date obtained combining SNeIa, BAO, cosmic chronometers and Gamma-Ray
+3.4
Bursts data gives . H0 = 67.2−3.2 km/s/Mpc(for a .ɅCDM cosmology, . H0 = 66.9+3.5
−3.4
km/s/Mpc for an open wCDM cosmology, [17]).
15 Addressing the Hubble Tension with Cosmic Chronometers 289

Table 15.2 Cosmological parameter constraints that can be obtained with present-day and future
CC measurements (as presented in Sect. 15.5) in two different cosmologies, a flat.ɅCDM cosmology
(upper rows) and a more general open .wCDM (lower rows). The Hubble constant H0 is given in
units km/s/Mpc
H0 % acc Ωm % acc Ωde % acc .w % acc
Flat .ɅCDM
Current dataset .66.7 ± 5.3 8% .0.33+0.08
−0.06 20.6% – – – –
Combined .69.0 ± 2.1 3% .0.3 ± 0.01 4.4% – – – –
Optimistic .68.2 ± 0.7 1% .0.29 ± 4.2% – – – –
0.01
Open .wCDM
Current dataset +8.8
.68.0−7.1 11.7% +0.15
.0.22−0.12 62% .0.51+0.24 48% .−1.6+0.8 52%
−0.26 −0.9
+3.1 +0.09 +0.18 +0.3
Combined .71.6−2.6 4% .0.29−0.09 31% .0.67−0.14 24% .−1.2−0.4 29%
Optimistic .70.8+2.8
−1.9 3.3% .0.3+0.08
−0.09 28% .0.68+0.18
−0.11 22% +0.3
.−1.2−0.4 29%

Acknowledgements MM acknowledges the grants ASI n.I/023/12/0 and AS I n.2018–23-HH.0.,

and support from MIUR, PRIN 2017 (grant 20179ZF5KS) and from grant MIUR, PRIN 2022 (grant
2022NY2ZRS_001).

References

1. E. Abdalla, G. Franco Abellán, A. Aboubrahim, A. Agnello, O. Akarsu, Y. Akrami, G. Alestas,

D. Aloni, L. Amendola, L.A. Anchordoqui et al., JHEAp 34, 49–211 (2022) https://doi.org/
10.1016/j.jheap.2022.04.002 ([arXiv:2203.06142 [astro-ph.CO]].)
2. N. Aghanim et al. [Planck], Astron. Astrophys. 641, A6 (2020) [erratum: Astron. Astrophys.
652, C4 (2021)]. https://doi.org/10.1051/0004-6361/201833910. ([arXiv:1807.06209 [astro-
ph.CO]] .)
3. R. Arjona, S. Nesseris, Phys. Rev. D 101(12), 123525 (2020). https://doi.org/10.1103/
PhysRevD.101.123525. ([arXiv:1910.01529 [astro-ph.CO]].)
4. S. Arnouts, E. Le Floc’h, J. Chevallard, B.D. Johnson, O. Ilbert, M. Treyer, H. Aussel, P. Capak,
D.B. Sanders, N. Scoville et al., Astron. Astrophys. 558, A67 (2013). https://doi.org/10.1051/
0004-6361/201321768. ([arXiv:1309.0008 [astro-ph.CO]].)
5. A. Bonilla, S. Kumar, R.C. Nunes, Eur. Phys. J. C 81(2), 127 (2021). https://doi.org/10.1140/
epjc/s10052-021-08925-z. ([arXiv:2011.07140 [astro-ph.CO]])
6. N. Borghi, M. Moresco, A. Cimatti, Astrophys. J. Lett. 928(1), L4 (2022). https://doi.org/10.
3847/2041-8213/ac3fb2 ([arXiv:2110.04304 [astro-ph.CO]].)
7. N. Borghi, M. Moresco, A. Cimatti, A. Huchet, S. Quai, L. Pozzetti, Astrophys. J. 927(2), 164
(2022). https://doi.org/10.3847/1538-4357/ac3240. ([arXiv:2106.14894 [astro-ph.GA]].)
8. D. Brout, D. Scolnic, B. Popovic, A.G. Riess, J. Zuntz, R. Kessler, A. Carr, T.M. Davis, S.
Hinton, D. Jones et al., Astrophys. J. 938(2), 110 (2022). https://doi.org/10.3847/1538-4357/
ac8e04. ([arXiv:2202.04077 [astro-ph.CO]].)
9. K.I. Caputi, J.S. Dunlop, R.J. McLure, J. Huang, G.G. Fazio, M.L.N. Ashby, M. Castellano,
A. Fontana, M. Cirasuolo, O. Almaini et al., Astrophys. J. Lett. 750, L20 (2012). https://doi.
org/10.1088/2041-8205/750/1/L20. ([arXiv:1202.0496 [astro-ph.CO]].)
10. A.C. Carnall, D.J. McLeod, R.J. McLure, J.S. Dunlop, R. Begley, F. Cullen et al.,
Mon. Not. R. Astron. Soc. 520, 3974–3985 (2023). https://doi.org/10.1093/mnras/stad369.
([arXiv:astro-ph/2208.00986 [astro-ph]].)
290 M. Moresco

11. A.C. Carnall, R.J. McLure, J.S. Dunlop, F. Cullen, D.J. McLeod et al., Mon. Not. R. Astron. Soc.
490, 417–439 (2019). https://doi.org/10.1093/mnras/stz2544. ([arXiv:astro-ph/1903.11082
[astro-ph]].)
12. A.C. Carnall, R.J. McLure, J.S. Dunlop, R. Davé, Mon. Not. R. Astron. Soc. 480, 4379–4401
(2018). https://doi.org/10.1093/mnras/sty2169. ([arXiv:astro-ph/1712.04452 [astro-ph]].)
13. J. Choi, C. Conroy, J. Moustakas, G.J. Graves, B.P. Holden, M. Brodwin, M.J.I. Brown, P.G.
van Dokkum, Astrophys. J. 792(2), 95 (2014). https://doi.org/10.1088/0004-637X/792/2/95.
([arXiv:1403.4932 [astro-ph.GA]].)
14. A. Cimatti and M. Moresco, [arXiv:2302.07899 [astro-ph.CO]]
15. A. Cimatti, E. Daddi, A. Renzini, P. Cassata, E. Vanzella, L. Pozzetti, S. Cristiani, A. Fontana,
G. Rodighiero, M. Mignoli et al., Nature 430, 184–187 (2004). https://doi.org/10.1038/
nature02668. ([arXiv:astro-ph/0407131 [astro-ph]].)
16. A. Citro, L. Pozzetti, M. Moresco, A. Cimatti, Astron. Astrophys. 592, A19 (2016). https://
doi.org/10.1051/0004-6361/201527772. ([arXiv:1604.07826 [astro-ph.CO]].)
17. F. Cogato, M. Moresco, L. Amati and A. Cimatti, [arXiv:2309.01375 [astro-ph.CO]]
18. E.Ó. Colgáin, M.M. Sheikh-Jabbari, Eur. Phys. J. C 81(10), 892 (2021). https://doi.org/10.
1140/epjc/s10052-021-09708-2. ([arXiv:2101.08565 [astro-ph.CO]].)
19. C. Conroy, G. Graves, P. van Dokkum, Astrophys. J. 780, 33 (2014). https://doi.org/10.1088/
0004-637X/780/1/33. ([arXiv:1303.6629 [astro-ph.CO]].)
20. E. Daddi, A. Cimatti, A. Renzini, J. Vernet, C. Conselice, L. Pozzetti, M. Mignoli, P. Tozzi,
T. J. Broadhurst and S. di Serego Alighieri et al., Astrophys. J. Lett. 600, L127–L130 (2004).
https://doi.org/10.1086/381020. ([arXiv:astro-ph/0308456 [astro-ph]].)
21. E. Daddi, A. Renzini, N. Pirzkal, A. Cimatti, S. Malhotra, M. Stiavelli, C. Xu, A. Pasquali, J.E.
Rhoads, M. Brusa et al., Astrophys. J. 626, 680–697 (2005). https://doi.org/10.1086/430104.
[arXiv:astro-ph/0503102 [astro-ph]].)
22. V. Estrada-Carpenter, C. Papovich, I. Momcheva, G. Brammer, J. Long, R.F. Quadri, J. Bridge,
M. Dickinson, H. Ferguson, S. Finkelstein et al., Astrophys. J. 870, 133–160 (2019). https://
doi.org/10.3847/1538-4357/aaf22e. ([arXiv:astro-ph/1810.02824 [astro-ph]].)
23. A. Favale, A. Gómez-Valent, M. Migliaccio, Mon. Not. R. Astron. Soc. 523(3), 3406–3422
(2023). https://doi.org/10.1093/mnras/stad1621. ([arXiv:2301.09591 [astro-ph.CO]].)
24. I. Ferreras, S. Charlot, J. Silk, Astrophys. J. 521, 81 (1999). https://doi.org/10.1086/307513.
([arXiv:astro-ph/9803235 [astro-ph]].)
25. A. Fontana, S. Salimbeni, A. Grazian, E. Giallongo, L. Pentericci, M. Nonino, F. Fontanot, N.
Menci, P. Monaco, S. Cristiani et al., Astron. Astrophys. 459, 745–757 (2006). https://doi.org/
10.1051/0004-6361:20065475. ([arXiv:astro-ph/0609068 [astro-ph]].)
26. A. Gallazzi, E.F. Bell, S. Zibetti, J. Brinchmann, D.D. Kelson, Astrophys. J. 788, 72 (2014).
https://doi.org/10.1088/0004-637X/788/1/72. ([arXiv:1404.5624 [astro-ph.GA]].)
27. A. Gallazzi, S. Charlot, J. Brinchmann, S.D.M. White, C.A. Tremonti, Mon. Not.
R. Astron. Soc. 362, 41–58 (2005). https://doi.org/10.1111/j.1365-2966.2005.09321.x.
([arXiv:astro-ph/0506539 [astro-ph]].)
28. G. Gavazzi, C. Bonfanti, G. Sanvito, A. Boselli, M. Scodeggio, Astrophys. J. 576, 1 (2002).
https://doi.org/10.1086/341730
29. A. Gómez-Valent, L. Amendola, JCAP 04, 051 (2018). https://doi.org/10.1088/1475-7516/
2018/04/051. ([arXiv:1802.01505 [astro-ph.CO]].)
30. B.S. Haridasu, V.V. Luković, M. Moresco, N. Vittorio, JCAP 10, 015 (2018). https://doi.org/
10.1088/1475-7516/2018/10/015. ([arXiv:1805.03595 [astro-ph.CO]].)
31. A. Heavens, B. Panter, R. Jimenez, J. Dunlop, Nature 428, 625 (2004). https://doi.org/10.1038/
nature02474. ([arXiv:astro-ph/0403293 [astro-ph]].)
32. E. Hubble, Proc. Nat. Acad. Sci. 15, 168–173 (1929). https://doi.org/10.1073/pnas.15.3.168
33. D. Huterer, D.L. Shafer, Rept. Prog. Phys. 81(1), 016901 (2018). https://doi.org/10.1088/1361-
6633/aa997e. ([arXiv:1709.01091 [astro-ph.CO]].)
34. O. Ilbert, P. Capak, M. Salvato, H. Aussel, H.J. McCracken, D.B. Sanders, N. Scoville, J.
Kartaltepe, S. Arnouts, E.L. Floc’h et al., Astrophys. J. 690, 1236–1249 (2009). https://doi.
org/10.1088/0004-637X/690/2/1236. ([arXiv:0809.2101 [astro-ph]].)
15 Addressing the Hubble Tension with Cosmic Chronometers 291

35. O. Ilbert, H.J. McCracken, O.L. Fevre, P. Capak, J. Dunlop, S. Arnouts, H. Aussel, K. Caputi,
J. Comparat, Q. Guo et al., Astron. Astrophys. 556, A55 (2013). https://doi.org/10.1051/0004-
6361/201321100. ([arXiv:1301.3157 [astro-ph.CO]].)
36. Ž Ivezić et al., LSST. Astrophys. J. 873(2), 111 (2019). https://doi.org/10.3847/1538-4357/
ab042c. ([arXiv:0805.2366 [astro-ph]].)
37. K. Jiao, N. Borghi, M. Moresco, T.J. Zhang, Astrophys. J. Suppl. 265(2), 48 (2023). https://
doi.org/10.3847/1538-4365/acbc77. ([arXiv:2205.05701 [astro-ph.CO]].)
38. R. Jimenez, A. Cimatti, L. Verde, M. Moresco, B. Wandelt, JCAP 03, 043 (2019). https://doi.
org/10.1088/1475-7516/2019/03/043. ([arXiv:1902.07081 [astro-ph.CO]].)
39. R. Jimenez, A. Loeb, Astrophys. J. 573, 37–42 (2002). https://doi.org/10.1086/340549.
([arXiv:astro-ph/0106145 [astro-ph]].)
40. R. Jimenez, M. Moresco, L. Verde, B.D. Wandelt, [arXiv:2306.11425 [astro-ph.CO]]
41. R. Jimenez, P. Thejll, U. Jorgensen, J. MacDonald, B. Pagel, Mon. Not. R. Astron. Soc. 282,
926–942 (1996). https://doi.org/10.1093/mnras/282.3.926. ([arXiv:astro-ph/9602132 [astro-
ph]].)
42. R. Jimenez, L. Verde, T. Treu, D. Stern, Astrophys. J. 593, 622–629 (2003). https://doi.org/10.
1086/376595. ([arXiv:astro-ph/0302560 [astro-ph]].)
43. L.M. Krauss, B. Chaboyer, Science 299, 65–70 (2003). https://doi.org/10.1126/science.
1075631
44. M. Kriek, S.H. Price, C. Conroy, K.A. Suess, L. Mowla, I. Pasha, R. Bezanson, P. Van Dokkum,
G. Barro, Astrophys. J. Lett. 880(2), L31 (2019). https://doi.org/10.3847/2041-8213/ab2e75.
([arXiv:astro-ph/1907.04327 [astro-ph]].)
45. R. Laureijs et al., [EUCLID], [arXiv:1110.3193 [astro-ph.CO]]
46. A. Marin-Franch, A. Aparicio, G. Piotto, A. Rosenberg, B. Chaboyer, A. Sarajedini, M. Siegel,
J. Anderson, L.R. Bedin, A. Dotter et al., Astrophys. J. 694, 1498–1516 (2009). https://doi.
org/10.1088/0004-637X/694/2/1498. ([arXiv:0812.4541 [astro-ph]].)
47. R.M. McDermid, K. Alatalo, L. Blitz, F. Bournaud, M. Bureau, M. Cappellari, A.F. Crocker,
R.L. Davies, T.A. Davis, P.T. de Zeeuw et al., Mon. Not. R. Astron. Soc. 448(4), 3484–3513
(2015). https://doi.org/10.1093/mnras/stv105. ([arXiv:1501.03723 [astro-ph.GA]].)
48. A. Mehrabi, M. Rezaei, Astrophys. J. 923(2), 274 (2021). https://doi.org/10.3847/1538-4357/
ac2fff. ([arXiv:2110.14950 [astro-ph.CO]].)
49. E. Merlin, F. Fortuni, M. Torelli, P. Santini, M. Castellano, A. Fontana, A. Grazian, L. Pentericci,
S. Pilo, K.B. Schmidt Mon. Not. R. Astron. Soc. 490(3), 3309–3328 (2019). https://doi.org/
10.1093/mnras/stz2615. ([arXiv:1909.07996 [astro-ph.GA]].)
50. M. Mignoli, G. Zamorani, M. Scodeggio, A. Cimatti, C. Halliday, S.J. Lilly, L. Pozzetti, D.
Vergani, C.M. Carollo, T. Contini et al., Astron. Astrophys. 493, 39 (2009). https://doi.org/10.
1051/0004-6361:200810520. ([arXiv:0810.2245 [astro-ph]].)
51. M. Moresco, Mon. Not. R. Astron. Soc. 450(1), L16–L20 (2015). https://doi.org/10.1093/
mnrasl/slv037. ([arXiv:1503.01116 [astro-ph.CO]].)
52. M. Moresco, L. Amati, L. Amendola, S. Birrer, J.P. Blakeslee, M. Cantiello, A. Cimatti, J. Dar-
ling, M. Della Valle, M. Fishbach et al., Living Rev. Rel. 25(1), 6 (2022). https://doi.org/10.
1007/s41114-022-00040-z). ([arXiv:2201.07241 [astro-ph.CO]].)
53. M. Moresco, A. Cimatti, R. Jimenez, L. Pozzetti, G. Zamorani, M. Bolzonella, J. Dunlop, F.
Lamareille, M. Mignoli, H. Pearce et al., JCAP 08, 006 (2012). https://doi.org/10.1088/1475-
7516/2012/08/006. ([arXiv:1201.3609 [astro-ph.CO]].)
54. M. Moresco, R. Jimenez, A. Cimatti, L. Pozzetti, JCAP 03, 045 (2011). https://doi.org/10.
1088/1475-7516/2011/03/045. ([arXiv:1010.0831 [astro-ph.CO]].)
55. M. Moresco, R. Jimenez, L. Verde, A. Cimatti, L. Pozzetti, Astrophys. J. 898(1), 82 (2020).
https://doi.org/10.3847/1538-4357/ab9eb0. ([arXiv:2003.07362 [astro-ph.GA]].)
56. M. Moresco, R. Jimenez, L. Verde, A. Cimatti, L. Pozzetti, C. Maraston, D. Thomas, JCAP
12, 039 (2016). https://doi.org/10.1088/1475-7516/2016/12/039. ([arXiv:1604.00183 [astro-
ph.CO]].)
57. M. Moresco, R. Jimenez, L. Verde, L. Pozzetti, A. Cimatti, A. Citro, Astrophys. J. 868(2), 84
(2018). https://doi.org/10.3847/1538-4357/aae829. ([arXiv:1804.05864 [astro-ph.CO]].)
292 M. Moresco

58. M. Moresco, L. Pozzetti, A. Cimatti, R. Jimenez, C. Maraston, L. Verde, D. Thomas, A. Citro,

R. Tojeiro, D. Wilkinson, JCAP 05, 014 (2016). https://doi.org/10.1088/1475-7516/2016/05/
014. ([arXiv:1601.01701 [astro-ph.CO]].)
59. M. Moresco, L. Pozzetti, A. Cimatti, G. Zamorani, M. Bolzonella, F. Lamareille, M. Mignoli,
E. Zucca, S.J. Lilly, C.M. Carollo et al., Astron. Astrophys. 558, A61 (2013). https://doi.org/
10.1051/0004-6361/201321797. ([arXiv:1305.1308 [astro-ph.CO]].)
60. M. Moresco, L. Verde, L. Pozzetti, R. Jimenez, A. Cimatti, JCAP 07, 053 (2012). https://doi.
org/10.1088/1475-7516/2012/07/053. ([arXiv:1201.6658 [astro-ph.CO]].)
61. A. Muzzin, D. Marchesini, M. Stefanon, M. Franx, H.J. McCracken, B. Milvang-Jensen, J.S.
Dunlop, J.P.U. Fynbo, O. Le Fevre, G. Brammer et al., Astrophys. J. 777, 18 (2013). https://
doi.org/10.1088/0004-637X/777/1/18. ([arXiv:1303.4409 [astro-ph.CO]].)
62. M. Onodera, C.M. Carollo, A. Renzini, M. Cappellari, C. Mancini, N. Arimoto, E. Daddi, R.
Gobat, V. Strazzullo, S. Tacchella et al., Astrophys. J. 808(2), 161 (2015). https://doi.org/10.
1088/0004-637X/808/2/161. ([arXiv:1411.5023 [astro-ph.GA]].)
63. M. Onodera, A. Renzini, M. Carollo, M. Cappellari, C. Mancini, V. Strazzullo, E. Daddi, N.
Arimoto, R. Gobat, Y. Yamada et al., Astrophys. J. 755, 26 (2012). https://doi.org/10.1088/
0004-637X/755/1/26. ([arXiv:1206.1540 [astro-ph.CO]].)
64. S. Perlmutter et al., Supernova cosmology project. Astrophys. J. 517, 565–586 (1999). https://
doi.org/10.1086/307221. ([arXiv:astro-ph/9812133 [astro-ph]].)
65. L. Pozzetti, F. Mannucci, Mon. Not. R. Astron. Soc. 317, L17 (2000). https://doi.org/10.1046/
j.1365-8711.2000.03829.x. ([arXiv:astro-ph/0006430 [astro-ph]].)
66. A.L. Ratsimbazafy, S.I. Loubser, S.M. Crawford, C.M. Cress, B.A. Bassett, R.C. Nichol, P.
Väisänen, Mon. Not. R. Astron. Soc. 467(3), 3239–3254 (2017). https://doi.org/10.1093/mnras/
stx301. ([arXiv:1702.00418 [astro-ph.CO]].)
67. A.G. Riess et al., Supernova Search Team. Astron. J. 116, 1009–1038 (1998). https://doi.org/
10.1086/300499. ([arXiv:astro-ph/9805201 [astro-ph]].)
68. A.G. Riess, W. Yuan, L.M. Macri, D. Scolnic, D. Brout, S. Casertano, D.O. Jones, Y. Murakami,
L. Breuval, T.G. Brink et al., Astrophys. J. Lett. 934(1), L7 (2022). https://doi.org/10.3847/
2041-8213/ac5c5b. ([arXiv:2112.04510 [astro-ph.CO]].)
69. A.J. Ross, J. Bautista, R. Tojeiro, S. Alam, S. Bailey, E. Burtin, J. Comparat, K.S. Dawson,
A. de Mattia, H. du Mas des Bourboux et al., Mon. Not. R. Astron. Soc. 498(2), 2354–2371
(2020). https://doi.org/10.1093/mnras/staa2416 ([arXiv:2007.09000 [astro-ph.CO]].)
70. M. Seikel, S. Yahya, R. Maartens, R. Clarkson, Phys. Rev. D 86, 083001 (2012). https://doi.
org/10.1103/PhysRevD.86.083001. ([arXiv:astro-ph/1205.3431 [astro-ph]].)
71. J. Simon, L. Verde, R. Jimenez, Phys. Rev. D 71, 123001 (2005). https://doi.org/10.1103/
PhysRevD.71.123001. ([arXiv:astro-ph/0412269 [astro-ph]].)
72. D. Stern, R. Jimenez, L. Verde, M. Kamionkowski, S.A. Stanford, JCAP 02, 008 (2010). https://
doi.org/10.1088/1475-7516/2010/02/008. ([arXiv:0907.3149 [astro-ph.CO]].)
73. D. Thomas, C. Maraston, K. Schawinski, M. Sarzi, J. Silk, Mon. Not. R. Astron. Soc. 404,
1775 (2010). https://doi.org/10.1111/j.1365-2966.2010.16427.x. ([arXiv:0912.0259 [astro-
ph.CO]].)
74. E. Tomasetti, M. Moresco, N. Borghi, K. Jiao, A. Cimatti, L. Pozzetti, A.C. Carnall,
R.J. McLure, L. Pentericci, [arXiv:2305.16387 [astro-ph.CO]]
75. S. Vagnozzi, F. Pacucci, A. Loeb, JHEAp 36, 27–35 (2022). https://doi.org/10.1016/j.jheap.
2022.07.004. ([arXiv:2105.10421 [astro-ph.CO]].)
76. L. Verde, T. Treu, A.G. Riess, Nat. Astron. 3, 891. https://doi.org/10.1038/s41550-019-0902-
0. ([arXiv:1907.10625 [astro-ph.CO]])
77. L.L. Wang, A.L. Luo, S.Y. Shen, W. Hou, X. Kong, Y.H. Song, J.N. Zhang, H. Wu, Z.H.Z.
Cao, Y.H. Hou, Y.F. Wang, Y. Zhang, Y.H. Zhao, Mon. Not. R. Astron. Soc. 474(2), 1873–1885
(2018). https://doi.org/10.1093/mnras/stx2798. ( [arXiv:1710.10611 [astro-ph.CO]].)
78. R.J. Williams, R.F. Quadri, M. Franx, P. van Dokkum, I. Labbe, Astrophys. J. 691, 1879–1895
(2009). https://doi.org/10.1088/0004-637X/691/2/1879. ([arXiv:0806.0625 [astro-ph]].)
79. G. Worthey, Astrophys. J. Suppl. 95, 107–149 (1994). https://doi.org/10.1086/192096
15 Addressing the Hubble Tension with Cosmic Chronometers 293

80. H. Yu, B. Ratra, F.Y. Wang, Astrophys. J. 856(1), 3 (2018). https://doi.org/10.3847/1538-4357/

aab0a2. ([arXiv:1711.03437 [astro-ph.CO]].)
81. C. Zhang, H. Zhang, S. Yuan, T.J. Zhang, Y.C. Sun, Res. Astron. Astrophys. 14(10), 1221–1233
(2014). https://doi.org/10.1088/1674-4527/14/10/002 ([arXiv:1207.4541 [astro-ph.CO]].)
82. G.B. Zhao, M. Raveri, L. Pogosian, Y. Wang, R.G. Crittenden, W.J. Handley, W.J. Percival, F.
Beutler, J. Brinkmann, C.H. Chuang et al., Nat. Astron. 1(9), 627–632 (2017). https://doi.org/
10.1038/s41550-017-0216-z. ([arXiv:1701.08165 [astro-ph.CO]].)
83. E. Zucca, S. Bardelli, M. Bolzonella, G. Zamorani, O. Ilbert, L. Pozzetti, M. Mignoli, K. Kovac,
S. Lilly, L. Tresse et al., Astron. Astrophys. 508, 1217 (2009). https://doi.org/10.1051/0004-
6361/200912665. ([arXiv:0909.4674 [astro-ph.CO]].)
Chapter 16
The Cosmic Microwave Background
and H0 .

Pablo Lemos and Paul Shah

16.1 The Cosmic Microwave Background

16.1.1 Introduction

The Cosmic Microwave Background (CMB) is a fundamental component of modern

cosmology, providing a wealth of information about the early universe. It is a faint
radiation that permeates the entire cosmos, and it is essentially the ‘afterglow’ of
the Big Bang. The CMB is composed of photons that were released approximately
.380, 000 years after the Big Bang when the universe had cooled sufficiently for atoms

to form (at redshift .z ∼ 1100). Before that time, the universe was a hot, dense plasma
where photons were constantly scattered by charged particles in a process known as
Thomson scattering.
The CMB was first detected in 1965 by Arno Penzias and Robert Wilson at the
Bell Telephone Laboratories [45]. They were conducting experiments using a large
horn antenna designed for satellite communication. Penzias and Wilson noticed a
persistent background noise in their measurements, which they initially attributed to
various sources of interference. However, they encountered difficulties in eliminating
the noise, and after careful consideration, they realized that they were detecting a
uniform radiation coming from all directions in the sky. They had inadvertently
stumbled upon the CMB. The discovery of the CMB provided strong support for

P. Lemos (B)
Department of Physics and Astronomy, University of Sussex, Brighton BN1 9QH, UK
Mila, Montreal, QC H2S 3H1, Canada
e-mail: P.Lemos@sussex.ac.uk
P. Shah
Department of Physics and Astronomy, University College London, Gower Street, London WC1E
6BT, UK
e-mail: paul.shah.19@ucl.ac.uk

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 295
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_16
296 P. Lemos and P. Shah

the Big Bang theory; as opposed to the steady state theory, stating that the Universe
remains unchanged [31].
The CMB radiation is highly isotropic, with a temperature of approximately .2.7
Kelvin. However, there is a lot of cosmological information on the tiny variations
in the temperature of the CMB across the sky. These fluctuations were first detected
by the Cosmic Background Explorer (COBE) satellite, launched in 1989 [62, 63].
In recent years, there have also been remarkable advancements in the detection of
CMB polarization. The polarization of the CMB arises from the Thomson scattering
of photons by electrons at the time of last scattering, and it carries independent
information to the temperature fluctuations. The first detection of CMB polarization
was announced by the Degree Angular Scale Interferometer (DASI, [36]) in 2002,
followed by the Wilkinson Microwave Anisotropy Probe (WMAP, [30]).
However, the most significant breakthroughs in CMB temperature and polariza-
tion measurements came from the Planck satellite, launched by the European Space
Agency (ESA) in 2009 [48]. The Planck mission provided unprecedented CMB data,
unveiling the intricate anisotropy patterns of the CMB on both large and small angu-
lar scales. These observations have shed light on the conditions of the universe during
the era of recombination and have provided insights into cosmological parameters,
inflationary models, and the nature of dark matter and dark energy.
The origin of the modern Hubble constant tension can be reasonably dated
to the first Planck release of cosmological parameters [48], although disagree-
ments about the value of the Hubble constant date decades back [58, 68]. While
WMAP values for . H0 were compatible with other results at the time, the Planck
value of . H0 = 67.4 ± 0.5km s−1 Mpc−1 is discrepant with the S. H0 ES result of
−1
. H0 = 73.04 ± 1.04km s Mpc−1 [56] at .4.9σ (although Planck is consistent with
WMAP as we discuss below). As such, CMB values for . H0 merit considerable
scrutiny. Fortunately, while local measurements focus purely on . H0 , the CMB deliv-
ers a “package deal” of cosmological parameters which is highly dependent on the
cosmological model. The cosmological model that is most supported by present-day
observations is .ΛCDM, consisting of a spatially-flat cosmology with a cosmological
constant and three species of neutrinos with small but non-zero masses. We will make
it clear when we depart from this model. As usual .h ≡ H0 /(100 km s.−1 Mpc.−1 . Con-
straints in the parameters of this model may be compared to the values from other
data sets to see if other parameters are discrepant too. Also, as we explain below,
the CMB is not exclusively an early universe result, as the anisotropy maps carry the
imprint of late-universe physics. If therefore the Planck value for . H0 is considered
an “extraordinary claim”, we discuss the evidence for it in this chapter.

16.1.2 Why Are There Anisotropies in the CMB?

The physical origin of density perturbations lies in the primordial universe, specifi-
cally during the period of cosmic inflation. Cosmic inflation is a theoretical concept
that proposes a rapid exponential expansion of space in the early universe [24, 41,
16 The Cosmic Microwave Background and . H0 297

42]. This expansion was driven by a hypothetical scalar field known as the inflaton.
During inflation, quantum fluctuations in the inflaton field lead to variations in the
energy density across space. These quantum fluctuations are inherently random and
occur on extremely small scales. However, due to the exponential expansion of space
during inflation, these fluctuations are stretched to cosmological scales, providing
the seed for density perturbations. After inflation, the universe enters a phase known
as reheating, during which the inflaton field decays into other particles and reestab-
lishes the thermal equilibrium of the universe. As the universe cools, the density
perturbations seeded by inflation begin to grow under the influence of gravity.
The theory of inflation is widely accepted in the cosmology community, despite
a lack of direct observational evidence. This is because it solves some theoretical
issues, such as the horizon and flatness problems; but more importantly because of
its ability to explain the origin of the density perturbations that seed the formation
of cosmic structures. Inflationary models make specific predictions for the spectrum
of these initial fluctuations. They are expected to be approximately scale invariant
(if inflation lasts long enough), but small deviations from scale invariance can occur
due to the dynamics of the inflation field. The predicted density perturbations are in
excellent agreement with observations, including the temperature anisotropies in the
CMB and the distribution of galaxies in the universe. Inflation, therefore, provides a
natural mechanism for generating the observed structure in the universe.
The density perturbations are responsible for the observed variations in the CMB
temperature across the sky. During the epoch of recombination (which occurred at the
same temperature everywhere) when the universe became transparent to radiation,
the photons of the CMB decoupled from the baryonic matter. At this point, the
temperature of the CMB was determined by the density of matter in different regions
of the universe. Regions with higher matter density experienced stronger gravitational
pull, causing photons to lose energy as they climbed out of the gravitational potential
wells. As a result, these regions appear slightly cooler in the CMB. Conversely,
regions with lower matter density experienced weaker gravitational pull, allowing
photons to retain more energy and appear slightly warmer in the CMB. In addition,
photons are Doppler-shifted by the velocity perturbations in the plasma.
During their long journey to our detectors, the CMB photons are perturbed by
effects known as secondary anisotropies. In addition, foreground emission of radi-
ation at the same frequencies as the CMB occurs both in our own galaxy and from
extra-galactic sources. For the purposes of cosmology, these foregrounds must be
subtracted, but we note they also carry interesting astrophysical information about
the properties of dust, magnetic fields and the inter-galactic medium.

16.1.3 The Temperature of the CMB

The average flux intensity of the CMB observed today (that is, the monopole in
the power spectrum), shown in Fig. 16.1 is a near-perfect black body spectrum (‘the
most perfect black body ever measured in nature’, [72]), demonstrating the plasma
298 P. Lemos and P. Shah

Fig. 16.1 The spectrum of

the CMB measured by
FIRAS. The error bars
shown are .±400σ . The solid
line is a .2.728 K blackbody.
Image credit: [61]

of baryons and photons were in thermal equilibrium on the surface of last scattering
(and therefore presumably at earlier times). The temperature was initially measured
in balloon experiments, but was more accurately measured by the FIRAS instrument
aboard the satellite COBE. The combination of data gives.T = 2.72548 ± 0.00057 K
[23]. Because the spectrum is thermal, the energy density of the CMB is determined
solely by the temperature as .U = 4c σ T 4 . This results in the physical density of
photons today as .ργ = 4.65 × 10−31 kg m.−3 , or in units of the critical density .Ωγ ≡
ργ /ρc = 5.4 × 10−5 , where the critical density represents the average mass or energy
per unit volume required to achieve a flat universe. It is worth pointing out that spectral
distortions in the CMB have been proposed as a powerful cosmological probe [35],
although more accurate measurements of the CMB spectrum would be required.
We use . Neff to refer to the effective number count of neutrinos or other relativis-
tic particles. Assuming there is no substantial addition to the photon density from
the decay of other particles or fields, knowledge of the CMB temperature today
determines the energy density of radiation at earlier times as

T0
ρ (z) = (1 + 0.227Neff )(1 + z)4 (
. r )4 × 4.65 × 10−31 kg m−3 , (16.1)
2.72548
where .T0 is the present-day CMB temperature.
In the standard model of particle physics, when .T ≪ 1 Mev, . Neff = 3.043 (it is
not an integer as the spectrum of neutrinos is not precisely a black-body; the energy
density of neutrinos is altered by a contribution from the annihilation of electrons and
positrons after their decoupling from the primordial plasma). However, in extensions
to the standard model . Neff is often treated as a free parameter (Fig. 16.2).
16 The Cosmic Microwave Background and . H0 299

Fig. 16.2 The binned Planck power spectrum for TT and residuals to the best fit .CDM model. The
vertical line delineates the different methodologies used to resolve the power spectrum at .l < 30.
Figure from [51]

16.1.4 The Power Spectrum and Its Features

To analyse the anisotropies in the CMB temperature and polarization, the map of the
fluctuations in the sky can be decomposed into a sum of spherical harmonics
ʃ
a
. m = ΔT (θ, φ) Y m (θ, φ)dΩ , (16.2)

where . is limited by the resolution of the instrument and .m runs over .(− , ). The
CMB sky represents a single sampling of the random distribution of primordial fluc-
tuations. If we want to know the “true” values of cosmological parameters, we should
average over a large ensemble of realisations—but we have only one! An approxi-
mation to ensemble averaging is obtained by averaging over the .2 + 1 harmonics
in each mode, which is likely to give a reliable answer at high-. but is of limited
accuracy at low-. , a floor referred to as cosmic variance.
The power spectra are then defined as

C X X = ⟨|a XmX |2 ⟩m ,
. (16.3)

where . X can reperesent the temperature (.T ), E-mode polarization (. E) or B-mode

polarization (. B). In the case of temperature, for convenience of display, the power
spectra are often presented as the dimensionless quantity
300 P. Lemos and P. Shah

( + 1) T T
.DT T = C . (16.4)
2π
This scaling factor is commonly used to highlight the characteristic angular scales
of the CMB fluctuations. At present, there is no evidence of non-Gaussanity in the
primordial fluctuations so the .C (or equivalently the 2-point correlation function)
capture all available information on primary anisotropies (although there are claims
of structure in low-. .a m —see below). However, gravitational lensing induces non-
zero higher-order correlation function values that can be measured separately (also
see below).

16.1.4.1 Primary Anisotropies and Length Scales

Primary anisotropies in the cosmic microwave background (CMB) radiation pro-

vide a wealth of information about the early universe and the processes that shaped
its evolution. These anisotropies refer to the small fluctuations in temperature and
polarization observed across the CMB sky. There are three main features of primary
anisotropies in the CMB:

1. The Large-Scale Plateau: On large angular scales, typically above a few degrees,
the primary anisotropies in the CMB predominantly exhibit a feature known as the
large-scale plateau. This plateau corresponds to scales that were not significantly
affected by pre-recombination physics and provides valuable information about
the primordial Universe.
2. Acoustic Oscillations: On smaller angular scales, below a few degrees, an addi-
tional feature becomes prominent in the primary anisotropies: the acoustic oscilla-
tions. These oscillations arise from the competition between gravitational attrac-
tion and radiation pressure in the early universe. The density and velocity pertur-
bations in the primordial plasma give rise to sound waves that propagate through
the medium, and these waves lead to periodic oscillations in the CMB power
spectrum, with peaks and troughs that correspond to the specific length scale
associated with the sound horizon at the time of photon decoupling (Fig. 16.3).
The position of the peaks provides a measurement of the extrema of these oscil-
lations at last scattering, and in particular of the sound horizon at radiation drag
.r d , which measures the distance sounds waves can travel before recombination
([71]):
ʃ ∞
cs (z)
.r d = z, (16.5)
z ⋆ H (z)

where .z ⋆ is the redshift at photon decoupling and .cs (z) is the sound speed in the
photon-baryon fluid ([32]) which by the standard thermodynamics of adiabatic
perturbations is
16 The Cosmic Microwave Background and . H0 301
[ ]
1 3 ρb (z) −1
.cs (z) = 1+ .
2
(16.6)
3 4 ργ (z)

The sound horizon serves as a standard ruler, and the position of the acoustic peaks
captures the history of . H (z) together with matter and radiation densities prior to
recombination. Additionally, the heights of the peaks represent the competition
between pressure (a function of the physical baryon density .Ωb h 2 ) and gravity (a
function of the physical matter density .Ωm h 2 ) as we show in Fig. 16.4 below.
3. Silk Damping Tail: On even smaller angular scales, the anisotropies in the CMB
exhibit a phenomenon known as the Silk damping tail. This effect arises from the
diffusion of photons caused by scattering off free electrons just prior to recombi-
nation. The diffusion process tends to erase small-scale temperature fluctuations
in the CMB, resulting in a damping of the power spectrum on scales below the
Silk scale:
√ √
λD ∼
. N λc ∼ η⋆ λc , (16.7)

where . N is the number of steps in the random walk, .λc is the mean-free path to
Thomson scattering, and.η⋆ is the conformal time at recombination. The Silk scale
corresponds to an angular scale around . ∼ 1000 and is determined by baryon
density and . H (z) just prior to recombination (e.g. [34]).

16.1.4.2 Secondary Anisotropies

Secondary anisotropies in the CMB refer to additional fluctuations observed in the

CMB sky, which arise from various astrophysical and cosmological phenomena. The
main sources of secondary CMB anisotropies are:

1. integrated Sachs-Wolfe (ISW) effect: The ISW effect [57] is caused by the
gravitational potential wells that photons traverse as they travel through evolving
large-scale structures in the universe. If the potentials change photons pass through
them, their energy is altered, leading to temperature fluctuations in the CMB. In
particular, the Late ISW effect provides constraints on the expansion rate of the
universe during the accelerated expansion era at .z < 0.6, albeit as it primarily
affects large scales this constraint is limited by cosmic variance.
2. CMB Lensing: The light from the CMB is deflected by foreground matter inho-
mogeneities, with the typical deflection angle being 3’ [40]. This is in the domain
of weak lensing, where just one image source is seen but the source is distorted
and magnified or de-magnified. The amount of lensing depends on the degree of
inhomogeneity (best captured by the combination .σ8 Ωm0.25 ) and a combination of
angular diameter distances between the observer, deflector and source (sometimes
called the lensing efficiency). For the CMB, the lensing efficiency peaks between
.1 < z < 2 meaning it is sensitive to structure just beyond the reach of present-day
galaxy weak lensing studies.
302 P. Lemos and P. Shah

Lensing has two linked effects: firstly, it smooths the observed power spectra,
most noticeably for .l > 800 multipoles. Secondly, observations of the four-point
correlation function of fluctuations can be used to reconstruct the (projected)
foreground matter distribution. As a consistency check we may compare (1) the
theory prediction of lensing given the amplitude of the fluctuations in the CMB
and growth of structure (2) the observed smoothing of the power spectrum, and
(3) the observed power spectrum of the lensing reconstruction.
3. Reionization: During the early stages of the universe, neutral hydrogen became
ionized due to the radiation from the first stars and galaxies, a process known as
reionization. This reionization process leaves characteristic imprints on the CMB
through the scattering and absorption of CMB photons. At .z ∼ 10 the formation
of the first stars ionizes the Universe. The presence of free electrons from this
point onward means that a fraction .(1 − e−τ ) of CMB photons undergo Thomson
scattering after reionization, where.τ is the optical depth to reionization defined as
the line-of-sight opacity of the CMB radiation with respect to Thomson scattering
with free electrons:
ʃ χre
.τ = σT n e dχ , (16.8)
0

where .χre is the comoving distance to reionization, .σT is the Thomson scattering
cross-section and.n e is the number density of free electrons. As a consequence, the
power spectrum on scales that entered the horizon before reionization is damped
by a factor of .e−2τ .
4. Sunyaev-Zel’dovich (SZ) effect: The SZ effect occurs when CMB photons inter-
act with the hot, ionized gas in galaxy clusters through inverse Compton scatter-
ing [65, 66, 73]. There are two types of SZ effect: thermal SZ and kinematic SZ.
The thermal SZ effect arises from the Compton scattering of CMB photons with
the electrons in the cluster gas, boosting the energy of the photons and distorting
their spectrum to higher frequencies. The kinematic SZ effect is subdominant
and arises from the bulk motion of the clusters. The thermal SZ effect can be
detected by its frequency dependence, while the kinematic SZ effect is frequency
independent.

16.1.5 Polarization

Temperature anisotropies are not the only source of information from CMB radia-
tion. CMB polarization provides valuable complementary information to temperature
anisotropies, offering further insights into the early Universe. The polarization pat-
tern arises due to scalar and tensor perturbations to the metric. Scalar perturbations
generate a quadrupolar component of the radiation field, leading to linear polariza-
tion through Thomson scattering during recombination. On the other hand, tensor
perturbations, originating from gravitational waves, also contribute to the polariza-
16 The Cosmic Microwave Background and . H0 303

Fig. 16.3 Planck 2018 temperature-cross E-mode polarization power spectrum (TE, left), and E-
mode autocorrelation power-spectrum (EE, right). The vertical axis on the left plot shows .D ≡
( + 1)C /(2π ). The red points are the observed data and the blue line shows the best fit .CDM
cosmology. Scales change from logarithmic to linear at . = 30. The lower panels show residuals
with respect to the best fits. Image credit: [50]

tion. The polarization can be decomposed into four Stokes parameters: intensity
(I), linear polarization (Q, U), and circular polarization (V). The intensity parame-
ter represents the total power, while the linear polarization parameters capture the
orientation and amplitude of the polarization. Circular polarization is absent in the
CMB due to Thomson scattering properties (for a more detailed discussion about
CMB polarization see [13, 33]).
The linear polarization field can be further classified into two types: E-modes and
B-modes. E-modes represent the curl-free component of the polarization field and
are thus primarily generated by scalar perturbations on the surface of last scattering.
B-modes arise from a combination of tensor perturbations (primordial gravitational
waves) and gravitational lensing of E-modes. B-modes are the curl component of the
polarization field and are particularly interesting as they provide a unique window
into the presence of primordial gravitational waves. Detection of B-mode polar-
ization, after accounting for the effects of gravitational lensing, would be a direct
confirmation of the existence of primordial gravitational waves, which are predicted
by some inflationary models. The Planck measurements of the E-mode power spec-
trum (.C E E ) and its cross-correlation with temperature (.C T E ) is shown in Fig. 16.3.
Cosmological parameters determined from the Planck polarization and temperature
power spectrums are consistent with each other (see Fig. 5 of [51]).

16.1.6 Foreground Removal

Foregrounds introduced by our own Milky Way, extragalactic sources and instru-
ment effects are primarily identified by a comparison of multiple frequency channels.
The CMB has a near-perfect blackbody spectrum, and therefore a precisely defined
“colour”. Foregrounds generated by non-thermal effects (such as synchrotron emis-
sion from dust) or thermal effects at different temperatures (such as the thermal
304 P. Lemos and P. Shah

Sunyaev-Zeldovich effect) exhibit different colours, enabling their individual iden-

tification.
Foregrounds are removed by a combination of sky masks, frequency modelling
and spectral templates .C ,fore (for unresolved foregrounds) which represent the con-
tribution of foreground emissions at different multipole moments. These templates
capture the spectral characteristics of foreground sources such as synchrotron emis-
sion from dust and the thermal Sunyaev-Zeldovich effect. The nuisance parameters
associated with these templates are marginalized over during the fitting process for
cosmological parameters (see [49] for a description of the process). Data from high-
resolution experiments like ACT and SPT are particularly useful to constrain the
nuisance parameters, and the final cleaned spectra can be compared across frequen-
cies as a consistency check. The details are not important for this review, but the key
point is: could the foreground model bias the inference of . H0 ?
The evidence is they do not. The foreground models provide an acceptable fit
across frequency ranges as measured by their .χ 2 , justifying the claim that they cor-
rectly model the physics [48]. Correlations between foreground nuisance parameters
and cosmological parameters are low [49]. Features such as the oscillating resid-
uals in the region .1000 < l < 1500 (see Fig. 16.2) and the low power for .l < 30
which play a role in driving . H0 for Planck lower than WMAP are present in mul-
tiple frequency channels; this suggests they are real features of the spectrum and
not foregrounds. The fit to foreground-cleaned temperature spectrum is consistent
with the polarisation spectrum, which depends differently on the foregrounds and
in particular, is not expected to be dominated by them at high-. [67]. Likelihoods
that use different foreground subtraction methods produce cosmological parameters
consistent with Planck [20, 64, 70].

16.1.7 The Planck Cosmology

The CMB is the only cosmological probe that can simultaneously

{ constrain all six}
parameters in the standard .CDM model of cosmology . Ωb h 2 , Ωc h 2 , H0 , τ, As , n s .
Note that different combinations of these six parameters are commonly used to
parameterize .CDM. The values measured by the Planck 2018 TT+TE+EE+lowE
likelihood are shown in Table 16.1, and the effect of varying them in the CMB power
temperature power spectrum is shown in Fig. 16.4.

16.1.8 How Does the CMB Determine . H0 ?

16.1.8.1 The Angular Size of the Sound Horizon

As we discussed in the section on primary anisotropies, the sound horizon is a

characteristic length imprinted on the CMB power spectrum, visible as the positions
16 The Cosmic Microwave Background and . H0 305

Table 16.1 Best fit values and .68% limits for the constraints from Planck 2018 TT+TE+EE+lowE
[50]
Parameter Symbol Best fit value and .68% limits
.Ωc h ± 0.0014
Cold dark matter density 2 .0.1202

.Ωb h ± 0.00015
Baryon density 2 .0.02236

Hubble parameter .h .0.6727 ± 0.0060

+0.0070
Optical depth to reionization .τ .0.0544−0.0081
( 10 )
Scalar spectrum amplitude .log 10 As .3.045 ± 0.016

Scalar spectral index .n s .0.9649 ± 0.0044

Fig. 16.4 Variations in the CMB temperature power spectrum, as different cosmological parameters
are changed, one at a time, while keeping the rest of the parameters set to the Planck best-fit values
shown in Table 16.1

of the multiple peaks and troughs (strictly, we should say the complete waveform of
the spectrum). This is observed as a ratio .θ⋆ = rs /d A (z ⋆ ) where .d A (z ⋆ ) is the angular
diameter distance to the surface of last scattering. It is the most accurately measured
parameter by Planck, and .100θ⋆ = 1.0411 ± 0.0003 [51].
The theory expectation can then be simply expressed as
ʃ∞ cs (z)
z H (z)
dz
θ = ʃ z⋆⋆
. ⋆
1
. (16.9)
0 H (z)
dz

At this stage, very few assumptions about cosmology have been made. The above
assumes only that the sound horizon has resulted from the propagation of adiabatic
pressure waves in plasma from initial seed densities, that recombination is instanta-
neous and photons propagate freely to us from the surface of last scattering (these
last two are straightforwardly generalized).
306 P. Lemos and P. Shah

It is convenient to define dimensionless physical densities .ωi = Ωi h 2 . In .ΛCDM

cosmology, we may accurately approximate the above as
ʃ∞ cs (z)
z ⋆ ((1+z)3 + ωωmr (1+z)4 )1/2 dz
.θ⋆ = ʃ z⋆ 1
. (16.10)
0 ((1+z)3 + ω )1/2 dz
ωm

The above looks rather complicated, so let us unpack it! The denominator is relatively
simple, with the two terms representing the matter density and dark energy density
respectively. The radiation density is mostly negligible (although it is still relevant
for the first dex of expansion after recombination).
In the numerator, we have assumed the universe is comprised only of matter
and radiation. Curvature and dark energy are physically negligible pre-combination.
As discussed, the radiation density .ωr (z, TCMB ) is determined by the observed CMB
temperature. The sound speed .cs (z, ωb ) is determined by the physical baryon density
.ωb and standard thermodynamics. The redshift of the surface of last scattering has

only a weak cosmology dependence: it is approximately the freeze-out redshift when

the scattering rate of photons and baryons matches the Hubble expansion rate. More
baryons increase the scattering rate but also make recombination happen earlier,
and the factors roughly cancel (the same argument is made for the Hubble rate).
It is thus mainly a function of primordial Helium abundance .Y P (which may be
either measured or calculated from the baryon density, nucleosynthesis rates, and
the Hubble expansion rate at that time).
As discussed above in the section on primary anisotropies, .Ωm and .ωb are deter-
mined by the relative heights of the peaks of the power spectrum, and by the Silk
damping scale at high-. . The only free parameter remaining to be adjusted is .h.
The situation is a little more complicated if we drop the assumption of spatial
flatness. In this case, we now have two free parameters, .ΩΛ and .h. However, .ΩΛ
is constrained by the secondary anisotropies of the CMB: the power spectrum is
smoothed on a characteristic scale by gravitational lensing at .z ∼ 1 − 2, and power
in .l < 30 modes is enhanced by the late-ISW effect (which depends on the expansion
history for .z < 0.6). We show in Fig. 16.5 a schematic of the redshift range of the
sensitivity of the CMB power spectra to . H (z) : each window function effectively
determines a ruler whose theoretically-determined length may be compared to the
observed angular size.

16.1.9 Consistency of . H0 Derived from the CMB

Cosmological parameters influence CMB measurements in multiple ways. It is there-

fore useful, as a test of .ΛCDM, to check the consistency of CMB-derived parameters
both internally and with non-CMB data sets.
16 The Cosmic Microwave Background and . H0 307

Fig. 16.5 A schematic of the. H (z) redshift windows to which the CMB power spectra are sensitive
to. The y-axis scale for each component is arbitrary and not shown. The growth of the sound horizon
(blue) takes place primarily in the last dex of expansion before recombination : solutions to the . H0 -
tension that change the sound horizon must therefore modify . H (z) at this time. The visibility
function (orange) represents the depth of the surface of last scattering, manifest in the Silk damping
scale and polarization spectra. The Early ISW effect (green) modifies the power spectrum near its
first peak due to the effect on gravitational potentials of the residual radiation density in the first
dex of expansion after .z eq , whereas the Late ISW effect (red) modifies the low-l power spectrum
close to the onset of accelerated expansion at .z ∼ 0.6. Finally lensing (purple) peaks at a redshift
corresponding to a distance that is midway (in comoving coordinates) between us and the surface
of last scattering

We should however remember that we are observing one sample of a Gaussian

random field, limited by cosmic variance. If we make multiple consistency tests, it
would be surprising if there were not some statistical excursions from the underlying
truth.

16.1.9.1 Consistency with Other CMB Experiments

The consistency among various cosmic microwave background (CMB) experiments

provides valuable insights into the fundamental properties of the universe, particu-
larly the Hubble constant. WMAP determined . H0 = 70.0 ± 2.2km s−1 Mpc−1 [10,
29] purely from the power spectrum of temperature fluctuations up to .l ∼ 800. This
is consistent with the equivalent Planck measurement over the same multipole range
(see Fig. 21 of [51]). Planck shifts . H0 lower due to its higher multipole range. At
least part of the reason why WMAP has a higher . H0 than Planck appears to be
due to low power in the .10 < l < 30 range (as also observed by Planck—more on
this below). Adding BAO information to WMAP brings the value closer to Planck
308 P. Lemos and P. Shah

(. H0 = 68.76 ± 0.84km s−1 Mpc−1 ). A remarkable result is a consistently low value

of . H0 from Planck lensing and BAO alone (. H0 = 68.76 ± 0.84km s−1 Mpc−1 [14]).
Two notable recent experiments are the Atacama Cosmology Telescope (ACT),
and the South Pole Telescope (SPT) which have played a crucial role in advanc-
ing our understanding of the CMB. ACT’s combined analysis of temperature and
polarization data has led to a determination of the Hubble constant, with a value
of . H0 = 67.9 ± 1.5km s−1 Mpc−1 (e.g. [4, 43]). Interestingly, initial polarization
analyses from the SPT yielded a value of . H0 = 71.3 ± 2.1km s−1 Mpc−1 [27], in
tension with Planck and ACT, and closer to the value obtained using direct mea-
surements. However, the more recent SPT-3G measurements including temperature
anisotropies recovered a low value . H0 = 68.3 ± 1.5km s−1 Mpc−1 [9], consistent
with other CMB experiments. Furthermore, possible issues in the analysis of [27]
were raised in [50], as well as in Chap. 4 of [37].
Therefore, there is remarkable consistency observed between . H0 measurements
from different CMB experiments, as shown in Fig. 16.6, despite small tensions in
other parameters [26]. This convergence of results from independent experiments
provides strong evidence for the low . H0 from the CMB and rules out the option of
explaining the tension as a systematic effect in one of the experiments.

16.1.9.2 Internal Consistency of Planck Data

There is a moderate tension between low (.2 < l < 800) and high (.801 < l < 2508)
multipole values for . H0 that has been noted by many authors (for example, see
[2]). The overlap between the two posteriors is consistent with the independent
polarization and lensing power spectra, and also by the addition of BAO to WMAP.
At least part of the difference appears to be caused by a dip in power for the multipole
range .10 < l < 30 (visible in Fig. 16.2 above). This seems to be a real feature of the
CMB sky.
The power spectrum of lensing reconstruction and amplitude of fluctuations are
consistent with each other, both internally in Planck (see Fig. 3 of [51]) and between
Planck and ACT [54]. However, the smoothing in the Planck power spectrum is not
consistent with its amplitude, captured by the parameter . A L = 1.180 ± 0.065 [51]
(where . A L = 1 indicates perfect consistency). The connection with . H0 is that allow-
ing . A L as an additional nuisance parameter in the .ΛCDM solution both pulls . H0
to higher values and broadens the error bands, so the tension with S. H0 ES measure-
ments is lowered, although it remains significant at .∼ 3.8σ (see Fig. 16.7). Despite
its definition, . A L is probably not due to lensing. It may be in part due to the dip
in power between .10 < l < 30: removing the .l < 30 part of the spectrum restores
consistency with. A L = 1 (and values of . H0 consistent with the overall Planck result).
It may also be due to unknown systematics or residual foregrounds in Planck. There
is no evidence for . A L > 1 from either ACT (. A L = 1.01 ± 0.11 [4]) or SPT-3G
(. A L = 0.87 ± 0.11 [9]). In addition, the likelihood for Planck from the CamSpec
analysis [22] including a larger area of the sky is consistent with . A L = 1 across the
full multipole range, implying it is not a real-sky effect.
16 The Cosmic Microwave Background and . H0 309

Fig. 16.6 Comparison of Hubble constant constraints between different CMB experiments, and
with the latest results from S. H0 ES [55]
310 P. Lemos and P. Shah

Fig. 16.7 Comparison of the Planck 2018 .68% and .95% marginalized posterior distributions for
different parameters, including and not including . A L . Note that the significance of this result
decreases significantly in the CamSpec analyses [22]. Image credit: [50]

Other anomalies have been noted in Planck data, for example, a lack of large
angle correlations (such as would be expected from the measured power spectrum),
alignment of the quadrupole and octopole, and angular variations of the spectral
power and . H0 across patches of the sky (see for example [5]). While some are
simulated to have less than .0.1% chance of occurrence in random realisations of the
CMB sky, the “look-elsewhere” effect1 cautions us to avoid simple interpretations
of these p-values. Nevertheless, such claims are at least at the level of curiosity; but
at present there is no apparent connection with the Hubble tension or any physical
process.

16.1.9.3 Consistency with Other Astrophysical Data

As we described above, the conversion of CMB spectral data into constraints on

cosmological parameters is highly model dependent. Since . H0 from the CMB in
.ΛCDM is discrepant with local values, other discrepancies would add evidence to
the argument .ΛCDM is not the correct model and may provide hints about how to
change it, especially where those parameters are correlated to . H0 (see Fig. 16.8 or
Fig. 5 of [51]).
The relative heights and positions of the peaks of the CMB spectrum derive from
the baryon density .ρb , the matter density .ρm and the pre-recombination expansion
history. Two key events pre-combination are nucleosynthesis (BBN) at .z BBN ∼ 1010
and matter-radiation equality at.z eq ∼ 3400. Additionally, . H (z) at the time of recom-
bination determines the shape of the Silk damping tail.
The primordial abundance of helium is sensitive to the expansion rate . H (z BBN )
as the availability of neutrons to form it depends on the time between the decou-
pling of weak interactions that keep protons and neutrons in equilibrium, and the
time when deuterium (D) can form. The abundance of D is very sensitive to the

1 We refer the reader to https://en.wikipedia.org/wiki/Six_nines_in_pi and https://xkcd.com/882/

for entertaining examples.
16 The Cosmic Microwave Background and . H0 311

baryon density: as D is rapidly processed into He, a higher baryon density implies
a faster interaction rate and hence less D. Primordial element abundance obser-
vations are extrapolative by nature, but the Planck values for .(Neff , ρb ) = (2.89 ±
0.37, 0.02224 ± 0.00022) (Tables 4 and 5 of P18) are consistent with observations of
.YP = 4n He /n b = 0.2449 ± 0.004 [8] and .YD = n D /n H × 10 = 2.52 ± 0.03 [15].
5

There is also excellent agreement with calculations using experimentally determined

D-burning rates [47], and Planck is therefore compatible with the standard model of
particle physics which predicts . Neff ≃ 3.043.
It is possible to go further and drop the CMB spectrum (but not its tem-
perature) by combining .ρb constraints from BBN with the standard ruler pro-
vided by baryon acoustic oscillations; these are an imprint of the sound horizon
on the late universe and effectively determine .Ωm [3]. In this case, one obtains
. H0 = 68.1 ± 1.1km s
−1
Mpc−1 ([17], see also [59]) and .ΩM = 0.300 ± 0.018, fully
compatible with Planck within .ΛCDM.
The power spectrum . P(k) of matter density fluctuations has a maximum at a scale
.k ∼ k eq which is the scale of modes entering the horizon at . z eq (the shape near the

maximum also depends weakly on .ρb as dark matter fluctuations are suppressed by
baryons). Faster expansion in the pre-recombination universe lowers the horizon and
increases .keq . Therefore, the shape of the power spectrum as measured from galaxy
surveys may be used to check compatibility with the Planck . H0 in .ΛCDM. The
evidence at present is in favour of compatibility with . H0 = 69.6 ± 1.2km s−1 Mpc−1
([46], see also the chapter in this book).
The expansion history of the universe from .z < 1.2 is strongly constrained by
Type Ia SN (the furthest observations reach to .z ∼ 2). Type Ia SN do vary in intrinsic
brightness by a factor of .×2, but the majority of this fluctuation is attributable to
the colour and duration of the explosion. The residual fluctuations may be as low as
.0.07 mag [11] and the evidence points to them being a homogeneous population

for the purposes of cosmology. While they cannot by themselves determine . H0

as this is degenerate with their fiducial absolute magnitude, the Hubble diagram
constructed from the Pantheon+ sample [60] is fully compatible with Planck with
.Ωm = 0.338 ± 0.018 in .ΛCDM [12]. Is it then is no surprise that calibrating the SN
Ia magnitudes with BAO in turn produces a . H0 also fully compatible with Planck
[44].
The values of . S8 = σ8 (Ωm /0.3)0.5 derived from Planck (via .ΛCDM assumptions
on structure growth) are high (at the level of .∼ 2.5σ ) compared to observations of
galaxy lensing [1, 28, 39]. The relevance of this for the Hubble tension is unclear, and
it has been proposed a resolution lies in correcting theory calculations of structure
growth in the quasi-linear regime [7]. However, we note if one were to attempt
to resolve the Hubble tension by speeding up the expansion of the universe pre-
recombination, one would need to compensate for the slower structure growth by
adding more matter, increasing the . S8 discrepancy and potentially creating a new
discrepancy with SN Ia.
In summary, within the framework of .ΛCDM and some simple one-parameter
extensions, the only CMB-derived cosmological parameters that are inconsistent
with other astrophysical datasets are . H0 , and to a lesser extent .σ8 . In particular,
312 P. Lemos and P. Shah

matter, baryon, curvature densities and number of particle species are compatible
with other observations.

16.1.10 The CMB in .ΛCDM Extensions

Extensions to the .ΛCDM model often involve modifications to the standard physics
of the early universe or the presence and nature of dark energy (either pre- or post-
recombination). These modifications can leave imprints on the CMB, allowing for
tests of these alternative scenarios. For instance, variations in the primordial power
spectrum, generated during cosmic inflation, can lead to specific patterns in the
CMB temperature and polarization anisotropies. Planck data has been instrumental
in placing constraints on such modifications, providing valuable insights into the
inflationary paradigm and the physics of the early universe [6], and finding strong
support for a featureless primordial power spectrum.
Another area of interest is the nature of dark energy itself. While the .ΛCDM
model assumes a cosmological constant .Λ to account for the observed accelerated
expansion of the universe, alternative theories propose dynamical dark energy models
or modifications to general relativity. These models can result in deviations from the
standard predictions for the CMB power spectrum and other statistical properties.
One common approach to parameterize dark energy deviations from .ΛCDM is with
a time-dependent equation of state .w(a) = w0 + (1 − a)wa (.ΛCDM assumes .w0 =
−1 and.wa = 0). Interestingly for the. H0 tension, Planck measurements are consistent
with a value .w0 < −1, which also raises the value of . H0 . This value falls under the
regime known as “phantom” dark energy [69], as they require a second scalar field, or
some other complex theoretical explanation (e.g. [16, 18]). However, the combination
of SN Ia, Planck and BAO (which may be expected to give the best constraints on
dark energy) does not favour phantom dark energy [12].
Another interesting extension to .ΛCDM is spatial curvature (.Ω K /= 0), which
+0.018
raised interest due to the Planck measurement .Ω K = −0.044−0.015 , which is over
.2σ away from zero [19, 25]. However, as discussed in [21], this value is strongly
correlated with the . A L measurement from Planck discussed above, and is equally
explained by statistical fluctuations in the temperature power spectra in the multi-
pole range .800 < < 1600. Similarly to . A L , the CamSpec likelihood [22] strongly
reduces the significance of this detection. Furthermore, a closed Universe leads to
even lower values of . H0 , further exacerbating the tension.
Finally, it is possible to explore the changes to∑neutrino physics. For example,
it is possible to vary the sum of neutrino masses . m ν or the effective number of
neutrino species . Ne f f introduced above. None of these changes, however, is enough
to alleviate the tension.
A more radical idea is to increase the size of the sound horizon by introducing
“early” dark energy (EDE) to the pre-recombination universe (possibly connected
with the onset of matter domination). EDE must be present at the level of .∼ 10%
of the energy density during the last dex of expansion prior to the surface of last
16 The Cosmic Microwave Background and . H0 313

Fig. 16.8 Comparison of Hubble constant constraints between different one parameter extensions
to the .ΛC D M model, using Planck TTTEEE + lowE in blue, and adding BAO in green. In red, are
the .68% and .95% confidence levels for . H0 . The black, dotted lines show the .ΛCDM value for each
parameter

scattering, as that is when most of the growth in the size of the sound horizon happens
(see Fig. 16.5). Silk damping and CMB polarization on small scales constrain its
presence at .z = 1100. In general, whilst EDE models cannot resolve the Hubble
tension, they can alleviate it to the level of .∼ 2σ [52]. EDE models also make
specific predictions for the .l > 1000 temperature and polarization spectrums, and so
future high-resolution CMB data will determine whether they are viable or not. EDE
models may seem somewhat fine-tuned. They do not predict a high late universe . H0
but they may explain it (see [53] for a review). An additional problem is the boost
to pre-recombination expansion requires more dark matter to reproduce the heights
of the peaks of the power spectrum (as dark energy suppresses growth of structure),
increasing the . S8 tension with galaxy weak lensing results.
Figure 16.8 shows the effect of the most popular one-parameter.ΛCDM extensions
in the Planck estimates of the Hubble parameter. As we see, there is no unambiguous
evidence from CMB data alone of a departure from Flat .ΛCDM.
Generally, while it is true that one possible solution to the . H0 tensions is a mod-
ification to the .ΛC D M cosmological model, it is very challenging to envision new
314 P. Lemos and P. Shah

physics that solves the tension. Indeed, as described in [38] to solve the tension with
new physics, it needs to lower the sound horizon by approximately .9% (to about
.135 Mpc) while preserving the structure of the temperature and polarization power
spectra described in this chapter; and also preserving the consistency between BBN
and observed abundances of light elements.

16.1.11 Conclusions and Future Prospects

The cosmic microwave background (CMB) observations within the framework of

the .ΛCDM model have demonstrated remarkable consistency with a wide range
of non-CMB astrophysical data at varying redshifts. The cosmological parameters
derived from Planck are consistent across frequencies and likelihood methodologies,
and with WMAP, ACT and SPT.
However, there are a few remaining tensions that warrant further investigation.
Notably, weak lensing measurements of the parameter .σ8 , which characterizes the
amplitude of matter fluctuations on large scales; and, more importantly for this
chapter, the local measurement of the Hubble constant (. H0 ) from the nearby universe
appear to deviate from the predictions of the CMB. An interesting discrepancy that
could be connected to the Hubble constant is the baseline Planck lensing amplitude
parameter. A L . However results elsewhere are consistent with. A L = 1 suggesting that
the observed .∼ 2σ Planck discrepancy may be a fluke. Further investigations and
independent corroborating measurements are necessary to better understand these
internal inconsistencies.
Proposed solutions that aim to resolve the Hubble constant discrepancy by accel-
erating the growth of the sound horizon during the pre-recombination era have shown
the potential to exacerbate the tension in the .σ8 parameter. These proposals, which
introduce variations in the early universe’s expansion rate, could have broader impli-
cations for our understanding of cosmological structures and the growth of matter
fluctuations. The challenge lies in reconciling these competing tensions and finding
a coherent theoretical framework that can simultaneously explain both the Hubble
constant and .σ8 measurements. A comprehensive analysis of late-time evolution,
combining observations from Type Ia supernovae and baryon acoustic oscillations
(BAOs), offers valuable insights into the cosmic expansion history and provides
constraints on different cosmological models.
Future CMB experiments hold immense promise for shedding light on unre-
solved questions and advancing our understanding of the universe. Next-generation
missions, such as the Simons Observatory, CMB-S4, and the LiteBIRD mission
(planned for launch in 2028) are poised to deliver groundbreaking observations.
These experiments will allow for higher precision measurements of the CMB tem-
perature and polarization anisotropies. This will allow tests of early dark energy
models and explore potential deviations from the standard .ΛCDM cosmology. The
combination of these next-generation CMB experiments with independent probes,
such as large-scale structure surveys and supernova observations, will provide a
16 The Cosmic Microwave Background and . H0 315

multi-faceted approach to addressing the outstanding tensions and advancing our

understanding of the universe’s evolution.

Acknowledgements We thank Antony Lewis for comments on an earlier version of this chapter.
PL acknowledges support from the Simons Foundation.

References

1. T.M.C. Abbott et al., Dark energy survey year 3 results: cosmological constraints from galaxy
clustering and weak lensing. Phys. Rev. D 105(2), 023520 (2022). https://doi.org/10.1103/
PhysRevD.105.023520. (arXiv:2105.13549 [astro-ph.CO])
2. G.E. Addison, Y. Huang, D.J. Watts, et al.: Quantifying Discordance in the 2015 Planck
CMB Spectrum. ApJ 818(2), 132 (2016). https://doi.org/10.3847/0004-637x/818/2/132.
arXiv:1511.00055
3. G.E. Addison, D.J. Watts, C.L. Bennett et al., Elucidating .ɅCDM: impact of baryon acoustic
oscillation measurements on the hubble constant discrepancy. ApJ 853(2), 119 (2018). https://
doi.org/10.3847/1538-4357/aaa1ed. arXiv:1707.06547
4. S. Aiola, E. Calabrese, L. Maurin et al., The Atacama cosmology telescope: DR4 maps and
cosmological parameters. J. Cosmology Astropart. Phys. 12, 047 (2020). https://doi.org/10.
1088/1475-7516/2020/12/047. (arXiv:2007.07288 [astro-ph.CO])
5. Y. Akrami, Y. Fantaye, A. Shafieloo et al., Power asymmetry in WMAP and Planck temperature
sky maps as measured by a local variance estimator. Astrophys. J. Lett. 784, L42 (2014). https://
doi.org/10.1088/2041-8205/784/2/L42. (arXiv:1402.0870 [astro-ph.CO])
6. Y. Akrami et al., Planck 2018 results X: constraints on inflation. Astron. Astrophys. 641, A10
(2020). https://doi.org/10.1051/0004-6361/201833887. (arXiv:1807.06211 [astro-ph.CO])
7. A. Amon, G. Efstathiou, A non-linear solution to the S.8 tension? MNRAS 516(4), 5355–5366
(2022). https://doi.org/10.1093/mnras/stac2429. (arXiv:2206.11794 [astro-ph.CO])
8. E. Aver, K.A. Olive, E.D. Skillman, The effects of He i .λ10830 on helium abundance determi-
nations. J. Cosmol. Astropart. Phys. 2015(7) (2015). https://doi.org/10.1088/1475-7516/2015/
07/011. (arXiv:1503.08146)
9. L. Balkenhol et al.: A measurement of the CMB temperature power spectrum and constraints
on cosmology from the SPT-3G 2018 TT/TE/EE data set (2022). (arXiv:2212.05642 [astro-
ph.CO])
10. C.L. Bennett, D. Larson, J.L. Weiland et al., Nine-year Wilkinson Microwave Anisotropy Probe
(WMAP) observations: final maps and results. ApJs 208(2), 20 (2013). https://doi.org/10.1088/
0067-0049/208/2/20. (arXiv:1212.5225 [astro-ph.CO])
11. D. Brout, D. Scolnic, It’s dust: solving the mysteries of the intrinsic scatter and host-
galaxy dependence of standardized type la Supernova Brightnesses. ApJ 909(1), 17 (2021).
arXiv:2004.10206
12. D. Brout, D. Scolnic, B. Popovic et al., The Pantheon+ analysis: cosmological constraints. ApJ
938(2), 110 (2022). https://doi.org/10.3847/1538-4357/ac8e04. (arXiv:2202.04077 [astro-
ph.CO])
13. P. Cabella, M. Kamionkowski, Theory of cosmic microwave background polarization. Astro-
physics. ArXiv e-prints astro-ph/0403392
14. J. Carron, M. Mirmelstein, A. Lewis, CMB lensing from Planck PR4 maps. JCAP 09, 039
(2022). https://doi.org/10.1088/1475-7516/2022/09/039. (arXiv:2206.07773 [astro-ph.CO])
15. R.J. Cooke, M. Pettini, C.C. Steidel, One percent determination of the primordial
deuterium abundance. ApJ 855, 102 (2018). https://doi.org/10.3847/1538-4357/aaab53
(arXiv:1710.11129)
316 P. Lemos and P. Shah

16. P. Creminelli, G. D’Amico, J. Noreña, F. Vernizzi, The effective theory of quintessence: the w
.< i-1 side unveiled. J. Cosmol. Astropart. Phys. 2, 018 (2009). https://doi.org/10.1088/1475-
7516/2009/02/018. (arXiv:0811.0827 [astro-ph])
17. A. Cuceu, J. Farr, P. Lemos, A. Font-Ribera, Baryon acoustic oscillations and the Hubble
constant: past, present and future. J. Cosmol. Astropart. Phys. 10, 044 (2019). https://doi.org/
10.1088/1475-7516/2019/10/044. (arXiv:1906.11628 [astro-ph.CO])
18. C. Deffayet, O. Pujolàs, I. Sawicki, A. Vikman, Imperfect dark energy from kinetic gravity
braiding. J. Cosmol. Astropart. Phys. 10, 026 (2010). https://doi.org/10.1088/1475-7516/2010/
10/026. (arXiv:1008.0048 [hep-th])
19. E. Di Valentino, A. Melchiorri, J. Silk, Planck evidence for a closed universe and a possible
crisis for cosmology. Nat. Astron. 4(2), 196–203 (2019). https://doi.org/10.1038/s41550-019-
0906-9. (arXiv:1911.02087 [astro-ph.CO])
20. C. Dickinson, CMB foregrounds—a brief review, in 51st Rencontres de Moriond on Cosmology
(2016), pp. 53–62. (arXiv:1606.03606 [astro-ph.CO])
21. G. Efstathiou, S. Gratton, The evidence for a spatially flat Universe. Mon. Not. R. Astron. Soc.
496(1), L91–L95 (2020). https://doi.org/10.1093/mnrasl/slaa093. (arXiv:2002.06892 [astro-
ph.CO])
22. G.P. Efstathiou, S. Gratton, A detailed description of the CamSpec likelihood pipeline and a
reanalysis of the Planck high frequency maps (2019). arXiv arXiv:1910.00483
23. D.J. Fixsen, The temperature of the cosmic microwave background. ApJ 707(2), 916–920
(2009). https://doi.org/10.1088/0004-637X/707/2/916. (arXiv:0911.1955)
24. A.H. Guth, Inflationary universe: a possible solution to the horizon and flatness problems. Phys.
Rev. D 23, 347–356 (1981). https://doi.org/10.1103/PhysRevD.23.347
25. W. Handley, Curvature tension: evidence for a closed universe. Phys. Rev. D 103(4), L041301
(2021). https://doi.org/10.1103/PhysRevD.103.L041301. (arXiv:1908.09139 [astro-ph.CO])
26. W. Handley, P. Lemos, Quantifying the global parameter tensions between ACT, SPT, and
Planck. Phys. Rev. D 103(6), 063529 (2021)
27. J.W. Henning, J.T. Sayre, C.L. Reichardt et al., Measurements of the temperature and E-mode
polarization of the CMB from 500 square degrees of SPTpol data. ApJ 852(2), 97 (2018).
https://doi.org/10.3847/1538-4357/aa9ff4. (arXiv:1707.09353)
28. C. Heymans et al., KiDS-1000 cosmology: multi-probe weak gravitational lensing and spec-
troscopic galaxy clustering constraints. Astron. Astrophys. 646, A140 (2021). https://doi.org/
10.1051/0004-6361/202039063. (arXiv:2007.15632 [astro-ph.CO])
29. G. Hinshaw, D. Larson, E. Komatsu et al., Nine-year Wilkinson Microwave Anisotropy Probe
(WMAP) observations: cosmological parameter results. Astrophys. J. Suppl. Ser. 208(2), 19
(2013a)
30. G. Hinshaw, D. Larson, E. Komatsu et al., Nine-year Wilkinson Microwave Anisotropy Probe
(WMAP) observations: cosmological parameter results. Astrophys. J. Suppl. Ser. 208, 19
(2013b). https://doi.org/10.1088/0067-0049/208/2/19. (arXiv:1212.5226)
31. F. Hoyle, G. Burbidge, J.V. Narlikar, A different approach to cosmology: from a static universe
through the big bang towards reality (Cambridge University Press, 2000)
32. W. Hu, M. White, Acoustic signatures in the cosmic microwave background. ApJ 471, 30
(1996). https://doi.org/10.1086/177951.astro-ph/9602019
33. W. Hu, M. White, A CMB polarization primer. New A2(2), 323–344 (1997). https://doi.org/
10.1016/S1384-1076(97)00022-5.astro-ph/9706147
34. W. Hu, M. White, The damping tail of cosmic microwave background anisotropies. ApJ 479,
568–579 (1997). https://doi.org/10.1086/303928.astro-ph/9609079
35. A. Kogut, M. Abitbol, J. Chluba et al., CMB spectral distortions: status and prospects (2019)
36. J.M. Kovac, E.M. Leitch, C. Pryke et al., Detection of polarization in the cosmic microwave
background using DASI. Nature 420, 772–787 (2002). https://doi.org/10.1038/nature01269.
astro-ph/0209478
37. P. Lemos, Tests of the Planck cosmology at high and low redshifts. PhD thesis, Cambridge
University (2018). https://doi.org/10.17863/CAM.36135
16 The Cosmic Microwave Background and . H0 317

38. P. Lemos, E. Lee, G. Efstathiou, S. Gratton, Model independent . H (z) reconstruction using the
cosmic inverse distance ladder (2018). ArXiv e-prints arXiv:1806.06781
39. P. Lemos et al., Assessing tension metrics with dark energy survey and Planck data. Mon.
Not. R. Astron. Soc. 505(4), 6179–6194 (2021). https://doi.org/10.1093/mnras/stab1670.
(arXiv:2012.09554 [astro-ph.CO])
40. A. Lewis, A. Challinor, Weak gravitational lensing of the CMB. Phys. Rep. 429(1), 1–65 (2006)
41. A.D. Linde, A new inflationary universe scenario: a possible solution of the horizon, flatness,
homogeneity, isotropy and primordial monopole problems. Phys. Lett. B 108, 389–393 (1982).
https://doi.org/10.1016/0370-2693(82)91219-9
42. A.D. Linde, Chaotic inflation. Phys. Lett. B 129, 177–181 (1983). https://doi.org/10.1016/
0370-2693(83)90837-7
43. T. Louis, E. Grace, M. Hasselfield et al., The Atacama cosmology telescope: two-season ACT-
Pol spectra and parameters. J. Cosmol. Astropart. Phys. 6, 031 (2017). https://doi.org/10.1088/
1475-7516/2017/06/031. (arXiv:1610.02360)
44. E. MacAulay, R.C. Nichol, D. Bacon et al., First cosmological results using type Ia supernovae
from the dark energy survey: measurement of the Hubble constant. MNRAS 486(2), 2184–2196
(2019). https://doi.org/10.1093/mnras/stz978. (arXiv:1811.02376)
45. A.A. Penzias, R.W. Wilson, A measurement of excess antenna temperature at 4080 Mc/s. ApJ
142, 419–421 (1965). https://doi.org/10.1086/148307
46. O.H.E. Philcox, M.M. Ivanov, BOSS DR12 full-shape cosmology: .Ʌ CDM constraints from
the large-scale galaxy power spectrum and bispectrum monopole. Phys. Rev. D 105(4), 043517
(2022). https://doi.org/10.1103/PhysRevD.105.043517. (arXiv:2112.04515 [astro-ph.CO])
47. O. Pisanti, G. Mangano, G. Miele, P. Mazzella, Primordial deuterium after LUNA: concor-
dances and error budget. J. Cosmol. Astropart. Phys. 4, 020 (2021). https://doi.org/10.1088/
1475-7516/2021/04/020. (arXiv:2011.11537 [astro-ph.CO])
48. Planck Collaboration, P.A.R. Ade, N. Aghanim et al., Planck 2013 results XXII con-
straints on inflation. A&A 571, A22 (2014). https://doi.org/10.1051/0004-6361/201321569.
(arXiv:1303.5082)
49. Planck Collaboration, P.A.R. Ade, N. Aghanim, et al.: Planck 2013 results XV CMB
power spectra and likelihood. A&A 571, A15 (2014b). https://doi.org/10.1051/0004-6361/
201321573. (arXiv:1303.5075 [astro-ph.CO])
50. Planck Collaboration, N. Aghanim, Y. Akrami et al., Planck 2018 results VI cosmological
parameters (2018). ArXiv e-prints arXiv:1807.06209
51. Planck Collaboration, N. Aghanim, Y. Akrami et al., Planck 2018 results VI: cos-
mological parameters. A&A 641 (2020). https://doi.org/10.1051/0004-6361/201833910.
(arXiv:1807.06209)
52. V. Poulin, T.L. Smith, T. Karwal, M. Kamionkowski, Early dark energy can resolve the Hubble
tension. Phys. Rev. Lett. 122(22), 221301 (2019). https://doi.org/10.1103/PhysRevLett.122.
221301. (arXiv:1811.04083)
53. V. Poulin, T.L. Smith, T. Karwal, The ups and downs of early dark energy solutions to the
hubble tension: a review of models, hints and constraints circa 2023 (2023). https://doi.org/10.
48550/arXiv.2302.09032. (arXiv:2302.09032 [astro-ph.CO])
54. F.J. Qu, B.D. Sherwin, M.S. Madhavacheril et al., The Atacama cosmology telescope: a mea-
surement of the DR6 CMB lensing power spectrum and its implications for structure growth
(2023). https://doi.org/10.48550/arXiv.2304.05202. (arXiv:2304.05202 [astro-ph.CO])
55. A.G. Riess, S. Casertano, W. Yuan et al., Cosmic distances calibrated to 1% precision with gaia
EDR3 parallaxes and Hubble space telescope photometry of 75 milky way Cepheids confirm
tension with lambda CDM. ApJ 908(1):1–11 (2021). (arXiv:2012.08534)
56. A.G. Riess, W. Yuan, L.M. Macri et al., A comprehensive measurement of the local value
of the Hubble constant with 1 km s.−1 Mpc.−1 uncertainty from the hubble space telescope
and the S. H0 ES team. ApJ 934(1), L7 (2022). https://doi.org/10.3847/2041-8213/ac5c5b.
(arXiv:2112.04510 [astro-ph.CO])
57. R.K. Sachs, A.M. Wolfe, Perturbations of a cosmological model and angular variations of the
microwave background. ApJ 147, 73 (1967). https://doi.org/10.1086/148982
318 P. Lemos and P. Shah

58. A. Sandage, Current problems in the extragalactic distance scale. ApJ 127, 513 (1958). https://
doi.org/10.1086/146483
59. N. Schöneberg, J. Lesgourgues, D. C. Hooper, The BAO+BBN take on the Hubble tension.
J. Cosmol. Astropart. Phys. 10, 029 (2019). https://doi.org/10.1088/1475-7516/2019/10/029.
(arXiv:1907.11594 [astro-ph.CO])
60. D. Scolnic, D. Brout, A. Carr et al., The Pantheon+ analysis: the full data set and
light-curve release. ApJ 938(2), 113 (2022). https://doi.org/10.3847/1538-4357/ac8b7a.
(arXiv:2112.03863 [astro-ph.CO])
61. G.F. Smoot, The CMB spectrum, in International School of Astrophysics, D. Chal-
lenge: 5th Course: Current Topics in Astro fundamental Physics (1997), pp. 407–440.
(arXiv:astro-ph/9705101)
62. G.F. Smoot, COBE observations and results, in 3K cosmology, vol. 476, ed. by L. Maiani,
F. Melchiorri, N. Vittorio (American Institute of Physics Conference Series, 1999), pp 1–10.
https://doi.org/10.1063/1.59326. (astro-ph/9902027)
63. G.F. Smoot, C.L. Bennett, A. Kogut et al., Structure in the COBE differential microwave
radiometer first-year maps. Astrophys. J. 396, L1–L5 (1992). https://doi.org/10.1086/186504
64. V. Sudevan, P.K. Aluri, S.K. Yadav, R. Saha, T. Souradeep, Improved diffuse foreground sub-
traction with the ILC Method: CMB map and angular power spectrum using planck and
WMAP observations. Astrophys. J. 842(1), 62 (2017). https://doi.org/10.3847/1538-4357/
aa7334. (arXiv:1612.03401 [astro-ph.CO])
65. R.A. Sunyaev, I.B. Zeldovich, The velocity of clusters of galaxies relative to the microwave
background—the possibility of its measurement. MNRAS 190, 413–420 (1980). https://doi.
org/10.1093/mnras/190.3.413
66. R.A. Sunyaev, Y.B. Zeldovich, The observations of relic radiation as a test of the nature of
X-ray radiation from the clusters of galaxies. Comments Astrophys. Space Phys. 4, 173 (1972)
67. M. Tucci, L. Toffolatti, The impact of polarized extragalactic radio sources on the detection of
CMB anisotropies in polarization. Adv. Astron. 2012, 624987 (2012). https://doi.org/10.1155/
2012/624987. (arXiv:1204.0427 [astro-ph.CO])
68. R. De Vaucouleurs, The cosmic Distance Scale and the Hubble Constant. Mount Stromlo and
Siding Spring Observatories, Australian National University (1982). https://books.google.ca/
books?id=zRSRzgEACAAJ
69. A. Vikman, Can dark energy evolve to the phantom? Phys. Rev. D 71(2), 023515 (2005). https://
doi.org/10.1103/PhysRevD.71.023515. (arXiv:astro-ph/0407107 [astro-ph])
70. S. Wagner-Carena, M. Hopkins, A. Diaz Rivero, C. Dvorkin, A novel CMB component separa-
tion method: hierarchical generalized morphological component analysis. Mon. Not. R. Astron.
Soc. 494(1), 1507–1529 (2020). https://doi.org/10.1093/mnras/staa744. (arXiv:1910.08077
[astro-ph.CO])
71. M. White, D. Scott, J. Silk, Anisotropies in the cosmic microwave background. ARA&A 32,
319–370 (1994). https://doi.org/10.1146/annurev.astro.32.1.319
72. M.J. White, Anisotropies in the CMB, in American Physical Society (APS) Meeting of the
Division of Particles and Fields (DPF 99) (1999). arXiv:astro-ph/9903232
73. Y.B. Zeldovich, R.A. Sunyaev, The interaction of matter and radiation in a hot-model universe.
Astrophys. Space Sci. 4, 301–316 (1969). https://doi.org/10.1007/BF00661821
Chapter 17
Measuring H0 with Spectroscopic
.

Surveys

Mikhail M. Ivanov and Oliver H. E. Philcox

17.1 Introduction

Spectroscopic surveys map out the large-scale structure (hereafter LSS) of the Uni-
verse through the distribution of galaxies. By measuring the positions and redshifts
of a wide variety of galaxies, we can build a three-dimensional map of the galaxy
overdensity, which, on comparatively large-scales, traces the underlying density of
dark matter (e.g., [1]). As such, spectroscopic surveys can be used to place constraints
on physics operating in the early Universe (which set the initial conditions for matter
clustering), and at late times (through the growth of structure). In this chapter, we
will discuss how they can be used to place competitive constraints on the Hubble
parameter, . H0 .
Since the late 1970s, there have been a wealth of galaxy redshift surveys, with the
current state-of-the-art public data (from the Sloan Digital Sky Survey [2]) containing
around one-million objects in the Northern sky. These are primarily analyzed through
clustering statistics, the simplest of which are the power spectrum, . P(k), or two-
point correlation function, .ξ(r ), encoding the distribution of pairs of objects on
some characteristic scale .r or .2π/k (.k denotes perturbations’ wavenumber). In the
coming decade the data volume will greatly increase, both due to ground-based
(DESI, and, potentially, others such as MegaMapper) and space-based (SPHEREx,
Euclid, Roman, and Rubin) missions [3–8]. With increasing survey volume comes
tighter constraints on cosmological parameters; as such, we may expect to yield sub-
percent measurements of the Hubble parameter from a variety of probes in the near
future.

M. M. Ivanov (B)
Institute for Advanced Study, Princeton, USA
e-mail: ivanov@ias.edu
O. H. E. Philcox (B)
Columbia University, New York and Simons Foundation, New York, USA
e-mail: ohep2@cantab.ac.uk

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 319
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_17
320 M. M. Ivanov and O. H. E. Philcox

As for the cosmic microwave background (CMB), spectroscopic . H0 constraints

arise from the combination of observed and physical scales, through the canoni-
cal ‘standard ruler’ mechanism. An important question, therefore, is which scales
should be used. To be an accurate probe of . H0 , a standard ruler must be: (a) pre-
cisely predicted form theory, and (b) well measured by experiment. In LSS, there
are two candidates: the particle horizon scale at the epoch of matter-radiation equal-
ity (at .z eq ≈ 3600) (e.g., [9]) and the sound horizon at recombination (strictly, at
baryon drag, .z d ≈ 1060) (e.g., [10, 11]) Each is straightforward to predict, and
leaves detectable signatures in the observed galaxy clustering statistics, primarily
(and fortuitously) concentrated in the linear regime. Whilst one could concoct other
physical scales (such as that of Silk damping, or dark-energy-matter equality), the
above are the simplest in practice in the context of galaxy surveys.
Mathematically, the directly observed angular size of a feature, .Δθ is related to
the theoretically predicted physical size, .r , via
rfeature
Δθfeature ≡
. , (17.1)
(1 + z)D A (z)

where . D A (z) is the angular diameter distance to the galaxy survey at some redshift
z. If the feature is primordial, .rfeature does not depend on physics operating at late
.
times, however . D A (z) does, thus the measurement is not strictly ‘early’ in nature.
Let us use the following heuristic argument to understand where the sensitivity
to . H0 comes from. Inserting the form of the angular diameter distance in Eq. (17.1)
and assuming a flat Universe (for simplicity alone), we find
{
Δθfeature z
dz '
. H0 = , (17.2)
rfeature 0 E(z ' )

where . E 2 (z) = Ωm,0 (1 + z)3 + Ωr,0 (1 + z)4 + ΩΛ ≡ H 2 (z)/H02 . Equation (17.2)

gives the rough idea on how the combination of a measured .rfeature and an observed
.Δθfeature yields. H0 (given the cosmological expansion history); equivalently, the angu-

lar measurements translate (roughly) to constraints on.rfeature H0 . In the below, we will

discuss our two possibilities for the standard ruler feature, including their physical
origin and non-linear evolution, and present the various associated. H0 measurements,
as well as their limitations and future prospects.1

1 We caution that for the minimal .ΛCDM model, and for probes of cosmological fluctuations such
as the CMB and LSS, .ΩΛ , Ωm etc. implicitely depend on . H0 , so that Eq. (17.2) does not actually
define the Hubble constant.
17 Measuring . H0 with Spectroscopic Surveys 321

17.2 Baryon Acoustic Oscillations

17.2.1 Theoretical Background

Prior to recombination, the Universe is filled with a strongly interacting photon-

baryon fluid, pressure-supported against. As with any fluid, this is filled with sound
waves, on some characteristic scale .rs , seeded by early Universe inhomogeneities.
As the Universe cools and free electrons combine with protons in recombination,
the fluid decouples (at the baryon drag redshift .z ≈ z d ∼ 1060) and the photon-
baryon oscillations cease to propagate, effectively ‘freezing’ the sound waves at the
sound-horizon scale, .rs (z d ) ≡ rd . During the subsequent evolution of the Universe,
this characteristic inhomogeneity is transferred to the dark matter distribution by
gravitational interactions, yielding an excess of clustering at late-times.
Physically, these features (dubbed ‘baryon acoustic oscillations’, hereafter BAO)
source a peak in the two-point correlation function at .r ∼ rd , or an oscillatory feature
in the galaxy power spectrum at .k ∼ 2π/rd . The precise sound horizon scale can be
computed explicitly:
{ ∞
cs (z)
r (z) =
. s dz ' , (17.3)
z H (z ' )
[ ]−1/2
with the sound speed .cs ≈ c 3 + 9/4ρb /ργ , where .ρb,γ are baryon and photon
densities. All quantities in the above can be estimated precisely given cosmological
parameters (such as those measured by Planck), and are principally sensitive to
the physical baryon density, .ωb , since the photon density is strongly constrained
from the CMB monopole. Assuming the Planck .ΛCDM cosmology, we find .rd ∼
105h −1 Mpc, with an approximate cosmology dependence of [12]
[ ]
55.154 exp −72.3(ων,0 + 0.0006)2
r ≈
. d Mpc , (17.4)
ωm,0 ωb,0
0.25351 0.12807

where.ωm,0 , ωb,0 , ων,0 denote the current physical densities of matter (i.e. dark matter
+ baryons), baryons, and neutrinos.
Whilst (17.3) accurately predicts the BAO scale at decoupling, various physi-
cal effects could alter its manifestation in the late-time galaxy correlators. Firstly,
the amplitude of oscillatory features decays with time due to bulk flows, i.e. the
large-scale motion of matter in the post-recombination Universe. At late-times, the
oscillatory part of the linear power spectrum, . PLosc , (which contains the BAO wig-
gles), is suppressed according to

PLosc (k, z) → e−k Σ (z) PLosc (k, z),

2 2
. (17.5)
{Λ
. Σ (z) = dp PL ( p, z)[1 − j0 ( prd ) + 2 j2 ( prd )] ,
2 1
6π 2 0
(17.6)
322 M. M. Ivanov and O. H. E. Philcox

where .Λ ≈ 0.1 h Mpc−1 is the IR cutoff scale, . jl (x) are spherical Bessel functions
and .Σ(z) is BAO damping scales extracted via a systematic theoretical procedure
called IR resummation [13–17] (also see [18, 19] for early work on the subject). When
extracting the BAO signal from data, it is important to include such a suppression
to avoid biased results. We note that this is not included in simple parametrizations
such as HaloFit.
Secondly, gravitational evolution modulates the clustering of matter at late times,
causing various physical effects such as halo formation, which impact the galaxy
power spectrum. This is important for . H0 determination, since non-linear physics
could alter the position of peaks in the galaxy power spectrum, from which the
absolute distance scale is extracted via (17.2). In practice, the effects of this are
small, though not negligible (e.g., [10, 20]). Non-linear effects yield a (cosmology-
dependent) .∼ 0.5% shift in the BAO scale whose small amplitude is itself somewhat
of an accident in .ΛCDM [16, 17, 21]. In practice, one often marginalizes over
non-linear contributions to the power spectrum, whilst assuming the BAO to be
determined only by linear physics (plus damping) (e.g. [22]). This is sufficiently
accurate for current surveys, but will not suffice in the future.

17.2.2 Extracting the BAO Feature

17.2.2.1 Template Analyses

To infer . H0 from the BAO feature, it has become commonplace to use template (or
scaling) analyses, based on the measurement of Alcock-Paczynski parameters (e.g.,
[22, 23]). This involves a theoretical templates for the BAO feature, such that the
oscillatory part of the power spectrum (or equivalently correlation function), takes
the form

. P obs (k|| , k⊥ ) ∝ P template (k|| /α|| , k⊥ /α⊥ ), (17.7)

where we parametrize the three-dimensional space by .k|| , k⊥ (relative to some line-

of-sight direction), where the Alcock-Paczynski parameters are defined as

H fid (z)rdfid D A (z)rdfid

α =
. || α⊥ = (17.8)
H (z)rd A (z)r d
D fid

These encode both the physical BAO scale, .rd , and the distance to the galaxy sample,
via . D A (z) (tangential) or . H (z) (radial, through velocities), and are normalized such
that .α||,⊥ = 1 if the reference cosmology (used to define the templates) matches the
truth. We caution that these parameters assume that the position of the BAO peak is
not altered by non-linear physics.
In practice, one must additionally account for non-oscillatory effects contributing
to the power spectrum, including galaxy bias, the fingers-of-God effect, and redshift-
17 Measuring . H0 with Spectroscopic Surveys 323

space distortions. Incorporating these effects (usually via some phenomenological

framework, or with perturbation theory) gives rise to a full theoretical model for the
galaxy power spectrum depending on a number of free parameters, including .α|| , α⊥ .
Fitting this model to data then yields the best-fit parameters, themselves encoding
. H (z)r d and . D A (z)/r d . Though these are usually said to be cosmology-independent,
we caution that their estimation does require theoretical modeling to define the tem-
plates. If the templates are far from the truth, or new physics is at work that alters the
BAO features, (17.7) breaks down, thus one cannot extract . H0 -dependent quantities
directly from the measured Alcock-Paczynski parameters. In practice, the former
effect is small, assuming the background cosmology to be relatively well known.
The output of this procedure is measurements of the cosmological quantities
. H (z)r d and . D A (z)/r d . Alone, these do not measure . H0 , due to a perfect . H0 r d degen-
eracy; practically, they are measuring only the angular size of the BAO feature. As
mentioned above, one requires also the physical size. Combining these constraints
with knowledge of .rd , within .ΛCDM or some other model (such as early dark
energy), one can break the degeneracy and measure . H0 directly. This is usually done
by combining with the CMB (e.g., [22]), or with Big Bang Nucleosynthesis probes
(e.g., [24]), both of which place strong constraints on .ωb , and hence .rd . The resulting
constraints on . H0 will be discussed in Sect. 17.4.
Whilst the above discussion has focused on measuring quantities from the
observed galaxy power spectrum, it has become commonplace for survey analyses
to include also data from the “reconstructed” power spectrum [25]. In this instance,
one first performs a catalog-level transformation of the data, effectively ‘undoing’
the large-scale bulk flows which lead to smearing of the BAO feature. This can
be performed robustly, and increases the signal-to-noise of oscillatory features by
reducing the damping given in (17.5). The resulting BAO peak is well described by
(weakly damped) linear theory (although we stress that the underlying physics relies
on approximate cancellations between non-linear contributions), and can be used to
enhance constraints on . H (z)rd and . D A (z)/rd . At low redshifts (where the damping
is significant), this can bolster constraints by up to .∼ 40%.

17.2.2.2 Full-Shape Analyses

In the above formalism, the galaxy power spectrum is analyzed using a phenomeno-
logical model, effectively discarding any information besides the BAO feature. An
alternative approach, which has found favor in recent years, is to instead model the
entire power spectrum theoretically (not just the wiggles), from which a range of
cosmological information can be extracted, including . H0 (e.g., [26, 27]). This is
made possible via theoretical efforts such as the Effective Field Theory of Large
Scale Structure (hereafter EFT, (e.g., [28–32])), which provides a first principles
prediction for the galaxy power spectrum, . Pgg (k; θ ), based on perturbation theory
and symmetry arguments, operating in the linear and mildly non-linear regime. This
depends on a parameter vector .θ, which encompasses cosmology, galaxy bias, and
other nuisance parameters. By fitting . Pgg (k; θ ) to the data directly, we obtain a pos-
324 M. M. Ivanov and O. H. E. Philcox

terior for .θ , and thus obtain constraints on the underlying cosmological parameters.
Since the theory model depends on the cosmological initial conditions and subse-
quent evolution, it naturally picks up . H0 information through the BAO feature (and
the equality scale, as discussed below), including both the physical size (from theory)
and the angular size (distorted due to some choice of reference cosmology).
Full-shape analyses are naturally dependent on the cosmological model; however,
this is not a limitation, since one can construct predictions in any framework, be it
.ΛCDM or beyond, with extension parameters including early dark energy, modified
gravity, and massive neutrinos. As far as . H0 is concerned, the approach is similar
to a template fitting, but the templates themselves are dynamically varied, and thus
depend on cosmological parameters such as .Ωm and (through the wiggle damping),
.ωb . Rather than predicting derived parameters such as . H (z)r d ), full-shape analyses
can predict the underlying . H0 parameter (or some other combination of cosmolog-
ical parameters, such as the CMB temperature, or angular sound horizon), since
the physical forms of . D A (z), . H (z), and .rd are baked into the model. To extract . H0
with template-based approaches, external information on these functions must be
supplied, thus BAO measurements are analyzed jointly with Planck, or some other
external data-set, which provide estimates of the.Ωm (.ωb ) parameters needed to deter-
mine . D A and . H (.rd ). In the full-shape case, both the parameters can be constrained
from the data directly (through coordinate distortions and peak heights), though it is
usually commonplace to supplement the survey with external.ωb information coming
from the CMB or Big Bang Nucleosynthesis (BBN) constraints (e.g., [33]). In the
limit of a well-measured BAO feature, the precision of our output constraint on . H0 is
fully limited by our knowledge of .ωb , through the .rd (ωb ) relation: if this is not mea-
sured with full-shape approaches (which themselves give only comparatively weak
constraints), we may not see large improvements in . H0 constraints if the external
BBN information does not increase in precision.
As for template analyses, it is possible to consider the reconstructed power spec-
trum in the full-shape context. Once again, our logic is that reconstruction acts to
sharpen the BAO oscillations by shifting information into the two-point function
from the higher-point correlators, thus capturing . H0 information that would usually
be found only in the bispectrum and above. Though several works have considered
the full-shape modeling of the reconstructed power spectrum [34–38], it is highly
non-trivial, and depends strongly on the particular flavor of reconstruction applied
(quickly becoming unwieldy for multi-step reconstruction procedures). Given that
the aim of reconstruction is to sharpen the BAO feature, and not to undo non-linear
effects in the bulk flow (which it is known to perform poorly at, with non-linear
formation being distorted by the necessary smoothing [39]), it is thus appropriate
to employ a hybrid approach, performing a template analysis of the reconstructed
power spectrum and a full-shape analysis of the data, pre-reconstruction (e.g., [40,
41]). This proceeds by measuring Alcock-Paczynski parameters from the former
(effectively acting as a data compression step), and performing a joint cosmologi-
cal analysis of these with the measured power spectrum multipoles in a .ΛCDM (or
beyond) context, including the covariance between the two. This has been shown to
17 Measuring . H0 with Spectroscopic Surveys 325

bolster . H0 information from full-shape analyses by up to .∼ 40% [40], and avoids the
difficulty in modeling broadband reconstruction artifacts.

17.3 Matter-Radiation Equality

17.3.1 Theoretical Background

Around .z ≈ 3600, the dynamics of the Universe change from being dominated by
radiation pressure, to being dominated by matter (both baryonic and dark). The
characteristic redshift is straightforward to determine by setting equal the density of
photons and matter, .Ωm,0 (1 + z)3 = Ωr,0 (1 + z)4 (treating neutrinos as relativistic),
where the radiation density is set by the current temperature of the CMB, as measured
by FIRAS.2 The horizon scale is set by the light horizon at this time,
{ ∞
c
r
. eq = dz ' , (17.9)
z eq H (z ' )

with the Fourier-space scale being approximately

−4
k
. eq = (2Ωm H02 z eq )1/2 , z eq = 2.5 × 104 Ωm h 2 Θ2.7 (17.10)

Eisenstein and Hu [43], where .Θ2.7 is the CMB temperature in .2.7K units. This is a
precisely known scale that acts as a second standard-ruler for. H0 analyses (dependent
only on the physical matter density .Ωm h 2 ).
Physically, the equality scale corresponds to a turn-over of the matter power spec-
trum at wavenumbers .k ∼ keq , as described by the Meszaros equation. Whilst the
turnover predominantly occurs on larger scales than can be robustly probed in many
spectroscopic analyses (except for quasar samples, and SPHEREx, though it will be
subject to significant cosmic variance and systematic penalities), its signature can be
seen also on somewhat larger scales, due to a .ln2 (k/keq ) feature in the matter power
spectrum at .k ≥ keq . These scales are concentrated in the linear regime, which sig-
nificantly simplifies modeling, especially given that the feature is spectrally smooth,
unlike the easy-to-distinguish BAO oscillations.
As before, we practically measure the ratio of the observed and theoretical equality
scale: by the above arguments, this imparts a degeneracy .h ∼ Δθeq keq , thus we in
practice constrain the combination .Ωm h, usually denoted .Γ . The notion that galaxy
surveys are sensitive to this scale arose significantly before the interest in the BAO
feature, with measurements of .Γ being a primary target of the spectroscopic surveys
of the 1990s and early 2000s [44, 45], though interest in its use as a BAO-independent
probe of . H0 is comparatively recent.

2Note that physical moment of matter-radiation equality is determined by the invariant matter-to-
photon ratio and does not actually depend on the current CMB temperature [42].
326 M. M. Ivanov and O. H. E. Philcox

17.3.2 Extracting Equality

As for the BAO feature, one can proceed to measure the equality scale via template
based approaches, by constructing some phenomenological model for the low-.k
galaxy power spectrum, or from full-shape analyses. In this instance, most of the
constraints have been derived using full-shape approaches, though there has been
some work employed to connect the equality scale to effects parameters in the ‘Shap-
eFit’ formalism [46]. Though one could proceed to analyze only the galaxy power
spectrum in the linear regime, it is usually preferred to utilize the whole EFTofLSS
full-shape methodology to extract . H0 from the equality scale, since this (a) calibrates
the galaxy bias parameters from quasi-linear scales, (b) naturally includes any non-
linear contributions relevant at low-.k, allowing their differentiation from the equality
features in the transfer function.
In practice, applying full-shape modeling to the galaxy power spectrum (and
beyond) picks up information from both the equality and sound-horizon scales, since
both enter the theory model . Pgg (k, θ ). To isolate the equality feature, one employ
measures to remove the BAO features; it is not sufficient to simply employ a scale cut,
since there is some overlap between the smooth equality feature and the first BAO
wiggle. Noting that the BAO-derived . H0 information depends strongly on sound-
horizon calibration through .ωb information, one option is to simply perform the
analysis without an external prior on .ωb [47]. This is sufficient for current surveys
(where the external information is much more precise than the.ωb measured internally,
e.g., by BAO wiggle heights), but this will cease to be the case in the future, due to
higher precision survey data. In this case, one may effectively marginalize over the
oscillatory features, yielding an equality-only constraint on . H0 [48]. Notably that
these measurements are naturally dependent on the choice of model used to analyze
them (since one requires a prediction for . Pgg (k) in terms of .Γ ≡ Ωm h), however,
this is rarely a limitation since one can straightforwardly test any theoretical model
of interest within the formalism, and, furthermore, there is very different dependence
on the cosmological model to BAO analyses, e.g., the measurement would be largely
unaffected by changes in sound-horizon physics.
As described above, equality-based measurements are practically sensitive to
.Γ ≡ Ωm h, thus . H0 determination requires some additional constraint on .Ωm (and,

to a lesser extent, .n s ). This can be done within the survey itself, principally through
coordinate distortions and redshift evolution, or externally. In the latter case, one
can place a prior on .Ωm from uncalibrated BAO measurements or supernovae dis-
tance constraints (both of which measure the relative expansion rate, without . H0
calibration). We finally note that an analogous procedure can be used to constrain
. H0 from photometric surveys or CMB lensing. The same does not hold for the BAO

feature, since the projection integrals tend to wash out such features; indeed this
allows straightforward baryon-independent . H0 constraints to be wrought.
17 Measuring . H0 with Spectroscopic Surveys 327

17.4 Measurements

Having discussed the physics behind spectroscopic . H0 constraints, we now present a

number of recent constraints from both the BAO and equality scales. For tractability,
we will specialize to analyses performed on BOSS data and beyond. A representative
(though not necessarily complete) set of constraints are shown in Figs. 17.1 and 17.2,
which indicate the plurality of work on the subject, particularly given that most results
are derived from the same data-sets (BOSS and/or eBOSS clustering). In all cases,
we assume a .ΛCDM cosmology, unless otherwise specified.
We first consider discuss sound-horizon-based constraints on . H0 , as derived via
template fitting. This is the preferred mode of analysis for the BOSS and eBOSS sur-
veys, with the major data-product being a set of . H (z)rd and . D A (z)/rd measurements
at the effective redshifts of the relevant data sets. As discussed above, these can be
combined with external ‘shape’ information from the cosmic microwave background
or Big Bang Nucleosynthesis (effectively fixing .ωb ), to break the .rd -degeneracy
and extract . H0 directly. The official BOSS analysis (combining with Planck ) finds
.(67.6 ± 0.5) km s
−1
Mpc−1 [22]; this is updated to .(67.6 ± 0.4) km s−1 Mpc−1 with
the inclusion of eBOSS data [49] (see also [12, 50]). Removing Planck information
leads to broader constraints (since Planck tightly constrains .rd and .Ωm as well as . H0
itself, through BAO features), with [24] finding.(67.5 ± 1.1) km s−1 Mpc−1 , reducing
to .(67.6 ± 1.0) km s−1 Mpc−1 with eBOSS [51] (see also [52]).
Full-shape analyses make use of both sound-horizon and matter-radiation-equality
standard rulers, which can yield tight constraints without explicit Planck calibration
(see also the ShapeFit approach, which captures similar information in a template-
based approach [46, 53], finding .(67.9+0.7 −1
−0.4 ) km s Mpc
−1
with BOSS, eBOSS and
Planck). A wide variety of full-shape . H0 constraints exist, predominantly differing
in their choice of cosmological priors and supplementary datasets. For example, [54]
find.(68.7 ± 1.5) km s−1 Mpc−1 from the BOSS DR12 power spectrum with a Planck
prior on the baryon density (see also [27, 55]). Constraining the spectral slope to that
measured by Planck improves constraints by .∼ 40%, with consistent results found
from both Planck and BBN .ωb priors, and three major analysis teams [26, 27, 41,
+0.7
56]. Combining with the full Planck dataset yields .(68.0−0.5 ) km s−1 Mpc−1 [57]).
The above results can be improved further by folding in information from the
reconstructed power spectrum (which is standard practice in template analyses).
As shown in [40], this improves constraints by another .∼ 40%, with BOSS DR12
+ BBN yielding .(67.8 ± 0.7) km s−1 Mpc−1 or .(68.6 ± 1.1) km s−1 Mpc−1 with and
without.n s priors (see also [41, 58]). A variety of extensions to these approaches have
been performed, for example the inclusion of the configuration-space correlation
function [41, 59], a proxy for the real-space power spectrum [60, 61] (also see [62–
65]), lensing-galaxy cross correlations [66], the bispectrum monopole [67], and the
bispectrum multipoles [68, 69]. The most recent result, combining the BOSS power
spectrum, reconstructed power spectrum, real-space power spectrum, and bispectrum
multipoles, yields.(68.2 ± 0.8) km s−1 Mpc−1 [68]. This is consistent with the official
BOSS results and those from Planck. Similar results can be obtained using eBOSS
328 M. M. Ivanov and O. H. E. Philcox

Fig. 17.1 Summary plot of recent. H0 measurements from spectroscopic data, alongisde the Planck
CMB-only posterior (black) and S. H0 ES results (blue). In each case, we plot the .1σ bound on . H0
from the given analysis. Results shown in blue include a BBN prior on.ωb , whilst those in green use
an .ωb prior from Planck. Measurements shown in red are combined with the full Planck dataset.
In each case we list the datasets used, with . P⊥ indicating real-space power spectrum proxies,
κg
. Bl indicating bispectrum multipoles, and .C l indicating a galaxy-lensing cross spectrum. Entries
denoted ‘+ SNe’ add supernova .Ωm information from Pantheon or PantheonPlus, whilst ‘+ .n s ’
indicates use of the Planck prior on .n s . This figure shows only full-shape results (which naturally
include both BAO and equality information; Fig. 17.2 shows the other measurements
17 Measuring . H0 with Spectroscopic Surveys 329

Fig. 17.2 As Fig. 17.1, but displaying results from other types of measurement; full-shape equality
measurements, template BAO constraints, and template equality results. The colors are as before.
Results obtained using the ‘ShapeFit’ compression scheme are indicated explicitly
330 M. M. Ivanov and O. H. E. Philcox

data, both from the ELG [70], QSO [70–74], and combined samples [74, 75], though
their individual constraining power is weaker than that of BOSS.
Finally, we consider. H0 constraints which derive information only from the equal-
ity scale. Since this ruler suffers from increased cosmic variance and requires .Ωm
calibration, the constraints are generally weaker than for the BAO, though can still
be used to facilitate interesting tests of the cosmological model [48]. From tem-
plate based analyses of the BOSS and eBOSS data with BBN information, we find
+1.9
.(70.1−2.1 ) km s
−1
Mpc−1 using the ‘ShapeFit’ formalism [76], whilst results from
the eBOSS quasars yield the broad constraint .(64.8+8.4 −1
−7.8 ) km s Mpc
−1
[77]. In full-
shape analyses, one can isolate the equality information either by omitting the .ωb
prior, or by explicit marginalization; combining BOSS with CMB lensing and ‘Pan-
theon+’ supernovae, yields the final constraint .(64.8+2.2 −1
−2.5 ) km s Mpc
−1
, which is
sound-horizon-independent but consistent with BAO constraints and Planck (see
also [9, 47, 48, 78].
Whilst the above results have been derived in a .ΛCDM framework, similar analy-
ses are possible in any other cosmological model of choice. For the template analyses,
one reinterprets the . H (z)rd and . D A (z)/rd constraints with the updated model for
growth and .rd , though we caution that this requires the desired cosmology to be
relatively close to Planck, since one assumed a .ΛCDM model to generate the tem-
plates. In contrast, full-shape analyses can be extended by changing the underlying
model used in computation of the linear power spectrum (and, if necessary, late-time
evolution). A wide variety of alternative cosmologies have been considered, such as
non-flat universes, varying dark energy, early dark energy, varying . Neff , interacting
dark energy, quintessence, et cetera [22, 49, 53], full-shape BOSS [usually with
Planck]: [79–87]. In general, these give consistent constraints to those in .ΛCDM but
with broader errors, due to the additional degrees of freedom in the model.

17.5 Summary and Future Prospects

Spectroscopic surveys can be utilized to provide stringent constraints on the Uni-

verse’s structure, composition, and expansion rate. In this chapter, we have discussed
their ability to measure the latter, through standard rulers provided by baryon acous-
tic oscillations and the transition from radiation- to matter-domination. In each case,
one observes a characteristic angular size in the galaxy distribution (either directly
or indirectly) which is compared to a theoretical computation, with the ratio giving
an (inverse) absolute distance measurement, at known redshift. Historically, interest
has focussed on the BAO signal analyzed by way of scaling templates; more recently,
theoretical and computational advances have led to . H0 being directly extracted from
the spectrum (through implicit standard ruler constraints) using full-shape method-
ologies. To date, these measurements have allowed for tight bounds on the Universe’s
expansion rate, with the BOSS survey yielding CMB-independent.∼ 1.6% constraints
on . H0 (e.g., [68]), which is of comparable width to the Planck posterior.
It is interesting to discuss how such measurements will evolve in the future (see
Fig. 17.3 for a selection of representative forecasts). In the coming years, both DESI
17 Measuring . H0 with Spectroscopic Surveys 331

Fig. 17.3 As Fig. 17.1, but displaying a collection of . H0 forecasts for future experiments from
the literature. As before, constraints marked in blue include (current) BBN priors on .ωb , those in
green add Planck priors on .ωb , whilst those in red combine with the full Planck posterior. Here, we
group forecasts by instrument, with the captions denoting the forecasting set-up. ‘WL’ indicates the
inclusion of weak lensing information, whilst SO and SKA refer to the Simons Observatory and
the Square Kilometer Array
332 M. M. Ivanov and O. H. E. Philcox

[3] and Euclid [6] will measure the redshifts of hundreds of millions of galaxies,
providing precise mapping of the BAO parameters . H (z)rd and . D A (z)/rd from low
to intermediate redshifts; moreover, surveys such as SPHEREx [5] will add sig-
nificant low-redshift measurements, and, more futuristically, proposed space mis-
sions including MegaMapper [4] can yield a tight high-redshift extension. Accord-
ing to [52], five years of DESI data will lead to a tight BAO-derived . H0 constraint
of .σ (H0 ) = 0.33 km s−1 Mpc−1 via template methods, working at fixed .ωb (not-
ing that the DESI white paper does not present . H0 constraints directly [3]), whilst
the same paper gives .σ (H0 ) = 0.21 km s−1 Mpc−1 from Euclid. As before, there
are a plethora of forecasted constraints, with varying analysis choices: for exam-
ple, the official Euclid forecast indicates a full-shape power spectrum constraint of
.0.63 km s
−1
Mpc−1 [88] (without external priors), tightening to .0.37 km s−1 Mpc−1
in conjunction with weak lensing and photometric surveys, or .0.12 km s−1 Mpc−1 ,
additionally adding in future CMB data [89] (see also [48, 90–93]). For SPHEREx,
no official . H0 forecast has been made, but one can expect similar results to DESI
and Euclid spectroscopy, but with most information coming from lower redshifts.
In the future, surveys such as MegaMapper could yield .σ (H0 ) = 0.36 km s−1 Mpc−1
alone, or three times tighter with CMB data, though the precise forecasts depend
on which parameters are being varied (e.g., . Neff , Mν ) [93, 94]. We note that many
of these results are naturally reliant on the .ωb calibration (with BBN yielding a
.≈ 1.6% constraint [33]), thus, if this is tightened further (for example from better

BBN measurements), one may expect a tighter increase still.

From the equality scale, we may also expect significant improvements in the
future. A forecast for Euclid found .σ (H0 ) = 0.72 km s−1 Mpc−1 [48], which is only
some .40% tighter than the BAO measurement (without reconstruction). We find that
the two scales will be more balanced in terms of constraining power in the future,
in the absence of tighter BBN calibration. As such, many of the above full-shape
constraints are importantly impacted by equality-scale information. Finally, we note
that a variety of other measurements will appear soon, both from additional galaxy
surveys, such as Rubin and Roman [7, 8], as well as other probes including 21cm
emission, line-intensity mapping, radio surveys, and beyond.
The coming measurements will open up a wealth of tests of the cosmological
model. Firstly, BAO analyses allow one to map the . H (z)rd and . D A (z)/rd derived
parameters across a large swathe of redshifts, facilitating robust consistency tests of
the low-redshift Universe; the same is true for full-shape derived constraints if we
search for trends in the . H0 posteriors from different redshift samples. Secondly, the
consistency of equality and BAO . H0 constraints has been previously shown to be a
powerful diagnostic test of new physics in the early Universe, due to the different
dependence on redshifts and recombination-scale physics. For now, the tests are
weak, but with future surveys such as Euclid, one could potentially rule out many
such models [48]. Furthermore, the two rulers are symbiotic; it is only by combining
all sets of information available that we can obtain the strongest bounds on . H0 , and
thus physics new and old.
17 Measuring . H0 with Spectroscopic Surveys 333

Acknowledgements We thank Eric Baxter, Giovanni Cabass, Gerrit Farren, Blake Sherwin, Marko
Simonovic, and Matias Zaldarriaga, with whom many of the above . H0 constraints were derived.
OHEP thanks Emirates for hosting a visit during which the majority of this draft was written. OHEP
is a Junior Fellow of the Simons Society of Fellows and thanks the Simons Foundation for support.

References

1. P.J.E. Peebles, Large scale clustering in the universe, in Large Scale Structures in the Universe,
ed. by M.S. Longair, J. Einasto (1978), vol. 79, p. 217
2. SDSS collaboration, SDSS-III: massive spectroscopic surveys of the distant universe, the milky
way galaxy, and extra-solar planetary systems. Astron. J. 142, 72 (2011) [1101.1529]. https://
doi.org/10.1088/0004-6256/142/3/72
3. DESI collaboration, The DESI Experiment Part I: Science,Targeting, and Survey Design,
1611.00036
4. D.J. Schlegel et al., Astro2020 APC white paper: the MegaMapper: a z > 2 spectroscopic
instrument for the study of inflation and dark energy. Bull. Am. Astron. Soc. 51, 229 (2019)
[1907.11171]
5. O. Doré et al., Cosmology with the SPHEREX all-sky spectral survey, 1412.4872
6. EUCLID collaboration, Euclid Definition Study Report, 1110.3193
7. LSST collaboration, Large synoptic survey telescope: overview. Proc. SPIE Int. Soc. Opt. Eng.
4836, 10 (2002) [astro-ph/0302102] https://doi.org/10.1117/12.456772
8. J. Green et al., Wide-Field InfraRed Survey Telescope (WFIRST) Final Report, 1208.4012
9. E.J. Baxter, B.D. Sherwin, Determining the hubble constant without the sound horizon
scale: measurements from CMB lensing. Mon. Not. Roy. Astron. Soc. 501, 1823 (2021)
[2007.04007]. https://doi.org/10.1093/mnras/staa3706
10. D.J. Eisenstein, H.-J. Seo, M.J. White, On the robustness of the acoustic scale in the low-
redshift clustering of matter. Astrophys. J. 664, 660 (2007) [astro-ph/0604361]. https://
doi.org/10.1086/518755
11. SDSS collaboration, Detection of the Baryon acoustic peak in the large-scale correlation func-
tion of SDSS luminous red galaxies. Astrophys. J. 633, 560 (2005) [astro-ph/0501171].
https://doi.org/10.1086/466512
12. A. Cuceu, J. Farr, P. Lemos, A. Font-Ribera, Baryon acoustic oscillations and the hubble
constant: past, present and future. JCAP 10, 044 (2019) [1906.11628]. https://doi.org/10.
1088/1475-7516/2019/10/044
13. L. Senatore, M. Zaldarriaga, The IR-resummed effective field theory of large scale structures,
JCAP 02, 013 (2015) [1404.5954]. https://doi.org/10.1088/1475-7516/2015/02/013
14. T. Baldauf, M. Mirbabayi, M. Simonović, M. Zaldarriaga, Equivalence principle and the
Baryon acoustic peak. Phys. Rev. D 92, 043514 (2015) [1504.04366]. https://doi.org/10.
1103/PhysRevD.92.043514
15. D. Blas, M. Garny, M.M. Ivanov, S. Sibiryakov, Time-sliced perturbation theory for large scale
structure I: general formalism, JCAP 07, 052 (2016) [1512.05807]. https://doi.org/10.1088/
1475-7516/2016/07/052
16. D. Blas, M. Garny, M.M. Ivanov, S. Sibiryakov, Time-sliced perturbation theory II: Baryon
acoustic oscillations and infrared resummation, JCAP 07, 028 (2016) [1605.02149]. https://
doi.org/10.1088/1475-7516/2016/07/028
17. M.M. Ivanov, S. Sibiryakov, Infrared resummation for biased tracers in redshift space. JCAP
07, 053 (2018) [1804.05080]. https://doi.org/10.1088/1475-7516/2018/07/053
18. M. Crocce, R. Scoccimarro, Renormalized cosmological perturbation theory. Phys. Rev. D 73,
063519 (2006) [astro-ph/0509418]. https://doi.org/10.1103/PhysRevD.73.063519
334 M. M. Ivanov and O. H. E. Philcox

19. M. Crocce, R. Scoccimarro. Nonlinear evolution of Baryon acoustic oscillations. Phys. Rev. D
77, 023533 (2008) [0704.2783]. https://doi.org/10.1103/PhysRevD.77.023533
20. R.E. Smith, R. Scoccimarro, R.K. Sheth, Eppur Si Muove: on the motion of the acoustic peak
in the correlation function. Phys. Rev. D 77, 043525 (2008) [astro-ph/0703620]. https://
doi.org/10.1103/PhysRevD.77.043525
21. B.D. Sherwin, M. Zaldarriaga, The shift of the Baryon acoustic oscillation scale: a simple
physical picture. Phys. Rev. D 85, 103523 (2012) [1202.3998]. https://doi.org/10.1103/
PhysRevD.85.103523
22. BOSS collaboration, The clustering of galaxies in the completed SDSS-III Baryon oscillation
spectroscopic survey: cosmological analysis of the DR12 galaxy sample. Mon. Not. Roy.
Astron. Soc. 470, 2617 (2017) [1607.03155]. https://doi.org/10.1093/mnras/stx721
23. C. Alcock, B. Paczynski, An evolution free test for non-zero cosmological constant. Nature
281, 358 (1979). https://doi.org/10.1038/281358a0
24. N. Schöneberg, J. Lesgourgues, D.C. Hooper, The BAO+BBN take on the hubble tension.
JCAP 10, 029 (2019) [1907.11594]. https://doi.org/10.1088/1475-7516/2019/10/029
25. D.J. Eisenstein, H.-J. Seo, E. Sirko, D. Spergel, Improving cosmological distance mea-
surements by reconstruction of the Baryon acoustic peak. Astrophys. J. 664, 675 (2007)
[astro-ph/0604362]. https://doi.org/10.1086/518712
26. M.M. Ivanov, M. Simonović, M. Zaldarriaga, Cosmological parameters from the BOSS galaxy
power spectrum. JCAP 05, 042 (2020) [1909.05277]. https://doi.org/10.1088/1475-7516/
2020/05/042
27. G. D’Amico, J. Gleyzes, N. Kokron, K. Markovic, L. Senatore, P. Zhang et al., The cosmological
analysis of the SDSS/BOSS data from the effective field theory of large-scale structure. JCAP
05, 005 (2020) [1909.05271]. https://doi.org/10.1088/1475-7516/2020/05/005
28. D. Baumann, A. Nicolis, L. Senatore, M. Zaldarriaga, Cosmological non-linearities as an
effective fluid. JCAP 07, 051 (2012) [1004.2488]. https://doi.org/10.1088/1475-7516/2012/
07/051
29. J.J.M. Carrasco, M.P. Hertzberg, L. Senatore, The effective field theory of cosmologi-
cal large scale structures. JHEP 09, 082 (2012) [1206.2926]. https://doi.org/10.1007/
JHEP09(2012)082
30. A. Chudaykin, M.M. Ivanov, O.H.E. Philcox, M. Simonović, Nonlinear perturbation theory
extension of the Boltzmann code CLASS. Phys. Rev. D 102, 063533 (2020) [2004.10607].
https://doi.org/10.1103/PhysRevD.102.063533
31. M.M. Ivanov, Effective field theory for large scale structure. 2212.08488
32. G. Cabass, M.M. Ivanov, M. Lewandowski, M. Mirbabayi, M. Simonović, Snowmass white
paper: effective field theories in cosmology, in 2022 Snowmass Summer Study, 3 (2022)
2203.08232
33. E. Aver, K.A. Olive, E.D. Skillman, The effects of He I λ 10830 on helium abundance deter-
minations. JCAP 07, 011 (2015) [1503.08146]. https://doi.org/10.1088/1475-7516/2015/
07/011
34. C. Hikage, K. Koyama, A. Heavens, Perturbation theory for BAO reconstructed fields: one-loop
results in the real-space matter density field. Phys. Rev. D 96, 043513 (2017) [1703.07878].
https://doi.org/10.1103/PhysRevD.96.043513
35. S.-F. Chen, Z. Vlah, M. White, The reconstructed power spectrum in the Zeldovich approxi-
mation. JCAP 09, 017 (2019) [1907.00043]. https://doi.org/10.1088/1475-7516/2019/09/
017
36. C. Hikage, K. Koyama, R. Takahashi, Perturbation theory for the redshift-space matter power
spectra after reconstruction. Phys. Rev. D 101, 043510 (2020) [1911.06461]. https://doi.
org/10.1103/PhysRevD.101.043510
37. A. Ota, H.-J. Seo, S. Saito, F. Beutler, Modeling iterative reconstruction and displacement field
in the large scale structure. Phys. Rev. D 104, 123508 (2021) [2106.46]. https://doi.org/10.
1103/PhysRevD.104.123508
38. M. Schmittfull, T. Baldauf, M. Zaldarriaga, Iterative initial condition reconstruction. Phys. Rev.
D 96, 023505 (2017) [1704.06634]. https://doi.org/10.1103/PhysRevD.96.023505
17 Measuring . H0 with Spectroscopic Surveys 335

39. M. Schmittfull, Y. Feng, F. Beutler, B. Sherwin, M.Y. Chu, Eulerian BAO reconstructions and
N-point statistics. Phys. Rev. D 92, 123522 (2015) [1508.06972]. https://doi.org/10.1103/
PhysRevD.92.123522
40. O.H.E. Philcox, M.M. Ivanov, M. Simonović, M. Zaldarriaga, Combining full-shape and BAO
analyses of galaxy power spectra: a 1.6% CMB-independent constraint on H0 . JCAP 05, 032
(2020) [2002.04035]. https://doi.org/10.1088/1475-7516/2020/05/032
41. S.-F. Chen, Z. Vlah, M. White, A new analysis of galaxy 2-point functions in the BOSS
survey, including full-shape information and post-reconstruction BAO. JCAP 02, 008 (2022)
[2110.05530]. https://doi.org/10.1088/1475-7516/2022/02/008
42. M.M. Ivanov, Y. Ali-Haïmoud, J. Lesgourgues, H0 tension or T0 tension?. Phys. Rev. D 102,
063515 (2020) [2005.10656]. https://doi.org/10.1103/PhysRevD.102.063515
43. D.J. Eisenstein, W. Hu, Baryonic features in the matter transfer function, Astrophys. J. 496,
605 (1998) [astro-ph/9709112]. https://doi.org/10.1086/305424
44. M. Tegmark, Measuring cosmological parameters with galaxy surveys, Phys. Rev. Lett. 79,
3806 (1997) [astro-ph/9706198]. https://doi.org/10.1103/PhysRevLett.79.3806
45. 2dFGRS collaboration, The 2dF galaxy redshift survey: the power spectrum and the matter con-
tent of the universe. Mon. Not. Roy. Astron. Soc. 327, 1297 (2001) [astro-ph/0105252].
https://doi.org/10.1046/j.1365-8711.2001.04827.x
46. S. Brieden, H. Gil-Marín, L. Verde, ShapeFit: extracting the power spectrum shape information
in galaxy surveys beyond BAO and RSD. JCAP 12, 054 (2021) [2106.07641]. https://doi.
org/10.1088/1475-7516/2021/12/054
47. O.H.E. Philcox, B.D. Sherwin, G.S. Farren, E.J. Baxter, Determining the hubble constant
without the sound horizon: measurements from galaxy surveys. Phys. Rev. D 103, 023538
(2021) [2008.08084]. https://doi.org/10.1103/PhysRevD.103.023538
48. G.S. Farren, O.H.E. Philcox, B.D. Sherwin, Determining the hubble constant without the
sound horizon: perspectives with future galaxy surveys. Phys. Rev. D 105, 063503 (2022)
[2112.10749]. https://doi.org/10.1103/PhysRevD.105.063503
49. eBOSS collaboration, Completed SDSS-IV extended Baryon oscillation spectroscopic sur-
vey: cosmological implications from two decades of spectroscopic surveys at the Apache
point observatory. Phys. Rev. D 103, 083533 (2021) [2007.08991]. https://doi.org/10.1103/
PhysRevD.103.083533
50. M. Blomqvist et al., Baryon acoustic oscillations from the cross-correlation of Lyα absorption
and quasars in eBOSS DR14. Astron. Astrophys. 629, A86 (2019) [1904.03430]. https://
doi.org/10.1051/0004-6361/201935641
51. N. Schöneberg, L. Verde, H. Gil-Marín, S. Brieden, BAO+BBN revisited—growing the hubble
tension with a 0.7 km/s/Mpc constraint. JCAP 11, 039 (2022) [2209.14330]. https://doi.
org/10.1088/1475-7516/2022/11/039
52. Y. Wang, L. Xu, G.-B. Zhao, A measurement of the hubble constant using galaxy redshift
surveys, Astrophys. J. 849, 84 (2017) [1706.09149]. https://doi.org/10.3847/1538-4357/
aa8f48
53. S. Brieden, H. Gil-Marín, L. Verde, Model-agnostic interpretation of 10 billion years of cosmic
evolution traced by BOSS and eBOSS data, JCAP 08, 024 (2022) [2204.11868]. https://
doi.org/10.1088/1475-7516/2022/08/024
54. T. Colas, G. D’amico, L. Senatore, P. Zhang, F. Beutler, Efficient cosmological analysis of the
SDSS/BOSS data from the effective field theory of large-scale structure. JCAP 06, 001 (2020)
[1909.07951]. https://doi.org/10.1088/1475-7516/2020/06/001
55. T. Tröster et al., Cosmology from large-scale structure: constraining ɅCDM with BOSS.
Astron. Astrophys. 633, L10 (2020) [1909.11006]. https://doi.org/10.1051/0004-6361/
201936772
56. D. Wadekar, M.M. Ivanov, R. Scoccimarro, Cosmological constraints from BOSS with analytic
covariance matrices. Phys. Rev. D 102, 123521 (2020) [2009.00622]. https://doi.org/10.
1103/PhysRevD.102.123521
57. M.M. Ivanov, M. Simonović, M. Zaldarriaga, Cosmological parameters and neutrino masses
from the final planck and full-shape BOSS data. Phys. Rev. D 101, 083504 (2020)
[1912.08208]. https://doi.org/10.1103/PhysRevD.101.083504
336 M. M. Ivanov and O. H. E. Philcox

58. G. D’Amico, L. Senatore, P. Zhang, H. Zheng, The hubble tension in light of the full-shape
analysis of large-scale structure data. JCAP 05, 072 (2021) [2006.12420]. https://doi.org/
10.1088/1475-7516/2021/05/072
59. P. Zhang, G. D’Amico, L. Senatore, C. Zhao, Y. Cai, BOSS correlation function analysis from
the effective field theory of large-scale structure. JCAP 02, 036 (2022) [2110.07539]. https://
doi.org/10.1088/1475-7516/2022/02/036
60. M.M. Ivanov, O.H.E. Philcox, M. Simonović, M. Zaldarriaga, T. Nischimichi, M. Takada, Cos-
mological constraints without nonlinear redshift-space distortions. Phys. Rev. D 105, 043531
(2022) [2110.00006]. https://doi.org/10.1103/PhysRevD.105.043531
61. G. D’Amico, L. Senatore, P. Zhang, T. Nishimichi, Taming redshift-space distortion effects in
the EFTofLSS and its application to data [2110.00016]
62. A.J.S. Hamilton, M. Tegmark, The real space power spectrum of the PSCz survey from 0.01 to
300 h Mpc**-1. Mon. Not. Roy. Astron. Soc. 330, 506 (2002) astro-ph/0008392. https://
doi.org/10.1046/j.1365-8711.2002.05033.x, https://arxiv.org/abs/astro-ph/0008392
63. M. Tegmark, A.J.S. Hamilton, Y.-Z. Xu, The power spectrum of galaxies in the 2dF 100k
redshift survey. Mon. Not. Roy. Astron. Soc. 335, 887 (2002) astro-ph/0111575. https://
doi.org/10.1046/j.1365-8711.2002.05622.x, https://arxiv.org/abs/astro-ph/0111575
64. SDSS collaboration, The 3-D power spectrum of galaxies from the SDSS, Astrophys. J.
606, 702 (2004) astro-ph/0310725. https://doi.org/10.1086/382125, https://arxiv.org/
abs/astro-ph/0310725
65. R. Scoccimarro, Redshift-space distortions, pairwise velocities and nonlinearities, Phys. Rev.
D 70, 083007 (2004) astro-ph/0407214. https://doi.org/10.1103/PhysRevD.70.083007,
https://arxiv.org/abs/astro-ph/0407214
66. S.-F. Chen, M. White, J. DeRose, N. Kokron, Cosmological analysis of three-dimensional
BOSS galaxy clustering and Planck CMB lensing cross correlations via Lagrangian perturba-
tion theory. JCAP 07, 041 (2022) [2204.10392]. https://doi.org/10.1088/1475-7516/2022/
07/041
67. O.H.E. Philcox, M.M. Ivanov, BOSS DR12 full-shape cosmology: ɅCDM constraints from
the large-scale galaxy power spectrum and bispectrum monopole. Phys. Rev. D 105, 043517
(2022) [2112.04515]. https://doi.org/10.1103/PhysRevD.105.043517
68. M.M. Ivanov, O.H.E. Philcox, G. Cabass, T. Nishimichi, M. Simonović, M. Zaldarriaga, Cos-
mology with the galaxy bispectrum multipoles: optimal estimation and application to BOSS
data [2302.04414]
69. G. D’Amico, Y. Donath, M. Lewandowski, L. Senatore, P. Zhang, The BOSS bispectrum
analysis at one loop from the effective field theory of large-scale structure [2206.08327]
70. M.M. Ivanov, Cosmological constraints from the power spectrum of eBOSS emission
line galaxies. Phys. Rev. D 104, 103514 (2021) [2106.12580]. https://doi.org/10.1103/
PhysRevD.104.103514
71. A. Semenaite et al., Cosmological implications of the full shape of anisotropic cluster-
ing measurements in BOSS and eBOSS. Mon. Not. Roy. Astron. Soc. 512, 5657 (2022)
[2111.03156]. https://doi.org/10.1093/mnras/stac829
72. A. Semenaite, A.G. Sánchez, A. Pezzotta, J. Hou, A. Eggemeier, M. Crocce et al., Beyond
ɅCDM constraints from the full shape clustering measurements from BOSS and eBOSS
[2210.07304]
73. A. Chudaykin, M.M. Ivanov, Cosmological constraints from the power spectrum of eBOSS
quasars [2210.17044]
74. T. Simon, P. Zhang, V. Poulin, Cosmological inference from the EFTofLSS: the eBOSS QSO
full-shape analysis [2210.14931]
75. R. Neveux et al., Combined full shape analysis of BOSS galaxies and eBOSS quasars using
an iterative emulator. Mon. Not. Roy. Astron. Soc. 516, 1910 (2022) [2201.04679]. https://
doi.org/10.1093/mnras/stac2114
76. S. Brieden, H. Gil-Marín, L. Verde, A tale of two (or more) h’s 2212.04522
77. B. Bahr-Kalus, D. Parkinson, E.-M. Mueller, Measurement of the matter-radiation equal-
ity scale using the extended Baryon oscillation spectroscopic survey quasar sample
[2302.07484]
17 Measuring . H0 with Spectroscopic Surveys 337

78. T.L. Smith, V. Poulin, T. Simon, Assessing the robustness of sound horizon-free determinations
of the Hubble constant [2208.12992]
79. A. Chudaykin, K. Dolgikh, M.M. Ivanov, Constraints on the curvature of the universe and
dynamical dark energy from the full-shape and BAO data. Phys. Rev. D 103, 023507 (2021)
[2009.10106]. https://doi.org/10.1103/PhysRevD.103.023507
80. A. Chudaykin, D. Gorbunov, N. Nedelko, Exploring ɅCDM extensions with SPT-3G and
Planck data: 4σ evidence for neutrino masses, full resolution of the hubble crisis by dark
energy with phantom crossing, and all that [2203.03666]
81. S. Vagnozzi, E. Di Valentino, S. Gariazzo, A. Melchiorri, O. Mena, J. Silk, The galaxy power
spectrum take on spatial curvature and cosmic concordance. Phys. Dark Univ. 33, 100851
(2021) [2010.02230]. https://doi.org/10.1016/j.dark.2021.100851
82. S. Kumar, R.C. Nunes, P. Yadav, Updating non-standard neutrinos properties with Planck-
CMB data and full-shape analysis of BOSS and eBOSS galaxies. JCAP 09, 060 (2022)
[2205.04292]. https://doi.org/10.1088/1475-7516/2022/09/060
83. A. Reeves, L. Herold, S. Vagnozzi, B.D. Sherwin, E.G.M. Ferreira, Restoring cosmological
concordance with early dark energy and massive neutrinos?. Mon. Not. Roy. Astron. Soc. 520,
3688 (2023) [2207.01501]. https://doi.org/10.1093/mnras/stad317
84. G. D’Amico, L. Senatore, P. Zhang, Limits on wCDM from the EFTofLSS with the PyBird
code. JCAP 01, 006 (2021) [2003.07956]. https://doi.org/10.1088/1475-7516/2021/01/006
85. G. D’Amico, Y. Donath, L. Senatore, P. Zhang, Limits on clustering and smooth quintessence
from the EFTofLSS [2012.07554]
86. M.M. Ivanov, E. McDonough, J.C. Hill, M. Simonović, M.W. Toomey, S. Alexander et al.,
Constraining early dark energy with large-scale structure. Phys. Rev. D 102, 103502 (2020)
[2006.11235]. https://doi.org/10.1103/PhysRevD.102.103502
87. T. Simon, P. Zhang, V. Poulin, T.L. Smith, Updated constraints from the effective field theory
analysis of BOSS power spectrum on early dark energy [2208.05930]
88. Euclid collaboration, Euclid preparation: VII. Forecast validation for Euclid cosmological
probes. Astron. Astrophys. 642, A191 (2020) [1910.09273]. https://doi.org/10.1051/0004-
6361/202038071
89. Euclid collaboration, Euclid preparation—XV. Forecasting cosmological constraints for the
Euclid and CMB joint analysis. Astron. Astrophys. 657, A91 (2022) [2106.08346]. https://
doi.org/10.1051/0004-6361/202141556
90. A. Chudaykin, M.M. Ivanov, Measuring neutrino masses with large-scale structure: Euclid
forecast with controlled theoretical error. JCAP 11, 034 (2019) [1907.06666]. https://doi.
org/10.1088/1475-7516/2019/11/034
91. J.-F. Zhang, L.-Y. Gao, D.-Z. He, X. Zhang, Improving cosmological parameter estimation with
the future 21 cm observation from SKA. Phys. Lett. B 799, 135064 (2019) [1908.03732].
https://doi.org/10.1016/j.physletb.2019.135064
92. T. Sprenger, M. Archidiacono, T. Brinckmann, S. Clesse, J. Lesgourgues, Cosmology in the
era of Euclid and the square Kilometre array. JCAP 02, 047 (2019) [1801.08331]. https://
doi.org/10.1088/1475-7516/2019/02/047
93. N. Sailer, E. Castorina, S. Ferraro, M. White, Cosmology at high redshift—a probe of fun-
damental physics. JCAP 12, 049 (2021) [2106.09713]. https://doi.org/10.1088/1475-7516/
2021/12/049
94. S. Ferraro, N. Sailer, A. Slosar, M. White, Snowmass2021 cosmic frontier white paper: cosmol-
ogy and fundamental physics from the three-dimensional large scale structure [2203.07506]
 Part III
Complications with Measurements
Chapter 18
Impact of Peculiar Velocities
on Measurements of H0 .

W. D’Arcy Kenworthy and Tamara M. Davis

18.1 Introduction

The most direct method to measure the Hubble constant, . H0 , is to compare the cur-
rent proper distance, . D0 , to the current recession velocity, .v0 , to get . H0 = v0 /D0 .
All direct measures of . H0 use some variant of this basic idea. However, to derive . H0
accurately it is important that measurements of the recession velocity are not contam-
inated by (unbudgeted) local motions induced by inhomogeneities in the Universe.
Motions that deviate from the homogeneous, isotropic expansion are known as pecu-
liar velocities. On large scales peculiar velocities arise primarily due to gravitational
acceleration. The impact of peculiar velocities on measurements of . H0 are the focus
of this chapter.
The most prominent example of a direct . H0 measurement involves luminosity
distances inferred from standard candles such as supernovae, calibrated using a ‘dis-
tance ladder’. These are local measurements that directly infer recession velocities
from observed redshifts and compare those to measured distances.
By contrast, indirect methods involve fitting an entire cosmological model to data
that has some dependence on . H0 . The most prominent example uses the patterns in
the cosmic microwave background (CMB), which are dependent on . H0 but strongly
correlated with other parameters, so . H0 only appears as one parameter in a fit to a
global cosmological model.
Both direct and indirect methods assume an underlying on-average homogeneous
and isotropic universe (even when they are using inhomogeneities as the signal!).
However, we know fluctuations exist—after all, we are here—and these fluctuations

W. D’Arcy Kenworthy (B)

Stockholm University, AlbaNova University Center, Fysikum, Stockholm 10691, Sweden
e-mail: darcykenworthy@me.com
T. M. Davis
School of Mathematics and Physics, The university of Queensland, Brisbane, QLD 4072, Australia
e-mail: tamarad@physics.uq.edu.au

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 341
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_18
342 W. D’Arcy Kenworthy and T. M. Davis

can cause biases in our measurements of cosmological parameters unless properly

accounted for. Currently there is a tension between direct and indirect methods of
measuring . H0 , so it is critical to understand whether any inhomogeneities are biasing
our cosmological inferences.
In this chapter we review the impact of peculiar velocities on our measurements
of . H0 . We begin by summarising the theoretical basis for peculiar velocities. We
then look at the potential impact peculiar velocities can have on cosmological mea-
surements, and ways to mitigate that impact to give as pure a measurement of . H0 as
possible.

18.2 Theoretical Basis

The distance in the Hubble-Lemaître law is usually split into two components, the
scalefactor . R and the comoving coordinate .χ, such that . D = Rχ1 . Any motion
due to a change in the scalefactor, which affects all distances, is a recession veloc-
ity .vrec = Ṙχ = H D (where an overdot represents differentiation with respect to
time2 ), and the Hubble parameter . H ≡ Ṙ/R. While any motion due to move-
ment through comoving coordinates is a peculiar velocity, .vpec = R χ̇. Although the
Hubble-Lemaître law holds at all times we often evaluate these values at the present
day, which we indicate by the subscript 0.
There are two primary sources of peculiar velocities to worry about. Namely,
our own motion, .vpec
Sun
, and the sources’ motion, .vpec
source
. Both are measured with
respect to the cosmic ‘rest frame’, which is the frame in which the cosmic microwave
background (CMB) radiation appears isotropic.

18.2.1 Inferring Peculiar Velocities

Peculiar velocities can either be used as a cosmological probe in their own right, or
measured in order to mitigate their impact on other cosmological probes.
Inferring peculiar velocities from observables needs two measured quantities—the
observed redshift .z obs and an observed distance measure . Dobs . The distance measure
could be any measure (e.g. luminosity or angular diameter distance) from which we
can infer the true cosmological redshift.
To derive the recession velocity (the velocity that appears in the Hubble-Lemaître
Law, defined above) from the cosmological redshift one needs to know the values of
the cosmological parameters that accelerate and decelerate the universe such as the
matter density .Ωm , dark energy density .Ωx , and dark energy equation of state, .w.
The basic equation for recession velocity as a function of cosmological redshift is,

1 Throughout this work we assume a flat cosmology on large scales. In non-flat cosmologies, the
analysis is only marginally more complex.
2 The time in the Friedmann-Lemaître-Robertson-Walker metric.
18 Impact of Peculiar Velocities on Measurements of . H0 343
{ z cos
dz
v
. rec =c (18.1)
0 E(z)

where from Friedmann’s equation we use . E(z) ≡ H (z)/H0 =

[ ]
2 1/2
∑
Ωm (1 + z) + Ωx (1 + z)
3 3(1+w)
+ Ωk (1 + z) , where.Ωk = 1 − i Ωi . Impor-
tantly, you can see that recession velocity as a function of redshift is independent of
. H0 . You only need . H0 if you want to know how velocity depends on distance.
A good approximation for the recession velocity that is often used is,
[ ]
cz cos 1 1
v
. rec = 1 + (1 − q0 )z cos − (1 − q0 − 3q02 + j0 )z cos
2
, (18.2)
1 + z cos 2 6

where in the standard cosmological model the deceleration parameter is .q0 ∼ −0.55,
and the jerk is . j0 ∼ 1.0. This approximation is what is used in the S. H0 ES papers
(e.g. [1, 2]) for deriving . H0 from supernova distances and redshifts.
However, we do not have direct access to the true recession redshift, .z cos . Instead
we have the observed redshift, .z obs , which is impacted by peculiar velocities—both
our own and that of the source.
The total velocity is the sum of the peculiar and recession velocities, .vtot = vrec +
vpec . So in many direct measurements of peculiar velocity you see this rearranged to
become .vpec = vtot − vrec ≈ cz obs − H0 D, where .vrec was replaced with . H0 D from
the Hubble-Lemaître law, and the total velocity was approximated by.cz obs . However,
this is a terrible approximation even by redshift .z ∼ 0.1 where it will give a .∼700
km s−1 error in the derived peculiar velocity [3].3
A more accurate equation for expressing velocity in terms of observed redshift
and observed distance is,
( )
z obs − z cos (Dobs )
.vpec = c , (18.3)
1 + z cos (Dobs )

where .z cos is the cosmological redshift that has been inferred from the observed dis-
tance measure . Dobs . Calculating .z cos from a distance measure typically requires a
cosmological model.4 However, at low redshift (.z ≾ 0.1) the cosmological depen-
dence is weak and any model that approximates the expansion evolution well (e.g.
Eq. 18.2 is sufficient).
Direct measures of peculiar velocities are expensive (because you need to mea-
sure distances) and have large errors (because distance measurements have large
uncertainties). Therefore a more precise (but modelling-dependent) way to predict
the peculiar velocity of distance sources is to reconstruct a peculiar velocity field

obs = z cos + z pec , which neglects

3 Using this equation is equivalent to approximating Eq. 18.5 by .z

a factor of .z cos z pec . It therefore gets worse linearly with redshift. Note that most modern analyses
do not actually use this approximation, even if it still appears in the standard introduction in many
papers.
4 Just to describe the shape of the magnitude redshift relation, i.e. parameters such as .Ω and .Ω
m Λ
but not a value of . H0 .
344 W. D’Arcy Kenworthy and T. M. Davis

from the density contrast, .δ(→r ) ≡ (ρ(→

r ) − ρ̄)/ρ̄, where .ρ̄ is the mean density of the
Universe and .ρ is the density at position .r→. Given a density field as measured by a
large redshift survey, the large-scale peculiar velocity of a point at .r→x can be derived
by integrating over the gravitational attraction of all other points,
{
f r→ − r→x 3
. v→(→
rx ) = δ(→
r) d r→, (18.4)
4π |→
r − r→x |3

where. f is the growth rate of structure. This equation is only valid in the linear regime,
that is when the density fluctuations are small perturbations on the mean density of the
Universe (.δ << 1). Thus a reconstruction from the density field will only accurately
reproduce large scale peculiar velocities. It will not be able to reproduce small-scale
motions, such as the orbits of galaxies in a galaxy cluster (see Sect. 18.3.4).
In Sect. 18.4.2 we show how these peculiar velocity reconstructions can be used
to mitigate the impact of peculiar velocities on our cosmological measurements.

18.2.2 The Impact of Peculiar Velocities on Observables

The impact of peculiar velocities on cosmological measurements arises in a number

of ways. Most significantly they generate an additional peculiar redshift .z pec which
causes the observed redshift, .z obs , to deviate from the true cosmological (recession)
redshift .z cos according to [4],

.1 + z obs = (1 + z cos )(1 + z pec ). (18.5)

The luminosity distance (defined as the distance that fulfills .Flux = Luminosity/
(4π DL2 )) differs from the proper distance by a factor of .(1 + z obs ) and curvature
terms. Ignoring curvature and concentrating only on the differences impacted by
peculiar velocity, we can write,

. DL = (1 + z obs )D. (18.6)

The extra factor of .(1 + z obs ) arises because the luminosity drops both due to time
dilation decreasing the arrival rate of photons and the redshift decreasing the energy
of each photon. Note that these factors both depend on the total redshift the photons
experience, both cosmological and peculiar, which is why the .(1 + z obs ) factor uses
the observed redshift. In the absence of peculiar velocities it reduces to . D L = (1 +
z cos )D.
Peculiar velocities also induce relativistic beaming—photons are focused in the
direction of travel—which means the flux of photons is higher in the direction of
motion than behind it. For a source with a peculiar velocity toward us we therefore
infer both a lower redshift (relative blueshift) and a lower distance (brighter) than a
comoving source at the same distance, so the effects slightly cancel (see [5], Fig. 1).
18 Impact of Peculiar Velocities on Measurements of . H0 345

However, the beaming impact is an order of magnitude lower than the impact of the
redshift shift, and is therefore negligible for most realistic peculiar velocities.
Beaming due to our own motion also slightly alters the distribution of distant
sources, such that there is a higher number density of sources in the direction of our
motion [6, 7]. This becomes relevant for high-precision baryon acoustic oscillation
(BAO) and CMB measurements of . H0 , and has been seen both in the number counts
of distant galaxies (e.g. [8, 9]) and in the CMB [10, 11] at a measurable level.
The impact of peculiar velocities is largest for low-redshift probes, and when
few sources are available. Such is the current case for gravitational wave standard
sirens. For example, the peculiar velocity correction that was applied to the neutron-
star merger measurement of . H0 lowered the estimated . H0 by .∼ 7 km s−1 Mpc−1
[12] relative to the uncorrected measurement, and the large value of this correction
motivated several reanalyses [13–15]. As the number of sources grows the impact of
peculiar velocities tends to reduce, as their random directions average out.

18.3 Power Spectrum of Peculiar Velocities

Peculiar velocities are correlated across a range of scales and patterns, from small-
scale infall around clusters, through correlations between large-scale structure, to
bulk flows and expansion. It is therefore convenient to discuss the impact of peculiar
velocities using a power spectrum, in which we separate the signal into multipoles.
Here we discuss the monopole (expansion or contraction), dipole (motion in one
direction), and higher-order multipoles (correlated motion on smaller scales).

18.3.1 Monopole

The lowest multipole is the monopole, which corresponds to spherical expansion or

contraction. This could occur due to local over- or under-densities. Over the years
evidence has appeared and disappeared for such a local ‘Hubble bubble’ (e.g. [16–
21]). This is usually very difficult to disentangle from the effect of the cosmological
expansion, and so is one of the most insidious potential peculiar velocity errors.
However, to find ourselves sitting at the centre of an abnormally large over- or
under-density is disfavoured by the cosmological principle, and if we sit off-centre
we would see asymmetric effects that would reveal its presence. Observationally,
recent measurements of the distance-redshift relation from type Ia supernovae show
no evidence for a abnormal monopole sufficient to explain the Hubble tension [20,
21]. However, other measurements continue to report potentially anomolous structure
(see Sect. 18.3.3 for examples). Measurements of the local ‘gravitational watershed’
of our Universe, known as Laniakea [22, 23], shows that we do likely sit at the edge
(not centre) of an over-density.
346 W. D’Arcy Kenworthy and T. M. Davis

In general, we can quantify the expected distribution of monopole effects from

different locations in the standard model of cosmology, and they would be expected to
be small see also Chap. 21 [24]. However, advocates of inhomogenous cosmological
solutions to the Hubble tension usually invoke a density fluctuation that is larger than
expected in the standard model.

18.3.2 Dipole

The second multipole is the dipole, which corresponds to our own motion with respect
to the cosmic microwave background ‘rest’ frame. Measurements of the dipole in the
CMB reveal that the Sun is moving at.369.82 ± 0.11 km s−1 with respect to the cosmic
rest frame [25]. Interestingly our galaxy is moving faster, at .552.2 ± 5.5 km s−1 , but
the Sun in rotation around the galactic centre is currently going backwards relative
to the galaxy’s motion, so the two partially cancel out.
The dipole can be measured not only with respect to the CMB, but also with
respect to shells of matter at a range of redshifts—e.g. radio galaxies at high-redshift
[8, 9, 26] or supernovae at low to intermediate redshifts [27, 28]. As the radius of
the shell approaches the radius of the observable Universe you expect the dipole
measurement to converge on that of the CMB. However, when the radius of the shell
is very small, you expect the dipole to disappear, as we are measuring with respect
to galaxies that share our local motion.
It is important (and straightforward, using Eq. 18.5) to remove the dipole from
the observed redshift before calculating cosmological distances, and this is standard
practice. Historically this has often been done using low-redshift approximations
(akin to Eq. 18.12, but given the accuracy of modern cosmological measurements it
is becoming important to use the full correction as in Eq. 18.1 [29, 30].

18.3.3 Bulk Flow

The bulk flow is also a manifestation of the second multipole, but is not to be confused
with the dipole. The bulk flow refers to the average motion of all the galaxies (or
total matter) in a volume of space. For example, in the standard cosmological model
one expects the galaxies in a sphere of 100 Mpc./ h to have an average speed of
approximately 180 km s−1 Mpc−1 relative to the CMB rest frame.
Many theoretical calculations predicting the magnitude of the bulk flow assume
it to be a spherical volume, but in most cases a bulk flow can only be measured using
data sets covering incomplete patches of the sky. Moreover, the completeness of a
survey as a function of redshift has a significant impact on the bulk flow prediction.
Mitigating the incomplete data is a major challenge to estimating accurate bulk flows
and comparing them to the expectation from theory. Recent measurements have
18 Impact of Peculiar Velocities on Measurements of . H0 347

confirmed a long-standing trend that measured bulk flows are larger than expected
in the standard cosmological model [31–35].
In contrast to the dipole, you expect the bulk flow to be significant at small scales,
but disappear as you increase the radius to encompass the whole observable Universe.
Interestingly bulk flows stay somewhat large even out to surprisingly large distances,
the expectation of a spherical bulk flow of the volume within .z ∼ 1 (about half the
radius of the observable Universe) would still be about 10 km s−1 .

18.3.4 Higher Order Motion

Higher-order multipoles can be quantified using the peculiar velocity power spec-
trum, which describes the amplitude of peculiar velocity correlations as a function of
scale. The auto-correlation can be supplemented with the cross-correlation between
peculiar velocities and densities, which contains additional information as galaxies
fall toward over-dense regions.
Perturbations on very large scales are well described by linear perturbations on
a smooth background. Linear perturbation theory allows you to essentially predict
analytically how large scale fluctuations will behave. This is relevant for predicting
the size of the bulk flow, for example. However, on small scales perturbation theory
breaks down, and we enter the non-linear regime. On scales small enough that the
Universe deviates significantly from homogeneity (.δ ≿ 1) we can no longer treat
density fluctuations as small perturbations, and we need numerical simulations to
estimate the pattern of density and velocity fluctuations.
Scales on which you see galaxies orbiting each other (such as galaxy clusters)
are highly non-linear, subject to the virial theorem. However, non-linear density
fluctuations begin to become important even on much larger scales than clusters;
often nonlinear effects are present at scales .k ≿ 0.1 hMpc−1 , where .k ≡ 2π/|→ r |,
corresponding to separations of .|r | ≾ 60 Mpc h −1 . When nonlinear scales dominate
(.k ≿ 1 hMpc−1 ) one cannot straightforwardly perform peculiar velocity reconstruc-
tion, limiting the resolution of our peculiar velocity maps (see Sect. 18.4.2).

18.3.5 Covariances

Non-linear behavior is, in general, more complicated than can be represented by

Gaussian statistics, and requires use of higher order statistics to fully model. However,
given that velocities are less sensitive to small-scale behavior as compared to densities
or number-counts by a factor of .k 2 , non-linear behaviors are comparatively less
impactful. If necessary, the effects of nonlinear contributions can also be evaluated
by reference to full. N -body simulations [24]. Nevertheless, in many analyses the non-
linear or virial contribution is the component that is easiest to fit to data, typically
representing merely a diagonal uncorrelated scatter on the Hubble diagram. It is
348 W. D’Arcy Kenworthy and T. M. Davis

Fig. 18.1 Expected, uncorrected peculiar velocity covariance between pairs of galaxy in the low-
redshift Pantheon+ sample [36]. Each point represents a pair of SN host galaxies, with their angular
separation represented by color, and physical distance shown on the x-axis. Red points indicate
points on the same line of sight, while green points are across the sky from one another. The power
spectrum was evaluated with Planck parameters [37] and halofit corrections [38]

thus a reasonable approximation to use Gaussian statistics to quantify the effect of

peculiar velocities on cosmography. The matter power spectrum is defined by the
relation
→∗ )δ̃(k)
. < δ̃(k → >= (2π)3 P(|k|)δ
→ 3 (k→ − k→∗ ) (18.7)

→ is a Fourier transform of the density contrast .δ = Δρ/ρ̄, and .δ 3 is the

where .δ̃(k)
Dirac delta function in 3D. Using the theory of Gaussian statistics, you can then
compute the covariance of the peculiar velocity between two objects at arbitrary
points as
{
dD || dD || dk
. < vi v j > = | | PΛCDM (k) × F(kχi , kχ j , cos(θ)) (18.8)
dτ χi dτ χ j 2π 2

with
∞
∑
. F(u, v, cos(θ)) = (2l + 1) jl' (u) jl' (v)Pl (cos(θ)) (18.9)
l=0
√ |
∂ 2 sinc( x 2 − 2x y cos(θ) + y 2 ) ||
. = (18.10)
∂x∂ y |
x=u,y=v

where .τ is conformal time, .D is the linear growth factor, .χ is comoving distance, and
.θ is the angular separation between the two objects across the sky [39, 40]. The term
. F(u, v, cos(θ)), which has been written here as a sum over all multipoles, can alter-
natively be expressed in closed form using the second definition. An example of this
covariance function is visualized in Fig. 18.1, where we show the covariance between
pairs of galaxies found in the Pantheon+ [36] catalog. Galaxies close together have
almost wholly correlated motions, while galaxies on the same line of sight but distant
18 Impact of Peculiar Velocities on Measurements of . H0 349

from one another are anti-correlated. Galaxies on the opposite side of the sky show
significant anti-correlation, depending on distance. These covariances are present
throughout the sample, and require mitigation through several techniques to yield an
accurate measurement.

18.4 Mitigations of the Impact of Peculiar Velocities on . H0

Measurements

While the effect of peculiar velocities on leading measurements is likely insuffi-

cient to explain the Hubble tension, verification of proposed theoretical solutions
will require a maximally robust, precise, and accurate measurement. Mitigating, cor-
recting for, and budgeting peculiar velocity uncertainties will thus be an essential
element of distance-scale measurements using data in the local (.z < 0.05) universe.
We discuss the strategies used by different teams below, along with suggestions for
improvements in future iterations.

18.4.1 Exclusion/Weighting of Data

The simplest means of mitigating the effect of peculiar velocities on direct measure-
ments of the Hubble constant is to exclude low-redshift data. Since the fractional error
in redshifts is the quantity of concern for a direct measurement of . H0 , removing data
below a given redshift drastically reduces the potential impact of these uncertainties,
without introducing additional dependence on modeling or external datasets. How-
ever relying entirely on high-redshift data may limit the statistical power of the sam-
ple, increase dependence on other systematic uncertainties such as evolution, and/or
increase the dependence of the result on knowledge of the higher-order terms of the
Hubble-Lemaître law. The S. H0 ES team cut their Hubble flow sample at .z > 0.023
for the primary measurement. As can be seen in Fig. 18.2, most of the constraining
power of the supernova sample is above this value, while the variance contribution
of peculiar velocities is no longer dominant. In contrast, the CCHP team [41] chose
to include all 99 supernovae from the Carnegie Supernova Project [42] that passed
quality cuts and were not themselves used as TRGB calibrators in their Hubble flow
sample. This approach effectively represents a cut at .z > 0.004. Similarly the Mega-
maser Cosmology Project [43], with only six objects to choose from, did not impose
any redshift cut on their final result. It is thus evident that the choice of a redshift
cut will depend on the demographics of a particular sample and the significance of
peculiar velocities in the overall error budget of a given measurement.
A less extreme approach (which may be combined with redshift cuts) than entirely
cutting low-redshift data is to downweight nearby data using a variance or covariance
matrix [5, 39]. This often takes the form of an assumed uncorrelated velocity scatter
350 W. D’Arcy Kenworthy and T. M. Davis

Fig. 18.2 Dependence of the uncertainty in the S. H0 ES estimate of . H0 from the type Ia supernova
Hubble diagram as a function of minimum redshift cut. Purple line shows total, budgeted uncertainty
in the measurement from the Hubble diagram measurement (not including uncertainties from the first
two rungs of the ladder), while light blue shows potential impact of uncorrected peculiar velocities
based on the .ΛCDM power spectrum. The power spectrum was evaluated with Planck parameters
[37] and halofit corrections [38]. As the real measurement includes corrections from reconstructed
peculiar velocity maps, PV uncertainty in the final measurement is significantly smaller, shown in
dark blue

of .σPV ≈ 150 − 300km s−1 , ignoring the correlations between velocities, which are
small for widely separated sources. Frequently no covariant or correlated uncertain-
ties will be included in the primary fit of the analysis, although the sensitivity of the
final value to various correction methodologies (as discussed in the next section) may
be investigated and ultimately included in systematic error budgets. Different stud-
ies have found quite different values for the total variance of peculiar velocities .σPV
[44–51] ranging from 150 to 250 km s−1 . As was discussed earlier in Sect. 18.3, the
magnitude of this quantity will include significant nonlinear and virialized contribu-
tions it is difficult to predict from theory. Accurate measurement of this term requires
both an understanding of peculiar velocities and the control of other elements of the
uncertainty budget. Assessment of the correlated contribution from linear theory
alone is comparatively much simpler [39], as measurements of the power spectrum
at scales .k < 0.1 Mpc h−1 are widely published and the relevant physics is thought to
be well understood. Use of an appropriate covariance matrix will reduce sensitivity
of the final result to peculiar velocities and improve the robustness of the result. How-
ever it can be a significant contributor to the final error budget. To further improve
measurements, direct correction of peculiar velocity noise is necessary.

18.4.2 Peculiar Velocity Corrections

Previously mentioned methods work by reducing the contribution of the most con-
taminated data in a linear fit to the Hubble diagram. A separate category of miti-
18 Impact of Peculiar Velocities on Measurements of . H0 351

gations would be the attempt to subtract the signal in the observed data by using
other datasets to predict the peculiar velocities. As peculiar velocities are gravita-
tionally sourced, measurements of the mass distribution in proximity to the Milky
Way provide information on the local peculiar velocity field. One can therefore use
the observed distribution of galaxies to ‘reconstruct’ the expected peculiar velocity
field.
Given a reconstructed velocity field with line-of-sight velocity .vpec
recon
(→
r ) as a func-
tion of comoving position, the velocity corrections for a given galaxy can be derived
by reference to Eq. 18.5. By substitution, with observed redshift in the CMB frame
. z obs and sky position .n̂ given as a unit vector this is evaluated as

1 + z obs
.(1 + z cos ) = , (18.11)
1 + vpec
recon (χ(z )n̂)/c
cos

although some works use the approximated form

z
. cos ≈ z obs − vpec
recon
(χ(z obs )n̂)/c (18.12)

where .χ(z) is the comoving distance as a function of redshift. This approximation

leaves out a term of .z pec z cos . Another common approximation is the use of .χ(z obs )
rather than .χ(z cos ) to determine the position .r→. As .vpec
recon
is a function of physical
position, the appropriate redshift to use here is the one corresponding to the inferred
cosmological distance, rather than the observed quantity. This can most easily be
corrected for by resampling the reconstruction into redshift-space instead of physical
space [30]. However, as discussed in Sect. 18.4.3, the complex structure of these maps
near large mass concentrations, as well as scatter in redshifts caused by small-scale
noise may require more sophisticated procedures.
The simplest peculiar velocity model may be the “VGAS” model used in the
Hubble Key Project [52], which has seen subsequent use by the the CCHP [41,
53]. This models the local peculiar velocity field as sourced by the three largest
concentrations of mass in the nearby universe, Virgo, Shapley, and the Great Attractor.
This model has the advantage of simplicity and explainability, and is likely to reduce
the overall systematics by some amount. However these objects combined account
for less than a part per billion of the mass within 200 Mpc of the Milky Way. Further,
an analysis of the associated errors/uncorrected motions is difficult to carry out due
to the ad-hoc construction of the model.
Researchers have also created full reconstructions of the local volume, which
provide an estimated description of all mass within a defined volume at resolutions
between 1 and 10 Mpc .h −1 . Uncertainty at scales larger than those covered by the
underlying data is represented by a fitted external bulk flow parameter, constrained by
both the redshift catalog as well as sometimes additional SN Ia data. The methodology
underpinning these reconstructions varies widely, from filter-based approaches which
convolve a defined kernel with observed data [49] to Bayesian inference of initial
conditions combined with a forward model [48]. However in many cases significant
portions of the underlying data are shared between these different teams, in particular
352 W. D’Arcy Kenworthy and T. M. Davis

the 2MASS Redshift Survey [54, 55]. The most widely used of these among local
measurements of . H0 is the “2M++ model” [46], a reconstruction primarily based
on the 2M++ catalog using data from the 2MASS Redshift Survey [54, 55] and the
6 Degree Field Survey [56]. The 2M++ model has been used in . H0 measurements
including those made with the standard siren GW170817 [12], Type II SNe [57], by
the S. H0 ES project [2, 58], and the Megamaser Cosmology Project [43]. Redshift
survey based approaches have provided some estimates of their uncertainties, around
150 km s−1 in variance. However these estimates are typically provided as either
general quantities or only referred to on plots, rather than made available as data
files.
Cosmicflows [59–62] is an effort to incorporate as many distance measurements
as possible into a single model of the nearby universe. The most recent results
include more than 55,000 distances, from the Tully-Fisher relation, the fundamental
plane, surface brightness fluctuations, core-collapse supernovae, type Ia supernovae,
masers, Cepheids, and the tip of the red giant branch. The methodology used to pro-
cess this data has evolved from a Wiener-filter based approach to more recent efforts
using Hamiltonian Monte-Carlo. However the heterogeneity of the underlying data
means that associated uncertainties are highly unclear.
The Pantheon+ and S. H0 ES team have analyzed the effectiveness of peculiar veloc-
ity corrections in their application to the SN Ia distance ladder. The Pantheon+ analy-
sis [50] examined the data primarily through looking at diagonal scatter of SN Ia data
on the Hubble diagram. Their results found strong evidence that the use of 2M++
[46] and 2MRS [49] reconstructions substantively improved low-redshift scatter on
the Hubble diagram. At .10 σ significance these two reconstructions were favored
as compared to uncorrected CMB frame redshifts. However the Pantheon+ team
found that there was no evidence use of corrections based on Cosmicflows-3 or the
VGAS model were effective. In the final covariance matrix for S. H0 ES/Pantheon+,
a covariant term based on the differences between the two favored reconstructions
was included. Further work by the S. H0 ES team [51] attempted to empirically quan-
tify the degree of accuracy of the 2M++ and 2MRS models in reducing correlated
noise. Cepheid observations from the S. H0 ES team confirmed the accuracy of these
two reconstructions at .∼ 3σ, and measured a factor of .∼ 4 reduction in covariant
uncertainties in the linear regime. These results are qualitatively consistent with the
estimated uncertainties presented by the reconstruction teams. Estimated errors have
been included in the S. H0 ES error budget, and are subdominant in the final measure-
ment, at the level of .∼ 0.25%.

18.4.3 Mathematical Treatment of Reconstructions

A direct substitution of observed redshift and reconstructed velocity along a given line
of sight into Eq. 18.11 to solve for cosmological redshift is the most straightforward
approach to calculating PV corrections. However, since the PV reconstructions are
not perfectly precise, the problem of determining corrections is a statistical problem.
18 Impact of Peculiar Velocities on Measurements of . H0 353

Fig. 18.3 Cartoon example of a triply-valued region. Near a dense cluster, the velocity as a function
of distance takes on a characteristic Z-shaped curve which flattens the distance-redshift relation.
This velocity profile is shown in the second panel. As a result, an object’s observed redshift is fully
consistent with any of three positions; the middle of the cluster, the near side, or the far side of it.
Intuitively, this means that objects near clusters have highly uncertain true distances

The scatter in redshifts sourced by small scale motions and the uneven shape of
the distance-redshift relation in the presence of inhomogeneity can bias corrected
redshifts obtained from direct substitution. In extreme cases near galaxy clusters, the
distance-redshift relation is bent into “triply-valued regions”, where a single observed
redshift is consistent with three underlying cosmological redshifts. An illustration of
the phenomenon is shown in Fig. 18.3.
True triply valued regions are rare, but the principle of curvature in the Hub-
ble diagram due to inhomogeneity is generally applicable. Due to scatter of the
observed redshifts by small-scale noise, estimation of the cosmological redshift given
an observed redshift is nontrivial, requiring either a Bayesian or frequentist analysis
of the likelihood. A simple example of this might be based on a likelihood function
for the observed redshift, given the cosmological redshift. Such a likelihood could
be estimated by comparing the predicted peculiar redshift from the reconstruction,
to the peculiar redshift expected from the ratio of the observed and cosmological
.(1 + z) factors,

( ( ) )
1 1 + z obs 2
. P(z obs |z cos ) = √ exp − (1 + vpec (χ(z cos )n̂)/c) − /(2σv ) .
recon 2
2πσv2 1 + z cos
(18.13)
where .σv represents uncertainty due to small-scale scatter, between .150 and
250 km s−1 depending on sample, use of group corrections, or other variables. As
.vpec
recon
is not a linear function, this likelihood function is not Gaussian as a function
of .z cos . A simple prior on .z cos may assume that the sample is a galaxy randomly
selected from a given volume
354 W. D’Arcy Kenworthy and T. M. Davis

dχ(z)
. P(z cos ) ∝ χ(z cos )2 · (1 + δ(z cos )) |z=zcos (18.14)
dz

where .δ is the matter density contrast in the redshift-space reconstruction and .χ(z)
is the comoving distance as a function of redshift. This prior assumes the sample is
proportional to the volume element times the relative number density of galaxies,
accounting for the inhomogeneous Malmquist bias [63]. However selection effects
may require the use of different priors or more complicated modeling. If this simple
model is appropriate, it is simple to sample from it using an MCMC, or even evalu-
ate it directly by one-dimensional integration over the line of sight. When correctly
accounted, each point will have different uncertainties based on the local environ-
ment, and the expectation of the cosmological redshift may differ from the maxi-
mum likelihood value found by direct substitution into Eq. 18.11. The Pantheon+
team determined that estimates integrated using this methodology are more accurate
[50], although ultimately the simpler maximum likelihood method was used in the
final analysis. Other selections of data may impose other considerations and should
be considered by groups making use of peculiar velocity corrections. These line-of-
sight statistical effects have been accounted for by some [14, 51, 64, 65], but not all,
works in cosmography.

18.4.4 Group Assignments

Galaxies can be determined to belong to gravitationally bound groups based on the

use of redshift catalogs. As the mean redshift of a group averages over virialized
motion within the group, the use of group redshifts should remove noise below the
resolution of full-scale reconstructions.
The Pantheon+ team [50] tested four sets of group assignments [66–69], evaluated
by their effect on the Hubble scatter of the low-redshift Pantheon+ sample. All four
were favored by the data at .∼ 8σ significance, although results by Tully [67] were
favored compared to the others by .∼ 4σ. Use of any one of these group catalogs is
likely to significantly improve precision in Hubble diagram redshifts. Group assign-
ments have been used in several more recent measurements of the Hubble constant
[2, 12, 13, 43].

18.5 Conclusions and Future Outlook

Ultimately, peculiar velocity uncertainties are significantly unlikely to be responsible

for the Hubble tension. However, correctly understanding these effects is critical to
ensure our measurements are robust. Impact on the full S. H0 ES three rung ladder
is estimated at the level of .∼ 0.25% from the released covariant uncertainties (see
Fig. 18.1). Evaluation of the theoretical maximum contribution under .ΛCDM based
18 Impact of Peculiar Velocities on Measurements of . H0 355

on the measured power spectrum yields a maximum contribution of .∼ 0.8%. In the

context of a more than 10% discrepancy, these uncertainties are subdominant. When
considering competing methods restricted to smaller numbers of objects at smaller
distances, the role of peculiar velocities becomes much more critical. Use of redshifts
based on galaxy groups, corrected for bulk motions by the use of peculiar velocity
reconstructions, is significantly favored by SN Ia data.
Existing peculiar velocity reconstructions are limited in the distances which they
can probe, as the completeness of most spectroscopic surveys methods limit the
distances at which high-resolution models can be efficiently determined beyond
. z < 0.05. Future results by the DESI [70] and Euclid [71] instruments may greatly

increase the distance over which such reconstructions can be made [72, 73]. However,
the impact of peculiar velocities is typically drowned out by distance uncertainties at
these redshifts regardless, and the most precise measurements of the Hubble constant
are unlikely to be limited by peculiar velocities at any point in the near future. Nev-
ertheless, improvements in modeling may significantly improve some probes, par-
ticularly the two rung Cepheid distance ladder [51], direct geometric measurements
from nuclear megamasers [43], and gravitational wave standard sirens [13–15].
The utility of these reconstructions may be limited by the distribution of best-
fit reconstructions, without public estimates of uncertainties at the voxel level. The
methodology of the 2MRS reconstruction [49] created multiple realizations of the
reconstruction, although final data products are published only as a single estimate of
the density distribution. However their code is publicly available, and estimated vari-
ances as a function of radial position are available from their paper. Fully Bayesian
approaches like BORG [48] offer fully quantified uncertainties with full covari-
ance, but neither the BORG code nor data products have been publicly released. We
strongly encourage the production of voxel-level and/or covariant uncertainties in
future results. Scaling relations imply distribution of these data products may impose
significant storage and/or computational costs, but release of multiple samples or
low-ranked matrix approximations may be a viable approach.
Future work on extremely low-redshift measurements of the Hubble constant
should focus on consistent modeling of peculiar velocities as an integrated component
of the measurement process.

Acknowledgements W. D’Arcy Kenworthy thanks the Knut and Alice Wallenberg Foundation
for their support from the research project grant ‘Understanding the Dynamic Universe’ under
Dnr KAW 2018.0067. T. M. Davis acknowledges the support of an Australian Research Council
Australian Laureate Fellowship (FL180100168) funded by the Australian Government. The authors
thank Khaled Said for useful comments.
356 W. D’Arcy Kenworthy and T. M. Davis

References

1. A.G. Riess, L.M. Macri, S.L. Hoffmann, D. Scolnic, S. Casertano, A.V. Filippenko, B.E.
Tucker, M.J. Reid, D.O. Jones, J.M. Silverman, R. Chornock, P. Challis, W. Yuan, P.J. Brown,
R.J. Foley. A 2.4% determination of the local value of the Hubble constant. APJ 826(1), 56
(2016)
2. A.G. Riess, W. Yuan, L.M. Macri, D. Scolnic, D. Brout, S. Casertano, D.O. Jones, Y. Murakami,
G.S. Anand, L. Breuval, T.G. Brink, A.V. Filippenko, S. Hoffmann, S.W. Jha, W. D’arcy
Kenworthy, J. Mackenty, B.E. Stahl, W. Zheng, A comprehensive measurement of the local
value of the Hubble constant with 1 km s.−1 Mpc.−1 uncertainty from the Hubble space telescope
and the S. H0 ES team. ApJL 934(1), L7 (2022)
3. Tamara M. Davis, Morag I. Scrimgeour, Deriving accurate peculiar velocities (even at high
redshift). MNRAS 442(2), 1117–1122 (2014). (August)
4. E.R. Harrison, Interpretation of redshifts of galaxies in clusters. ApJL 191, L51 (1974). (July)
5. Lam Hui, Patrick B. Greene, Correlated fluctuations in luminosity distance and the importance
of peculiar motion in supernova surveys. PhRvD 73(12), 123526 (2006). (June)
6. G.F.R. Ellis, J.E. Baldwin, On the expected anisotropy of radio source counts. MNRAS 206,
377–381 (1984). (January)
7. Tobias Nadolny, Ruth Durrer, Martin Kunz, Hamsa Padmanabhan, A new way to test the cosmo-
logical principle: measuring our peculiar velocity and the large-scale anisotropy independently.
JCAP 2021(11), 009 (2021). (November)
8. M. Rubart, D.J. Schwarz, Cosmic radio dipole from NVSS and WENSS. A&A 555, A117
(2013). (July)
9. J.D. Wagenveld, H.R. Klöckner, D.J. Schwarz, The cosmic radio dipole: Bayesian estimators
on new and old radio surveys. A&A 675, A72 (2023). (July)
10. Planck Collaboration, N. Aghanim, C. Armitage-Caplan, M. Arnaud, M. Ashdown, F. Atrio-
Barandela, J. Aumont, C. Baccigalupi, A.J. Banday, R.B. Barreiro, J.G. Bartlett, K. Benabed,
A. Benoit-Lévy, J.P. Bernard, M. Bersanelli, P. Bielewicz, J. Bobin, J.J. Bock, J.R. Bond, J. Bor-
rill, F.R. Bouchet, M. Bridges, C. Burigana, R.C. Butler, J.F. Cardoso, A. Catalano, A. Challinor,
A. Chamballu, H.C. Chiang, L.Y. Chiang, P.R. Christensen, D.L. Clements, L.P.L. Colombo,
F. Couchot, B.P. Crill, A. Curto, F. Cuttaia, L. Danese, R.D. Davies, R.J. Davis, P. de Bernardis,
A. de Rosa, G. de Zotti, J. Delabrouille, J.M. Diego, S. Donzelli, O. Doré, X. Dupac, G. Efs-
tathiou, T.A. Enßlin, H.K. Eriksen, F. Finelli, O. Forni, M. Frailis, E. Franceschi, S. Galeotta,
K. Ganga, M. Giard, G. Giardino, J. González-Nuevo, K.M. Górski, S. Gratton, A. Gregorio,
A. Gruppuso, F.K. Hansen, D. Hanson, D.L. Harrison, G. Helou, S.R. Hildebrandt, E. Hivon,
M. Hobson, W.A. Holmes, W. Hovest, K.M. Huffenberger, W.C. Jones, M. Juvela, E. Keihänen,
R. Keskitalo, T.S. Kisner, J. Knoche, L. Knox, M. Kunz, H. Kurki-Suonio, A. Lähteenmäki,
J.M. Lamarre, A. Lasenby, R.J. Laureijs, C.R. Lawrence, R. Leonardi, A. Lewis, M. Liguori,
P.B. Lilje, M. Linden-Vørnle, M. López-Caniego, P.M. Lubin, J.F. Macías-Pérez, N. Mandolesi,
M. Maris, D.J. Marshall, P.G. Martin, E. Martínez-González, S. Masi, M. Massardi, S. Matar-
rese, P. Mazzotta, P.R. Meinhold, A. Melchiorri, L. Mendes, M. Migliaccio, S. Mitra, A. Moneti,
L. Montier, G. Morgante, D. Mortlock, A. Moss, D. Munshi, P. Naselsky, F. Nati, P. Natoli, H.U.
Nørgaard-Nielsen, F. Noviello, D. Novikov, I. Novikov, S. Osborne, C.A. Oxborrow, L. Pagano,
F. Pajot, D. Paoletti, F. Pasian, G. Patanchon, O. Perdereau, F. Perrotta, F. Piacentini, E. Pier-
paoli, D. Pietrobon, S. Plaszczynski, E. Pointecouteau, G. Polenta, N. Ponthieu, L. Popa, G.W.
Pratt, G. Prézeau, J.L. Puget, J.P. Rachen, W.T. Reach, M. Reinecke, S. Ricciardi, T. Riller,
I. Ristorcelli, G. Rocha, C. Rosset, J.A. Rubiño-Martín, B. Rusholme, D. Santos, G. Savini,
D. Scott, M.D. Seiffert, E.P.S. Shellard, L.D. Spencer, R. Sunyaev, F. Sureau, A.S. Suur-Uski,
J.F. Sygnet, J.A. Tauber, D. Tavagnacco, L. Terenzi, L. Toffolatti, M. Tomasi, M. Tristram,
M. Tucci, M. Türler, L. Valenziano, J. Valiviita, B. Van Tent, P. Vielva, F. Villa, N. Vittorio,
L.A. Wade, B.D. Wandelt, M. White, D. Yvon, A. Zacchei, J.P. Zibin, A. Zonca. Planck 2013
results. XXVII. doppler boosting of the CMB: Eppur si muove. A&A 57, A27 (2014)
11. Planck Collaboration, Y. Akrami, M. Ashdown, J. Aumont, C. Baccigalupi, M. Ballardini,
A.J. Banday, R.B. Barreiro, N. Bartolo, S. Basak, K. Benabed, J.P. Bernard, M. Bersanelli,
18 Impact of Peculiar Velocities on Measurements of . H0 357

P. Bielewicz, J.R. Bond, J. Borrill, F.R. Bouchet, C. Burigana, E. Calabrese, J.F. Cardoso,
B. Casaponsa, H.C. Chiang, C. Combet, D. Contreras, B.P. Crill, F. Cuttaia, P. de Bernardis,
A. de Rosa, G. de Zotti, J. Delabrouille, E. Di Valentino, J.M. Diego, O. Doré, M. Douspis,
X. Dupac, T.A. Enßlin, H.K. Eriksen, R. Fernandez-Cobos, F. Finelli, M. Frailis, E. Franceschi,
A. Frolov, S. Galeotta, S. Galli, K. Ganga, R.T. Génova-Santos, M. Gerbino, J. González-
Nuevo, K.M. Górski, A. Gruppuso, J.E. Gudmundsson, W. Handley, D. Herranz, E. Hivon,
Z. Huang, A.H. Jaffe, W.C. Jones, E. Keihänen, R. Keskitalo, K. Kiiveri, J. Kim, T.S. Kisner,
N. Krachmalnicoff, M. Kunz, H. Kurki-Suonio, J.M. Lamarre, M. Lattanzi, C.R. Lawrence,
M. Le Jeune, F. Levrier, M. Liguori, P.B. Lilje, V. Lindholm, M. López-Caniego, J.F. Macías-
Pérez, D. Maino, N. Mandolesi, A. Marcos-Caballero, M. Maris, P.G. Martin, E. Martínez-
González, S. Matarrese, N. Mauri, J.D. McEwen, A. Mennella, M. Migliaccio, D. Molinari,
A. Moneti, L. Montier, G. Morgante, A. Moss, P. Natoli, L. Pagano, D. Paoletti, F. Perrotta,
V. Pettorino, F. Piacentini, G. Polenta, J.P. Rachen, M. Reinecke, M. Remazeilles, A. Renzi,
G. Roha, C. Roset, J.A. Rubiño-Martín, B. Ruiz-Granados, L. Salvati, M. Savelainen, D. Scott,
C. Sirignano, G. Sirri, L.D. Spencer, R.M. Sullivan, R. Sunyaev, A.S. Suur-Uski, J.A. Tauber,
D. Tavagnacco, M. Tenti, L. Toffolatti, M. Tomasi, T. Trombetti, J. Valiviita, B. Van Tent,
P. Vielva, F. Villa, N. Vittorio, I.K. Wehus, A. Zacchei, A. Zonca, Planck intermediate results.
LVI. Detection of the CMB dipole through modulation of the thermal Sunyaev-Zeldovich
effect: Eppur si muove II. A&A 644, A100 (2020)
12. B.P. Abbott, R. Abbott, T.D. Abbott, F. Acernese, K. Ackley, C. Adams, T. Adams, P. Addesso,
R.X. Adhikari, V.B. Adya, C. Affeldt, M. Afrough, B. Agarwal, M. Agathos, K. Agatsuma,
N. Aggarwal, O.D. Aguiar, L. Aiello, A. Ain, P. Ajith, B. Allen, G. Allen, A. Allocca, P.A.
Altin, A. Amato, A. Ananyeva, S.B. Anderson, W.G. Anderson, S.V. Angelova, S. Antier, S.
Appert, K. Arai, M.C. Araya, J.S. Areeda, N. Arnaud, K.G. Arun, S. Ascenzi, G. Ashton, M.
Ast, S.M. Aston, P. Astone, D.V. Atallah, P. Aufmuth, C. Aulbert, K. Aultoneal, C. Austin,
A. Avila-Alvarez, S. Babak, P. Bacon, M.K.M. Bader, S. Bae, P.T. Baker, F. Baldaccini, G.
Ballardin, S.W. Ballmer, S. Banagiri, J.C. Barayoga, S.E. Barclay, B.C. Barish, D. Barker, K.
Barkett, F. Barone, B. Barr, L. Barsotti, M. Barsuglia, D. Barta, J. Bartlett, I. Bartos, R. Bassiri,
A. Basti, J.C. Batch, A gravitational-wave standard siren measurement of the Hubble constant.
Nature 551(7678), 85–88 (2017). (November)
13. Cullan Howlett, Tamara M. Davis, Standard siren speeds: improving velocities in gravitational-
wave measurements of H.0 . MNRAS 492(3), 3803–3815 (2020). (March)
14. Suvodip Mukherjee, Guilhem Lavaux, François R. Bouchet, Jens Jasche, Benjamin D. Wandelt,
Samaya Nissanke, Florent Leclercq, Kenta Hotokezaka, Velocity correction for Hubble constant
measurements from standard sirens. A&A 646, A65 (2021). (February)
15. Constantina Nicolaou, Ofer Lahav, Pablo Lemos, William Hartley, Jonathan Braden, The
impact of peculiar velocities on the estimation of the Hubble constant from gravitational wave
standard sirens. MNRAS 495(1), 90–97 (2020). (June)
16. Idit Zehavi, Adam G. Riess, Robert P. Kirshner, Avishai Dekel, A local Hubble bubble from
type Ia supernovae? APJ 503(2), 483–491 (1998). (August)
17. Saurabh Jha, Adam G. Riess, Robert P. Kirshner, Improved distances to type Ia supernovae
with multicolor light-curve shapes: MLCS2k2. APJ 659(1), 122–148 (2007). (April)
18. A. Conley, R.G. Carlberg, J. Guy, D.A. Howell, S. Jha, A.G. Riess, M. Sullivan, Is there
evidence for a Hubble bubble? The nature of type Ia supernova colors and dust in external
galaxies. ApJL 664(1), L13–L16 (2007). (July)
19. Benjamin Sinclair, Tamara M. Davis, Troels Haugbølle, Residual Hubble-bubble effects on
supernova cosmology. APJ 718(2), 1445–1455 (2010). (August)
20. W. D’Arcy Kenworthy, D. Scolnic, A. Riess, The local perspective on the Hubble tension: local
structure does not impact measurement of the Hubble constant. APJ 875(2), 145 (2019)
21. Sveva Castello, Marcus Högås, Edvard Mörtsell, A cosmological underdensity does not solve
the Hubble tension. JCAP 2022(7), 003 (2022). (July)
22. R. Brent Tully, Hélène Courtois, Y. Hoffman, D. Pomarède, The Laniakea supercluster of
galaxies. Nature 513(7516), 71–73 (2014)
358 W. D’Arcy Kenworthy and T. M. Davis

23. A. Dupuy, H.M. Courtois, Dynamic cosmography of the local Universe: Laniakea and five
more watershed superclusters (2023). arXiv e-prints arXiv:2305.02339
24. Wu. Hao-Yi, Dragan Huterer, Sample variance in the local measurements of the Hubble con-
stant. MNRAS 471(4), 4946–4955 (2017). (November)
25. Planck Collaboration, N. Aghanim, Y. Akrami, F. Arroja, M. Ashdown, J. Aumont, C. Bac-
cigalupi, M. Ballardini, A.J. Banday, R.B. Barreiro, N. Bartolo, S. Basak, R. Battye, K. Ben-
abed, J.P. Bernard, M. Bersanelli, P. Bielewicz, J.J. Bock, J.R. Bond, J. Borrill, F.R. Bouchet,
F. Boulanger, M. Bucher, C. Burigana, R.C. Butler, E. Calabrese, J.F. Cardoso, J. Carron,
B. Casaponsa, A. Challinor, H.C. Chiang, L.P.L. Colombo, C. Combet, D. Contreras, B.P.
Crill, F. Cuttaia, P. de Bernardis, G. de Zotti, J. Delabrouille, J.M. Delouis, F.X. Désert, E. Di
Valentino, C. Dickinson, J.M. Diego, S. Donzelli, O. Doré, M. Douspis, A. Ducout, X. Dupac,
G. Efstathiou, F. Elsner, T.A. Enßlin, H.K. Eriksen, E. Falgarone, Y. Fantaye, J. Fergusson,
R. Fernandez-Cobos, F. Finelli, F. Forastieri, M. Frailis, E. Franceschi, A. Frolov, S. Gale-
otta, S. Galli, K. Ganga, R.T. Génova-Santos, M. Gerbino, T. Ghosh, J. González-Nuevo,
K.M. Górski, S. Gratton, A. Gruppuso, J.E. Gudmundsson, J. Hamann, W. Handley, F.K.
Hansen, G. Helou, D. Herranz, S.R. Hildebrandt, E. Hivon, Z. Huang, A.H. Jaffe, W.C. Jones,
A. Karakci, E. Keihänen, R. Keskitalo, K. Kiiveri, J. Kim, T.S. Kisner, L. Knox, N. Krachmal-
nicoff, M. Kunz, H. Kurki-Suonio, G. Lagache, J.M. Lamarre, M. Langer, A. Lasenby, M. Lat-
tanzi, C.R. Lawrence, M. Le Jeune, J.P. Leahy, J. Lesgourgues, F. Levrier, A. Lewis, M. Liguori,
P.B. Lilje, M. Lilley, V. Lindholm, M. López-Caniego, P.M. Lubin, Y.Z. Ma, J.F. Macías-Pérez,
G. Maggio, D. Maino, N. Mandolesi, A. Mangilli, A. Marcos-Caballero, M. Maris, P.G. Martin,
M. Martinelli, E. Martínez-González, S. Matarrese, N. Mauri, J.D. McEwen, P.D. Meerburg,
P.R. Meinhold, A. Melchiorri, A. Mennella, M. Migliaccio, M. Millea, S. Mitra, M.A. Miville-
Deschênes, D. Molinari, A. Moneti, L. Montier, G. Morgante, A. Moss, S. Mottet, M. Münch-
meyer, P. Natoli, H.U. Nørgaard-Nielsen, C.A. Oxborrow, L. Pagano, D. Paoletti, B. Partridge,
G. Patanchon, T.J. Pearson, M. Peel, H.V. Peiris, F. Perrotta, V. Pettorino, F. Piacentini, L. Polas-
tri, G. Polenta, J.L. Puget, J.P. Rachen, M. Reinecke, M. Remazeilles, C. Renault, A. Renzi,
G. Rocha, C. Rosset, G. Roudier, J.A. Rubiño-Martín, B. Ruiz-Granados, L. Salvati, M. Sandri,
M. Savelainen, D. Scott, E.P.S. Shellard, M. Shiraishi, C. Sirignano, G. Sirri, L.D. Spencer,
R. Sunyaev, A.S. Suur-Uski, J.A. Tauber, D. Tavagnacco, M. Tenti, L. Terenzi, L. Toffolatti,
M. Tomasi, T. Trombetti, J. Valiviita, B. Van Tent, L. Vibert, P. Vielva, F. Villa, N. Vittorio, B.D.
Wandelt, I.K. Wehus, M. White, S.D.M. White, A. Zacchei, A. Zonca, Planck 2018 results. I.
Overview and the cosmological legacy of Planck. A&A 641, A1 (2020)
26. Chris Blake, Jasper Wall, A velocity dipole in the distribution of radio galaxies. Nature
416(6877), 150–152 (2002). (March)
27. U. Feindt, M. Kerschhaggl, M. Kowalski, G. Aldering, P. Antilogus, C. Aragon, S. Bailey, C.
Baltay, S. Bongard, C. Buton, A. Canto, F. Cellier-Holzem, M. Childress, N. Chotard, Y. Copin,
H.K. Fakhouri, E. Gangler, J. Guy, A. Kim, P. Nugent, J. Nordin, K. Paech, R. Pain, E. Pecontal,
R. Pereira, S. Perlmutter, D. Rabinowitz, M. Rigault, K. Runge, C. Saunders, R. Scalzo, G.
Smadja, C. Tao, R.C. Thomas, B.A. Weaver, C. Wu, Measuring cosmic bulk flows with type
Ia supernovae from the nearby supernova factory. A&A 560, A90 (2013). (December)
28. Nick Horstmann, Yannic Pietschke, Dominik J. Schwarz, Inference of the cosmic rest-frame
from supernovae Ia. A&A 668, A34 (2022). (December)
29. Tamara M. Davis, Samuel R. Hinton, Cullan Howlett, Josh Calcino, Can redshift errors bias
measurements of the Hubble constant? MNRAS 490(2), 2948–2957 (2019). (December)
30. Anthony Carr, Tamara M. Davis, Dan Scolnic, Khaled Said, Dillon Brout, Erik R. Peterson,
Richard Kessler, The Pantheon+ analysis: improving the redshifts and peculiar velocities of
type Ia supernovae used in cosmological analyses. PASA 39, e046 (2022). (October)
31. Richard Watkins, Hume A. Feldman, Michael J. Hudson, Consistently large cosmic flows
on scales of 100h.−1 Mpc: a challenge for the standard .ΛCDM cosmology. MNRAS 392(2),
743–756 (2009). (January)
32. K. Migkas, F. Pacaud, G. Schellenberger, J. Erler, N.T. Nguyen-Dang, T.H. Reiprich, M.E.
Ramos-Ceja, L. Lovisari, Cosmological implications of the anisotropy of ten galaxy cluster
scaling relations. A&A 649, A151 (2021). (May)
18 Impact of Peculiar Velocities on Measurements of . H0 359

33. H.M. Courtois, A. Dupuy, D. Guinet, G. Baulieu, F. Ruppin, P. Brenas, Gravity in the local
universe: density and velocity fields using CosmicFlows-4. A&A 670, L15 (2023). (February)
34. R. Watkins, T. Allen, C.J. Bradford, A. Ramon, A. Walker, H.A. Feldman, R. Cionitti, Y. Al-
Shorman, E. Kourkchi, R.B. Tully, Analysing the large-scale bulk flow using cosmicflows4:
increasing tension with the standard cosmological model. MNRAS 524(2), 1885–1892 (2023)
35. A.M. Whitford, C. Howlett, T.M. Davis, Evaluating bulk flow estimators for CosmicFlows-4
measurements (2023). arXiv e-prints arXiv:2306.11269
36. D. Brout, D. Scolnic, B. Popovic, A.G. Riess, A. Carr, J. Zuntz, R. Kessler, T.M. Davis,
S. Hinton, D. Jones, W. D’Arcy Kenworthy, E.R. Peterson, K. Said, G. Taylor, N. Ali, P.
Armstrong, P. Charvu, A. Dwomoh, C. Meldorf, A. Palmese, H. Qu, B.M. Rose, B. Sanchez,
C.W. Stubbs, M. Vincenzi, C.M. Wood, P.J. Brown, R. Chen, K. Chambers, D.A. Coulter,
M. Dai, G. Dimitriadis, A.V. Filippenko, R.J. Foley, S.W. Jha, L. Kelsey, R.P. Kirshner, A.
Möller, J. Muir, S. Nadathur, Y.-C. Pan, A. Rest, C. Rojas-Bravo, M. Sako, M.R. Siebert, M.
Smith, B.E. Stahl, P. Wiseman, The Pantheon+ analysis: cosmological constraints. APJ 938(2),
110 (2022)
37. Planck Collaboration, N. Aghanim, Y. Akrami, M. Ashdown, J. Aumont, C. Baccigalupi,
M. Ballardini, A.J. Banday, R.B. Barreiro, N. Bartolo, S. Basak, R. Battye, K. Benabed, J.P.
Bernard, M. Bersanelli, P. Bielewicz, J.J. Bock, J.R. Bond, J. Borrill, F.R. Bouchet, F. Boulanger,
M. Bucher, C. Burigana, R.C. Butler, E. Calabrese, J.F. Cardoso, J. Carron, A. Challinor, H.C.
Chiang, J. Chluba, L.P.L. Colombo, C. Combet, D. Contreras, B.P. Crill, F. Cuttaia, P. de
Bernardis, G. de Zotti, J. Delabrouille, J.M. Delouis, E. Di Valentino, J.M. Diego, O. Doré,
M. Douspis, A. Ducout, X. Dupac, S. Dusini, G. Efstathiou, F. Elsner, T.A. Enßlin, H.K.
Eriksen, Y. Fantaye, M. Farhang, J. Fergusson, R. Fernandez-Cobos, F. Finelli, F. Forastieri,
M. Frailis, A.A. Fraisse, E. Franceschi, A. Frolov, S. Galeotta, S. Galli, K. Ganga, R.T. Génova-
Santos, M. Gerbino, T. Ghosh, J. González-Nuevo, K.M. Górski, S. Gratton, A. Gruppuso, J.E.
Gudmundsson, J. Hamann, W. Handley, F.K. Hansen, D. Herranz, S.R. Hildebrandt, E. Hivon,
Z. Huang, A.H. Jaffe, W.C. Jones, A. Karakci, E. Keihänen, R. Keskitalo, K. Kiiveri, J. Kim, T.S.
Kisner, L. Knox, N. Krachmalnicoff, M. Kunz, H. Kurki-Suonio, G. Lagache, J.M. Lamarre,
A. Lasenby, M. Lattanzi, C.R. Lawrence, M. Le Jeune, P. Lemos, J. Lesgourgues, F. Levrier,
A. Lewis, M. Liguori, P.B. Lilje, M. Lilley, V. Lindholm, M.López-Caniego, P.M. Lubin, Y.Z.
Ma, J.F. Macías-Pérez, G. Maggio, D. Maino, N. Mandolesi, A. Mangilli, A. Marcos-Caballero,
M. Maris, P.G. Martin, M. Martinelli, E. Martínez-González, S. Matarrese, N. Mauri, J.D.
McEwen, P.R. Meinhold, A. Melchiorri, A. Mennella, M. Migliaccio, M. Millea, S. Mitra, M.A.
Miville-Deschênes, D. Molinari, L. Montier, G. Morgante, A. Moss, P. Natoli, H.U. Nørgaard-
Nielsen, L. Pagano, D. Paoletti, B. Partridge, G. Patanchon, H.V. Peiris, F. Perrotta, V. Pettorino,
F. Piacentini, L. Polastri, G. Polenta, J.L. Puget, J.P. Rachen, M. Reinecke, M. Remazeilles,
A. Renzi, G. Rocha, C. Rosset, G. Roudier, J.A. Rubiño-Martín, B. Ruiz-Granados, L. Salvati,
M. Sandri, M. Savelainen, D. Scott, E.P.S. Shellard, C. Sirignano, G. Sirri, L.D. Spencer,
R. Sunyaev, A.S. Suur-Uski, J.A. Tauber, D. Tavagnacco, M. Tenti, L. Toffolatti, M. Tomasi,
T. Trombetti, L. Valenziano, J. Valiviita, B. Van Tent, L. Vibert, P. Vielva, F. Villa, N. Vittorio,
B.D. Wandelt, I.K. Wehus, M. White, S.D.M. White, A. Zacchei, A. Zonca, lanck 2018 results.
VI. Cosmological parameters. A&A 641, A6 (2020)
38. J.A. Peacock, R.E. Smith, HALOFIT: nonlinear distribution of cosmological mass and galaxies.
Astrophys. Source Code Library, record ascl:1402.032 (2014)
39. Tamara M. Davis, Lam Hui, Joshua A. Frieman, Troels Haugbølle, Richard Kessler, Benjamin
Sinclair, Jesper Sollerman, Bruce Bassett, John Marriner, Edvard Mörtsell, Robert C. Nichol,
Michael W. Richmond, Masao Sako, Donald P. Schneider, Mathew Smith, The effect of peculiar
velocities on supernova cosmology. APJ 741(1), 67 (2011). (November)
40. Dragan Huterer, Daniel L. Shafer, Fabian Schmidt, No evidence for bulk velocity from type Ia
supernovae. JCAP 2015(12), 033 (2015). (December)
41. W.L. Freedman, B.F. Madore, D. Hatt, T.J. Hoyt, I. Sung Jang, R.L. Beaton, C.R. Burns,
M. Gyoon Lee, A.J. Monson, J.R. Neeley, M.M. Phillips, J.A. Rich, M. Seibert, The Carnegie-
Chicago Hubble program. VIII. An independent determination of the Hubble constant based
on the tip of the red giant branch. APJ 882(1), 34 (2019)
360 W. D’Arcy Kenworthy and T. M. Davis

42. K. Krisciunas, C. Contreras, C.R. Burns, M.M. Phillips, M.D. Stritzinger, N. Morrell, M.
Hamuy, J. Anais, L. Boldt, L. Busta, A. Campillay, S. Castellón, G. Folatelli, W.L. Freedman,
C. González, E.Y. Hsiao, W. Krzeminski, S.E. Persson, M. Roth, F. Salgado, J. Serón, N.B.
Suntzeff, S. Torres, A.V. Filippenko, W. Li, B.F. Madore, D.L. DePoy, J.L. Marshall, J.-P.
Rheault, S. Villanueva, The Carnegie supernova project. I. Third photometry data release of
low-redshift type Ia supernovae and other white dwarf explosions. AJ 154(5), 211 (2017)
43. D.W. Pesce, J.A. Braatz, M.J. Reid, A.G. Riess, D. Scolnic, J.J. Condon, F. Gao, C. Henkel,
C.M.V. Impellizzeri, C.Y. Kuo, K.Y. Lo, The Megamaser cosmology project XIII. Combined
Hubble constant constraints. ApJL 891(1), L1 (2020). (March)
44. R.O. Marzke, M.J. Geller, L.N. da Costa, J.P. Huchra, Pairwise velocities of galaxies in the
CfA and SSRS2 redshift surveys. AJ 110, 477 (1995)
45. M. Betoule, R. Kessler, J. Guy, J. Mosher, D. Hardin, R. Biswas, P. Astier, P. El-Hage, M.
Konig, S. Kuhlmann, J. Marriner, R. Pain, N. Regnault, C. Balland, B.A. Bassett, P.J. Brown,
H. Campbell, R.G. Carlberg, F. Cellier-Holzem, D. Cinabro, A. Conley, C.B. D’Andrea, D.L.
DePoy, M. Doi, R.S. Ellis, S. Fabbro, A.V. Filippenko, R.J. Foley, J.A. Frieman, D. Fouchez,
L. Galbany, A. Goobar, R.R. Gupta, G.J. Hill, R. Hlozek, C.J. Hogan, I.M. Hook, D.A. Howell,
S.W. Jha, L. Le Guillou, G. Leloudas, C. Lidman, J.L. Marshall, A. Möller, A.M. Mourão,
J. Neveu, R. Nichol, M.D. Olmstead, N. Palanque-Delabrouille, S. Perlmutter, J.L. Prieto,
C.J. Pritchet, M. Richmond, A.G. Riess, V. Ruhlmann-Kleider, M. Sako, K. Schahmaneche,
D.P. Schneider, M. Smith, J. Sollerman, M. Sullivan, N.A. Walton, C.J. Wheeler, Improved
cosmological constraints from a joint analysis of the SDSS-II and SNLS supernova samples.
Astron. Astrophys. 568, A22 (2014). (Aug)
46. J. Carrick, S.J. Turnbull, G. Lavaux, M.J. Hudson, Cosmological parameters from the compar-
ison of peculiar velocities with predictions from the 2M++ density field. Monthly Notices of
the Royal Astron. Soc. 450(1), 317–332 (2015)
47. D.M. Scolnic, D.O. Jones, A. Rest, Y.C. Pan, R. Chornock, R.J. Foley, M.E. Huber, R. Kessler,
G. Narayan, A.G. Riess, S. Rodney, E. Berger, D.J. Brout, P.J. Challis, M. Drout, D. Finkbeiner,
R. Lunnan, R.P. Kirshner, N.E. Sanders, E. Schlafly, S. Smartt, C.W. Stubbs, J. Tonry, W.M.
Wood-Vasey, M. Foley, J. Hand, E. Johnson, W.S. Burgett, K.C. Chambers, P.W. Draper, K.W.
Hodapp, N. Kaiser, R.P. Kudritzki, E.A. Magnier, N. Metcalfe, F. Bresolin, E. Gall, R. Kotak, M.
McCrum, K.W. Smith, The complete light-curve sample of spectroscopically confirmed SNe Ia
from Pan-STARRS1 and cosmological constraints from the combined Pantheon sample. APJ
859(2), 101 (2018). (June)
48. J. Jasche, G. Lavaux, Physical Bayesian modelling of the non-linear matter distribution: new
insights into the nearby universe. A&A 625, A64 (2019). (May)
49. Robert Lilow, Adi Nusser, Constrained realizations of 2MRS density and peculiar velocity
fields: growth rate and local flow. MNRAS 507(2), 1557–1581 (2021). (October)
50. E.R. Peterson, W. D’Arcy Kenworthy, D. Scolnic, A.G. Riess, D. Brout, A. Carr, Hélène
Courtois, T. Davis, A. Dwomoh, D.O. Jones, B. Popovic, B.M. Rose, K. Said, The Pantheon+
analysis: evaluating peculiar velocity corrections in cosmological analyses with nearby type Ia
supernovae. APJ 938(2), 112 (2022)
51. W. D’Arcy Kenworthy, A.G. Riess, D. Scolnic, W. Yuan, J. Luis Bernal, D. Brout, S. Casertano,
D.O. Jones, L. Macri, E.R. Peterson, Measurements of the Hubble constant with a two-rung
distance ladder: two out of three ain’t bad. APJ 935(2), 83 (2022)
52. J.R. Mould, J.P. Huchra, W.L. Freedman, Jr. R.C. Kennicutt, L. Ferrarese, H.C. Ford, B.K. Gib-
son, J.A. Graham, S.M.G. Hughes, G.D. Illingworth, D.D. Kelson, L.M. Macri, B.F. Madore,
S. Sakai, K.M. Sebo, N.A. Silbermann, P.B. Stetson, The Hubble space telescope key project
on the extragalactic distance scale. XXVIII. Combining the constraints on the Hubble constant.
APJ 529(2), 786–794 (2000)
53. W.L. Freedman, B.F. Madore, B.K. Gibson, L. Ferrarese, D.D. Kelson, S. Sakai, J.R. Mould,
Jr. R.C. Kennicutt, H.C. Ford, J.A. Graham, J.P. Huchra, S.M.G. Hughes, G.D. Illingworth,
L.M. Macri, P.B. Stetson, Final results from the Hubble space telescope key project to measure
the Hubble constant. APJ 553(1), 47–72 (2001)
18 Impact of Peculiar Velocities on Measurements of . H0 361

54. John P. Huchra, Lucas M. Macri, Karen L. Masters, Thomas H. Jarrett, Perry Berlind, Michael
Calkins, Aidan C. Crook, Roc Cutri, Pirin Erdoǧdu, Emilio Falco, Teddy George, Conrad M.
Hutcheson, Ofer Lahav, Jeff Mader, Jessica D. Mink, Nathalie Martimbeau, Stephen Schneider,
Michael Skrutskie, Susan Tokarz, Michael Westover, The 2MASS redshift survey—description
and data release. ApJS 199(2), 26 (2012). (April)
55. L.M. Macri, R.C. Kraan-Korteweg, T. Lambert, M. Victoria Alonso, P. Berlind, M. Calkins,
P. Erdoğdu, E.E. Falco, T.H. Jarrett, J.D. Mink, The 2MASS redshift survey in the zone of
avoidance. ApJS 245(1), 6 (2019)
56. D. Heath Jones, M.A. Read, W. Saunders, M. Colless, T. Jarrett, Q.A. Parker, A.P. Fairall, T.
Mauch, E.M. Sadler, F.G. Watson, D. Burton, L.A. Campbell, P. Cass, S.M. Croom, J. Dawe,
K. Fiegert, L. Frankcombe, M. Hartley, J. Huchra, D. James, E. Kirby, O. Lahav, J. Lucey,
G.A. Mamon, L. Moore, B.A. Peterson, S. Prior, D. Proust, K. Russell, V. Safouris, K.-I.
Wakamatsu, E. Westra, M. Williams, The 6dF galaxy survey: final redshift release (DR3) and
southern large-scale structures. MNRAS 399(2), 683–698 (2009)
57. T. de Jaeger, L. Galbany, A.G. Riess, B.E. Stahl, B.J. Shappee, A.V. Filippenko, W. Zheng,
A 5 percent measurement of the Hubble-Lemaître constant from type II supernovae. MNRAS
514(3), 4620–4628 (2022). (August)
58. Adam G. Riess, Weidong Li, Peter B. Stetson, Alexei V. Filippenko, Saurabh Jha, Robert P.
Kirshner, Peter M. Challis, Peter M. Garnavich, Ryan Chornock, Cepheid calibrations from the
Hubble space telescope of the luminosity of two recent type Ia supernovae and a redetermination
of the Hubble constant. APJ 627(2), 579–607 (2005). (July)
59. H.M. Courtois, Y. Hoffman, R. Brent Tully, S. Gottlöber, Three-dimensional velocity and
density reconstructions of the local universe with Cosmicflows-1. APJ 744(1), 43 (2012)
60. R. Brent Tully, H.M. Courtois, A.E. Dolphin, J. Richard Fisher, P. Héraudeau, B.A. Jacobs,
I.D. Karachentsev, D. Makarov, L. Makarova, S. Mitronova, L. Rizzi, E.J. Shaya, J.G. Sorce,
P.-F. Wu, Cosmicflows-2: the data. AJ 146(4), 86 (2013)
61. R. Brent Tully, H.M. Courtois, J.G. Sorce. Cosmicflows-3. AJ 152(2), 50 (2016)
62. R. Brent Tully, E. Kourkchi, H.M. Courtois, G.S. Anand, J.P. Blakeslee, D. Brout, T. de Jaeger,
A. Dupuy, D. Guinet, C. Howlett, J.B. Jensen, D. Pomarède, L. Rizzi, D. Rubin, K. Said, D.
Scolnic, B.E. Stahl. Cosmicflows-4. APJ 944(1), 94 (2023)
63. R.W. Pike, M.J. Hudson, Cosmological parameters from the comparison of the 2MASS gravity
field with peculiar velocity surveys. APJ 635(1), 11–21 (2005)
64. Supranta S. Boruah, Michael J. Hudson, Guilhem Lavaux, Peculiar velocities in the local
universe: comparison of different models and the implications for H.0 and dark matter. MNRAS
507(2), 2697–2713 (2021). (October)
65. W. Rahman, R. Trotta, S.S. Boruah, M.J. Hudson, D.A. van Dyk, New constraints on anisotropic
expansion from supernovae type Ia. Monthly Notices of the Royal Astron. Soc. 514(1), 139–163
(2022)
66. Aidan C. Crook, John P. Huchra, Nathalie Martimbeau, Karen L. Masters, Tom Jarrett, Lucas
M. Macri, Groups of galaxies in the two micron all sky redshift survey. APJ 655(2), 790–813
(2007). (February)
67. R. Brent Tully, Galaxy groups. AJ 149(2), 54 (2015)
68. S.H. Lim, H.J. Mo, Y. Lu, H. Wang, X. Yang, Galaxy groups in the low-redshift universe.
MNRAS 470(3), 2982–3005 (2017)
69. T.S. Lambert, R.C. Kraan-Korteweg, T.H. Jarrett, L.M. Macri, The 2MASS redshift sur-
vey galaxy group catalogue derived from a graph-theory based friends-of-friends algorithm.
MNRAS 497(3), 2954–2973 (2020)
70. M. Levi, L.E. Allen, A. Raichoor, C. Baltay, S. BenZvi, F. Beutler, A. Bolton, F.J. Castander,
C.-H. Chuang, A. Cooper, J.-G. Cuby, A. Dey, D. Eisenstein, X. Fan, B. Flaugher, C. Frenk,
A.X. Gonzalez-Morales, O. Graur, J. Guy, S. Habib, K. Honscheid, S. Juneau, J.-P. Kneib, O.
Lahav, D. Lang, A. Leauthaud, B. Lusso, A. de la Macorra, M. Manera, P. Martini, S. Mao,
J.A. Newman, N. Palanque-Delabrouille, W.J. Percival, C. Allende Prieto, C.M. Rockosi, V.
Ruhlmann-Kleider, D. Schlegel, H.-J. Seo, Y.-S. Song, G. Tarle, R. Wechsler, D. Weinberg, C.
Yeche, Y. Zu, The dark energy spectroscopic instrument (DESI). Bull. Am. Astron. Soc. 51,
57 (2019)
362 W. D’Arcy Kenworthy and T. M. Davis

71. Luca Amendola, Stephen Appleby, Anastasios Avgoustidis, David Bacon, Tessa Baker, Marco
Baldi, Nicola Bartolo, Alain Blanchard, Camille Bonvin, Stefano Borgani, Enzo Branchini,
Clare Burrage, Stefano Camera, Carmelita Carbone, Luciano Casarini, Mark Cropper, Clau-
dia de Rham, Jörg. P. Dietrich, Cinzia Di Porto, Ruth Durrer, Anne Ealet, Pedro G. Fer-
reira, Fabio Finelli, Juan García-Bellido, Tommaso Giannantonio, Luigi Guzzo, Alan Heavens,
Lavinia Heisenberg, Catherine Heymans, Henk Hoekstra, Lukas Hollenstein, Rory Holmes,
Zhiqi Hwang, Knud Jahnke, Thomas D. Kitching, Tomi Koivisto, Martin Kunz, Giuseppe La
Vacca, Eric Linder, Marisa March, Valerio Marra, Carlos Martins, Elisabetta Majerotto, Dida
Markovic, David Marsh, Federico Marulli, Richard Massey, Yannick Mellier, Francesco Mon-
tanari, David F. Mota, Nelson J. Nunes, Will Percival, Valeria Pettorino, Cristiano Porciani,
Claudia Quercellini, Justin Read, Massimiliano Rinaldi, Domenico Sapone, Ignacy Sawicki,
Roberto Scaramella, Constantinos Skordis, Fergus Simpson, Andy Taylor, Shaun Thomas,
Roberto Trotta, Licia Verde, Filippo Vernizzi, Adrian Vollmer, Yun Wang, Jochen Weller, Tom
Zlosnik, Cosmology and fundamental physics with the Euclid satellite. Living Rev. Relativity
21(1), 2 (2018). (April)
72. C. Saulder, C. Howlett, K.A. Douglass, K. Said, S. BenZvi, S. Ahlen, G. Aldering, S. Bailey, D.
Brooks, T.M. Davis, A. de la Macorra, A. Dey, A. Font-Ribera, J.E. Forero-Romero, S. Gontcho,
A. Gontcho, K. Honscheid, A.G. Kim, T. Kisner, A. Kremin, M. Landriau, M.E. Levi, J. Lucey,
A.M. Meisner, R. Miquel, J. Moustakas, A.D. Myers, N. Palanque-Delabrouille, W. Percival,
C. Poppett, F. Prada, F. Qin, M. Schubnell, G. Tarlé, M. Vargas Magaña, B. Alan Weaver, R.
Zhou, Z. Zhou, H. Zou, Target selection for the DESI peculiar velocity survey. MNRAS 525(1),
1106–1125 (2023)
73. Adam Andrews, Jens Jasche, Guilhem Lavaux, Fabian Schmidt, Bayesian field-level inference
of primordial non-Gaussianity using next-generation galaxy surveys. MNRAS 520(4), 5746–
5763 (2023). (April)
Chapter 19
The Impact of Dust on Cepheid and Type
Ia Supernova Distances

Dillon Brout and Adam Riess

19.1 Introduction

Cosmic dust particles play a crucial role in cosmological observations because they
absorb, scatter, and re-emit starlight, leading to extinction, attenuation, and an overall
reddening of light from distant objects. The combination of Milky-Way dust and
extragalactic dust affect nearly all measurements in cosmology. From measurements
of the cosmic microwave background, to Cepheids and Type Ia Supernovae (SNe
Ia), to weak lensing and large-scale structure, to optical counterpart identification
of gravitational-wave multimessenger phenomena. In this chapter we focus on the
impact of dust specifically for two of the rungs of the S. H0 ES [1] cosmic distance
ladder: Cepheids and SNe Ia.
Cosmic dust grains are composed of tiny solid particles, ranging in size from a
few nanometers to several micrometers, that are generated by a variety of processes,
including stellar nucleosynthesis, supernova explosions, and collisions between solid
bodies. While the structure and composition of dust particles depend on the specific
astrophysical environment in which they are formed, in general, dust grains are made
up of silicon, carbon, and large hydrocarbon molecules (see Fig. 19.1 Top), each of
which have slightly different effects due to their composition, shape, and reflectivity.
The particles can have a wide range of shapes, from spherical to elongated, though
their shape and size distributions are not measured directly, but rather are instead
inferred through their effects on extinction, polarization, and albedo.
To measure for dust extinction in the Milky Way, specific backlights such as
stars and quasars serve as reference sources with well-characterized intrinsic spectral

D. Brout (B)
Boston University, 685 Commonwealth Ave, Boston, MA 02215, USA
e-mail: dbrout@bu.edu
A. Riess
Space Telescope Science Institute, 3700 San Martin Drive, Baltimore, MD 21218, USA
Department of Physics and Astronomy, Johns Hopkins University, Baltimore, MD 21218, USA

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 363
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_19
364 D. Brout and A. Riess

Fig. 19.1 Top: Schematic

representation [2] of an
example non-normalized
attenuation or extinction
curve. Bottom: Schematic
representation [3] of the
different effects of grain
compositions

energy distributions (SEDs). By comparing the observed attenuated SEDs to the unat-
tenuated counterparts, one can estimate the extinction as a function of wavelength.
In this chapter, we will explore the mathematical relationships between relevant
parameters describing the physics and impact of cosmic dust, including extinction,
reddening, and dust laws, with a particular focus on their impact on the cosmic
distance ladder and measurements of cosmological parameters.

19.2 Dust Formalism

Dust extinction and attenuation are processes that cause dimming and reddening of
light. While they arise from different but related physical mechanisms, they have
the same effects on cosmological observables. Extinction is due to dust along the
line of sight to objects that causes absorption or scattering away from the line of
sight. Attenuation includes the effects of extinction but also includes scattering
back into the line of sight and the contribution to the light by unobscured stars. It is
the net loss of light from a complex, spatially extended population and depends on
19 The Impact of Dust on Cepheid and Type Ia Supernova Distances 365

both extinction and the star-dust environment geometry. The wavelength dependent
extinction or attenuation can be expressed mathematically as:

. A λ = m λ − m λ0 (19.1)

where .m λ0 represents the unextinguished magnitude of an astronomical object at

a given wavelength, .m λ represents the observed magnitude, and . Aλ describes the
amount of extinction.
The attenuation or extinction curve describes a normalized shape of . Aλ as a
function of wavelength (see Fig. 19.1 Bottom). The overall normalization (in most
cases set by . A V ) of the attenuation or extinction curve is set by the column density
of grains and the overall shape is set by the mixture of grain types and grain size
distribution, and grain optical properties along the line of sight. It is difficult to relate
the extinction curve directly to dust physics, so such curves are largely a measured
quantity using known background sources and for the purposes of cosmology this is
sufficient.
There are several observed features of a typical dust extinction curve (shown in
Fig. 19.1). The Ultra-violet (UV)—optical slope (. A1500 /A3000 ) and the strength of
the UV bump (B) is seen to vary significantly within and across galaxies and provides
a large lever arm to constrain the overall dust properties as it provides information
about the ratio of the star formation rate (SFR) and the dust column density. The
optical slope (. A B /A V ) is important to account for in many cosmological constraints
that are observed in the optical, e.g. SNe Ia. The near-IR power-law slope (.βNIR )
which is the smallest of these features as dust has the least impact and variation at
larger wavelengths.
The color excess . E(B − V ) characterizes the reddening of an object’s spectrum
due to dust extinction. It is defined (Eq. 19.2) as the difference between the extinction
in two bands (B-V).

. E(B − V ) = A B − A V (19.2)

By comparing the magnitudes in different bands, one can estimate the amount of
reddening caused by dust and gain insights into the properties of the intervening
material. Differential measurements of E(B-V) are extremely useful, though to know
the true color excess one must have some knowledge of the intrinsic color of the
source object. Thus, differential maps of E(B-V) across the Milky-Way galaxy are
quite accurate, though knowing the absolute value/normalization of the color excess
is a more difficult task (see Sect. 19.3).
Lastly, a useful quantity that is often used in analyses of distance indicators is the
‘total-to-selective extinction ratio’ . R:

. RV ≡ A V /E(B − V ) (19.3)
366 D. Brout and A. Riess

where this represents the linear slope of the extinction curve at a given band (e.g.
RV ). These ratios provide quantitative information about the differential extinction in
.
the relevant (.V -optical) bands/wavelengths and are used in cosmological constraints
of Sects. 19.5.1 and 19.5.2.

19.3 Milky Way Dust

Accurate estimates of the Milky Way dust are exceptionally important in cosmology
as all electromagnetic observations must contend with it at some level to correct
for the observed brightnesses and colors of extra-galactic sources and to understand
spatially dependent selection effects.
Reddening: Dust maps can be derived in two main ways: emission-based maps and
extinction-based maps. Emission-based maps utilize the observed far-infrared ther-
mal emission from dust to estimate the dust column density. The Schlegel-Finkbeiner-
Davis map (SFD: [4]) is based on 100.μm data from the Infrared Astronomy Satellite
(IRAS), with corrections to remove zodiacal light and the cosmic IR background
based on maps from the Diffuse Infrared Background Experiment (DIRBE). This
SFD 100.μm map has a resolution of 6.1 arcminutes FWHM. Because hotter dust
emits more 100.μm radiation per column density, the SFD 100.μm map is modulated
by a temperature correction based on the DIRBE 100.μm/240.μm color temperature.
This modulation is on a 1.3 degree scale where the dust is sufficiently bright, and
transitions to a 7 degree scale at high Galactic latitude.
The overall calibration of the SFD map was updated in 2011 [5] based on spectra
of main sequence turn-off stars. This followed [6], which used Sloan Digital Sky
Survey data of the “blue tip” of the stellar locus (a prominent blue edge in the stellar
color distribution) that can be used to estimate extinction along the line of sight
and is sensitive to .∼10–20 mmag color changes. Using this method [6] find a 14%
scaling of the overall SFD map normalization and a preference for the Fitzpatrick
1999 [7] color law over O’Donnell [8] and Cardelli et al. [9] (see next section).
[5] follow up on this work using a different technique that uses stellar spectra and
modeling codes to find an additional 4% calibration normalization relative to [6] and
re-confirm a preference for the Fitzpatrick color law in the optical wavelengths. This
4% difference between modern methodologies has been adopted for the systematic
uncertainty in numerous cosmology analyses. Though it is important to emphasize
that these uncertainty estimates were best constrained from data within the galactic
plane and are not guaranteed to hold true out of the plane of the galaxy where
data is more sparse and relative constraints are worse (though with lower reddening
amplitude).
After the SFD work, maps based on Planck [10] became available. Planck obser-
vations of dust at several frequencies (217, 353, 545, 857 GHz) are used, but in order
to constrain the high-frequency side of the thermal emission spectrum, they used
the SFD 100.μm. In other words, the same deficiencies in the SFD zodiacal light
19 The Impact of Dust on Cepheid and Type Ia Supernova Distances 367

correction and zero point affect the Planck maps. Furthermore, the Planck analysis
attempts to infer the emissivity power-law index .β along with dust temperature and
optical depth .τ , introducing additional noise at high latitude where the emission is
to faint to constrain 3 parameters. So although the Planck data are superior to IRAS
in general, the derived dust map shares some shortcomings with SFD.
Another method for mapping dust involves utilizing inferred stellar extinctions
while considering a presumed or fixed reddening law. Empirical constraints on extinc-
tion require some knowledge of the unextinguished spectrum, either via models or
empirical pairs of more/less reddened stars. Estimations of reddening along the line
of sight to these stars is somewhat challenging, but not an overwhelming burden. Fur-
thermore, since each star provides information about the extinction up to its distance,
these stellar reddening-based maps allow for the reconstruction of dust extinction
in three dimensions [11–13]. However, these maps may suffer from contamination
due to incomplete removal of galaxies and quasi-stellar objects (QSOs) from stellar
catalogs [14]. This contamination can lead to systematic errors, causing both over-
estimation and underestimation of extinction at the positions of galaxies and QSOs.
Thus, these maps are not perfect and there is the need to motivate further progress
for the next generation of experiments.
Dust Laws: For use in broadband photometry fits like that of SN Ia and Cepheid
luminosity measurements, it is important that attenuation curves are parameterized
for use in fitting. Common parameterizations distill attenuation to single parameter
characterizing the slope of of a predefined curve, though some include additional
parameters describing the UV bump (see Fig. 19.1). Such parameterizations do not
need to be unique to the Milky Way galaxy and can extend to extra-galactic dust
measurements for comparison. While the variations in attenuation curves has thus
far been difficult to relate to dust physics, for the purposes of precision cosmological
constraints the empirical relations described below are adequate.
Fitzpatrick [7] provide a single-family parameterization for extinction curves
(originally intended for the Milky-Way, though now extended beyond) from the
optical to near-IR. They used a comprehensive dataset of reddened stars with known
intrinsic colors and distances, coupled with detailed spectroscopic measurements to
compute reddening effects across a wide range of wavelengths and characterize R(V).
Fitzpatrick [7] showed significant variations in the extinction law along different sight
lines and several notable studies have extensively confirmed and characterized these
variations [5, 6, 15]. The impact of this variation is discussed in Sect. 19.5.2.

19.4 Extragalactic Dust

For many cosmic distance probes, it is also important to understand the dust in the
galaxy to which distances are being measured. As . RV varies for different dust grain
sizes and composition, and galaxies have different dust properties, it is well known
that different galaxies and different regions within galaxies exhibit a wide range of
368 D. Brout and A. Riess

. RV values. In fact, while the Milky Way galaxy has an . RV on average .∼ 3.1, it has a
distribution of at least .σ RV = 0.2 [16]. Additionally, different parts of the LMC and
SMC have been found to have . RV values with a range of . RV ∼ 2 − 5 [17, 18]. The
origin of these dramatically different extinction curves is still debated as it could be
driven by metallicity, radiation field strength, or both.
Furthermore, [19] study the dust attenuation curves of 230,000 individual galaxies
in the local universe, using GALEX, SDSS, and WISE photometry calibrated on the
Herschel ATLAS, and they find quiescent galaxies, which are typically high-mass,
have a mean. RV = 2.61 and star-forming galaxies, which are lower-mass on average,
have a mean . RV = 3.15.
. R V has also been measured through large SN sample statistics and detailed studies
of individual SNe, though often with varying sets of assumptions. [20] compiled 13
various studies of SN Ia samples from the literature which determined a range of . RV
values from .∼ 1 to .∼ 3.5. [20] itself determined . RV from nearby SNe and for 21 SNe
Ia observed in Sab-Sbp galaxies and 34 SNe in Sbc-Scp they find . RV = 2.71 ± 1.58
and . RV = 1.70 ± 0.38 respectively. While so many past analyses have recovered
. R V < 2 for studies of individual SNe (e.g. [21, 22]), these were often SNe Ia with
high . E(B − V ), and it was postulated . RV may decrease with . E(B − V ). However,
[23] found from a sample of modestly reddened (. E(B − V ) < 0.25 mag) SNe Ia,
a small value of . RV ∼ 1 and more recently, [24] analyzed high-quality UV-NIR
spectra of 6 SNe and found that SNe with high reddening indicated . RV ’s ranging
from .∼ 1.4 to .∼ 2.8 and SNe with low amounts of reddening also indicated . RV ’s of
.∼ 1.4 and .∼ 2.8. Importantly, [24] stressed that the observed diversity in . R V is not
accounted for in analyses that measure the cosmological expansion of the universe.
Since the low . RV values (.< 2) are not found in studies of the Milky Way, this has
motivated various SN Ia studies to ascribe the dust to circumstellar dust around the
progenitor at the time of the explosion [25, 26]. However, an alternative interpretation
could be that the low . RV values are caused by dust in the interstellar medium [27].
This understanding has been supported by [28, 29], which constrained the location
of the dust that caused the reddening in the SN Ia spectra to be, for the majority of the
SNe that they observed, on scales of the interstellar medium, rather than circumstellar
surroundings.

19.5 Impact on Distance Indicators

19.5.1 Impact on Cepheid Magnitudes

While the impact of dust on measurements of galactic and extra-galactic Cepheids

is minimal due to the fact that in the distance ladder approaches they are observed
in the Hubble Space Telescope (HST) Near Infrared (NIR) bands, we detail here
the possible impact and its relevance for the inference of the Hubble Constant. For
the case of nearby first and second rung Cepheids in the S. H0 ES distance ladder,
19 The Impact of Dust on Cepheid and Type Ia Supernova Distances 369

they are all at extremely low redshift (.z < 0.01), and therefore both Milky Way and
extra-galactic dust affect the same range of wavelength and their effects are treated
simultaneously below.
Extra-galactic Cepheids are observed in regions of recent star formation in late-
type galaxies and are found to have a mean reddening of . E(V − I ) .∼ 0.3mag. The
S. H0 ES [1] observations in HST NIR F160W (. H -Band) reduces the impact of and
correct (‘deredden’) for the effects of interstellar dust reddening.
For the distance ladder, the goal is to measure the relative distance between two
sets of Cepheids (rungs) free from reddening which by definition is .(m H,i − A H,i ) −
(m H, j − A H, j ) for the .ith and . jth samples where . A H is the extinction from dust and
comes from the color excess,. E(V − I ) multiplied by a reddening ratio,. R H , so. A H =
R H × E(V − I ). However, in practice we do not directly measure . E(V − I ) which
would require knowledge of the intrinsic color, .(m V − m I )0 . Rather, we observe the
apparent color, .(m V − m I )0 . A useful, shortcut to measuring the distance difference
using only the apparent quantities is to use so-called “Wesenheit” magnitudes [30]
(.m W
H ) following:

.H = m H − R H × (m V − m I )
mW (19.4)

This measure is based only on directly observed quantities. Below we denote the
components of the Cepheid model that make up .(m V − m I ). Wesenheit magnitudes
W
.m H are not true magnitudes in the sense that they do not represent the brightness.
Rather, their utility is in the comparison between two sets of such magnitudes, which
results in the reddening-free magnitude difference.
Their utility can be seen when we decompose the apparent color .(m V − m I ) into
its constituents. The main component is .(m V − m I )0 which would be the “unred-
dened” color at the middle of the Cepheid instability strip (typically this term is
. V − I ∼ 1). The displacement in color from the midline of the instability strip is
given by .Δ(m V − m I ) which corresponds to an intrinsic property of each Cepheid.
Finally, the color excess due to dust is given by . E(V − I ) which is reddening by
dust:

m V − m I = (m V − m I )0 + ΔV,I + E(V − I )
. (19.5)

Thus, multiplying (.m V − m I ) of Eq. 19.4 by a single . R H for two samples .i and
.j (i.e. two rungs of the distance ladder) results in the following when applied to the
decomposition in 19.5. The terms .(m V − m I )0 multiplied by . R H will be the same for
different galaxies, this being an intrinsic property of Cepheids, and it is important to
use the same value of . R H (whatever that is) where multiplying by the intrinsic color.
When multiplying . R H on the .ΔV,I term, because the mean of .ΔV,I is at or near zero
(by definition) for large samples of Cepheids that fill the instability strip, . R H ΔV,I
will also average to zero for a statistically large sample. However, because Cepheids,
like most stars, are dimmer when they are redder (cooler), a positive value of . R H
can also reduce the apparent dispersion of the instability strip. In fact, observations
of stars show the change in intrinsic color with brightness is of the same order as the
370 D. Brout and A. Riess

reddening law, . R, a happy coincidence, so that the term . R H reduces variance in two
ways. Therefore, it is common to choose the value of . R H to be the reddening ratio
between the . H -band and the .V − I color, as derived from a reddening law. Lastly,
the only term that does not cancel or we expect to not be consistent across galaxies
is . R H × E(V − I ) because . E(V − I ) varies within and across Cepheid samples.
Thus, we can now rewrite the observable of Eq. 19.4 as .m H − R H × E(V − I ).
Finally, taking the differential between two samples leaves us with our desired result
.m H,i − m H, j = (m H,i − R H × E(V − I )i ) − (m H, j − R H × E(V − I ) j ).
W W

We get the exact difference in distance when . RW,i = RW, j (and with a large
Cepheid sample, .ΔV,I = 0). We note here that the “shortcut” Eq. 19.4 only works if
we require . RW,i = RW, j . If we do not want to require . RW,i = RW, j , then we must
decompose 19.4 into the terms in 19.5 and then we need to define and model the
intrinsic standardizable color of Cepheids from another sample and only allow for
. R W to vary across for the dust excess component [1, 31, 32].
In the case of correcting NIR magnitudes (.m H ) following Eq. 19.4, we have thus
far assumed a single . R H . Fortunately, . R H is well-characterized to be small (.∼ 0.4)
for Milky Way-like extinction. This value is much smaller than for correcting optical
magnitudes (. RV ∼ 3) and demonstrates the comparatively low sensitivity to dust in
the S. H0 ES analysis. This dust law insensitivity is also important when considering
potential variation in the dust curve across the population of Cepheid host galaxies.
While . RV can range from 2 to 6, . R H correspondingly ranges from 0.35 to 0.5 with
a definitive lower bound at .∼0.2 due to the Rayleigh scattering limit or 0.25 for any
known reddening law.
While . R H has been determined independently, . R H can also be determined from
the Cepheids themselves as the value that minimizes the observed scatter in the
observed Period-Luminosity relation. From the highest quality Milky Way Cepheids
. R H = 0.363 ± 0.038 [33] which matches independently determined values in the

literature. Additionally, . R H determined from the entire distance ladder of Cepheids

is . R H = 0.34 ± 0.02 [1]. For this value, . H0 increases by 0.3.
It is possible to consider different values of the reddening ratio for each individual
host. However, because with different . R H the intrinsic color term .(m V − m I )0 no
longer cancels across rungs, the intrinsic term loses its meaning, and defies the
interpretation of Cepheids as standard candles. So this cannot be done directly with
Wesenheit magnitudes. Therefore, to vary the reddening ratio, one must first subtract
the color excess using an empirical period-color relation, such as .(V − I )0 = 0.5 +
0.25 log P (empirical, Milky Way or Large Magellanic Cloud) from .(V − I ), or this
can lead to a spurious result [32]. When . R H is fit individually for each host, . H0 has
been shown to change at the 0.8 km/s/Mpc level, as shown in [31]. In practice, the
Cepheid data is not very constraining of individual reddening ratios in the SN Ia
hosts, so an alternative and better approach is to assign individual host reddening
ratios based on the host star formation rate and mass which predict .2.8 < RV < 4.2
[1].
Observing in the NIR to mitigate the impact of dust is not without its own chal-
lenges. On HST, the resolution in NIR is 2–3 times worse than that of the optical
and in the NIR the background of ubiquitous red giant stars is an order of magnitude
19 The Impact of Dust on Cepheid and Type Ia Supernova Distances 371

higher. However, with the advent of the James Webb Space Telescope (JWST), we
have been further reassured that there is not a confounding issue when it comes to
the impact of dust for Cepheids. Both [34] and [35] show that JWST provides signif-
icantly improved resolution, improved sensitivity in the NIR, and even lower impact
of interstellar dust in comparison to HST. They find good agreement between the
HST and JWST systems and no evidence for bias beyond the 0.04mag level, which is
significantly better than the 0.2mag size of the . H0 tension and is expected to further
improve with future observations.

19.5.2 Impact on Type Ia Supernova Distance Moduli

In order to compute distances, Type Ia Supernova (SN Ia) photometric light curves
are fit to an underlying spectral time-series model (e.g. SALT3 [36]). These fits output
the following summary statistics that characterize each light curve: the peak B-band
brightness (.m B ), the duration ‘stretch’ (.x1 ), and the color (.c). The latter of which, .c,
characterizes the combination of intrinsic SN Ia color and dust reddening because at
the photometric light curve level such properties are degenerate. SN Ia light-curve
fit parameters are then combined empirically to compute distances (.μobs ) where the
nuisance parameters (.α, .β, .γ ) are single numbers that correlate light-curve and host
properties with an intrinsic luminosity and are determined such that they minimize
the scatter in distances of SNe Ia [37]. The distance moduli are computed following
a modified version of the Tripp [38] equation:

.μobs = m B + αx1 − βc + γ G host − M B + μbias , (19.6)

where .G host is each host-galaxy’s mass that has been shown to correlate with SN Ia
brightness [39–42], .μbias is a correction for instrumental and survey selection effects
for each SN based on careful survey simulations [43], and .M B is a single number
representing the intrinsic magnitude of a SN Ia as determined by the S. H0 ES distance
ladder.
For SNe Ia, galactic and extra-galactic dust both reddens and obscures light in
the rest-frame UV to optical wavelengths. With the exception of the future NASA
Roman Space telescope, large area supernova surveys are typically observed from the
ground in the optical wavelengths (due to atmospheric transmission). The analysis
of Pantheon+ [45] and S. H0 ES [1] account for Milky Way galactic dust by applying
the Schlafly et al. 2011 stellar locus map [5]. A second correction is applied with
a 4% absolute scale offset following [6] found using the stellar spectral modeling
approach. This difference is evaluated in the Pantheon+ SN Ia covariance matrix in
order to characterize the systematic uncertainty.
Dust in the host galaxy of the supernova is degenerate with the light-curve color
and is hypothesized to be the major cause of the so called intrinsic scatter of stan-
dardized SN Ia brightnesses (the excess scatter of SN Ia distance residuals compared
to a best-fit cosmology after accounting for measurement noise). Recent promising
372 D. Brout and A. Riess

Fig. 19.2 Top: Observed SALT light curve fit parameter color (.c) from the full data sample of [44],
with a symmetric Gaussian overlaid (dashed) for comparison. Skewness in the observed dataset to
the red is evident and indicative of dust. Middle: RMS scatter of Hubble Diagram residuals as a
function of color. The RMS is calculated after Tripp standardization and after subtracting the mean
Hubble residual bias. SNe Ia with bluer colors exhibit significantly better standardizability than
their redder SN Ia counterparts. This is indicative of unmodeled variation in dust properties for
each SN Ia or host galaxy. Bottom: Binned Hubble diagram residuals as a function of color, after
Tripp standardization using a single best fit.β. The bluest SNe Ia exhibit a different color-luminosity
relation than the reddest; indicative of an intrinsic SN Ia distribution and a secondary distribution
characterized by dust. In the bottom two panels, the significance of the deviation from a flat line is
show in the bottom corner

models have now explained the majority of SN Ia intrinsic scatter using dust. The size
of SN Ia intrinsic scatter has recently been found to depend on SN Ia color and host
galaxy properties (BS21: [44]). Recently, BS21 introduced dust extinction to the SN
Ia spectral model by forward modeling simultaneously 1) an intrinsic SN color pop-
ulation, 2) extrinsic dust populations, 3) variation in the dust properties from galaxy
to galaxy, and 4) correlations between the dust properties of galaxies with different
host properties (.G host ). BS21 marks a major improvement over previous intrinsic
scatter models commonly used in SN Ia cosmology analyses (e.g. grey scatter: Guy
et al. 2010 [46] and spectral scatter: Conley 2011 [47]. While the Tripp Eq. 19.6 does
not disentangle intrinsic color and extrinsic dust properties, by forward modeling
this process and analyzing the forward modeled simulations with the Tripp equation,
the model can be compared with and fit to the data in a hierarchical approach (see
Fig. 19.2).
19 The Impact of Dust on Cepheid and Type Ia Supernova Distances 373

The major successes of the BS21 model are that solves two key problems with
one simple solution. (1) It allows for . RV ’s in high and low mass host galaxies which
explains the observed correlation with host properties. (2) It explains the observed
color dependent intrinsic scatter with variation in dust properties. It also makes key
predictions, that the typical SN Ia dust (whether it be local to the SN or global to the
host galaxy) corresponds to an . RV σ 2 and with a large inter-galaxy dispersion .∼ 1
(BS21 and [48]) which has similarly been seen in numerous followup studies [49–
51]. Continuing upon the improved modeling developed in BS21, [52] implement
the dust model used in cosmological analyses. [52] correct the Pantheon+ distance
moduli on a population level and [48] address the computational limitations in the
hierarchical model fitting process in order to provide robust model uncertainties
that include correlations in uncertainties between dust parameters and SN intrinsic
parameters.
Note that alternative models for the correlation between SN Ia luminosity and
host-galaxy properties have been proposed that involve progenitor differences [53]
or different subclasses [54] instead of dust. However, in light of recent findings
in Pantheon+ for the very strong color dependence of intrinsic scatter and .β, such
alternative scenarios must explain now a three path mechanism to explain the obser-
vations of SN Ia color skewness to the red, SN Ia one-sided color-dependant scatter,
and both scatter and .β dependencies on host properties (not depicted in Fig. 19.2).
The large statistics of the Pantheon+ sample illuminating these dependencies makes
such alternative approaches challenging.
A new SN-host correlations model was recently explored in the Dark Energy Sur-
vey Supernova Program that accurately and realistically models physically motivated
relations between SN age, SN host galaxy properties, and SN stretch (.x1 ). Vincenzi
et al. (in prep) find that this model does not explain the host-mass correlation with
luminosity, so they use this model in conjunction with a BS21 dust based scatter
model and find that inferred distances to SN Ia remain relatively unchanged relative
to the approach taken in Pantheon+. Finally, further evidence for BS21 is substanti-
ated by the fact that dust properties are also found to correlate with other physical
parameters of interest in galaxy surveys. [19] find lower mass galaxies tend to have
steeper attenuation curves (. RV ).

19.6 Impact on Cosmological Constraints

19.6.1 Dust Impact (or Lack-Thereof) on the . H0 Tension

The S. H0 ES methodology of obtaining the Hubble constant is one that leverages

differential measurements between the first and second rungs (for Cepheids), and
the second and third rungs (for SNe Ia). Here we emphasized that only differences
in dust properties between rungs is what could potentially contribute to biases in . H0 .
374 D. Brout and A. Riess

Fig. 19.3 Difference between the observed color excess of the Cepheids for the first (Anchor) and
second (SN Host) rungs of the distance ladder for different underlying dust law (. R V ). . H0 Tension
of 0.2 mag is shown for comparison to the 0.03 mag differences observed in the sample

This is because the S. H0 ES likelihood and methodology is a differential measurement

for Cepheids and SNe Ia, meaning that offsets that are consistent across rungs cancel.
Cepheids:
Figure 19.3 visualizes the impact of dust on the Cepheids in the first two rungs of the
distance ladder (Geometric “Anchor” Cepheids and “SN Host” Cepheids). For the
Cepheids in both rungs (red and blue points), color excess . E(V − I ) is measured
relative to the intrinsic Cepheid color at the center of the instability strip and for a
typical Milky Way dust law (. RV = 3.3) the corresponding extinction (. A V ) values
are inferred from the color excess. As discussed in Sect. 19.4, there are numerous
extinction laws not only within a galaxy but also across galaxies. Therefore, Fig. 19.3
also visualizes other possible dust laws ranging from . RV of 2–4.
The difference in mean color excess and inferred extinction is the most relevant
quantity for assessing the impact on the . H0 tension. Thus, we show the Cepheid
color excess for each rung (First: blue, Second: red) and the average difference
(.ΔE(V − I ) ∼ 0.08) in Fig. 19.3. If the means between rungs are identical, . H0 is
completely insensitive to dust by nature of the distance ladder methodology, as only
differences in the means of each rung can propagate to changes in . H0 .
The entire . H0 tension between S. H0 ES and Planck (6km/s/Mpc) can be expressed
as 0.18mag in .ΔA V (shown for reference on Fig. 19.3). However, because the mean
color excess between the first and second rungs is .ΔE(V − I ) ∼ 0.08 and because a
Milky Way dust law of. RV =3.3 corresponds to an. R H =0.4 in the infrared (because the
19 The Impact of Dust on Cepheid and Type Ia Supernova Distances 375

Cepheid luminosities are computed in HST H-band), this corresponds to a 0.03 . A H

difference for ‘Milky Way-like’ dust. If the typical dust law is not Milky Way-Like,
but rather . RV =2.3 (. R H = 0.35) or in a worse case scenario of . RV =4.0 (. R H = 0.35),
the difference in extinction . A H could range from 0.025 to 0.04, however, in no
case does the expected . A H approach the 0.18mag needed to resolve the . H0 tension.
Additionally, analyses such as [31] and [32] have allowed for. R H to vary from galaxy
to galaxy and find that, despite the extra model freedom and parameters, because
. R H is small and the color excess are small, variations in . R H across Cepheid galaxies
cannot explain the. H0 tension. Finally, in Table 19.1 are depicted the S. H0 ES variants;
S. H0 ES explore numerous attenuation laws (e.g. Cardelli et al. [9] and Fitzpatrick
[7]), attenuation law slopes and dependencies, and even remove the dust treatment
altogether and show that they all have either minimal impact on . H0 or move the
central value up to make the tension worse.
SNe Ia:
As shown in Fig. 13 of Pantheon+, the impact of a 4% absolute scaling of galactic
dust maps on . H0 is negligible. This is because, first, SNe Ia are typically discovered
in low extinction regions of the sky. Second, the average Milky Way. E(B − V ) of the
second rung is 0.050 and of the third rung is 0.046; multiplying the difference of 0.004
by the typical Milky Way color law of .∼ 3 results in an average brightness difference
between the rungs of 0.012 mag. Given that the . H0 difference is .∼0.20 mag, this
suggests very minimal impact on . H0 from differences in Milky Way . E(B − V )
across rungs (See Table 19.1 for .ΔH0 ). Lastly, it is important to note that small
errors in the dust maps will result in different inferred SALT color parameter (.c) of
each light curve. Because the SALT color-luminosity relation .β ∼ 3 is very similar
to that Milky Way dust attenuation law .∼ 3, dust map differences are fortuitously
‘absorbed’ by the Tripp estimator Eq. 19.6.
Pantheon+ incorporate distance uncertainties in a covariance built from numerous
systematic perturbations to the analysis, including perturbations characterizing the
uncertainty in the underlying model of dust. The inferred. H0 for each of the systematic
perturbations is shown in Fig. 13 of Pantheon+ and shows a 0.2 km/s/Mpc scatter
in . H0 resulting from the uncertainty in the dust population parameters, which is
well below the level of the . H0 tension. This is bolstered by the fact that because the
Cepheids are only visible in edge-on late type/spiral galaxies in the second rung, and
robust determinations of . H0 select for the similar types of host galaxies in the third
rung of the distance ladder. While the underlying dust properties and differences
between spiral galaxies and other galaxy types remain an open question, restricting
the third rung of the distance ladder in S. H0 ES to only late type galaxies mitigates
possible differences in galaxy properties between rungs. The S. H0 ES analysis explore
an additional variant in which they allow all types of Hubble flow host galaxies (and
therefore dust properties) and in this case find a .ΔH0 = +0.3, exacerbating the
tension.
The .ΔH0 relative to the BS21 dust model is shown in Table 19.1 for the Guy
et al. [46] and Conley [47] intrinsic scatter models. Instead of dust, the Guy et al.
[46] model describes SN Ia scatter with a grey luminosity scatter. While this model
376 D. Brout and A. Riess

Table 19.1 Variants to the Pantheon+ and S. H0 ES analysis that assess the impact of dust. Negative
.ΔH0 is in the direction of reducing tension
Dust-Related Systematic .ΔH0 (km/s/Mpc)

Cepheids
Milky Way (MW) Color Law from Fitzpatrick (1999), . R V = 2.5 +0.20
MW Color Law from Cardelli, Clayton, & Mathis (1989), . R V = 3.1 +0.05
. R W free global +0.20
Intrinsic color subtracted & MW Color Law from Fitzpatrick (1999) & . R V = +0.09
3.3
Intrinsic color subtracted & MW Color Law from Fitzpatrick (1999) & . R V = +0.30
free
Intrinsic color subtracted & . R V (host mass-SFR) +0.81
No dust treatment (. A H values assumed to cancel) +0.74
SNe Ia
Milky Way (MW) Dust Map Absolute Scaling (4%) .−0.02

MW Color Law from Cardelli, Clayton, & Mathis (1989) +0.05

Grey Scatter Model from Conley (2010) .−0.20

Spectral Scatter Model from Guy et al. (2011) .−0.22

SN Ia/Host Dust Model Parameter Uncertainty 1 .−0.02

SN Ia/Host Dust Model Parameter Uncertainty 2 .−0.12

SN Ia/Host Dust Model Parameter Uncertainty 3 .−0.13

All host galaxy morphologies/types in third rung +0.28

Hubble flow high mass hosts only .log10 (Mstellar ) > 10 .−0.07

does not perform as well on the most recent Pantheon+ data (see Fig. 5 of BS21),
nonetheless while this model is agnostic to host-galaxy dust it only achieves a.ΔH0 of
.−0.2 km/s/Mpc. Likewise, Conley (2011) characterizes SN Ia scatter with empirical
spectral variations and finds .ΔH0 of .−0.2 km/s/Mpc.
There have also been numerous independent crosschecks on the SNe Ia and Pan-
theon+ dust treatment. First, [55] derive SN Ia distances using infrared light curves
thus reducing the impact of dust. They find reduced correlations with host-galaxy
properties and SN Ia luminosity, they suggest that dust drives the correlations and
their infrared observations explain the lack of correlation. Ultimately [55] report a
value of the . H0 = 73.2 ± 2.3 km/s/Mpc.
Most recently, [56] assessed the impact of dust on a subset of the SNe Ia for which
additional wavelengths from UV to NIR were available and with an improved SN
Ia model were able to disentangle the dust and SN properties for 37 calibrator SNe
and 67 Hubble flow SNe. While only for a subset of the total Pantheon+ SNe, this
importantly meant [56] could simultaneously compute an independent check on the
Pantheon+ dust modeling and find . H0 = 74.82 ± 1.28.
A third major crosscheck has been made in [57] on the S. H0 ES methodology
of restricting SN Ia hosts in the Hubble flow to match that of the second rung of
19 The Impact of Dust on Cepheid and Type Ia Supernova Distances 377

Cepheid hosts. To avoid the SN Ia dust systematics imposed by restricting to late-

type star-forming galaxies in both the second and third rungs of the distance ladder,
[57] use what is colloquially called a 4-rung distance ladder. Using galaxy groups,
[57] transfer the Cepheid calibration late-type galaxies to elliptical galaxies using
the surface brightness fluctuations (SBF) technique of determining relative distances
of early-type galaxies. These galaxies have different distributions of mass and dust
and also host different populations of SNe Ia. While the introduction of an additional
and less precise SBF rung of the distance ladder results in a larger uncertainty,
the recovered central value is an important crosscheck on many host dust related
systematics. [57] find . H0 = 74.6 ± 2.85.

19.6.2 Dust Impact on SN Ia Constraints of Dark Energy

and Dark Matter

While the above has shown that impact on inference of. H0 from Milky-Way and host-
galaxy dust is minimal, the same cannot be said for inference of other cosmological
parameters: Matter+Dark Matter Density (.Ω M ) and Equation of State of Dark Energy
(.w). The . H0 insensitivity is largely due to similarities between the second and third
rung SNe Ia and hosts, but it is also attributable to the small redshift range in the
Hubble flow (third) rung of the distance ladder.0.023 < z < 0.15, which corresponds
to a relatively small span of .<1.5 billion years of cosmic history over which dust
properties have little time to evolve.
For inference of.Ω M and.w, a much larger span of cosmic history is used to compile
sufficient SN Ia statistics and the constraints of those models are also best performed
at relatively higher redshifts. Samples of SNe Ia covering a very large span of cosmic
history (currently 11 billion years) opens up the possibility for the evolution of dust
parameters or the environments in which SNe Ia are found. Both [44] and [48] show
that if uncertainties in dust populations of SN Ia hosts can evolve with redshift and
this results in the largest systematic uncertainty in SN Ia cosmology today (on par
with photometric calibration). Though, if correlations between dust properties and
host-galaxy photometric properties (i.e. stellar mass, star formation rate, color) can
be well constrained at low-redshift, then potential population evolution can be traced
photometrically for future high redshift surveys. We note here that while there is
much effort on survey calibration for future telescopes, there is comparatively less
effort being placed on Milky Way and host-galaxy dust modeling, both of which are
tractable issues with external datasets and effort.
Future surveys such as The Vera Rubin Observatory Legacy Survey of Space and
Time (LSST) and The Nancy Grace Roman Space Telescope (Roman) will be forced
to tackle this challenge. LSST observes in optical wavelengths and will be more
susceptible to the impact of host-galaxy dust and population evolution similarly to
that of the Dark Energy Survey analysis. However, unlike the Dark Energy Survey,
LSST will survey the entire southern sky (at much larger values of Milky Way
378 D. Brout and A. Riess

E(B − V )) and will have to contend with increased systematics due to Milky Way
.

dust. It will be important to improve uncertainties on Milky Way dust maps in regions
of the sky where extinction is larger. Comparatively, Roman will observe in the
infrared, which will make it less sensitive to dust for SNe Ia at moderate redshifts,
though Roman will contend with the same dust population evolution systematics as
other surveys when observing SNe Ia at redshift 3 (rest-frame optical).

19.7 Conclusion

In conclusion, we have explored the impact of dust on Cepheid and Type Ia Supernova
distance indicators and the resulting impact on cosmological constraints, focusing
on the measurement of the Hubble Constant (. H0 ). We show that the effects of dust
propagate to changes in . H0 only if there are significant differences between dif-
ferent rungs of the distance ladder. We discuss the impact of Milky-Way dust and
host-galaxy dust and the associated uncertainties in these properties. In both cases,
the differential measurements employed in the distance ladder approach for both
Cepheids and SNe Ia largely cancel out the effects of dust, making them robust tools
for measuring . H0 .
For Cepheids, which are observed in the Hubble Space Telescope (HST) Near
Infrared (NIR) bands, the impact of dust is minimal. The use of Wesenheit magnitudes
allows for the correction of dust effects based on directly observed quantities and the
small value of the reddening ratio (. R H ) in the NIR reduces the sensitivity to dust,
which has been well-characterized for Milky Way-like extinction. For SNe Ia, we
discuss the application of dust corrections to SNe Ia as a function of host-galaxy
properties and as an important component of accounting for SN Ia intrinsic scatter.
These dust-based corrections are applied consistently across rungs and uncertainties
in the dust model parameters are characterized and accounted for and do not account
for the observed Hubble tension.
Overall, the analysis presented in this paper reaffirms the robustness of current
. H0 measurements and suggests that dust-related effects are not a significant source
of tension in the determination of the Hubble Constant. However, we emphasize the
importance of continued refinement in our understanding of dust properties and dust
variation for cosmological studies beyond . H0 measurements.

Acknowledgements Brout and Riess would like to thank Doug Finkbeiner for his useful com-
ments on this chapter and discussions on the topic in general. This is a preprint of the following
chapter: Dillon Brout and Adam Riess, The Impact of Dust on Cosmic Distance Indicators, pub-
lished in The Hubble Constant Tension, edited by Eleonora DiValentino and Dillon Brout, expected
2024, Springer Nature reproduced with permission of Springer Nature Singapore Pte Ltd. The final
authenticated version will be available online at: https://doi.org/10.48550/arXiv.2311.08253.
19 The Impact of Dust on Cepheid and Type Ia Supernova Distances 379

References

1. A.G. Riess, W. Yuan, L.M. Macri, D. Scolnic, D. Brout, S. Casertano, D.O. Jones, Y. Murakami,
G.S. Anand, L. Breuval, T.G. Brink, A.V. Filippenko, S. Hoffmann, S.W. Jha, W. D’arcy
Kenworthy, J. Mackenty, B.E. Stahl, W.K. Zheng, A comprehensive measurement of the local
value of the Hubble constant with 1 km s.−1 Mpc.−1 uncertainty from the Hubble space telescope
and the S. H0 ES team. ApJ 934(1), L7 (2022)
2. S. Salim, D. Narayanan, The dust attenuation law in galaxies. Ann. Rev. Astron. Astrophys.
58, 529–575 (2020)
3. J.C. Weingartner, B.T. Draine, Dust grain–size distributions and extinction in the milky way,
large magellanic cloud, and small magellanic cloud. The Astrophys. J. 548(1), 296–309 (2001)
4. D.J. Schlegel, D.P. Finkbeiner, M. Davis, Maps of dust infrared emission for use in estimation of
reddening and cosmic microwave background radiation foregrounds. The Astrophys. J. 500(2),
525–553 (1998)
5. E.F. Schlafly, D.P. Finkbeiner, Measuring reddening with Sloan digital sky survey stellar spectra
and recalibrating SFD. ApJ 737(2), 103 (2011)
6. E.F. Schlafly, D.P. Finkbeiner, D.J. Schlegel, M. Jurić, Željko Ivezić, R.R. Gibson, G.R. Knapp,
B.A. Weaver, The blue tip of the stellar locus: measuring reddening with the Sloan digital sky
survey. The Astrophys. J. 725(1), 1175–1191 (2010)
7. E.L. Fitzpatrick, Correcting for the effects of interstellar extinction. Publications of the Astro-
nomical Society of the Pacific 111(755), 63–75 (1999)
8. J.E. O’Donnell, R v-dependent optical and near-ultraviolet extinction. ApJ 422, 158 (1994)
9. J.A. Cardelli, G.C. Clayton, J.S. Mathis, The relationship between infrared, optical, and ultra-
violet extinction. ApJ 345, 245 (1989)
10. Planck Collaboration, N. Aghanim, M. Ashdown, J. Aumont, C. Baccigalupi, M. Ballardini,
A.J. Banday, R.B. Barreiro, N. Bartolo, S. Basak, K. Benabed, J.-P. Bernard, M. Bersanelli,
P. Bielewicz, L. Bonavera, J.R. Bond, J. Borrill, F.R. Bouchet, F. Boulanger, C. Burigana, E.
Calabrese, J.-F. Cardoso, J. Carron, H.C. Chiang, L.P.L. Colombo, B. Comis, F. Couchot, A.
Coulais, B.P. Crill, A. Curto, F. Cuttaia, P. de Bernardis, G. de Zotti, J. Delabrouille, E. Di
Valentino, C. Dickinson, J.M. Diego, O. Doré, M. Douspis, A. Ducout, X. Dupac, S. Dusini, F.
Elsner, T.A. Enßlin, H.K. Eriksen, E. Falgarone, Y. Fantaye, F. Finelli, F. Forastieri, M. Frailis,
A.A. Fraisse, E. Franceschi, A. Frolov, S. Galeotta, S. Galli, K. Ganga, R.T. Génova-Santos,
M. Gerbino, T. Ghosh, Y. Giraud-Héraud, J. González-Nuevo, K.M. Górski, A. Gruppuso, J.E.
Gudmundsson, F.K. Hansen, G. Helou, S. Henrot-Versillé, D. Herranz, E. Hivon, Z. Huang,
A.H. Jaffe, W.C. Jones, E. Keihänen, R. Keskitalo, K. Kiiveri, T.S. Kisner, N. Krachmalnicoff,
M. Kunz, H. Kurki-Suonio, J.-M. Lamarre, M. Langer, A. Lasenby, M. Lattanzi, C.R. Lawrence,
M. Le Jeune, F. Levrier, P.B. Lilje, M. Lilley, V. Lindholm, M. López-Caniego, Y.-Z. Ma,
J.F. Macías-Pérez, G. Maggio, D. Maino, N. Mandolesi, A. Mangilli, M. Maris, P.G. Martin,
E. Martínez-González, S. Matarrese, N. Mauri, J.D. McEwen, A. Melchiorri, A. Mennella,
M. Migliaccio, M.-A. Miville-Deschênes, D. Molinari, A. Moneti, L. Montier, G. Morgante,
A. Moss, P. Natoli, C.A. Oxborrow, L. Pagano, D. Paoletti, G. Patanchon, O. Perdereau, L.
Perotto, V. Pettorino, F. Piacentini, S. Plaszczynski, L. Polastri, G. Polenta, J.-L. Puget, J.P.
Rachen, B. Racine, M. Reinecke, M. Remazeilles, A. Renzi, G. Rocha, C. Rosset, M. Rossetti,
G. Roudier, J.A. Rubiño-Martín, B. Ruiz-Granados, L. Salvati, M. Sandri, M. Savelainen,
D. Scott, C. Sirignano, G. Sirri, J.D. Soler, L.D. Spencer, A.-S. Suur-Uski, J.A. Tauber, D.
Tavagnacco, M. Tenti, L. Toffolatti, M. Tomasi, M. Tristram, T. Trombetti, J. Valiviita, F.
Van Tent, P. Vielva, F. Villa, N. Vittorio, B.D. Wandelt, I.K. Wehus, A. Zacchei, A. Zonca,
Planck intermediate results—xlviii. Disentangling galactic dust emission and cosmic infrared
background anisotropies. A&A 596, A109 (2016)
11. Gregory M. Green, Edward Schlafly, Catherine Zucker, Joshua S. Speagle, Douglas Finkbeiner,
A 3d dust map based on Gaia, Pan-STARRS 1, and 2MASS. The Astrophys. J. 887(1), 93 (2019)
12. J.L. Vergely, R. Lallement, N.L.J. Cox, Three-dimensional extinction maps: Inverting inter-
calibrated extinction catalogues. Astron. Astrophys. 664, A174 (2022)
380 D. Brout and A. Riess

13. G. Edenhofer, C. Zucker, P. Frank, A.K. Saydjari, J.S. Speagle, D. Finkbeiner, T. Enßlin, A
parsec-scale galactic 3d dust map out to 1.25 kpc from the sun (2023)
14. Y.-K. Chiang, B. Mé nard, Extragalactic imprints in galactic dust maps. The Astrophys. J.
870(2), 120 (2019)
15. B.T. Draine. Scattering by interstellar dust grains. I. Optical and ultraviolet. The Astrophys. J.
598(2), 1017 (2003)
16. E.F. Schlafly, A.M. Meisner, A.M. Stutz, J. Kainulainen, J.E.G. Peek, K. Tchernyshyov, H.W.
Rix, D.P. Finkbeiner, K.R. Covey, G.M. Green, E.F. Bell, W.S. Burgett, K.C. Chambers, P.W.
Draper, H. Flewelling, K.W. Hodapp, N. Kaiser, E.A. Magnier, N.F. Martin, N. Metcalfe, R.J.
Wainscoat, C. Waters, The optical-infrared extinction curve and its variation in the milky way.
ApJ 821(2), 78 (2016)
17. J. Gao, B.W. Jiang, A. Li, M.Y. Xue, The mid-infrared extinction law in the large Magellanic
cloud. ApJ 776(1), 7 (2013)
18. P.Y.Merica-Jones, K.M. Sandstrom, L.C. Johnson, J. Dalcanton, A.E. Dolphin, K. Gordon, J.
Roman-Duval, D.R. Weisz, B.F. Williams, The small Magellanic cloud investigation of dust
and gas evolution (SMIDGE): the dust extinction curve from red clump stars. ApJ 847(2), 102
(2017)
19. S. Salim, M. Boquien, J.C. Lee, Dust attenuation curves in the local universe: demographics
and new laws for star-forming galaxies and high-redshift analogs. ApJ 859(1), 11 (2018)
20. A. Cikota, S. Deustua, F. Marleau, Determining type Ia supernova host galaxy extinction
probabilities and a statistical approach to estimating the absorption-to-reddening ratio R.V .
ApJ 819(2), 152 (2016)
21. X. Wang, L. Wang, Y.-Q. Lou, X. Zhou, Z. Li, A novel color parameter as a luminosity calibrator
for type Ia supernovae. ApJ 620(2), L87–L90 (2005)
22. K. Krisciunas, J.L. Prieto, P.M. Garnavich, J.-L. Riley, A. Rest, C. Stubbs, R. McMillan,
Photometry of the type Ia supernovae 1999cc, 1999cl, and 2000cf. AJ 131(3), 1639–1647
(2006)
23. S. Nobili, A. Goobar, The colour-lightcurve shape relation of type ia supernovae and the
reddening law. Astron. Astrophys. 487(1), 19–31 (2008)
24. R. Amanullah, J. Johansson, A. Goobar, R. Ferretti, S. Papadogiannakis, T. Petrushevska, P.J.
Brown, Y. Cao, C. Contreras, H. Dahle, N. Elias-Rosa, J.P.U. Fynbo, J. Gorosabel, L. Guaita,
L. Hangard, D.A. Howell, E.Y. Hsiao, E. Kankare, M. Kasliwal, G. Leloudas, P. Lundqvist,
S. Mattila, P. Nugent, M.M. Phillips, A. Sand berg, V. Stanishev, M. Sullivan, F. Taddia,
G. Östlin, S. Asadi, R. Herrero-Illana, J.J. Jensen, K. Karhunen, S. Lazarevic, E. Varenius,
P. Santos, S. Seethapuram Sridhar, S.H.J. Wallström, J. Wiegert, Diversity in extinction laws
of type Ia supernovae measured between 0.2 and 2 .μm. Mon. Not. Roy. Astron. Soc. 453(3),
3300–3328 (2015)
25. L. Wang, Dust around type Ia supernovae. ApJ 635(1), L33–L36 (2005)
26. Ariel Goobar, Low R.V from Circumstellar Dust around Supernovae. ApJ 686(2), L103 (October
2008)
27. M.M. Phillips, J.D. Simon, N. Morrell, C.R. Burns, N.L.J. Cox, R.J. Foley, A.I. Karakas,
F. Patat, A. Sternberg, R.E. Williams, A. Gal-Yam, E.Y. Hsiao, D.C. Leonard, S.E. Persson,
M. Stritzinger, I.B. Thompson, A. Campillay, C. Contreras, G. Folatelli, W.L. Freedman, M.
Hamuy, M. Roth, G.A. Shields, N.B. Suntzeff, L. Chomiuk, I.I. Ivans, B.F. Madore, B.E.
Penprase, D. Perley, G. Pignata, G. Preston, A.M. Soderberg, On the source of the dust extinction
in type Ia supernovae and the discovery of anomalously strong Na I absorption. ApJ 779(1),
38 (2013)
28. M. Bulla, A. Goobar, R. Amanullah, U. Feindt, R. Ferretti, Estimating dust distances to type Ia
supernovae from colour excess time evolution. Mon. Not. Roy. Astron. Soc. 473(2), 1918–1929
(2018)
29. M. Bulla, A. Goobar, S. Dhawan, Shedding light on the type Ia supernova extinction puzzle:
dust location found. Mon. Not. Roy. Astron. Soc. 479(3), 3663–3674 (2018)
30. B.F. Madore, The period-luminosity relation. IV. Intrinsic relations and reddenings for the large
Magellanic cloud cepheids. ApJ 253, 575–579 (1982)
19 The Impact of Dust on Cepheid and Type Ia Supernova Distances 381

31. B. Follin, L. Knox, Insensitivity of the distance ladder Hubble constant determination to cepheid
calibration modelling choices. Monthly Notices of the Royal Astron. Soc. 477(4), 4534–4542
(2018)
32. E. Mörtsell, A. Goobar, J. Johansson, S. Dhawan, Sensitivity of the Hubble constant determi-
nation to Cepheid calibration. The Astrophys. J. 933(2), 212 (2022)
33. A.G. Riess, S. Casertano, W. Yuan, J.B. Bowers, L. Macri, J.C. Zinn, D. Scolnic. Cosmic
distances calibrated to 1% precision with gaia EDR3 parallaxes and Hubble space telescope
photometry of 75 milky way Cepheids confirm tension with .λCDM. The Astrophys. J. Lett.
908(1), L6 (2021)
34. W. Yuan, A.G. Riess, S. Casertano, L.M. Macri, A first look at Cepheids in a type ia supernova
host with JWST. The Astrophys. J. Lett. 940(1), L17 (2022)
35. A.G. Riess, G.S. Anand, W. Yuan, S. Casertano, A. Dolphin, L.M. Macri, L. Breuval, D.
Scolnic, M. Perrin, R.I. Anderson, Crowded no more: the accuracy of the Hubble constant
tested with high resolution observations of Cepheids by JWST (2023)
36. W.D. Kenworthy, D.O. Jones, M. Dai, R. Kessler, D. Scolnic, D. Brout, M.R. Siebert, J.D.R.
Pierel, K.G. Dettman, G. Dimitriadis, R.J. Foley, S.W. Jha, Y.C. Pan, A. Riess, S. Rodney, C.
Rojas-Bravo, SALT3: an improved type Ia supernova model for measuring cosmic distances.
ApJ 923(2), 265 (2021)
37. J. Marriner, J.P. Bernstein, R. Kessler, H. Lampeitl, R. Miquel, J. Mosher, R.C. Nichol, M. Sako,
D.P. Schneider, M. Smith. A more general model for the intrinsic scatter in type Ia supernova
distance moduli. The Astrophys. J. 740(2), 72 (2011)
38. R. Tripp, A two-parameter luminosity correction for type IA supernovae. A&A 331, 815–820
(1998)
39. M. Hicken, P. Challis, S. Jha, R.P. Kirshner, T. Matheson, M. Modjaz, A. Rest, W.M. Wood-
Vasey, G. Bakos, E.J. Barton, P. Berlind, A. Bragg, C. Briceño, W.R. Brown, N. Caldwell, M.
Calkins, R. Cho, L. Ciupik, M. Contreras, K.-C. Dendy, A. Dosaj, N. Durham, K. Eriksen, G.
Esquerdo, M. Everett, E. Falco, J. Fernandez, A. Gaba, P. Garnavich, G. Graves, P. Green, T.
Groner, C. Hergenrother, M.J. Holman, V. Hradecky, J. Huchra, B. Hutchison, D. Jerius, A.
Jordan, R. Kilgard, M. Krauss, K. Luhman, L. Macri, D. Marrone, J. McDowell, D. McIntosh,
B. McNamara, T. Megeath, B. Mochejska, D. Munoz, J. Muzerolle, O. Naranjo, G. Narayan,
M. Pahre, W. Peters, D. Peterson, K. Rines, B. Ripman, A. Roussanova, R. Schild, A. Sicilia-
Aguilar, J. Sokoloski, K. Smalley, A. Smith, T. Spahr, K.Z. Stanek, P. Barmby, S. Blondin,
C.W. Stubbs, A. Szentgyorgyi, M.A.P. Torres, A. Vaz, A. Vikhlinin, Z. Wang, M. Westover, D.
Woods, P. Zhao, CfA3: 185 type Ia supernova light curves from the CfA. ApJ 700, 331–357
(2009)
40. M. Sullivan, A. Conley, D.A. Howell, J.D. Neill, P. Astier, C. Balland, S. Basa, R.G. Carlberg,
D. Fouchez, J. Guy, D. Hardin, I.M. Hook, R. Pain, N. Palanque-Delabrouille, K.M. Perrett,
C.J. Pritchet, N. Regnault, J. Rich, V. Ruhlmann-Kleider, S. Baumont, E. Hsiao, T. Kronborg,
C. Lidman, S. Perlmutter, E.S. Walker, The dependence of type Ia supernovae luminosities on
their host galaxies. Mon. Not. Roy. Astron. Soc. 406, 782–802 (2010)
41. H. Lampeitl, M. Smith, R.C. Nichol, B. Bassett, D. Cinabro, B. Dilday, R.J. Foley, J.A. Frieman,
P.M. Garnavich, A. Goobar, M. Im, S.W. Jha, J. Marriner, R. Miquel, J. Nordin, L. Östman,
A.G. Riess, M. Sako, D.P. Schneider, J. Sollerman, M. Stritzinger, The effect of host galaxies
on type Ia supernovae in the SDSS-II supernova survey. ApJ 722, 566–576 (2010)
42. M. Childress, G. Aldering, P. Antilogus, C. Aragon, S. Bailey, C. Baltay, S. Bongard, C. Buton,
A. Canto, F. Cellier-Holzem, N. Chotard, Y. Copin, H.K. Fakhouri, E. Gangler, J. Guy, E.Y.
Hsiao, M. Kerschhaggl, A.G. Kim, M. Kowalski, S. Loken, P. Nugent, K. Paech, R. Pain,
E. Pecontal, R. Pereira, S. Perlmutter, D. Rabinowitz, M. Rigault, K. Runge, R. Scalzo, G.
Smadja, C. Tao, R.C. Thomas, B.A. Weaver, C. Wu, Host galaxies of type Ia supernovae from
the nearby supernova factory. ApJ 770, 107 (2013)
43. R. Kessler, D. Brout, S. Crawford et al., First cosmology results using type Ia supernova from the
dark energy survey: simulations to correct supernova distance biases. Mon. Not. Roy. Astron.
Soc. 485, 1171–1187 (2019)
382 D. Brout and A. Riess

44. Dillon Brout, Daniel Scolnic, It’s dust: solving the mysteries of the intrinsic scatter and host-
galaxy dependence of standardized type Ia supernova brightnesses. Astrophys. J. 909(1), 26
(2021)
45. D. Brout, D. Scolnic, B. Popovic, A.G. Riess, A. Carr, J. Zuntz, R. Kessler, T.M. Davis,
S. Hinton, D. Jones, W. D’Arcy Kenworthy, E.R. Peterson, K. Said, G. Taylor, N. Ali, P.
Armstrong, P. Charvu, A. Dwomoh, C. Meldorf, A. Palmese, H. Qu, B.M. Rose, B. Sanchez,
C.W. Stubbs, M. Vincenzi, C.M. Wood, P.J. Brown, R. Chen, K. Chambers, D.A. Coulter,
M. Dai, G. Dimitriadis, A.V. Filippenko, R.J. Foley, S.W. Jha, L. Kelsey, R.P. Kirshner, A.
Möller, J. Muir, S. Nadathur, Y.-C. Pan, A. Rest, C. Rojas-Bravo, M. Sako, M.R. Siebert, M.
Smith, B.E. Stahl, P. Wiseman, The Pantheon+ analysis: cosmological constraints. ApJ 938(2),
110 (2022)
46. J. Guy, M. Sullivan, A. Conley, N. Regnault, P. Astier, C. Balland, S. Basa, R.G. Carlberg, D.
Fouchez, D. Hardin, I.M. Hook, D.A. Howell, R. Pain, N. Palanque-Delabrouille, K.M. Perrett,
C.J. Pritchet, J. Rich, V. Ruhlmann-Kleider, D. Balam, S. Baumont, R.S. Ellis, S. Fabbro,
H.K. Fakhouri, N. Fourmanoit, S. González-Gaitán, M.L. Graham, E. Hsiao, T. Kronborg, C.
Lidman, A.M. Mourao, S. Perlmutter, P. Ripoche, N. Suzuki, E.S. Walker, The supernova legacy
survey 3-year sample: type Ia supernovae photometric distances and cosmological constraints.
A&A 523, A7 (2010)
47. A. Conley, J. Guy, M. Sullivan, N. Regnault, P. Astier, C. Balland, S. Basa, R.G. Carlberg, D.
Fouchez, D. Hardin et al., Supernova constraints and systematic uncertainties from the first
three years of the supernova legacy survey. The Astrophys. J. Suppl. Series 192(1), 1 (2010)
48. B. Popovic, D. Brout, R. Kessler, D. Scolnic, The pantheon+ analysis: forward-modeling the
dust and intrinsic colour distributions of type Ia supernovae, and quantifying their impact on
cosmological inferences (2021)
49. R. Chen, D. Scolnic, E. Rozo, E. S. Rykoff, B. Popovic, R. Kessler, M. Vincenzi, T. M.
Davis, P. Armstrong, D. Brout, L. Galbany, L. Kelsey, C. Lidman, A. Möller, B. Rose,
M. Sako, M. Sullivan, G. Taylor, P. Wiseman, J. Asorey, A. Carr, C. Conselice, K. Kuehn,
G.F. Lewis, E. Macaulay, M. Rodriguez-Monroy, B.E. Tucker, T.M.C. Abbott, M. Aguena,
S. Allam, F. Andrade-Oliveira, J. Annis, D. Bacon, E. Bertin, S. Bocquet, D. Brooks, D.L.
Burke, A. Carnero Rosell, M. Carrasco Kind, J. Carretero, R. Cawthon, M. Costanzi, L.N.
da Costa, M.E.S. Pereira, S. Desai, H.T. Diehl, P. Doel, S. Everett, I. Ferrero, B. Flaugher,
D. Friedel, J. Frieman, J. Garcí a-Bellido, M. Gatti, E. Gaztanaga, D. Gruen, S.R. Hinton,
D.L. Hollowood, K. Honscheid, D.J. James, O. Lahav, M. Lima, M. March, F. Menanteau,
R. Miquel, R. Morgan, A. Palmese, F. Paz-Chinchón, A. Pieres, A.A. Plazas Malagón, J. Prat,
A.K. Romer, A. Roodman, E. Sanchez, M. Schubnell, S. Serrano, I. Sevilla-Noarbe, M. Smith,
M. Soares-Santos, E. Suchyta, G. Tarle, D. Thomas, C. To, D.L. Tucker, T.N. Varga, Measur-
ing cosmological parameters with type Ia supernovae in redMaGiC galaxies. The Astrophys.
J. 938(1), 62 (2022)
50. C. Meldorf, A. Palmese, D. Brout, R. Chen, D. Scolnic, L. Kelsey, L. Galbany, W.G. Hart-
ley, T.M. Davis, A. Drlica-Wagner, M. Vincenzi, J. Annis, M. Dixon, O. Graur, C. Lidman,
A. Möller, P. Nugent, B. Rose, M. Smith, S. Allam, D.L. Tucker, J. Asorey, J. Calcino, D. Car-
ollo, K. Glazebrook, G.F. Lewis, G. Taylor, B.E. Tucker, A.G. Kim, H. T Diehl, M. Aguena,
F. Andrade-Oliveira, D. Bacon, E. Bertin, S. Bocquet, D. Brooks, D.L. Burke, J. Carretero,
M. Carrasco Kind, F.J. Castander, M. Costanzi, L.N. da Costa, S. Desai, P. Doel, S. Everett,
I. Ferrero, D. Friedel, J. Frieman, J. Garcí a-Bellido, M. Gatti, D. Gruen, J. Gschwend, G. Gutier-
rez, S.R. Hinton, D.L. Hollowood, K. Honscheid, D.J. James, K. Kuehn, M. March, J.L. Mar-
shall, F. Menanteau, R. Miquel, R. Morgan, F. Paz-Chinchón, M.E.S. Pereira, A.A. Plazas
Malagón, E. Sanchez, V. Scarpine, I. Sevilla-Noarbe, E. Suchyta, G. Tarle, T.N. Varga, The
dark energy survey supernova program results: type ia supernova brightness correlates with
host galaxy dust. Monthly Notices of the Royal Astron. Soc. 518(2), 1985–2004 (2022)
51. S. Thorp, K.S. Mandel, Constraining the SN Ia host galaxy dust law distribution and mass step:
hierarchical Bayesn analysis of optical and near-infrared light curves. Monthly Notices of the
Royal Astron. Soc. 517(2), 2360–2382 (2022)
19 The Impact of Dust on Cepheid and Type Ia Supernova Distances 383

52. B. Popovic, D. Brout, R. Kessler, D. Scolnic, L. Lu, Improved treatment of host-galaxy corre-
lations in cosmological analyses with type Ia supernovae (2021)
53. M. Rigault, Y. Copin, G. Aldering, P. Antilogus, C. Aragon, S. Bailey, C. Baltay, S. Bongard,
C. Buton, A. Canto, F. Cellier-Holzem, M. Childress, N. Chotard, H.K. Fakhouri, U. Feindt,
M. Fleury, E. Gangler, P. Greskovic, J. Guy, A.G. Kim, M. Kowalski, S. Lombardo, J. Nordin,
P. Nugent, R. Pain, E. Pécontal, R. Pereira, S. Perlmutter, D. Rabinowitz, K. Runge, C. Saun-
ders, R. Scalzo, G. Smadja, C. Tao, R.C. Thomas, B.A. Weaver, Evidence of environmental
dependencies of type Ia supernovae from the nearby supernova factory indicated by local H.α.
A&A 560, A66 (2013)
54. A. Polin, P. Nugent, D. Kasen, Observational predictions for sub-chandrasekhar mass explo-
sions: Further evidence for multiple progenitor systems for type Ia supernovae. The Astrophys.
J. 873(1), 84 (2019)
55. C.R. Burns, E. Parent, M.M. Phillips, M. Stritzinger, K. Krisciunas, N.B. Suntzeff, E.Y. Hsiao,
C. Contreras, J. Anais, L. Boldt, L. Busta, A. Campillay, S. Castelló n, G. Folatelli, W.L.
Freedman, C. González, M. Hamuy, P. Heoflich, W. Krzeminski, B.F. Madore, N. Morrell, S.E.
Persson, M. Roth, F. Salgado, J. Serón, S. Torres, The Carnegie supernova project: absolute
calibration and the Hubble constant. The Astrophys. J. 869(1), 56 (2018)
56. S. Dhawan, S. Thorp, K.S. Mandel, S.M. Ward, G. Narayan, S.W. Jha, T. Chant, A BayeSN
distance ladder: . H0 from a consistent modelling of type ia supernovae from the optical to the
near-infrared. Monthly Notices of the Royal Astron. Soc. 524(1), 235–244 (2023)
57. P. Garnavich, C.M. Wood, P. Milne, J.B. Jensen, J.P. Blakeslee, P.J. Brown, D. Scolnic, B. Rose,
D. Brout, Connecting infrared surface brightness fluctuation distances to type Ia supernova
hosts: testing the top rung of the distance ladder. The Astrophys. J. 953(1), 35 (2023)
 Part IV
Theoretical Proposed Solutions
Chapter 20
(Introduction to the Second Part
of the Book) What About the Solutions?

Eleonora Di Valentino

Out of various cosmological models proposed in literature, the Lambda cold dark
matter (.ɅCDM) stands out as the standard model, chosen for its simplicity and
remarkable ability to accurately describe a wide array of astrophysical and cosmo-
logical observations.
However, despite its success, .ɅCDM grapples with numerous unknowns, strug-
gling to elucidate fundamental concepts pertaining to the structure and evolution of
the universe. These conceptual challenges arise from three elusive ingredients, unsup-
ported by theoretical first principles or laboratory experiments, but rather inferred
from cosmological and astrophysical observations.
These three enigmatic components are: inflation, dark matter (DM), and dark
energy (DE). In the realm of.ɅCDM, inflation is given by a single, slow-rolling scalar
field; DM is postulated to interact only through gravity, be cold and pressureless, and
lack direct evidence of its existence; DE is represented by the cosmological constant
term .Ʌ, offering little in the way of a robust physical explanation.
While.ɅCDM stands as the preferred model for its adeptness in accurately describ-
ing observed phenomena, it is not immune to theoretical shortcomings. The model,
characterized by its six parameters, lacks a foundation in deep-rooted physical prin-
ciples and should, at best, be regarded as an approximation of an as-yet-undiscovered
underlying physical theory.
As observations multiply and attain greater accuracy, it is anticipated that devi-
ations from the .ɅCDM model will come to light [1, 2]. Indeed, discrepancies in
pivotal cosmological parameters, such as . H0 [3–6], have already emerged across
various observations, each carrying distinct statistical significance.

E. Di Valentino (B)
School of Mathematics and Statistics, University of Sheffield, Hounsfield Road, Sheffield S3
7RH, Sheffield, UK
e-mail: e.divalentino@sheffield.ac.uk

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 387
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_20
388 E. Di Valentino

While certain tensions observed in cosmological measurements may originate

from systematic errors, their persistent manifestation across multiple probes sug-
gests potential flaws in the conventional cosmological framework. This raises the
intriguing possibility that new physics might be essential to elucidate these observa-
tional discrepancies. Therefore, the enduring nature of these tensions raises questions
about the adequacy of the established .ɅCDM model.
Considering the diverse array of the Hubble constant measurements detailed in
prior chapters, where late universe observations are consistently higher than early
ones and vice versa, pinpointing a singular systematic error to consistently explain
the observed disparities becomes challenging. This complexity makes it difficult to
envisage a single source of error that uniformly accounts for the discrepancies across
the broad spectrum of measurements encountered earlier, therefore addressing the
Hubble constant tension.
Persisting at a significance level of 5–6.3 .σ [7–9], even after eliminating mea-
surements related to specific object types, teams, or calibrations, the Hubble constant
tension remains robust. Identifying a singular error to reconcile this tension becomes
an intricate task. While the possibility of multiple independent systematic errors
offers some flexibility in resolving the tension, such occurrences are inherently less
probable.
Lastly, recognizing that the indirect early universe constraints are model-
dependent, there exists an opportunity to explore the prospect of expanding the
cosmological framework. Examining potential extensions and their capacity to rec-
oncile the observed discrepancies among various cosmological probes becomes a
promising avenue for further investigation.
In the subsequent chapters of this book, it will become evident that while the seem-
ingly straightforward resolution-such as the existence of a local void to address the. H0
tension-is discredited at a significance level of approximately .20σ [10], alternative
propositions can be categorized into two broad groups: those involving modifications
to the expansion history during early times [11–14] (Chap. 22) (prior to recombina-
tion) and those occurring during late times [15] (post-recombination). Additionally,
there is the option of considering adjustments to the recombination epoch itself [16,
17] or the inflationary period (Chap. 34). Regardless of the chosen avenue, it is
essential to note that, at present, no specific proposal has emerged as a compelling
frontrunner, demonstrating a significantly higher likelihood or superiority over oth-
ers. The exploration of these possibilities in subsequent chapters will shed light on
the nuanced complexities of each proposal and their implications for resolving the
persistent challenges posed by the . H0 tension.
Indeed, while late-time solutions exhibit an increase in . H0 by reducing the expan-
sion history at intermediate redshifts, they fall short in providing a satisfactory fit
to low-redshift measurements, such as Baryon Acoustic Oscillations (BAO) or Type
Ia Supernovae data [18]. Conversely, early-time solutions exhibit promising adjust-
ments to both parameters, but they are unable to completely resolve the . H0 tension
between Planck Cosmic Microwave Background (CMB) data [19] and the S. H0 ES
collaboration measurement [20] without introducing an additional Hubble constant
prior. These solutions tend to enhance the consistency between data sets primarily
20 (Introduction to the Second Part of the Book) What About the Solutions? 389

due to a volume effect but exacerbate the . S8 tension [22] between the CMB and weak
lensing data [21]. While the subsequent chapters explore also intriguing proposals
involving negative dark energy densities (Chap. 28), interactions between differ-
ent components [23], and modified gravity theories (Chaps. 30 and 31), it becomes
increasingly evident that relying solely on early or late-time solutions proves insuffi-
cient [24]. The complexities of the observed tensions necessitate a more comprehen-
sive approach that may involve a combination of these strategies or the exploration
of alternative theoretical frameworks.
Adding to the intricacies, caution is warranted when delving into exotic models.
The correlation between parameters in such models may yield artificial signals that
are not present in nature, introducing the risk of false detections for new physics [25].
Additionally, the hidden model dependence of certain datasets must be considered,
as some datasets assume the .ɅCDM model in various steps of the data reduction
process, exemplified in techniques like BAO [26]. Notably, the use of a . H0 prior
for late-time transitions resembling a hockey stick shape is under scrutiny, with
suggestions for its replacement by the . M B value [27]. Further complexity arises from
anomalies in CMB data, including the Planck. A Lens problem [28], and inconsistencies
between different CMB experiments that may produce significantly disparate results
in extended cosmologies [29].
The Hubble constant tension not only emphasizes the need for new observations
but also serve as catalysts for probing alternative theoretical models and solutions.
The current state of the field suggests that while the game is still open, a paradigm shift
in cosmology is not beyond the realm of possibility in the near future. The ongoing
investigation into these challenges holds the potential to reshape our understanding
of the universe.

Acknowledgements EDV is supported by a Royal Society Dorothy Hodgkin Research Fellowship.

References

1. E. Abdalla, G. Franco Abellán, A. Aboubrahim, A. Agnello, O. Akarsu, Y. Akrami, G. Alestas,

D. Aloni, L. Amendola, L.A. Anchordoqui et al., JHEAp 34, 49–211 (2022). https://doi.org/
10.1016/j.jheap.2022.04.002 [arXiv:2203.06142 [astro-ph.CO]]
2. L. Perivolaropoulos, F. Skara, New Astron. Rev. 95, 101659 (2022). https://doi.org/10.1016/j.
newar.2022.101659 [arXiv:2105.05208 [astro-ph.CO]]
3. L. Verde, T. Treu, A.G. Riess, Nature Astron. 3, 891. https://doi.org/10.1038/s41550-019-
0902-0 [arXiv:1907.10625 [astro-ph.CO]]
4. E. Di Valentino, L.A. Anchordoqui, O. Akarsu, Y. Ali-Haimoud, L. Amendola, N. Arendse, M.
Asgari, M. Ballardini, S. Basilakos, E. Battistelli et al., Astropart. Phys. 131, 102605 (2021).
https://doi.org/10.1016/j.astropartphys.2021.102605 [arXiv:2008.11284 [astro-ph.CO]]
5. E. Di Valentino, O. Mena, S. Pan, L. Visinelli, W. Yang, A. Melchiorri, D. F. Mota, A. G. Riess,
J. Silk, Class. Quant. Grav. 38(15), 153001 (2021). https://doi.org/10.1088/1361-6382/ac086d
[arXiv:2103.01183 [astro-ph.CO]]
6. M. Kamionkowski, A.G. Riess, Ann. Rev. Nucl. Part. Sci. 73, 153–180 (2023)
[arXiv:2211.04492 [astro-ph.CO]]
390 E. Di Valentino

7. A.G. Riess, Nature Rev. Phys. 2(1), 10-12 (2019). https://doi.org/10.1038/s42254-019-0137-

0 [arXiv:2001.03624 [astro-ph.CO]]
8. E. Di Valentino, Mon. Not. Roy. Astron. Soc. 502(2), 2065–2073 (2021). https://doi.org/10.
1093/mnras/stab187 [arXiv:2011.00246 [astro-ph.CO]]
9. E. Di Valentino, Universe 8(8), 399 (2022). https://doi.org/10.3390/universe8080399
10. D. Huterer, H.Y. Wu, [arXiv:2309.05749 [astro-ph.CO]]
11. F. Niedermann, M.S. Sloth, [arXiv:2307.03481 [hep-ph]]
12. S. Gariazzo, O. Mena, [arXiv:2306.15067 [astro-ph.CO]]
13. A. Nygaard, E.B. Holm, T. Tram, S. Hannestad, [arXiv:2307.00418 [astro-ph.CO]]
14. K. Naidoo, [arXiv:2308.13617 [astro-ph.CO]]
15. M. Raveri, [arXiv:2309.06795 [astro-ph.CO]]
16. K. Jedamzik, L. Pogosian, [arXiv:2307.05475 [astro-ph.CO]]
17. J. Chluba, L. Hart, [arXiv:2309.12083 [astro-ph.CO]]
18. L. Knox, M. Millea, Phys. Rev. D 101(4), 043533 (2020). https://doi.org/10.1103/PhysRevD.
101.043533 [arXiv:1908.03663 [astro-ph.CO]]
19. N. Aghanim et al., [Planck], Astron. Astrophys. 641, A6 (2020) [erratum: Astron. Astrophys.
652, C4 (2021)]. https://doi.org/10.1051/0004-6361/201833910 [arXiv:1807.06209 [astro-
ph.CO]]
20. A.G. Riess, W. Yuan, L. M. Macri, D. Scolnic, D. Brout, S. Casertano, D.O. Jones, Y. Murakami,
L. Breuval, T.G. Brink et al., Astrophys. J. Lett. 934(1), L7 (2022). https://doi.org/10.3847/
2041-8213/ac5c5b [arXiv:2112.04510 [astro-ph.CO]]
21. E. Di Valentino, L.A. Anchordoqui, Ö. Akarsu, Y. Ali-Haimoud, L. Amendola, N. Arendse, M.
Asgari, M. Ballardini, S. Basilakos, E. Battistelli et al., Astropart. Phys. 131, 102604 (2021).
https://doi.org/10.1016/j.astropartphys.2021.102604 [arXiv:2008.11285 [astro-ph.CO]]
22. K. Jedamzik, L. Pogosian, G.B. Zhao, Commun. in Phys. 4, 123 (2021). https://doi.org/10.
1038/s42005-021-00628-x [arXiv:2010.04158 [astro-ph.CO]]
23. S. Pan, W. Yang, [arXiv:2310.07260 [astro-ph.CO]]
24. S. Vagnozzi, Universe 9(9), 393 (2023). https://doi.org/10.3390/universe9090393
[arXiv:2308.16628 [astro-ph.CO]]
25. E. Di Valentino. O. Mena, Mon. Not. Roy. Astron. Soc. 500(1), L22-L26 (2020) https://doi.
org/10.1093/mnrasl/slaa175 [arXiv:2009.12620 [astro-ph.CO]]
26. S. Anselmi, G.D. Starkman, A. Renzi, Phys. Rev. D 107(12), 123506 (2023). https://doi.org/
10.1103/PhysRevD.107.123506 [arXiv:2205.09098 [astro-ph.CO]]
27. D. Camarena, V. Marra, [arXiv:2307.02434 [astro-ph.CO]]
28. G.E. Addison, C.L. Bennett, M. Halpern, G. Hinshaw, J.L. Weiland, [arXiv:2310.03127 [astro-
ph.CO]]
29. W. Giarè, [arXiv:2305.16919 [astro-ph.CO]]
Chapter 21
Not Empty Enough: A Local Void
Cannot Solve the H0 Tension .

Dragan Huterer and Hao-Yi Wu

21.1 Introduction

In the currently favored flat .ɅCDM cosmological model, the local measurement of
the Hubble constant . H0 from the distance ladder [1–3] and the global measurement
from the CMB anisotropies [4–7] should agree. The fact that they do not may well
be a harbinger of new physics—or, at the very least, unexpectedly large systematic
errors in either of the two principal measurements.
Hubble tension has sparked huge interest in cosmology, and a dizzying number
of explanations for it have been put forth (see Ref. [8], which refers to more than
1000 papers in this regard). However, no single explanation has yet proven to be
compelling. Most of the theoretical explanations try to change the global CMB . H0
fit from .∼ 67 km s−1 Mpc−1 to .∼ 73 km s−1 Mpc−1 , the value favored by the distance
ladder with SNIa. While doing so, these new models have to preserve the excellent
fit to CMB, baryon acoustic oscillations, and other cosmological data which—in an
unmodified, standard cosmological model—favor . H0 ∼ 67 km s−1 Mpc−1 . Even in
reasonably successful models, it has proven difficult to raise the Hubble constant
all the way to .∼ 73 km s−1 Mpc−1 (rather than halfway there, to .∼ 70 km s−1 Mpc−1 )
while at the same time improving the goodness of fit to the data to justify the new
parameter(s) of the models in question.
In this Chapter, we review arguably the simplest explanation for the Hubble
tension—the possibility that we live in a void empty enough to resolve the Hub-
ble tension. The logic of this explanation goes as follows. There are underdensities

D. Huterer (B)
Department of Physics and Leinweber Center for Theoretical Physics, University of Michigan,
450 Church St, Ann Arbor, MI 48103, USA
e-mail: huterer@umich.edu
H.-Y. Wu (B)
Department of Physics, Boise State University, Boise, ID 83725, USA
e-mail: hywu@boisestate.edu
© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 391
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_21
392 D. Huterer and H.-Y. Wu

and overdensities in the universe. If we live in an underdensity, the local Hubble

constant, . H0loc , will be higher than the global Hubble constant, as less mass in the
void implies a higher expansion rate. This would imply a possible mismatch between
loc
. H0 and the global, “true” . H0 . In addition, given that the local SNIa sample is highly

sparse and inhomogeneous in the 3D spatial distribution, the error bar on the local
. H0 might be larger than previously thought. The two effects—local matter density
and sample sparsity—constitute the sample variance (or cosmic variance) in the
local . H0 measurements.
Our goal is to precisely quantify the sample variance, and we do so in Sect. 21.4.
First, however, we review the physics behind the void explanation in Sect. 21.2, and
briefly review the observational evidence for a void in Sect. 21.3. We conclude in
Sect. 21.5.

21.2 The Impact of Underdensity on the Hubble Constant:

Theoretical Considerations

The “Hubble bubble” hypothesis refers to the possibility that Milky Way resides in an
underdense region in the universe, and the underdensity makes the local expansion
rate significantly different from the global expansion rate. If we live in such a void,
nearby galaxies would preferentially have positive peculiar velocities (moving away
from us faster than expected from the global expansion rate). This effect can cause
the local Hubble constant to be higher than the global value.
The effect of the local density of the universe on the locally measured Hubble
constant can be quantified as follows. In an underdense region, the expansion rate is
higher than the mean (and vice versa in an overdense region) [9–18]. Namely, the
relation between the local . H0 shift, .ΔH0 , and the local density contrast, .δ, is

ΔH0 1
. = − δ f (Ω M )Θ(δ, Ω M ), (21.1)
H0 3

where . f (Ω M ) ≃ Ω M (a)0.55 is the growth rate of density perturbations (see Ref. [19]
for a review), and .Θ = 1 − O(δ) is a non-linear correction which is small for under-
densities of typical size. Therefore, an underdensity with .δ < 0 will automatically
lead to a higher measured local expansion rate, that is, .ΔH0 > 0. The effect of an
under/overdensity described here is qualitatively guaranteed, as the local density
alone leads to location-dependent local measurements of . H0 . This makes the Hubble
bubble an appealing model to potentially explain why local measurements of the
Hubble constant disagree with the global measurement given by Planck.
Is it possible that sample variance provides a sufficiently large effect to explain
the Hubble tension? Over the past decade or so, the answer has been arrived at
with increasingly realistic numerical and analytical estimates. The effect of sample
variance on the local measurements of . H0 was pioneered by Refs. [20–22] using
21 Not Empty Enough: A Local Void Cannot Solve the . H0 Tension 393

numerical simulations. These analyses have strongly hinted that the underdensity
required to explain the Hubble tension is extremely unlikely. In parallel, theorists
have investigated the Hubble bubble scenario, relying on simplified analytical models
of overdensity. For example, some work adopted the top-hat spherical model [23]
or the Lemaitre–Tolman–Bondi (LTB) metric, a variant of the FLRW metric with
a spherically symmetric underdensity or overdensity [24–26].1 All of these results
indicate that it is extremely unlikely that we live in a void that is sufficiently large in
size and devoid of matter to explain the Hubble tension.
However, most of these past analyses of sample variance—both numerical and
theoretical—have modeled the presence of spherical underdensity but have not
attempted to mimic the local . H0 measurements particularly closely. Specifically,
these studies have been assuming spherical symmetry in the SNIa measurements.
However, merely visualizing the highly inhomogeneous distribution of SNIa in the
sky brings into question the accuracy of the homogeneity assumption that underlies
most past numerical and theoretical treatments. For example, SNIa distribution is
denser in some regions due to historically available observations (e.g. Sloan Digital
Sky Survey’s Stripe 82; see also, for example, Fig. 1 in Ref. [29]).
Before discussing our work addressing sample inhomogeneity, we briefly review
why a direct measurement of the Hubble bubble is challenging.

21.3 Do We Live in a Hubble Bubble? Observations

of the Local Density Contrast

As far as the observational evidence for the local density contrast is concerned, the
following quote by Arthur Eddington
Never trust an experimental result until it has been confirmed by theory
is perhaps very relevant here. This is because observational evidence for the Hubble
bubble must necessarily rely on the estimate of mass density, which is plagued by
many observational uncertainties. The principal challenge is twofold: to convert
from light in some electromagnetic band (visible, X-ray, radio, etc.) to mass, and
then to map out the three-dimensional distribution of the mass (rather than just the
2D projection on the sky). Despite the massive efforts by observers and modelers,
we are still not at the level of accuracy where such direct measurements can be made
to make robust claims about the Hubble bubble.
Some observations do hint that the Milky Way might be in an underdense region
in the universe, but the results have been controversial [30–37]. Several observations
use large-area surveys of galaxy number density to map the radial distribution of
mass centered around our location in the universe. For example, Ref. [34] uses the
600 deg.2 galaxy catalog from UKIDSS-LAS, finding a significant local underdensity
.δ = −0.3 within 200 Mpc. Ref. [31] uses the full-sky 2M++ catalog and finds no

1 Note that some studies combine theoretical and numerical tools [27, 28].
394 D. Huterer and H.-Y. Wu

evidence for the local underdensity out to 120 . h −1 Mpc. Ref. [27] uses the distance–
redshift relation derived from .∼ 1000 SNe from Pantheon, Foundation, and Carnegie
Supernova Project and derived that .|δ| < 27% (5.σ constraint) on scales larger than
69 . h −1 Mpc (.z = 0.023). In contrast, Ref. [36] find that the local density within
. z < 0.075 is 20% lower than the mean density of the universe, which would lead to
a 3% correction of the local Hubble constant.
Clearly, on the front of direct observations, things are unsettled, and there is a
wide divergence of interpretations of observational measurements. Nevertheless, in
the remainder of this review, we demonstrate that, even if we take these results at
face value, none of the claimed observed underdensities can alleviate the Hubble
tension. As will be described in the next section, resolving the Hubble tension would
require an underdensity .δ ≈ −0.8 within a radius of 120 . h −1 Mpc [38]. Such an
underdensity is incompatible with any of the observational results and is impossible
to achieve within a .ɅCDM framework. In sum, quantifying the local density contrast
is an interesting research subject per se, but it is unlikely to lead to a viable solution
for the Hubble tension.

21.4 Investigating Density Contrasts in Large-Volume

N-Body Simulations

We investigate the effect of under/overdensities with a controlled experiment, using

an N-body simulation of the large-scale structure of the universe that covers many
realizations of the supernova sample volume.
Let us start with some preliminaries. In the presence of a local void, galaxies tend
to move away from the center of the void and exhibit positive peculiar velocities.
Let’s assume we have. N standard candles at comoving distances.ri and radial peculiar
velocities .vr,i (i.e. radial velocities of galaxies with the global expansion rate of the
universe taken out). The shift in the local Hubble constant caused by these peculiar
velocities is given by
1 ∑ vr,i
N
.ΔH0 = .
loc
(21.2)
N i=1 ri

Clearly, positive net radial peculiar velocities will lead to a positive shift in . H0 , and
vice versa for negative velocities. Equation (21.2) forms a basis for how one can
estimate the statistical fluctuations of the observed . H0 .2
Here we recapitulate the results of Wu and Huterer [38]. We address two aspects of
the sample variance: (1) variance caused by the local density contrast, and (2) variance

2 Note that, in SNIa observations, the peculiar velocities of galaxies are usually removed; for exam-

ple, Ref. [2] has used the density field reconstructed by the 2M++ survey to correct for the peculiar
velocities. Here we discuss the maximum impact of the local density by assuming that the peculiar
velocities are not corrected for. Our results thus serve as a conservative upper limit of the sample
variance.
21 Not Empty Enough: A Local Void Cannot Solve the . H0 Tension 395

caused by the inhomogeneous selection of the supernova sample “Supercal” [39, 40],
which is used in the Riess et al. [2] analysis. The former effect is irreducible, while
the latter effect can be reduced with future larger samples.
To address the contribution of the local density contrast to the Hubble constant,
we perform our study on a large N-body simulation. Our general idea is to place the
observer at many independent locations in the simulation, mimic the SNIa observa-
tions around each observer, and measure the statistical distribution of the inferred
Hubble constant with respect to the global value (i.e. in .ΔH0 from Eq. 21.1). The
advantage of this approach is that we can mimic the SNIa observations precisely by
selecting halos in our simulation that stand for SNIa with their (near-)exact three-
dimensional locations around the observer. This enables measurements of .ΔH0 that
account for the sample variance of actual SNIa observations very precisely. Given
the size of the numerical simulation and our consequent ability to place many inde-
pendent observed volumes in it, we also have great statistical power at our disposal
to quantify the effect of sample variance.
We use the public release of the Dark Sky simulations [41]. Specifically, we use the
−1
.(8 h Gpc)3 volume with .1012 particles and a mass resolution .3.9 × 1010 h −1 Mʘ .
The simulation assumes a standard flat .ɅCDM model consistent with Planck 2016
[42]. We divide this .(8 h −1 Gpc)3 volume into 512 subvolumes of .(1 h −1 Gpc)3 . We
then choose the halo with virial mass . Mvir ∈ [1012.3 , 1012.4 ] Mʘ closest to the center
of each subvolume as our observer. This choice simulates 512 independent observers
living in Milky Way-mass halos, each with a separate subvolume of the large-scale
structure out to the distance of interest (.z max = 0.15). For the host halos of SNIa,
we also use Milky Way-mass halos with . Mvir ∈ [1012.3 , 1012.4 ] Mʘ , and we have
explicitly checked that using all halos above .1013 Mʘ does not change the result.
We then simulate the Supercal supernova observation in each subvolume.
Our procedure then goes as follows (see also Fig. 21.1):

1. Divide the 8. h −1 Gpc Dark Sky box into 512 subvolumes as mentioned above;
each Milky Way-mass halo in every subvolume is a possible host of an SNIa.
Then, for each subvolume.
2. Select a random orientation of the overall SNIa sky coordinates with respect to
the simulation’s Cartesian coordinates.
3. Assign each SNIa to the closest halo in the subvolume based on its 3D coordinates
reported in the Supercal dataset release [39, 40].
4. Calculate the .ΔH0loc from the radial velocities of SNIa host halos using the
relation
1 ∑ vr,i
N
.ΔH0 = ,
loc
(21.3)
N i=1 ri

where .ri and .vr,i are the comoving distance to, and the peculiar velocity of,
the halo that hosts the .i th SNIa. We average over the .ΔH0loc of all supernovae
from this particular orientation for this observer. (The actual analysis includes
additional details; see Ref. [38].)
396 D. Huterer and H.-Y. Wu

Fig. 21.1 Sketch of the simulation procedure in Ref. [38]. We divide the 8 . h −1 Gpc N-body
simulation into 512 subvolumes. In each subvolume, we place an observer at the center, use .∼3000
rotations of the global coordinates of the observed SNIa (not shown in this sketch), and find the
closest Milky Way-mass halo to each SNIa. We then use the known properties of that halo—its
distance from the observer and its peculiar velocity—as proxies for those of SNIa, and use Eq. 21.2
to estimate the Hubble constant shift

5. Go to step 2., repeat the measurements for many different coordinate-system

orientations and obtain the histogram of .ΔH0loc of different orientations from
this particular subvolume.
6. Go to step 1., repeat the measurements for the 512 non-overlapping subvolumes
and obtain the distribution of .ΔH0loc from all subvolumes and all orientations.

The 512 subvolumes account for the variance in the local density in a .ɅCDM
universe, and the .∼3000 SNIa coordinate-system orientations in each subvolume
account for the skewed redshift distribution and the sparse angular distribution of the
SNIa sample. The full procedure therefore corresponds to .∼1.5 million simulations
of inferring the Hubble constant from local SNIa based on the Supercal sample
(.0.023 < z < 0.15).
Figure 21.2 shows the histogram of the distributions of.ΔH0 for all subvolumes and
SNIa coordinate-system orientations. The overall distribution in the Hubble constant
shifts, shown with the dark blue histogram, indicates that the sample variance in the
local . H0loc measurements is much smaller than the amount that could explain the
Hubble tension (i.e. one that would reach out to the Riess et al. value, shown as the
vertical beige bar). For example, even a void that is underdense at the .∼ 2σ level
relative to the mean (the green histogram) would lead to the locally measured Hubble
constant, . H0loc , that is only slightly larger than the mean value, assumed here to be
Planck’s . H0 ≃ 67 km s−1 Mpc−1 .
We have, in fact, been able to measure the standard deviation of the measured
Hubble constant very precisely,

σ
. sample variance = 0.31 km s−1 Mpc−1 . (21.4)
21 Not Empty Enough: A Local Void Cannot Solve the . H0 Tension 397

Fig. 21.2 Sample variance in.ΔH0loc from the simulations of Wu and Huterer [38], compared to the
Planck and distance-ladder ([2], “Riess” in the plot) error bars, assuming Planck’s . H0 [6] to be the
true global value. The blue histogram shows 3240 rotations of the SNIa coordinate system from 512
subvolumes derived from the 8 . h −1 Gpc Dark Sky box, corresponding to .∼1.5 million SN-to-halo
coordinate-system configurations. The green histogram shows a particularly underdense subvolume
with a high .ΔH0loc at the 2-.σ level relative to the mean of all subvolumes. The sample variance in
loc
. H0 is much smaller than the difference between SNIa and CMB measurements. Adopted from
Ref. [38]

In other words, .|H0Planck − H0loc |/σsample variance ≃ 20. Therefore, sample variance
is about 20 times too small to explain the Hubble tension. Interestingly, using
completely different approaches based on perturbation theory, Refs. [23, 28] find a
very similar result: a sample variance of.∼ 0.4 km s−1 Mpc−1 for the Pantheon sample
(also see [43]).
How “empty” should a local void be to resolve the Hubble tension? Fig. 21.3
shows the relation between the change of the local Hubble constant, .ΔH0loc , and
the local density contrast, .δ. For each of the 512 subvolumes, we measure the dark
matter density contrast at .z < 0.04 (.120 h −1 Mpc) and perform mock measurements
of . H0loc using the Supercal supernova sample (.z < 0.15). The small green error bars
correspond to the 512 simulation subvolumes, and the black line is the linear fit.
The dashed curve includes the non-linear correction. This slope is different from that
expected from the perturbation theory because we explicitly consider the sample
selection, which simple perturbation-theory calculations do not account for.
The horizontal line shows that to account for the 6 . km s−1 Mpc−1 shift requires
an underdensity of .δ ≈ −0.8 with a radius of 120 . h −1 Mpc, which is incompatible
with the density fluctuations in the .ɅCDM model. The blue arrows correspond
to observational results from Refs. [31, 34, 35], and none of them can cause a 6
−1
. km s Mpc−1 shift.
Why does the sample variance of local measurements with SNIa have such a
small effect? The reason is in the fact the Supercal SNIa sample goes out to .z = 0.15
(corresponding to distances of about .∼ 500 h −1 Mpc). Averaged over such a large
volume, overdensities and underdensities are not very pronounced, and the variations
in the locally determined Hubble constant are correspondingly small.
398 D. Huterer and H.-Y. Wu

Fig. 21.3 Shift in local Hubble constant versus local density contrast derived from the mock
Supercal sample created from the 8 . h −1 Gpc Dark Sky box (green error bars). The blue arrows
mark the local density inferred from Refs. [31, 34, 35]. The black solid line shows the linear fit,
and the dashed curve shows the non-linear correction. A local Hubble constant shift of .ΔH0loc ≃
6 km s−1 Mpc−1 required to explain the Hubble tension would consequently require an underdensity
of .δ ≈ −0.8 with a radius of 120 . h −1 Mpc, which is incompatible with the cosmic structure in a
.ɅCDM universe. Adopted from Ref. [38]

This result, and the related ones cited above, put a nail in the coffin of the sample-
variance explanations for the Hubble tension. Because the sample-variance (or void)
explanation was arguably the simplest one, it leads to an exciting situation that the
true explanation, whatever it is, is likely to only be more “exotic” and, overall, more
unexpected.

21.5 Summary

We have reviewed what is logically the simplest explanation for the Hubble tension—
the notion that we might be simply seeing the effect of sample variance in the distance-
ladder-based measurements. Sample variance affects the last rung of the distance lad-
der, whose distance–redshift relation can be impacted by local density fluctuations.
In the sample-variance explanation scenario, (1) we live in an underdense region of
the universe, which has an enhanced locally measured Hubble constant, and (2) the
local supernova sample is sparse and highly inhomogeneous so that the error bars
for the local Hubble constant might be larger than previously estimated.
21 Not Empty Enough: A Local Void Cannot Solve the . H0 Tension 399

We have first illustrated the formalism that links the local density to the measured
Hubble constant and have reviewed the methodology adopted by the community
over the past decade. We have then reviewed recent observational results of the local
matter density contrast, which have been inconclusive because it is difficult to infer
the three-dimensional distribution of mass density from galaxies.
Proceeding to discuss our own work on the subject [38], we set out our goal: to
use an N-body simulation to mimic the actual SNIa observations used for . H0 mea-
surements. In the simulation, we place an observer in many independent subvolumes,
each containing the 3D distribution of SNIa that precisely matches that used in actual
observations. This automatically takes into account the effect of over/underdensities,
as well as the inhomogeneous distribution of supernovae. Moreover, each SNIa is
represented by a host halo, whose peculiar velocity is known in the simulation.
Therefore, we have access to a simulated distance-ladder procedure that takes full
account of sample variance, including how it mimics actual SNIa observations. The
procedure is also statistically powerful—we use a total of 1.5 million realizations of
SNIa observations; see Sect. 21.4.
Combining large-volume N-body .ɅCDM simulations with recent SNIa data, we
have found that the effect of the sample variance is small. Specifically, the resulting
standard deviation in the Hubble constant (the square root of sample variance) is

σ
. sample variance = 0.31 km s−1 Mpc−1 . (21.5)

To explain the 6 . km s−1 Mpc−1 Hubble tension would require a .20σ effect in sample
variance. Sample variance is therefore an extremely unlikely explanation for the
Hubble tension.
In addition, we have found that resolving the 6 . km s−1 Mpc−1 Hubble tension
requires an underdensity .δ = −0.8 out to a radius 120 . h −1 Mpc. The underdensity
required to resolve the Hubble tension is so low that it is incompatible with the large-
scale structure in a .ɅCDM universe. Although the observational evidence for local
underdensity is still controversial, even the most extreme observational claims for
the local underdensity could not explain the Hubble tension.
Ruling out the local void is an exciting development in the nascent field of Hubble
tension, as it removes this very simple explanation and a “guaranteed" effect in
cosmology, leaving more exotic or surprising solutions. Future research should focus
on the systematics of supernova and CMB measurements, as well as the exploration
of new physics beyond the standard model of cosmology.

Acknowledgements DH is supported by NASA grant under contract 19-ATP19-0058 and DOE

under contract DE-FG02-95ER40899. HW is supported by the DOE grant under contract DE-
SC0021916.
400 D. Huterer and H.-Y. Wu

References

1. W.L. Freedman, B.F. Madore, V. Scowcroft, C. Burns, A. Monson, S.E. Persson, M. Seibert, J.
Rigby, Carnegie Hubble program: a mid-infrared calibration of the Hubble constant. APJ 758,
24 (2012)
2. A.G. Riess, L.M. Macri, S.L. Hoffmann, D. Scolnic, S. Casertano, A.V. Filippenko, B.E.
Tucker, M.J. Reid, D.O. Jones, J.M. Silverman, R. Chornock, P. Challis, W. Yuan, P.J. Brown,
R.J. Foley. A 2.4% determination of the local value of the Hubble constant. APJ 826, 56 (2016)
3. A.G. Riess et al., A comprehensive measurement of the local value of the Hubble constant with
1 km s.−1 Mpc.−1 uncertainty from the Hubble space telescope and the S. H0 ES team. Astrophys.
J. Lett. 934(1), L7 (2022)
4. N. Aghanim et al., Planck 2018 results. VI. Cosmological parameters. Astron. Astrophys., 641,
A6 (2020) [Erratum: Astron.Astrophys. 652, C4 (2021)]
5. G. Hinshaw, D. Larson, E. Komatsu, D.N. Spergel, C.L. Bennett, J. Dunkley, M.R. Nolta, M.
Halpern, R.S. Hill, N. Odegard, L. Page, K.M. Smith, J.L. Weiland, B. Gold, N. Jarosik, A.
Kogut, M. Limon, S.S. Meyer, G.S. Tucker, E. Wollack, E.L. Wright, Nine-year Wilkinson
microwave anisotropy probe (WMAP) observations: cosmological parameter results. ApJS
208, 19 (2013)
6. X.I.I.I. Planck Collaboration, Planck 2015 results. XIII. Cosmological parameters. A&A 594,
A13 (2016)
7. X.V.I. Planck Collaboration, Planck 2013 results. XVI. Cosmological parameters. A&A 571,
A16 (2014)
8. E. Di Valentino, O. Mena, S. Pan, L. Visinelli, W. Yang, A. Melchiorri, D.F. Mota, A.G. Riess,
J. Silk, In the realm of the Hubble tension—a review of solutions. Class. Quant. Grav. 38(15),
153001 (2021)
9. I. Ben-Dayan, R. Durrer, G. Marozzi, D.J. Schwarz, Value of H.0 in the inhomogeneous universe.
Phys. Rev. Lett. 112(22), 221301 (2014)
10. A. Cooray, R.R. Caldwell, Large-scale bulk motions complicate the Hubble diagram. PhRvD
73(10), 103002 (2006)
11. H.M. Courtois, D. Pomarède, R.B. Tully, Y. Hoffman, D. Courtois, Cosmography of the local
universe. AJ 146, 69 (2013)
12. P. Fleury, C. Clarkson, R. Maartens, How does the cosmic large-scale structure bias the Hubble
diagram? JCAP 3, 062 (2017)
13. L. Hui, P.B. Greene, Correlated fluctuations in luminosity distance and the importance of
peculiar motion in supernova surveys. PhRvD 73(12), 123526 (2006)
14. L.A. Martinez-Vaquero, G. Yepes, Y. Hoffman, S. Gottlöber, M. Sivan. Constrained simulations
of the local universe—II. The nature of the local Hubble flow. MNRAS 397, 2070–2080 (2009)
15. X. Shi, M.S. Turner, Expectations for the difference between local and global measurements
of the Hubble constant. APJ 493, 519–522 (1998)
16. B. Sinclair, T.M. Davis, T. Haugbølle, Residual Hubble-bubble effects on supernova cosmology.
APJ 718, 1445–1455 (2010)
17. E.L. Turner, R. Cen, J.P. Ostriker, The relation of local measures of Hubble’s constant to its
global value. AJ 103, 1427–1437 (1992)
18. L. Wang, P.J. Steinhardt, Cluster abundance constraints for cosmological models with a time-
varying, spatially inhomogeneous energy component with negative pressure. APJ 508, 483–490
(1998)
19. D. Huterer. Growth of cosmic structure 12 (2022)
20. V. Marra, L. Amendola, I. Sawicki, W. Valkenburg, Cosmic variance and the measurement of
the local Hubble parameter. Phys. Rev. Lett. 110(24), 241305 (2013)
21. I. Odderskov, S. Hannestad, T. Haugbølle, On the local variation of the Hubble constant. JCAP
10, 028 (2014)
22. R. Wojtak, A. Knebe, W.A. Watson, I.T. Iliev, S. Heß, D. Rapetti, G. Yepes, S. Gottlöber,
Cosmic variance of the local Hubble flow in large-scale cosmological simulations. MNRAS
438, 1805–1812 (2014)
21 Not Empty Enough: A Local Void Cannot Solve the . H0 Tension 401

23. Z. Zhai, W.J. Percival, Sample variance for supernovae distance measurements and the Hubble
tension. PhRvD 106(10), 103527 (2022)
24. D. Camarena, V. Marra, Z. Sakr, C. Clarkson, A void in the Hubble tension? The end of the
line for the Hubble bubble. Class. Quantum Gravity 39(18), 184001 (2022)
25. S. Castello, M. Högås, E. Mörtsell, A cosmological underdensity does not solve the Hubble
tension. JCAP 2022(7), 003 (2022)
26. V. Marra, T. Castro, D. Camarena, S. Borgani, A. Ragagnin, The BEHOMO project: .Ʌ
Lemaître-Tolman-Bondi N-body simulations. A&A 664, A179 (2022)
27. W. D’Arcy Kenworthy, D. Scolnic, A. Riess, The local perspective on the Hubble tension:
local structure does not impact measurement of the Hubble constant. Astrophys. J. 875(2), 145
(2019)
28. Z. Zhai, W.J. Percival, The effective volume of supernovae samples and sample variance. arXiv
e-prints (2023) [arXiv:2303.05717]
29. J. Soltis, A. Farahi, D. Huterer, C. Michael Liberato, Percent-level test of isotropic expansion
using type Ia supernovae. Phys. Rev. Lett. 122(9), 091301 (2019)
30. H. Böhringer, G. Chon, C. A. Collins, Observational evidence for a local underdensity in the
universe and its effect on the measurement of the Hubble constant. A&A 633, A19 (2020)
31. J. Carrick, S.J. Turnbull, G. Lavaux, M.J. Hudson, Cosmological parameters from the compari-
son of peculiar velocities with predictions from the 2M++ density field. MNRAS 450, 317–332
(2015)
32. W.J. Frith, G.S. Busswell, R. Fong, N. Metcalfe, T. Shanks, The local hole in the galaxy
distribution: evidence from 2MASS. MNRAS 345(3), 1049–1056 (2003)
33. S. Jha, A.G. Riess, R.P. Kirshner, Improved distances to type Ia supernovae with multicolor
light-curve shapes: MLCS2k2. APJ 659, 122–148 (2007)
34. R.C. Keenan, A.J. Barger, L.L. Cowie, Evidence for a .∼300 Megaparsec Scale Under-density
in the local galaxy distribution. APJ 775, 62 (2013)
35. J.R. Whitbourn, T. Shanks, The local hole revealed by galaxy counts and redshifts. MNRAS
437, 2146–2162 (2014)
36. J. H. Wong, T. Shanks, N. Metcalfe, J.R. Whitbourn, The local hole: a galaxy underdensity
covering 90 per cent of sky to .≈200 Mpc. MNRAS 511(4), 5742–5755 (2022)
37. I. Zehavi, A.G. Riess, R.P. Kirshner, A. Dekel, A local Hubble bubble from type IA supernovae?
APJ 503, 483–491 (1998)
38. W. Hao-Yi, D. Huterer, Sample variance in the local measurements of the Hubble constant.
Mon. Not. Roy. Astron. Soc. 471(4), 4946–4955 (2017)
39. D. Scolnic, S. Casertano, A. Riess, A. Rest, E. Schlafly, R.J. Foley, D. Finkbeiner, C. Tang,
W.S. Burgett, K.C. Chambers, P.W. Draper, H. Flewelling, K.W. Hodapp, M.E. Huber, N.
Kaiser, R.P. Kudritzki, E.A. Magnier, N. Metcalfe, C.W. Stubbs, Supercal: cross-calibration of
multiple photometric systems to improve cosmological measurements with type Ia supernovae.
APJ 815, 117 (2015)
40. D. Scolnic, A. Rest, A. Riess, M.E. Huber, R.J. Foley, D. Brout, R. Chornock, G. Narayan,
J.L. Tonry, E. Berger, A.M. Soderberg, C.W. Stubbs, R.P. Kirshner, S. Rodney, S.J. Smartt, E.
Schlafly, M.T. Botticella, P. Challis, I. Czekala, M. Drout, M.J. Hudson, R. Kotak, C. Leibler,
R. Lunnan, G.H. Marion, M. McCrum, D. Milisavljevic, A. Pastorello, N.E. Sanders, K. Smith,
E. Stafford, D. Thilker, S. Valenti, W.M. Wood-Vasey, Z. Zheng, W.S. Burgett, K.C. Chambers,
L. Denneau, P.W. Draper, H. Flewelling, K.W. Hodapp, N. Kaiser, R.-P. Kudritzki, E.A. Mag-
nier, N. Metcalfe, P.A. Price, W. Sweeney, R. Wainscoat, C. Waters, Systematic uncertainties
associated with the cosmological analysis of the first Pan-STARRS1 type Ia supernova sample.
APJ 795, 45 (2014)
41. S.W. Skillman, M.S. Warren, M.J. Turk, R.H. Wechsler, D.E. Holz, P.M. Sutter, Dark sky
simulations: early data release. ArXiv e-prints (2014)
42. Planck Collaboration Int. XLVI. Planck intermediate results. XLVI. Reduction of large-scale
systematic effects in HFI polarization maps and estimation of the reionization optical depth.
A&A 596, A107 (2016)
43. D. Camarena, V. Marra, Impact of the cosmic variance on H.0 on cosmological analyses. PhRvD
98(2), 023537 (2018)
Chapter 22
Resolving the Hubble Tension with Early
Dark Energy

Vivian Poulin and Tristan L. Smith

22.1 Introduction: Why Early Dark Energy?

The notion of “Early Dark Energy” was originally introduced as an attempt at tack-
ling the cosmic coincidence problem [1, 2], i.e., the question of why dark matter
and dark energy have such similar densities, in particular since the vacuum energy
prediction from quantum field theory for the value of the cosmological constant is
so far off compared to the observation. If dark energy is a scalar-field, it may have
an equation of state different from a cosmological constant, and it may even leave a
small contribution already at the last-scattering surface [3–5].
Over the last few years, with the rise of the Hubble tension, the interest for
models of EDE has dramatically increased, as it can be used to boost the value
of the expansion rate in the decade of redshift prior to recombination to reduce the
size of the sound horizon .rs (z ∗ ). Indeed, a naive increase in . H0 would decrease the
angular diameter distance and in turn increase the angular size of the sound horizon
.θs . This must be compensated for by a decrease in .r s (z ∗ ), in order to keep .θs fixed to
the very precise value measured by Planck.
The Hubble tension is, in fact, more than just a tension in the value of . H0 : it
also requires an increase in the physical matter density. Under the assumption of a
flat .ɅCDM Universe, the expansion history measurements derived from Supernovae

V. Poulin (B)
Laboratoire Univers and Particules de Montpellier (LUPM), CNRS and Université de Montpellier
(UMR-5299), Place Eugène Bataillon 34095 Montpellier Cedex 05, France
e-mail: vivian.poulin@umontpellier.fr
T. L. Smith
Department of Physics and Astronomy, Swarthmore College, 500 College Ave., Swarthmore, PA
19081, USA
e-mail: tsmith2@swarthmore.edu
Center for Cosmology and Particle Physics, Department of Physics, New York University, New
York, NY 10003, USA
© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 403
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_22
404 V. Poulin and T. L. Smith

Type Ia (SNIa) and Baryon Acoustic Oscillations (BAO) impose a tight constraint on
Ωm (the ratio of matter density .ρm to the critical density1 .ρcrit ). When combined with
.
the S. H0 ES determination of . H0 , they lead to a physical matter density .ωm (.≡ Ωm h 2 )
of.0.1603 ± 0.0063, a value that is approximately.3σ larger than the one inferred from
Planck observations assuming the .ɅCDM model (.ωm = 0.14228+0.00079 −0.00091 ). Adjusting
cosmological data therefore requires an increase in.ωm that would spoil the amplitude
of the peaks measured by Planck.
Remarkably, with a single mechanism, EDE manages to reduce the size of the
sound horizon and compensate for the effect of an increase of .ωcdm on the CMB.
Indeed, the rise in . H (z) induced by EDE contributes to an augmentation of Hubble
friction, resulting not only in a smaller sound horizon but also in a slowdown of the
growth of the Cold Dark Matter (CDM) density contrast. EDE therefore manages to
maintain a good fit to the combination of CMB, BAO and SN1a data, while adjusting
the value of . H0 (or rather the SN1a magnitude) measured by S. H0 ES.
In the following, we review the results of analyses of EDE, focusing on the
toy-model of axion-like EDE as a representative, well-studied, example of EDE.
Sect. 22.2 introduces the model and present results of analyses of EDE with a suite
of cosmological data (in particular Planck ACT and SPT data). Sect. 22.3 discusses
the impact of EDE beyond the Hubble tension, from observational impacts on large-
scale structure (LSS) data in particular, to theoretical “fine-tuning” issues and model-
building efforts. We present our conclusions in Sect. 22.4.

22.2 Resolving the Hubble Tension with Early Dark Energy

22.2.1 Modelling Early Dark Energy

Initially suggested as a potential mechanism to explain the Hubble tension in Refs. [6,
7], EDE has since then undergone various refinements in the modelling and additional
studies by numerous authors [8–38].
EDE models typically propose the existence of a scalar field, and may differ in
the choice of scalar field potential, the number of scalar field involve in the model,
and whether or not the scalar field is minimally coupled. In this chapter, we focus on
an axion-like potential of the form2 [7, 8]

. V (θ ) = m 2 f 2 [1 − cos(θ )]n , (22.1)

where .m represents the axion mass, . f the axion decay constant, and .θ ≡ φ/ f is a re-
normalized field variable defined such that .−π ≤ θ ≤ π. The next chapter focuses
on an alternative realization of an Early Dark Energy model named ‘New Early

1Here .ρcrit ≡ 3H02 /(8π G), and .h is given by .h ≡ H0 /(100 km/s/Mpc).

2This potential was implemented in a modified version of the CLASS code [39] available at https://
github.com/PoulinV/AxiCLASS.
22 Resolving the Hubble Tension with Early Dark Energy 405

Dark Energy’ [14, 15]. The scalar field background dynamics3 can be understood
by looking at the homogeneous Klein-Gordon equation of motion:
d V (φ)
.φ̈ + 3 H φ̇ + = 0. (22.2)
dφ

Due to the initially large Hubble friction, the background field is held fixed at
some value, resulting in a constant background energy density. When . H ∼ ∂θ2 V (θ )
the field becomes dynamical, rolling down its potential to eventually oscillate about
the minimum. Although this potential is not tied to a fundamental theory, it provides
a flexible EDE model with a rate of dilution set by the value of the exponent .n, and
this is what we are most interested in here. Taken at face-value, this potential may
be generated by higher-order instanton corrections [41–43] that require fine-tuning
to cancel the leading contributions [30, 35], or through the interaction of an axion
and a dilaton [11] and is therefore non-generic [44]. Related constructions of this
potential in the context of ‘Natural Inflation’ were also proposed in Refs. [45–47].
Interestingly, recent works have paved the way to embed this potential in a string-
theory framework [30, 48], showing that the theory challenges for building an EDE
with such a potential are not particularly different from that of other cosmological
models, such as inflation, dark energy, or fuzzy dark matter [48].
In full generality, this model therefore has 4 free parameters. Yet, most studies
trade the ‘theory parameters’, .m and . f , for ‘phenomenological parameters’ that are
better constrained by the data. These are the critical redshift .z c at which the field
becomes dynamical and the fractional energy density . f EDE (z c ) contributed by the
field at the critical redshift. The third parameter, .n, controls the equation of state
.w ≡ (n − 1)/(n + 1) when the field dilutes. Finally, the last parameter is the initial
field value .θi , which has two main impacts: it changes the asymmetry of the energy
injection and controls the effective sound speed .cs2 and thus the dynamics of the
perturbations (mostly). Note that it is assumed that the field always starts in slow-
roll (as enforced by the very high value of the Hubble rate at early times), and
without loss of generality one can restrict .0 ≤ θi ≤ π. We illustrate the effect of
each parameter on the energy injection . f EDE (z) in Fig. 22.1.

22.2.2 EDE in Light of Planck Data

We now review results of analyses of the axion-like EDE model in light of Planck
CMB power spectra. The baseline analysis includes the full Planck 2018 TT,TE,EE
and lensing power spectra [49], BAO measurements from BOSS DR12 [50], SDSS
DR7 [51] and 6dFGS [52], and a compilation of uncalibrated luminosity distances to
SN1a from Pantheon [53]. Reconstructed parameters are provided in Table 22.1. We
refer to this dataset as “BaseTTTEEE”. All the analyses we present follow the Planck

3 For a discussion about EDE perturbed dynamics, we refer to the review [40].
406 V. Poulin and T. L. Smith

Fig. 22.1 Effect of varying the different EDE parameters on the energy injection . f EDE (z). The
fiducial model has {. f EDE (z c ) = 0.1, log10 (z c ) = 3.5, θi = 2.8, n = 3}

Table 22.1 Mean (best-fit) .±1σ (or .2σ for one-sided bounds) of reconstructed parameters in
the EDE model confronted to PlanckTTTEEE+Lensing+BAO+Pantheon. We report constraints
that include and leave out the S. H0 ES prior, and compute the residual tension using the . Q DMAP
statistics.
Planck+BAO+Pantheon +S. H0 ES
+0.030
. f EDE (z c ) .< 0.091(0.088) .0.109(0.122)−0.024
+0.029
.log10 (z c ) unconstrained .(3.55) .3.599(3.568)−0.081
+0.22
.θi unconstrained (2.8) .2.65(2.73)−0.025
+0.006
.h .0.688(0.706)−0.011 .0.715(0.719) ± 0.009
+0.011
. S8 .0.831(0.839)−0.013 .0.839(0.843) ± 0.012
.Ωm .0.3084(0.3041) ± 0.0058 .0.3008(0.3005) ± 0.0048
+0.0018
.ωcdm .0.1227(0.1281)−0.0036 .0.1303(0.1319) ± 0.0035
2ω +0.018
.10 b .2.258(2.266)−0.020 .2.283(2.303) ± 0.020

.2.122(2.135) ± 0.032 .2.153(2.145) ± 0.032

9
.10 As
+0.0053
.n s .0.9734(0.9823)−0.0076 .0.9883(0.9895) ± 0.0060
+0.0069
.τreio .0.0570(0.0574)−0.0076 .0.0582(0.0579) ± 0.0075

Total .χmin
2 3799.2 3802.9
.Δχmin .−3.8 .−23.7
2

. Q DMAP 1.9.σ
22 Resolving the Hubble Tension with Early Dark Energy 407

Fig. 22.2 2D posteriors for .h vs .{ f EDE (z c ), log10 (z c ), θi , n s , ωcdm , S8 }

conventions for the treatment of neutrinos and inclue two massless and one massive
species with with .m ν = 0.06 eV [49]. For the .ɅCDM parameters, large flat priors
on the dimensionless baryon energy density .ωb , the dimensionless cold dark matter
energy density .ωcdm , the Hubble parameter today . H0 , the logarithm of the variance
.ln(10 As ) of curvature perturbations centered around the pivot scale .k p = 0.05
10
−1
Mpc. (according to the Planck convention), the scalar spectral index .n s , and the re-
ionization optical depth .τreio were imposed. For the EDE parameters, a logarithmic
prior on .z c , and flat priors for . f EDE (z c ) and .θi as .log10 (z c ) ∈ [3, 4], . f EDE (z c ) ∈
[0, 0.3], .θi ∈ [0, π ] were considered. For most of our discussion, the exponent .n (or
equivalently the equation of state once the field rolls .w f = (n − 1)/(n + 1)) is fixed
to .n = 3 to reduce the dimensionality, without affecting conclusions.
We start by presenting results of analyses that include a direct measurements of. H0
by S. H0 ES as modeled by a Gaussian prior on the Hubble parameter .h = 0.7304 ±
0.001 where .h ≡ H/(100 km/s/Mpc) [54]. We show the posterior distribution of .h
vs.{ f EDE (z c ), log10 (z c ), θi , n s , ωcdm , S8 } in the EDE and.ɅCDM model (Fig. 22.2). In
this analysis, we find. f EDE (z c ) = 0.109+0.030 −0.024 , around the critical redshift .log10 (z c ) =
+0.029
3.599(3.568)−0.081 with.h = 0.715 ± 0.009, and a.Δχ 2 ≡ χEDE 2
− χɅCDM
2
= −23.7,
which is a strong preference in favor of the EDE model (a simple estimate gives.∼ 4σ
preference, assuming .Δχ 2 is .χ 2 −distributed with 3 degrees of freedom). We stress
that it is not correct to estimate the tension with S. H0 ES by using the reconstructed
value of.h from this analysis, given that the S. H0 ES prior was included in the analysis.
Rather, we report the following tension metric [55, 56]
√
. Q DMAP ≡ χ 2 (w/ SH0 ES) − χ 2 (w/ SH0 ES) , (22.3)

where .χ 2 (w/o SH0 ES) is obtained from a fit to the same data set but removing
the S. H0 ES prior. This metric can easily be interpreted intuitively, as for a single
additional data point (S. H0 ES here), the .χ 2 value of a “good” model should not
degrade by more than .∼ 1 (for a rigorous discussion, see Ref. [55]). Applying this
to EDE, one finds the tension metric . Q DMAP = 1.9σ , while the metric gives 4.8.σ
in .ɅCDM. Therefore, the EDE model is able to accommodate large values of . H0 ,
while providing a good fit to Planck+BAO+SN1a. From these results, one concludes
that a contribution of .10% of EDE at a redshift close to matter-radiation equality
(recalling .z eq ∼ 3500) can resolve the Hubble tension.
408 V. Poulin and T. L. Smith

Fig. 22.3 Constraints on EDE “microphysics”:.n controls the EDE equation of state when it dilutes
(displayed as.w = (n − 1)/(n + 1)) and.θi controls the asymmetry of the injection and the dynamics
of perturbations

22.2.3 Constraints on EDE “Microphysics”

We show in Fig. 22.3, that on top of constraining the peak and location of the energy
injection, Planck data also provide constraints on the “microphysics” of the energy
injection. When .n is let free to vary,4 one finds .n = 3.37+0.41
−0.99 (68% C.L.) [8, 40],
with a fairly sharp constraint .n ≾ 5.2 (95% C.L.). This translates into a constraint
on .w = (n − 1)/(n + 1) = 0.535 ± 0.086 (68% C.L.) and a limit .w < 0.7 (95%
C.L.), i.e., a very fast dilution with a phase of kinetic energy domination (.w = 1) is
disfavored by the data in the axion model. In addition, they impose tight constraints
on the initial field value .θi (in units of . f , the axion decay constant). As shown in
Sect. 22.2.1, .θi at the background level controls the shape of the energy injection,
which in turn affects the sound horizon and damping scales. At the perturbation
level, its dominant impact is setting the effective sound speed of propagation of EDE
perturbations.
For an oscillating EDE with a potential of the form.V ∝ φ 2n around the minimum,
an approximate expression5 for the sound speed can be derived by “cycle-averaging”
the sound-speed [57]:

⟨δ P⟩ 2a 2 (n − 1)ϖ 2 + k 2
c2 (k, a) =
. s ≃ 2 , (22.4)
⟨δρ⟩ 2a (n + 1)ϖ 2 + k 2

where .ϖ is the angular frequency of the oscillating background field [8, 57, 58]
√
π ┌( 1+n ) θ n−1 (a)
.ϖ (a) ≃ 3 H (z c ) ( 2n
) 2−(1+n)/2 √ env ,
┌ 1 + 2n1
|E n,θθ (θi )|

4 For numerical reason, a lower bound.n > 2 is imposed in this analysis. Part of the parameter space
below .n = 2 may be allowed, although as one gets closer to the matter case .n = 1, it is expected
that constraints will become strong.
5 Note that this approximate sound speed only applies when.ϖ /H >> 1– i.e. when the field oscillates

several times per Hubble time.

22 Resolving the Hubble Tension with Early Dark Energy 409

Fig. 22.4 2D posteriors for . f EDE (z c ) vs .{h, log10 (z c ), θi , n s , ωcdm , S8 } when analyzing
PlanckTTTEEE+BAO+Pantheon with and without the S. H0 ES prior (. H0 = 73.04 ± 1.04 [54]),
as analyzed in Ref. [60]

where .┌(x) is the Euler Gamma function. The envelope of the background field
(.θenv ≡ φenv / f ) once it is oscillating is roughly .φenv (a) = φc (ac /a)3/(n+1) and we
have written the scalar field potential as .Vn (φ) = m 2 f 2 E n (θ = φ/ f ). The angu-
lar frequency .ϖ defines a scale that governs the sound speed that affects a given
mode: For .a >> ac , .ϖ = 0, so that .cs2 = 1; when .a << ac and .k << aϖ , .cs,eff 2
=
(n − 1)/(n + 1) = w; when .a << ac and .k >> aϖ , .cs = 1. 2

Data prefer an EDE for which modes within the horizon around .z c have effective
sound speed .≾ 0.9 [8, 10]. This can be achieved √ if the potential is flat around the
initial field value .θi such that the term .θin−1 / E n,θθ (θi ) >> 1 in Eq. 22.4. Since the
range of .k−modes within the horizon is a sharp function of the initial field value .θi ,
data provide a fairly strong constraint on .θi ∼ π (see Fig. 22.4).
These constraints on the dynamics, and as a result on the shape of the potential,
explain why simpler power-law potentials (as in the Rock’n’Roll model [9]) fair less
favorably in resolving the tension.
Finally, translating the phenomenological parameters back into the theory param-
eters, the EDE model suggests the existence of an axion-like field with mass
−27
.m ≃ 10 H0 = 2.13 h × 10 eV, decay constant . f ≃ 0.15Mpl , and initial field
6

value .φi / f ∼ π . This prompted studies to understand whether the near-Planckian

values of the decay constant and field excursion can lead to additional phenomenol-
ogy (as in the EDS model [27, 59]) and constraints, as it might violate the axion
weak gravity conjecture [35, 44].

22.2.4 “Barefoot Analyses”: Bayesian or Frequentist?

Concerns often arise regarding the inclusion of S. H0 ES data for EDE constraints and
for comparison to .ɅCDM: ideally, a model that resolves the Hubble tension when
analyzed with CMB data alone should predict a higher . H0 , improve the .χ 2 more
than the additional degrees of freedom and be preferred over .ɅCDM. It is therefore
interesting to look at “barefoot” analyses of EDE, i.e., analyses without the S. H0 ES
prior.
We show in Fig. 22.4 the 2D reconstructed posteriors for the parameters . f EDE (z c )
vs .{h, log10 (z c ), θi , n s , ωcdm , S8 } when analyzing PlanckTTTEEE+BAO+Pantheon
410 V. Poulin and T. L. Smith

with and without the S. H0 ES prior [60]. In the absence of a prior on . H0 , it is striking
that the data combination PlanckTTTEEE+BAO+Pan18 provides an apparent upper
limit on the EDE contribution. f EDE (z c ) < 0.091 (at 95% C.L.), with.h = 0.688+0.006
−0.011 .
In fact, taken at face value, this constraint seems to exclude the value of . f EDE (z c ) ∼
0.12 favored in the analysis that includes S. H0 ES.
This apparent inconsistency is the sign of ‘prior-volume effects’ [56, 61–67]:
when the parameter . f EDE ∼ 0, the other EDE parameters (.z c , .θi ) lose all influence
on the dynamics and have therefore no impact on the model predictions, which
become identical to .ɅCDM regardless of the value of these two parameters. This
leads to a large prior volume at . f EDE ≃ 0, which impacts the results of the MCMCs
and skews the posterior distributions towards . f EDE → 0 [63–65]. These posteriors
however do not trace the true data likelihood in the extended.ɅCDM+EDE parameter
space. To test for the impact of prior-volume effects, one can compute the “profile
likelihood” of a given parameter by successively optimizing .χ 2 while keeping the
parameter of interest fixed at different values within a given range. Following the
Frequentist approach, confidence intervals are then extracted by computing the .Δχ 2
with respect to the bestfit point, with .Δχ 2 ≃ 1 giving the .1σ region in the case of a
Gaussian profile far from the parameter prior boundaries [68].6
In Fig. 22.5, we compare the MCMC 1D posteriors for . H0 and . f EDE in .ɅCDM
and EDE to the profile likelihoods of the same parameters obtained when analyz-
ing Planck 2018 CMB TTTEEE data. First, it is clear that there is little difference
between the posteriors and the likelihood profiles of.ɅCDM from the two approaches
[70]. This shows that for .ɅCDM, conclusions derived from the Bayesian analyses
are perfectly robust. For the EDE model however, the profile likelihood shows much
wider constraints to both . H0 and . f EDE than the Bayesian posteriors. Strikingly, while
one would conclude from the MCMC analysis that . f EDE is consistent with 0, the pro-
file likelihood analysis shows preference for . f EDE > 0 at over .2σ , without including
the S. H0 ES . H0 prior.

22.2.5 Imprints of EDE in the CMB: Target for Future

Surveys

The predictions for the residuals of the CMB TT and EE power spectra with respect
to .ɅCDM in the best-fit EDE cosmology are shown with the solid black line. To
emphasize the irreducible effects imprinted by EDE in data (i.e. those that cannot
be absorbed by a re-shuffling of .ɅCDM parameters) that future surveys can target,
we show in Fig. 22.6 the residuals of 100 samples randomly drawn from the MCMC
chains, normalized to the .ɅCDM best-fit to Planck. The color code indicates the

6 In a more generic (non-Gaussian) case, one can use the Feldman-Cousins prescription to define

confidence intervals [69]. For EDE, these were found to match the simple Gaussian prescription
[65].
22 Resolving the Hubble Tension with Early Dark Energy 411

Fig. 22.5 Posteriors obtained from a profile likelihood (dashed) and MCMC methods (solid) are
shown for .ɅCDM (blue) and EDE (red) for . H0 and . f EDE , for Planck 2018 CMB TTTEEE data.
For the profile likelihoods, we plot the normalised curves of .e−χ /2 . While the posterior and profile
2

likelihood match under .ɅCDM, the profile likelihood indicates both a higher . H0 and . f EDE under
the EDE model

Fig. 22.6 Residuals of 100 samples randomly drawn from the MCMC chains in the EDE model (fit
to Planck+BAO+Pantheon+S. H0 ES), normalized to the .ɅCDM best-fit to Planck. The color code
indicates the value of . H0 in each sample

value of . H0 in each sample. Note that we use a combination of logarithmic and

linear spacing in order to clearly show both large and small scale features.
First, the spectra show a bump-like feature around .l ∼ 1500 with increased power
toward large .l in TT and EE that is present in essentially all the samples, and is due to
the larger .n s . This feature can be further probed with high-resolution ground-based
measurements, such as ACT and SPT, and in the future, the Simons Observatory
(SO) [71] and CMB-S4 [72]. Multipoles .l ∼ 100 − 500 are particularly sensitive
412 V. Poulin and T. L. Smith

to details of the EDE dynamics, as these are the multipoles entering the horizon
concurrently with the largest fractional contribution of the EDE energy density. Yet,
the data require no strong departure from the .ɅCDM prediction in that multipole
range, and the samples oscillate around zero. Finally, one can also see large residual
patterns at large angular scales (.l < 30) in all three power spectra. The increased
depth of the Sachs-Wolfe plateau in TT is (mostly) a consequence of slightly different
values of . As and .n s , and is common to all samples, albeit at different amplitudes. On
the other hand, the large bump/dip in TE and EE, which is due to small differences
in .τ and .ωb , should not be overinterpreted, as the amplitudes of these vary greatly
within the random samples of 100 points. It is therefore possible to easily adjust any
forthcoming large-scale EE measurements with a slight shift in parameters and this
signal is not a constraining feature for EDE. Note also that the correlation between the
value of. H0 and the size of the residuals is stronger in polarization than in temperature,
with the EE and TE spectra showing sharper oscillations when . H0 increases. On the
other hand, in TT it is possible to find samples of points with large . H0 values that
show only small deviations from .ɅCDM (and vice versa).

22.2.6 Detecting EDE in the CMB

At the moment, one can conservatively conclude that Planck data alone do not
strongly favor, nor disfavor, EDE over .ɅCDM. This raises a couple of questions:
should we have expected these data to be able to detect . f ∼ 0.12, .h ∼ 0.72, as the
best-fit model extracted from analyses that include S. H0 ES suggest? If not, can future
CMB experiment unambiguously detect such a signal?
These questions can be addressed by running mock analyses that include an EDE
signal, as performed in Ref. [8], and illustrated in Fig. 22.7. On the left panel, it shows
that Planck data alone cannot detect even a .11.5% contribution of EDE at .z c ∼ 3500
with .h = 0.72 and only lead to upper limits on the EDE contribution. However, an
experiment like CMB-S4 would detect this signal at more than .10σ . In addition, the
dashed lines on the right panel shows that the biases introduced by analysing the
mock Planck data assuming .ɅCDM leads to parameters that are in good agreement
with what is found in real data, with a degradation in .χ 2 of the same order as to what
is found in real data as well. Looking forward, when analysing data like CMB-S4, the
.ɅCDM model should provide a bad fit to the data, with a fit improved by.∼ 500 under

EDE, such that any statistical measure should unambiguously favor EDE. .ɅCDM
parameters should become strongly biased (solid lines on the right panel), with big
differences in parameter values compared to what is currently found with Planck. Let
us stress that the current disagreement between Bayesian and Frequentist approach
is purely due to the lack of precision of current CMB data. In the case of CMB-S4,
the data likelihood overwhelmingly dominates over the prior, and both approaches
agree.
22 Resolving the Hubble Tension with Early Dark Energy 413

Fig. 22.7 Reconstruction from mock data with fiducial model .{h = 0.72, f EDE (z c ) =
0.115, log10 (z c ) = 3.53}. Left panel: 2D Posterior distributions in the EDE model from mock Planck
data and CMB-S4. Right panel: 1D Posterior distributions of . H0 and .ωcdm from mock Planck data
(dashed lines) and CMB-S4 (full lines) in either the.ɅCDM (blue) or EDE (red) model. Taken from
Ref. [8]

Fig. 22.8 Left panel: Posterior distributions of .{H0 , f EDE (z c ), θi , log10 (z c )} reconstructed from
ACT data (with and without PlanckTT data upto .l ≤ 650). Right panel: Residuals between EDE
and .ɅCDM when fit to ACT DR4 data alone and in combination with PlanckTT650. We also show
the residuals of ACT DR4, SPT3G and PlanckTT650 in the .ɅCDM model

22.2.7 Early Dark Energy in Light of ACT and SPT Data

The next-generation surveys, the Atacama Cosmology Telescope (ACT) [73] and
the South Pole Telescope (SPT) [74, 75] can already provide additional constraining
power on EDE. Interestingly, recent analyses have shown that the ACT DR4 data
favor EDE at 2–3 .σ over .ɅCDM, with . H0 values that are in good agreement with
the S. H0 ES determination. We review the results from ground-based telescope data
analyses, the hints of EDE, and contrasts it with the results from the Planck satellite
just presented.
We show in Fig. 22.8 left panel the posterior distributions of
.{H0 , f EDE (z c ), θi , log10 (z c )} in the EDE cosmology, with ACT alone (in blue)
and in combination with PlanckTT restricted data, that is essentially equivalent to
the data from WMAP [76]. ACT data favor a non-zero contribution of EDE at .≿ 2σ ,
+0.055 +0.036
with . f EDE (z c ) = 0.152−0.092 and .h = 0.742−0.049 , improving the .Δχ 2 = −9.3 with
414 V. Poulin and T. L. Smith

Fig. 22.9 Left panel: Posterior distributions of .{H0 , f EDE (z c ), θi , log10 (z c )} reconstructed from
the analysis of ACT and SPT data with PlanckTT data up to .l ≤ 650. Right panel: Same as left
panel but now comparing the use of PlanckTT650TEEE or full PlanckTTTEEE. From Ref. [79]
Left panel: Posterior distributions of .{H0 , f EDE (z c ), θi , log10 (z c )} reconstructed from the analysis
of ACT and SPT data with PlanckTT data up to .l ≤ 650. Right panel: Same as left panel but now
comparing the use of PlanckTT650TEEE or full PlanckTTTEEE. From Ref. [79]

respect to .ɅCDM (see also Ref. [77] for the ACT collaboration analysis yielding
similar results with a different analysis pipeline).
Ref. [77] established that the preference for EDE in ACT data is due to a residual
pattern at a multipole value of around 500 in the EE spectrum. This can be observed
in the right panel of Fig. 22.8, which shows the residual between EDE and the best-fit
.ɅCDM model to the ACT DR4 data. One can see that the pattern is accompanied
by a similar bump at .l ∼ 350 in TT, that is not observed in Planck data. As a result,
when the large-scale PlanckTT650 data is included, the residual features around
the multipole value of 500 in both the EE and TT spectra become less prominent.
Instead, a residual pattern of oscillations is observed in all spectra, which still provides
a significantly better fit to the ACT data, equally supported by the TT, TE, and EE
spectra, and in fact the preference for EDE increases to .∼ 3σ , with .Δχ 2 = −15.4.
The inclusion of SPT TEEE data is shown in Fig. 22.9 left panel. SPT data prefer
a small deficit of power in EE around .∼ 350 where ACT prefers a small increase,
and thereby slighlty reduce the preference for EDE. Analyses with full SPT TT data
have yet to appear in the literature.
Finally, we consider separately the role of PlanckTT and PlanckTEEE on
constraining the hints for EDE seen in ACT. Remarkably, the right panel of
Fig. 22.9 shows that the analyses of ACT+SPT with PlanckTT650TEEE data
detects . f EDE (z c ) = 0.163(0.179)+0.047 +0.028
−0.04 at .log10 (z c ) = 3.526(3.528)−0.024 and .θi =
+0.098
2.784(2.806)−0.093 , with a .Δχ = −16.2 (3.3.σ preference) in favor of EDE over
2
+1.9
.ɅCDM [78]. In addition, the model predicts the value of . H0 = 74.2−2.1 km/s/Mpc,

in good agreement with direct determinations.

22 Resolving the Hubble Tension with Early Dark Energy 415

Fig. 22.10 The linear (left) and non linear (right) matter power spectra in an EDE cosmology and
in .ɅCDM. Taken from Ref. [82]

However, these results are in some tension with the analysis of full Planck data
presented earlier, which showed only upper limits (in the MCMC analyses) in the
absence of a prior on . H0 . One can see in the right panel of Fig. 22.9 that the recon-
structions from PlanckTTTEEE (in blue) and PlanckTT650TEEE+ACT+SPT (in
red) show a mild inconsistency, though compatible with a statistical fluke at .∼ 2σ 7 .
Consequently, the combination of all datasets PlanckTTTEEE+ACT+SPT (in black)
leads to the constraint . f EDE < 0.128, significantly weaker than that obtained from
analyzing PlanckTTTEEE alone (. f EDE < 0.091). Future data from ACT, and later
on from SO and CMB-S4, are eagerly awaited to shed light on the situation.

22.3 Consequences of Early Dark Energy

22.3.1 Constraining EDE with Galaxy Clustering Data

We show in Fig. 22.10 the linear (left) and non linear (right) matter power spectra in
the EDE cosmology and in.ɅCDM as computed in Ref. [82] for a fiducial analysis that
includes Planck+BAO+Pantheon+S. H0 ES data. One can see that the EDE cosmology
results in a clear increase of power at scales .k ≿ 0.1 h/Mpc, although the differences
are smaller in the non-linear power spectrum. This is mostly due to the reshuffling
of the .ɅCDM parameters necessary to preserve the fit to CMB data, in particular
the increase in .ωcdm and .n s [83]. As stressed in introduction, the increase in .ωcdm ,
that is the main driver of the larger power at small scales, is necessary in order to
keep .Ωm ≡ ωcdm / h 2 fixed and adjust the late-time measurements of the shape of
the expansion rate from BAO and SN1a. The effect of EDE partly compensates that
increase, though not perfectly, resulting in a larger value of .σ8 . As we discuss below,
the change in the matter power spectrum and in .σ8 can be probed with LSS data.

7See Refs. [80, 81] for further examination of the statistical agreement between these three CMB
data sets.
416 V. Poulin and T. L. Smith

The use of BOSS data analyzed using the one-loop prediction of the galaxy
power spectrum in redshift space from the Effective Field Theory of Large-Scale
Structures8 [84–89] has been suggested to constrain EDE [90, 91]. When analyzing
Planck+BAO+Pantheon+EFTBOSS, Refs. [92, 93] originally derived . f EDE (z c ) <
0.08 and . f EDE (z c ) < 0.072 respectively, with two different implementations of the
EFT. However, the apparent constraining power from the BOSS full-shape analysis
is partly amplified by the impact of the prior volume artificially favoring .ɅCDM in
the Bayesian context. This was explicitly verified with a profile-likelihood approach
[65, 94] which indicated that EDE is favored by the combination of Planck+BOSS
at the .∼ 2σ level, with . f EDE (z c ) = 0.072 ± 0.036 (see discussion in Sect. 22.2.4).
The most recent analysis, that corrects an issue in the normalization of BOSS data
window function [95], showed that the inclusion of EFTBOSS leads to . f EDE (z c ) <
0.083, which is a marginal.∼ 5% improvement over the constraints with conventional
BAO and redshift space distortion data [60]. The Hubble tension is reduced to the
.2.1σ level in the EDE cosmology (.1.9σ without EFTBOSS) compared to .4.8σ in
the .ɅCDM model, and one finds . f EDE (z c ) = 0.103+0.027 +255
−0.023 at .z c = 3970−205 when the
S. H0 ES prior is included.
Nevertheless, Fig. 22.11 right panel shows that EFTBOSS data however do fur-
ther restrict the preference for EDE seen in ACT data, leading to a mild constraint
on . f EDE (z c ) < 0.172 with .Δχ 2 (EDE − ɅCDM) = −11.1, to be compared with
+0.064
. f EDE (z c ) = 0.128−0.039 without EFTBOSS, with .Δχ (EDE − ɅCDM) = −14.6.
2

When full Planck data are included, the constraints become . f EDE (z c ) < 0.110, and
one can estimate that the Hubble tension is reduced to .1.5σ in the EDE model,
to be compared with .4.7σ in .ɅCDM. The inclusion of the S. H0 ES prior leads to
+0.028 +220
. f EDE (z c ) = 0.108−0.021 at . z c = 3565−495 . Therefore, EFTBOSS data do not rule out
EDE as a resolution to the Hubble tension, although they do constrain very high EDE
fractions, as seen from analyzing ACT DR4 data. Future data from DESI, Euclid and
the Vera Rubin Observatory will be pivotal in testing for the presence of EDE [93].

22.3.2 EDE and the . S8 Tension

Inferences of . S8 from weak lensing surveys such as the CFHTLenS [96], DES [97],
KiDS [98], and HSC [99] as well as from Planck SZ cluster abundances [49, 100]
are about 2–3 .σ smaller than that from the CMB. We have already argued that a
generic feature of early-universe models proposed to resolve the Hubble tension is
to predict a greater . S8 and increase this tension, simply because one cannot keep .Ωm
fixed if .h increases without increasing .ωm , and therefore yielding a larger onset of
matter domination and a larger .σ8 [101, 102].
Due to the impact on . S8 , Ref. [83] found that the combination of DES data, with
priors on . S8 as measured by KIDS and HSC, leads to the constraint . f EDE (z c ) <
+0.032
0.06 without S. H0 ES, and . f EDE (z c ) = 0.062−0.033 with S. H0 ES. However, the profile

8 We refer to this dataset as ‘EFTBOSS’ for brevity.

22 Resolving the Hubble Tension with Early Dark Energy 417

Fig. 22.11 Left panel: 2D posterior distributions from BaseEFTBOSS and BaseTTT-
TEEE+Planck Lensing+S. H0 ES, with and without EFTBOSS data. BaseTTTTEEE refers to
Planck TTTEE+BAO+Pan18, while BaseEFTBOSS refers to EFTBOSS+BBN+Planck Lens-
ing+BAO+Pan18. Right Panel: 2D posterior distributions from ACT+Lens+EFTBOSS in com-
bination with either BaseTT650TEEE, or BaseTTTEEEE with and without S. H0 ES

likelihood analysis of Planck+DES leads to . f EDE (z c ) = 0.061+0.035

−0.034 , . H0 = 70.28 ±
1.33 km/s/Mpc [63, 103], showing that prior-volume effects still play a role in the
apparent constraints from weak lensing data.
It is worth noting that the most recent cosmic shear data that combines both KiDS
and DES are in less tension with Planck [104], and therefore it is expected that up-
to-date constraints on EDE would be weaker than reported here, though a dedicated
analysis is still lacking. Nevertheless, it is clear that low-. S8 measurements cannot be
explained by the presence of EDE [105], and future LSS surveys (e.g. DESI, Euclid,
LSST) will be decisive in testing for the existence of EDE.
If EDE does not explain a low . S8 , one may wonder whether one (or more) addi-
tional degrees of freedom that reduces the amplitude of fluctuations at small scales
could help further, whether it is connected to the EDE dynamics itself, or indepen-
dent of it. For instance, massive neutrinos lead to a power suppression∑ at small scales
which decreases the value of.σ8 . However, the sum of neutrino masses. m ν ∼ 0.3eV
required to resolve the tension is excluded by Planck data (e.g. [49, 106]), although
constraints on the sum of neutrino masses are substantially broadened in extended
cosmologies (e.g. [107, 108]). In the context of EDE, simply adding the neutrino
mass as a free parameter does not change the reconstructed value of . S8 [63, 109].
It was shown however that, when EDE is combined to another proposal that is able
to alleviate the . S8 tension, one can explain both tension independently. For instance,
the existence of a dark sector, where DM is composed of .∼ 5% ultra-light axions
(ULAs), can help in relieving the . S8 tension without affecting the success of the EDE
solution in explaining a high-. H0 [110]. Such a dark sector could be the manifestation
of the presence of numerous scalar-fields that are ubiquitous in string theory [111,
418 V. Poulin and T. L. Smith

112]. Ref. [113] obtained a similar result considering a decaying DM model instead
of the ULA component. Though these models may seem ‘ad hoc’, they demonstrate
that . H0 and . S8 solutions can come from different (potentially disconnected) sectors
and therefore that one should not dismiss EDE models on the basis that they cannot
simultaneously explain both tensions.
Efforts have been made to establish a connection between the resolutions of. S8 and
. H0 within a unified theoretical framework [22, 26, 27, 29]. Examples include EDE

models coupled to DM as a consequence of the Swampland Distance Conjecture

(SDC) [27], or in a chameleon-inspired EDE in Ref. [26], that can induce a “fifth
force” that has the potential of reducing the growth of structure and relieving the
tension with LSS data. Let us also mention the recent progress made in the framework
of the Cold NEDE model, discussed in more details in the next chapter, where the
presence of an axion-like field with mass .∼ 10−27 eV that resolves the . S8 tension
also triggers a phase-transition in the EDE field at precisely the right time to explain
the Hubble tension [114].

22.3.3 Impact on Higher-Redshift LSS Probes

The impact of EDE on the matter power spectrum can profoundly affect the predicted
halo mass function. By running N-body simulations, Ref. [82] established that EDE
predicts only a small (1–10%) increase in the number of halos at .z = 0 for very
massive clusters, but at larger redshifts, that difference can increase tremendously.
For example, the EDE model predicts about 50% more massive clusters of mass. M =
(3 − 5) × 1014 h −1 Mʘ at.z = 1, and the difference is expected to keep increasing with
increasing redshifts, such that differences will soon be tested by JWST observations
[82, 115]. Although a dedicated analysis of the halo mass function in EDE model
at such large reshift is still lacking, it has been pointed out that the first publicly
available data show a somewhat unexpectedly high number of massive galaxies at
high redshift that may be better explained when invoking the presence of EDE [116].
It has also been shown that Lyman-.α data from eBOSS [117], the XQ-100 [118]
and MIKE/HIRES [119] quasar samples are in tension with the EDE prediction,
leading to the strongest constraints to date on . f EDE [120]. It is however important
to bare in mind that even under .ɅCDM, Lyman-.α data favor significantly different
value (at the .3σ level) of the power spectrum tilt and amplitude at small scales
(.k ∼ 1 h/Mpc) and redshift .z = 3 than reconstructed from analysis of CMB data,
which (barring a simple statistical fluctuation) may indicate either a systematic effect,
or some new physics is at play such as the running of the spectral tilt [121] or a new
type of DM properties [122].
22 Resolving the Hubble Tension with Early Dark Energy 419

22.3.4 A Second Cosmic Coincidence Problem

The fact that the EDE contribution become significant precisely close to matter-
radiation equality when it is most impactful on the CMB, leads to an obvious ‘coin-
cidence problem’ known as the ‘why-then’ problem. This echoes the ‘why now’ or
’cosmic coincidence’ problem, that arises from the observation that the densities of
dark energy and matter are similar today, when one would naturally expect very dif-
ferent energy scale. Interestingly, this new ‘why then’ problem may provide insight
into the long-standing ‘why now’ problem.
This new coincidence problem motivates a model-building effort towards dynami-
cally explaining the fact that.z eq is so close to matter-radiation equality. An interesting
starting point is ‘tracking dark energy’ [1, 123], which explains the ‘why now’ prob-
lem by postulating that the density of a scalar field is always a fixed fraction of the
energy density of the dominant component. The models provide an interesting way to
connect various eras of accelerated expansion in the universe. Assisted quintessence
[2], which realizes tracking models through a collection of scalar fields with the same
potential, was recently applied in the EDE context. However, authors of Ref. [21]
found that it could not alleviate the tension.
Other models connect the dynamics of the EDE model to the onset of DM domi-
nation by coupling the EDE scalar field to DM. These include the chameleon EDE
[26] and trigger EDS [59] models. In these model, the dilution of EDE is precisely
triggered around matter-radiation equality, and EDE can affect the growth of matter
perturbations, with interesting implications for the . S8 tension. Alternatively, EDE
may be coupled to another species experiencing a non-relativistic transition, such
as neutrinos. Refs. [12, 124] showed that a conformal coupling between EDE and
neutrinos lead to a modified KG equation, that now counts an additional source term
proportional to the trace of the neutrino energy-momentum tensor. The effect of the
neutrino coupling is to kick the scalar out of its minimum and up its potential right
when the neutrinos become non-relativistic, .Tν ∼ m ν , independently of the value
of the scalar field mass and its initial condition. This model sill lacks a dedicated
analysis in light of up-to-date cosmological data. Interestingly it requires individual
masses of .O(0.1 − 0.5) eV, and it is expected that it can be probed with current and
near-future CMB and LSS observations.
Finally, the last avenue is that this coincidence appears from mere chance. For
instance, the “Axiverse” scenario [112] suggests that many scalar fields with similar
potentials exist in the universe. In Ref. [125], it was argued that the probability
that one of such fields explains DE today while leaving observables unaffected is
not small, thereby solving the “why now” problem. In this context, it is therefore
possible that there exist additional EDE-like fields in the Universe that may have
been relevant (or not) at different eras, with a non-negligible probability that one
such field becomes dynamical close to matter-radiation equality. The EDE field
responsible for the Hubble tension would therefore be a member of a family of such
fields, while other member may play the role of Dark Energy, inflation, and more of
such fields may be detected in the future, depending on whether they left an imprint
420 V. Poulin and T. L. Smith

Fig. 22.12 Constraints on

inflation in the .n s − r plane
from the combination of
Planck+Bicep/Keck+BAO
data, in the .ɅCDM and EDE
models. constraints are
shifted towards larger .n s ∼ 1
in the EDE model, but the
bound on .r is unaffected

on observables (see also Refs. [19, 123, 126–128]). In that context, it is important to
note that it is not unexpected that that an EDE field acting around matter-radiation
equality was detected first precisely because our cosmological data gains most of
their sensitivity right around recombination. If the existence of EDE is confirmed
with future data, the question of connecting EDE with DE and inflation will become
particularly relevant.

22.3.5 Impact of EDE for Inflation

The presence of EDE has important consequences for inflation, as the value of .n s
reconstructed from the fit to CMB data is significantly larger than in .ɅCDM. Under
the .ɅCDM model, current Planck data which favor .n s ∼ 0.96 and .r ≾ 0.1 are in
perfect agreement with predictions from Starobinsky (or . R 2 ) inflation, while they
disfavor power low (or convex) potentials [129]. However, if the Hubble tension
persists, and the preference for EDE models with .n s ≃ 1 increases, it can reshape
our understanding of cosmic inflation [130, 131]. We illustrate bounds on .n s − r
in the .ɅCDM and EDE model on Fig. 22.12. All constraints in the standard .n s − r
plane are shifted towards larger .n s , while the bound on .r is essentially unaffected
by the presence of EDE [131]. In the EDE case, convex potentials can generally
be rescued, with the important consequence that Starobinsky inflation is disfavored
compared to power-law potential, or the curvaton model.
It is important to note that the preference for .n s ∼ 1 is very generic to models
that resolve the Hubble tension by increasing the expansion rate pre-recombination
22 Resolving the Hubble Tension with Early Dark Energy 421

Fig. 22.13 The age of the

universe .tU vs .{h, f EDE (z c )}
in the EDE cosmology. The
purple bands mark the recent
measurements from GCs,
.tU = 13.5 ± 0.27 Gyrs
[133–135]. The dashed black
line shows the prediction
from Eq. 22.6, accounting for
the small difference in .Ωm
between EDE and .ɅCDM

(such as an extra radiation species .ΔNeff ). It follows from the fact that the angular
scale .θ D corresponding to damping roughly depends on

H0
θ (z ∗ ) ∼ √
. D , (22.5)
τ̇ (z ∗ )H (z ∗ )

where .τ̇ (z ∗ ) = n e (z ∗ )xe (z ∗ )σT /(1 + z ∗ ), .n e is the electron density, .xe is the ioniza-
tion fraction, and .σT is the Thomson cross section. If the indirect . H0 increases by a
fraction . f , with .θs (z ∗ ) fixed, .θ D (z ∗ ) increases by . f 1/2 , leading to a suppression of
power, which can be compensated for by increasing .n s . Although details of specific
models will change by how much the value of .n s must increase, this degeneracy
is very generic, and should therefore be considered as an important caveat to our
current understanding of cosmic inflation (regardless of the cosmological nature of
the Hubble tension).

22.3.6 The Age of the Universe

Models that resolve the Hubble tension by modifying the pre-recombination era, such
as Early Dark Energy, generically predict a significantly smaller age of the universe
than .ɅCDM. Indeed, assuming flat-.ɅCDM and radiation is sub-dominant, the age
of the Universe today is approximately [132]
( )−0.28 ( )
Ωm h pl
t ≡ t (a = 1) ≃ tpl
. U . (22.6)
Ωm,pl h

The freedom to adjust .Ωm is limited given the tight constraints on .Ωm from BAO
and Pantheon data. Therefore, one generically expects that increasing . H0 without
changing .Ωm leads to a universe .∼ 7.5% younger than .ɅCDM predicts. This is
illustrated in Fig. 22.13 in the context of EDE.
This raises a potential issue for EDE and other “early-universe” models, as they
can run into tension with the measured ages of old objects such as globular clusters
422 V. Poulin and T. L. Smith

of stars (GCs) [132, 135, 136]. More precisely, recent measurements from GCs
give .tU = 13.5 ± 0.27 Gyrs [133–135] (see also Ref. [137–139] for a discussion of
other measurements), while the EDE cosmology fit to Planck+S. H0 ES predicts .tU =
13.17+0.14
−0.16 Gyr. We illustrate this potential tension in Fig. 22.13, for the EDE model
reconstructed from our PlanckTT650TEEE+ACT and Planck+S. H0 ES analyses, with
the purple band marking the GC measurement that lies somewhat larger than the
model predicts.
The dashed black line shows the prediction from Eq. 22.6, accounting for the
small difference in .Ωm in the EDE model compared to .ɅCDM. Any model that only
modifies the early universe would follow a similar degeneracy line, and therefore
can be tested with more accurate measurements of .tU . Resolving the Hubble tension
while adjusting both measurements of the age of the universe and measurements of
the shape of the expansion rate may require both an early and late-time modification
[135].
Similar considerations apply to cosmic chronometers (e.g. Ref. [137, 140]) which
are currently too imprecise to probe the EDE scenarios [56], but could provide
interesting tests in the future. In fact, Ref. [132] points out that the entire time-
redshift relation is affected by the EDE model, differing from the base .ɅCDM model
by at least .∼ 4% at all .z and .t that can have significant implications for the era of
reionization.

22.4 Conclusions

Early Dark Energy, a new form of dark energy active at early times, has emerged as
a promising way to explain the Hubble tension. In this chapter, we have reviewed
the status of this solution, paying particular attention to the axion-like EDE model,
a toy-model whose flexibility enables us to extract the generic properties required to
explain the Hubble tension. EDE is motivated by two observations: first, that one must
decrease the sound horizon in order to keep the angular-size of the sound horizon
fixed in the CMB and BAO data and predict a magnitude of SN1a in agreement with
the measurements from S. H0 ES. Second, that a higher .h, at fix .Ωm (as imposed by
the BAO and SN1a measurements of the shape of the late-time expansion) requires
an increase in .ωcdm which, if uncompensated, would spoil the fit to CMB data. By
briefly contributing to the Hubble rate in the pre-recombination era, EDE precisely
manages to “kill two birds with one stone”, and provide a good fit to combinations
of up-to-date cosmological data from CMB, BAO and S. H0 ES-calibrated SN1a.
Focusing on the “axion-like” EDE model as a representative example, we have
reviewed current analyses in details in Sect. 22.2, showing that EDE can reduce the
Hubble tension to below the .2σ level. In addition, Planck data constraint the specific
‘microphysics’ of EDE, ultimately tied to the shape of the energy injection, and the
dynamics of perturbations in the EDE component. We have also discussed recent
indications of EDE in ACT data, that are further supported by Planck polarization
data. However, Planck high-.l temperature data are mildly inconsistent with ACT
22 Resolving the Hubble Tension with Early Dark Energy 423

and restricts the preference. Forthcoming data from ACT and SPT will be crucial in
confirming or dismissing the current indications of EDE. Looking ahead, it is antic-
ipated that high-precision measurements at high-.l values (.≿1000) will be sensitive
to the distinct increase in power caused by EDE (primarily due to the reshuffling of
.ɅCDM parameters), while improvements at intermediate .l values (.∼50–500) can
provide insights into the detailed dynamics of EDE perturbations.
There are many interesting implications of EDE that we have reviewed in
Sect. 22.3. First, the EDE cosmology shows an increase of power in the matter
power spectrum (at the 5.−10% level in the range .k ∼ 0.01 − 1 h/Mpc) compared to
.ɅCDM. This gives an opportunity to probe the model with LSS data such as BOSS,

although current data are not constraining enough to rule out EDE. Nevertheless, EDE
cannot explain the low-. S8 measured by current galaxy weak-lensing surveys such as
KiDS, DES and HSC, triggering a large model-building effort towards embedding
a solution to the . S8 tension within the EDE framework. Second, EDE predicts a
much younger universe, which may also require some late-time modification if mea-
surements of the age of universe become more precise. Third, the existence of an
EDE phase may have important consequences for inflation, as the value of .n s recon-
structed in EDE models that resolve the Hubble tension favor curvaton models over
Starobinsky inflation [130, 131]. At the theoretical level, EDE also triggers some
fine-tuning issues in particular a ‘why-then’ problem that resembles the ‘why-now’
problem and suggests that EDE may shed light on this long-standing problem. In
fact, connections between inflation, late DE and early DE are still largely unexplored,
and may provide an interesting path forward for model-building with undetermined
consequences for EDE models.
In this chapter, we have argued that EDE is a promising mechanism to resolve
the . H0 tension, but still lacks some missing pieces, for various observational and
theoretical reasons. Future work will be of utmost important in order to find an
unambiguous solutions to current cosmic tensions, in particular given the numerous
surveys to come within the next decade (DESI, Euclid, SO, among others), and
hopefully shed light on the unknown nature of dark matter and dark energy.

Acknowledgements This project has received support from the European Union’s Horizon 2020
research and innovation program under the Marie Skodowska-Curie grant agreement No 860881-
HIDDeN. This project has received funding from the European Research Council (ERC) under the
European Union’s HORIZON-ERC-2022 (Grant agreement No. 101076865).

References

1. S. Dodelson, M. Kaplinghat, E. Stewart, Solving the coincidence problem?: tracking oscil-

lating energy. Phys. Rev. Lett. 85, 5276–5279 (2000)
2. S.A. Kim, A.R. Liddle, S. Tsujikawa, Dynamics of assisted quintessence. Phys. Rev. D 72,
043506 (2005)
3. M. Doran, M.J. Lilley, J. Schwindt, C. Wetterich, Quintessence and the separation of CMB
peaks. Astrophys. J. 559, 501–506 (2001)
424 V. Poulin and T. L. Smith

4. C. Wetterich, Phenomenological parameterization of quintessence. Phys. Lett. B 594, 17–22

(2004)
5. M. Doran, G. Robbers, Early dark energy cosmologies. JCAP 06, 026 (2006)
6. T. Karwal, M. Kamionkowski, Dark energy at early times, the Hubble parameter, and the
string axiverse. Phys. Rev. D 94(10), 103523 (2016)
7. V. Poulin, T.L. Smith, T. Karwal, M. Kamionkowski, Early dark energy can resolve the Hubble
tension. Phys. Rev. Lett. 122(22), 221301 (2019)
8. T.L. Smith, V. Poulin, M.A. Amin, Oscillating scalar fields and the Hubble tension: a resolution
with novel signatures. Phys. Rev. D 101(6), 063523 (2020)
9. P. Agrawal, F.-Y. Cyr-Racine, D. Pinner, L. Randall, Rock ’n’ roll solutions to the Hubble
tension (2019)
10. M.-X. Lin, G. Benevento, W. Hu, M. Raveri, Acoustic dark energy: potential conversion of
the Hubble tension. Phys. Rev. D 100(6), 063542 (2019)
11. S. Alexander, E. McDonough, Axion-Dilaton destabilization and the Hubble tension. Phys.
Lett. B 797, 134830 (2019)
12. J. Sakstein, M. Trodden, Early dark energy from massive neutrinos as a natural resolution of
the Hubble tension. Phys. Rev. Lett. 124(16), 161301 (2020)
13. A. Gogoi, R. Kumar Sharma, P. Chanda, S. Das, Early mass-varying neutrino dark energy:
nugget formation and Hubble anomaly. Astrophys. J. 915, 132 (2021)
14. F. Niedermann, M.S. Sloth, New early dark energy. Phys. Rev. D 103(4), L041303 (2021)
15. F. Niedermann, M.S. Sloth, Resolving the Hubble tension with new early dark energy. Phys.
Rev. D 102(6), 063527 (2020)
16. F. Niedermann, M.S. Sloth, Hot new early dark energy. Phys. Rev. D 105(6), 063509 (2022)
17. G. Ye, Y.-S. Piao, Is the Hubble tension a hint of AdS phase around recombination? Phys.
Rev. D 101(8), 083507 (2020)
18. K.V. Berghaus, T. Karwal, Thermal friction as a solution to the Hubble tension. Phys. Rev. D
101(8), 083537 (2020)
19. K. Freese, M.W. Winkler, Chain early dark energy: a proposal for solving the Hubble tension
and explaining today’s dark energy. Phys. Rev. D 104(8), 083533 (2021)
20. M. Braglia, W.T. Emond, F. Finelli, A.E. Gumrukcuoglu, K. Koyama, Unified framework for
early dark energy from .α-attractors. Phys. Rev. D 102(8), 083513 (2020)
21. V.I. Sabla, R.R. Caldwell, No. H0 assistance from assisted quintessence. Phys. Rev. D 103(10),
103506 (2021)
22. V.I. Sabla, R.R. Caldwell, Microphysics of early dark energy. Phys. Rev. D 106(6), 063526
(2022)
23. A. Gómez-Valent, Z. Zheng, L. Amendola, V. Pettorino, C. Wetterich, Early dark energy in
the pre- and postrecombination epochs. Phys. Rev. D 104(8), 083536 (2021)
24. A. Moss, E. Copeland, S. Bamford, T. Clarke, A model-independent reconstruction of dark
energy to very high redshift 9 (2021)
25. E. Guendelman, R. Herrera, D. Benisty, Unifying inflation with early and late dark energy
with multiple fields: Spontaneously broken scale invariant two measures theory. Phys. Rev.
D 105(12), 124035 (2022)
26. T. Karwal, M. Raveri, B. Jain, J. Khoury, M. Trodden, Chameleon early dark energy and the
Hubble tension. arXiv e-prints, arXiv:2106.13290 (2021)
27. E. McDonough, M.-X. Lin, J.C. Hill, W. Hu, S. Zhou, The early dark sector, the Hubble
tension, and the swampland 12 (2021)
28. H. Wang, Y.-S. Piao, A fraction of dark matter faded with early dark energy? 9 (2022)
29. S. Alexander, H. Bernardo, M.W. Toomey, Addressing the Hubble and . S8 tensions with a
kinetically mixed dark sector 7 (2022)
30. E. McDonough, M. Scalisi, Towards early dark energy in string theory 8 (2022)
31. S. Nakagawa, F. Takahashi, W. Yin, Early dark energy by dark Higgs, and axion-induced
non-thermal trapping 9 (2022)
32. A. Gómez-Valent, Z. Zheng, L. Amendola, C. Wetterich, V. Pettorino, Coupled and uncoupled
early dark energy, massive neutrinos and the cosmological tensions 7 (2022)
22 Resolving the Hubble Tension with Early Dark Energy 425

33. H. Mohseni Sadjadi, V. Anari, Early dark energy and the screening mechanism 5 (2022)
34. K. Kojima, Y. Okubo, Early dark energy from a higher-dimensional gauge theory. Phys. Rev.
D 106(6), 063540 (2022)
35. T. Rudelius, Constraints on early dark energy from the axion weak gravity conjecture 3 (2022)
36. V.K. Oikonomou, Unifying inflation with early and late dark energy epochs in axion . F(R)
gravity. Phys. Rev. D 103(4), 044036 (2021)
37. S.X. Tian, Z.-H. Zhu, Early dark energy in .k-essence. Phys. Rev. D 103(4), 043518 (2021)
38. M. Maziashvili, Inflaton-driven early dark energy. Astropart. Phys. 145, 102792 (2023)
39. D. Blas, J. Lesgourgues, T. Tram, The cosmic linear anisotropy solving system (CLASS) II:
approximation schemes. JCAP 1107, 034 (2011)
40. V. Poulin, T.L. Smith, T. Karwal, The ups and downs of early dark energy solutions to the
Hubble tension: a review of models, hints and constraints circa 2023 2 (2023)
41. H. Abe, T. Kobayashi, H. Otsuka, Natural inflation with and without modulations in type IIB
string theory. JHEP 04, 160 (2015)
42. K. Choi, H. Kim, Aligned natural inflation with modulations. Phys. Lett. B 759, 520–527
(2016)
43. R. Kappl, H.P. Nilles, M.W. Winkler, Modulated natural inflation. Phys. Lett. B 753, 653–659
(2016)
44. N. Kaloper, Dark energy, . H0 and weak gravity conjecture. Int. J. Mod. Phys. D 28(14),
1944017 (2019)
45. M. Czerny, T. Higaki, F. Takahashi, Multi-natural inflation in supergravity and BICEP2. Phys.
Lett. B 734, 167–172 (2014)
46. D. Croon, V. Sanz, Saving natural inflation. JCAP 02, 008 (2015)
47. T. Higaki, F. Takahashi, Elliptic inflation: interpolating from natural inflation to R.2 -inflation.
JHEP 03, 129 (2015)
48. M. Cicoli, M. Licheri, R. Mahanta, E. McDonough, F.G. Pedro, M. Scalisi, Early dark energy
in type IIB string theory 3 (2023)
49. N. Aghanim et al., Planck 2018 results. VI. Cosmological parameters. Astron. Astrophys.
641, A6 (2020) [Erratum: Astron.Astrophys. 652, C4 (2021)]
50. S. Alam et al., The clustering of galaxies in the completed SDSS-III Baryon oscillation
spectroscopic survey: cosmological analysis of the DR12 galaxy sample. Mon. Not. Roy.
Astron. Soc. 470(3), 2617–2652 (2017)
51. A.J. Ross, L. Samushia, C. Howlett, W.J. Percival, A. Burden, M. Manera, The clustering of
the SDSS DR7 main Galaxy sample—I. A 4 percent distance measure at.z = 0.15. Mon. Not.
Roy. Astron. Soc. 449(1), 835–847 (2015)
52. F. Beutler, C. Blake, M. Colless, D.H. Jones, L. Staveley-Smith, L. Campbell, Q. Parker,
W. Saunders, F. Watson, The 6dF galaxy survey: Baryon acoustic oscillations and the local
Hubble constant. Mon. Not. Roy. Astron. Soc. 416, 3017–3032 (2011)
53. D.M. Scolnic et al., The complete light-curve sample of spectroscopically confirmed SNe
Ia from Pan-STARRS1 and cosmological constraints from the combined pantheon sample.
Astrophys. J. 859(2), 101 (2018)
54. A.G. Riess et al., A comprehensive measurement of the local value of the Hubble constant
with 1 km/s/Mpc uncertainty from the Hubble space telescope and the S. H0 ES team 12 (2021)
55. M. Raveri, W. Hu, Concordance and discordance in cosmology (2018)
56. N. Schöneberg, G. Franco Abellán, A. Pérez Sánchez, S.J. Witte, V. Poulin, J. Lesgourgues,
The . H0 Olympics: a fair ranking of proposed models 7 (2021)
57. V. Poulin, T.L. Smith, D. Grin, T. Karwal, M. Kamionkowski, Cosmological implications of
ultralight axionlike fields. Phys. Rev. D 98(8), 083525 (2018)
58. M.C. Johnson, M. Kamionkowski, Dynamical and gravitational instability of oscillating-field
dark energy and dark matter. Phys. Rev. D 78, 063010 (2008)
59. M.-X. Lin, E. McDonough, J.C. Hill, W. Hu, A dark matter trigger for early dark energy
coincidence 12 (2022)
60. T. Simon, P. Zhang, V. Poulin, T.L. Smith, Updated constraints from the effective field theory
analysis of BOSS power spectrum on early dark energy. Phys. Rev. D. In press, 8 (2022)
426 V. Poulin and T. L. Smith

61. A. Lewis, S. Bridle, Cosmological parameters from CMB and other data: A Monte Carlo
approach. Phys. Rev. D 66, 103511 (2002)
62. B. Audren, J. Lesgourgues, K. Benabed, S. Prunet, Conservative constraints on early cosmol-
ogy: an illustration of the Monte Python cosmological parameter inference code. JCAP 02,
001 (2013)
63. R. Murgia, G.F. Abellán, V. Poulin, Early dark energy resolution to the Hubble tension in
light of weak lensing surveys and lensing anomalies. Phys. Rev. D 103(6), 063502 (2021)
64. T.L. Smith, V. Poulin, J.L. Bernal, K.K. Boddy, M. Kamionkowski, R. Murgia, Early dark
energy is not excluded by current large-scale structure data. Phys. Rev. D 103(12), 123542
(2021)
65. L. Herold, E.G.M. Ferreira, E. Komatsu, New constraint on early dark energy from Planck
and BOSS data using the profile likelihood 12 (2021)
66. A. Gómez-Valent, Fast test to assess the impact of marginalization in Monte Carlo analyses
and its application to cosmology. Phys. Rev. D 106(6), 063506 (2022)
67. B. Hadzhiyska, K. Wolz, S. Azzoni, D. Alonso, C. García-García, J. Ruiz-Zapatero, A. Slosar,
Cosmology with 6 parameters in the stage-IV era: efficient marginalisation over nuisance
parameters 1 (2023)
68. J. Neyman, Outline of a theory of statistical estimation based on the classical theory of
probability. Phil. Trans. Roy. Soc. Lond. A 236(767), 333–380 (1937)
69. G.J. Feldman, R.D. Cousins, A unified approach to the classical statistical analysis of small
signals. Phys. Rev. D 57, 3873–3889 (1998)
70. P.A.R. Ade et al., Planck intermediate results. XVI. Profile likelihoods for cosmological
parameters. Astron. Astrophys. 566, A54 (2014)
71. P. Ade et al., The Simons observatory: science goals and forecasts. JCAP 02, 056 (2019)
72. K.N. Abazajian et al., CMB-S4 Science Book, vol. 10, 1st edn. (2016)
73. S. Aiola et al., The Atacama cosmology telescope: DR4 maps and cosmological parameters.
JCAP 12, 047 (2020)
74. D. Dutcher et al., Measurements of the E-mode polarization and temperature-E-mode corre-
lation of the CMB from SPT-3G 2018 data. arXiv e-prints, arXiv:2101.01684 (2021)
75. L. Balkenhol et al., A measurement of the CMB temperature power spectrum and constraints
on cosmology from the SPT-3G 2018 TT/TE/EE data set 12 (2022)
76. Y. Huang, G.E. Addison, J.L. Weiland, C.L. Bennett, Assessing consistency between WMAP
9-year and Planck 2015 temperature power spectra. Astrophys. J. 869(1), 38 (2018)
77. J.C. Hill et al., The Atacama cosmology telescope: constraints on pre-recombination early
dark energy 9 (2021)
78. T.L. Smith, V. Poulin, T. Simon, Assessing the robustness of sound horizon-free determina-
tions of the Hubble constant 8 (2022)
79. T.L. Smith, M. Lucca, V. Poulin, G.F. Abellan, L. Balkenhol, K. Benabed, S. Galli, R. Murgia,
Hints of early dark energy in Planck, SPT, and ACT data: new physics or systematics? 2 (2022)
80. W. Handley, P. Lemos, Quantifying the global parameter tensions between ACT, SPT and
Planck. Phys. Rev. D 103(6), 063529 (2021)
81. E. Di Valentino, W. Giarè, A. Melchiorri, J. Silk, A health checkup test of the standard cos-
mological model in view of recent Cosmic Microwave Background Anisotropies experiments
9 (2022)
82. A. Klypin, V. Poulin, F. Prada, J. Primack, M. Kamionkowski, V. Avila-Reese, A. Rodriguez-
Puebla, P. Behroozi, D. Hellinger, T.L. Smith, Clustering and halo abundances in early dark
energy cosmological models. Mon. Not. Roy. Astron. Soc. 504(1), 769–781 (2021)
83. J.C. Hill, E. McDonough, M.W. Toomey, S. Alexander, Early dark energy does not restore
cosmological concordance. Phys. Rev. D 102(4), 043507 (2020)
84. D. Baumann, A. Nicolis, L. Senatore, M. Zaldarriaga, Cosmological non-linearities as an
effective fluid. JCAP 07, 051 (2012)
85. J.J.M. Carrasco, M.P. Hertzberg, L. Senatore, The effective field theory of cosmological large
scale structures. JHEP 09, 082 (2012)
22 Resolving the Hubble Tension with Early Dark Energy 427

86. L. Senatore, M. Zaldarriaga, The IR-resummed effective field theory of large scale structures.
JCAP 1502(02), 013 (2015)
87. L. Senatore, Bias in the effective field theory of large scale structures. JCAP 1511(11), 007
(2015)
88. L. Senatore, M. Zaldarriaga, Redshift space distortions in the effective field theory of large
scale structures (2014)
89. A. Perko, L. Senatore, E. Jennings, R.H. Wechsler, Biased tracers in redshift space in the EFT
of large-scale structure (2016)
90. G. D’Amico, J. Gleyzes, N. Kokron, D. Markovic, L. Senatore, P. Zhang, F. Beutler, H. Gil-
Marín, The cosmological analysis of the SDSS/BOSS data from the effective field theory of
large-scale structure. JCAP 05, 005 (2020)
91. M.M. Ivanov, M. Simonovi, M. Zaldarriaga, Cosmological parameters from the BOSS galaxy
power spectrum (2019)
92. G. D’Amico, L. Senatore, P. Zhang, H. Zheng, The Hubble tension in light of the full-shape
analysis of large-scale structure data. JCAP 05, 072 (2021)
93. M.M. Ivanov, E. McDonough, J.C. Hill, M. Simonović, M.W. Toomey, S. Alexander, M.
Zaldarriaga, Constraining early dark energy with large-scale structure. Phys. Rev. D 102(10),
103502 (2020)
94. A. Reeves, L. Herold, S. Vagnozzi, B.D. Sherwin, E.G.M. Ferreira, Restoring cosmological
concordance with early dark energy and massive neutrinos? 7 (2022)
95. T. Simon, G. Franco Abellán, P. Du, V. Poulin, Y. Tsai, Constraining decaying dark matter
with BOSS data and the effective field theory of large-scale structures. Phys. Rev. D. 106(2),
023516 (2022)
96. C. Heymans et al., CFHTLenS: the Canada-France-Hawaii telescope lensing survey. Mon.
Not. Roy. Astron. Soc. 427, 146 (2012)
97. T.M.C. Abbott et al., Dark energy survey year 3 results: cosmological constraints from galaxy
clustering and weak lensing. Phys. Rev. D 105(2), 023520 (2022)
98. C. Heymans et al., KiDS-1000 cosmology: multi-probe weak gravitational lensing and spec-
troscopic galaxy clustering constraints 7 (2020)
99. C. Hikage et al., Cosmology from cosmic shear power spectra with Subaru hyper suprime-
cam first-year data. Publ. Astron. Soc. Jap. 71(2), Publications of the Astronomical Society
of Japan, vol. 71, no. 2, April 2019, p. 43. https://doi.org/10.1093/pasj/psz010 (2019)
100. P.A.R. Ade et al., Planck 2015 results. XXIV. cosmology from Sunyaev-Zeldovich cluster
counts. Astron. Astrophys. 594, A24 (2016)
101. K. Jedamzik, L. Pogosian, G.-B. Zhao, Why reducing the cosmic sound horizon alone can
not fully resolve the Hubble tension. Commun. in Phys. 4, 123 (2021)
102. S. Vagnozzi, Consistency tests of .ɅCDM from the early ISW effect: implications for early-
time new physics and the Hubble tension. arXiv e-prints, arXiv:2105.10425 (2021)
103. L. Herold, E.G.M. Ferreira, Resolving the Hubble tension with early dark energy 10 (2022)
104. T.M.C. Abbott et al., DES Y3 + KiDS-1000: consistent cosmology combining cosmic shear
surveys 5 (2023)
105. L.F. Secco, T. Karwal, W. Hu, E. Krause, The role of the Hubble scale in the weak lensing
versus CMB tension 9 (2022)
106. G.F. Abellan, R. Murgia, V. Poulin, J. Lavalle, Hints for decaying dark matter from . S8 mea-
surements 8 (2020)
107. Z. Chacko, A. Dev, P. Du, V. Poulin, Y. Tsai, Cosmological limits on the neutrino mass and
lifetime. JHEP 04, 020 (2020)
108. I.M. Oldengott, G. Barenboim, S. Kahlen, J. Salvado, D.J. Schwarz, How to relax the cosmo-
logical neutrino mass bound. JCAP 04, 049 (2019)
109. E. Fondi, A. Melchiorri, L. Pagano, No evidence for EDE from Planck data in extended
scenarios. Astrophys. J. Lett. 931, L18 (2022)
110. I.J. Allali, M.P. Hertzberg, F. Rompineve, Dark sector to restore cosmological concordance.
Phys. Rev. D 104(8), L081303 (2021)
111. P. Svrcek, E. Witten, Axions in string theory. JHEP 06, 051 (2006)
428 V. Poulin and T. L. Smith

112. A. Arvanitaki, S. Dimopoulos, S. Dubovsky, N. Kaloper, J. March-Russell, String axiverse.

Phys. Rev. D 81, 123530 (2010)
113. S.J. Clark, K. Vattis, J. Fan, S.M. Koushiappas, The . H0 and . S8 tensions necessitate early and
late time changes to .ɅCDM 10 (2021)
114. J.S. Cruz, F. Niedermann, M.S. Sloth, Cold new early dark energy pulls the trigger on the . H0
and . S8 tensions: a simultaneous solution to both tensions without new ingredients 5 (2023)
115. R. Endsley, P. Behroozi, D.P. Stark, C.C. Williams, B.E. Robertson, M. Rieke, S. Gottlöber,
G. Yepes, Clustering with JWST: constraining galaxy host halo masses, satellite quenching
efficiencies, and merger rates at z= 4-10. Monthly Notices of the Royal Astron. Soc. 493,
1178–1196, 02 (2020)
116. M. Boylan-Kolchin, Stress testing .ɅCDM with high-redshift galaxy candidates 8 (2022)
117. S. Chabanier et al., The one-dimensional power spectrum from the SDSS DR14 Ly.α forests.
JCAP 07, 017 (2019)
118. V. Iršič, M. Viel, T.A.M. Berg, V. D’Odorico, M.G. Haehnelt, S. Cristiani, G. Cupani, T.-
S. Kim, S. López, S. Ellison, G.D. Becker, L. Christensen, K.D. Denney, G. Worseck, J.S.
Bolton, The Lyman ? forest power spectrum from the XQ-100 legacy survey. Monthly Notices
of the Royal Astron. Soc. 466, 4332–4345, 12 (2016)
119. M. Viel, G.D. Becker, J.S. Bolton, M.G. Haehnelt, Warm dark matter as a solution to the
small scale crisis: new constraints from high redshift Lyman-.α forest data. Phys. Rev. D 88,
043502 (2013)
120. S. Goldstein, J.C. Hill, V. Iršič, B.D. Sherwin, Canonical Hubble-tension-resolving early dark
energy cosmologies are inconsistent with the Lyman-.α forest 3 (2023)
121. N. Palanque-Delabrouille, C. Yèche, N. Schöneberg, J. Lesgourgues, M. Walther, S. Chaban-
ier, E. Armengaud, Hints, neutrino bounds and WDM constraints from SDSS DR14 Lyman-.α
and Planck full-survey data. JCAP 04, 038 (2020)
122. D.C. Hooper, M. Lucca, Hints of dark matter-neutrino interactions in Lyman-.α data. Phys.
Rev. D 105(10), 103504 (2022)
123. K. Griest, Toward a possible solution to the cosmic coincidence problem. Phys. Rev. D 66,
123501 (2002)
124. M. Carrillo González, Q. Liang, J. Sakstein, M. Trodden, Neutrino-assisted early dark energy:
theory and cosmology. JCAP 04, 063 (2021)
125. M. Kamionkowski, J. Pradler, D.G.E. Walker, Dark energy from the string axiverse. Phys.
Rev. Lett. 113(25), 251302 (2014)
126. E.V. Linder, T.L. Smith, Dark before light: testing the cosmic expansion history through the
cosmic microwave background. JCAP 1104, 001 (2011)
127. J. Samsing, E.V. Linder, T.L. Smith, Model independent early expansion history and dark
energy. Phys. Rev. D 86, 123504 (2012)
128. A. Hojjati, E.V. Linder, J. Samsing, New constraints on the early expansion history of the
universe. Phys. Rev. Lett. 111(4), 041301 (2013)
129. Y. Akrami et al., Planck 2018 results. X. Constraints on inflation. Astron. Astrophys. 641,
A10 (2020)
130. F. Takahashi, W. Yin, Cosmological implications of ns 1 in light of the Hubble tension. Phys.
Lett. B 830, 137143 (2022)
131. J.S. Cruz, F. Niedermann, M.S. Sloth, A grounded perspective on new early dark energy using
ACT, SPT, and BICEP/Keck 9 (2022)
132. M. Boylan-Kolchin, D.R. Weisz, Uncertain times: the redshift-time relation from cosmology
and stars. Mon. Not. Roy. Astron. Soc. 505(2), 2764–2783 (2021)
133. D. Valcin, J.L. Bernal, R. Jimenez, L. Verde, B.D. Wandelt, Inferring the age of the universe
with globular clusters. JCAP 12, 002 (2020)
134. D. Valcin, R. Jimenez, L. Verde, J.L. Bernal, B.D. Wandelt, The age of the universe with
globular clusters: reducing systematic uncertainties. JCAP 08, 017 (2021)
135. J.L. Bernal, L. Verde, R. Jimenez, M. Kamionkowski, D. Valcin, B.D. Wandelt, The trouble
beyond . H0 and the new cosmic triangles. Phys. Rev. D 103(10), 103533 (2021)
22 Resolving the Hubble Tension with Early Dark Energy 429

136. S. Vagnozzi, F. Pacucci, A. Loeb, Implications for the Hubble tension from the ages of the
oldest astrophysical objects. JHEAp 36, 27–35 (2022)
137. M. Moresco et al., Unveiling the universe with emerging cosmological probes. Living Rev.
Rel. 25(1), 6 (2022)
138. A. Cimatti, M. Moresco, Revisiting oldest stars as cosmological probes: new constraints on
the Hubble constant 2 (2023)
139. J.M. Ying, B. Chaboyer, E.M. Boudreaux, C. Slaughter, M. Boylan-Kolchin, D. Weisz, The
absolute age of M92. Astron. J. 166(1), 18 (2023)
140. S. Vagnozzi, A. Loeb, M. Moresco, Eppur è piatto? The cosmic chronometers take on spatial
curvature and cosmic concordance. Astrophys. J. 908(1), 84 (2021)
Chapter 23
New Early Dark Energy as a Solution
to the H0 and S8 Tensions
. .

Florian Niedermann and Martin S. Sloth

23.1 Introduction

New Early Dark Energy (NEDE) is a promising framework for resolving the Hubble
and the S8 tension using well-motivated standard particle physics methods. Within
the NEDE framework, the Hubble tension is resolved by a fast-triggered phase tran-
sition around the eV energy scale just before matter-radiation equality and the sub-
sequent recombination time. The fast-triggered phase transition, distinct from other
EDE-like models (see [1] and for reviews [2, 3]), leads to the quick decay of an early
dark energy component, resolving the Hubble tension.
The trigger of the phase transition can be realized by different microphysical
degrees of freedom in different NEDE models. NEDE is, therefore, not a single
model but an entire framework for addressing the Hubble tension, depending on the
trigger. Simple cases discussed so far are the cases where the trigger of the phase
transition is a second scalar field (Cold NEDE [4, 5]) or a non-vanishing temperature
of the dark sector (Hot NEDE [6, 7]). Since the CMB and LSS are very precise probes
of perturbations, different NEDE models differ at the phenomenological level, as well
as being distinguishable from other EDE-type models, and we expect that we will
be able to identify the correct model with future data.

F. Niedermann (B)
Nordita, KTH Royal Institute of Technology and Stockholm University, Hannes Alfvéns väg 12,
106 91 Stockholm, Sweden
e-mail: florian.niedermann@su.se
M. S. Sloth
Universe-Origins, University of Southern Denmark, Campusvej 55, 5230 Odense M, Denmark
e-mail: sloth@sdu.dk

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 431
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_23
432 F. Niedermann and M. S. Sloth

The physics of NEDE has some similarities with old inflation models1 [8], in which
inflation ends in a first-order phase transition, as opposed to the second-order slow-
roll over phase transition of new inflation [9, 10]. The “new” in NEDE, therefore,
refers to the similar but opposite role of NEDE versus other EDE-like models when
compared to old versus new inflation. The NEDE framework differs as much from
other EDE-like models as old inflation differs from new inflation.
First-order phase transitions happen in nature at virtually all energy scales, like
the boiling or freezing of water or the QCD phase transition. Since we know that we
have an extended dark sector consisting of at least both dark energy and dark matter,
it is compelling to think that similar phase transitions can occur in the dark sector.
This makes NEDE conceptually appealing at the theoretical level. In fact, in the Cold
NEDE model, the two fields are naturally realized as axions [11], which might reside
in the axiverse of string theory [12, 13], one with a mass at the eV scale, around the
mass scale of also the QCD axion, and the other an ultralight axion with a mass of
order .10−27 eV playing the role of a small fuzzy Dark Matter (DM) component [14–
19]. In the landscape [20], or in a dynamical solution to the cosmological constant
problem, phase transitions triggered by multiple axions may indeed be expected
and be consistent [21] with expectations based on swampland conjectures, which
requires the DE sector to have non-trivial dynamics [22, 23]. On the other hand, in
the Hot NEDE model, an eV phase transition could explain how the neutrinos get
their mass in a Higgs-like mechanism, where the NEDE boson plays a similar role
that the Higgs boson plays in the electroweak phase-transition in giving mass to all
the other Standard Model (SM) particles [6, 7].
On the phenomenological level, NEDE distinguishes itself from other EDE-like
models, not only by its fast-triggered phase transition on the background level but
particularly at the perturbation level. The discontinuous change in the equation of
state at the background level, characteristic of a first-order phase transition, is also
felt by the perturbations. Since the transition happens at different times in different
places due to the unavoidable adiabatic perturbations in the trigger field, the evolution
of perturbations across the phase transition is non-trivial and distinct for the NEDE
model. As these perturbations get imprinted in the CMB and LSS today, measuring
the detailed perturbations is our primary way to test different NEDE models against
each other and other EDE-like models (see [2, 5, 24–27] for a first phenomenological
model comparison).
One prominent way the differences between models belonging to the NEDE
framework and other EDE-like models manifest themselves at the perturbation level
is through their ability to address the S8 problem in tandem with the Hubble tension.
In the Cold NEDE model, it was originally assumed that the energy density of the
light trigger field is completely sub-dominant (in that case the . S8 tension remained
similar significant as in.ɅCDM [27, 28]). Still, if one allows it to contribute just.0.5%
of the energy density of the universe, it can resolve the S8 tension while also improv-
ing the ability of NEDE to solve the . H0 tension [11]. This enables Cold NEDE to
solve both the Hubble tension and the S8 tension with only four new degrees of free-

1 Although a fast triggered second order phase transition, called Hybrid NEDE, is also possible [5].
23 New Early Dark Energy as a Solution to the . H0 and . S8 Tensions 433

dom compared to .ɅCDM (AxiEDE model also has four extra parameters but cannot
address the S8 problem without adding more [29, 30]). In the Hot NEDE model, the
non-vanishing dark temperature, different from the temperature of the visible sector,
requires an extended dark sector. The extended thermal dark sector allows for non-
trivial Dark Matter (DM) and Dark Radiation (DR) evolution and interactions, which
can also lead to resolutions of the S8 problem. In both Cold NEDE and Hot NEDE,
the trigger required for solving the Hubble tension is also central to resolving the
S8 tension. This sets NEDE apart from other EDE-like models, which do not have a
trigger mechanism like NEDE.
A novel feature of NEDE at the more technical level is that the phase transition
happens when the smallest length scales relevant to the CMB have already entered
the horizon. In other cosmological phase transitions, like the QCD phase transition,
all relevant cosmological scales are super-horizon at the time of the phase transition,
and the matching of the perturbations across the discontinuous phase transition is
trivial as the co-moving curvature perturbation is conserved on super-horizon scales.
In the case of NEDE, the matching of perturbations on sub-horizon scales across
the phase transition has therefore required novel theoretical developments of the
perturbation theory, which have been implemented in a numerical Boltzmann code
called TriggerCLASS.2
The treatment of perturbations through a fast-triggered phase transition is what
we describe as a phenomenological model below. The most important input for deter-
mining the phenomenological NEDE model is the choice of trigger (scalar field in
Cold NEDE versus temperature in Hot NEDE), as well as the details of the NEDE
fluid after the transition. It is the phenomenological model that has been imple-
mented into the Boltzmann solver. It is crucial that the phenomenological model has
its motivation in a fundamental microphysical realization, and the relation of NEDE
to particle physics is through the microphysical description. In a minimal micro-
scopic model, there is a one-to-one mapping between the fundamental parameters of
the microphysical description and the degrees of freedom in the phenomenological
model fitted to data (more generally, at least a surjective mapping). Using cosmo-
logical observations and constraints on the phenomenological model, this mapping
enables us also to constrain the underlying fundamental particle physics. From par-
ticle physics, we will have separate constraints on the microscopic particle physics
parameters, which makes the NEDE framework highly predictive, promising to take
precision particle cosmology to a new level.

23.2 A Phenomenological Model

We start by outlining a purely phenomenological model describing a fast-triggered

decay of a new phase of dark energy before discussing two possible microscopic
realizations in Sects. 23.3 and 23.4. As we will argue in more detail later, the pres-
ence of the trigger is crucial for having a fast phase transition that completes quickly

2 https://github.com/NEDE-Cosmo/TriggerCLASS.
434 F. Niedermann and M. S. Sloth

on cosmological time scales set by .1/H . In particular, this avoids large inhomo-
geneities that could be produced in a slow phase transition and would conflict with
the homogeneity of the CMB.

23.2.1 Fluid Description

Our phenomenological model introduces an ideal fluid, which, on background level,

is characterized by an equation of state parameter .wNEDE (τ ), relating its pressure
and energy density, . p̄NEDE (τ ) = wNEDE (τ )ρ̄NEDE (τ ). General for a first-order phase
transition is a discontinuous change in the properties of the microscopic state at the
phase boundary, reflected in our case in the sharp change in the equation of state.
Initially, NEDE behaves as vacuum energy or a cosmological constant equiva-
lently and, accordingly, .wNEDE = −1 or .ρ̄NEDE = const. After the phase transition
at (conformal) time .τ∗ , we require the fluid to subside at least as fast as radiation.
This corresponds to an equation of state parameter3 .wNEDE (τ ) ≥ 1/3, which, at this
stage, can be a general time-dependent function. Integrating the energy conservation
'
equation .ρ̄NEDE + 3H (ρ̄NEDE + p̄NEDE ) = 0, fixes the time-dependence of .ρNEDE ,
┌ ʃ ]
[ ]
.ρ̄NEDE (τ ) = ρ̄NEDE (τ∗ ) exp −3 dτ̃ 1 + wNEDE (τ̃ ) H (τ̃ ) , (23.1)

where .H = a ' (τ )/a(τ ) = a H is the conformal Hubble parameter and .a(τ ) is the
scale factor. We thus model the phase transition as a discontinuity in .wNEDE at time
.τ∗ , explicitly

wNEDE = −1
. for τ < τ∗ ,
1/3 ≤wNEDE (τ ) ≤ 1 for τ ≥ τ∗ , (23.2)

where the upper bound on .wNEDE ≤ 1 ensures that the fluid is compatible with the
dominant energy condition. The abrupt change of .wNEDE (τ ) at .τ∗ , which is a distinc-
tive feature of NEDE, underlies the assumption that the phase transition completes
quickly on cosmological time scales. With these choices, the NEDE fluid never dom-
inates the energy budget and instead gives rise to a short energy injection that peaks
at the time of the phase transition. The larger .wNEDE , the quicker the fall-off of .ρNEDE
and the more localized the energy injection becomes. As a rule of thumb, the NEDE
phenomenology favors .wNEDE ≈ 2/3 just after the phase transition (this is consistent
with the findings for EDE [1]). We plot the fractional energy density of NEDE as
the red curve in Fig. 23.1 for a typical NEDE cosmology. The maximal fraction is
defined as

f
. NEDE = ρ̄NEDE (τ∗ )/ρ̄tot (τ∗ ) (23.3)

3 We note that this condition can be relaxed after the onset of matter domination to .wNEDE (τ ) ≥ 0.
23 New Early Dark Energy as a Solution to the . H0 and . S8 Tensions 435

Fig. 23.1 NEDE background evolution in the case of Cold NEDE for one of the best-fit cosmologies
(using CMB, BAO, supernovae, and LSS data) in [11]. NEDE leads to a sharp energy injection
shortly before matter-radiation equality. The phase transition at time .τ∗ (dotted vertical line) is
modeled as a discontinuity in the equation of state parameter. The trigger is sub-dominant (yet not
negligible) throughout the evolution

and corresponds to the peak labeled “NEDE” in Fig. 23.1. The second peak, labeled
“CC”, corresponds to the (final) cosmological constant contribution to the energy
budget.
Another distinctive feature in our setting is the presence of a sub-dominant trigger,
plotted in purple in Fig. 23.1. It is the physical agent that fixes the time of the phase
transition. Without committing to a particular microphysical implementation, we can
describe it in terms of a function .q(t, x). The decay time .τ∗ (x), or transition surface
equivalently, is then defined implicitly through

.q(t∗ (x), x) ≡ q∗ (= const) , (23.4)

where.q∗ is the threshold value at which the phase transition becomes efficient. Due to
the spatial variations in .q(t, x), also .τ∗ (x) is in general position dependent. To make
this more concrete, we will later discuss the case where .q(τ, x) is identified with
either a scalar field (Cold NEDE) or a temperature (Hot NEDE). Then the threshold
value .q∗ corresponds to a particular field value or temperature, respectively.

23.2.2 Perturbations

Except for the sharper transition and presence of the trigger field, the above model
shares many similarities with the original EDE model proposed in [1]. However,
more differences arise on the perturbation level. In short, NEDE density perturba-
436 F. Niedermann and M. S. Sloth

tions .δρNEDE (τ, x) = ρNEDE (τ, x) − ρ̄NEDE (τ ) are generated at the time of the phase
transition .τ∗ (for earlier times, .τ < τ∗ , .ρNEDE (τ, x) = const with vanishing adiabatic
perturbations). To be precise, their initial amplitude is determined by adiabatic per-
turbations in the trigger .δq(τ∗ , x) = q(τ∗ , x) − q̄(τ∗ ), where .q̄(τ ) is the trigger’s
background value. The picture is simple: Without trigger perturbations, the phase
transition would be an instantaneous event on large scales. As will become clearer
later, tiny differences in its timing could still arise on short length scales that char-
acterize the microphysics of the phase transition. For example, in a first-order tran-
sition, such a scale is the typical size of vacuum bubbles when they collide. Still, on
much larger cosmological scales the transition would occur at constant time .τ∗ . Once
we consider perturbations in the trigger—we know there always will be adiabatic
perturbations—this is no longer true and the transition time acquires a small spa-
tial dependence, i.e. .τ∗ → τ∗ (x) = τ̄∗ + δτ∗ (x), where.δτ∗ (x) = −δq(τ∗ , x)/q̄ ' (τ∗ ).
The last formula can be understood intuitively: If .q̄ ' (τ∗ ) > 0, at a fixed position .x,
the threshold value .q∗ is reached earlier if .δq(τ∗ , x) > 0 and later if .δq(τ∗ , x) < 0,
corresponding to .δτ∗ (x) < 0 and .δτ∗ (x) > 0, respectively. Now, these time varia-
tions imply that the decay of NEDE will start at slightly different times at different
positions in space, which, in turn, creates perturbations.δρNEDE (t∗ ). Simply speaking,
an earlier decay corresponds to an underdensity .δρNEDE (τ̄∗ , x) < 0 and a later decay
to an overdensity .δρNEDE (τ̄∗ , x) > 0. A more detailed matching calculation, based on
work by Deruelle and Mukhanov [31], that matches the cosmological perturbations
across the transition surface yields (in momentum space) [5]

δρNEDE (τ̄∗ , k)
. = 3H (τ̄∗ ) [1 + wNEDE (τ̄∗ )] δτ∗ (k) , (23.5a)
ρ̄NEDE (τ¯∗ )
.θNEDE (τ¯∗ ) = −k δτ∗ (k) .
2
(23.5b)

We defined the divergence of the NEDE fluid velocity .θ as .ik i (δTNEDE )i0 ≡ (1 +
wNEDE )ρ̄NEDE θNEDE and its density contrast as.δNEDE .≡.δρNEDE (τ̄∗ , k)ρ̄NEDE (τ¯∗ ). The
derivation is based on Israel’s matching equations that relate the extrinsic curvature
on both sides of the transition surface [32]. The expressions in (23.5) provide the
initial conditions for the subsequent evolution of the fluid perturbations, described
by the usual dynamical equations for an interacting fluid with vanishing viscosity
parameter. In synchronous gauge they are [33, 34]
( )
' h'
.δNEDE = −(1 + wNEDE ) θNEDE + − 3(cs2 − wNEDE )H δNEDE
2
θNEDE
− 9(1 + wNEDE )(cs2 − ca2 )H 2 , (23.6a)
k2
cs2 k 2
θ'
. NEDE = −(1 − 3cs2 )H θNEDE + δNEDE (23.6b)
1 + wNEDE

where .h is the trace of spatial metric perturbation, and .ca (τ ) and .cs (τ, k) are the
adiabatic and the rest-frame sound speed of the decaying NEDE fluid, respectively.
23 New Early Dark Energy as a Solution to the . H0 and . S8 Tensions 437

In summary, the full dynamical system is given by the background solution in (23.1)
and the perturbation equations in (23.6), subject to the boundary conditions (23.5).
The above description is very general and requires fixing. fNEDE [which determines
.ρ̄NEDE (t∗ )], .wNEDE (τ ), .cs (τ, k) and .q(τ∗ , k) [or .δτ∗ (k)] to close the dynamical sys-

tem. At this stage, it is clear that we need further input to reduce the amount of freedom
we have. After all, .wNEDE (τ ) alone gives infinite freedom for the background evolu-
tion. Therefore, in the next section, we propose a set of simple assumptions to close
the system.

23.2.3 A Simple Model for Extending .∧CDM

The aim of this section is to highlight the phenomenological prospects of a triggered

decay scenario. To that end, we close the dynamical system by imposing a set of
simple assumptions. We will present a microscopic scenario to motivate them later.
1. The trigger is given by an ultra-light scalar field .φ(τ, k) with mass .m and adia-
batic perturbations .δφ(τ, k), explicitly .q(τ, k) ≡ φ(τ, k). We then assume that
the decay of NEDE is triggered when .φ drops out of slow-roll, corresponding
to . H (τ∗ )/m ≈ 0.2. This then fixes the decay time .τ∗ as a function of .m. In
other words, the lighter the mass, the later the phase transition is triggered. A
microscopic realization in terms of a tunneling field .ψ that couples to .φ will
be presented in Sect. 23.3. In any event, this introduces the trigger’s initial field
value, .φini , as a new parameter. Moreover, if .φini is sufficiently large, the trig-
ger field makes a contribution to dark matter .Ωφ = ρ0,φ /(3Mpl2 H02 ) through its
coherent oscillations at late times. It can be related to the other model parameters
through [11]
( )( )2
1 + z∗ φini
Ωφ ≃ 0.4 ×
. (1 − f NEDE ) . (23.7)
5000 MPl

where .z ∗ = 1/a(τ∗ ) − 1 is the decay redshift. The trigger field can therefore
make a sizeable contribution to the energy budget provided its initial field value
is near (yet still below) the Planck scale.
2. The equation of state parameter after the phase transition is constant and, in
accordance with (23.2), lies in the range .1/3 ≤ wNEDE ≤ 1. This can be under-
∗
stood as keeping the first term in a Taylor expansion, .wNEDE (τ ) = wNEDE +
∗
dwNEDE
dτ
(τ − τ∗ ) + . . ., where .τ∗ is the decay time. This is a reasonable truncation
∗
dwNEDE
provided. H1 ∗ dτ ≪ 1, which ensures that.ρNEDE has become negligible before
cosmological observables become sensitive to the running of .wNEDE . Otherwise,
higher-order corrections need to be included.
3. The rest-frame sound speed equals the adiabatic sound speed, explicitly
438 F. Niedermann and M. S. Sloth

'
p̄NEDE
c2 (τ, k) = ca2 (τ ) =
. s
'
ρ̄NEDE
1 dwNEDE
= wNEDE −
3(1 + wNEDE )H dτ
∗
≈ wNEDE . (23.8)

In other words, this assumes that the NEDE fluid does not produce sizeable
entropy perturbations on cosmological scales. We will formulate a corresponding
necessary condition on the microscopic model in the next section.
These three assumptions are enough to reduce the NEDE framework to a 4-
parameter extension, supplementing the six .ɅCDM parameters with . f NEDE , .m,
∗
.wNEDE and .φini , where .m can be traded for the decay redshift . z ∗ and .φini for
.Ωφ = ρ0,φ /(3Mpl H0 ), quantifying the trigger contribution to the energy budget
2 2

today. This model has been implemented in the state-of-the-art Boltzmann code
.TriggerCLASS, which extends the Cosmic Linear Anisotropy Solving System
(.CLASS [37]). It can be fitted to CMB (Planck 2018 temperature, polarization,
and lensing [38]), baryonic acoustic oscillations (6dFGS [39], MGS [40], and
BOSS [41]), and supernovae data (Pantheon [42]). The result of a recent analysis
performed in [11] is shown in Fig. 23.2. The blue contour depicts the marginalized
.ɅCDM constraints, which is discrepant with the distance ladder result of. H0 obtained
by the SH.0 ES team (grey vertical band [35]) and at a milder level also with the value
of . S8 inferred from weak lensing measurements of KiDS, VIKING, and DES (col-
ored vertical band [36]). Specifically, for .ɅCDM the Gaussian tensions are .4.8σ for
. H0 and .2.3σ for . S8 .

The main observation is then that within NEDE (red contour), both the . H0 and . S8
tension are reduced below.2σ . This makes a combined analysis possible that includes
the weak lensing and the distance ladder constraint implemented as a Gaussian prior
+0.032
on . S8 and . H0 (dotted contour). In that case, we obtain . f NEDE = 0.134−0.025 (.68%
C.L.), which corresponds to a Gaussian evidence for a non-vanishing fraction of
NEDE that exceeds .5σ . We also report the new concordance values . H0 = 71.71 ±
0.88 km sec−1 Mpc−1 (.68% C.L.) and. S8 = 0.793 ± 0.018 (.68% C.L.). A selection of
cosmological parameters for the four data set combinations is presented in Table 23.1.
The NEDE phenomenology can be best understood in terms of different parameter
correlations depicted in Fig. 23.3:

• There is the usual anticorrelation between . f NEDE and the sound horizon .rs typical
for EDE-type models caused by the energy injection in the primordial fluid (see the
NEDE feature in Fig. 23.1 and first panel in Fig. 23.3) [1]. The angular scale of .rs
(as measured by BAO and CMB observations at high precision) is held constant by
raising . H0 . This is the core mechanism responsible for addressing the . H0 tension
(second panel).
• There is a positive correlation between . f NEDE and the amount of cold dark matter
.wcdm ≡ Ωc h (third panel). It is generically expected for models that reduce the
2

sound horizon [45, 46]. In terms of the CMB power spectrum, this increase is
23 New Early Dark Energy as a Solution to the . H0 and . S8 Tensions 439

Fig. 23.2 NEDE addresses the . H0 and . S8 tension simultaneously. The red and blue contours
were obtained for the baseline (b.l.) datasets (Planck, BAO, uncalibrated supernovae). The vertical
(gray) and horizontal (magenta) shaded region corresponds to the measured value . H0 = 73.04 ±
+0.025
1.04 kms−1 Mpc−1 (S. H0 ES [35]) and . S8 = 0.762−0.024 (KiDS+VIKING+DES [36]), respectively.
The dotted contour was obtained by including these values as Gaussian priors in the MCMC analysis.
Here and henceforth, the light and dark-shaded regions correspond to the .68% and .95% C.L.,
respectively

Fig. 23.3 The NEDE model introduces approximate degeneracies between. f NEDE and (.rs (at matter
drag), . H0 , .ωcdm , .n s ). The gray band represents the . H8 constraint from S. H0 ES
440 F. Niedermann and M. S. Sloth

Table 23.1 Constraints on all .(6 + 4) NEDE parameters (means with .1σ error bars and best-fit
in parenthesis) obtained from an analysis with the baseline datasets (CMB, BAO, uncalibrated
supernovae) and one supplemented by weak lensing and distance ladder priors on . S8 and . H0 ,
respectively. The values are cited from [11], and . S8 has been added as a derived parameter. The
last three rows quantify the . H0 and . S8 tension and state the overall .χ 2 improvement compared to
.ɅCDM

Baseline Baseline + . H0 + . S8
+0.0320
. f NEDE .< 0.130 .(0.1140) .0.1340−0.0250 .(0.1460)
+1000 +500
.z decay .4911−2000 .(4441) .4414−800 .(4626)
∗ +0.12
.3ωNEDE .> 1.41 .(2.09) .2.05−0.18 .(2.03)
+0.0025
.Ωφ .< 0.0068 .(0.0010) .0.0057−0.0029 .(0.0060)
+0.0002 +0.0002
.Ωb h .0.0227−0.0003 .(0.0230) .0.0230−0.0002 .(0.0230)
2
+0.0023 +0.0030
.Ωc h2 .0.1232−0.0044 .(0.1290) .0.1284−0.0030 .(0.1300)
+0.0160 +0.0160
s) .3.0570−0.0160 .(3.0670) .3.0700−0.0160 .(3.0690)
.log(10
10 A
+0.0060 +0.0058
.n s .0.9739−0.0090 .(0.9840) .0.9868−0.0058 .(0.9880)
+0.0074 +0.0070
.τreio .0.0572−0.0074 .(0.0550) .0.0594−0.0080 .(0.0560)
+0.78 +0.88
. H0 [km/s/Mpc] .69.06−1.40 .(70.84) .71.71−0.88 .(71.74)
+0.0180 +0.0180
. S8 .0.8160−0.0150 .(0.8330) .0.7930−0.0180 .(0.7990)

. H0 tension.a 3.05.σ (1.4.σ ) –

. S8 tension 1.9.σ –
.∆χ .−3.7 .−20.1
2

a
. The Gaussian tension measure is unreliable in the case of. H0 due to non-Gaussian posteriors caused
by prior volume effects [5, 24]. A way out is offered by the DMAP criterion (in parenthesis), which
is prior-independent [43]. This was also supported by the recent maximum likelihood analysis
in [27] (see also [44] for EDE)

compensated by the presence of dark sector acoustic oscillations, whose positive

pressure leads to an excess decay of the gravitational potential [5, 47].
• Lowering .rs generically leads to enhanced diffusion damping on small scales [5].
This is balanced by increasing the spectral tilt .n s and amplitude . As . As argued
in [48], the correlation between . f NEDE and .n s (fourth panel) has far-reaching
implications for inflationary model-building. Specifically, it makes simple models
of inflation such as the curvaton [49–51], or perhaps power-law inflation, viable
again (see f.ex. [48] for a more detailed discussion of this point).
• The presence of the ultralight trigger field allows for a sub-percent fraction of non-
thermal dark matter.Ωφ . It suppresses the matter power spectrum on scales relevant
for determining the . S8 parameter due to acoustic oscillations in its perturbations.
The anti-correlation between .Ωφ and . S8 is then responsible for resolving the . S8
tension (see Fig. 23.4 and [11] for more details).
23 New Early Dark Energy as a Solution to the . H0 and . S8 Tensions 441

Fig. 23.4 The trigger field acts like axion dark matter with a mass.m ∼ 10−27 eV and suppresses the
matter power spectrum on scales relevant for the determination of . S8 due to the Jeans stabilization
of its perturbations. The magenta band represents the . S8 constraint from weak lensing

23.3 Cold New Early Dark Energy

The model of Cold New Early Dark Energy (Cold NEDE) relies on a two-field
mechanism to explain the sudden transition in the equation of state proposed in (23.2).
Here, the trigger field.φ, which we introduced before as defining the transition surface
.q(τ, k), initiates the vacuum tunneling of a scalar field .ψ when it drops out of slow-
roll. The tunneling corresponds to a first-order phase transition that converts vacuum
energy into a new form of energy that decays with redshift (for another first-order
model, see [52]). A schematic illustration is provided in Fig. 23.5. A similar model
has been proposed in the early 90’s at higher energies as a mechanism to exit from
inflation in a first-order phase transition and yet avoid typical problems with old
inflation [53–56]. In any event, this idea can be realized in terms of the renormalizable
two-field potential
442 F. Niedermann and M. S. Sloth

Fig. 23.5 Illustration of the Cold NEDE potential (not-to-scale). The field starts at the orange dot
but is initially frozen due to the Hubble friction. It is separated from the true minimum (white dot)
through a high potential barrier. Shortly before matter-radiation equality, when.m ∼ H ∼ 10−27 eV,
the field starts rolling along the orange line. As a result, the barrier between the field and the true
minimum shrinks, and sub-barrier tunneling becomes admissible after the field passes the blue dot.
The tunneling rate increases quickly, leading to the percolation of space with the true vacuum. In
its final state, the field oscillates around the true minimum (red arrow)

λ 4 β 2 2 1 1 1
. V (ψ, φ) = ψ + M ψ − α Mψ 3 + m 2 φ 2 + λ̃φ 2 ψ 2 . . . , (23.9)
4 2 3 2 2

where.λ,.α, and.λ̃ are dimensionless constants, and the ellipses denote higher terms in
an effective field theory expansion that are suppressed at low energies. This is a two-
scale potential with . M ∼ eV being of the order of the critical energy density at the
same time of matter-radiation equality and .m ∼ M 2 /Mpl ∼ 10−27 eV an ultralight-
scale set by the value of the Hubble parameter at the same time. This could have a
UV completion in a model with two axions, one with a mass close to the QCD axion
mass at the eV scale and an Ultra-Light Axion (ULA) with a mass of .10−27 eV [11].
For .λ < 2α 2 /(9β), this potential admits a false minimum at .(ψ, φ)False = (0, 0) and
a true one at
( ] )
M [ √
.(ψ, φ)True = α + α − 4λβ , 0 .
2 (23.10)
2λ

Initially the field is frozen at .(ψ, φ)False = (0, φini ) due to a large Hubble friction.
In particular, it is blocked from tunneling to the true minimum (23.10) by a huge
potential barrier created by the term .λ̃φ 2 ψ 2 in (23.9) (which requires .φini to be
23 New Early Dark Energy as a Solution to the . H0 and . S8 Tensions 443

sufficiently large). Eventually, when . H ∼ m during the CMB epoch close to matter-
radiation equality, .φ starts rolling towards zero. This gradually removes the barrier
and increases the tunneling probability per spacetime volume denoted as .┌. In our
case, it can be approximated as [57]

┌ ∼ M 4 e−SE
. (23.11)

where

4 π2 ( )
S ≈
. E (2 − δeff )−3 α1 δeff + α2 δeff
2
+ α3 δeff
3
, (23.12)
3λ
is the Euclidian action corresponding to the so-called “bounce solution” [58, 59] with
numerically determined coefficients [60] .α1 = 13.832, .α2 = −10.819, and .α3 =
2.0765. Its time dependence is captured through
( )
λ φ 2 (τ )
δ (τ ) = 9
. eff β + λ̃ . (23.13)
α2 M2

Tunneling becomes efficient when.┌ outperforms the Hubble expansion, correspond-

ing to .┌ ≳ H 4 . For . M ∼ eV and . H ∼ m ∼ 10−27 eV, this happens when . S E ≲ 250
(this condition is always met as .φ → 0, provided .9β/α 2 ≲ 0.1 and .λ < 1), which
in turn fixes the decay time .τ∗ (or .z ∗ alternatively). A more quantitative argument
finds [5] . H (τ∗ )/m ≈ 0.2, which will be used as a reference value below. Moreover,
the fact that .┌ is exponentially sensitive to the value of .φ (through .δeff ) justifies that
the phase transition occurs across a surface of constant .φ.
For .τ > τ∗ , the space is quickly filled with bubbles of the true vacuum. These
are spherical configurations of the scalar field .ψ with .ψ ≈ ψTrue in their interior and
.ψ ≈ ψFalse in their exterior. Both phases are separated by a bubble wall across which
the field interpolates between both vacua. The percolation time scale is given by the
inverse of .β̄ = −dS E /dt, where .t denotes cosmological time (.dt = adτ ). After a
time .∆t ∼ β̄ −1 , the tunneling rate has increased so much that all of space has been
converted to the true vacuum. Since the vacuum bubbles expand with almost the
speed of light, .β̄ −1 also sets the typical size of bubbles at the time when they collide.
These bubbles introduce large anisotropies in the cosmic fluid (corresponding to a
density contrast of the order of . f NEDE ). Requiring that we cannot resolve them in our
cosmological probes, which have an angular resolution .θmin , sets an upper bound on
−1
.β̄ [5]:
( )( ( )
−1 θmin z ∗ ) ( g∗ )1/2 0.7
. H∗ β̄ < 0.3 × (1 − f NEDE )−1/2 , (23.14)
0.14◦ 5000 3.4 h

where .g∗ is the number of relativistic degrees of freedom at the time of the transition
and . H∗ = H (τ∗ ). We normalized the expression with respect to a typical CMB
resolution of .θmin ∼ 2π/2500 ≃ 0.14◦ . By relating .β̄ to the microscopic parameters,
444 F. Niedermann and M. S. Sloth

it can be shown that .λ can be used as a dial to suppress . H∗ β̄ −1 . A sufficient condition

to satisfy (23.14) is

( )( ( )/ ∗ ┌ ]
θmin z ∗ ) ( g∗ )1/2 0.7 δeff − δ d(λS E )
λ < 2.1 ×
. ,
0.14◦ 5000 3.4 h 1 − f NEDE dδeff
(23.15)

∗
where .δeff = δeff (τ∗ ) and .δ = δeff |φ=0 = 9λβ/α 2 . This is also a necessary condition
to justify our effective fluid description introduced in Sect. 23.2.3, which relies on the
assumption that the scalar field condensate after the phase transition is homogenous
and isotropic on cosmological scales. Moreover, this is a necessary condition for the
absence of entropy perturbations in the large-scale fluid, which would be created if
the phase transition were too slow and structures emerged at cosmological scales.
The liberated vacuum energy after the phase transition is given as the difference
in potential energy between the true and false vacuum, .ρNEDE (τ∗ ) = V (0, φ(τ∗ )) −
V (ψTrue , 0), amounting to

ρNEDE (τ∗ ) cδ α 4 M 4
f
. NEDE = = , (23.16)
ρtot (τ∗ ) 36 λ3 MPl
2
H∗2

where we introduced the function

1 ( √ )2 ( √ )
c =
. δ 3 + 9 − 4δ 3 − 2δ + 9 − 4δ , (23.17)
216
which is of order unity for .δ < 1. We finally provide a dictionary between the micro-
scopic and the phenomenological parameters (. f NEDE , .z ∗ and .Ωφ ):
( )( )
1 f NEDE /(1 − f NEDE ) ( g∗ ) ( z ∗ )4
λ3 α −4
. M 4 ≃ (0.4 eV)4 ,
cδ 0.01 0.1 3.4 5000
(23.18a)
( g )1/2 ( z )2 ( 0.2 )
−27 ∗ ∗
.m ≃ 1.7 × 10 eV (1 − f NEDE )−1/2 ,
3.4 5000 H∗ /m
(23.18b)
( )2 ( )−1 ( )
φini 1 + z∗ Ωφ
. ≃ 0.13 × (1 − f NEDE )−1 . (23.18c)
MPl 5000 0.005

Here, (23.18a) and (23.18c) follow directly from (23.16) and (23.7), respectively, and
we assumed radiation domination to obtain (23.18b). Different terms are normalized
with respect to their phenomenologically relevant values (cf. Table 23.1). The upshot
from this is that Cold NEDE requires an ultralight trigger field and an .eV-scale
tunneling field. Moreover, the initial field value for the trigger field .φini remains sub-
23 New Early Dark Energy as a Solution to the . H0 and . S8 Tensions 445

Planckian even if we require that it makes a contribution to dark matter large enough
to resolve the . S8 tension.
Relating .wNEDE to the parameters in (23.9) is more complicated, and more work is
needed to achieve this. In general terms, however, there are at least two contributions
to .ρNEDE that have the potential to explain the observational fact that .wNEDE > 1/3.
First, the small-scale anisotropic stress, which corresponds to the colliding bubble
walls, sources tensor, vector, and scalar shear. While tensor shear is equivalent to
gravitational radiation and hence associated with.w = 1/3, the vector and scalar shear
are known to behave like a stiff fluid; in particular, their energy density averaged over
large scales decays as .1/a 6 [61]. This possibility requires the gravitational sourcing
to be very efficient, which corresponds to the regime where . H∗ β̄ −1 is close to its
upper bound in (23.14). Second, the scalar field condensate after the decay, despite
being highly fragmented, oscillates in an effective field theory potential that contains
higher order terms such as .∝ ψ 6 /M 2 or .∝ ψ 8 /M 4 [indicated by the ellipses in
(23.9)]. If these operators are probed within the perturbative regime, the expectation
is that this will increase .wNEDE initially (similar to the fact that coherent oscillations
in higher order potentials lead to a stiffer fluid when averaged over cycles).
An entirely different possibility has been prosed in [6] as “scenario B”. Here the
idea is that the vacuum bubbles during the collision phase quickly dissipate into
relativistic light particles that become non-relativistic shortly after (before recombi-
nation). The net effect is a reduction of the dark matter density at early times which
is then replenished at later times. In particular, at background level, this was shown
to reproduce the shape of the energy injection in Fig. 23.1 without relying on a stiff
fluid component.
Another question regards the ultraviolet completion of (23.9). A successful phe-
nomenology requires a large separation of scales where .m ≪ M. For this choice to
be radiative stable at low energies, the coupling .λ̃ has to be very weak, approxi-
mately [5] .λ̃ ≤ O(1) × 103 m 2 /(β M 2 ). This raises the question of how this can be
achieved in a more fundamental theory. As mentioned earlier, a preliminary answer
was given in [11] recently. The idea is to explain the smallness of .m and .λ̃ through the
breaking of a continuous shift symmetry, .φ → φ + const, down to a discrete sym-
metry, .φ → φ + 2π f , which occurs at an energy scale .Ʌ ≪ f . To be specific, in
this axion-like setup, we find .m = Ʌ2 / f , which for .Ʌ ∼ eV and . f ∼ MPl recovers
the right value of the trigger mass.

23.4 Hot New Early Dark Energy

Since the cosmological electroweak and QCD phase transitions are both known to be
triggered by the temperature of the primordial plasma dropping below some critical
temperature, it is natural to consider the possibility that a temperature triggers the
NEDE phase transition. In that case, when the ultralight scalar field trigger is replaced
by a dark sector temperature .Td , there is only one energy scale, the eV energy scale,
controlling the phase transition.
446 F. Niedermann and M. S. Sloth

This scenario has been introduced as Hot NEDE in [6, 7]. Correspondingly, in
(23.4), we identify .q(τ, x) ≡ Td (τ, x). Moreover, the dark sector is assumed to be
1/4
colder than the visible sector, .Tvis > Td . In fact, we require .ξ ≡ Td /Tvis < 0.6/grel,d ,
where.grel,d is the effective number of relativistic degrees of freedom in the dark sector.
This avoids bounds on the equivalent number of neutrino species4 , which receives
a contribution [64] .∆Neff = 47 ( 114
)4/3 grel,d ξ 4 < 0.29. As we will see, it also makes
sure that for . f NEDE ∼ 0.1, the corresponding dark radiation fluid is subdominant
compared to the released vacuum energy measured by .ρNEDE , implying a strong
first-order phase transition in agreement with the sharp energy injection in Fig. 23.1.
On a microphysical level, the idea is that a complex NEDE field.ψ is charged under
a local .U(1)D with gauge coupling .gNEDE and self-coupling .λ. The corresponding
gauge bosons form a plasma of temperature .Td and the NEDE field receives thermal
corrections of the form (valid in the perturbative regime where [6, 65] .λ ≲ gNEDE 3
)

.δV (ψ; Td ) = 3Td4 K (gNEDE ψ/Td ) e−gNEDE ψ/Td , (23.19)

√
where .ψ = 2|ψ| and . K (a) can be approximated within the range .0 < a ≡
gNEDE ψ/Td < 30 as

. K (a) = −0.1134 (1 + a) − 0.0113 a 2

+ 4.32 × 10−6 ln (a) a 3.58 + 0.0038 e−a(a−1) . (23.20)

For small argument, we can expand . K (a)e−a ≃ −π 2 /90 + a 2 /24 + . . . and recover
the usual result in the literature valid for gauge boson masses .gd ψ ≪ Td [66]. The
full potential then reads

1 2 λ
. V (ψ; Td ) = − gNEDE T◦2 ψ 2 + ψ 4 + δV (ψ; Td ) + V1−loop + . . . , (23.21)
8 4
where the first two terms correspond to the vacuum potential, and.T◦ is a characteristic
temperature scale, which is related to the mass scale .μ through .T◦ = 2μ/gNEDE . The
term .V1−loop denotes 1-loop corrections to the vacuum potential [67, 68]. While they
can introduce changes to the precise shape of the potential, we do not expect them
to change the conclusion about the character and strength of the phase transition (for
a more detailed discussion see [69]). The tachyonic character of the quadratic term
reflects the fact that the symmetric state with .ψ = 0 is unstable at zero temperature.
To describe the temperature dependence of .V (ψ; Td ), it is useful to introduce the
parameter
( )
T◦2
.δ̃eff (Td ) = π 1− 2 . (23.22)
Td

4To be specific, big bang nucleosynthesis and the CMB give rise to the bounds .∆Neff < 0.39 [62]
and .∆Neff < 0.3 [63], respectively.
23 New Early Dark Energy as a Solution to the . H0 and . S8 Tensions 447

Fig. 23.6 The NEDE scalar field potential .Ṽ = 4π V /Td4 as a function of .Td and .ψ̃ =
√
gNEDE γ ψ/Td for .γ = 0.02. Initially, .Td > Tc , and the symmetric vacuum at .ψ̃ = 0 minimizes
the potential energy. As the dark sector cools further, the potential develops a second minimum at
.ψ̃ > 0. It becomes degenerate with the symmetric vacuum at . Td = Tc . At this point, the tunneling
rate .┌(Td ) becomes non-vanishing. Percolation becomes possible later on when .┌(Td ) ≈ H 4 is
met, which implicitly defines the temperature .Td∗ at which the NEDE phase transition occurs

We depict different stages of the potential’s evolution as a function of .Td in

Fig. 23.6. Initially, for .Td ≫ T0 (or .δ̃eff ≈ π equivalently), we are in the symmetric
phase where the global minimum lies at .ψ = 0. However, as the temperature cools,
the potential develops a second minimum. Both minima become degenerate at a
critical temperature .Tc . Finally, at an even lower temperature .Td∗ < Tc , the phase
transition occurs. We are mostly interested in the supercooled regime for which.Td∗ ≪
Tc . In that case, we have a strong first-order phase transition where the difference in
vacuum energy dominates over the energy density of the radiation fluid, a necessary
requirement to have a localized energy injection. It can be shown that this regime is
realized for [6] [requiring the general form of the thermal corrections in (23.19)]

4π λ
γ ≡
.
4
≲ 1. (23.23)
gNEDE
√
As before, we define .ρNEDE (τ∗ ) = V (0) − V (ψTrue ), where .ψTrue ≃ μ/ λ ≡ vψ .
The maximal fraction of NEDE then evaluates to
448 F. Niedermann and M. S. Sloth
( )2
π δ̃ ∗ Td∗4
f
. NEDE = 1 − eff , (23.24)
16γ π ρtot (τ∗ )

∗
where .δeff ≡ δeff (τ∗ ). In accordance with Fig. 23.6, and as explicitly shown in [6],
∗
strong supercooling corresponds to the regime where.δ̃eff /π ≪ 1 (or.Td∗ ≪ Tc equiv-
∗4
alently). Together with .ρtot (τ∗ ) ≈ 1.1 × Tvis /(1 − f NEDE ), this implies
┌ ]
f NEDE /(1 − f NEDE )
ξ 4 ≡ (Td /Tvis )4 |τ =τ∗ ≃ 0.6 × γ
. ∗ . (23.25)
0.1

We note that for .γ ≪ 1, . f NEDE can be sizeable, although we assume .ξ∗ <
1/4
0.6/grel,d . As mentioned before, this regime avoids bounds on .∆Neff and ensures
that the dark radiation energy density, .∝ Td4 , is smaller than .ρNEDE . In other words,
like in Cold NEDE, the trigger remains sub-dominant. In absolute terms, we derive
the dark sector temperature to be
┌ ]┌ ]4
∗4 f NEDE /(1 − f NEDE ) 1 + z∗
. Td ≃ (0.7eV) γ4
. (23.26)
0.1 5000

This then constitutes a central relation between the phenomenological parameters,

f
. NEDE and .z ∗ , and the more fundamental parameters, .Td∗ and .γ . As with Cold NEDE,
the phase transition is required to complete quickly on cosmological time scales. The
percolation time sale .β̄ −1 was derived to be [6]

. H∗ β̄ −1 ∼ 10−2 gNEDE
2
(valid if γ ≲ 1 and gNEDE ≲ 0.1) , (23.27)

which for a sufficiently weak gauge coupling can easily satisfy the phenomenological
bound in (23.14).
To summarize, Hot NEDE can provide us with a sizeable fraction of (false) vac-
uum energy that, in the supercooled regime, dominates over the dark radiation com-
ponent. While the former provides us with a sizeable amount of (decaying) early
dark energy, the latter plays the role of a subdominant trigger component in agree-
ment with Fig. 23.1. In particular, we can work in a regime where .ξ = Td /Tvis ≪ 1.
This property sets Hot NEDE apart from dark radiation approaches to the Hubble
tension—see, for example, the models that introduce a mass-threshold [70–74]—
which require .ξ to be of order unity in order to be able to address the . H0 tension.
This typically leads to challenges with big bang nucleosynthesis (BBN), which gives
rise to an upper bound on [64] .∆Neff (and hence .ξ ) or requires a post-BBN ther-
malization of the dark sector, which comes with its own model-building challenges
[75, 76].
Phenomenological differences with Cold NEDE will arise due to the fact that the
NEDE perturbations in (23.5) are now seeded by dark sector temperature pertur-
bations, .δτ∗ = −δTd (τ∗ , x)/Td' (τ∗ ), as opposed to perturbations of a slowly rolling
23 New Early Dark Energy as a Solution to the . H0 and . S8 Tensions 449

scalar field in the Cold NEDE case. A detailed phenomenological study is planned
for future work.

23.4.1 Neutrino Mass Generation

It has been pointed out long ago in the context of the inverse seesaw mechanism [77–
80] that the active neutrino masses can be created through a mass mixing with a
set of sterile neutrinos .νs and right-handed neutrinos .ν R . Besides two high-energy
Dirac mixings .νs ↔ ν R and .ν R ↔ νe,μτ , the crucial ingredient is a Majorana mass
term .∝ m s (νs )c (νs ) , where .m s is of order (or larger) than .O(eV). The mechanism
is theoretically compelling because the active neutrino masses are protected against
radiative corrections through a Lepton symmetry that is recovered in the limit of
vanishing neutrino mass. This is similar to the electron mass .m e , which is radiatively
stable due to a chiral symmetry that is recovered in the limit.m e → 0 [81]. However, in
contrast to the electron, which receives its mass in the electroweak phase transition at
a temperature.Tvis ∼ 160 GeV [82] through the Higgs mechanism, the inverse seesaw
mechanism does not provide a dynamical explanation for how the Majorana mass
term is created. It is, therefore, natural to ask if the sterile mass .m s was generated
during the NEDE phase transition when .Td ∼ eV. In other words, we can ask if there
was as dark sector analog of the electroweak phase transition.
To make these statements more concrete, we first review the inverse seesaw mech-
anism in its original form (following [80]). For simplicity, we only consider one
neutrino generation. Besides one sterile neutrino .νs , we introduce the Dirac spinor
of a right-handed neutrino .ν R , and denote the left-handed neutrino as .ν L (this could
be .νe , .νμ , or .ντ ). Defining . N ≡ (ν L , ν Rc , νs )T , we can introduce the mass mixing

1
.Lν = − N T C M N + h.c. (23.28)
2
where
⎛ ⎞
0d 0
. M = ⎝d 0 n ⎠ (23.29)
0 n ms

is a (complex-valued) mass matrix, and .C denotes charge conjugation. Using

this notation, off-diagonal and diagonal elements correspond to Dirac and Majo-
rana mass terms, respectively. For concreteness, we assume that .m s = O(100eV),
.d = O(100GeV) and .n = O(10TeV) (other choices are possible too). The crucial
observation is that this mass matrix, upon diagonalization, admits a light eigenmode
with mass
450 F. Niedermann and M. S. Sloth

|d|2
m light =
. |m s | + O(|m s |2 ) , (23.30)
|d|2 + |n|2

which, for the numerical example provided above, amounts to .m light = O(0.05eV)
and thus can play the role of one of the active neutrinos. In addition, there is a
pair of super-TeV pseudo-Dirac fermions, which are irrelevant for the low-energy
dynamics. As described in [80], this mechanism can reproduce the observed mixing
angles and mass spectrum when generalized to three generations. Moreover, only
the Majorana mass term .∝ m s breaks Lepton number, and thus the symmetry gets
restored for .m s → 0. Consequently, .m s and, due to (23.30) also .m light , can be small
in a technically natural way.
Now, the idea is to introduce an interaction between the sterile neutrino and the
NEDE scalar .ψ with Yukawa coupling .gs ,

1 ∑
L ⊃ −√
. gs ψ(νs )c (νs ) + h.c. , (23.31)
2 ij

which is invariant under a global Lepton symmetry .U(1) L if we assign a charge

L = −2 to .ψ and . L = 1 to .νs . In addition, we assume appropriate charges √
. under
our local .U(1)D to preserve gauge invariance. In any event, when .ψ → vψ / 2 in
the NEDE phase transition, we indeed generate the Majorana mass entry in (23.29);
more explicitly, we have .m s = gs vψ . In particular, the vacuum expectation value .vψ
can be related to the phenomenological NEDE parameters through
┌ ]1/4 ┌ ]
gs f NEDE /(1 − f NEDE ) 1 + z∗
m s ≈ (1.2 eV) ×
. × . (23.32)
(4π λ)1/4 0.1 5000

As a result, we find that we can create a (super-) eV-scale mass entry .m s as required
by the inverse seesaw mechanism, provided the NEDE self-coupling is sufficiently
small. To be precise, we need .4π λ < gs4 . Due to the mixing between the active
neutrino .ν L and .νs of order of .O(d 2 /n 2 ), there is a bound .gs < 10−7 O(n 2 /d 2 ) [83],
which makes sure that the Boltzmann evolution of the active neutrinos is preserved. It
translates to .λ < 10−20 (10−2 eV/m light )4 . As a consistency check, such a small value
of .λ is consistent with the supercooling bound in (23.23), if the gauge coupling is not
to small (.gNEDE > 10−5 ). There is one concern, though, with√ .λ being that small. It
can lead to a large vacuum expectation value because .vψ = μ/ λ and hence a heavy
gauge boson mass .∝ gNEDE vψ , suppressing the gauge boson number density in the
broken phase. However, for the thermal description to be applicable, the gauge field
and .ψ must be in thermal equilibrium. As argued in [7], this translates into a lower
bound .ξ∗ > 0.3, ensuring the dark sector is populated strongly enough to maintain
thermal equilibrium. This, in turn, makes the neutrino scenario more constrained,
1/4
leading to the finite range .0.3 < ξ∗ < 0.6/grel,d .
Before ending this section, we note that this mechanism only explains the eV-scale
entry in (23.29). A more complete mechanism was suggested in [6, 7]. It creates the
23 New Early Dark Energy as a Solution to the . H0 and . S8 Tensions 451

d entry in the electroweak phase transition and postulates a new, super-TeV phase
.

transition to generate the .n(> d) term5 . The latter phase transition takes place again
in the dark sector and corresponds to a breaking of a .SU(2) D × U(1)Y D → U(1)D . If
the higher breaking scale is sufficiently large, the same phase transition could give
mass to a super-TeV dark matter, which, due to its gravitational interactions, could be
produced at early times through the freeze-in mechanism [85, 86]. Overall, the Hot
NEDE proposal should be understood as an attempt to develop a more fundamental
and comprehensive dark sector model guided by the aim of resolving the cosmic
tension.

23.5 Conclusion

New Early Dark Energy is a framework for a triggered low-scale phase transition in
the dark sector. As discussed in Sect. 23.2, it admits a very general phenomenological
description in terms of an abruptly decaying early dark energy component. The
corresponding sharp energy injection into the cosmic fluid can resolve the Hubble
tension, making use of the well-established degeneracy between . H0 and the sound
horizon .rs (see Fig. 23.3). Its unique feature is a trigger sector. It sets off the phase
transition and seeds a new type of dark sector acoustic oscillations carried by the
decaying NEDE fluid. The trigger leads to unique signatures in the CMB and matter
power spectra, which are different from other EDE models and is crucial for the
model’s phenomenological success in reducing the . H0 and . S8 tension below the two-
sigma level. When identified with an ultralight scalar field, it was recently shown
that the trigger could resolve the . S8 tension when it makes a small contribution to
dark matter through its coherent oscillations at late times [11]. When testing the
model against CMB, BAO, SNe data and including a local prior on the value of . S8
and . H0 (which has become statistically viable within the NEDE framework), we
+0.032
obtain . f NEDE = 0.134−0.025 , which corresponds to a Gaussian evidence for a non-
vanishing fraction of NEDE that exceeds .5σ along with the new concordance values
. H0 = 71.71 ± 0.88 km sec
−1
Mpc−1 and . S8 = 0.793 ± 0.018.
As mentioned before, NEDE admits different microscopical descriptions with
their own unique signatures. Here we have presented two possibilities. One where
the trigger is an ultralight scalar field (Cold NEDE) and one where it is identified
with a dark sector temperature (Hot NEDE)—each of which comes with its own
opportunities and challenges. In Cold NEDE, the primary theoretical tasks involve
linking the equation of state parameter .wNEDE to the effective field theory parameters
in the potential and explaining the hierarchy between the trigger scale .m and the
tunneling scale . M ≫ m. While we believe the latter can be achieved within a multi-

5As mentioned above, this is a generic choice of energy scales, which avoids any additional tuning
of the .d entry relative to the electroweak scale. It is, however, possible that both the .d and the .n
energy scales are lower, bringing the new UV dark phase transition, generating the .n scale, within
an energy scale relevant for a possible interpretation of the recent NANOgrav observation [69, 84].
452 F. Niedermann and M. S. Sloth

axion framework, the former might require more detailed numerical studies. On a
phenomenological level, new CMB data, such as future ACT releases [87] and the
CMB-S4 experiment [88], as well as additional LSS data from observations of galaxy
abundances and clustering or spectroscopic galaxy surveys, such as the James Web
Telescope [89] or Euclid [90], will help discriminate the model from its competitors
and constrain its unique signatures (for first studies including ground-based CMB
and full-shape matter power spectrum data see [27, 28, 48, 91]).
For Hot NEDE, a full cosmological data extraction is still outstanding. Here, we
expect changes arising from the difference in trigger perturbations. Such an analysis
is important because it can reveal how predictive the physics of the trigger sector is.
Moreover, in the case where we include the neutrino mass generation, we could look
for additional signatures related to the physics of the inverse seesaw mechanism (for
a detailed list see [7]).
Although the preferred microscopic scenario remains uncertain at the moment—
in fact, other triggered scenarios different from cold and Hot NEDE, such as hybrid
NEDE, are possible [5]—the crucial aspect is that, due to the low-energy scale of
the transition, any scenario will be highly constrained by our cosmological probes.
So, NEDE is a promising and natural framework for resolving the Hubble tension,
and if the Hubble tension really is the signature of a triggered phase transition in the
dark sector, we will soon know.

Acknowledgements M. S. Sloth is supported by Independent Research Fund Denmark grant 0135-

00378B. The work of F. Niedermann was supported by VR Starting Grant 2022-03160 of the
Swedish Research Council.

References

1. V. Poulin, T.L. Smith, T. Karwal, M. Kamionkowski, Early dark energy can resolve the Hubble
tension. Phys. Rev. Lett. 122, 221301 (2019). https://doi.org/10.1103/PhysRevLett.122.221301
[arXiv:1811.04083]
2. V. Poulin, T.L. Smith, T. Karwal, The ups and downs of early dark energy solutions to the
Hubble tension: a review of models, hints and constraints circa 2023. arXiv:2302.09032
3. M. Kamionkowski, A.G. Riess, The Hubble tension and early dark energy. arXiv:2211.04492
4. F. Niedermann, M.S. Sloth, New early dark energy. Phys. Rev. D 103, L041303 (2021). https://
doi.org/10.1103/PhysRevD.103.L041303 [arXiv:1910.10739]
5. F. Niedermann, M.S. Sloth, Resolving the Hubble tension with new early dark
energy. Phys. Rev. D 102, 063527 (2020). https://doi.org/10.1103/PhysRevD.102.063527
[arXiv:2006.06686]
6. F. Niedermann, M.S. Sloth, Hot new early dark energy. Phys. Rev. D 105, 063509 (2022).
https://doi.org/10.1103/PhysRevD.105.063509 [arXiv:2112.00770]
7. F. Niedermann, M.S. Sloth, Hot new early dark energy: towards a unified dark sector of neutri-
nos, dark energy and dark matter. Phys. Lett. B 835, 137555 (2022). https://doi.org/10.1016/j.
physletb.2022.137555 [arXiv:2112.00759]
8. A.H. Guth, The inflationary universe: a possible solution to the horizon and flatness problems.
Phys. Rev. D 23, 347 (1981). https://doi.org/10.1103/PhysRevD.23.347
23 New Early Dark Energy as a Solution to the . H0 and . S8 Tensions 453

9. A.D. Linde, A new inflationary universe scenario: a possible solution of the horizon, flatness,
homogeneity, isotropy and primordial monopole problems. Phys. Lett. B 108, 389 (1982).
https://doi.org/10.1016/0370-2693(82)91219-9
10. A. Albrecht, P.J. Steinhardt, Cosmology for grand unified theories with radiatively induced
symmetry breaking. Phys. Rev. Lett. 48, 1220 (1982). https://doi.org/10.1103/PhysRevLett.
48.1220
11. J.S. Cruz, F. Niedermann, M.S. Sloth, Cold new early dark energy pulls the trigger on
the . H0 and . S8 tensions: a simultaneous solution to both tensions without new ingredients.
arXiv:2305.08895
12. P. Svrcek, E. Witten, Axions in string theory. JHEP 06, 051 (2006). https://doi.org/10.1088/
1126-6708/2006/06/051 [arXiv:hep-th/0605206]
13. A. Arvanitaki, S. Dimopoulos, S. Dubovsky, N. Kaloper, J. March-Russell, String axi-
verse. Phys. Rev. D 81, 123530 (2010). https://doi.org/10.1103/PhysRevD.81.123530
[arXiv:0905.4720]
14. J. Preskill, M.B. Wise, F. Wilczek, Cosmology of the invisible axion. Phys. Lett. B 120, 127
(1983) . https://doi.org/10.1016/0370-2693(83)90637-8
15. M.S. Turner, Coherent scalar field oscillations in an expanding universe. Phys. Rev. D 28, 1243
(1983). https://doi.org/10.1103/PhysRevD.28.1243
16. R.H. Brandenberger, Cosmological perturbations in a universe dominated by a coherent scalar
field. Phys. Rev. D 32, 501 (1985). https://doi.org/10.1103/PhysRevD.32.501
17. B. Ratra, P.J.E. Peebles, Cosmological consequences of a rolling homogeneous scalar field.
Phys. Rev. D 37, 3406 (1988). https://doi.org/10.1103/PhysRevD.37.3406
18. J.E. Kim, Light pseudoscalars, particle physics and cosmology. Phys. Rept. 150, 1 (1987).
https://doi.org/10.1016/0370-1573(87)90017-2
19. D.J.E. Marsh, Axion cosmology. Phys. Rept. 643, 1 (2016). https://doi.org/10.1016/j.physrep.
2016.06.005 [arXiv:1510.07633]
20. L. Susskind, The anthropic landscape of string theory. arXiv:hep-th/0302219
21. L.F. Abbott, A mechanism for reducing the value of the cosmological constant. Phys. Lett. B
150, 427 (1985). https://doi.org/10.1016/0370-2693(85)90459-9
22. P. Agrawal, G. Obied, P.J. Steinhardt, C. Vafa, On the cosmological implications of the string
swampland. Phys. Lett. B 784, 271 (2018). https://doi.org/10.1016/j.physletb.2018.07.040
[arXiv:1806.09718]
23. E. Palti, The swampland: introduction and review. Fortsch. Phys. 67, 1900037 (2019). https://
doi.org/10.1002/prop.201900037 [arXiv:1903.06239]
24. N. Schöneberg, G. Franco Abellán, A. Pérez Sánchez, S.J. Witte, V. Poulin, J. Lesgourgues,
The . H0 Olympics: a fair ranking of proposed models. Phys. Rept. 984, 1 (2022). https://doi.
org/10.1016/j.physrep.2022.07.001 [arXiv:2107.10291]
25. V. Poulin, T.L. Smith, A. Bartlett, Dark energy at early times and ACT data: a larger Hubble
constant without late-time priors. Phys. Rev. D 104, 123550 (2021). https://doi.org/10.1103/
PhysRevD.104.123550 [arXiv:2109.06229]
26. B.S. Haridasu, H. Khoraminezhad, M. Viel, Scrutinizing early dark energy models through
CMB lensing. arXiv:2212.09136
27. J.S. Cruz, S. Hannestad, E.B. Holm, F. Niedermann, M.S. Sloth, T. Tram, Profiling cold new
early dark energy. arXiv:2302.07934
28. F. Niedermann, M.S. Sloth, New early dark energy is compatible with current LSS
data. Phys. Rev. D 103, 103537 (2021). https://doi.org/10.1103/PhysRevD.103.103537
[arXiv:2009.00006]
29. R. Murgia, G.F. Abellán, V. Poulin, Early dark energy resolution to the Hubble tension in light
of weak lensing surveys and lensing anomalies. Phys. Rev. D 103, 063502 (2021). https://doi.
org/10.1103/PhysRevD.103.063502 [arXiv:2009.10733]
30. T.L. Smith, V. Poulin, J.L. Bernal, K.K. Boddy, M. Kamionkowski, R. Murgia, Early dark
energy is not excluded by current large-scale structure data. Phys. Rev. D 103, 123542 (2021).
https://doi.org/10.1103/PhysRevD.103.123542 [arXiv:2009.10740]
454 F. Niedermann and M. S. Sloth

31. N. Deruelle, V.F. Mukhanov, On matching conditions for cosmological perturbations. Phys.
Rev. D 52, 5549 (1995). https://doi.org/10.1103/PhysRevD.52.5549 [arXiv:gr-qc/9503050]
32. W. Israel, Singular hypersurfaces and thin shells in general relativity. Nuovo Cim. B 44S10, 1
(1966). https://doi.org/10.1007/BF02710419
33. C.-P. Ma, E. Bertschinger, Cosmological perturbation theory in the synchronous and con-
formal Newtonian gauges. Astrophys. J. 455, 7 (1995). https://doi.org/10.1086/176550
[arXiv:astro-ph/9506072]
34. W. Hu, Structure formation with generalized dark matter. Astrophys. J. 506, 485 (1998). https://
doi.org/10.1086/306274 [arXiv:astro-ph/9801234]
35. A.G. Riess et al., A comprehensive measurement of the local value of the Hubble constant with
1 km s.−1 Mpc.−1 uncertainty from the Hubble space telescope and the S. H0 ES team. Astrophys.
J. Lett. 934, L7 (2022). https://doi.org/10.3847/2041-8213/ac5c5b [arXiv:2112.04510]
36. S. Joudaki et al., KiDS+VIKING-450 and DES-Y1 combined: cosmology with cos-
mic shear. Astron. Astrophys. 638, L1 (2020). https://doi.org/10.1051/0004-6361/
201936154[arXiv:1906.09262]
37. D. Blas, J. Lesgourgues, T. Tram, The cosmic linear anisotropy solving system (CLASS) II:
approximation schemes. JCAP 07, 034 (2011). https://doi.org/10.1088/1475-7516/2011/07/
034 [arXiv:1104.2933]
38. Planck collaboration, Planck 2018 results. VI. Cosmological parameters. Astron. Astrophys.
641, A6 (2020). https://doi.org/10.1051/0004-6361/201833910 [arXiv:1807.06209]
39. F. Beutler, C. Blake, M. Colless, D.H. Jones, L. Staveley-Smith, L. Campbell et al., The 6dF
galaxy survey: Baryon acoustic oscillations and the local Hubble constant. Mon. Not. R. Astron
Soc. 416, 3017 (2011). https://doi.org/10.1111/j.1365-2966.2011.19250.x
40. A.J. Ross, L. Samushia, C. Howlett, W.J. Percival, A. Burden, M. Manera, The clustering of
the SDSS DR7 main galaxy sample—I. A 4 percent distance measure at .z = 0.15. Mon. Not.
Roy. Astron. Soc. 449, 835 (2015). https://doi.org/10.1093/mnras/stv154 [arXiv:1409.3242]
41. BOSS collaboration, The clustering of galaxies in the completed SDSS-III Baryon oscillation
spectroscopic survey: cosmological analysis of the DR12 galaxy sample. Mon. Not. Roy.
Astron. Soc. 470, 2617 (2017). https://doi.org/10.1093/mnras/stx721 [arXiv:1607.03155]
42. Pan-STARRS1 collaboration, The complete light-curve sample of spectroscopically confirmed
SNe Ia from Pan-STARRS1 and cosmological constraints from the combined Pantheon sample.
Astrophys. J. 859, 101 (2018). https://doi.org/10.3847/1538-4357/aab9bb [arXiv:1710.00845]
43. M. Raveri, W. Hu, Concordance and discordance in cosmology. Phys. Rev. D 99, 043506
(2019). https://doi.org/10.1103/PhysRevD.99.043506 [arXiv:1806.04649]
44. L. Herold, E.G.M. Ferreira, E. Komatsu, New constraint on early dark energy from Planck and
BOSS data using the profile likelihood. Astrophys. J. Lett. 929, L16 (2022). https://doi.org/10.
3847/2041-8213/ac63a3 [arXiv:2112.12140]
45. K. Jedamzik, L. Pogosian, G.-B. Zhao, Why reducing the cosmic sound horizon alone can
not fully resolve the Hubble tension. Commun. Phys. 4, 123 (2021). https://doi.org/10.1038/
s42005-021-00628-x [arXiv:2010.04158]
46. S. Vagnozzi, Consistency tests of .Ʌ CDM from the early integrated Sachs-Wolfe effect: impli-
cations for early-time new physics and the Hubble tension. Phys. Rev. D 104, 063524 (2021).
https://doi.org/10.1103/PhysRevD.104.063524 [arXiv:2105.10425]
47. M.-X. Lin, G. Benevento, W. Hu, M. Raveri, Acoustic dark energy: potential conversion of
the Hubble tension. Phys. Rev. D 100, 063542 (2019). https://doi.org/10.1103/PhysRevD.100.
063542 [arXiv:1905.12618]
48. J.S. Cruz, F. Niedermann, M.S. Sloth, A grounded perspective on new early dark energy using
ACT, SPT, and BICEP/Keck. JCAP 02, 041 (2023). https://doi.org/10.1088/1475-7516/2023/
02/041 [arXiv:2209.02708]
49. K. Enqvist, M.S. Sloth, Adiabatic CMB perturbations in pre-big bang string cosmol-
ogy. Nucl. Phys. B 626, 395 (2002). https://doi.org/10.1016/S0550-3213(02)00043-3
[arXiv:hep-ph/0109214]
50. D.H. Lyth, D. Wands, Generating the curvature perturbation without an inflaton. Phys. Lett. B
524, 5 (2002). https://doi.org/10.1016/S0370-2693(01)01366-1 [arXiv:hep-ph/0110002]
23 New Early Dark Energy as a Solution to the . H0 and . S8 Tensions 455

51. T. Moroi, T. Takahashi, Effects of cosmological moduli fields on cosmic microwave back-
ground. Phys. Lett. B 522, 215 (2001). https://doi.org/10.1016/S0370-2693(01)01295-3
[arXiv:hep-ph/0110096]
52. K. Freese, M.W. Winkler, Chain early dark energy: a proposal for solving the Hubble tension
and explaining today’s dark energy. Phys. Rev. D 104, 083533 (2021). https://doi.org/10.1103/
PhysRevD.104.083533 [arXiv:2102.13655]
53. A.D. Linde, Eternal extended inflation and graceful exit from old inflation without Jordan-
Brans-Dicke. Phys. Lett. B 249, 18 (1990). https://doi.org/10.1016/0370-2693(90)90521-7
54. F.C. Adams, K. Freese, Double field inflation. Phys. Rev. D 43, 353 (1991). https://doi.org/10.
1103/PhysRevD.43.353 [arXiv:hep-ph/0504135]
55. E.J. Copeland, A.R. Liddle, D.H. Lyth, E.D. Stewart, D. Wands, False vacuum inflation with
Einstein gravity. Phys. Rev. D 49, 6410 (1994). https://doi.org/10.1103/PhysRevD.49.6410
[arXiv:astro-ph/9401011]
56. M. Cortês, A.R. Liddle, Viable inflationary models ending with a first-order phase
transition. Phys. Rev. D 80, 083524 (2009). https://doi.org/10.1103/PhysRevD.80.083524
[arXiv:0905.0289]
57. A.D. Linde, Decay of the false vacuum at finite temperature. Nucl. Phys. B 216, 421 (1983).
https://doi.org/10.1016/0550-3213(83)90072-X
58. S.R. Coleman, The fate of the false vacuum. 1. Semiclassical theory. Phys. Rev. D 15, 2929
(1977). https://doi.org/10.1103/PhysRevD.16.1248
59. C.G. Callan, Jr., S.R. Coleman, The fate of the false vacuum. 2. First quantum corrections.
Phys. Rev. D 16, 1762 (1977). https://doi.org/10.1103/PhysRevD.16.1762
60. F.C. Adams, General solutions for tunneling of scalar fields with quartic potentials. Phys. Rev.
D 48, 2800 (1993). https://doi.org/10.1103/PhysRevD.48.2800 [arXiv:hep-ph/9302321]
61. B. Xue, P.J. Steinhardt, Evolution of curvature and anisotropy near a nonsingular bounce. Phys.
Rev. D 84, 083520 (2011). https://doi.org/10.1103/PhysRevD.84.083520 [arXiv:1106.1416]
62. B.D. Fields, K.A. Olive, T.-H. Yeh, C. Young, Big-Bang nucleosynthesis after Planck. JCAP
03, 010 (2020). https://doi.org/10.1088/1475-7516/2020/03/010 [arXiv:1912.01132]
63. Planck collaboration, Planck 2018 results. VI. Cosmological parameters. Astron. Astrophys.
641, A6 (2020). https://doi.org/10.1051/0004-6361/201833910 [arXiv:1807.06209]
64. M.A. Buen-Abad, G. Marques-Tavares, M. Schmaltz, Non-Abelian dark matter and dark
radiation. Phys. Rev. D 92, 023531 (2015). https://doi.org/10.1103/PhysRevD.92.023531
[arXiv:1505.03542]
65. P.B. Arnold, O. Espinosa, The effective potential and first order phase transitions: beyond
leading-order. Phys. Rev. D 47, 3546 (1993). https://doi.org/10.1103/PhysRevD.47.3546
[arXiv:hep-ph/9212235]
66. M. Dine, R.G. Leigh, P.Y. Huet, A.D. Linde, D.A. Linde, Towards the theory of the elec-
troweak phase transition. Phys. Rev. D 46, 550 (1992). https://doi.org/10.1103/PhysRevD.46.
550 [arXiv:hep-ph/9203203]
67. S.R. Coleman, E.J. Weinberg, Radiative corrections as the origin of spontaneous symmetry
breaking (1888). Phys. Rev. D 7 (1973). https://doi.org/10.1103/PhysRevD.7.1888
68. D.A. Kirzhnits, A.D. Linde, Symmetry behavior in Gauge theories. Annals Phys. 101, 195
(1976). https://doi.org/10.1016/0003-4916(76)90279-7
69. J.S. Cruz, F. Niedermann, M.S. Sloth, NANOGrav meets hot new early dark energy and the
origin of neutrino mass. arXiv:2307.03091
70. D. Aloni, A. Berlin, M. Joseph, M. Schmaltz, N. Weiner, A step in understanding the Hub-
ble tension. Phys. Rev. D 105, 123516 (2022). https://doi.org/10.1103/PhysRevD.105.123516
[arXiv:2111.00014]
71. N. Schöneberg, G. Franco Abellán, A step in the right direction? Analyzing the Wess Zumino
dark radiation solution to the Hubble tension. JCAP 12, 001 (2022). https://doi.org/10.1088/
1475-7516/2022/12/001 [arXiv:2206.11276]
72. N. Schöneberg, G. Franco Abellán, T. Simon, A. Bartlett, Y. Patel, T.L. Smith, The weak,
the strong and the ugly—a comparative analysis of interacting stepped dark radiation.
arXiv:2306.12469
456 F. Niedermann and M. S. Sloth

73. M.A. Buen-Abad, Z. Chacko, C. Kilic, G. Marques-Tavares, T. Youn, Stepped partially acoustic
dark matter: likelihood analysis and cosmological tensions. arXiv:2306.01844
74. I.J. Allali, F. Rompineve, M.P. Hertzberg, Dark sectors with mass thresholds face cosmological
datasets. arXiv:2305.14166
75. A. Berlin, N. Blinov, S.W. Li, Dark sector equilibration during nucleosynthesis. Phys. Rev. D
100, 015038 (2019). https://doi.org/10.1103/PhysRevD.100.015038 [arXiv:1904.04256]
76. D. Aloni, M. Joseph, M. Schmaltz, N. Weiner, Dark radiation from neutrino mixing after big
bang nucleosynthesis. arXiv:2301.10792
77. R.N. Mohapatra, J.W.F. Valle, Neutrino mass and Baryon number nonconservation in super-
string models. Phys. Rev. D 34, 1642 (1986). https://doi.org/10.1103/PhysRevD.34.1642
78. M.C. Gonzalez-Garcia, J.W.F. Valle, Fast decaying neutrinos and observable flavor violation in
a new class of Majoron models. Phys. Lett. B 216, 360 (1989). https://doi.org/10.1016/0370-
2693(89)91131-3
79. F. Deppisch, J.W.F. Valle, Enhanced lepton flavor violation in the supersymmetric inverse
seesaw model. Phys. Rev. D 72, 036001 (2005). https://doi.org/10.1103/PhysRevD.72.036001
[arXiv:hep-ph/0406040]
80. A. Abada, M. Lucente, Looking for the minimal inverse seesaw realisation. Nucl. Phys. B 885,
651 (2014). https://doi.org/10.1016/j.nuclphysb.2014.06.003 [arXiv:1401.1507]
81. G. ’t Hooft, Naturalness, chiral symmetry, and spontaneous chiral symmetry breaking. NATO
Sci. Ser. B 59, 135 (1980). https://doi.org/10.1007/978-1-4684-7571-5_9
82. M. D’Onofrio, K. Rummukainen, Standard model cross-over on the lattice. Phys. Rev. D 93,
025003 (2016). https://doi.org/10.1103/PhysRevD.93.025003 [arXiv:1508.07161]
83. M. Archidiacono, S. Hannestad, Updated constraints on non-standard neutrino interac-
tions from Planck. JCAP 07, 046 (2014). https://doi.org/10.1088/1475-7516/2014/07/046
[arXiv:1311.3873]
84. NANOGrav collaboration, The NANOGrav 15-year data set: search for signals from
new physics. Astrophys. J. Lett. 951 (2023). https://doi.org/10.3847/2041-8213/acdc91
[arXiv:2306.16219]
85. M. Garny, M. Sandora, M.S. Sloth, Planckian interacting massive particles as dark mat-
ter. Phys. Rev. Lett. 116, 101302 (2016). https://doi.org/10.1103/PhysRevLett.116.101302
[arXiv:1511.03278]
86. M. Garny, A. Palessandro, M. Sandora, M.S. Sloth, Theory and phenomenology of Planckian
interacting massive particles as dark matter. JCAP 02, 027 (2018). https://doi.org/10.1088/
1475-7516/2018/02/027 [arXiv:1709.09688]
87. ACT collaboration, The Atacama cosmology telescope: DR4 maps and cosmological parame-
ters. J. Cosmol. Astropart. Phys. 12, 047 (2020). https://doi.org/10.1088/1475-7516/2020/12/
047 [arXiv:2007.07288]
88. K. Abazajian et al., CMB-S4 science case, reference design, and project plan. arXiv:1907.04473
89. J.P. Gardner et al., The James Webb space telescope. Space Sci. Rev. 123, 485 (2006). https://
doi.org/10.1007/s11214-006-8315-7 [arXiv:astro-ph/0606175]
90. L. Amendola et al., Cosmology and fundamental physics with the Euclid satellite. Living Rev.
Rel. 21, 2 (2018). https://doi.org/10.1007/s41114-017-0010-3 [arXiv:1606.00180]
91. T.L. Smith, M. Lucca, V. Poulin, G.F. Abellan, L. Balkenhol, K. Benabed et al., Hints of early
dark energy in Planck, SPT, and ACT data: New physics or systematics?. Phys. Rev. D 106,
043526 (2022). https://doi.org/10.1103/PhysRevD.106.043526 [arXiv:2202.09379]
Chapter 24
On the Dark Radiation Role
in the Hubble Constant Tension

Stefano Gariazzo and Olga Mena

24.1 Introduction

Dark radiation in our universe provides a unique window to new mechanisms able to
solve many open questions in particle physics and cosmology nowadays. The particle
content of the minimal Standard Model (SM) scenario may be enlarged with addi-
tional species that feebly couple to ordinary matter, i.e. dark radiation. The putative
existence of these species is strongly motivated, both from the pure theoretical and
phenomenological/experimental perspectives. The explanation of neutrino masses
within the SM and the short-baseline neutrino oscillation anomalies may need the
existence of additional sterile neutrino species. The so-called strong CP problem may
imply the existence of axions in our universe, which could contribute significantly to
the (observationally required) dark matter component. Last, but not least, Majoron
models provide very appealing scenarios where lepton number may be spontaneously
broken and rare events such as neutrinoless double beta decay are possible.
Currently, there are several open anomalies that can not be fully understood in
the minimal cosmological constant plus cold dark matter (.ɅCDM) scenario. The
most significant, .5σ disagreement is related to the value of the Hubble constant
. H0 extracted from model-independent, direct measurements of local distances and
redshifts in the nearby Universe (. H0 = 73.04 ± 1.04 km s−1 Mpc−1 ) [1] and the one
indirectly inferred from model-dependent Cosmic Microwave Background (CMB)
observations (. H0 = 67.4 ± 0.5 km s−1 Mpc−1 ) [2], see also [1, 3–6]. The goal of
this manuscript is to analyze the role of dark radiation models in the Hubble constant
tension. We start in Sect. 24.2 with a brief introduction which summarizes the main
properties of the dark radiation component, such as its thermal abundance and its
impact on the cosmological observables, focusing mostly on the CMB. Sect. 24.3
describes the most simple models involving dark radiation species ad their role in

S. Gariazzo (B)
Istituto Nazionale di Fisica Nucleare (INFN), Sezione di Torino, Via P. Giuria 1, 10125 Turin, Italy
e-mail: gariazzo@to.infn.it
O. Mena
Instituto de Física Corpuscular (CSIC-Universitat de València), 46980 Paterna, Spain
e-mail: omena@ific.uv.es

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 457
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_24
458 S. Gariazzo and O. Mena

the . H0 conundrum. Sect. 24.4 presents more sophisticated and refined dark radiation
scenarios, which may provide a resolution of the. H0 issue. We conclude in Sect. 24.5.

24.2 Dark Radiation in the Early Universe

One of the most important phases of the Universe’s evolution is radiation domination.
During this epoch, the largest contribution to the energy density of the Universe
comes from relativistic particles, i.e. mainly photons, neutrinos and electrons by the
time the temperature of the fluid is below 100 MeV. During radiation domination,
several interesting processes take place: (i) neutrinos decouple when the temperature
of the plasma is approximately 2 MeV; (ii) electrons become non-relativistic shortly
after and transfer their energy density to photons while annihilating away, and (iii)
light nuclei are produced in the Big-Bang Nucleosynthesis (BBN) process until the
photon temperature drops below .∼ 0.05 MeV. Shortly after the end of radiation
domination (at .T ∼ 1 eV), the last scattering of photons occurs (.T ∼ 0.3 eV) and
the CMB radiation is produced. The expansion speed at all these times, proportional
to the square root of the radiation energy density, is crucial because it affects the
observables we measure today.
After electron-positron pairs annihilate into photons, neutrinos and photons are
the only SM particles behaving as radiation. Other relativistic particles may exist, as
we will discuss in the following. Such non-standard particles are normally grouped
under the name “dark radiation”, since they do not take part into electroweak inter-
actions. The amount of the dark radiation energy density, .ρR , can be conveniently
parameterized by means of the effective number of relativistic neutrino-like species,
Neff , by ( )
( )
7 4 4/3
.ρR = ργ 1+ Neff , (24.1)
8 11

where .ργ represents the photon energy density. If only three standard neutrinos
which underwent instantaneous decoupling exist, Neff would be equal to 3. In case
the neutrino energy density is different from the instantaneous decoupling one, Neff
may deviate from 3 even in presence of only three standard neutrino families. The
presence of additional contributions to the radiation energy density, moreover, would
correspond to an increased value of Neff . In other words, . Neff /= 3 might originate
either because of new degrees of freedom which have nothing to do with standard
neutrinos, or from a non-standard momentum distribution for the three neutrinos.
Our theoretical knowledge of their decoupling, however, states that standard neu-
trinos did not undergo instantaneous decoupling and that their momentum distribu-
tion function slightly deviates from a pure Fermi-Dirac one. When taking into account
the momentum dependence of neutrino-electron interactions and the evolution of the
thermal plasma around electron-positron annihilation, the standard neutrino energy
std
density is computed to be . Neff = 3.044 [7–9], see also [10]. This number was pre-
24 On the Dark Radiation Role in the Hubble Constant Tension 459

viously claimed to be a bit higher [11, 12], but state-of-the-art calculations confirm
that the (theoretical and numerical) error on this value is at the level of .10−4 . The
presence of new physics related to the three standard neutrinos, such as non-standard
interactions (NSI, see e.g. [13]) between neutrinos and electrons [14, 15], or a non-
std
unitary neutrino mixing matrix [16] can lead to deviations from Neff . These effects
are rather small and do not significantly impact the . H0 value, although they may be
tested by next-generation CMB measurements [17, 18]. As we shall see, the situa-
tion changes significantly if we consider the presence of additional particles (axions,
sterile neutrinos, and so on) or more complicated neutrino interactions, for example
with dark matter. In such cases, it is common to have much larger contributions
to Neff , usually denoted as .ΔNeff ≡ Neff − Neff std
. An exhaustive description of the
effect of ΔNeff on the different cosmological observables can be found in Ref. [19].
Current limits on .ΔNeff arise primarily from observables at two epochs: (i) at the
BBN period, and, (ii) at the Recombination epoch. Here we briefly summarize the
main impact of this parameter on both BBN light element abundances and on the
CMB power spectrum, as the impact on the matter power spectrum is subdominant.
BBN refers to the formation of the first light nuclei (heavier than the lightest
isotope of hydrogen) in the very first three minutes of our universe’s lifetime. BBN
abundances are computed by employing a solid understanding of the nuclear inter-
actions involved in the production of elements, providing a natural laboratory where
to test extensions of the SM of elementary particles that involve additional rela-
tivistic species ΔNeff . Any extra contribution to the dark radiation of our universe
will increase the expansion rate . H (z) and will shift towards higher temperatures the
freeze-out epoch of the weak interactions, implying a higher neutron-to-proton ratio
and therefore a larger fraction of primordial Helium and Deuterium (as well as a
higher fraction of other primordial elements) with respect to hydrogen. This makes
BBN a powerful tool for constraining the total amount of relativistic species and
other beyond-the-SM physics frameworks: given a concrete model, we can solve
numerically the set of differential equations that regulate the nuclear interactions
in the primordial plasma, see e.g. [20–22], compute the light element abundances
and compare the results to the values inferred by astrophysical and cosmological
observations. Given current uncertainties, the standard BBN predictions show a
good agreement with direct measurements of primordial abundances [23–26] lim-
iting .ΔNeff ≾ 0.3 − 0.4 at 95% CL. Notice also that the BBN predictions for the
Helium abundance can impact the CMB angular spectra because they can be used to
estimate the baryon energy density through a simple formula [27]:
( )3
1 − 0.007125 Y pBBN TCMB
Ωb h =
.
2
η10 . (24.2)
273.279 2.7255 K

Here, .η10 ≡ 1010 n b /n γ is the photon-baryon ratio today, .TCMB is the CMB tempera-
ture at the present time and .Y pBBN ≡ 4n He /n b is the Helium nucleon fraction, defined
as the ratio of the 4-Helium number density to the total baryon one.
460 S. Gariazzo and O. Mena

Concerning the CMB temperature power spectrum, first of all, varying Neff
changes the redshift of the matter radiation equivalence, .z eq , inducing an enhance-
ment of the early Integrated Sachs Wolfe (ISW) effect which increases the CMB
spectrum around the first acoustic peak. Nevertheless, this is a sub-dominant effect.
The authors of Ref. [28] carefully explained that the most relevant impact of chang-
ing Neff is located at high multipoles, i.e. at the damping tail. If ΔNeff increases, the
Hubble parameter . H during radiation domination will increase as well. The delay in
matter radiation equality will also modify the sound speed and the comoving sound
horizon, proportional to .1/H , which reads as:
{ τ' { a
da
r =
. s dτ cs (τ ) = cs (a) ,
0 0 a2 H

and it is proportional to the inverse of the expansion rate .rs ∝ 1/H . The consequence
is a reduction in the angular scale of the acoustic peaks .θs = rs /DA , where . DA is the
angular diameter distance, causing a horizontal shift of the peak positions towards
higher multipoles. Furthermore, a vertical shift also affects the amplitude of the
peaks at high multipoles, where the ISW effect is negligible, and it is related to the
so-called Silk damping. This results from the fact that the baryon-photon decoupling
is not an instantaneous process, leading to a diffusion damping of oscillations in
the plasma. If decoupling starts at .τd and ends at .τls , during .Δτ the radiation free
streams on scale .λd = (λΔτ )1/2 where .λ is the photon mean free path and .λd is
shorter than the thickness of the last scattering surface. As a consequence, temperature
fluctuations on scales smaller than .λd are damped, because on such scales photons
can spread freely both from overdensities and underdensities. The overall result is
√ scale .θd = rd /DA is proportional to the square root of the
that the damping angular
expansion rate .θd ∝ H and consequently it increases with ΔNeff which therefore
induces a suppression of the peaks and a smearing of the oscillations that intensifies
at the CMB damping tail.
The three aforementioned effects caused by a non-zero ΔNeff (redshift of equiva-
lence, the size of the sound horizon at recombination, and the damping tail suppres-
sion) can be easily compensated by varying other cosmological parameters, includ-
ing the Hubble constant . H0 . Notice that the horizontal shift towards smaller angular
scales caused by an increased value of Neff can be compensated by decreasing . DA ,
which can be automatically satisfied by increasing . H0 . The effect of Neff on the
damping tail is however more difficult to mimic via . H0 , as it is mostly degenerated
with the helium fraction which enters directly in .rd , i.e. the mean square diffusion
distance at recombination via .n e , the number density of free electrons.
There is however one effect induced by Neff which cannot be mimicked by other
cosmological parameters: the neutrino anisotropic stress [29, 30], related to the fact
that neutrinos are free-streaming particles propagating at the speed of light, faster
than the sound speed in the photon fluid, suppressing the oscillation amplitude of
CMB modes that entered the horizon in the radiation epoch. The effect on the CMB
power spectrum is therefore located at scales that cross the horizon before the matter-
24 On the Dark Radiation Role in the Hubble Constant Tension 461

radiation equivalence by an increase in power of .5/(1 + 154

f ν ) [31], where . f ν is the
fraction of radiation density contributed by free-streaming particles.
All in all, our current knowledge confirms that Neff is close to 3 as measured
by CMB observations (. Neff = 2.99+0.34
−0.33 at 95% confidence level (CL) [2]) or BBN
abundances (e.g. . Neff = 2.87+0.24
−0.21 at 68% CL [21]) independently.

24.3 Tree-level Dark Radiation Solutions to the . H0 Tension

Let us now analyse the simplest dark radiation candidate that can be used to extend
the standard three-neutrino picture and discuss their effect on the . H0 tension.

24.3.1 Sterile Neutrinos

One simple dark radiation candidate is the sterile neutrino, intended as a right-handed
fermion which cannot take part into SM interactions but that participates in neutrino
oscillations. In order to be considered as dark radiation, the sterile neutrino must be
relativistic in the early universe, indicatively at the time of BBN or CMB epochs.
This excludes sterile neutrinos with masses much larger than the MeV scale, which
are usually proposed in order to build a seesaw model and explain the smallness
of active neutrino masses, see e.g. [32–34]. Sterile neutrinos at the keV scale can
play a significant role in the universe as warm dark matter [35, 36], but since they
become non-relativistic during BBN, they are not considered as dark radiation com-
ponents. When considering masses at the eV scale, sterile neutrinos would become
non-relativistic approximately at the time of matter-radiation equality, thus being rel-
ativistic during the entire radiation-dominated era, although already non-relativistic
at CMB decoupling (.T ∼ 0.3 eV). Sterile neutrinos corresponding to much smaller
mass splittings with respect to active neutrinos (mass splittings smaller than the solar
or atmospheric ones) are also considered in quasi-Dirac or pseudo-Dirac scenarios,
see e.g. [37, 38]. In this case, they may remain relativistic until today, depending on
the mass of the lightest neutrino. In the following, however, we will focus primarily
on the eV-scale sterile neutrino.
Sterile neutrinos at the eV scale have been introduced to give a possible neutrino-
oscillation explanation to the so-called Short-Baseline (SBL) anomalies [39–42],
which originally included the excess appearance of electron antineutrinos measured
by LSND [43], later confirmed also by MiniBooNE [44, 45], and the anomalous
disappearance of electron neutrinos at Gallium experiments [46, 47] and of electron
antineutrinos from nuclear reactors [48]. The original anomalies have been studied
at different experimental probes and at the theoretical level for more than ten years.
Here we will briefly summarize the present status.
462 S. Gariazzo and O. Mena

Concerning the appearance channel, the LSND results have been tested by sev-
eral experiments apart from MiniBooNE. None of these experiments has observed an
anomalous signal, including KARMEN [49] and OPERA [50], which therefore put
only upper bounds on active-sterile mixing angles. In more recent years, the Micro-
BooNE experiment is testing the MiniBooNE anomaly, and the first results seem
to indicate that the excess of events is not compatible with active-sterile neutrino
oscillations [51, 52].
In the disappearance channel, we can separate the electron and muon (anti)neutrino
channels. In the former case, the Gallium and reactor anomalies have been studied
both experimentally and theoretically. Recent revisitations of the Gallium cross sec-
tions appear to reduce the significance of the Gallium anomaly with respect to the
original calculations [53, 54], but the BEST experiment observed a disappearance
with stronger significance [55, 56] with respect to previous GALLEX and SAGE
results, see also [57]. If neutrino oscillations are responsible for the BEST results, part
of the preferred region is in tension with SBL reactor antineutrino experiments [58–
61]. On the other hand, although most of the reactor experiments do not confirm the
original reactor anomaly [58–63], the Neutrino-4 experiment claims a best-fit at a
high mass splitting and mixing angle [64], whose significance may be overestimated
due to the treatment of the .χ 2 [65]. Neutrino-4 oscillation results are perfectly com-
patible with BEST constraints, which however have a much broader allowed range
for the new mass splitting. Recent theoretical calculations of the reactor antineu-
trino flux [66–68] differ slightly from the two that produced the original reactor
anomaly [69, 70], which is therefore significantly reduced or even disappeared if the
most updated flux estimations are considered [68, 71].
Finally, concerning muon (anti)neutrino disappearance, most of the experiments,
for example MINOS/MINOS+ [72], only put upper limits on the mixing parameters
because no SBL oscillations are observed. Only in recent years, IceCube reported a
mild preference in favor of active-sterile neutrino oscillations by analysing 8 years of
muon neutrino events [73], with a broad uncertainty on the preferred mass splitting
and mixing angles.
Assuming that all the above-mentioned anomalies are generated by active-sterile
neutrino mixing, when one attempts to combine all these appearance and disappear-
ance results in order to produce a global fit of active-sterile oscillation parameters,
it seems impossible to reconcile the strong upper bounds from disappearance exper-
iments with the observations at appearance probes. The tension was already strong
a few years ago [40, 74, 75], at the point that a combination of appearance and dis-
appearance constraints has no statistical meaning. In summary, the sterile neutrino
searches from different probes seem unable to pin out a single preferred region for
the active-sterile oscillation parameters, and possibly a consistent explanation of the
anomalies requires different kinds of new physics. In the following, therefore, we
will consider the potential impact of eV-scale sterile neutrinos on the cosmological
environment and ignore the problems in terrestrial searches of sterile neutrinos.
The thermalization of sterile neutrinos in the early universe has been studied for a
long time, see e.g. [76–82], or [83] for a review of early studies. The most important
result in these early calculations is to show that the production of sterile neutrinos
24 On the Dark Radiation Role in the Hubble Constant Tension 463

occurs mostly in a non-resonant way through neutrino oscillations. Reference [81]

provides a rather simple formula that determines the relation between active-sterile
mixing parameters and the approximated value of ΔNeff they generate. In more
recent years, precise numerical calculations have been developed in order to deter-
mine the contribution of eV-scale sterile neutrinos to ΔNeff . Calculations have been
first implemented in a “one active plus one sterile neutrino” scenario but keeping
the full momentum dependence, see e.g. [84–86], and then extended to include mul-
tiple flavors (two active and one sterile neutrino) [87–89]. Finally, more accurate
calculations at the numerical level have been proposed in recent years [90, 91].
The amount of sterile neutrinos produced through oscillations in the early universe
depends significantly on the values of the active-sterile mixing parameters. Because
the dense plasma blocks neutrino oscillations at very early times by maintaining
neutrinos into a flavor state, oscillations cannot take place until rather late. Since this
effect depends on the oscillation frequency, higher mass splittings correspond to an
earlier start of oscillations, that in turn means that there is more time to produce sterile
neutrinos. Lower mass splittings, instead, generate oscillations that are blocked for
a longer time, and are less efficient in producing sterile states. On the other hand,
a large mixing angle allows to have a faster conversion between active and sterile
flavors. Finally, it has been noticed that the new active-sterile mixing angles are
equivalent at the time of producing neutrino oscillations in the early universe. More
specifically, for a .Δm 241 ∼ 1 eV, it is sufficient to have any of the mixing angles
larger than .∼ 10−3 to generate a sterile state in full thermal equilibrium with the
active neutrinos, or .ΔNeff ≃ 1 [90].
This is of course in tension with the constraints we obtain from CMB and BBN
observables. Cosmological analyses normally parameterize the presence of ster-
ile neutrinos by means of its contribution ΔNeff to the effective number of neu-
trinos when relativistic, and its contribution to the matter energy density, propor-
tional to the effective mass .m eff,sterile , when it becomes non-relativistic. In case of
non-resonant production of the sterile state [92], the relation .m eff,sterile = m s ΔNeff
applies, where .m s is the sterile neutrino physical mass. From the sterile neutrino
search, Planck obtains upper limits . Neff < 3.30 and .m eff,sterile < 0.652 eV at 95%
CL (TTTEEE+lensing+BAO) [2]. Notice that in this case, the analysis also yields
+1.3 −1
. H0 = 67.8−1.2 km s Mpc−1 at 95% CL, meaning that the Hubble tension is not
alleviated. The constraints are particularly strong thanks to the combined effect
of Neff and . H0 on both temperature and polarization. If one only considers the
TT+lensing+BAO dataset, Planck reports . Neff < 3.52, .m eff,sterile < 0.551 eV and
+2.1 −1
. H0 = 68.2−1.7 km s Mpc−1 , which is still in tension with a fully thermalized
fourth neutrino, but with a slightly reduced significance. Notice that the bounds on
. Neff being smaller than 4 as reported by Planck are also confirmed by ACT and
SPT, which respectively yield . Neff < 3.38, .m eff,sterile < 0.561 eV and . Neff < 3.86,
.m eff,sterile < 0.232 eV when analysed in combination with WMAP 9-year data and

other low-redshift probes [93]. Notice that in the SPT case, the bound on m eff,sterile is
much stronger, while the bound on Neff is relaxed. This arises from the fact that heav-
ier states are allowed only if their contribution to Neff is rather small (otherwise their
contribution to the non-relativistic energy density, proportional to .m eff,sterile , would
464 S. Gariazzo and O. Mena

be too high), while for lighter states a larger value of Neff is allowed because these
particles mostly act as radiation. Even in the SPT case, however, these results show
that none of the available cosmological data sets currently allows for the presence of
a fully thermalized sterile neutrino, and the allowed parameter ranges do not permit
a full solution of the . H0 tension.

24.3.2 Neutrino Asymmetries

In the standard big-bang model, the neutrino asymmetry .ηνα ≡ (n να − n ν̄α )/n γ is
assumed to be negligibly small, comparable to the tiny-but-crucial baryon asymmetry.
That is, while standard baryogenesis models involving sphalerons suggest that .ηνα
should be of the order of the baryon asymmetry .η B = n B /n γ ≃ 6.1 × 10−10 [94,
95], other proposed models [96–98] manage to combine a large lepton asymmetry
with the value of .η B . The initial asymmetry in a given neutrino flavour .α, defined as
the difference between the neutrino and antineutrino comoving densities, is related
to the neutrino chemical potential .μα . It is possible to compute the constraints on this
parameter, or, as it is usually done in the literature, on the dimensionless degeneracy
parameter .ξα ≡ μα /Tν . A non-vanishing value of the electron neutrino chemical
potential .ξe affects the neutron/proton freeze-out, modifying BBN predictions [99].
Notice that a priori, the individual neutrino flavour chemical potentials can have
different values. Depending on the neutrino mixing angles as well as on the initial
values of the neutrino chemical potentials there could be or not an equilibration of the
different .ξα [100].1 A significant lepton asymmetry in the early universe, represented
by a nonvanishing chemical potential .ξαi , for thermal distributions of two light mass
eigenstates .νi , will contribute to the dark radiation of the universe as [100]
( ( )2 )
∑ 15 ( ξ )2 xi
ΔNeff
. = 2+ . (24.3)
i=1,2
7 π π

The authors of Ref. [100] show that CMB data alone lead to a 95% CL limit of
ξ < 0.77, associated to .ΔNeff = 0.3 and . H0 = 67.71 ± 0.95 km s−1 Mpc−1 , and
.
therefore not alleviating the . H0 tension. However, a combination with Supernovae I
data provides weak evidence for a non negligible value of .ξ , together with a .95% CL
lower bound for the Hubble parameter of . H0 > 69.8 km s−1 Mpc−1 , larger than
the value without Supernovae observations and in a better agreement with local
measurements of the Hubble constant.

1 When assuming a unique common degeneracy parameter .ξ for the three neutrino degeneracy
parameters due to neutrino oscillations [101–103], one can derive bounds from BBN observa-
tions [94, 104], from CMB data [2, 105], or from a combination of the former two measurements,
providing .ξ = 0.001 ± 0.016 [106, 107].
24 On the Dark Radiation Role in the Hubble Constant Tension 465

24.3.3 Axions

Axions are the dynamical pseudo-scalar fields providing one of the most compelling
solutions to the strong CP problem [108, 109]. In addition, axions also may be consid-
ered as a natural candidate for the dark matter component in our universe. Apart from
non-thermal mechanisms, axions can be copiously produced in the early Universe
also via scattering and decays of particles belonging to the primordial bath [110–118],
contributing to the radiation energy density Neff in the same way as massive neu-
trinos do when they are still relativistic. Eventually, axions become non-relativistic
and provide a hot and sub-dominant dark matter component. The presence of such
a cosmic axion background can leave distinct and detectable imprints, and current
cosmological data put bounds on the axion mass and interactions [119–132]. Axion
couplings are proportional to the mass of the axion itself, and thermal production
channels are efficient only if the axion mass is large enough. On the contrary, the
cold axion dark matter density is a decreasing function of the mass. Therefore, a sig-
nificant thermal population of axions would be possible only if cold axions provide
a sub-dominant component to the observed cosmic dark matter abundance. In terms
of . f a , known as the axion decay constant, the axion mass reads as [133–135]
( )
1012 GeV
m a ≃ 5.7 μeV
. . (24.4)
fa

The landscape of axion models is broad [136, 137]. Nevertheless, it is pos-

sible to divide them into two main classes according to the origin of the color
anomaly: the Kim-Shifman-Vainshtein-Zakharov (KSVZ) [138, 139] and Dine-
Fischler-Srednicki-Zhitnitsky (DFSZ) [140, 141] frameworks, leading to the very
same axion mass, see Eq. (24.4). However, the interactions with electroweak gauge
bosons and SM fermions are different. Scatterings are the only thermal production
channel available for the minimal KSVZ and DFSZ frameworks due to the flavor-
conserving axion interactions. If flavor-violating couplings are allowed then decays
also contribute [118]. Initially, axions may belong to the thermal bath if during or
after inflationary reheating there are efficient mechanisms to produce them. Even if
they are not present at the onset of the radiation domination epoch, they are produced
by thermal scatterings and their interaction strength could bring them to equilibrium.
A quick method to estimate the axion relic density is via the instantaneous decou-
pling approximation, assuming that axions at some point reach thermal equilibrium,
and that they suddenly decouple when the bath had a temperature .TD identified by
the relation
.┌a (TD ) = H (TD ) . (24.5)

The asymptotic axion comoving density results in

45ζ (3)
Y ∞ = Ya (T ≾ TD ) =
. a , (24.6)
2π 4 g★s (TD )
466 S. Gariazzo and O. Mena

with .ζ (3) ≃ 1.2 the Riemann zeta function. The asymptotic comoving density
depends on the decoupling temperature only through the factor.g★s (TD ) in the denom-
inator. Axions decoupling at later times are a more significant fraction of the ther-
mal bath and therefore their comoving density is larger. Notice that the criterion
in Eq. (24.5) leads to a convenient estimate of the decoupling epoch but it is far
from being rigorous. As it is manifest from Eq. (24.6), the final axion abundance is
sensitive to the detailed value of the decoupling temperature only if .g★s is changing
around the decoupling time, see Ref. [130], which provides a robust computation
of the axion relic density calculation in terms of the Boltzmann equation, in order
to find solid cosmological bounds on the axion mass, finding that the actual axion
relic density is substantially larger than previously estimated. Once the relic density
is computed, it is possible to evaluate the axion contribution to the effective number
( )4/3
of additional neutrino species via the relation .ΔNeff ≃ 75.6 Ya∞ , assuming that
the phase-space distribution is thermal. Therefore, as long as thermal axions remain
relativistic particles (.Ta >> m a ), they behave as radiation in the early Universe and
their cosmological effects are those produced via their contribution to the effec-
tive number of neutrino species Neff . By means of BBN light element abundances
data, Ref. [130] finds for the KSVZ axion .ΔNeff < 0.33 and an axion mass bound
.m a < 0.53 eV (i.e., a bound on the axion decay constant. f a > 1.07 × 10 GeV), both
7

at 95% CL. These BBN bounds are improved to .ΔNeff < 0.14 and .m a < 0.16 eV
(. f a > 3.56 × 107 GeV) if a prior on the baryon energy density from CMB data is
assumed. When cosmological observations from the CMB temperature, polarization
and lensing from the Planck satellite are combined with large scale structure data
limits of .ΔNeff < 0.23 and .m a < 0.28 eV are found, both at 95% CL. Very simi-
lar results are quoted for the DFSZ axion. The authors of Ref. [114] showed that a
model in which hot axions are produced from the coupling with muons leads to an
alleviation of the Hubble tension at .3σ .

24.4 Higher Order Dark Radiation Corrections to the . H0

Tension

In the following, we shall present proposed scenarios which imply extra interactions
beyond the SM paradigm, involving also in some cases exotic particles, such as sterile
neutrino species, the dark matter component, or Majorons.

24.4.1 Interacting Scenarios

Both neutrinos and dark matter provide evidence for physics beyond the SM
of elementary particles and possible interactions among them can induce impor-
tant changes in the canonical evolution within the minimal .ɅCDM model, see
Refs. [142–150]. In particular, in these models, the value of . Neff will generically
24 On the Dark Radiation Role in the Hubble Constant Tension 467

be increased at photon decoupling while will remain unchanged during the BBN
period. Since the effective number of relativistic degrees of freedom is proportional
to the ratio of the neutrino and photon temperatures, if the photon temperature .Tγ is
reduced, or, alternatively, the neutrino temperature is increased, . Neff will be larger
than in the canonical scenario. If there exists a dark matter particle that remains in
thermal equilibrium with neutrinos after their weak decoupling processes, these inter-
actions will effectively reheat the neutrino sector with respect to the electromagnetic
plasma. This neutrino reheating will take place after standard neutrino decoupling
once the dark matter particles are no longer relativistic, implying they should have
a mass of tens of MeV [151, 152]. It has been explored if such a scenario could
alleviate the . H0 tension [148, 153]. Including Planck polarization data, however,
dilutes such a possibility, although this very same scenario is able to significantly
relax the lower bounds on the value of the clustering parameter .σ8 inferred in the
context of .ɅCDM from the Planck data, leading to agreement within .1 − 2σ with
weak lensing estimates of the .σ8 parameter.

24.4.2 Self-interacting Active Neutrinos

Free-streaming neutrinos travel supersonically through the photon-baryon plasma

at early times, inducing a net phase shift in the CMB power spectra towards larger
scales (smaller multipoles), leading to a physical size of the photon sound horizon
at last scattering that is slightly larger [29, 154–157]. Self-interacting neutrinos, see
e.g. [158, 159] shift the power spectrum towards smaller scales and boost their fluctu-
ation amplitude, reducing the physical size of photon sound horizon at last scattering:
a smaller value of the angular diameter distance would be required, implying a higher
value of . H0 [160]. If the mass of the mediator is heavy, the process can be expressed
in terms of a four-fermion interaction .G eff ν ν̄ν ν̄ [160], with .G eff >> G F , the former
governing weak-interaction processes. Neutrinos experience self-scatterings after
weak decoupling, and increasing .G eff delays neutrino free-streaming. The authors
of Ref. [160] found a strongly interacting mode with .G eff = 2.5+0.8 −0.5 × 10 GeV
4 −2

and associated it to a much larger value of . H0 than in the standard .ɅCDM model,
−1
. H0 = 72.3 ± 1.4 km s Mpc−1 . This very same strongly interacting neutrino cos-
mology prefers . Neff = 4.02 ± 0.29. Despite being a very interesting result, the for-
mer analysis was restricted to Planck observations of temperature power spectrum
data: the authors of Ref. [161] realized that when high polarisation data from the
Planck 2018 release are included in the fit, the strongly interacting neutrino mode
was disfavoured, even if it could not be completely excluded. As a consequence,
strong neutrino self-interactions do not lead to a high value of the Hubble constant
and therefore such a model is not a viable solution to the current . H0 discrepancy
when considering the full Planck 2018 data, see also Refs. [159, 162–168].
Among the classes of theoretical neutrino model building, the self-interacting
scenario discussed here includes an important class of models related to the expla-
nation of the smallness of neutrino masses. Why are the neutrino masses within
468 S. Gariazzo and O. Mena

the SM much smaller than those of the charged fermions in the very same family?
One of the most elegant and complete benchmarks is the so-called seesaw mech-
anism [32–34], in which additional right handed neutrinos are added via a Majo-
rana mass term, naturally suppressing the (light, active) neutrino mass scale. These
models imply a lepton number spontaneous symmetry breaking, with an associated
pseudo-Goldstone boson, the Majoron [169, 170], a light scalar with very weak inter-
actions with visible matter [171, 172]. Nevertheless, Majoron-neutrino interactions
will change the cosmological observables at both the BBN and CMB periods, as they
contribute to both ΔNeff and to the neutrino anistotropic stress, reducing neutrino
free-streaming [29, 173–176]. With Planck 2015 data, the Majoron models were
shown to predict a higher value of the . H0 parameter, especially when . Neff is free
in the analysis [177]. Updated studies demonstrated that the flexibility in explaining
the . H0 tension was reduced when considering non-CMB data together with Planck
[178], as the . H0 values obtained in the combined analysis of Planck 2015 TTTEEE
plus external data is .68.13 ± 0.48 km s−1 Mpc−1 . Considering Planck 2018 data and
an updated local measurement of . H0 , instead, the authors of Ref. [179] show that
Majorons could alleviate the Hubble constant tension: as these particles contribute
naturally to the dark radiation component of the universe, with .ΔNeff ∼ 0.11, they
are able to reduce by a significant level the . H0 discrepancy. The combined fit of
CMB, BAO and . H0 measurements improves by .Δχ 2 = −12.2 [179] when consid-
ering the self-interacting scenario with respect to the standard .ɅCDM model, see
also Refs. [180–182], although the CMB .χ 2 alone is slightly worsened. Exploiting
the very same mechanism, i.e. a very light scalar Majoron coupled to neutrinos, the
authors of Ref. [183] found a better agreement among high redshift estimates and
local, direct measurements of . H0 .

24.4.3 Self-interacting Sterile Neutrinos

Another class of self-interacting scenarios involves the coupling between a sterile

neutrino and a new pseudo-scalar particle, see e.g. [42] and references therein. This
secret interaction would induce a large matter potential that suppresses active-sterile
oscillations in the early universe. During the universe expansion, when the matter
potential becomes similar to the vacuum oscillation frequency, the sterile neutrino
may encounter a resonance [184]. Depending on the mass of the mediator, the reso-
nance can occur at different times. If the new particle is much lighter than the sterile
neutrino, the resonance occurs much later than BBN, leaving the abundances unaf-
fected, and the sterile state is produced through vacuum neutrino oscillations. At
some point of the evolution, moreover, the sterile neutrino can decay into the new
mediator and the mass bounds are avoided [185–187]. If the mediator is heavier than
the MeV scale [188, 189], instead, the sterile neutrino might be produced through a
resonant process. If this happens before BBN, the light element abundances would
be significantly altered. If the sterile neutrino production occurs after BBN, even
24 On the Dark Radiation Role in the Hubble Constant Tension 469

ignoring the resonant mechanism, a copious abundance of sterile neutrinos is gener-

ated by active-sterile conversions. In such case, the scenario would be disfavored by
cosmological bounds on the neutrino energy density.
When considering a light pseudoscalar, the most recent analyses performed
with Planck data [190] show that the self-interacting model naturally corresponds
to a high value for . H0 , namely . H0 = 71.6+1.1−1.6 km s
−1
Mpc−1 , which is very
close to the local observed value. This result corresponds to a slightly better fit
(.Δχ 2 = −1.0) of PlanckTTTEEE + . H0 constraints than a simple sterile neutrino
scenario, although the CMB .χ 2 alone is slightly worsened with respect to the
non-interacting case. When fitting PlanckTTTEEE+lensing+BAO data, the model
+0.7
allows to obtain . H0 = 70.0−1.1 km s−1 Mpc−1 , which is still much higher than
the CMB prediction in the .ɅCDM model. When considering high-multipoles
CMB data, the light pseudoscalar scenario is instead preferred over the .ɅCDM
model by ACT data alone (.Δχ 2 = −5.3), moreover with a preferred value of
+2.0 −1
. H0 = 72.7−2.8 km s Mpc−1 , although the combination of Planck+ACT prefers
the simpler .ɅCDM model with a .Δχ 2 = 17.4 against the pseudoscalar case and a
lower value of . H0 = 70.6+0.7
−0.9 km s
−1
Mpc−1 [191].
Concerning secret sterile neutrino interactions with a heavy mediator, the situation
for the . H0 tension is instead worsened [184]. When the sterile neutrino mass is
enforced to obey SBL constraints and be close to 1 eV, the study reports indeed
that . H0 is even smaller in the self-interacting scenario than in the .ɅCDM model,
because of the lower value of Neff enforced by the interactions and the high value
of .m s . Since higher neutrino masses are correlated with lower . H0 values and a
small Neff also corresponds to a decrease in . H0 , both effects drive . H0 even down to
−1
.59.56 ± 0.88 km s Mpc−1 [184], which of course increases the already strong . H0
tension. This class of model is therefore ruled out by cosmological constraints on the
sum of neutrino masses and on free-streaming of active neutrinos, unless additional
phenomena are considered to evade the neutrino mass bound, for example a fast
decay mode from the sterile state into active neutrinos [192].

24.5 Summary

Dark radiation models provide a natural environment where to explain many theo-
retical (QCD strong CP problem, origin of neutrino masses) and observational (short
baseline neutrino oscillation anomalies) open problems in the Standard Model of
elementary particles. Some of these new particles can also play a very important role
in astrophysics and cosmology, e.g. contributing to the hot dark matter component of
the universe, stellar evolution and more. The question we have explored throughout
this review is whether the extra dark radiation component can also play a relevant
role in resolving the so-called Hubble constant tension.
The standard and commonly exploited way of parameterizing the mass-energy
density in the new sector makes use of the Neff parameter, the effective number
of relativistic degrees of freedom. A larger value of Neff can be obtained in very
470 S. Gariazzo and O. Mena

simple models, such as those with sterile neutrinos or with non-vanishing neutrino
asymmetries. However, to solve the . H0 tension, more elaborated models leading to a
non-zero .ΔNeff are required. Interacting neutrino scenarios with a Majoron or with
extra sterile species are examples in which the Hubble constant tension is alleviated.
In addition, some of these models are also able to explain some of the aforementioned
open questions within the SM, and therefore should be regarded as very appealing
scenarios, both from the theoretical and observational perspectives. Upcoming data,
not only from future cosmological probes, but also from laboratory, man-made par-
ticle beams and/or from astroparticle physics searches (neutrino, gamma-ray and
cosmic ray telescopes in the case of axions and Majorons, for instance) may shed
light on the role of extra dark radiation and the Hubble constant problem.

Acknowledgements This work has been supported by the MCIN/AEI/10.13039/501100011033

of Spain under grant PID2020-113644GB-I00, by the Generalitat Valenciana of Spain under grant
PROMETEO/2019/083 and by the European Union’s Framework Programme for Research and
Innovation Horizon 2020 (2014-2020) under grant agreement 754496 (FELLINI) and 860881 (HID-
DeN).

References

1. A.G. Riess et al., A comprehensive measurement of the local value of the Hubble
constant with 1 km s.−1 Mpc.−1 uncertainty from the Hubble space telescope and the
S. H0 ES team. Astrophys. J. Lett. 934, L7 (2022). https://doi.org/10.3847/2041-8213/ac5c5b
[https://arxiv.org/abs/2112.04510 2112.04510]
2. Planck collaboration, Planck 2018 results. VI. Cosmological parameters.
Astron. Astrophys. 641, A6 (2020). https://doi.org/10.1051/0004-6361/201833910
[https://arxiv.org/abs/1807.06209 1807.06209]
3. L. Verde, T. Treu, A. Riess, Tensions Between the Early and the Late Universe, vol. 3,
p. 891 (2019). https://doi.org/10.1038/s41550-019-0902-0 [https://arxiv.org/abs/1907.10625
1907.10625]
4. E. Di Valentino et al., Snowmass2021—letter of interest cosmology intertwined II: the
Hubble constant tension. Astropart. Phys. 131, 102605 (2021). https://doi.org/10.1016/j.
astropartphys.2021.102605 [https://arxiv.org/abs/2008.11284 2008.11284]
5. E. Di Valentino et al., In the realm of the Hubble tension—a review of solu-
tions. Class. Quant. Grav. 38, 153001 (2021). https://doi.org/10.1088/1361-6382/ac086d
[https://arxiv.org/abs/2103.01183 2103.01183]
6. N. Schöneberg et al., The H.0 Olympics: a fair ranking of proposed mod-
els. Phys. Rep. 984, 2228 (2022). https://doi.org/10.1016/j.physrep.2022.07.001
[https://arxiv.org/abs/2107.10291 2107.10291]
7. K. Akita, M. Yamaguchi, A precision calculation of relic neutrino decoupling. JCAP 08, 012
(2020). https://doi.org/10.1088/1475-7516/2020/08/012 [https://arxiv.org/abs/2005.07047
005.07047]
8. J. Froustey, C. Pitrou, M.C. Volpe, Neutrino decoupling including flavour oscillations and
primordial nucleosynthesis. JCAP 12, 015 (2020). https://doi.org/10.1088/1475-7516/2020/
12/015 [https://arxiv.org/abs/2008.01074 2008.01074]
9. J.J. Bennett et al., Towards a precision calculation of . Neff in the standard model II: neutrino
decoupling in the presence of flavour oscillations and finite-temperature QED. JCAP 04, 073
24 On the Dark Radiation Role in the Hubble Constant Tension 471

(2021). https://doi.org/10.1088/1475-7516/2021/04/073 [https://arxiv.org/abs/2012.02726

2012.02726]
10. M. Cielo, M. Escudero, G. Mangano, O. Pisanti, Neff in the standard model at NLO is 3.043.
https://arxiv.org/abs/2306.05460 2306.05460
11. G. Mangano, G. Miele, S. Pastor, T. Pinto, O. Pisanti, P.D. Serpico, Relic neutrino decou-
pling including flavor oscillations. Nucl. Phys. B729, 221 (2005). https://doi.org/10.1016/j.
nuclphysb.2005.09.041 [https://arxiv.org/abs/hep-ph/0506164 hep-ph/0506164]
12. P.F. de Salas, S. Pastor, Relic neutrino decoupling with flavour oscillations
revisited. JCAP 07, 051 (2016). https://doi.org/10.1088/1475-7516/2016/07/051
[https://arxiv.org/abs/1606.06986 1606.06986]
13. Y. Farzan, M. Tortola, Neutrino oscillations and non-standard interactions. Front. Phys.
6, 10 (2018). https://doi.org/10.3389/fphy.2018.00010 [https://arxiv.org/abs/1710.09360
1710.09360]
14. Y. Du, J.-H. Yu, Neutrino non-standard interactions meet precision measure-
ments of N.e f f . JHEP 05, 058 (2021). https://doi.org/10.1007/JHEP05(2021)058
[https://arxiv.org/abs/2101.10475 2101.10475]
15. P.F. de Salas, S. Gariazzo, P. Martínez-Miravé, S. Pastor, M.Tórtola, Cosmological radi-
ation density with non-standard neutrino-electron interactions. Phys. Lett. B 820, 136508
(2021). https://doi.org/10.1016/j.physletb.2021.136508 [https://arxiv.org/abs/2105.08168
2105.08168]
16. S. Gariazzo, P. Martínez-Miravé, O. Mena, S. Pastor, M. Tórtola, Non-unitary three-neutrino
mixing in the early universe. JCAP 03, 046 (2023). https://doi.org/10.1088/1475-7516/2023/
03/046 [https://arxiv.org/abs/2211.10522 2211.10522]
17. Simons observatory collaboration, The Simons observatory: science goals and
forecasts. JCAP 02, 056 (2019). https://doi.org/10.1088/1475-7516/2019/02/056
[https://arxiv.org/abs/1808.07445 1808.07445]
18. CMB-S4 collaboration, CMB-S4 Science Book, 1st edn. https://arxiv.org/abs/1610.02743
1610.02743
19. M. Archidiacono, E. Giusarma, S. Hannestad, O. Mena, Cosmic dark radiation and neu-
trinos. Adv. High Energy Phys. 2013, 191047 (2013). https://doi.org/10.1155/2013/191047
[https://arxiv.org/abs/1307.0637 1307.0637]
20. O. Pisanti et al., PArthENoPE: public algorithm evaluating the nucleosynthesis of primordial
elements. Comput. Phys. Commun. 178, 956 (2008). https://doi.org/10.1016/j.cpc.2008.02.
015 [https://arxiv.org/abs/0705.0290 0705.0290]
21. R. Consiglio, P.F. de Salas, G. Mangano, G. Miele, S. Pastor, O. Pisanti, PArthENoPE
reloaded. Comput. Phys. Commun. 233, 237 (2018). https://doi.org/10.1016/j.cpc.2018.06.
022 [https://arxiv.org/abs/1712.04378 1712.04378]
22. S. Gariazzo, P.F. de Salas, O. Pisanti, R. Consiglio, PArthENoPE revolutions. Com-
put. Phys. Commun. 271, 108205 (2022). https://doi.org/10.1016/j.cpc.2021.108205
[https://arxiv.org/abs/2103.05027 2103.05027]
23. C. Pitrou, A. Coc, J.-P. Uzan, E. Vangioni, A new tension in the cosmological model from
primordial deuterium? Mon. Not. Roy. Astron. Soc. 502, 2474 (2021). https://doi.org/10.
1093/mnras/stab135 [https://arxiv.org/abs/2011.11320 2011.11320]
24. V. Mossa et al., The Baryon density of the Universe from an improved rate of deuterium
burning. Nature 587, 210 (2020). https://doi.org/10.1038/s41586-020-2878-4
25. O. Pisanti, G. Mangano, G. Miele, P. Mazzella, Primordial deuterium after LUNA: concor-
dances and error budget. JCAP 04, 020 (2021). https://doi.org/10.1088/1475-7516/2021/04/
020 [https://arxiv.org/abs/2011.11537 2011.11537]
26. T.-H. Yeh, K.A. Olive, B.D. Fields, The impact of new .d( p, γ )3 rates on Big
Bang nucleosynthesis. JCAP 03, 046 (2021). https://doi.org/10.1088/1475-7516/2021/03/
046[https://arxiv.org/abs/2011.13874 2011.13874]
27. P.D. Serpico et al., Nuclear reaction network for primordial nucleosynthesis: a detailed analy-
sis of rates, uncertainties and light nuclei yields. JCAP 12, 010 (2004). https://doi.org/10.1088/
1475-7516/2004/12/010 [https://arxiv.org/abs/astro-ph/0408076 astro-ph/0408076]
472 S. Gariazzo and O. Mena

28. Z. Hou, R. Keisler, L. Knox, M. Millea, C. Reichardt, How Massless neutrinos affect the
cosmic microwave background damping tail. Phys. Rev. D 87, 083008 (2013). https://doi.
org/10.1103/PhysRevD.87.083008 [https://arxiv.org/abs/1104.2333 1104.2333]
29. S. Bashinsky, U. Seljak, Neutrino perturbations in CMB anisotropy and matter clus-
tering. Phys. Rev. D 69, 083002 (2004). https://doi.org/10.1103/PhysRevD.69.083002
[https://arxiv.org/abs/astro-ph/0310198 astro-ph/0310198]
30. S. Hannestad, Structure formation with strongly interacting neutrinos–implications for the
cosmological neutrino mass bound. JCAP 02, 011 (2005). https://doi.org/10.1088/1475-7516/
2005/02/011
31. W. Hu, N. Sugiyama, Small scale cosmological perturbations: an analytic approach.
Astrophys. J. 471, 542 (1996). https://doi.org/10.1086/177989 [https://arxiv.org/abs/astro-
ph/9510117 astro-ph/9510117]
32. S.F. King, Neutrino Mass and Mixing in the Seesaw Playground, vol. 908, pp. 456–466 (2016).
https://doi.org/10.1016/j.nuclphysb.2015.12.015DOI [https://arxiv.org/abs/1511.03831
1511.03831]
33. O.G. Miranda, J.W.F. Valle, Neutrino oscillations and the seesaw origin of neutrino
mass. Nucl. Phys. B908, 436 (2016). https://doi.org/10.1016/j.nuclphysb.2016.03.027
https://arxiv.org/abs/1602.00864 1602.00864]
34. Z.-Z. Xing, Z.-H. Zhao, The minimal seesaw and leptogenesis models. Rep. Prog. Phys. 84,
066201 (2021). https://doi.org/10.1088/1361-6633/abf086 [https://arxiv.org/abs/2008.12090
2008.12090]
35. A. Boyarsky, M. Drewes, T. Lasserre, S. Mertens, O. Ruchayskiy, Sterile neutrino dark
matter. Prog. Part. Nucl. Phys. 104, 1 (2019). https://doi.org/10.1016/j.ppnp.2018.07.004
[https://arxiv.org/abs/1807.07938 1807.07938]
36. R. Adhikari et al., A white paper on keV sterile neutrino dark matter. JCAP 01, 025
(2017). https://doi.org/10.1088/1475-7516/2017/01/025 [https://arxiv.org/abs/1602.04816
1602.04816]
37. G. Anamiati, V. De Romeri, M. Hirsch, C.A. Ternes, M. Tórtola, Quasi-Dirac neutrino oscil-
lations at DUNE and JUNO. Phys. Rev. D 100, 035032 (2019). https://doi.org/10.1103/
PhysRevD.100.035032 [https://arxiv.org/abs/1907.00980 1907.00980]
38. K. Carloni, I. Martinez-Soler, C.A. Arguelles, K. Babu, P.B. Dev, Probing pseudo-dirac neu-
trinos with astrophysical sources at IceCube. https://arxiv.org/abs/2212.00737 2212.00737
39. S. Gariazzo, C. Giunti, M. Laveder, Y.F. Li, E.M. Zavanin, Light sterile neutri-
nos. J. Phys. G 43, 033001 (2016). https://doi.org/10.1088/0954-3899/43/3/033001
[https://arxiv.org/abs/1507.08204 1507.08204]
40. C. Giunti, T. Lasserre, eV-scale sterile neutrinos. Ann. Rev. Nucl. Part. Sci. 69, 163 (2019).
https://doi.org/10.1146/annurev-nucl-101918-023755 [https://arxiv.org/abs/1901.08330
1901.08330]
41. S. Böser et al., Status of light sterile neutrino searches. Prog. Part. Nucl. Phys. 111, 103736
(2020). https://doi.org/10.1016/j.ppnp.2019.103736 [https://arxiv.org/abs/1906.01739
1906.01739]
42. M. Archidiacono, S. Gariazzo, Two sides of the same coin: sterile neutrinos and dark radiation,
status and perspectives. Universe 8, 175 (2022). https://doi.org/10.3390/universe8030175
[https://arxiv.org/abs/2201.10319 2201.10319]
43. LSND collaboration, Evidence for neutrino oscillations from the observation of.ν̄e appearance
in a.ν̄μ beam. Phys. Rev. D 64, 112007 (2001). https://doi.org/10.1103/PhysRevD.64.112007
[https://arxiv.org/abs/hep-ex/0104049 hep-ex/0104049]
44. MiniBooNE collaboration, Unexplained excess of electron-like events from a 1-GeV neutrino
beam. Phys. Rev. Lett. 102, 101802 (2009). https://doi.org/10.1103/PhysRevLett.102.101802
[https://arxiv.org/abs/0812.2243 0812.2243]
45. MiniBooNE collaboration, Significant excess of electron like events in the MiniBooNE short-
baseline neutrino experiment. Phys. Rev. Lett. 121, 221801 (2018). https://doi.org/10.1103/
PhysRevLett.121.221801 [https://arxiv.org/abs/1805.12028 1805.12028]
24 On the Dark Radiation Role in the Hubble Constant Tension 473

46. J. Abdurashitov et al., Measurement of the response of a Ga solar neutrino experiment to

neutrinos from an Ar-37 source. Phys. Rev. C 73, 045805 (2006). https://doi.org/10.1103/
PhysRevC.73.045805 [https://arxiv.org/abs/nucl-ex/0512041 nucl-ex/0512041]
47. C. Giunti, M. Laveder, Statistical significance of the gallium anomaly. Phys. Rev. C 83, 065504
(2011). https://doi.org/10.1103/PhysRevC.83.065504 [https://arxiv.org/abs/1006.3244
1006.3244]
48. G. Mention, M. Fechner, T. Lasserre, T.A. Mueller, D. Lhuillier, M. Cribier et al., The reactor
antineutrino anomaly. Phys. Rev. D 83, 073006 (2011). https://doi.org/10.1103/PhysRevD.
83.073006 [https://arxiv.org/abs/1101.2755 1101.2755]
49. KARMEN collaboration, Upper limits for neutrino oscillations muon-anti-neutrino
.→ electron-anti-neutrino from muon decay at rest. Phys. Rev. D 65, 112001
(2002). https://doi.org/10.1103/PhysRevD.65.112001 [https://arxiv.org/abs/hep-ex/0203021
hep-ex/0203021]
50. OPERA collaboration, Final results of the search for .νμ → νe oscillations with the OPERA
detector in the CNGS beam. JHEP 06, 151 (2018). https://doi.org/10.1007/JHEP06(2018)151
[https://arxiv.org/abs/1803.11400 1803.11400]
51. MicroBooNE collaboration, Search for neutrino-induced neutral-current .Δ radiative decay
in MicroBooNE and a first test of the MiniBooNE low energy excess under a single-photon
hypothesis. Phys. Rev. Lett. 128, 111801 (2022). https://doi.org/10.1103/PhysRevLett.128.
111801 [https://arxiv.org/abs/2110.00409 2110.00409]
52. MicroBooNE collaboration, First constraints on light sterile neutrino oscillations from
combined appearance and disappearance searches with the MicroBooNE detector.
Phys. Rev. Lett. 130, 011801 (2023). https://doi.org/10.1103/PhysRevLett.130.011801
[https://arxiv.org/abs/2210.10216 2210.10216]
53. C. Giunti, Y. Li, C. Ternes, O. Tyagi, Z. Xin, Gallium anomaly: critical view from the
global picture of .νe and .ν e disappearance. JHEP 10, 164 (2022). https://doi.org/10.1007/
JHEP10(2022)164 [https://arxiv.org/abs/2209.00916 2209.00916]
54. C. Giunti, Y. Li, C. Ternes, Z. Xin, Inspection of the detection cross section dependence of
the Gallium anomaly. Phys. Lett. B 842, 137983 (2023). https://doi.org/10.1016/j.physletb.
2023.137983 [https://arxiv.org/abs/2212.09722 2212.09722]
55. V. Barinov et al., Results from the Baksan experiment on sterile transitions (BEST).
Phys. Rev. Lett. 128, 232501 (2022). https://doi.org/10.1103/PhysRevLett.128.232501
[https://arxiv.org/abs/2109.11482 2109.11482]
56. V. Barinov et al., Search for electron-neutrino transitions to sterile states in the BEST
experiment. Phys. Rev. C 105, 065502 (2022). https://doi.org/10.1103/PhysRevC.105.065502
[https://arxiv.org/abs/2201.07364 2201.07364]
57. S.R. Elliott, V. Gavrin, W. Haxton, The gallium anomaly. https://arxiv.org/abs/2306.03299
2306.03299
58. STEREO collaboration, Improved sterile neutrino constraints from the STEREO experiment
with 179 days of reactor-on data. Phys. Rev. D 102, 052002 (2020). https://doi.org/10.1103/
PhysRevD.102.052002 [https://arxiv.org/abs/1912.06582 1912.06582]
59. PROSPECT collaboration, Improved short-baseline neutrino oscillation search and energy
spectrum measurement with the PROSPECT experiment at HFIR. Phys. Rev. D 103, 032001
(2021). https://doi.org/10.1103/PhysRevD.103.032001 [https://arxiv.org/abs/2006.11210
2006.11210]
60. DANSS collaboration, Search for sterile neutrinos at the DANSS experiment.
Phys. Lett. B 787, 56 (2018). https://doi.org/10.1016/j.physletb.2018.10.038
[https://arxiv.org/abs/1804.04046 1804.04046]
61. RENO, NEOS collaboration, Search for sterile neutrino oscillations using RENO and NEOS
data. Phys. Rev. D 105, L111101 (2022). https://doi.org/10.1103/PhysRevD.105.L111101
https://arxiv.org/abs/2011.00896 2011.00896]
62. DANSS collaboration, Recent Results from DANSS, vol. NOW2022, p. 017 (2023). https://
doi.org/10.22323/1.421.0017 [https://arxiv.org/abs/2305.07417 2305.07417]
474 S. Gariazzo and O. Mena

63. NEOS collaboration, Sterile neutrino search at the NEOS experiment. Phys.
Rev. Lett. 118, 121802 (2017). https://doi.org/10.1103/PhysRevLett.118.121802
[https://arxiv.org/abs/1610.05134 1610.05134]
64. A. Serebrov, R. Samoilov, M. Chaikovskii, O. Zherebtsov, The result of the Neutrino-4
experiment, sterile neutrinos and dark matter, the fourth neutrino and the Hubble constant.
https://arxiv.org/abs/2302.09958 2302.09958
65. C. Giunti, Y. Li, C. Ternes, Y. Zhang, Neutrino-4 anomaly: oscillations or fluctua-
tions? Phys. Lett. B 816, 136214 (2021). https://doi.org/10.1016/j.physletb.2021.136214
[https://arxiv.org/abs/2101.06785 2101.06785]
66. M. Estienne et al., Updated summation model: an improved agreement with the Daya
Bay antineutrino fluxes. Phys. Rev. Lett. 123, 022502 (2019). https://doi.org/10.1103/
PhysRevLett.123.022502 [https://arxiv.org/abs/1904.09358 1904.09358]
67. V. Kopeikin, Y.N. Panin, A. Sabelnikov, Measurement of the Ratio of Cumulative Spectra
of Beta Particles from .235 U and .239 Pu Fission Products for Solving Problems of Reactor-
Antineutrino Physics, vol. 84, pp. 1–10 (2021). https://doi.org/10.1134/S1063778821010129
68. L. Perissé et al., A comprehensive revision of the summation method for the prediction of
reactor antineutrino fluxes and spectra. https://arxiv.org/abs/2304.14992 2304.14992
69. T.A. Mueller et al., Improved predictions of reactor antineutrino spectra.
Phys. Rev. C 83, 054615 (2011). https://doi.org/10.1103/PhysRevC.83.054615
https://arxiv.org/abs/1101.2663 1101.2663]
70. P. Huber, On the determination of anti-neutrino spectra from nuclear reactors.
Phys. Rev. C 84, 024617 (2011). https://doi.org/10.1103/PhysRevC.84.024617
[https://arxiv.org/abs/1106.0687 1106.0687]
71. C. Giunti, Y. Li, C. Ternes, Z. Xin, Reactor antineutrino anomaly in light of recent flux
model refinements. Phys. Lett. B 829, 137054 (2022). https://doi.org/10.1016/j.physletb.
2022.137054 [https://arxiv.org/abs/2110.06820 2110.06820]
72. MINOS+ collaboration, Search for sterile neutrinos in MINOS and MINOS+ using a two-
detector fit. Phys. Rev. Lett. 122, 091803 (2019). https://doi.org/10.1103/PhysRevLett.122.
091803 [https://arxiv.org/abs/1710.06488 1710.06488]
73. IceCube collaboration, eV-scale sterile neutrino search using eight years of atmospheric
muon neutrino data from the IceCube neutrino observatory. Phys. Rev. Lett. 125, 141801
(2020). https://doi.org/10.1103/PhysRevLett.125.141801 [https://arxiv.org/abs/2005.12942
2005.12942]
74. S. Gariazzo, C. Giunti, M. Laveder, Y.F. Li, Updated global 3+1 analysis of short-BaseLine
neutrino oscillations. JHEP 06, 135 (2017). https://doi.org/10.1007/JHEP06(2017)135
[https://arxiv.org/abs/1703.00860 1703.00860]
75. M. Dentler et al., Updated global analysis of neutrino oscillations in the presence of eV-
scale sterile neutrinos. JHEP 08, 010 (2018). https://doi.org/10.1007/JHEP08(2018)010
[https://arxiv.org/abs/1803.10661 1803.10661]
76. R. Barbieri, A. Dolgov, Bounds on sterile-neutrinos from nucleosynthesis. Phys. Lett. B 237,
440 (1990). https://doi.org/10.1016/0370-2693(90)91203-N
77. R. Barbieri, A. Dolgov, Neutrino oscillations in the early universe. Nucl. Phys. B 349, 743
(1991). https://doi.org/10.1016/0550-3213(91)90396-F
78. K. Kainulainen, Light singlet neutrinos and the primordial nucleosynthesis. Phys. Lett. B 244,
191 (1990). https://doi.org/10.1016/0370-2693(90)90054-A
79. K. Enqvist, K. Kainulainen, J. Maalampi, Resonant neutrino transitions and nucleosynthesis.
Phys. Lett. B 249, 531 (1990). https://doi.org/10.1016/0370-2693(90)91030-F
80. K. Enqvist, K. Kainulainen, M.J. Thomson, Stringent cosmological bounds on inert neutrino
mixing. Nucl. Phys. B 373, 498 (1992). https://doi.org/10.1016/0550-3213(92)90442-E
81. A.D. Dolgov, F.L. Villante, BBN bounds on active sterile neutrino mixing. Nucl. Phys. B679,
261 (2004). https://doi.org/10.1016/j.nuclphysb.2003.11.031 [https://arxiv.org/abs/hep-
ph/0308083 hep-ph/0308083]
82. M. Cirelli, G. Marandella, A. Strumia, F. Vissani, Probing oscillations into ster-
ile neutrinos with cosmology, astrophysics and experiments. Nucl. Phys. B 708,
24 On the Dark Radiation Role in the Hubble Constant Tension 475

215 (2005). https://doi.org/10.1016/j.nuclphysb.2004.11.056 [https://arxiv.org/abs/hep-

ph/0403158 hep-ph/0403158]
83. A.D. Dolgov, Neutrinos in cosmology. Phys. Rep. 370, 333 (2002). https://doi.org/10.1016/
S0370-1573(02)00139-4 [https://arxiv.org/abs/hep-ph/0202122 hep-ph/0202122]
84. S. Hannestad, I. Tamborra, T. Tram, Thermalisation of light sterile neutrinos in the
early universe. JCAP 07, 025 (2012)[. https://doi.org/10.1088/1475-7516/2012/07/025
https://arxiv.org/abs/1204.5861 1204.5861]
85. S. Hannestad, R.S. Hansen, T. Tram, Can active-sterile neutrino oscillations lead to chaotic
behavior of the cosmological lepton asymmetry? JCAP 04, 032 (2013). https://doi.org/10.
1088/1475-7516/2013/04/032 [https://arxiv.org/abs/1302.7279 1302.7279]
86. S. Hannestad, R.S. Hansen, T. Tram, Y.Y.Y. Wong, Active-sterile neutrino oscillations in the
early universe with full collision terms. JCAP 08, 019 (2015). https://doi.org/10.1088/1475-
7516/2015/08/019 https://arxiv.org/abs/1506.05266 1506.05266]
87. A. Mirizzi, N. Saviano, G. Miele, P.D. Serpico, Light sterile neutrino production in the early
universe with dynamical neutrino asymmetries. Phys. Rev. D 86, 053009 (2012). https://doi.
org/10.1103/PhysRevD.86.053009 [https://arxiv.org/abs/1206.1046 1206.1046]
88. A. Mirizzi, G. Mangano, N. Saviano, E. Borriello, C. Giunti, G. Miele et al., The strongest
bounds on active-sterile neutrino mixing after Planck data. Phys. Lett. B726, 8 (2013). https://
doi.org/10.1016/j.physletb.2013.08.015 [https://arxiv.org/abs/1303.5368 1303.5368]
89. N. Saviano, A. Mirizzi, O. Pisanti, P.D. Serpico, G. Mangano, G. Miele, Multi-momentum
and multi-flavour active-sterile neutrino oscillations in the early universe: role of neutrino
asymmetries and effects on nucleosynthesis. Phys. Rev. D 87, 073006 (2013). https://doi.org/
10.1103/PhysRevD.87.073006 [ https://arxiv.org/abs/1302.1200 1302.1200]
90. S. Gariazzo, P.F. de Salas, S. Pastor, Thermalisation of sterile neutrinos in the early universe
in the 3+1 scheme with full mixing matrix. JCAP 07, 014 (2019). https://doi.org/10.1088/
1475-7516/2019/07/014 [ https://arxiv.org/abs/1905.11290 1905.11290]
91. L. Mastrototaro, P.D. Serpico, A. Mirizzi, N. Saviano, Massive sterile neutrinos in the early
Universe: from thermal decoupling to cosmological constraints. Phys. Rev .D 104, 016026
(2021). https://doi.org/10.1103/PhysRevD.104.016026 [ https://arxiv.org/abs/2104.11752
2104.11752]
92. S. Dodelson, L.M. Widrow, Sterile-neutrinos as dark matter. Phys. Rev. Lett. 72,
17 (1994). https://doi.org/10.1103/PhysRevLett.72.17 https://arxiv.org/abs/hep-ph/9303287
hep-ph/9303287]
93. E. Di Valentino, S. Gariazzo, W. Giarè, O. Mena, Weighing neutrinos at the damping
tail.https://arxiv.org/abs/2305.12989 2305.12989
94. B.D. Fields, K.A. Olive, T.-H. Yeh, C. Young, Big-bang nucleosynthesis after
Planck. JCAP 03, 010 (2020). https://doi.org/10.1088/1475-7516/2020/03/010
[https://arxiv.org/abs/1912.01132 1912.01132]
95. Particle Data Group collaboration, Review of particle physics. PTEP 2022, 083C01 (2022).
https://doi.org/10.1093/ptep/ptac097
96. P.-H. Gu, Large Lepton asymmetry for small Baryon asymmetry and warm dark
matter. Phys. Rev. D 82, 093009 (2010). https://doi.org/10.1103/PhysRevD.82.093009
[https://arxiv.org/abs/1005.1632 1005.1632]
97. J. March-Russell, H. Murayama, A. Riotto, The small observed Baryon asymmetry from a
large lepton asymmetry. JHEP 11, 015 (1999). https://doi.org/10.1088/1126-6708/1999/11/
015 [https://arxiv.org/abs/hep-ph/9908396 hep-ph/9908396]
98. J. McDonald, Naturally large cosmological neutrino asymmetries in the MSSM.
Phys. Rev. Lett. 84, 4798 (2000). https://doi.org/10.1103/PhysRevLett.84.4798
[https://arxiv.org/abs/hep-ph/9908300 hep-ph/9908300]
99. J. Froustey, C. Pitrou, Incomplete neutrino decoupling effect on big bang nucleosyn-
thesis. Phys. Rev. D 101, 043524 (2020). https://doi.org/10.1103/PhysRevD.101.043524
[https://arxiv.org/abs/1912.09378 1912.09378]
100. G. Barenboim, W.H. Kinney, W.-I. Park, Flavor versus mass eigenstates in neutrino asym-
metries: implications for cosmology. Eur. Phys. J. C 77, 590 (2017). https://doi.org/10.1140/
epjc/s10052-017-5147-4 [https://arxiv.org/abs/1609.03200 1609.03200]
476 S. Gariazzo and O. Mena

101. G. Mangano, G. Miele, S. Pastor, O. Pisanti, S. Sarikas, Updated BBN bounds on the cos-
mological lepton asymmetry for non-zero .θ13 . Phys. Lett. B 708, 1 (2012). https://doi.org/10.
1016/j.physletb.2012.01.015 [https://arxiv.org/abs/1110.4335 1110.4335]
102. A. Dolgov, S. Hansen, S. Pastor, S. Petcov, G. Raffelt, D. Semikoz, Cosmological bounds on
neutrino degeneracy improved by flavor oscillations. Nucl. Phys. B 632, 363 (2002). https://
doi.org/10.1016/S0550-3213(02)00274-2
103. Y.Y.Y. Wong, Analytical treatment of neutrino asymmetry equilibration from flavor oscil-
lations in the early universe. Phys. Rev. D 66, 025015 (2002). https://doi.org/10.1103/
PhysRevD.66.025015 [https://arxiv.org/abs/hep-ph/0203180 hep-ph/0203180]
104. V. Simha, G. Steigman, Constraining the universal Lepton asymmetry. JCAP 08, 011
(2008). https://doi.org/10.1088/1475-7516/2008/08/011 [https://arxiv.org/abs/0806.0179
0806.0179]
105. I.M. Oldengott, D.J. Schwarz, Improved constraints on lepton asymmetry from the cosmic
microwave background. Europhys. Lett. 119, 29001 (2017). https://doi.org/10.1209/0295-
5075/119/29001 [https://arxiv.org/abs/1706.01705 1706.01705]
106. C. Pitrou, A. Coc, J.-P. Uzan, E. Vangioni, Precision big bang nucleosynthesis with improved
Helium-4 predictions. Phys. Rep. 04, 005 (2018). https://doi.org/10.1016/j.physrep.2018.04.
005 [https://arxiv.org/abs/1801.08023 1801.08023]
107. J. Froustey, C. Pitrou, Primordial neutrino asymmetry evolution with full mean-field
effects and collisions. JCAP 03, 065 (2022). https://doi.org/10.1088/1475-7516/2022/03/
065 [https://arxiv.org/abs/2110.11889 2110.11889]
108. R.D. Peccei, H.R. Quinn, CP conservation in the presence of instantons. Phys. Rev. Lett. 38,
1440 (1977). https://doi.org/10.1103/PhysRevLett.38.1440
109. R.D. Peccei, H.R. Quinn, Constraints imposed by CP conservation in the presence of instan-
tons. Phys. Rev. D 16, 1791 (1977). https://doi.org/10.1103/PhysRevD.16.1791
110. M.S. Turner, Thermal production of not SO invisible axions in the early universe. Phys. Rev.
Lett. 59, 2489 (1987). https://doi.org/10.1103/PhysRevLett.59.2489
111. Z.G. Berezhiani, A.S. Sakharov, M.Y. Khlopov, Primordial background of cosmological
axions. Sov. J. Nucl. Phys. Series 55, 1063 (1992)
112. C. Brust, D.E. Kaplan, M.T. Walters, New light species and the CMB. JHEP 12, 058 (2013).
https://doi.org/10.1007/JHEP12(2013)058 [ https://arxiv.org/abs/1303.5379 1303.5379]
113. D. Baumann, D. Green, B. Wallisch, New target for cosmic axion searches.
Phys. Rev. Lett. 117, 171301 (2016). https://doi.org/10.1103/PhysRevLett.117.171301 [
https://arxiv.org/abs/1604.08614 1604.08614]
114. F. D’Eramo, R.Z. Ferreira, A. Notari, J.L. Bernal, Hot axions and the. H0 tension. JCAP 11, 014
(2018). https://doi.org/10.1088/1475-7516/2018/11/014 [https://arxiv.org/abs/1808.07430
1808.07430]
115. F. Arias-Aragón, F. D’eramo, R.Z. Ferreira, L. Merlo, A. Notari, Cosmic imprints of
XENON1T axions. JCAP 11, 025 (2020). https://doi.org/10.1088/1475-7516/2020/11/025
https://arxiv.org/abs/2007.06579 2007.06579]
116. F. Arias-Aragón, F. D’Eramo, R.Z. Ferreira, L. Merlo, A. Notari, Production of thermal axions
across the ElectroWeak phase transition. JCAP 03, 090 (2021). https://doi.org/10.1088/1475-
7516/2021/03/090 [https://arxiv.org/abs/2012.04736 2012.04736]
117. D. Green, Y. Guo, B. Wallisch, Cosmological implications of axion-matter cou-
plings. JCAP 02, 019 (2022). https://doi.org/10.1088/1475-7516/2022/02/019
[https://arxiv.org/abs/2109.12088 2109.12088]
118. F. D’Eramo, S. Yun, Flavor violating axions in the early universe. Phys. Rev. D 105, 075002
(2022). https://doi.org/10.1103/PhysRevD.105.075002 [https://arxiv.org/abs/2111.12108
2111.12108]
119. S. Hannestad, A. Mirizzi, G. Raffelt, New cosmological mass limit on thermal
relic axions. JCAP 07, 002 (2005). https://doi.org/10.1088/1475-7516/2005/07/002
https://arxiv.org/abs/hep-ph/0504059 hep-ph/0504059]
120. A. Melchiorri, O. Mena, A. Slosar, An improved cosmological bound on the thermal
axion mass. Phys. Rev. D 76, 041303 (2007). https://doi.org/10.1103/PhysRevD.76.041303
[https://arxiv.org/abs/0705.2695 0705.2695]
24 On the Dark Radiation Role in the Hubble Constant Tension 477

121. S. Hannestad, A. Mirizzi, G.G. Raffelt, Y.Y. Wong, Cosmological constraints on neutrino
plus axion hot dark matter. JCAP 08, 015 (2007). https://doi.org/10.1088/1475-7516/2007/
08/015 [https://arxiv.org/abs/0706.4198 0706.4198]
122. S. Hannestad, A. Mirizzi, G.G. Raffelt, Y.Y. Wong, Cosmological constraints on neutrino plus
axion hot dark matter: update after WMAP-5. JCAP 04, 019 (2008). https://doi.org/10.1088/
1475-7516/2008/04/019 [https://arxiv.org/abs/0803.1585 0803.1585]
123. S. Hannestad, A. Mirizzi, G.G. Raffelt, Y.Y. Wong, Neutrino and axion hot dark matter
bounds after WMAP-7. JCAP 08, 001 (2010). https://doi.org/10.1088/1475-7516/2010/08/
001 [https://arxiv.org/abs/1004.0695 1004.0695]
124. M. Archidiacono, S. Hannestad, A. Mirizzi, G. Raffelt, Y.Y. Wong, Axion hot dark matter
bounds after Planck. JCAP 10, 020 (2013). https://doi.org/10.1088/1475-7516/2013/10/020
https://arxiv.org/abs/1307.0615 1307.0615]
125. E. Giusarma, E. Di Valentino, M. Lattanzi, A. Melchiorri, O. Mena, Relic neutrinos, thermal
axions and cosmology in early 2014. Phys. Rev. D 90, 043507 (2014). https://doi.org/10.
1103/PhysRevD.90.043507 [https://arxiv.org/abs/1403.4852 1403.4852]
126. E. Di Valentino, S. Gariazzo, E. Giusarma, O. Mena, Robustness of cosmological axion
mass limits. Phys. Rev. D D 91, 123505 (2015). https://doi.org/10.1103/PhysRevD.91.123505
[https://arxiv.org/abs/1503.00911 1503.00911]
127. E. Di Valentino, E. Giusarma, M. Lattanzi, O. Mena, A. Melchiorri, J. Silk, Cosmolog-
ical Axion and neutrino mass constraints from Planck 2015 temperature and polariza-
tion data. Phys. Lett. B 752, 182 (2016). https://doi.org/10.1016/j.physletb.2015.11.025
https://arxiv.org/abs/1507.08665 1507.08665]
128. M. Archidiacono, T. Basse, J. Hamann, S. Hannestad, G. Raffelt, Y.Y.Y. Wong, Future cosmo-
logical sensitivity for hot dark matter axions. JCAP 05, 050 (2015). https://doi.org/10.1088/
1475-7516/2015/05/050 [https://arxiv.org/abs/1502.03325 1502.03325]
129. W. Giarè, E. Di Valentino, A. Melchiorri, O. Mena, New cosmological bounds on hot relics:
axions and neutrinos. Mon. Not. Roy. Astron. Soc. 505, 2703 (2021). https://doi.org/10.1093/
mnras/stab1442 [https://arxiv.org/abs/2011.14704 2011.14704]
130. F. D’Eramo et al., Cosmological bound on the QCD axion mass, redux. JCAP 09, 022
(2022). https://doi.org/10.1088/1475-7516/2022/09/022 [https://arxiv.org/abs/2205.07849
2205.07849]
131. W. Giarè, F. Renzi, A. Melchiorri, O. Mena, E. Di Valentino, Cosmological forecasts on ther-
mal axions, relic neutrinos, and light elements. Mon. Not. Roy. Astron. Soc. 511, 1373 (2022).
https://doi.org/10.1093/mnras/stac126 [https://arxiv.org/abs/2110.00340 2110.00340]
132. E. Di Valentino et al., Novel model-marginalized cosmological bound on the QCD axion
mass. Phys. Rev. D 107, 103528 (2023). https://doi.org/10.1103/PhysRevD.107.103528
[https://arxiv.org/abs/2212.11926 2212.11926]
133. W.A. Bardeen, S.H.H. Tye, J.A.M. Vermaseren, Phenomenology of the new light higgs Boson
search. Phys. Lett. B 76, 580 (1978). https://doi.org/10.1016/0370-2693(78)90859-6
134. G. Grilli di Cortona, E. Hardy, J. Pardo Vega, G. Villadoro, The QCD axion, precisely. HEP 01,
034 (2016). https://doi.org/10.1007/JHEP01(2016)034 J [https://arxiv.org/abs/1511.02867
1511.02867]
135. M. Gorghetto, G. Villadoro, Topological susceptibility and QCD axion mass: QED
and NNLO corrections. JHEP 03, 033 (2019). https://doi.org/10.1007/JHEP03(2019)033
[https://arxiv.org/abs/1812.01008 1812.01008]
136. J.E. Kim, G. Carosi, Axions and the strong CP problem. Rev. Mod. Phys. 82, 557 (2010).
https://doi.org/10.1103/RevModPhys.82.557 [https://arxiv.org/abs/0807.3125 0807.3125]
137. L. Di Luzio, M. Giannotti, E. Nardi, L. Visinelli, The landscape of QCD
axion models. Phys. Rep. 870, 1 (2020). https://doi.org/10.1016/j.physrep.2020.06.002
[https://arxiv.org/abs/2003.01100 2003.01100]
138. J.E. Kim, Weak interaction singlet and strong CP invariance. Phys. Rev. Lett. 43, 103 (1979).
https://doi.org/10.1103/PhysRevLett.43.103
139. M.A. Shifman, A.I. Vainshtein, V.I. Zakharov, Can confinement ensure natural CP invari-
ance of strong interactions? Nucl. Phys. B 166, 493 (1980). https://doi.org/10.1016/0550-
3213(80)90209-6
478 S. Gariazzo and O. Mena

140. M. Dine, W. Fischler, M. Srednicki, A simple solution to the strong CP problem with a harmless
axion. Phys. Lett. B 104, 199 (1981). https://doi.org/10.1016/0370-2693(81)90590-6
141. A.R. Zhitnitsky, On possible suppression of the axion hadron interactions. Sov. J. Nucl. Phys.
31, 260 (1980). ((in Russian))
142. G. Mangano, A. Melchiorri, P. Serra, A. Cooray, M. Kamionkowski, Cosmological bounds on
dark matter-neutrino interactions. Phys. Rev. D 74, 043517 (2006). https://doi.org/10.1103/
PhysRevD.74.043517 [https://arxiv.org/abs/astro-ph/0606190 astro-ph/0606190]
143. P. Serra, F. Zalamea, A. Cooray, G. Mangano, A. Melchiorri, Constraints on neutrino—
dark matter interactions from cosmic microwave background and large scale struc-
ture data. Phys. Rev. D 81, 043507 (2010). https://doi.org/10.1103/PhysRevD.81.043507
[https://arxiv.org/abs/0911.4411 0911.4411]
144. R. Diamanti, E. Giusarma, O. Mena, M. Archidiacono, A. Melchiorri, Dark radiation and
interacting scenarios. Phys. Rev. D 87, 063509 (2013). https://doi.org/10.1103/PhysRevD.
87.063509 [https://arxiv.org/abs/1212.6007 1212.6007]
145. R.J. Wilkinson, C. Boehm, J. Lesgourgues, Constraining dark matter-neutrino interactions
using the CMB and large-scale structure. JCAP 05, 011 (2014). https://doi.org/10.1088/1475-
7516/2014/05/011 [https://arxiv.org/abs/1401.7597 1401.7597]
146. A. Olivares-Del Campo, C. Bœhm, S. Palomares-Ruiz, S. Pascoli, Dark matter-neutrino
interactions through the lens of their cosmological implications. Phys. Rev. D 97, 075039
(2018). https://doi.org/10.1103/PhysRevD.97.075039 [https://arxiv.org/abs/1711.05283
1711.05283]
147. J. Stadler, C. Bœhm, O. Mena, First numerical study of neutrino-dark matter
mixed damping. JCAP 08, 014 (2019). https://doi.org/10.1088/1475-7516/2019/08/014
[https://arxiv.org/abs/1903.00540 1903.00540]
148. M.R. Mosbech et al., The full Boltzmann hierarchy for dark matter-massive neu-
trino interactions. JCAP 03, 066 (2021). https://doi.org/10.1088/1475-7516/2021/03/066
[https://arxiv.org/abs/2011.04206 2011.04206]
149. M. Escudero, Neutrino decoupling beyond the standard model: CMB constraints on the dark
matter mass with a fast and precise . Neff evaluation. JCAP 02, 007 (2019). https://doi.org/10.
1088/1475-7516/2019/02/007 [https://arxiv.org/abs/1812.05605 1812.05605]
150. M. Escudero, Precision early universe thermodynamics made simple:. Neff and neutrino decou-
pling in the standard model and beyond. JCAP 05, 048 (2020). https://doi.org/10.1088/1475-
7516/2020/05/048 https://arxiv.org/abs/2001.04466 2001.04466]
151. C. Boehm, M.J. Dolan, C. McCabe, Increasing Neff with particles in thermal equilib-
rium with neutrinos. JCAP 12, 027 (2012). https://doi.org/10.1088/1475-7516/2012/12/027
[https://arxiv.org/abs/1207.0497 1207.0497]
152. C. Boehm, M.J. Dolan, C. McCabe, A lower bound on the mass of cold thermal dark
matter from Planck. JCAP 08, 041 (2013). https://doi.org/10.1088/1475-7516/2013/08/041
[https://arxiv.org/abs/1303.6270 1303.6270]
153. E. Di Valentino, C.B.e. hm, E. Hivon, F.R. Bouchet, Reducing the . H0 and .σ8 tensions with
dark matter-neutrino interactions. Phys. Rev. D 97, 043513 (2018). https://doi.org/10.1103/
PhysRevD.97.043513 [https://arxiv.org/abs/1710.02559 1710.02559]
154. B. Follin, L. Knox, M. Millea, Z. Pan, First detection of the acoustic oscillation phase shift
expected from the cosmic neutrino background. Phys. Rev. Lett. 115, 091301 (2015). https://
doi.org/10.1103/PhysRevLett.115.091301 [https://arxiv.org/abs/1503.07863 1503.07863]
155. D. Baumann, D. Green, J. Meyers, B. Wallisch, Phases of new physics in
the CMB. JCAP 01, 007 (2016). https://doi.org/10.1088/1475-7516/2016/01/007
[https://arxiv.org/abs/1508.06342 1508.06342]
156. D. Baumann, D. Green, M. Zaldarriaga, Phases of new physics in the BAO
spectrum. JCAP 11, 007 (2017). https://doi.org/10.1088/1475-7516/2017/11/007
[https://arxiv.org/abs/1703.00894 1703.00894]
157. G. Choi, C.-T. Chiang, M. LoVerde, Probing decoupling in dark sectors with the cosmic
microwave background. JCAP 06, 044 (2018). https://doi.org/10.1088/1475-7516/2018/06/
044 [https://arxiv.org/abs/1804.10180 1804.10180]
24 On the Dark Radiation Role in the Hubble Constant Tension 479

158. J.M. Berryman et al., Neutrino Self-interactions: A White Paper, vol. 42, p. 101267,
3 (2023). https://doi.org/10.1016/j.dark.2023.101267 [https://arxiv.org/abs/2203.01955
2203.01955]
159. P. Taule, M. Escudero, M. Garny, Global view of neutrino interactions in cosmology: The
free streaming window as seen by Planck. Phys. Rev. D 106, 063539 (2022). https://doi.org/
10.1103/PhysRevD.106.063539 [https://arxiv.org/abs/2207.04062 2207.04062]
160. C.D. Kreisch, F.-Y. Cyr-Racine, O. Doré, Neutrino puzzle: anomalies, interactions, and cos-
mological tensions. Phys. Rev. D 101, 123505 (2020). https://doi.org/10.1103/PhysRevD.
101.123505 [https://arxiv.org/abs/1902.00534 1902.00534]
161. S. Roy Choudhury, S. Hannestad, T. Tram, Updated constraints on massive neutrino self-
interactions from cosmology in light of the . H0 tension. JCAP 03, 084 (2021). https://doi.org/
10.1088/1475-7516/2021/03/084 [https://arxiv.org/abs/2012.07519 2012.07519]
162. T. Brinckmann, J.H. Chang, M. LoVerde, Self-interacting neutrinos, the Hubble parameter
tension, and the cosmic microwave background. Phys. Rev. D 104, 063523 (2021). https://
doi.org/10.1103/PhysRevD.104.063523 [https://arxiv.org/abs/2012.11830 2012.11830]
163. A. Das, S. Ghosh, Self-interacting neutrinos as a solution to the Hubble tension?, in Euro-
pean Physical Society Conference on High Energy Physics 2021, vol. EPS-HEP2021,
p. 124, 9 (2022). https://doi.org/10.22323/1.398.0124 [https://arxiv.org/abs/2109.03263
2109.03263]
164. T. Brinckmann, J.H. Chang, P. Du, M. LoVerde, Confronting interacting dark radiation sce-
narios with cosmological data. https://arxiv.org/abs/2212.13264 2212.13264
165. S. Sandner, M. Escudero, S.J. Witte, Precision CMB constraints on eV-scale bosons coupled
to neutrinos. https://arxiv.org/abs/2305.01692 2305.01692
166. H.G. Escudero, J.-L. Kuo, R.E. Keeley, K.N. Abazajian, Early or phantom dark energy, self-
interacting, extra, or massive neutrinos, primordial magnetic fields, or a curved universe: an
exploration of possible solutions to the . H0 and .σ8 . Phys. Rev. D 106, 103517 (2022). https://
doi.org/10.1103/PhysRevD.106.103517 [https://arxiv.org/abs/2208.14435 2208.14435]
167. C.D. Kreisch et al., The Atacama cosmology telescope: the persistence of neutrino self-
interaction in cosmological measurements. https://arxiv.org/abs/2207.03164 2207.03164
168. J. Venzor, G. Garcia-Arroyo, A. Pérez-Lorenzana, J. De-Santiago, Resonant neutrino self-
interactions and the . H0 tension. https://arxiv.org/abs/2303.12792 2303.12792
169. Y. Chikashige, R.N. Mohapatra, R. Peccei, Spontaneously Broken Lepton Number and Cos-
mological Constraints on the Neutrino Mass. Spectrum (1926). Phys. Rev. Lett. Series 45
(1980). https://doi.org/10.1103/PhysRevLett.45.1926
170. Y. Chikashige, R.N. Mohapatra, R. Peccei, Are there real goldstone bosons associated
with broken Lepton number? Phys. Lett. B 98, 265 (1981). https://doi.org/10.1016/0370-
2693(81)90011-3
171. V.D. Barger, W.-Y. Keung, S. Pakvasa, Majoron emission by neutrinos. Phys. Rev. D 25, 907
(1982). https://doi.org/10.1103/PhysRevD.25.907
172. K. Akita, M. Niibo, Updated constraints and future prospects on majoron dark matter.
https://arxiv.org/abs/2304.04430 2304.04430
173. Z. Chacko, L.J. Hall, T. Okui, S.J. Oliver, CMB signals of neutrino mass gener-
ation. Phys. Rev. D 70, 085008 (2004). https://doi.org/10.1103/PhysRevD.70.085008
[https://arxiv.org/abs/hep-ph/0312267 hep-ph/0312267]
174. M. Lattanzi, J.W.F. Valle, Decaying warm dark matter and neutrino masses.
Phys. Rev. Lett. 99, 121301 (2007). https://doi.org/10.1103/PhysRevLett.99.121301
[https://arxiv.org/abs/0705.2406 0705.2406]
175. M. Lattanzi, S. Riemer-Sorensen, M. Tortola, J.W.F. Valle, Updated CMB and x- and .γ -ray
constraints on Majoron dark matter. Phys. Rev. D 88, 063528 (2013). https://doi.org/10.1103/
PhysRevD.88.063528 [https://arxiv.org/abs/1303.4685 1303.4685]
176. C. Biggio, L. Calibbi, T. Ota, S. Zanchini, Type-II Majoron Dark Matter.
https://arxiv.org/abs/2304.12527 2304.12527
177. F. Forastieri, M. Lattanzi, P. Natoli, Constraints on secret neutrino interactions
after Planck. JCAP 07, 014 (2015). https://doi.org/10.1088/1475-7516/2015/07/014
[https://arxiv.org/abs/1504.04999 1504.04999]
480 S. Gariazzo and O. Mena

178. F. Forastieri, M. Lattanzi, P. Natoli, Cosmological constraints on neutrino self-interactions

with a light mediator. Phys. Rev. D 100, 103526 (2019). https://doi.org/10.1103/PhysRevD.
100.103526 [https://arxiv.org/abs/1904.07810 1904.07810]
179. M. Escudero, S.J. Witte, A CMB search for the neutrino mass mechanism and its relation to
the Hubble tension. Eur. Phys. J. C 80, 294 (2020). https://doi.org/10.1140/epjc/s10052-020-
7854-5 [https://arxiv.org/abs/1909.04044 1909.04044]
180. F. Arias-Aragon, E. Fernandez-Martinez, M. Gonzalez-Lopez, L. Merlo, Neutrino masses
and Hubble tension via a Majoron in MFV. Eur. Phys. J. C 81, 28 (2021). https://doi.org/10.
1140/epjc/s10052-020-08825-8 [https://arxiv.org/abs/2009.01848 2009.01848]
181. G.-Y. Huang, W. Rodejohann, Solving the Hubble tension without spoiling Big Bang Nucle-
osynthesis. Phys. Rev. D 103, 123007 (2021). https://doi.org/10.1103/PhysRevD.103.123007
https://arxiv.org/abs/2102.04280 2102.04280]
182. M. Escudero, S.J. Witte, The Hubble tension as a hint of leptogenesis and neutrino mass
generation. Eur. Phys. J. C 81, 515 (2021). https://doi.org/10.1140/epjc/s10052-021-09276-
5 [https://arxiv.org/abs/2103.03249 2103.03249]
183. G. Barenboim, U. Nierste, Modified majoron model for cosmological anoma-
lies. Phys. Rev. D 104, 023013 (2021). https://doi.org/10.1103/PhysRevD.104.023013
[https://arxiv.org/abs/2005.13280 2005.13280]
184. F. Forastieri et al., Cosmic microwave background constraints on secret interactions among
sterile neutrinos. JCAP 07, 038 (2017). https://doi.org/10.1088/1475-7516/2017/07/038
[https://arxiv.org/abs/1704.00626 1704.00626]
185. M. Archidiacono, S. Hannestad, R.S. Hansen, T. Tram, Cosmology with self-interacting sterile
neutrinos and dark matter—a pseudoscalar model. Phys. Rev. D 91, 065021 (2015). https://
doi.org/10.1103/PhysRevD.91.065021 [https://arxiv.org/abs/1404.5915 1404.5915]
186. M. Archidiacono, S. Hannestad, R.S. Hansen, T. Tram, Sterile neutrinos with pseudoscalar
self-interactions and cosmology. Phys. Rev. D 93, 045004 (2016). https://doi.org/10.1103/
PhysRevD.93.045004 [ https://arxiv.org/abs/1508.02504 1508.02504]
187. M. Archidiacono, S. Gariazzo, C. Giunti, S. Hannestad, R. Hansen, M. Laveder
et al., Pseudoscalar–sterile neutrino interactions: reconciling the cosmos with neu-
trino oscillations. JCAP 08, 067 (2016). https://doi.org/10.1088/1475-7516/2016/08/067 [
https://arxiv.org/abs/1606.07673 1606.07673]
188. N. Saviano, O. Pisanti, G. Mangano, A. Mirizzi, Unveiling secret interactions among sterile
neutrinos with big-bang nucleosynthesis. Phys. Rev. D 90, 113009 (2014). https://doi.org/10.
1103/PhysRevD.90.113009 https://arxiv.org/abs/1409.1680 1409.1680]
189. A. Mirizzi, G. Mangano, O. Pisanti, N. Saviano, Collisional production of sterile neutrinos via
secret interactions and cosmological implications. Phys. Rev. D 91, 025019 (2015). https://
doi.org/10.1103/PhysRevD.91.025019 [https://arxiv.org/abs/1410.1385 1410.1385]
190. M. Archidiacono, S. Gariazzo, C. Giunti, S. Hannestad, T. Tram, Sterile neutrino self-
interactions: . H0 tension and short-baseline anomalies. JCAP 12, 029 (2020). https://doi.org/
10.1088/1475-7516/2020/12/029 [https://arxiv.org/abs/2006.12885 2006.12885]
191. M.A. Corona et al., Pseudoscalar sterile neutrino self-interactions in light of Planck,
SPT and ACT data. JCAP 06, 010 (2022). https://doi.org/10.1088/1475-7516/2022/06/010
[https://arxiv.org/abs/2112.00037 2112.00037]
192. X. Chu, B. Dasgupta, M. Dentler, J. Kopp, N. Saviano, Sterile neutrinos with secret
interactions—cosmological discord? JCAP 11, 049 (2018). https://doi.org/10.1088/1475-
7516/2018/11/049 https://arxiv.org/abs/1806.10629 1806.10629]
Chapter 25
Decaying Dark Matter and the Hubble
Tension

Andreas Nygaard, Emil Brinch Holm, Thomas Tram, and Steen Hannestad

25.1 Introduction

Although the nature of dark matter remains unknown, a brief look at the Standard
Model contents of the Universe reveals that a majority of the known particles are
unstable and decay [1, 2]. By analogy, a natural question to ask is whether dark
matter may decay on cosmological timescales. Decays of dark matter into electro-
magnetically interacting particles are strongly constrained by CMB observations [3].
Decays into a dark sector, so-called invisible decays, on the other hand, are much
less constrained because no direct observation channel exists. Nonetheless, there are
strong constraints on models that assume all of dark matter to decay on cosmological
timescales (e.g., the simple observation that we observe it today) [4, 5]. However,
these constraints may always be evaded by considering a scenario where only a frac-
tion of the dark matter decays invisibly. It is this class of models we study in this
chapter.
There exist several phenomenological models of invisibly decaying dark matter,
largely varying in their assumptions on the decaying particle (cold or warm) and on
the decay products (massive or massless, two- or many-body decays). In this chapter,
we review constraints on the three most studied models:
DCDM.→DR: Decaying cold dark matter (DCDM) decaying into dark radiation
(DR). Presented in Sect. 25.2.
DCDM.→DR+WDM: Decaying cold dark matter decaying into warm dark matter
(WDM) and dark radiation. Presented in Sect. 25.3.

A. Nygaard (B) · E. B. Holm · T. Tram · S. Hannestad

Department of Physics and Astronomy, Aarhus University, 8000 Aarhus C, Denmark
e-mail: andreas@phys.au.dk
E. B. Holm
e-mail: ebholm@phys.au.dk
T. Tram
e-mail: thomas.tram@phys.au.dk
S. Hannestad
e-mail: steen@phys.au.dk
© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 481
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_25
482 A. Nygaard et al.

Fig. 25.1 Hubble parameter . H as a function of scale factor .a, relative to its value in the .ɅCDM
model, for models of decaying cold dark matter (DCDM), decaying warm dark matter (DWDM),
and additional relativistic degrees of freedom, .ΔNeff . Taken from Ref. [1]

DWDM.→DR: Decaying warm dark matter (DWDM) decaying into dark radia-
tion. Presented in Sect. 25.4.

Why are these models relevant to the Hubble tension? Fig. 25.1 shows the rela-
tive change in the Hubble parameter . H (a) as a function of the scale factor .a for
the DWDM.→DR and DCDM.→DR models, as well as a model with additional
relativistic degrees of freedom, .ΔNeff . The abundances of the species have been
fixed such that they all contribute .ΔNeff = 0.5 to the number of relativistic degrees
of freedom at .a = 1, the acoustic scale is fixed (thus letting . H0 vary) to simulate
observational constraints, and the DWDM and DCDM models have the same decay
rate .┌.
In the top panel, .┌ is such that both models decay before recombination. In
this case, the energy density is greater than that of .ɅCDM, increasing the value
of the Hubble parameter. As the species decay, their values of . H converge to that
of the . Neff model. In all three models, the increase in . H (a) prior to recombination
decreases the sound horizon at recombination, .rs (a∗ ). Since observations essentially
fix the acoustic scale to .θs = rs (a∗ )/D A (a∗ ), where . D A (a∗ ) is the angular diameter
distance to recombination, this results in a decrease of . D A (a∗ ), which manifests in
an increased value of . H0 , as can be seen in the figure.1 The model of extra relativistic
degrees of freedom obtains the largest increase in . H0 but is known to have too strong
an impact on the CMB spectrum to satisfactorily solve the Hubble tension [6]. The
motivation for the decaying models, then, is that they may be able to circumvent these
constraints by virtue of injecting the radiation energy density more locally around
recombination.

1 However, in the bottom panel, where .┌ is such that both models decay after recombination, only
a comparatively negligible increase in . H0 is seen.
25 Decaying Dark Matter and the Hubble Tension 483

25.2 DCDM.→DR

The most simple model involving decaying dark matter is decaying cold dark matter
(DCDM), where a cold mother particle decays into massless particles denoted as dark
radiation (DR), which is a type of radiation that does not interact electromagnetically.
We can simply write this process as

. X dcdm → γdr .

As mentioned in the introduction, the scenario where all dark matter is decaying is
heavily constrained, so we consider a model where just a fraction,

Ωini
f
. dcdm ≡ dcdm
, (25.1)
Ωini
dcdm + Ωscdm

of the dark matter is unstable. Here .Ωscdm is the density parameter of the stable
cold dark matter, and .Ωini
dcdm is the density parameter of the decaying component
as it would be today if none of it had decayed [7]. This partial decay can also be
interpreted as all CDM decaying into a stable CDM particle and DR [7].
The equations for the homogeneous and isotropic energy densities in this model
resemble the normal background equations for CDM and radiation but with additional
source terms dependent on the decay rate, .┌dcdm (w.r.t. proper time) [7],

' a'
ρdcdm = −3 ρdcdm − a┌dcdm ρdcdm ,
.
a (25.2)
'
' a
ρdr = −4 ρdr + a┌dcdm ρdcdm ,
a
where the prime denotes derivatives w.r.t. conformal time.
The equations in first-order general relativistic perturbation theory for the cold
mother particle are exactly the same as for normal CDM, where only the density per-
turbation is non-trivial. For DR, the Boltzmann hierarchy resembles that of massless
neutrinos but with additional source terms in the lowest modes [4].
It is useful to distinguish between two regions of the parameter space when dealing
with this model, since the two limits .┌dcdm → 0 and .┌dcdm → ∞ correspond to
the .ɅCDM model, with the first having no constraints on the fractional amount of
DCDM, . f dcdm , and the latter having a very weak constraint of . f dcdm /= 1. This leads
to a large parameter space volume of good fits to data in both limiting cases, and
this funnel-shaped likelihood is difficult to sample simultaneously with an MCMC
algorithm. We therefore split the parameter space into a long-lived region, where
−1
.┌dcdm < 10 Gyr
3
, and a short-lived region, where .┌dcdm > 103 Gyr −1 . The long-
lived regime corresponds to a very late decay long after recombination and the
short-lived regime corresponds to an early decay before recombination. The latter is
probably the most interesting in relation to the Hubble tension, since it can impact the
484 A. Nygaard et al.

Fig. 25.2 Triangle plot of posteriors in the short-lived regime using Planck 2018 data with and
without BAO peak data. The figure shows the model-specific parameters, .ωdcdm
ini and .┌dcdm , along
with . H0 , and has been produced using the same MCMC runs used in Ref. [7]

physics around recombination, and hence the formation of the CMB. The constraints
on . H0 from Ref. [7] are different for the two regimes, with the long-lived regime
yielding a 95.45% credible interval of . H0 = 67.7+1.0
−0.9 km s
−1
Mpc−1 and the short-
+1.2 −1 −1
lived one yielding one of . H0 = 68.6−1.4 km s Mpc . As we might expect, the
short-lived regime alleviates the Hubble tension better of the two with almost one
standard deviation (compared to a value inferred from local measurements of . H0 =
73.04 ± 1.04 km s−1 Mpc−1 from Ref. [2]). We will therefore only consider the
short-lived regime from hereon.
The DCDM model has been investigated in numerous papers with Bayesian statis-
tics [8–15] and recently also with frequentist statistics and profile likelihoods [16].
Figure 25.2 shows the posteriors for the short-lived regime in the model-specific
parameters (along with . H0 ) using high-.l and low-.l temperature and polarization as
well as lensing data from Planck 2018 [17] (from now on just referred to as Planck
25 Decaying Dark Matter and the Hubble Tension 485

Fig. 25.3 Top panel: One-dimensional profile likelihood of .log10 ┌/Gyr. Bottom panel: The abun-
dance of decaying cold dark matter .ωdcdm
ini associated with every point in the profile above. Dashed
lines indicate the .Δχ 2 = 1 intersections, giving the .1σ CIs. Taken from Ref. [16]

2018 data) with and without the BOSS DR12 BAO peak data [18] (from now on
referred to as BAO peak data). The posterior of the decay rate, .┌dcdm , shows a vol-
ume effect towards higher values, where a large part of the parameter space fits the
data well. Because of this, it is not possible to assign a reasonable credible interval to
the decay rate since any such interval would be prior dependent. It is, however, also
apparent that a region around .log10 (┌/Gyr) ∈ [4.5, 5.5] includes a separate effect in
the posterior, visualized as a small “bump” or plateau, which marks this as a region of
interest. By investigating the same region in the decay rate with profile likelihoods,
we find that a peak in the likelihood (a “well” in.χ 2 (┌) ≡ −2 log(L(┌)/ max(L(┌)))
arises here, as shown in Fig. 25.3. This study has been performed with the same data
(Planck 2018 data and BAO peak data) along with low redshift BAO data from the
6dF survey [19] and the BOSS main galaxy sample [20] (the inclusion of which
will be referred to as the full BOSS DR12 data set). The approximate .68% confi-
dence interval obtained is .log10 ┌ Gyr−1 = 4.763+0.214−0.290 , while it is unconstrained at
.95%. The global best-fit of the model lies in this region at .Δχ = 2.8 relative to the
2

.ɅCDM model, hinting at a mild preference for the DCDM model over .ɅCDM. Inter-

estingly, the best-fitting parameters .log10 ┌ Gyr−1 = 4.763 and .ωdcdm ini
= 0.00429
correspond to a scenario where about .3% of the cold dark matter decays just prior
to recombination, in support of the tendency of early time solutions to the Hubble
tension to preferentially modify physics temporally close to recombination [6, 21].
Despite the stronger signature of DCDM in the frequentist analysis, the resulting
+0.54 −1
.68% constraint . H0 = 68.14−0.49 km s. Mpc.−1 solidifies the conclusion that the
DCDM.→DR model is unable to solve the . H0 tension (Fig. 25.4).
486 A. Nygaard et al.

Fig. 25.4 One-dimensional profile likelihoods of . H0 using Planck 2018 data and the full BOSS
DR12 data set (see text). The two profiles are with and without an additional S. H0 ES prior. Taken
from Ref. [16]

25.3 DCDM.→DR+WDM

A reasonable increase in model complexity is to allow one of the daughter particles

to be massive. We still have a cold mother particle decaying into dark radiation, but
now also accompanied by a massive daughter particle acting as warm dark matter
(WDM). We can write this process as

. X dcdm → γdr + Ywdm .

This model has the same model parameters as the DCDM.→DR model (abundance
and decay rate), but it also has a new parameter: The mass ratio of the WDM particle
to the DCDM particle, .m ~. Using conservation of energy and momentum, one can
relate this mass ratio to the WDM velocity and the fraction of energy transferred to
the massless DR particle in the decay, .∈, through the following two equations [22],

1 ∈2
∈=
. (1 − m
~2 ), βwdm
2
= , (25.3)
2 (1 − ∈)2

where .βwdm is the velocity of the massive daughter in natural units.

The background and perturbation equations are the same for DCDM and DR as
in the DCDM.→DR model, except for an additional factor of .∈ on the source term
in the background equation for DR. The equations for the massive daughter can be
expressed in terms of the equation-of-state parameter, .wwdm (a), which can be shown
to have the following form [22],
{
1 ┌β 2 a(t)
e−┌t D dln(a D )
.wwdm (a) = , (25.4)
3 1 − e−┌t aini H D [(a/a D )2 (1 − β 2 ) + β 2 ]
25 Decaying Dark Matter and the Hubble Tension 487

where .aini refers to some initial value of .a where our numerical integration begins,
and the subscript “. D” refers to the integration variable. The background equation
for the massive daughter particle is then

( ) a'
ρ̇
. wdm = −3 1 + wwdm (a) ρwdm + (1 − ∈)a┌ρ0 . (25.5)
a
The perturbation equations resemble those of massive neutrinos but can be rewritten
in terms of .wwdm in a similar manner [23]. An accurate implementation of the full
Boltzmann hierarchy fast enough for MCMC runs to be feasible is still missing, but
results have been produced using a fluid approximation [5, 6, 23–25].
Figure 25.5 shows the posteriors for the model-specific parameters,.┌ and.∈, along
with . H0 . These results assume that all dark matter is decaying, i.e., the abundance
ratio,. f dcdm , is fixed to unity. Because of this, the results are within the very long-lived
regime, and the energy transfer ratio, .∈, is rather low as well, resulting in a heavy
WDM daughter particle acting much like CDM. In particular, if.log10 (∈) ≾ −2.7, the
decay rate becomes unconstrained because the massive daughter particle becomes
virtually indistinguishable from the cold mother particle. The 68.27% credible inter-
val for the Hubble constant from this analysis is . H0 = 67.71+0.42 −0.43 km s
−1
Mpc−1 ,
which is similar to the result from the long-lived regime in the simple DCDM.→DR
model. We would expect that an analysis with a fractional decay in the short-lived
regime would yield a higher value of . H0 . In order to do this accurately, the full
hierarchy should be solved for WDM instead of using a fluid approximation.

25.4 DWDM.→DR

Another extension of the simplest decaying dark matter model in Sect. 25.2 is to allow
the decaying particle a non-negligible momentum, making it a warm dark matter
species. The model of a decaying warm dark matter (DWDM) species decaying to
dark radiation was studied in Refs. [1, 26]. This model has several interesting particle
physics realizations, such as decaying neutrinos [27] and majoron decays [28]. In
particular, in [29] it was used to constrain the lifetime of the active neutrino species.
The fundamental characteristic that separates the DWDM model from the other
decaying dark matter models is that the non-negligible momentum increases the
lifetime by a factor of . E/m through time dilation, where . E and .m denote the energy
and mass of the decaying particle, respectively. In the general case, the model contains
three parameters: The decay constant, .┌ (or, equivalently, the lifetime .τ ≡ 1/ ┌), the
DWDM mass, .m, and the abundance of the DWDM species, specified either at final
or initial time. The background equations are [1]
488 A. Nygaard et al.

Fig. 25.5 Triangle plot of the decay rate, the energy transfer ratio, and the Hubble constant. The red
lines and contours are from a background-only MCMC run with the full BOSS DR12 data set and
a S. H0 ES prior, while the blue lines and contours are from an MCMC run including perturbations
(with the fluid approximation) and also the Planck 2018 data in addition to the other data sets.
The gray contours are the . H0 measurement by the S. H0 ES collaboration. The figure is taken from
Ref. [23]

' a'
ρdwdm = −3 (ρdwdm + pdwdm ) − a┌mn dwdm ,
.
a (25.6)
' a'
ρdr = −4 ρdr + a┌dcdm mn dwdm ,
a
where .ρi and . pi denote the energy and pressure density of the .i’th species and .n dwdm
the number density of the decaying species. Apart from the addition of the non-zero
DWDM pressure, . pdwdm , these equations are identical to the case of a cold decaying
species (25.2) up to the substitution .ρ → mn, i.e., it is the rest mass and not the total
energy that drives the decay.
25 Decaying Dark Matter and the Hubble Tension 489

Fig. 25.6 Two-dimensional marginalized posteriors for the Hubble constant, . H0 , and the three
parameters of the DWDM.→DR model: The initial abundance, . Neff,ini,dwdm , parametrized as its
contribution to the effective number of relativistic degrees of freedom, the DWDM lifetime, .τ , and
the DWDM mass, .m. Taken from Ref. [1]

As in the DCDM.→DR model, the perturbations of the DWDM species are iden-
tical to those of massive neutrinos [30]. Moreover, the decay product perturbations
are influenced by a collision term, which captures the fact that the DWDM species
decays preferentially at low momenta [1].
Figure 25.6, adopted from [1], shows contours of the two-dimensional marginal-
ized posterior distributions for . H0 and three parameters of the DWDM.→DR model
using Planck 2018 data combined with the full BOSS DR12 data set. Here, the initial
abundance is parametrized as . Neff,ini,dwdm , its contribution to the effective number
of relativistic degrees of freedom. The prior on the lifetime studied here is roughly
equivalent to the short-lived regime explored earlier.
Evidently, although the maximum a posteriori estimate favors the .ɅCDM limit, a
considerable amount of the DWDM species is permitted by the data. Ref. [1] finds a
.68% credible upper bound of . Neff,ini,dwdm < 1.05. Furthermore, as seen in Fig. 25.6,

there is a mild correlation between . H0 and the DWDM abundance, indicating that the
DWDM.→DR model may admit a larger value of . H0 than .ɅCDM. A .68% credible
interval of . H0 = 68.73+0.81
−1.3 km s
−1
Mpc−1 , at a .2.7σ Gaussian tension with the
representative local value . H0 = 73.04 ± 1.04 km s−1 Mpc−1 [2], is presented in this
study. Thus, the broad conclusion is that there is no evidence from CMB data that
the DWDM.→DR model solves the Hubble tension.
Although the constraints on the lifetime, .τ , and mass, .m, are strongly driven by
the bounds of the uniform prior chosen, in broad terms, the data prefers small masses
and small lifetimes. Unfortunately, in all treatments of the DWDM.→DR model to
date, the effects of inverse decays have been neglected. The inverse decay process
is kinematically allowed when the DWDM species decays while still relativistic,
which is exactly the scenario in the preferred region of parameter space [1]. Thus, a
complete understanding of the DWDM.→DR model inevitably requires a numerical
implementation of the inverse decay processes and their quantum statistical correc-
tions [27]. Finally, we also note that the Bayesian constraints on the DWDM.→DR
model are expected to be strongly influenced by prior volume effects [1] since it
490 A. Nygaard et al.

reduces to the .ɅCDM model in the limit of vanishing abundance, making the life-
time and mass unconstrained and thereby storing a large probability mass in the
posterior distribution. At this time, a frequentist analysis of the model has not been
conducted.

25.5 Discussion and Conclusion

The three decay models presented in this chapter all show an ability to slightly
alleviate the Hubble tension. The simplest of the models (DCDM.→DR), has been
exhaustively studied, and it appears that it has reached its limits regarding how
much it can alleviate the tension. The other two models, however, still have poten-
tial for further investigation. The DCDM.→DR+WDM model still needs a fast and
accurate implementation of the full Boltzmann hierarchy, and with a fractional
decay, this model can also be studied in the short-lived regime. The DWDM.→DR
model also prefers slightly larger values of . H0 , but a study including the inverse
decay process is needed for a definite conclusion. Further investigation of the lat-
ter two models, along with improvements, could potentially alleviate the Hubble
tension further. Ultimately, at the time of writing, we cannot definitively say which
decaying dark matter model has the strongest alleviation of the Hubble tension,
although there is a slight preference for the DWDM.→DR model with the .68% C.I.
+0.81
. H0 = 68.73−1.3 km s
−1
Mpc−1 at a .2.7σ Gaussian tension with the representative
local value . H0 = 73.04 ± 1.04 km s−1 Mpc−1 [2]. Thus, although decaying dark
matter models are somewhat more natural than several other proposed solutions to
the . H0 discrepancy, they only accomplish a mild alleviation of the latter [6].
The next natural step to take in the hierarchy of numerical complexity would be
a fully general two-body decay from a mother particle with a definite mass to two
daughter particles with different masses, i.e., DWDM.→WDM.(1) +WDM.(2) . This is a
challenging task since it combines all the most difficult aspects of the previous mod-
els while also introducing new ones. Although the full set of perturbation equations
has been derived [27], an implementation in a numerical Einstein–Boltzmann solver
code still remains. Furthermore, we expect the full model to be very computation-
ally expensive to evaluate, possibly making it unfeasible for immediate inference
purposes. Nevertheless, a fully general decay scheme like this would, apart from
other use cases like neutrino decays, possibly be able to affect the physics around
recombination in just the right way for the Hubble tension to be relieved.

Acknowledgements A. Nygaard, E. B. Holm, and T. Tram were supported by a research grant

(29337) from VILLUM FONDEN. We would like to thank Guillermo Franco Abellán (first author
of Ref. [23]) for letting us use Fig. 25.5 in this chapter.
25 Decaying Dark Matter and the Hubble Tension 491

References

1. E.B. Holm, T. Tram, S. Hannestad, Decaying warm dark matter revisited. JCAP 08(08),
044 (2022). https://doi.org/10.1088/1475-7516/2022/08/044, http://arxiv.org/abs/2205.13628
[astro-ph.CO]
2. A.G. Riess et al., A comprehensive measurement of the local value of the hubble constant with 1
km s.−1 Mpc.−1 uncertainty from the hubble space telescope and the S. H0 ES team. Astrophys. J.
Lett. 934(1), L7(2022). https://doi.org/10.3847/2041-8213/ac5c5b, http://arxiv.org/abs/2112.
04510 [astro-ph.CO]
3. L. Zhang, X. Chen, M. Kamionkowski, Z.-G. Si, Z. Zheng, Constraints on radiative dark-matter
decay from the cosmic microwave background. Phys. Rev. D 76, 061301 (2007). https://doi.
org/10.1103/PhysRevD.76.061301, http://arxiv.org/abs/0704.2444 [astro-ph]
4. B. Audren, J. Lesgourgues, G. Mangano, P.D. Serpico, T. Tram, Strongest model-independent
bound on the lifetime of dark matter. JCAP 12, 028 (2014). https://doi.org/10.1088/1475-7516/
2014/12/028, http://arxiv.org/abs/1407.2418 [astro-ph.CO]
5. T. Simon, G. Franco Abellán, P. Du, V. Poulin, Y. Tsai, Constraining decaying dark mat-
ter with BOSS data and the effective field theory of large-scale structuresPhys. Rev. D 106
(2), 023516 (2022). https://doi.org/10.1103/PhysRevD.106.023516, http://arxiv.org/abs/2203.
07440 [astro-ph.CO]
6. N. Schöneberg, G. Franco Abellán, A. Pérez Sánchez, S. J. Witte, V. Poulin, J. Lesgourgues,
The . H0 Olympics: A fair ranking of proposed models. Phys. Rept. 984, 1–55 (2022). https://
doi.org/10.1016/j.physrep.2022.07.001, http://arxiv.org/abs/2107.10291 [astro-ph.CO]
7. A. Nygaard, T. Tram, S. Hannestad, Updated constraints on decaying cold dark matter. JCAP 05,
017 (2021). https://doi.org/10.1088/1475-7516/2021/05/017, http://arxiv.org/abs/2011.01632
[astro-ph.CO]
8. K. Ichiki, M. Oguri, K. Takahashi, WMAP constraints on decaying cold dark matter. Phys.
Rev. Lett. 93, 071302 (2004). https://doi.org/10.1103/PhysRevLett.93.071302, https://arxiv.
org/abs/astro-ph/0403164
9. Z. Berezhiani, A.D. Dolgov, I.I. Tkachev, Reconciling Planck results with low redshift
astronomical measurements. Phys. Rev. D 92(6), 061303 (2015). https://doi.org/10.1103/
PhysRevD.92.061303, http://arxiv.org/abs/1505.03644 [astro-ph.CO]
10. A. Chudaykin, D. Gorbunov, I. Tkachev, Dark matter component decaying after recombination:
Lensing constraints with Planck data. Phys. Rev. D 94, 023528 (2016). https://doi.org/10.1103/
PhysRevD.94.023528, http://arxiv.org/abs/1602.08121 [astro-ph.CO]
11. V. Poulin, P.D. Serpico, J. Lesgourgues, A fresh look at linear cosmological constraints on a
decaying dark matter component. JCAP 08, 036 (2016). https://doi.org/10.1088/1475-7516/
2016/08/036, http://arxiv.org/abs/1606.02073 [astro-ph.CO]
12. I.M. Oldengott, D. Boriero, D.J. Schwarz, Reionization and dark matter decay. JCAP 08,
054 (2016). https://doi.org/10.1088/1475-7516/2016/08/054, http://arxiv.org/abs/1605.03928
[astro-ph.CO]
13. A. Chudaykin, D. Gorbunov, I. Tkachev, Dark matter component decaying after recombina-
tion: sensitivity to baryon acoustic oscillation and redshift space distortion probes. Phys. Rev.
D 97(8), 083508 (2018). https://doi.org/10.1103/PhysRevD.97.083508, http://arxiv.org/abs/
1711.06738 [astro-ph.CO]
14. K.L. Pandey, T. Karwal, S. Das, Alleviating the . H0 and .σ8 anomalies with a decaying dark
matter model. JCAP 07, 026 (2020). https://doi.org/10.1088/1475-7516/2020/07/026, http://
arxiv.org/abs/1902.10636 [astro-ph.CO]
15. L. Xiao, L. Zhang, R. An, C. Feng, B. Wang, Fractional Dark Matter decay: cosmological
imprints and observational constraints. JCAP 01, 045 (2020). https://doi.org/10.1088/1475-
7516/2020/01/045, http://arxiv.org/abs/1908.02668 [astro-ph.CO]
16. E.B. Holm, L. Herold, S. Hannestad, A. Nygaard, T. Tram, Decaying dark matter with profile
likelihoods. Phys. Rev. D 107(2), L021303 (2023). https://doi.org/10.1103/PhysRevD.107.
L021303, http://arxiv.org/abs/2211.01935 [astro-ph.CO]
492 A. Nygaard et al.

17. Planck Collaboration, N. Aghanim et al., Planck 2018 results. VI. Cosmological parameters.
Astron. Astrophys. 641, A6 (2020). https://doi.org/10.1051/0004-6361/201833910, http://
arxiv.org/abs/1807.06209 [astro-ph.CO]. [Erratum: Astron.Astrophys. 652, C4 (2021)]
18. BOSS Collaboration, S. Alam et al., The clustering of galaxies in the completed SDSS-III
Baryon oscillation spectroscopic survey: cosmological analysis of the DR12 galaxy sample.
Mon. Not. Roy. Astron. Soc. 470(3), 2617–2652 (2017). https://doi.org/10.1093/mnras/stx721,
http://arxiv.org/abs/1607.03155 [astro-ph.CO]
19. BOSS Collaboration, S. Alam et al., The clustering of galaxies in the completed SDSS-III
Baryon oscillation spectroscopic survey: cosmological analysis of the DR12 galaxy sample.
Mon. Not. Roy. Astron. Soc. 470(3), 2617–2652 (2017). https://doi.org/10.1093/mnras/stx721,
http://arxiv.org/abs/1607.03155 [astro-ph.CO]
20. A.J. Ross, L. Samushia, C. Howlett, W.J. Percival, A. Burden, M. Manera, The clustering of
the SDSS DR7 main galaxy sample – I. A 4 per cent distance measure at .z = 0.15. Mon. Not.
Roy. Astron. Soc. 449(1), 835–847 (2015). https://doi.org/10.1093/mnras/stv154, http://arxiv.
org/abs/1409.3242 [astro-ph.CO]
21. L. Knox, M. Millea, Hubble constant hunter’s guide. Phys. Rev. D 101(4), 043533
(2020). https://doi.org/10.1103/PhysRevD.101.043533, http://arxiv.org/abs/1908.03663
[astro-ph.CO]
22. G. Blackadder, S.M. Koushiappas, Dark matter with two- and many-body decays and super-
novae type Ia. Phys. Rev. D 90(10), 103527 (2014). https://doi.org/10.1103/PhysRevD.90.
103527, http://arxiv.org/abs/1410.0683 [astro-ph.CO]
23. G. Blackadder, S.M. Koushiappas, Dark matter with two- and many-body decays and super-
novae type Ia. Phys. Rev. D 90(10), 103527 (2014). https://doi.org/10.1103/PhysRevD.90.
103527, http://arxiv.org/abs/1410.0683 [astro-ph.CO]
24. G. Franco Abellán, R. Murgia, V. Poulin, J. Lavalle, Implications of the. S8 tension for decaying
dark matter with warm decay products. Phys. Rev. D 105(6), 063525 (2022). https://doi.org/
10.1103/PhysRevD.105.063525, http://arxiv.org/abs/2008.09615 [astro-ph.CO]
25. L. Fuß, M. Garny, Decaying Dark Matter and Lyman-.α forest constraints. http://arxiv.org/abs/
2210.06117 [astro-ph.CO]
26. N. Blinov, C. Keith, D. Hooper, Warm decaying dark matter and the hubble tension. JCAP 06,
005 (2020). https://doi.org/10.1088/1475-7516/2020/06/005, http://arxiv.org/abs/2004.06114
[astro-ph.CO]
27. G. Barenboim, J.Z. Chen, S. Hannestad, I.M. Oldengott, T. Tram, Y.Y.Y. Wong, Invisible
neutrino decay in precision cosmology. JCAP 03, 087 (2021). https://doi.org/10.1088/1475-
7516/2021/03/087, http://arxiv.org/abs/2011.01502 [astro-ph.CO]
28. M. Escudero, S.J. Witte, The hubble tension as a hint of leptogenesis and neutrino mass gen-
eration. Eur. Phys. J. C 81(6), 515 (2021). https://doi.org/10.1140/epjc/s10052-021-09276-5,
http://arxiv.org/abs/2103.03249 [hep-ph]
29. G. Franco Abellán, Z. Chacko, A. Dev, P. Du, V. Poulin, Y. Tsai, Improved cosmological
constraints on the neutrino mass and lifetime. JHEP 08, 076 (2022). https://doi.org/10.1007/
JHEP08(2022)076, http://arxiv.org/abs/2112.13862 [hep-ph]
30. C.-P. Ma, E. Bertschinger, Cosmological perturbation theory in the synchronous and conformal
Newtonian gauges. Astrophys. J. 455, 7–25 (1995). https://doi.org/10.1086/176550, https://
arxiv.org/abs/astro-ph/9506072
Chapter 26
Signs of a Non-zero Equation-of-State
for Dark Matter

Krishna Naidoo

26.1 Tensions and Anomalies

The standard model of cosmology ɅCDM has been extremely successful in describ-
ing cosmological observations, including the cosmic microwave background [15],
galaxy clustering [2] and distance ladder [18] measurements. This simple model is
comprised of roughly 5% normal matter, 25% dark matter and 70% dark energy.
Despite it’s many successes the model remains incomplete. Almost 95% of our cur-
rent universe is made of stuff we have never directly measured, with few signs that
it’s properties will be detected in the near future. In ɅCDM we assume dark matter
interacts exclusively through gravity and is cold, meaning it’s particles have low
velocities that allow them to sink and form large clumps. On the other hand, dark
energy is assumed to be a cosmological constant, a scalar field which has a single
value across all of space-time. While general relativity allows such a solution, fun-
damental physics has found explaining its small value extremely difficult. This latter
point is often the premise for considering models proposing modifications to gravity.
Over the last decade tensions have emerged between the determination of param-
eters from the early universe and the late universe. The most egregious of these
discrepancies is the Hubble tension, a tension between the expansion rate of the
universe today measured from cepheid variables [16] and those derived from CMB
measurements [15]. Typically measurements from the late universe find a larger
Hubble constant than those derived from CMB and early universe physics (includ-
ing Baryonic Acoustic Oscillations from large scale structure). Combined with the
. S8 tension—a discrepancy between measurements of the clustering amplitude from
the early universe and late-time measurements of weak lensing have placed intense
scrutiny on the standard model of cosmology. Are these signs that the simple but

K. Naidoo (B)
University College London, Gower Street, London WC1E 6BT, UK
e-mail: k.naidoo@ucl.ac.uk

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 493
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_26
494 K. Naidoo

powerful assumptions of ɅCDM are starting to crack, are we seeing symptoms of

known limitations or are these just the result of systematics?
A lesser known discrepancy is the Integrated Sachs-Wolfe (ISW) void anomaly.
The ISW is a secondary anisotropy which impacts the light from the CMB and is
caused by the decay of gravitational potentials, which either draw or impart energy
to photons traversing a gravitational potential well (for an overdensity) or trough
(for an underdensity). This effect means large scale structure distorts the CMB we
observe. There are a number of ways to measure this effect, the simplest approach
is to cross-correlate large scale structure observations with the CMB, but this signal
is typically quite weak and difficult to measure (see for e.g. [14]). Another approach
is to use stacking, typically carried out with voids (see [5, 9]). From the latter, cos-
mologists have measured a surprising anomaly, in that the measurements of void
ISW is often significantly larger than what is expected from ɅCDM. This discrep-
ancy is strongest when measured from photometric voids, which are elongated and
preferentially aligned with the line-of-sight. What makes this result particularly puz-
zling is observations of the lensing signal show no large discrepancies, and therefore
no obvious systematics. Instead these measurements are either in agreement with
ɅCDM or slightly lower.

26.2 The Dark Matter Solution

ɅCDM assumes the presence of cold dark matter, massive uninteractive particles that
seed structure formation through their gravitational effects. However, little is known
about what dark matter really is; is it a single particle or many, is the microphysics
simple or complicated, and how does it fit in with the standard model of particle
physics? Its presumed simplicity has proved to be extremely powerful in making
predictions but have we reached the limit of these simplistic assumptions, i.e. has
precision cosmology highlighted our ignorance—could cosmological tensions actu-
ally be manifestations of unknown dark matter physics?
Extensions to dark matter often only depart from ɅCDM on very small scales,
when the mass of the dark matter particle starts to become relevant. For example
‘warm’ dark matter models [3], made of moderately light particles, prevent the growth
of structures on the very smallest scales due to their larger streaming velocities. Other
extensions such a fuzzy dark matter [7] similarly depart from ɅCDM on small scales,
but in this case are caused by extremely light particles with de Broglie wavelengths on
kiloparsec scales, causing quantum mechanical interference patterns on the matter
distribution field. The tendency for dark matter extensions to be only relevant on
small scales has created a perception that dark matter extensions are irrelevant for
any solutions to cosmological tensions. However, since we are yet to detect dark
matter directly there is plenty of room for unknown physics—physics that could be
relevant on large scales. For example what if dark matter is not a single particle but
many, with complicated interactions and decays? What if a subset of the particles are
unstable and decay, what if dark matter interacts with other dark sector components—
26 Signs of a Non-zero Equation-of-State for Dark Matter 495

i.e. dark energy and/or neutrinos? What approach do we take to explore these exotic
dark matter properties? One approach is to take each extension and explore them
individually, however this is a highly ineffective exploration of the parameter space.
An alternative approach is to explore this with phenomenological descriptions. This
allows us to explore the parameter space more effectively but comes at the cost of
losing specificity, since the dark matter properties will be generally defined. The
benefit of such an approach is that we can rely on data to dictate what dark matter
models are most viable and relevant for observations. With this in mind we turn to
the phenomenological generalised dark matter model [6]. In this model dark matter
is modelled as a fluid with free functions for the Equation-of-State (EoS), speed of
sound and shear viscosity. Since we are interested in the role this type of model
can have on cosmological tensions we will use previous works, in particular [8,
19], to limit the parameters and functional forms considered. In both [19] and [8] a
constant EoS for dark matter is constrained extremely tightly to .< 10−4 from CMB
experiments. For this reason we will assume at very early times dark matters EoS
is effectively zero. While we will assume the speed of sound is adiabatic and shear
viscosity is zero.

26.2.1 A Non-zero Equation-of-State

With this in mind we will be exploring the consequence of a dark matter model with
an EoS that begins at zero at early times and at some scale factor .anz transitions to a
monotonic function over cosmic time (specifically, linear as a function of the scale
factor .a); { ( )
wdm,0 a−a nz
, for a ≥ anz ,
.wdm (a) =
1−anz
(26.1)
0, otherwise,

where .wdm (a) is the EoS, .wdm,0 the EoS today, and .anz is the scale factor at which
the EoS transitions from zero to non-zero. In ɅCDM dark matter’s EoS is .wdm = 0.
From the continuity equation we can define the time evolution of the density as
ρdm,0
ρ (a) =
. dm W (a), (26.2)
a3
where ( { )
1
wdm (a ' ) '
. W (a) = exp 3 da . (26.3)
a a'

The total matter distribution is then defined as

ρm,0
ρ (a) =
. m ┌(a), (26.4)
a3
496 K. Naidoo

┌(a) = f b + f ν + f dm W (a),
. (26.5)

where . f i is the fraction of component .i to the matter density today (.b for baryons,
ν for non-relativistic neutrinos and .dm for dark matter). The addition of the time-
.

varying function .W (a) changes the time evolution of gravitational potentials in the
Poisson equation, which becomes

3 2 2
∇ 2 Φ(x, a) =
. H a Ωm (a)δ(x, a). (26.6)
2 0
In the linear regime the time evolution and spatial components of the density contrast
can be separated to .δ(x, a) = D(a)δ(x, 1) where . D(a) is the linear growth function.
This means the time evolution of the gravitational potential is proportional to

D(a)
Φ(a) ∝
. ┌(a), (26.7)
a

where as in ɅCDM .┌(a) = 1. This means the time-derivative of the gravitational

potential is given by

Φ̇(a) = H (a) [ f (a) − 1 + ϒ(a)] Φ(a),

. (26.8)

where . H (a) is the Hubble expansion rate, . f (a) = d ln D/d ln a the growth rate and

d┌(a) wdm (a)W (a)

ϒ(a) =
. = 3 f dm . (26.9)
d ln a ┌(a)

In ɅCDM, or any cosmological model where .Ωm ∝ a −3 , this last term (.ϒ) is zero.
A non-zero value is a unique signature of the non-zero EoS for dark matter.

26.2.2 Impact on the Integrated Sachs-Wolfe and Lensing

Signals

Consider now what effect this will have on the ISW and lensing convergence. The
ISW is defined as {
TISW (η̂) 2 χLS
. = 3 Φ̇(χ η̂, a)a dχ , (26.10)
TCMB c 0

and is sensitive to departures from the ɅCDM prediction that .ϒ(a) = 0 (see
Eq. 26.8). While the lensing convergence is defined as
{ χs
3H02 χ D(χ )
.κ(η̂) = Ωm,0 (χs − χ ) ┌(χ )δ(χ )dχ . (26.11)
2c2 0 χs a(χ )
26 Signs of a Non-zero Equation-of-State for Dark Matter 497

Fig. 26.1 The Hubble function . H (z), dark matter EoS .wdm (z) and power spectrum . P(k) are
displayed for models of ɅeDM as a function of the Hubble constant . H0 . Early universe physics are
fixed to CMB constraints. A negative dark mater EoS results in a larger . H0 and a lower .σ8 (shown
in the inset plot on the right). Therefore, ɅeDM naturally couples the two tensions, providing an
explanation for both the . H0 and .σ8 tensions simultaneously. Profiles for ɅCDM (assuming early
universe CMB constraints) are shown in dark blue (solid lines) and ɅeDM (assuming late universe
constraints on . H0 ) are shown in red (dashed lines). The figure was originally presented in [11]

and is therefore sensitive to departures from the ɅCDM prediction .┌(a) = 1.

By construction, ɅeDM is identical to ɅCDM at early times, since in both cases
dark matter EoS is zero, .wdm = 0. We can use this property to fix the early universe
to the conditions of ɅCDM inferred from Planck [15]. This allows us to focus on the
impact of.w(a > anz ) /= 0. In Fig. 26.1 we fix.anz = 0.5 (equivalent to redshift.z = 1)
and vary .wdm,0 showing a negative EoS for dark matter results in a larger . H0 and
smaller .σ8 . The higher . H0 is the results of a slower decay in the dark matter density,
while the lower .σ8 is the result of suppressed structure growth. The coupling of
these two effects shows the model can provide an explanation for both cosmological
tensions simultaneously.
In a similar fashion we can explore the effect of ɅeDM on the ISW and the lensing
profiles from voids. We look at three types of voids: spectroscopic voids—spherical
and relatively small [12], photometric voids—large and elongated along the line-of-
sight [9] and the Eridanus voids—voids founds along the line-of-sight of the cold
spot anomaly [10] (see [11] for specific void profile properties). In Fig. 26.2 we show
the effect of the same models considered in Fig. 26.1, showing a negative EoS leads to
much deeper ISW signals from voids, an effect which is particularly strong for large
elongated photometric and Eridanus voids. Rather interestingly this is coupled with
equal or slightly lower lensing signals. This is perfectly in keeping with observations,
where void ISW signals have been found to be much larger than predicted by ɅCDM
but with lensing signals either lower or in agreement with ɅCDM. This unique
prediction of ɅeDM is able to explain this long standing anomaly, which makes this
model of extreme interest should these anomalies persists in the future.
498 K. Naidoo

Fig. 26.2 The ISW (top panels) and lensing convergence .κ (bottom panels) angular profiles for
three void types (spec-z on the left, photo-z in the middle and Eridanus voids on the right) are shown
as a function of . H0 in ɅeDM with all other parameters fixed to early universe CMB constraints. In
the inset plots we show the amplitude of the ISW and.κ profiles with respect to ɅCDM, shown with a
.σ8 bias correction (solid purple lines) and without a bias correction (dashed purple lines). Measured
values are indicated with horizontal grey bands. A higher . H0 resolves the void-ISW anomaly and
recovers lower .κ amplitudes. For reference the profiles for ɅCDM assuming early universe CMB
constraints are shown with dark blue solid lines and the profiles for ɅeDM assuming late universe
constraints on . H0 are shown with red dashed lines. The figure was originally presented in [11]

26.2.3 Observational Constraints

In the prior sections we have described how an evolution of the dark matter EoS at
late-times can effect cosmological observables, with a keen focus on those relevant
to cosmological tensions and the ISW-void anomaly. We can now take this a step
further and constrain cosmological parameters with the EoS for dark matter allowed
to vary. To avoid exploring largely degenerate parameters we fix .anz = 0.5 and allow
.wdm,0 to vary between .−1 and .0 (the CDM limit). In Fig. 26.3 we show the constraints
on the standard 6 parameter model with the addition of .wdm,0 using observation of
the CMB from Planck [15], supernova from Pantheon [18] with S. H0 ES [16] Cepheid
constraints on the absolute magnitude . M B and BAO and redshift space distortions
from 6dF, SDSS DR7 and BOSS [1, 2, 17]. The constraints on the base ɅCDM
parameters are virtually identical since these are mostly set at early redshift where
the models are equivalent. Note, in the case of the dark matter fractional density, we
define it according to its asymptotic value at early times, which can be evaluated by
calculating
.Ωdm = Ωdm,0 W (anz ),
init
(26.12)

where .Ωdm,0 is the fractional density of dark matter today. This is a measure of what
the dark matter fractional density would be today in ɅCDM and is much easier to
26 Signs of a Non-zero Equation-of-State for Dark Matter 499

sample as this removes a strong degeneracy between .Ωdm,0 and .wdm,0 . With Planck
alone .Ωdm,0 is poorly constrained, unsurprisingly since this model considers a late
change to ɅCDM, CMB measurements can only constrain this model through their
impact on the sound horizon scale. With the addition of supernovae and BAO the
model is better constrained, with SN constraints preferring higher values of . H0 ,
pushing constraints to higher .wdm,0 . However, this preference is not definitive and is
still consistent with ɅCDM. In Fig. 26.4 we show the constraints on ɅeDM for the
derived parameters. H0 and.σ8 . This figure shows that there is a strong anti-correlation
for high-. H0 and low-.σ8 . In other words a small increase in . H0 is balanced with a
significantly larger shift in .σ8 . So while this model does correlate both tensions,
raising . H0 > 70 Km s−1 Mpc−1 , as measured by [16], are strongly disfavoured due
to the resulting low .σ8 .
A negative EoS for dark matter at late-time is a promising solution to tensions and
the ISW-void anomaly. Currently, cosmological data shows a marginal preference
for ɅCDM [11], with more precise measurements from void ISW and lensing (or
more typical galaxy and weak lensing auto and cross-correlations) required to make
a more definitive statement. Current best-fit constraints on ɅeDM, with a negative
EoS for dark matter at late-times, predict a dampening of the void lensing signal by
a factor of .∼ 0.9 with respect to ɅCDM and an increase in the void ISW signal by
a factor of .∼ 2 with respect to ɅCDM. These predictions fit perfectly with observa-
tional measurements from voids. To our knowledge, this is the only model that can
simultaneously provide a solution to the Hubble and . S8 tension, and the void-ISW
anomaly.

26.2.4 Implications: An Interacting Dark Sector

A negative EoS for dark matter is a significant departure from the cold dark matter
paradigm. So what does a negative EoS for dark matter actually mean? Generally,
the EoS for the different constituents in the universe are as follows, radiation has
an EoS = 1/3, cold dark matter and baryons have an EoS = 0 and a cosmological
constant has an EoS = .−1. Imagine now that dark matter was unstable and decayed
into radiation, over time its effective EoS would be slightly positive, since it would
describe a combination of dark matter and radiation decay products. Similarly, a
slightly negative EoS can be thought of as being an effective description of dark matter
that decays or interacts with dark energy. For this reason the model is degenerate
with models of interacting dark matter and dark energy [4]. To follow the behaviour
of ɅeDM, the interactions would need to be relevant only at very late-times. This
could be achieved if this interaction/decay only occurred at low energies or densities,
or if the dark matter particle was unstable and decayed with a very long half-life.
This behaviour is similar to the properties of the evanescent matter model of [13].
Should this turn out to be true, this would mark a significant departure from ɅCDM
and imply that dark energy is not a cosmological constant but rather something
else. Furthermore it would mean that cosmology could provide a window into new
500 K. Naidoo

Fig. 26.3 Constraints on the standard 6 parameter cosmological model with the addition of the
dark matter EoS today (.wdm,0 ). Constraints on the standard 6 parameters are identical to those
of ɅCDM, since by construction ɅeDM does not alter the early universe. We see that Planck
is very unconstraining on the additional .wdm,0 parameter. The addition of Pantheon supernovae
(with S. H0 ES constraints on the absolute magnitude . M B ) and BAO, significantly constrain .wdm,0 ,
and while the value is consistent with a zero there is a slight preference for a negative EoS of
.wdm,0 = −0.2

physics in the dark sector, physics which has remained elusive to direct detection
from particle physics experiments.

26.3 Future Work and Predictions for Experiment

A non-zero EoS for dark matter can alleviate tensions in cosmology and provide
an explanation for the ISW-void anomaly. Whether this models holds up against
26 Signs of a Non-zero Equation-of-State for Dark Matter 501

Fig. 26.4 Constraints on

ɅeDM on the Hubble
constants . H0 and .σ8
parameters. This shows a
direct anti-correlation
between high-. H0 and low-.σ8

the test of time depends on a number of factors: (1) that tensions and the ISW-
void anomaly remain, without clear systematics in observations and/or methodology
being discovered and (2) that limitations to the model are resolved.
Generalised dark matter is a good way to explore exotic dark matter models,
however the model is limited to linear perturbation theory and has not been explored
in great detail in high-resolution simulations and on small scales. Exploring these
model in simulations will allow us to test whether such a model is compatible with
observations and demonstrate how these models behave on small scales.
A unique consequence of this model is its prediction of larger ISW at low redshift.
Improved ISW measurements from voids, including better understanding of void
galaxy bias, will help to clearly establish the strength of the void-ISW anomaly—it’s
persistent will be the clearest sign of a negative EoS for dark matter at late times.

References

1. F. Beutler, C. Blake, M. Colless et al., MNRAS 416, 3017 (2011). https://doi.org/10.1111/j.

1365-2966.2011.19250.x
2. S. Alam et al., MNRAS 470, 2617–2652 (2017). https://doi.org/10.1093/mnras/stx721
3. S. Colombi, S. Dodelson, L.M. Widrow, ApJ 458, 1 (1996). https://doi.org/10.1086/176788
4. G.R. Farrar, P.J.E. Peebles, ApJ 604, 1 (2004)
5. B.R. Granett, M.C. Neyrinck, I. Szapudi, ApJ 683, L99 (2008). https://doi.org/10.1086/591670
6. W. Hu, ApJ 506, 485 (1998). https://doi.org/10.1086/306274
7. W. Hu, R. Barkana, A. Gruzinov, Phys. Rev. Lett. 85, 1158–1161 (2000). https://doi.org/10.
1103/PhysRevLett.85.1158
8. M. Kopp, C. Skordis, D.B. Thomas et al., Phys. Rev. Lett. 120, 221102 (2018). https://doi.org/
10.1103/PhysRevLett.120.221102
9. A. Kovács, C. Sánchez, J. García-Bellido et al., MNRAS 484, 5267 (2019). https://doi.org/10.
1093/mnras/stz341
502 K. Naidoo

10. R. Mackenzie, T. Shanks, M.N. Bremer et al., MNRAS 470, 2328 (2017). https://doi.org/10.
1093/mnras/stx931
11. K. Naidoo, M. Jaber, W.A. Hellwing, M. Bilicki, (2022). https://doi.org/10.48550/arXiv.2209.
08102, arXiv:2209.08102
12. S. Nadathur, R. Crittenden, ApJL 830, L19 (2016). https://doi.org/10.3847/2041-8205/830/1/
L19
13. P.J.E. Peebles, Annalen der Physik 524(9), 591 (2012)
14. Planck Collaboration, et al., 2016. A&A, 594. doi:https://doi.org/10.1051/0004-6361/
201525831
15. Planck Collaboration et al., A&A 641, A6 (2020). https://doi.org/10.1051/0004-6361/
201833910
16. A.G. Riess, W. Yuan, L.M. Macri et al., ApJL 934, L7 (2022). https://doi.org/10.3847/2041-
8213/ac5c5b
17. A.J. Ross, L. Samushia, C. Howlett, W.J. Percival, A. Burden, M. Manera, MNRAS 449,
835–847 (2015). https://doi.org/10.1093/mnras/stv154
18. Scolnic, D. M., et al., 2018, ApJ, 859. doi:https://doi.org/10.3847/1538-4357/aab9bb
19. L. Xu, Y. Chang, Phys. Rev. D 88, 127301 (2013). https://doi.org/10.1103/PhysRevD.88.
127301
Chapter 27
Resolving the Hubble Tension at Late
Times with Dark Energy

Marco Raveri

27.1 Introduction

In this chapter, we review attempts at solving the Hubble constant tension with non-
standard properties of Dark Energy (DE) at late times. The empirical evidence for
the existence of DE during the late cosmic epochs has been well-established through
numerous observational studies, starting from the discovery of cosmic accelera-
tion [1, 2]. Little progress has been made towards knowing its physical nature and
properties, that, to this day, remain unknown. So far, the best description we have for
DE is a cosmological constant, .Λ, which is one of the cornerstones of the standard
.ΛCDM cosmological model.

DE is a natural candidate to try to explain existing cosmological tensions between

the early and late time Universe. All physical components of the standard .ΛCDM
model, but DE, are showcased in the early universe, in particular at recombina-
tion. Dark matter, baryons, radiation, and neutrinos are all relevant at times when
anisotropies in the Cosmic Microwave Background (CMB) form, and these obser-
vations can be used to study their properties. DE, on the other hand, is, at least in
the standard model, completely irrelevant at those times. At late times, in turn, DE
becomes relevant and is the only difference between the physical components of the
early and late time universe. This is the empirical motivation to try to tie the physical
properties of DE, beyond a cosmological constant, to the resolution of cosmological
tensions.
In this chapter, we review these attempts, and the discussion is structured as
follows: we first show, in Sect. 27.2, how DE enters the cosmological observables
that are relevant for the Hubble constant tension; in Sect. 27.3 we discuss how DE can

M. Raveri (B)
Department of Physics, INFN and INAF at the University of Genova, Via Dodecaneso 33, 16146
Genova, Italy
e-mail: marco.raveri@unige.it

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 503
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_27
504 M. Raveri

be used to relieve the tension by changing the calibration of low redshift observables;
in Sect. 27.4 we discuss attempts at solving the Hubble tension by altering the late
times expansion history of the Universe with non-standard DE properties.

27.2 Dark Energy and the Hubble Constant

The observable universe is characterized by large-scale homogeneity and isotropy,

properties well-captured by the flat Friedmann-Lemaître-Robertson-Walker (FLRW)
metric with the line element .ds 2 = −dt 2 + a(t)2 dx2 . Here, .a(t) represents the scale
factor governing the background expansion of the universe, with.t being cosmic time.
The scale factor is directly related to redshift by .z = 1/a(t) − 1.
The temporal evolution of the scale factor is governed by the Friedmann equation,
given by
[ ]
. H 2 (a) = H02 Ωr a −4 + Ωm a −3 + Ωde (a) , (27.1)

where . H (a) ≡ ȧ/a represents the time-dependent Hubble expansion rate, in cosmic
time. In this equation, . H0 denotes the present-day Hubble constant, while .Ωr and .Ωm
correspond to the current relative energy densities of relativistic and non-relativistic
particle species, respectively. The term .Ωde (a) ≡ ρde (a)/ρtoday
critical
describes the effec-
tive density contribution of DE, .ρde , relative to present day critical density. Note
that, from Eq. (27.1), at present (.a = 1), the constraint .Ωr + Ωm + Ωde = 1 has to
be satisfied.
In the context of the.ΛCDM model, DE is described with a cosmological constant,
.Λ, resulting in a constant value for .Ωde . However, in a broader context, .Ωde (a) could

encompass the cumulative contribution from all components other than radiation and
matter densities. This includes possible modifications to gravity, which would lead
to a modified Friedmann equation, as well as the potential existence of a non-zero
curvature term, .Ωk a −2 . In all these cases .Ωde (a) becomes explicitly time dependent.
A crucial test for the standard .ΛCDM model involves verifying whether .Ωde remains
constant throughout cosmic history.
Friedman equations written as in Eq. (27.1) are very general and can in principle
describe any density contribution. For this reason, this approach is especially useful
to probe empirical discrepancies and span different families of models.
Numerous investigations in literature [3, 4] focus on the equation of state of DE,
which is given by

1 d ln Ωde
wde (a) =
. − 1, (27.2)
3 d ln a
looking for deviations from the standard model value of .wde = −1, which character-
izes a cosmological constant. We note here that this approach significantly restricts
the family of models that can be studied in the framework, as Eq. (27.2) is well
27 Resolving the Hubble Tension at Late Times with Dark Energy 505

defined only for positive .Ωde . This constraint is not necessarily satisfied in the con-
text of Modified Gravity (MG) theories [5], in which .Ωde is an effective quantity—a
placeholder—for structural modifications of Friedman equations. This constraint is
also not satisfied by curvature effective density contributions and a negative cosmo-
logical constant. For these reasons, we phrase, whenever possible, the discussion in
this chapter in terms of DE effective density.
The Hubble rate determines the redshift dependence of observable cosmological
distances, and in particular luminosity distance that is defined as:
/ { z
L dz
. D L (z) ≡ = (1 + z) (27.3)
4π S 0 H (z)

where . L and . S are the observed luminosity and flux of a source and we have assumed
spatial flatness, from here onward.
Luminosity distances in the universe at late times can be measured with Super-
novae (SN) observations that, once standardized, give us the apparent magnitude of
an event, as a function of luminosity distance:
( )
D L (z)
m(z) ≡ 5 log10
. + M0 (27.4)
10 pc

where . M0 is a calibration parameter that, crucially, is the same for all (standardized)
events of a given type and is redshift independent. This procedure is discussed in
detail, for a state-of-the-art SN survey, in [6]. Note that in Eq. (27.4), the value
of the Hubble constant enters as an additive parameter and, as such, is completely
degenerate with the determination of . M0 .
Distances can also be inferred, at late times, and in full analogy with the CMB,
by measuring the angular size of the Baryon Acoustic Oscillation (BAO) feature in
the galaxy correlation function. The reference physical scale of the BAO feature is
the sound horizon, .rs , of the Thompson scattering coupled photo-baryon fluid, at the
time when baryons stopped being dragged by photons. Crucially, at late times, this
quantity does not evolve with redshift and is fully determined by pre-recombination
physics. At a given redshift this scale projects on a feature at an angular separation
of:
rs
θ
. BAO = (27.5)
D A (z)

where . D A (z) is the redshift-dependent angular diameter distance, related to lumi-

nosity distance by . D L = (1 + z)2 D A . The value of the Hubble constant enters multi-
plicatively in Eq. (27.5), and, as such, is completely degenerate with the determination
of the sound horizon.
Both SN and BAO measurements have to be calibrated, in different ways, to
yield distance measurements. This is what we show in Fig. 27.1 where we can see:
SN measurements from the Pantheon compilation [11], with the redshift binning
506 M. Raveri

Fig. 27.1 The luminosity distance residuals with respect to the .ΛCDM model best fit to CMB
observations from Planck [7]. BAO [8–10] data points, in red, are calibrated by the Planck deter-
mination of the physical scale of the sound horizon in the .ΛCDM model. Pantheon SN [11] data
points, in black, are calibrated by the S. H0 ES [12] measurement of absolute SN magnitude,. M0 . The
upper colored band shows the uncertainty in the sound horizon calibration while the lower band
shows the errors in the measurements of . M0 . The separation between the two bands shows in full
display the Hubble tension. The gap between the data points of the two data sets shows the difficulty
of solving the tension with late times modifications to the expansion history: SN and BAO data
overlap in redshift, but, once calibrated, do not overlap in distance space. Figure adapted from [13]

used in [13, 14], calibrated by the measurement of . M0 trough the distance ladder,
from [12], which is often phrased as a measurement of . H0 ; BAO measurements
from the eBOSS [8], MGS [9] and 6dF [10] surveys, calibrated by the measurement
of the physical scale of the sound horizon, inferred from CMB observations of the
Planck satellite [7]. For both data sets, we show luminosity distance residuals, with
respect to the CMB best fit .ΛCDM cosmological model for visual clarity. BAO
angular diameter distances are converted to luminosity distances using the effective
redshift of the survey tomographic bins. The gray/red bands show the uncertainty in
the calibration of the two data sets respectively and the difference between the two
showcases the Hubble tension. Figure 27.1 illustrates the core issue in addressing the
Hubble constant tension at late times. The calibrated SN and BAO data sets share a
common redshift range, but their distance measurements do not coincide. Any given
cosmological model, that preserves the behavior of either calibrator, gives a unique
distance-redshift relation—is a line in Fig. 27.1—and cannot be in two different
positions at the same time.
For this reason achieving a complete resolution of the Hubble tension at late times
seems unattainable, although several models have been proposed to, at least, alleviate
the tension.
27 Resolving the Hubble Tension at Late Times with Dark Energy 507

A crucial aspect highlighting the challenge of fully addressing the Hubble constant
tension at late times arises from the joint use of . H0 /M0 measurements, SN, BAO,
and CMB data. In the context of this chapter, we assume that all four datasets have
no residual systematic effects when we discuss DE model performances at solving
the tension. It might be possible to achieve a better resolution of the Hubble constant
tension by excluding some parts of these data sets.

27.3 Dark Energy and Low Redshift Calibration

In this Section, we review attempts at solving the Hubble constant tension by using
DE to alter the low redshift calibration of distances.
This calibration is usually reported as a constraint on . H0 and, in the standard
approach, the local value of . H0 is obtained using the SN Ia magnitude-redshift
relation, as in Eq. 27.3, calibrated with Cepheid variables [15], with Tip of the Red
Giant Branch (TRGB) measurements [16], and others, see [17] for a review. To
strike a balanced tradeoff between sensitivity to expansion history, at high redshift,
and contamination from peculiar velocities, at low redshift, only a subset of SN Ia is
used in this determination of . H0 . For example, in [15], only SN with .z > 0.023 and
. z < 0.15 are used. Across this narrow redshift range the sensitivity to the shape of

the distance-redshift relation is limited and, in the standard analysis of [15], it is fit
with a template, by fixing the value of the deceleration parameter.
The . H0 calibration has then some residual dependency on the low redshift expan-
sion rate and its impact on the determination of the Hubble constant was studied in
detail in [18]. Several models altering the low redshift expansion history were studied,
with an emphasis on DE models. None of these models were found to give significant
changes to the inferred value of . H0 , with a maximum difference of .ΔH0 = 0.47 Km
s.−1 Mpc.−1 , or a shift of .0.6%, with an expansion history constrained to be very close
to the .ΛCDM one.
In retrospect, this conclusion is compatible with residuals in Fig. 27.1. The SN
calibrator sample includes only the first few low redshift data points and, once the
best fitting .ΛCDM expansion history is subtracted, there is no indication of redshift-
dependent residuals that could significantly change the estimate of the residuals
mean.
We mention here the special case of a very late transition in the DE density, as
was studied in [19]. If such a transition occurs very close to the present epoch (i.e.
. z < 0.1 or lower) it would go unnoticed in observations at higher redshifts, like

the CMB [20]. For these types of models it is crucial to treat the . H0 measurement
properly, as a constraint on SN calibration and in [19, 21] we discussed how to
convert back a constraint on . H0 from calibrated SN to a constraint on . M0 . A late
DE transition, before the first SN at .z < 0.01 can increase the local expansion rate,
pushing . H0 to values compatible with local measurements of [12] and even higher.
However, just as in void scenarios, the problem with such a solution is that raising
. H0 does not resolve the origin of the Hubble tension, as shown in Fig. 27.1, which
508 M. Raveri

resides in the calibration of SN in the Hubble flow. In other words, at redshift lower
than the Hubble flow SN sample, .z < 0.023, the expansion history is never relevant
as the distance-redshift relation is never used in the fit to derive the value of the
Hubble constant, which is effectively extrapolated to .z = 0.
Another, radically different, way of changing the SN Ia calibration is to change
the behavior of a step in the distance ladder calibration, especially Cepheids, by
altering gravity through DE, as discussed in [22]. A modification of gravity through
the introduction of a fifth force is a common outcome of theories that couple the
degree of freedom of DE with matter, propose novel interactions between objects,
or attempt to dynamically explain the nature of dark energy [5, 23–25]. Given the
absence of any observed fifth force in tests of gravity within the solar system, these
theories require a mechanism to maintain cosmological relevance while remaining
hidden within the solar system. The concept of screening—which involves weakening
the strength of the fifth force in regions of strong gravitational fields—was initially
introduced to ensure the consistency of scalar-tensor theories of gravity with tests
of the equivalence principle, the inverse square law, and post-Newtonian gravity
tests within the confines of the solar system [26, 27]. Known screening mechanisms
include: chameleon [28, 29], symmetron [30], dilaton [31], K-mouflage [32] and
Vainshtein [33] screening.
Since screening mechanisms generally depend on the local environment they
could alter the period-luminosity relation of Cepheids in different host galaxies. In
particular, this could happen between the Milky Way and N4258 and Cepheids in the
SN Ia calibrator sample, changing the inferred value of the Hubble constant. In [22],
the authors assume that the Cepheids used for calibrating the period-luminosity
relationship are subject to screening, and compute the expected change in . H0 , as
a function of the residual strength of the fifth force. Overall changes are not suffi-
cient to fully resolve the Hubble tension: models in which the cores of Cepheids are
unscreened (leading to increased luminosities) bring the results of the distance ladder
closer to, at most, 2–3 .σ from the Planck value when local variations in the effec-
tive gravitational constant, .ΔG/G are about 5%; models that exclusively unscreen
Cepheid envelopes (like chameleon models) require a higher 30–40% variations in
the gravitational constant to achieve a similar level of improvement.

27.4 Dark Energy and the Expansion History

In this Section, we review attempts at solving the Hubble constant tension by using
DE to change the expansion history at low to intermediate redshifts. This means
in particular that we consider changes in the expansion history at .z > 0.01, which
allows us to treat the calibration of SN Ia as a constraint on the Hubble constant.
As previously discussed, all these models incur in the no-go argument shown in
Sect. 27.2 so we generally expect to be able to somewhat relieve the Hubble tension,
without resolving it fully.
27 Resolving the Hubble Tension at Late Times with Dark Energy 509

27.4.1 Models for the de Equation of State

The simplest DE model we can consider, called wCDM, has a constant equation of
state, .wde = w0 , in analogy with matter and radiation components. In this case, the
value of the equation of state fixes the time dependence of the DE density as .Ωde ∝
a −3(1+w0 ) . To increase the expansion rate at low redshifts, with respect to constant
.Ωde , we need .Ωde to increase in time, rather than diluting, and hence we would need
.w0 < −1. As discussed in [34], because of the CMB geometric degeneracy, this might

look like an empirically viable model when considering CMB-only measurements.

Once the degeneracy is broken though, by the use of either SN or BAO data, the
model cannot solve the Hubble constant tension, leaving it qualitatively unchanged.
Similar conclusions apply to other phenomenological models for .wde , as reviewed
in [34]. If we consider the combination of Planck 2018, SN and BAO measurements,
as in Fig. 27.1, the baseline.ΛCDM value for the Hubble constant is,. H0 = 67.4 ± 0.8
Km s.−1 Mpc.−1 , in .∼ 4σ tension with results of . H0 = 73.04 ± 1.04 in [15]. The
wCDM model achieves, when fit to the same data combination . H0 = 68.34 ± 0.82
Km s.−1 Mpc.−1 , as reported in [34], which is a .∼ 0.5σ reduction in the tension.
We could consider as an extension of the wCDM model the Chevallier—
Polarski—Linder parameterization (CPL) [35, 36] for the equation of state, in which
.wde = w0 + wa (1 − a). When fitting our benchmark combination of data this model
yields . H0 = 68.35 ± 0.84 Km s.−1 Mpc.−1 , which is virtually identical to the wCDM
case. Other parametrizations have been tested in [34], in particular: the Jassal-Bagla-
Padmanabhan parameterization (JBP) [37] model; the Logarithmic DE equation
of state proposed by [38]; the BA parameterization put forward by Barboza and
Alcaniz [39]. All these models would not qualitatively change the results with respect
to the wCDM model. Some of these models might slightly change quantitatively the
tension, by either slightly shifting the central value of the . H0 estimate or inflating
the error bars of the final estimate. None would be on target and we refer the reader
to the discussion in [34] to gauge different model performances.
In this section, we have discussed only phenomenological approaches and not con-
crete physical model constructions. On the one hand, it is always possible to build a
model that is physically viable and reproduces any given expansion history [40]. On
the other hand, considering specific models would impose requirements of physical
viability on phenomenological quantities [41], hence restricting the available param-
eter space for solutions and necessarily degrading the performances of a model,
towards the solution of the Hubble tension. We refer the reader to [34, 42] for an
overview of DE model constraints in the context of the Hubble constant tension.

27.4.2 Reconstructed Dark Energy Models

While we have seen several models that alleviate the Hubble constant tension, it
is natural to ask what are the ultimate performances that can be achieved by DE
510 M. Raveri

models at this task. This requires going beyond testing specific model realizations
or simple parametric forms for .wde (z) and estimating .Ωde (z) directly from data.
Indeed, employing simpler constructions can introduce a bias into the results and
potentially lead to the loss of important information about ways in which tensions
can be relieved.
Multiple strategies exist for the non-parametric reconstruction of cosmological
functions like .wde (z) or .Ωde (z). Widely used techniques encompass discretizing
the functions at multiple redshifts, adopting Gaussian Processes (GP) [43–45], and
employing the correlated prior method [3, 46–48]. Employing a binning strategy,
where the function is assumed to be constant and independent within each bin or
smoothly interpolated between redshift nodes, leads to results that rely on an unphys-
ical implicit assumption of smoothness. Opting for a limited number of bins could
potentially introduce bias into the reconstruction while increasing their number risks
fitting noise features. On the other hand, the GP approach doesn’t impose restrictions
on bin numbers but enforces Gaussianity of the space where it is performed. This
space, incorporates a Gaussian prior that correlates the function across neighbor-
ing redshifts. The selection of the GP prior is fundamentally empirical, devoid of
any connection to a physical theory. Moreover, the prior’s parameters are typically
marginalized, which can obscure the Bayesian interpretation of the resulting recon-
struction. The correlated prior method similarly introduces a correlation between
neighboring redshifts, but unlike the GP approach, it employs a fixed prior covari-
ance matrix derived from theoretical considerations. This explicit prior enables a
clear articulation of the extent to which the data improves upon the prior and facili-
tates computation of Bayesian evidence that can be compared to that of the .ΛCDM
model, as discussed in [14].
In this section, we show what happens to the Hubble constant tension when we
opt for approaches 1 and 3, following [13, 14]. We do not pursue the GP strategy as
we want to reconstruct .Ωde (z). Distance measurements, for both SN and BAO, are
Gaussian distributed in data space but their measurements are non-linearly related
to .Ωde (z), which means that we generally expect the posterior distribution of .Ωde (z)
to be non-Gaussian. In the reconstruction .Ωde (z) is represented by its value at 11
discrete points (nodes) in redshift, with a cubic spline used to interpolate between
them. The first 10 nodes are distributed uniformly in the interval .a ∈ [1, 0.25] while
the last node is positioned at.a = 0.2. At higher redshift.Ωde (z) was made to approach
its .ΛCDM value, even though it is possible to study, with the same framework earlier
times deviations from the standard model [49].
On its own, the cubic spline introduces an implicit smoothing (and correlation)
between values of .Ωde (z) at different redshifts, as shown in [14]. In addition the
time correlation of .Ωde (z) is set from the theoretical correlation matrix obtained
in [50, 51] for the general family of Horndeski DE models [52]. This theory prior
has a much longer correlation time, with respect to the spline model, and acts as a
filter, discouraging but not completely prohibiting fast variations in .Ωde (z). Since the
correlation time is much longer than the spline model the theory prior reconstruction
is independent of the number of bins used in the reconstruction.
27 Resolving the Hubble Tension at Late Times with Dark Energy 511

Fig. 27.2 Left panel: luminosity distance residuals, as in Fig. 27.1, with the two best fitting recon-
structed DE models, with and without theoretical (Horndeski) priors. Right panel: reconstructed
.Ωde (z) for the data sets discussed in this chapter, with and without theory priors. The bands show
the .68% confidence level regions. The two vertical lines show the redshifts of equality between the
matter and DE densities, .z eq , and the beginning of cosmic acceleration, .z acc , in the best fit .ΛCDM
model. Figure adapted from [13]

We note that the discussion of the Hubble constant tension, in Sect. 27.2, relies
only on cosmological background arguments. For this reason, we show results for
.Ωde (z) and do not discuss the behavior of inhomogeneities beyond the standard
model. We refer the reader to [13, 14] for an in-depth discussion and how it relates
to other tensions.
As in [13, 14] we jointly fit all data discussed in Sect. 27.2, and the resulting
reconstruction of .Ωde (z) is shown in Fig. 27.2. The reconstruction is performed with
and without theoretical priors and the difference between the two showcases the
important role played by the theory priors. Without theory priors one can see oscil-
lations, in the redshift range .0 < z < 0.6, in the reconstructed .Ωde (z), which are
mainly driven by the scatter of BAO and SN data. These oscillations do not sig-
nificantly improve the fit and are hence suppressed by the theory prior. Figure 27.2
demonstrates the general difficulty of solving the Hubble tension by changing the
expansion history at late times. The CMB anchors the expansion rate at high redshift
and then the reconstruction struggles to strike a balance between fitting jointly SN
and BAO data points.
For the spline reconstruction, which deploys the largest fitting freedom we con-
sider, the inferred value for the Hubble constant, in the reconstructed DE model, is
−1
. H0 = (69.9) 69.3 ± 1.4 Km s. Mpc.−1 . In parentheses, hereafter, we have indicated
the value for the best-fitting model, which is significantly higher than the mean due
512 M. Raveri

to the non-Gaussianity of the posterior that we had anticipated. If the calibration of

SN is included in the fit, to force the reconstruction to struggle to increase the esti-
mate of the Hubble constant, then we obtain . H0 = (71.34) 70.9 ± 1.4. From these
two results, we can gauge the extent to which it is possible to relieve the Hubble
constant tension at late times. In the first case, the significance is slightly lowered,
to about 2 .σ , by a slight shift and significantly enlarged error bars. In the second
case, the significance of the tension is not changed—as the joint dataset contains
the SN calibration [53] but it is helpful to gauge the maximum performances that
are attainable. Note that both reconstructions achieve a value of the inferred Hubble
constant that is significantly higher than the DE models we discussed in the previous
section, because of the much larger fitting freedom.
When we include the DE theoretical priors in the reconstruction the ability of the
model to relieve the Hubble constant tension is significantly lowered. When fitting
only to CMB, BAO, and SN we can obtain at most . H0 = (67.7)67.8+1.1 −1.4 Km s.
−1
−1
Mpc. , a slight improvement over the standard .ΛCDM result of . H0 = 67.4 ± 0.8
Km s.−1 Mpc.−1 .

27.5 Summary

In this chapter, we have discussed how DE at late times could be invoked to solve the
tension over the value of the Hubble constant, between early and late time Universe
measurements. DE is a natural candidate to provide a resolution to this conundrum:
its physical properties are unknown; it’s the only physical constituent that is irrelevant
at early times and important at late times.
We have seen how DE at late times enters into the discussion over the Hubble
constant tension, how it influences the cosmological observables that are relevant
to the tension, and how this limits the capability of DE models to fully solve it.
Changing the expansion history, or the laws of gravity, at very late times, as we have
discussed in Sect. 27.3, could help alleviate the tension by changing the very way in
which the distance to Hubble flow SN is calibrated. Changing the expansion history at
intermediate redshifts, in the range .0.01 < z < 3, as we have reviewed in Sect. 27.4,
can also help alleviate the tension. For this family of models, the main limitation to a
full tension resolution is that the distance redshift relation is tomographically probed
by both SN and BAO observations. Both probes constrain the expansion history, in
most of this redshift interval, better than what is needed to solve the tension. We have
seen how this translates to the impossibility of achieving a full tension resolution,
even when deploying very general techniques that reconstruct, rather than constrain,
DE models at late times. These approaches are so general that they provide an upper
bound to the performances that can be achieved, which in this case are at about
. H0 = (71.34)70.9 ± 1.4, thus solving the tension only halfway.

Acknowledgements We thank Levon Pogosian for helpful discussions and comments. M.R.
acknowledges financial support from the INFN InDark initiative.
27 Resolving the Hubble Tension at Late Times with Dark Energy 513

References

1. A.G. Riess et al., Observational evidence from supernovae for an accelerating universe and a
cosmological constant. Astron. J. 116, 1009–1038 (1998)
2. S. Perlmutter et al., Measurements of .Ω and .Λ from 42 high redshift supernovae. Astrophys.
J. 517, 565–586 (1999)
3. G.-B. Zhao et al., Dynamical dark energy in light of the latest observations. Nature Astron.
1(9), 627–632 (2017)
4. L.A. Escamilla, W. Giarè, E. Di Valentino, R.C. Nunes, S. Vagnozzi, The state of the dark
energy equation of state circa 2023 7 (2023)
5. T. Clifton, P.G. Ferreira, A. Padilla, C. Skordis, Modified gravity and cosmology. Phys. Rept.
513, 1–189 (2012)
6. D. Brout et al., The Pantheon+ analysis: cosmological constraints. Astrophys. J. 938(2), 110
(2022)
7. N. Aghanim et al., Planck 2018 results. V. CMB power spectra and likelihoods. Astron. Astro-
phys. 641, A5 (2020)
8. S. Alam et al., Completed SDSS-IV extended Baryon oscillation spectroscopic survey: cosmo-
logical implications from two decades of spectroscopic surveys at the Apache point observatory.
Phys. Rev. D 103(8), 083533 (2021)
9. A.J. Ross, L. Samushia, C. Howlett, W.J. Percival, A. Burden, M. Manera, The clustering of
the SDSS DR7 main Galaxy sample—I. a 4 percent distance measure at .z = 0.15. Mon. Not.
Roy. Astron. Soc. 449(1), 835–847 (2015)
10. F. Beutler, C. Blake, M. Colless, D.H. Jones, L. Staveley-Smith, L. Campbell, Q. Parker, W.
Saunders, F. Watson, The 6dF Galaxy survey: Baryon acoustic oscillations and the local Hubble
constant. Mon. Not. Roy. Astron. Soc. 416, 3017–3032 (2011)
11. D.M. Scolnic et al., The complete light-curve sample of spectroscopically confirmed SNe
Ia from Pan-STARRS1 and cosmological constraints from the combined Pantheon sample.
Astrophys. J. 859(2), 101 (2018)
12. A.G. Riess, S. Casertano, W. Yuan, J.B. Bowers, L. Macri, J.C. Zinn, D. Scolnic, Cosmic
distances calibrated to 1% precision with Gaia EDR3 parallaxes and Hubble space telescope
photometry of 75 milky way Cepheids confirm tension with.ΛCDM. Astrophys. J. Lett. 908(1),
L6 (2021)
13. L. Pogosian, M. Raveri, K. Koyama, M. Martinelli, A. Silvestri, G.-B. Zhao, J. Li, S. Peirone,
A. Zucca, Imprints of cosmological tensions in reconstructed gravity. Nature Astron. 6(12),
1484–1490 (2022)
14. M. Raveri, L. Pogosian, M. Martinelli, K. Koyama, A. Silvestri, G.-B. Zhao, J. Li, S. Peirone,
A. Zucca, Principal reconstructed modes of dark energy and gravity. JCAP 02, 061 (2023)
15. A.G. Riess et al., A comprehensive measurement of the local value of the Hubble constant with
1 km s.−1 Mpc.−1 uncertainty from the Hubble space telescope and the S. H0 ES team. Astrophys.
J. Lett. 934(1), L7 (2022)
16. S.A. Uddin et al., Carnegie supernova project-I and -II: measurements of . H0 using Cepheid,
TRGB, and SBF distance calibration to type Ia supernovae 8 (2023)
17. A.G. Riess, L. Breuval, The local value of H.0 8 (2023)
18. S. Dhawan, D. Brout, D. Scolnic, A. Goobar, A.G. Riess, V. Miranda, Cosmological model
insensitivity of local . H0 from the Cepheid distance ladder. Astrophys. J. 894(1), 54 (2020)
19. G. Benevento, W. Hu, M. Raveri, Can late dark energy transitions raise the Hubble constant?
Phys. Rev. D 101(10), 103517 (2020)
20. M. Mortonson, W. Hu, D. Huterer, Hiding dark energy transitions at low redshift. Phys. Rev.
D 80, 067301 (2009)
21. M. Raveri, G. Zacharegkas, W. Hu, Quantifying concordance of correlated cosmological data
sets. Phys. Rev. D 101(10), 103527 (2020)
22. H. Desmond, B. Jain, J. Sakstein, Local resolution of the Hubble tension: the impact of screened
fifth forces on the cosmic distance ladder. Phys. Rev. D 100(4), 043537 (2019) [Erratum:
Phys.Rev.D 101, 069904 (2020), Erratum: Phys.Rev.D 101, 129901 (2020)]
514 M. Raveri

23. K. Koyama, Cosmological tests of modified gravity. Rept. Prog. Phys. 79(4), 046902 (2016)
24. A. Joyce, L. Lombriser, F. Schmidt, Dark energy versus modified gravity. Ann. Rev. Nucl. Part.
Sci. 66, 95–122 (2016)
25. C. Burrage, J. Sakstein, Tests of chameleon gravity. Living Rev. Rel. 21(1), 1 (2018)
26. J. Khoury, Theories of dark energy with screening mechanisms 11 (2010)
27. B. Jain, J. Khoury, Cosmological tests of gravity. Annals Phys. 325, 1479–1516 (2010)
28. J. Khoury, A. Weltman, Chameleon fields: awaiting surprises for tests of gravity in space. Phys.
Rev. Lett. 93, 171104 (2004)
29. J. Khoury, A. Weltman, Chameleon cosmology. Phys. Rev. D 69, 044026 (2004)
30. K. Hinterbichler, J. Khoury, Symmetron fields: screening long-range forces through local sym-
metry restoration. Phys. Rev. Lett. 104, 231301 (2010)
31. P. Brax, C. van de Bruck, A.-C. Davis, D. Shaw, The Dilaton and modified gravity. Phys. Rev.
D 82, 063519 (2010)
32. E. Babichev, C. Deffayet, R. Ziour, k-Mouflage gravity. Int. J. Mod. Phys. D 18, 2147–2154
(2009)
33. A.I. Vainshtein, To the problem of nonvanishing gravitation mass. Phys. Lett. B 39, 393–394
(1972)
34. E. Di Valentino, O. Mena, S. Pan, L. Visinelli, W. Yang, A. Melchiorri, D.F. Mota, A.G. Riess,
J. Silk, In the realm of the Hubble tension—a review of solutions. Class. Quant. Grav. 38(15),
153001 (2021)
35. M. Chevallier, D. Polarski, Accelerating universes with scaling dark matter. Int. J. Mod. Phys.
D 10, 213–224 (2001)
36. E.V. Linder, Exploring the expansion history of the universe. Phys. Rev. Lett. 90, 091301 (2003)
37. H.K. Jassal, J.S. Bagla, T. Padmanabhan, Observational constraints on low redshift evolution
of dark energy: how consistent are different observations? Phys. Rev. D 72, 103503 (2005)
38. G. Efstathiou, Constraining the equation of state of the universe from distant type Ia supernovae
and cosmic microwave background anisotropies. Mon. Not. Roy. Astron. Soc. 310, 842–850
(1999)
39. E.M. Barboza Jr., J.S. Alcaniz, A parametric model for dark energy. Phys. Lett. B 666, 415–419
(2008)
40. N. Frusciante, L. Perenon, Effective field theory of dark energy: a review. Phys. Rept. 857,
1–63 (2020)
41. S. Peirone, M. Martinelli, M. Raveri, A. Silvestri, Impact of theoretical priors in cosmological
analyses: the case of single field quintessence. Phys. Rev. D 96(6), 063524 (2017)
42. N. Schöneberg, G. Franco Abellán, A. Pérez Sánchez, S.J. Witte, V. Poulin, J. Lesgourgues,
The . H0 Olympics: a fair ranking of proposed models. Phys. Rept. 984, 1–55 (2022)
43. A. Shafieloo, A.G. Kim, E.V. Linder, Gaussian process cosmography. Phys. Rev. D 85, 123530
(2012)
44. M. Seikel, C. Clarkson, M. Smith, Reconstruction of dark energy and expansion dynamics
using Gaussian processes. JCAP 06, 036 (2012)
45. F. Gerardi, M. Martinelli, A. Silvestri, Reconstruction of the dark energy equation of state from
latest data: the impact of theoretical priors. JCAP 07, 042 (2019)
46. R.G. Crittenden, G.-B. Zhao, L. Pogosian, L. Samushia, X. Zhang, Fables of reconstruction:
controlling bias in the dark energy equation of state. JCAP 02, 048 (2012)
47. G.-B. Zhao, R.G. Crittenden, L. Pogosian, X. Zhang, Examining the evidence for dynamical
dark energy. Phys. Rev. Lett. 109, 171301 (2012)
48. Y. Wang, L. Pogosian, G.-B. Zhao, A. Zucca, Evolution of dark energy reconstructed from the
latest observations. Astrophys. J. Lett. 869, L8 (2018)
49. M.-X. Lin, M. Raveri, W. Hu, Phenomenology of modified gravity at recombination. Phys.
Rev. D 99(4), 043514 (2019)
50. M. Raveri, P. Bull, A. Silvestri, L. Pogosian, Priors on the effective dark energy equation of
state in scalar-tensor theories. Phys. Rev. D 96(8), 083509 (2017)
51. J. Espejo, S. Peirone, M. Raveri, K. Koyama, L. Pogosian, A. Silvestri, Phenomenology of large
scale structure in scalar-tensor theories: joint prior covariance of .wDE , .∑ and .μ in Horndeski.
Phys. Rev. D 99(2), 023512 (2019)
27 Resolving the Hubble Tension at Late Times with Dark Energy 515

52. G.W. Horndeski, Second-order scalar-tensor field equations in a four-dimensional space. Int.
J. Theor. Phys. 10, 363–384 (1974)
53. M. Raveri, W. Hu, Concordance and discordance in cosmology. Phys. Rev. D 99(4), 043506
(2019)
Chapter 28
Cosmological Models with Negative Λ .

Anjan A. Sen

28.1 Introduction

The evolution of Universe always presents us with many surprises and puzzles. Till
date, cosmologists are still trying to find answers to two biggest puzzles related to our
Universe: the dark matter (DM) and dark energy (DE). For dark matter, the answer
may lie in some form of extension of Standard model of particle physics. But for
dark energy, we currently have no clue about its actual nature originating from a field
theory prescription, instead we are working with different kind of phenomenological
characterisations of dark energy. This is due to the fact that the energy scale of the
dark energy is so abnormally low compared to any reasonable quantum field theory
scale [33, 34, 40], it is extremely difficult to model dark energy within realm of
standard model of particle physics and its various extensions.
Given this, the simplest candidate for dark energy is the cosmological constant .Λ,
the ground state or vacuum energy for the fields. Although we can not get the correct
value for .Λ (as needed to be consistent with observations) from any reasonable
field theory set up, still it is the best option due to its simple form (especially while
considering the perturbed Universe) and its unmatched consistency with almost all
the cosmological observations. This confirms the presence of a field (in the simplest
case, a canonical scalar field) at the non zero positive minimum of its potential which
confirms the presence of a stable or meta stable de-Sitter (dS) vacuum depending on
the nature of the minimum of the potential. We can also construct scenarios where
the scalar field is away from the minimum of the potential and is slowly rolling at
the positive part of the potential. This scenario gives rise to a DE behaviour, which
unlike.Λ, is not a constant but evolves with time giving rise to an evolving equation of
state .wde = pde /ρde for the DE (for .Λ, wde = −1). Such evolving DE component is
termed as “Quintessence” [17, 45]. Most of the quintessence scalar fields, discussed
in the context of DE, have a dS minima. Hence whether DE is .Λ or an evolving
scalar field, the Universe always settles in a dS phase in asymptotic future.

A. A. Sen (B)
Centre for Theoretical Physics, Jamia Millia Islamia, New Delhi 110025, India
e-mail: aasen@jmi.ac.cin

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 517
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_28
518 A. A. Sen

But this is not always true. A scalar field rolling over the positive part of a potential
which is sufficiently flat at present energy scale and also contains an Anti de-Sitter
(AdS) minima is fully consistent with a late time accelerating Universe. In fact,
constructing such potentials with AdS minima is not difficult in string theory context
as compared to that with dS minima [16, 18, 19, 43], although the right value for
the energy scale consistent with the late time Universe, has not been obtained as yet.
But still there are not many attempts in the literature for cosmological studies with
scalar fields having AdS minima.
In recent years, cosmological observations have unveiled tensions in the value
of present√day Hubble constant . H0 [6, 24, 36], the amplitude of matter fluctuations
. S8 := σ8 Ωm /0.3 that is related to the galaxy shear measurements using weak lens-
ing [1, 6, 10, 28, 44] as well as anomalies in the lensing parameter . Alens and/or
curvature .Ωk [3, 6, 25, 46]. All these tensions and anomalies are primarily present
in the concordance flat .ΛCDM model. The recent observations of very massive high
redshift galaxies by JWST are also difficult to explain under the assumption of a
concordance flat .ΛCDM model as our observable Universe [2, 13, 29].
All these indicate that the flat .ΛCDM model needs modifications. In this regard,
in recent years, cosmological models in the presence of a tiny negative cosmological
constant (with the energy density at the same order as dark energy) has been tested
with cosmological observations, particularly in the context of Hubble tension. It has
been shown that not only such models are consistent with cosmological observations,
in some cases they can be better in statistical sense, compared to flat .ΛCDM model.
Moreover, they may also help to alleviate different cosmological tensions.
With such motivation, in this chapter, we shall review the recent studies related
to cosmological models containing a negative cosmological constant and the scope
of alleviate some of the cosmological tensions in such models.

28.2 Universe with Negative .Λ

28.2.1 Possibility of Accelerating Universe with a Negative .Λ

In contrast to positive cosmological constant, a negative cosmological constant results

an attractive gravitational effect and hence can not result in an accelerated expanding
Universe on its own. But it was first shown by Hartle, Hawking and Hertog [26] that
a fundamental theory of quantum cosmology with negative cosmological constant
can give rise to classical low energy theory with positive cosmological constant that
can drive the accelerated expansion in the Universe. This follows from a underlying
symmetry of the model called “signature reversal symmetry” which is present in
both quantum and classical theories. Later it was argued by Mithani and Vilenkin
[31] that such theories are unstable as the mass of the corresponding field becomes
tachyonic.
28 Cosmological Models with Negative .Ʌ 519

28.2.2 Early Observational Hints for a Negative Dark Energy

The first observational hint that there may be a negative contribution in the Dark
Energy behaviour, especially at redshifts .z > 1 was from the Baryon Acoustic Oscil-
lation measurements in the Lyman-.α forest of high redshift quasars by BOSS DR11
[23]. With their measurement of . H (z = 2.34) = 222 ± 7 Km/s/Mpc, and using the
precise values of .rd and .Ωm h 2 by Planck, they obtained the constraint on dark energy
density at .z = 2.34:
ρde (z = 2.34)
. = −1.2 ± 0.8, (28.1)
ρde (z = 0)

which is in tension of around 2.5 .σ with a value .1 (that is needed for .Λ to be dark
energy). If this results persists, it shows that the dark energy density at .z = 2.34 is
lower than that at .z = 0 and perhaps with an opposite sign.
This negative dark energy behaviour at higher redshifts was further corroborated
by another study by Poulin, Boddy, Bird and Kamionkowski (PBBK) [35] who did
a model independent reconstruction of the dark energy density using a large set
of cosmological data including CMB, LSS, BAO, SnIa as well as Lyman-.α BAO
measurements at .z = 2.34. They showed that when one relaxes the prior .ρde > 0,
the reconstructed .ρde behaviour evolves towards negative values at higher redshifts.
Similar conclusions were arrived by Sahni, Shafieloo and Starobinsky (SSS) [39]
using the . Om diagnostics. For .ΛCDM model .Ωm h 2 is constant and is tightly con-
strained by Planck: .Ωm h 2 = 0.1426 ± 0.0025. Using the relation

. H 2 (z) = κ[ρde (z) + ρm0 (1 + z)3 ], (28.2)

together with the constraint . H (z = 2.34) = 222 ± 7 Km/s/Mpc by BOSS-DR11

with Lyman-.α measurments, one gets .h 2 (z)/(1 + z)3 = 0.132 ± 0.008 (where
.h(z) = H (z)/H (z = 0)). If one now assumes .Ωm h = 0.1426 ± 0.0025 as con-
2

strained by CMB observations by Planck, one can see .ρde (z = 2.34) < 0 (for details
see [39]) showing the possibility of a negative dark energy behaviour at higher red-
shifts. We should also mention this may also happen if the matter energy density
does not obey the standard conservation equation due to some interactions with
other component as well as for modified gravity models.
In a later work, Wang, Pogosian, Zhao and Zucca (WPZZ) [49] applied a different
model independent reconstruction method to reconstruct the dark energy evolution
using updated data from Planck as well as measurements from SnIa, BAO (including
the Lyman-.α BAO measurement by BOSS-DR11), cosmic chronometer as well as
the more recent . H0 measurement by S. H0 ES and showed the same negative dark
energy behaviour at higher redshifts.
All these studies independently confirmed the possible hint of a negative dark
energy behaviour at higher redshift primarily driven by the Lyman-.α BAO measure-
ment by BOSS-DR11.
520 A. A. Sen

28.2.3 Recent Observational Hint for the Presence

of a Negative .Λ

• Evidence of Negative Cosmological Constant from Reconstructed H(z) using

Pade Approximation
In recent years, the first observational hint for presence of a negative cosmological
constant in the energy budget of the Universe came from the work by Dutta et
al. [22]. They studied a model independent but parametrized form for the Hub-
ble parameter . H (z) in order to reconstruct the dark energy density. For this, they
assumed the Pade approximation . P2,2 for the Hubble parameter . H (z) and used
various low-redshift data to constrain the . H (z) behaviour. They obtained the simi-
lar negative dark energy behaviour at higher redshifts as obtained by earlier works
by PBBK, SSS and WPZZ discussed in the previous section. This reconstruction is
valid till .z ∼ 3 as there are no low redshift data beyond that redshift. Subsequently
they imposed the condition that this low-redshift behaviour of . H (z) should match
with the Planck constrained . H (z) behaviour for .ΛCDM at redshift .z ∼ 4 and
beyond. The rationale for such assumption is that for models with non interacting
dark energy, the Universe should behave as Einstein-de Sitter at higher redshifts
(because dark energy is negligible at higher redshifts and Universe is mostly dom-
inated by ordinary matter and radiation) irrespective of the dark energy model.
Together with this condition, the S. H0 ES prior was also imposed and the . H (z) was
reconstructed beyond .z > 4. From the reconstructed . H (z), one can now obtain
the dark energy redshift dependence as follows:

H 2 (z)
. = Ωm0 (1 + z)3 + (1 − Ωm0 ) f (z), (28.3)
H02

where . f (z) is the redshift dependence of the dark energy density. With this, one
can reconstruct the . f (z) and it was shown that . f (z) has a negative minima at
higher redshifts which can be modelled with the presence of small cosmological
constant with a value .Λ = 3H02 (1 − Ωm0 ) f min where . f min < 0. As evident, the
value of this small cosmological constant is of the same order of the dark energy
density at present and hence it does not involve any new energy scale. The authors
argued that (a) the BAO Lyman-.α data at .z = 2.34, (b) the S. H0 ES prior on . H0
(R16) as well as (c) the constraint that the. H (z) should reproduce the.ΛCDM. H (z)
behaviour as constrained by Planck-2018 at higher redshifts, all three together play
a crucial role for the presence of tiny negative cosmological constant in the energy
budget of the Universe.
Although the study involves a particular parametrization for . H (z), namely the
Pade approximation, the fact that reconstructed behaviour for .ρde (z) upto .z ∼ 3
using the low redshift data is consistent with earlier low-redshifts reconstructions
of .ρde (z) by PBBK and WPZZ, supports the validity of the reconstructed . H (z) by
Dutta et al.
28 Cosmological Models with Negative .Ʌ 521

• Dark energy with a constant Equation of state on top of negative Cosmological

Constant
Visinelli, Vagnozzi and Danielsson (VVD) [48] considered a string inspired exten-
sion for .ΛCDM model where dark energy with a constant equation of state is
considered on top of a negative cosmological constant. The equation of state for
the dark energy can be phantom in nature which can be possibly modelled through
string inspired multiverse scenario where multiple scalar fields with AdS vacua
can give an effective phantom equation of state. They used only low-redshift obser-
vational data from BAO and SnIA observations together with the two anchors: one
is .rd from Planck-2018 and the other is . H0 from S. H0 ES measurements.
There was no definite evidence for the existence of a negative cosmological con-
stant although at .68% confidence level, the constraints are consistent with the
presence of a non zero negative cosmological constant.
• What happens when one includes CMB measurement with low redshift mea-
surements?
Recently Sen, Adil and Sen (SAS) [42] have done a comprehensive analysis on
cosmological constraint for models with negative cosmological constant. Similar
to the work by VVD, they considered dark energy models with constant equation
of state top of a negative cosmological constant. They also considered dark energy
energy models with CPL parametrization for the equation of state together with a
negative cosmological constant. The role plays by negative cosmological constant
can be understood in the following way:
CMB measurement give extremely tight constraint on the angular scale.θs of sound
horizon at last scattering:
rd (z d )
.θs = , (28.4)
D A (z∗)

where .z∗ = 1100 is the redshift at last scattering and .z d ∼ z∗ is the redshift at
drag epoch. .rd is the sound horizon at .z d and . D A is the angular diameter distance
to the last scattering. As CMB gives tight constraint on .θs from the peak position,
rd (z d )
.
D A (z∗)
needs to be very tightly constrained. If there is no new physics at the early
Universe, .rd will not change also, and hence . D A is very tightly constrained. The
expression of . D A is as follows:
{ z∗
c 1 dz c 1
.DA = = F(z∗), (28.5)
H0 (1 + z) 0 E(z) H0 (1 + z)
{ z∗ dz
where . E(z) = HH(z)
0
and . F(z∗) = 0 E(z) . So with tightly constrained . D A , if we
need to increase . H0 without any change in peak locations in CMB, we need to
increase . F(z∗) proportionately. . F(z∗) contains the dark energy information at
late times. With a simple case, where dark energy is parametrized by a constant
equation of state, it is not difficult to check that a phantom type equation of state
can result the increase in . F(z∗) in comparison to .ΛCDM model. In this regard,
the presence of a negative cosmological constant plays a crucial role. As shown by
SAS [42], presence of a .−ve .Λ increases the . F(z∗) for a fixed phantom like .wde .
522 A. A. Sen

So one can obtain a required increase in . F(z∗) (in order to increase . H0 ) with less
phantom equation of state but with a certain contribution from .−ve .Λ whereas
without the contribution from .−ve .Λ, one needs much larger phantom equation
state for the dark energy, which may be difficult to obtain theoretically (see Fig. 1
in SAS [42]).
Subsequently a dark energy model with constant dark energy equation of state
together with a negative cosmological constant (WCDMCC) was confronted
with various cosmological datasets including CMB likelihood by Planck-2018
(TTTEEE+LowTT+LowEE+Lensing) [6], BAO [9, 11, 37] as well as Pantheon
data for SnIA [41] together with S. H0 ES-2019 prior (R19) [36]. With CMB+BAO
data alone, a dark energy model (with constant equation of state) together with a
+1.3
.−ve .Λ (WCDMCC) results . H0 = 69.6−1.5 Km/s/Mpc which is at much reduced
tension (.∼ 2σ ) with the S. H0 ES R19 prior for . H0 . The constraint on the .rd is
.147.17 ± 0.22Mpc showing no change in the early Universe physics. The con-
straint on the dark energy equation of state is .wde = −1.061+0.037
−0.015 showing a very
mild phantom behaviour. The constraint on the amount of negative cosmological
constant is .ΩΛ = −0.88+1.3
−0.46 . Although the zero value of .ΩΛ is consistent within
.68% confidence level, the results shows that the CMB+BAO data is fully consistent

with a dark energy model with - ve .Λ. Most interestingly, the WCDMCC model
has better evidence in terms of both AIC as well as Bayesian Evidence criteria.
As shown in the SAS paper [42], there is a reduction of around .16 in AIC for
WCDMCC model compared to .ΛCDM and an improvement of around .7 in the
Bayesian Evidence in WCDMCC model compared to .ΛCDM for the CMB+BAO
data. This shows that the WCDMCC is surely a viable cosmological model.
Subsequenty the Pantheon data for SnIA measurements as well as the prior on . H0
+0.82
by R19 were added to CMB+BAO data. This gives . H0 = 70.34−0.77 Km/s/Mpc,
+1.90 +0.035
.ΩΛ = −2.48−0.29 and .wde = −1.02−0.013 . For this full data set, the improvement

in AIC in WCDMCC model is around.8.15 compared to.ΛCDM and corresponding

improvement in Bayesian Evidence is .3.79 in WCDMCC compared to .ΛCDM.
Both of these numbers give a strong preference of WCDMCC model over .ΛCDM
when considering the full CMB+BAO+SnIA+. H0 data.
• Graduated Dark Energy
Inspired by the “Graduated Inflationary Universe” model proposed by Barrow in
1990, recently a “Graduated Dark Energy” model has been proposed by Akarsu,
Barrow, Escamilla and Vazquez (ABEV) [8]. The basic idea is to model the inertial
mass of the dark energy .ρintertia = ρ + p as .ρiner tia = γρ0 ( ρρ0 )λ . Here .γ and .λ
are two parameters of the model and .ρ0 is positive constant with its value at
present. It was claimed that this set up can describe a dark energy model where the
cosmological constant switches sign from a negative value in the early Universe
to a positive value in late Universe. While confronting with data from Planck
(compressed likelihood), SnIa, BAO (including the Lyman-.α measurement at .z −
2.34) as well as the . H0 measurement by S. H0 ES, the sign switching happens at
around .z ∼ 2.3. The model results a higher value for . H0 and results a better .χ 2
compared to.ΛCDM. This is another independent study where it was shown that the
28 Cosmological Models with Negative .Ʌ 523

presence of a small negative cosmological constant is consistent with cosmological

observations.
This sign switching Cosmological constant mode .Λs CDM was further studied in
another recent work by Akarsu et al. [7] where different cosmological observations
were taken into account including the full CMB likelihood from Planck-2018 as
well as various BAO measurements and SnIa measurement by Pantheon together
with the S. H0 ES prior. It was observed that such sign switching cosmological
constant model can address several cosmological tensions like . H0 , . S8 , . M B etc in a
single model set up. The model is also better fit to the cosmological observations
compared to .ΛCDM. This model also gives independent hint of the presence of a
negative cosmological constant in the energy budget of the Universe, especiallly
in the higher redshifts.
• Possibility of Negative Cosmological Constant in Dark Sector
This is another independent study by Calderon, Gannouji, L’Huillier and Polarski
[12]. While modelling dark energy density, they considered both the constant
equation state as well as CPL parametrization for this purpose. Moreover they
also considered various piecewise continuous functions for the equation of state
of the dark energy. Together this, they also considered the presence of a nega-
tive cosmological constant in the energy budget of the Universe. Using various
cosmological observations, they showed that the best models are those with phan-
tom behaviour at .z > 1 together with a higher evidence for the non-zero negative
cosmological constant.
• Dark Energy with Phantom Crossing: Omnipotent Dark Energy
Recently Di Valentino, Mukherjee and Sen (DMS) [47] have proposed a
parametrization for the dark energy density that involves a phantom crossing
behaviour. The parametrization is as follows:

1 + α(a − am )2 + β(a − am )3
ρ (z) = ρde (z = 0)
. de . (28.6)
1 + α(1 − am )2 + β(1 − am )3

The model has three parameters:.am ,.α and.β. The absence of the linear term ensures
that the .ρde has an extremum at the scale factor .a = am . This results the phantom
crossing at .a = am . In recent work, it have been shown that this parametrization
can unify a host of dark energy behaviours and is termed as “Omnipotent Dark
Energy”.
DMS showed that this model can result higher . H0 for both CMB+BAO as well
as CMB+SN data combinations, practically resolving the Hubble tension. Later
Chudaykin et al. [14] have shown by adding the SPT-3G [21] and SPTPOL [50]
data as well as information from BAO measurements [9, 11, 37] to the Planck-
2018 likelihood and Pantheon SnIa data, that this phantom crossing model can
resolve both . H0 and . S8 tensions simultaneosly and also can result a better fit to
the observational data compared to .ΛCDM model.
More recently, Adil et al. [4] have shown that the constrained Omnipotent dark
energy behaviour is equivalent to a phantom dark energy model together with
a small negative cosmological constant. One should note that the Omnipotent
524 A. A. Sen

parametrization for.ρde does not necessarily involve the presence of a negative cos-
mological constant. But as it is constrained by observational data, the constrained
behaviour for .ρde eventually shows the existence of a negative cosmological con-
stant which is very interesting. This is another independent hint for the existence
of a negative cosmological constant in the energy budget of the Universe.
• AdS Early Dark Energy
A cosmological phase of Early Dark Energy before recombination can also help
to resolve the Hubble Tension. Recently Jiang and Piao [27] have shown that a
short AdS period just before recombination can help the early dark energy model to
work better to solve the Hubble tension problem. With low-l Planck-TT likelihood
together with polarization data from SPTpol and lensing as well as with BAO and
Pantheon data (without SHOES) prior, it was shown an AdS-early dark energy
model can result . H0 = 71.9 Km/s/Mpc which resolves the . H0 tension. Inclusion
of galaxy shear and clustering data gives . S8 = 0.785, also resolving the . S8 tension.
• Multiple Transitions Vacuum Dark Energy
In another work, Moshafi, Firouzjahi and Talebian [32]has considered multiple
transition vacuum dark energy model, where vacuum energy undergoes multiple
transition both in early time as well as in late time. The transition can be dS
to AdS or vice versa. They showed while constraining the model using various
cosmological observations that a transient late time AdS-like vacuum energy can
give rise larger . H0 but at the same results a scale invariant Harrison Zeldovich type
priomordial power spectrum.
All these models, discussed above give independent hints for the existence of small
negative cosmological constant in the Universe. We should emphasis that modelling
the presence of a small negative cosmological constant is not difficult. The simplest
way to include a negative cosmological constant in the dark energy sector is include
an AdS minima in the quintessence field potential that gives rise to the dark energy
behaviour. One such potential was obtained by Cicoli et al. [15]. in models with
multiple axion: [ ( )]
φ
. V (φ) = V0 1 + pcos . (28.7)
f

Here for . p > 1, the potential has a AdS minima. This scalar field potential has
been confronted with variety of observational data and it was shown that . p > 1 case
which has an AdS minima in the potential, is consistent with observational data and
at the same time is better fit to the data than the .ΛCDM model [38]. But this model
can not result a phantom type dark energy. In most of the studies that resolve Hubble
tension using late time physics, it is known that phantom type dark energy is more
suitable. In such cases, this multi-axions scalar field model needs to be interacting
with other components in the Universe, like dark matter, to result an effective dark
energy equation of state that is phantom.
In another recent work, Malekjani et a. [30] have shown that, considering latest
Pantheon+ sample for the SnIa, there is a redshift evolution for the cosmological
parameters . H0 and .Ωm0 . In particular, there is an increasing .Ωm0 trend with redshift
28 Cosmological Models with Negative .Ʌ 525

that leads to .Ωm0 > 1 for .z > 1. If this is true, this may correspond to the existence
of a negative dark energy at .z > 1.
Finally, in a recent work, Demirtas et al. [20] have constructed a supersymmetric
vacuum in string theory where the magnitude of the cosmological constant is .∼
10−123 in Planck unit and it is negative. This is consistent with all the recent studies
where a small negative cosmological constant is considered to ameliorate different
cosmological tensions. This can have interesting consequences in modelling dark
energy models with small negative cosmological constant that may be a viable option
to resolve different cosmological tensions.

28.3 Summary

To summarize, presence of small negative cosmological constant in the energy budget

of the Universe is not unusual and can be modelled with a scalar field rolling over
potential that have an AdS minima. This is also very relevant in the context of dark
energy model building in string theory. As long as the scalar field is slowly rolling
at the positive part of the potential around present time and is also dominating in the
energy budget of the Universe, it can always result a late time accelerating Universe.
The non-zero AdS minima in the potential will result a small contribution to the
energy budget of the Universe in the form of a negative cosmological constant. As
shown in the various recent studies, discussed in this chapter, this contribution can
have interesting signatures in the cosmological evolution of the Universe. It can help
to resolve some of the cosmological tensions. Such models can also be statistically
a better fit of the data compared to .ΛCDM. Moreover the presence of this small
negative cosmological constant (due to its attractive gravitational effect), will reduce
the Hubble friction in the evolution of matter density fluctuations, thereby enhancing
the growth of matter fluctuations, especially at higher redshifts and the nonlinear
regime may start earlier than the usual .ΛCDM model. This may have interesting
implications in the context of recent JWST observations of very massive galaxies at
higher redshifts.
Hence cosmological models where dark energy energy contains the contribution
from a small negative cosmological constant, is definitely an interesting option in the
current context of different cosmological tensions and anomalies. Already a good
number of research works have shown the potential for such models in the context
of observational data and such models should be pursued further, especially in the
context of structure formation in the Universe.

Acknowledgements The author acknowledges the funding from SERB, Govt of India under the
research grant no: CRG/2020/004347.
526 A. A. Sen

References

1. T. Abbott, M. Aguena, A. Alarcon, S. Allam, O. Alves, A. Amon, F. Andrade-Oliveira, J. Annis,

S. Avila, D. Bacon, E. Baxter, K. Bechtol, M. Becker, G. Bernstein, S. Bhargava, S. Birrer,
J. Blazek, A. Brandao-Souza, S. Bridle, D. Brooks, E. Buckley-Geer, D. Burke, H. Camacho,
A. Campos, A.C. Rosell, M.C. Kind, J. Carretero, F. Castander, R. Cawthon, C. Chang, A.
Chen, R. Chen, A. Choi, C. Conselice, J. Cordero, M. Costanzi, M. Crocce, L. da Costa, M.
da Silva Pereira, C. Davis, T. Davis, J.D. Vicente, J. DeRose, S. Desai, E.D. Valentino, H. Diehl,
J. Dietrich, S. Dodelson, P. Doel, C. Doux, A. Drlica-Wagner, K. Eckert, T. Eifler, F. Elsner,
J. Elvin-Poole, S. Everett, A. Evrard, X. Fang, A. Farahi, E. Fernandez, I. Ferrero, A. Ferté,
P. Fosalba, O. Friedrich, J. Frieman, J. García-Bellido, M. Gatti, E. Gaztanaga, D. Gerdes, T.
Giannantonio, G. Giannini, D. Gruen, R. Gruendl, J. Gschwend, G. Gutierrez, I. Harrison, W.
Hartley, K. Herner, S. Hinton, D. Hollowood, K. Honscheid, B. Hoyle, E. Huff, D. Huterer, B.
Jain, D. James, M. Jarvis, N. Jeffrey, T. Jeltema, A. Kovacs, E. Krause, R. Kron, K. Kuehn,
N. Kuropatkin, O. Lahav, P.-F. Leget, P. Lemos, A. Liddle, C. Lidman, M. Lima, H. Lin, N.
MacCrann, M. Maia, J. Marshall, P. Martini, J. McCullough, P. Melchior, J. Mena-Fernández,
F. Menanteau, R. Miquel, J. Mohr, R. Morgan, J. Muir, J. Myles, S. Nadathur, A. Navarro-
Alsina, R. Nichol, R. Ogando, Y. Omori, A. Palmese, S. Pandey, Y. Park, F. Paz-Chinchón, D.
Petravick, A. Pieres, A.P. Malagón, A. Porredon, J. Prat, M. Raveri, M. Rodriguez-Monroy, R.
Rollins, A. Romer, A. Roodman, R. Rosenfeld, A. Ross, E. Rykoff, S. Samuroff, C. Sánchez,
E. Sanchez, J. Sanchez, D.S. Cid, V. Scarpine, M. Schubnell, D. Scolnic, L. Secco, S. Serrano,
I. Sevilla-Noarbe, E. Sheldon, T. Shin, M. Smith, M. Soares-Santos, E. Suchyta, M. Swanson,
M. Tabbutt, G. Tarle, D. Thomas, C. To, A. Troja, M. Troxel, D. Tucker, I. Tutusaus, T. Varga,
A. Walker, N. Weaverdyck, R. Wechsler, J. Weller, B. Yanny, B. Yin, Y. Zhang, J.Z And, Dark
energy survey year 3 results: cosmological constraints from galaxy clustering and weak lensing.
Phys. Rev. D 105, 2 (2022)
2. N.J. Adams, C.J. Conselice, L.Ferreira, D. Austin, J.A.A. Trussler, I. Juodzbalis, S.M. Wilkins,
J. Caruana, P. Dayal, A. Verma, A.P. Vijayan, Discovery and properties of ultra-high redshift
galaxies (.9 < z < 12) in the jwst ero smacs 0723 field. Monthly Notices of the Royal Astron.
Soc. 4755–4766
3. G.E. Addison, Y. Huang, D.J. Watts, C.L. Bennett, M. Halpern, G. Hinshaw, J.L. Weiland,
Quantifying discordance in the 2015 Planck CMB spectrum. The Astrophys. J. 818(2), 132
(2016)
4. S.A. Adil, O. Akarsu, E.D. Valentino, R.C. Nunes, E. Ozulker, A.A. Sen, E. Specogna, Omnipo-
tent dark energy: a phenomenological answer to the Hubble tension. arXiv e-prints 2306.08046
(2023)
5. S.A. Adil, U. Mukhopadhyay, A.A. Sen, S. Vagnozzi, Dark energy in light of the early jwst
observations: case for a negative cosmological constant? arXiv e-prints 2307.12763 (2023)
6. N. Aghanim, Y. Akrami, M. Ashdown, J. Aumont, C. Baccigalupi, M. Ballardini, A.J. Banday,
R.B. Barreiro, N. Bartolo, S. Basak, R. Battye, K. Benabed, J.-P. Bernard, M. Bersanelli, P.
Bielewicz, J.J. Bock, J.R. Bond, J. Borrill, F.R. Bouchet, F. Boulanger, M. Bucher, C. Burigana,
R.C. Butler, E. Calabrese, J.-F. Cardoso, J. Carron, A. Challinor, H.C. Chiang, J. Chluba,
L.P.L. Colombo, C. Combet, D. Contreras, B.P. Crill, F. Cuttaia, P. de Bernardis, G. de Zotti,
J. Delabrouille, J.-M. Delouis, E.D. Valentino, J.M. Diego, O. Doré , M. Douspis, A. Ducout,
X. Dupac, S. Dusini, G. Efstathiou, F. Elsner, T.A. Enßlin, H.K. Eriksen, Y. Fantaye, M.
Farhang, J. Fergusson, R. Fernandez-Cobos, F. Finelli, F. Forastieri, M. Frailis, A.A. Fraisse,
E. Franceschi, A. Frolov, S. Galeotta, S. Galli, K. Ganga, R.T. Génova-Santos, M. Gerbino,
T. Ghosh, J. González-Nuevo, K.M. Górski, S. Gratton, A. Gruppuso, J.E. Gudmundsson, J.
Hamann, W. Handley, F.K. Hansen, D. Herranz, S.R. Hildebrandt, E. Hivon, Z. Huang, A.H.
Jaffe, W.C. Jones, A. Karakci, E. Keihänen, R. Keskitalo, K. Kiiveri, J. Kim, T.S. Kisner, L.
Knox, N. Krachmalnicoff, M. Kunz, H. Kurki-Suonio, G. Lagache, J.-M. Lamarre, A. Lasenby,
M. Lattanzi, C.R. Lawrence, M.L. Jeune, P. Lemos, J. Lesgourgues, F. Levrier, A. Lewis, M.
Liguori, P.B. Lilje, M. Lilley, V. Lindholm, M. López-Caniego, P.M. Lubin, Y.-Z. Ma, J.F.
28 Cosmological Models with Negative .Ʌ 527

Macías-Pérez, G. Maggio, D. Maino, N. Mandolesi, A. Mangilli, A. Marcos-Caballero, M.

Maris, P.G. Martin, M. Martinelli, E. Martínez-González, S. Matarrese, N. Mauri, J.D. McEwen,
P.R. Meinhold, A. Melchiorri, A. Mennella, M. Migliaccio, M. Millea, S. Mitra, M.-A. Miville-
Deschênes, D. Molinari, L. Montier, G. Morgante, A. Moss, P. Natoli, H.U. Nørgaard-Nielsen,
L. Pagano, D. Paoletti, B. Partridge, G. Patanchon, H.V. Peiris, F. Perrotta, V. Pettorino, F.
Piacentini, L. Polastri, G. Polenta, J.-L. Puget, J.P. Rachen, M. Reinecke, M. Remazeilles, A.
Renzi, G. Rocha, C. Rosset, G. Roudier, J.A. Rubiño-Martín, B. Ruiz-Granados, L. Salvati,
M. Sandri, M. Savelainen, D. Scott, E.P.S. Shellard, C. Sirignano, G. Sirri, L.D. Spencer, R.
Sunyaev, A.-S. Suur-Uski, J.A. Tauber, D. Tavagnacco, M. Tenti, L. Toffolatti, M. Tomasi, T.
Trombetti, L. Valenziano, J. Valiviita, B.V. Tent, L. Vibert, P. Vielva, F. Villa, N. Vittorio, B.D.
Wandelt, I.K. Wehus, M. White, S.D.M. White, A. Zacchei, A. Zonca, Planck 2018 results. Vi.
Cosmological parameters. Astron. Astrophys. 641, A6 (2020)
7. O. Akarsu, S. Kumar, E. Ozulker, J.A. Vazquez, A. Yadav, Relaxing cosmological tensions
with a sign switching cosmological constant: improved results with Planck, bao and pantheon
data (2022). arXiv e-prints 2211.05742
8. Ö. Akarsu, J.D. Barrow, L.A. Escamilla, J.A. Vazquez, Graduated dark energy: Observational
hints of a spontaneous sign switch in the cosmological constant. Phys. Rev. D 101, 6 (2020)
9. S. Alam, M. Ata, S. Bailey, F. Beutler, D. Bizyaev, J.A. Blazek, A.S. Bolton, J.R. Brownstein, A.
Burden, C.-H. Chuang, J. Comparat, A.J. Cuesta, K.S. Dawson, D.J. Eisenstein, S. Escoffier, H.
Gil-Marín, J.N. Grieb, N. Hand, S. Ho, K. Kinemuchi, D. Kirkby, F. Kitaura, E. Malanushenko,
V. Malanushenko, C. Maraston, C.K. McBride, R.C. Nichol, M.D. Olmstead, D. Oravetz,
N. Padmanabhan, N. Palanque-Delabrouille, K. Pan, M. Pellejero-Ibanez, W.J. Percival, P.
Petitjean, F. Prada, A.M. Price-Whelan, B.A. Reid, S.A. Rodríguez-Torres, N.A. Roe, A.J. Ross,
N.P. Ross, G. Rossi, J.A. Rubiño-Martín, S. Saito, S. Salazar-Albornoz, L. Samushia, A.G.
Sánchez, S. Satpathy, D.J. Schlegel, D.P. Schneider, C.G. Scóccola, H.-J. Seo, E.S. Sheldon, A.
Simmons, A. Slosar, M.A. Strauss, M.E.C. Swanson, D. Thomas, J.L. Tinker, R. Tojeiro, M.V.
Magaña, J.A. Vazquez, L. Verde, D.A. Wake, Y. Wang, D.H. Weinberg, M. White, W.M. Wood-
Vasey, C. Yèche, I. Zehavi, Z. Zhai, G.-B. Zhao, The clustering of galaxies in the completed
SDSS-III Baryon oscillation spectroscopic survey: cosmological analysis of the DR12 galaxy
sample. Monthly Notices of the Royal Astron. Soc. 470(3), 2617–2652 (2017)
10. M. Asgari, C.-A. Lin, B. Joachimi, B. Giblin, C. Heymans, H. Hildebrandt, A. Kannawadi,
B. Stölzner, T. Tröster, J.L. van den Busch, A.H. Wright, M. Bilicki, C. Blake, J. de Jong, A.
Dvornik, T. Erben, F. Getman, H. Hoekstra, F. Köhlinger, K. Kuijken, L. Miller, M. Radovich,
P. Schneider, H. Shan, E. Valentijn, KiDS-1000 cosmology: cosmic shear constraints and
comparison between two point statistics. Astron. Astrophys. 645, A104 (2021)
11. F. Beutler, C. Blake, M. Colless, D.H. Jones, L. Staveley-Smith, L. Campbell, Q. Parker, W.
Saunders, F. Watson, The 6dF galaxy survey: Baryon acoustic oscillations and the local Hubble
constant. Monthly Notices of the Royal Astron. Soc. 416(4), 3017–3032 (2011)
12. R. Calderón, R. Gannouji, B. L’Huillier, D. Polarski, Negative cosmological constant in the
dark sector? Phys. Rev. D 103, 2 (2021)
13. M. Castellano, A. Fontana, T. Treu, P. Santini, E. Merlin, N. Leethochawalit, M. Trenti, E.
Vanzella, U. Mestric, A. Bonchi, D. Belfiori, M. Nonino, D. Paris, G. Polenta, G. Roberts-
Borsani, K. Boyett, Marus L15
14. A. Chudaykin, D. Gorbunov, N. Nedelko, Exploring.ɅCDM extensions with spt-3g and Planck
data: 4.σ evidence for neutrino masses and implications of extended dark energy models for
cosmological tensions (2023). arXiv e-prints 2203.03666
15. M. Cicoli, S. de Alwis, A. Maharana, F. Muia, F.. De. Quevedo, Sitter versus quintessence in
string theory. Fortschritte der Physik 67(1–2), 1800079 (2018)
16. M. Cicoli, S. de Alwis, A. Maharana, F. Muia, F.. De. Quevedo, Sitter versus quintessence in
string theory. Fortschritte der Physik 67(1–2), 1800079 (2019)
17. E.J. Copeland, M. Sami, S. Tsujikawa, Dynamics of dark energy. Int. J. Modern Phys. D 15(11),
1753–1935 (2006)
18. U.H. Danielsson, P. Koerber, T. van Riet, Universal de sitter solutions at tree-level. J. High
Energy Phys. 2010, 90 (2010)
528 A. A. Sen

19. U.H. Danielsson, G. Shiu, T. Van Riet, T. Wrase, A note on obstinate tachyons in classical dS
solutions. J. High Energy Phys. 2013, 138 (2013)
20. M. Demirtas, M. Kim, L. McAllister, J. Moritz, A. Rios-Tascon, Exponentially small cosmo-
logical constant in string theory. Phys. Rev. Lett
21. D. Dutcher, L. Balkenhol, P. Ade, Z. Ahmed, E. Anderes, A. Anderson, M. Archipley, J. Avva,
K. Aylor, P. Barry, R.B. Thakur, K. Benabed, A. Bender, B. Benson, F. Bianchini, L. Bleem,
F. Bouchet, L. Bryant, K. Byrum, J. Carlstrom, F. Carter, T. Cecil, C. Chang, P. Chaubal, G.
Chen, H.-M. Cho, T.-L. Chou, J.-F. Cliche, T. Crawford, A. Cukierman, C. Daley, T. de Haan,
E. Denison, K. Dibert, J. Ding, M. Dobbs, W. Everett, C. Feng, K. Ferguson, A. Foster, J. Fu, S.
Galli, A. Gambrel, R. Gardner, N. Goeckner-Wald, R. Gualtieri, S. Guns, N. Gupta, R. Guyser,
N. Halverson, A. Harke-Hosemann, N. Harrington, J. Henning, G. Hilton, E. Hivon, G. Holder,
W. Holzapfel, J. Hood, D. Howe, N. Huang, K. Irwin, O. Jeong, M. Jonas, A. Jones, T. Khaire,
L. Knox, A. Kofman, M. Korman, D. Kubik, S. Kuhlmann, C.-L. Kuo, A. Lee, E. Leitch, A.
Lowitz, C. Lu, S. Meyer, D. Michalik, M. Millea, J. Montgomery, A. Nadolski, T. Natoli, H.
Nguyen, G. Noble, V. Novosad, Y. Omori, S. Padin, Z. Pan, P. Paschos, J. Pearson, C. Posada,
K. Prabhu, W. Quan, S. Raghunathan, A. Rahlin, C. Reichardt, D. Riebel, B. Riedel, M. Rouble,
J. Ruhl, J. Sayre, E. Schiappucci, E. Shirokoff, G. Smecher, J. Sobrin, A. Stark, J. Stephen, K.
Story, A. Suzuki, K. Thompson, B. Thorne, C. Tucker, C. Umilta, L. Vale, K. Vanderlinde, J.
Vieira, G. Wang, N. Whitehorn, W. Wu, V. Yefremenko, K. Yoon, M.Y., Measurements of the
e-mode polarization and temperature-e-mode correlation of the CMB from SPT-3g 2018 data.
Phys. Rev. D 104, 2 (2021)
22. K. Dutta, A. Roy, A.A. Ruchika Sen, M.M. Sheikh-Jabbari, Beyond .λcdm with low and high
redshift data: implications for dark energy. General Relativity and Gravitation 52, 2 (2020)
23. T. Delubac, J.E. Bautista, J. Rich, D. Kirkby, S. Bailey, A. Font-Ribera, A. Slosar, K.G. Lee,
M.M. Pieri, J.C. Hamilton, E. Aubourg, Baryon acoustic oscillations in the ly-.α forest of BOSS
DR11 quasars. Astron. Astrophys. 574, A59 (2015)
24. W.L. Freedman, Measurements of the Hubble constant: tensions in perspective. The Astrophys.
J. 919(1), 16 (2021)
25. W. Handley, Curvature tension: evidence for a closed universe. Phys. Rev. D 103, 4 (2021)
26. J.B. Hartle, S.W. Hawking, T. Hertog, Accelerated expansion from negative .λ (2012). arXiv
e-prints 1205.3807
27. J.-Q. Jiang, Y.-S. Piao, Testing AdS early dark energy with planck, SPTpol, and LSS data.
Phys. Rev. D 104, 10 (2021)
28. S. Joudaki, C. Blake, C. Heymans, A. Choi, J. Harnois-Deraps, H. Hildebrandt, B. Joachimi, A.
Johnson, A. Mead, D. Parkinson, M. Viola, L. van Waerbeke, CFHTLenS revisited: assessing
concordance with Planck including astrophysical systematics. Monthly Notices of the Royal
Astron. Soc. 465(2), 2033–2052 (2016)
29. I. Labbé, P. van Dokkum, E. Nelson, R. Bezanson, K.A. Suess, J. Leja, G. Brammer, K.
Whitaker, E. Mathews, M. Stefanon, B. Wang, A population of red candidate massive galaxies
.∼ 600 myr after the big bang. Nature 616(7956), 266–269 (2023)
30. M. Malekjani, R.M. Conville, E.O. Colgain, S. Pourojaghi, M.M. Sheikh-Jabbari, Negative dark
energy density from high redshift pantheon+ supernovae (2023). arXiv e-prints 2301.12725
31. A.T. Mithani, A. Vilenkin, Inflation with negative potentials and the signature reversal sym-
metry. J. Cosmol. Astroparticle Phys. 2013(04), 024–024 (2013)
32. H. Moshafi, H. Firouzjahi, A. Talebian, Multiple transitions in vacuum dark energy and . H0
tension. The Astrophys. J. 940(2), 121 (2022)
33. T. Padmanabhan, Cosmological constant-the weight of the vacuum. Phys. Rep. 380(5–6), 235–
320 (2003)
34. P.J.E. Peebles, B. Ratra, The cosmological constant and dark energy. Rev. Mod. Phys. 75,
559–606 (2003)
35. V. Poulin, K.K. Boddy, S. Bird, M. Kamionkowski, Implications of an extended dark energy
cosmology with massive neutrinos for cosmological tensions. Phys. Rev. D 97, 12 (2018)
36. A.G. Riess, W. Yuan, L.M. Macri, D. Scolnic, D. Brout, S. Casertano, D.O. Jones, Y. Murakami,
G.S. Anand, L. Breuval, T.G. Brink, A.V. Filippenko, S. Hoffmann, S.W. Jha, W.D. Kenworthy,
28 Cosmological Models with Negative .Ʌ 529

J. Mackenty, B.E. Stahl, W. Zheng, A comprehensive measurement of the local value of the
Hubble constant with 1 km/s/mpc uncertainty from the Hubble space telescope and the S. H0 ES
team. The Astrophys. J. Lett. 934(1), L7 (2022)
37. A.J. Ross, L. Samushia, C. Howlett, W.J. Percival, A. Burden, M. Manera, The clustering of
the SDSS DR7 main galaxy sample—I. A 4 percent distance measure at z = 0.15. Monthly
Notices of the Royal Astron. Soc. 449(1), 835–847 (2015)
38. Ruchika, S.A. Adil, K. Dutta, A. Mukherjee, A.A. Sen, Observational constraints on axion(s)
dark energy with a cosmological constant. Phys. Dark Univ. 40, 101199 (2023)
39. V. Sahni, A. Shafieloo, A.A. Starobinsky, Model-independent evidence for dark energy evolu-
tion from Baryon acoustic oscillations. The Astrophys. J. Lett. 793(2), L40 (2014)
40. V. Sahni, A. Starobinsky, The case for a positive cosmological .Ʌ-term. Int. J. Modern Phys. D
9(4), 373–443 (2000)
41. D.M. Scolnic, D.O. Jones, A. Rest, Y.C. Pan, R. Chornock, R.J. Foley, M.E. Huber, R. Kessler,
G. Narayan, A.G. Riess, S. Rodney, E. Berger, D.J. Brout, P.J. Challis, M. Drout, D. Finkbeiner,
R. Lunnan, R.P. Kirshner, N.E. Sanders, E. Schlafly, S. Smartt, C.W. Stubbs, J. Tonry, W.M.
Wood-Vasey, M. Foley, J. Hand, E. Johnson, W.S. Burgett, K.C. Chambers, P.W. Draper, K.W.
Hodapp, N. Kaiser, R.P. Kudritzki, E.A. Magnier, N. Metcalfe, F. Bresolin, E. Gall, R. Kotak, M.
McCrum, K.W. Smith, The complete light-curve sample of spectroscopically confirmed SNe
Ia from Pan-STARRS1 and cosmological constraints from the combined Pantheon sample.
Astrophys. J. 859(2), 101 (2018)
42. A.A. Sen, S.A. Adil, S. Sen, Do cosmological observations allow a negative.λ? Monthly Notices
of the Royal Astron. Soc. 518(1), 1098–1105 (2022)
43. E. Silverstein, Simple de sitter solutions. Phys. Rev. D 77, 106006 (2008)
44. M. Troxel, N. MacCrann, J. Zuntz, T. Eifler, E. Krause, S. Dodelson, D. Gruen, J. Blazek,
O. Friedrich, S. Samuroff, J. Prat, L. Secco, C. Davis, A. Ferté, J. DeRose, A. Alarcon, A.
Amara, E. Baxter, M. Becker, G. Bernstein, S. Bridle, R. Cawthon, C. Chang, A. Choi, J.D.
Vicente, A. Drlica-Wagner, J. Elvin-Poole, J. Frieman, M. Gatti, W. Hartley, K. Honscheid, B.
Hoyle, E. Huff, D. Huterer, B. Jain, M. Jarvis, T. Kacprzak, D. Kirk, N. Kokron, C. Krawiec, O.
Lahav, A. Liddle, J. Peacock, M. Rau, A. Refregier, R. Rollins, E. Rozo, E. Rykoff, C. Sánchez,
I. Sevilla-Noarbe, E. Sheldon, A. Stebbins, T. Varga, P. Vielzeuf, M. Wang, R. Wechsler, B.
Yanny, T. Abbott, F. Abdalla, S. Allam, J. Annis, K. Bechtol, A. Benoit-Lévy, E. Bertin, D.
Brooks, E. Buckley-Geer, D. Burke, A.C. Rosell, M.C. Kind, J. Carretero, F. Castander, M.
Crocce, C. Cunha, C. D’Andrea, L. da Costa, D. DePoy, S. Desai, H. Diehl, J. Dietrich, P.
Doel, E. Fernandez, B. Flaugher, P. Fosalba, J. García-Bellido, E. Gaztanaga, D. Gerdes, T.
Giannantonio, D. Goldstein, R. Gruendl, J. Gschwend, G. Gutierrez, D. James, T. Jeltema, M.
Johnson, M. Johnson, S. Kent, K. Kuehn, S. Kuhlmann, N. Kuropatkin, T. Li, M. Lima, H. Lin,
M. Maia, M. March, J. Marshall, P. Martini, P. Melchior, F. Menanteau, R. Miquel, J. Mohr,
E. Neilsen, R. Nichol, B. Nord, D. Petravick, A. Plazas, A. Romer, A. Roodman, M. Sako, E.
Sanchez, V. Scarpine, R. Schindler, M. Schubnell, M. Smith, R. Smith, M. Soares-Santos, F.
Sobreira, E. Suchyta, M. Swanson, G. Tarle, D. Thomas, D. Tucker, V. Vikram, A. Walker, J.
Weller, Y.Z., Dark energy survey year 1 results: cosmological constraints from cosmic shear.
Phys. Rev. D 98, 4 (2018)
45. S. Tsujikawa, Quintessence: a review. Class. Quantum Gravity 30(21), 214003 (2013)
46. E.D. Valentino, A. Melchiorri, J. Silk, Planck evidence for a closed universe and a possible
crisis for cosmology. Nat. Astron. 4(2), 196–203 (2019)
47. E.D. Valentino, A. Mukherjee, A.A. Sen, Dark energy with phantom crossing and the . H0
tension. Entropy 23(4), 404 (2021)
48. L. Visinelli, S. Vagnozzi, U. Danielsson, Revisiting a negative cosmological constant from
low-redshift data. Symmetry 11(8), 1035 (2019)
49. Y. Wang, L. Pogosian, G.-B. Zhao, A. Zucca, Evolution of dark energy reconstructed from the
latest observations. The Astrophys. J. 869(1), L8 (2018)
50. W.L.K. Wu, L.M. Mocanu, P.A.R. Ade, A.J. Anderson, J.E. Austermann, J.S. Avva, J.A. Beall,
A.N. Bender, B.A. Benson, F. Bianchini, L.E. Bleem, J.E. Carlstrom, C.L. Chang, H.C. Chiang,
R. Citron, C.C. Moran, T.M. Crawford, A.T. Crites, T. de Haan, M.A. Dobbs, W. Everett, J.
530 A. A. Sen

Gallicchio, E.M. George, A. Gilbert, N. Gupta, N.W. Halverson, N. Harrington, J.W. Henning,
G.C. Hilton, G.P. Holder, W.L. Holzapfel, Z. Hou, J.D. Hrubes, N. Huang, J. Hubmayr, K.D.
Irwin, L. Knox, A.T. Lee, D. Li, A. Lowitz, A. Manzotti, J.J. McMahon, S.S. Meyer, M. Millea,
J. Montgomery, A. Nadolski, T. Natoli, J.P. Nibarger, G.I. Noble, V. Novosad, Y. Omori, S.
Padin, S. Patil, C. Pryke, C.L. Reichardt, J.E. Ruhl, B.R. Saliwanchik, J.T. Sayre, K.K. Schaffer,
C. Sievers, G. Simard, G. Smecher, A.A. Stark, K.T. Story, C. Tucker, K. Vanderlinde, T. Veach,
J.D. Vieira, G. Wang, N. Whitehorn, V. Yefremenko, A measurement of the cosmic microwave
background lensing potential and power spectrum from .500deg 2 of SPTpol temperature and
polarization data. The Astrophys. J. 884(1), 70 (2019)
Chapter 29
On the Interacting Dark Energy
Scenarios—The Case for Hubble
Constant Tension

Supriya Pan and Weiqiang Yang

29.1 Introduction

The dark sector of our universe is highly mysterious and it has become a very appeal-
ing topic for modern cosmology and cosmologists. According to a large amount of
observational evidences, the dark sector of the universe is comprised of two heavy
fluids, namely, dark matter (DM) and dark energy (DE) together which occupy nearly
96% of the total energy budget of our universe. The DM sector is responsible for the
observed structure formation of our universe while the role of DE is to drive the accel-
erating expansion of the universe [1, 2]. While the origin, nature and evolution of this
dark picture is not fully understood at this moment, but on the report of a large number
of astronomical observations, .Λ-Cold Dark Matter model, a very simple cosmolog-
ical model constructed in the context of Einstein’s General Relativity (GR) where
.Λ > 0 acts as the source of DE and DM being pressure-less, has been quite success-

ful in describing the current observational evidences. In this simplest cosmological

picture, DE and DM are conserved independently, that means, evolution of DM does
not affect the evolution of DE, and vice-versa. Despite marvellous successes,.ΛCDM
cosmological model has some limitations. The cosmological constant problem is one
of the biggest issues that still needs an answer [3]. On the other hand, the cosmic
coincidence problem (the so-called “why now” problem) [4] is another cosmolog-
ical issue associated with this cosmological model. Apart from the above issues, it
was reported that the estimations of some key cosmological parameters using the
cosmic microwave background (CMB) observations by Planck (within the .ΛCDM
paradigm) are in tension at many standard deviations with the local measurements.

S. Pan (B)
Department of Mathematics, Presidency University, 86/1 College Street, 700073 Kolkata, India
e-mail: supriya.maths@presiuniv.ac.in
W. Yang
Department of Physics, Liaoning Normal University, 116029 Dalian, People’s Republic of China
e-mail: d11102004@163.com

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 531
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_29
532 S. Pan and W. Yang

In particular, the measurements of the Hubble constant by the cosmic microwave

background observations by Planck within the .ΛCDM paradigm (. H0 = 67.4 ± 0.5
km/s/Mpc at 68% CL) [5] and S. H0 ES (Supernovae and . H0 for the Equation of State
of dark energy) collaboration (. H0 = 73.0 ± 1.0 km/s/Mpc at 68% CL, labeled as
R21) [6] are in tension with .≳ 5σ confidence. Additionally, some further estimations
by the S. H0 ES team indicate that this tension can increase up to .5.3σ [7]. Moreover,
the . S8 parameter, defined as a combination of the amplitude√of the matter power
spectrum .σ8 with the matter density at present .Ωm (. S8 = σ8 Ωm /0.3) exhibits a
tension at the level of 2–3 .σ between Planck (within .ΛCDM paradigm) and low
redshift probes, e.g. weak gravitational lensing and galaxy clustering (e.g. [8–17]).
Thus, even if .ΛCDM is extremely successful with a series of observational data,
however, based on the aforementioned limitations, it is argued that perhaps .ΛCDM
is an approximation of a more realistic scenario that is yet to be discovered. Follow-
ing this, several modifications of the .ΛCDM cosmological scenario were proposed
by many authors, see the recent reviews [18–21].
In this chapter we focus on a very generalized cosmological scenario where DE and
DM interact with each other in a non-gravitationally way. That means an exchange
of energy and/or momentum between DE and DM is allowed. This class of cosmo-
logical models is known as Interacting DE (or coupled DE or Interacting DE-DM).
According to the historical records, the idea of interaction between gravity and a
scalar field with exponential potential was proposed by Wetterich to explain the
tiny value of the cosmological constant [22]. However, in Wetterich’s proposal [22],
explicit DE-DM interaction was not introduced. Soon after the discovery of the late-
time accelerating expansion of our universe, when the dynamical DE in the name
of quintessence was proposed and the cosmic coincidence problem emerged in the
cosmological domain [4], Amendola wrote an article on the “Coupled quintessence”
[23] in which he formally introduced an interaction between the matter sector and
a scalar field with exponential potential acting as a source of DE where the author
pointed out that such an interaction can solve the long debating “cosmic coincidence”
problem. This coincidence problem was further investigated by other authors and it
was found that interacting DE models can excellently solve the cosmic coincidence
problem [24–30]. This was the first achievement of the theory of interacting DE
and this influenced various investigators to work with the interacting DE models. As
a result, over the last several years a number of interacting DE models have been
investigated with some interesting outcomes [31–63, 65, 66] (also see the review
articles on the interacting DE scenarios [67, 68]). In particular, it was observed that
an interaction between DE-DM can lead to a phantom DE scenario [69–72] without
invoking any scalar field theory with negative kinetic term which may lead to insta-
bilities in classical and quantum levels. In the context of the cosmological tensions,
interacting DE plays a very promising role. There are many evidences showing that
an interaction between DE and DM can increase the . H0 value and thus such sce-
narios can alleviate the . H0 tension between Planck (within the .ΛCDM paradigm)
and S. H0 ES collaborations [73–82]. It has been also noticed that the interacting DE
has the potentiality to alleviate the . S8 tension [74, 78, 83, 84]. Based on the above
observations, it is clear that the cosmological scenarios allowing a non-gravitational
29 On the Interacting Dark Energy Scenarios—The Case for Hubble Constant Tension 533

interaction between DE and DM are highly appealing and they should get more
attention in the forthcoming years.
This chapter is organized in the following way. In Sect. 29.2 we give an overview
of the interacting DE-DM models and present the evolution equations of a general
interaction model at the level of background and perturbations. In Sect. 29.3 we
discuss various models of interaction that have been investigated in the literature to
alleviate the Hubble tension. In particular, in Sect. 29.3.1 we discuss the interacting
models in which the DE equation of state is constant; in Sect. 29.3.2 we present those
results where the DE equation of state is dynamical; and in Sect. 29.3.3 we discuss a
special kind of interaction models in which the coupling parameter of the interaction
scenarios is assumed to be time dependent. In Sect. 29.5 we conclude the chapter
with a brief summary of the entire work.

29.2 Interacting DE

In the large scales, our universe is almost homogeneous and isotropic. This homoge-
neous and isotropic configuration of the universe is well described by the Friedmann-
Lemaître-Robertson-Walker (FLRW) line element
[ ]
dr 2
.ds 2 = −dt 2 + a 2 (t) + r 2 (dθ 2 + sin2 θ dφ 2 ) (29.1)
1 − kr 2

where (.t, r, θ, φ) are the co-moving coordinates; .a(t) is the expansion scale factor of
the universe; .k denotes the spatial geometry of the universe which can take three dis-
tinct values, namely, .0, +1, −1 corresponding to three distinct spatial geometries of
the universe—flat (.k = 0), closed (.k = +1) and open (.k = −1). We further assume
that the gravitational sector of the universe is well described by GR and the mat-
ter distribution is minimally coupled to gravity. In the matter distribution, one can
consider a variety of components, such as radiation, baryons, neutrinos, dark mat-
ter, dark energy. Now, within this framework, the Einstein’s gravitational equations
. G μν = 8π G Tμν can be recast as

k 8π G ∑
. H2 + 2
= ρi , (29.2)
a(t) 3 i
k ∑
.2 Ḣ + 3 H + = −8π pi ,
2
G (29.3)
a(t)2 i

where an overhead dot represents the derivative with respect to the cosmic time;
H ≡ ȧ(t)/a(t) is the Hubble rate of the FLRW universe;
.
∑. pi and .ρi∑are respectively
the pressure and energy density of the .i-th fluid. Hence . i ρi and . i pi denote the
total energy density and total pressure of the matter distribution, respectively. Using
534 S. Pan and W. Yang

the Bianchi’s identity .∇ μ Tμν = 0, or, alternatively, using Eqs. (29.2) and (29.3), one
can derive the conservation equation of the total fluid as follows
∑ ∑
. ρ̇i + 3 H (ρi + pi ) = 0. (29.4)
i i

We assume that DM and DE are coupled to each other, but the remaining fluids
are independently conserved. That means, in the context of the FLRW universe, the
conservation equations for DM and DE become,

ρ̇
. dm + 3 H (1 + wdm )ρdm = −Q(t), (29.5)
ρ̇ + 3 H (1 + wde )ρde = +Q(t),
. de (29.6)

where .ρdm and .ρde are respectively the energy density of DM and DE; .wdm ≥ 0,
.wde < −1/3, are respectively the barotropic equation of state parameters of DM and
DE. The quantity . Q denotes the rate of energy or/and momentum exchange between
these dark fluids and it is the heart of the theory of interacting dark energy. For
. Q(t) > 0, the energy flows from DM to the DE while for. Q(t) < 0, energy flow takes

place from DE to DM. In the limiting case . Q → 0, the non-interacting cosmological

scenario as a special case is recovered. Now, given a specific interaction function,
one can in principle determine the dynamics of the universe at the background level
using the conservation Eqs. (29.5), (29.6) together with the Hubble equation (29.2).
The interaction function also affects the evolution of the universe at the perturbation
level. However, there is no such guiding principle to select the interaction func-
tion. Although several attempts have been made to derive the interaction function
from some action principle [85–94], but no decisive conclusion has been made yet.
Thus, in most of the cases, the choice of the interaction function is made from the
phenomenological ground and hence there is no unique choice of the interaction
function. The existing interaction functions are usually divided into two classes: the
interaction functions including explicitly the Hubble rate . H of the FLRW universe,
e.g. . Q(t) = H f (ρdm , ρde ), where . f is an analytic function of the energy densities of
the dark components. A general form of the interaction function in this category that
α β
recovers many interactions is the following . Q = 3H ξρdm ρde (ρdm + ρde )γ , where
.α + β + γ = 1 and .ξ is a dimensionless coupling parameter. Note that for different
values of .(α, β, γ ), one can produce a cluster of interaction functions with only one
coupling parameter. The well known interaction function . Q = 3H ξρde is obtained
for .α = 0 = γ . We stress that the above choice is a generalized interaction function
in a specific direction and this does not include other kind of interaction models,
e.g. the interaction models with two coupling parameters, see for instance [49, 57],
or the interaction functions involving the exponential or trigonometric functions
[66]. Secondly, one can consider another class of interaction functions where the
Hubble rate does not appear explicitly, i.e. of the form . Q(t) = f (ρdm , ρde ). A gen-
α β
eralized interaction function can be of the following: . Q = Γ ρdm ρde (ρdm + ρde )γ ,
where .α + β + γ = 1 but here .Γ has the dimension equal to the dimension of the
29 On the Interacting Dark Energy Scenarios—The Case for Hubble Constant Tension 535

Hubble parameter.1 It has been argued in many places, see for instance, Ref. [95] that
as the interaction between DE and DM is a local phenomenon, therefore, it should
depend only on the local quantities and the appearance of the Hubble parameter . H ,
the global expansion rate, in the interaction functions is mainly designed for the
mathematical simplicity. While as argued in Ref. [96], explicit appearance of . H in
the interaction functions can be motivated from some well established cosmological
theories. Thus, the mathematical simplicity does not strongly hold in such cases. On
the other hand, it has been observed in Ref. [97] that a particular choice of the inter-
action function may lead to negative energy density of DM or (and) DE component
either in the past or in the future. This is a striking issue since the consequences of
the negative energy density of the dark components are not clearly understood at this
moment.
Notice that depending on the nature of the equation of state parameters, namely,
.wdm , .wde , various interacting scenarios can be considered. In most of the cases, .wdm
is set to zero due to the abundances of the cold DM in the universe sector2 , and the
nature of .wde remains unspecified, that means it can be either constant or dynamical.
Notice that the conservation Eqs. (29.5) and (29.6) can alternatively be expressed
as

ρ̇
. dm + 3 H (1 + wdm
eff
)ρdm = 0, (29.7)
.ρ̇de + 3 H (1 + wde )ρde = 0,
eff
(29.8)

eff eff
where .wde and .wdm are termed as the effective equation-of-state parameters of DM
and DE, and they are given by

Q(t)
.
eff
wdm = wdm + . (29.9)
3Hρdm
Q(t)
.
eff
wde = wde − . (29.10)
3Hρde

Thus, we see that even though we started with an interacting scenario between DE
and DM quantified through the interaction function. Q(t), but this interacting scenario
can always be expressed as a non-interacting scenario described by the conservation
equations in Eqs. (29.7) and (29.8) in terms of two effective equation of state param-
eters of DM and DE displayed in Eqs. (29.9) and (29.10). Now, if the DM sector is
cold, i.e. .wdm = 0, then .wdm
eff
= 3Hρ
Q(t)
dm
. We remark that the presence of . Q(t) in eqns.
(29.9) and (29.10) can significantly affect their nature. For example, in an expanding
universe (. H > 0), an interacting quintessence model in which .wde > −1 can offer
an effective phantom-like behaviour (.wde < −1) provided . Q(t) > 0. This is indeed

1 In the statistical analysis, we thus consider the dimensionless parameter .Γ /H0 .

2 However, as the nature of DE and DM is still yet to be discovered, and since the sensitivity of the
observational data is increasing over time, therefore, there is certainly no guiding rule to consider
.wdm = 0. In fact, one may consider .wdm to be a free parameter in the underlying cosmological
scenarios and let the observational data to decide its fate, see for instance, Refs. [98–104].
536 S. Pan and W. Yang

an interesting observation since a phantom equation-of-state in DE can increase the

expansion rate of the universe faster and as a result, the Hubble constant can take
higher values [105]. In this context, an interaction which can offer a phantom DE
equation of state through (29.10) can do this job.
As we noted, the evolution of the interacting scenario at the background level
is affected by presence of the interaction function . Q(t). Similarly, the evolution
of the interacting scenario at the level of perturbations is also affected due to the
presence of . Q(t). Following the perturbed FLRW metric in any gauge [106–108],
one can determine how the interacting scenario behaves in presence of the interaction
function, see for instance Refs. [41, 42, 95].

29.3 Different Interacting Scenarios and Their Roles

in Alleviating the . H0 Tension

As already argued, the interaction function, . Q, plays the key role in the interacting
dynamics. However, as the choice of . Q is not unique, therefore, one can freely
choose any interaction model and test its viability with the observational data. In
the following we describe how the interaction models help in alleviating the Hubble
constant tension. Throughout this section we have assumed that the curvature of the
universe is zero.

29.3.1 IDE with Constant Equation of State of DE

The interacting DE scenarios with constant equation of state in DE,.wde , have received
tremendous attention in the community. In this context, the most simplest scenario
is the one where the equation of state of DE mimics the cosmological constant, i.e.
.wde ∼ −1. On the other hand, one can assume a constant .wde other than .−1 and
investigate the consequences of the underlying interacting scenarios. In this section
we shall discuss how different interacting models with constant .wde return different
values of . H0 and in all cases we consider the spatial flatness of the FLRW universe.
In Ref. [82], an interaction between DE and CDM has been investigated using
multiple CMB datasets from various experiments where the DE equation of state has
been fixed at .wde = −0.999 and the interaction function has the following form

. Q = H ξρde , (29.11)

where .ξ is the coupling parameter of the interaction function. Using different CMB
data from multiple surveys, the constraints on . H0 are found to be as follows [82]:
for Planck 2018 alone, . H0 = 71.6 ± 2.1 km/s/Mpc at 68% CL and hence the tension
with respect to R21 is reduced down to .0.6 σ ; . H0 = 72.6+3.4
−2.6 km/s/Mpc at 68% CL
29 On the Interacting Dark Energy Scenarios—The Case for Hubble Constant Tension 537

(for ACT)3 and hence the tension with respect to R21 is reduced down to .0.1σ which
means in this case the . H0 tension is solved; for ACT+WMAP, we find . H0 = 71.3+2.6−3.2
km/s/Mpc at 68% CL and hence the tension is reduced down to .0.6σ ; and for Planck
+2.5
2018+ACT, . H0 = 71.4−2.8 km/s/Mpc at 68% CL and in this case the tension is
reduced down to .0.6σ . We note that in all the cases, we find an evidence of .ξ /= 0 at
more than 68% CL for all these datasets [82].
In Ref. [64], the authors investigated an interaction scenario between CDM and DE
with the same interaction function (29.11), but .wde was fixed at .−1 which exactly
mimics the vacuum energy sector. In this case, Planck 2018 alone leads to . H0 =
+4.26
70.84−2.50 km/s/Mpc at 68% CL with an evidence of a non-zero coupling (.ξ =
+0.142
0.132−0.077 at 68% CL) at more than 68% CL, and hence with respect to R21, the
tension on . H0 is reduced down to .0.6σ . This interacting scenario is quite appealing
because even if in presence of massive neutrinos the tension in . H0 is significantly
alleviated [64].
In Ref. [65], another interacting scenario between CDM and DE with .wde = −1
has been investigated for the interaction function

. Q = Γ ρde , (29.12)

which does not involve the Hubble parameter but it depends on the intrinsic nature
of the dark sector, e.g. the energy density of the DE density. Here, .Γ is the coupling
parameter of the interaction function having the dimension of the Hubble parameter
and hence .Γ /H0 which is dimensionless. For CMB alone from Planck 2018, this
interaction model leads to . H0 = 70.3+3.3
−2.0 km/s/Mpc at 68% CL which thus alleviates
+1.4
the tension with R21 at .1σ . For Planck 2018+BAO, . H0 = 69.0−1.8 km/s/Mpc at 68%
CL, and as a consequence, the tension with R21 is reduced down to .2.1σ .
In Ref. [81], the authors considered the same interaction function (29.11) describ-
ing an interaction between CDM and DE where.wde /= −1 is a constant and it could lie
either in the quintessence regime (.wde > −1) or in the phantom regime .(wde < −1).
For the quintessence case, Planck 2018 alone leads to . H0 = 69.8+4 −2.5 km/s/Mpc at
68% CL (hence, the tension with R21 is reduced down to .0.9σ ) with an evidence
of a non-zero coupling in the dark sector at more than .3σ significance [81]. When
.wde has a phantom nature, then the 68% CL lower bound on the Hubble constant
is found to be, . H0 > 70.4 km/s/Mpc with .wde = −1.59+0.18−0.33 (at 68% CL for Planck
2018 alone). In this case the resolution of the . H0 tension appears due to the strong
degeneracy between . H0 -.wde , not due to the interaction between CDM and DE.
In Ref. [61], a special kind of an interaction scenario between CDM and DE
(with constant .wde ) in which the interaction function may change its sign during
the evolution of the universe, known as sign changeable interaction function4 was
investigated for two variants of the interaction function. The interacting scenarios

3 Here ACT refers to the full Atacama Cosmology Telescope temperature and polarization DR4
likelihood [109], assuming a conservative Gaussian prior on .τ = 0.065 ± 0.015 as in [110].
4 For sign changeable interaction functions we refer to Refs. [111–115].
538 S. Pan and W. Yang

with .wde /= −1 are found to increase the . H0 values and therefore here we shall report
the constraints for this particular case. Considering the following interaction function

. Q = 3 H ξ(ρdm − ρde ), (29.13)

where.ξ is the coupling parameter of the interaction function, it was found that Planck
+0.93
2015+BAO returns . H0 = 69.12−1.39 km/s/Mpc at 68% CL [61] which thus alleviates
the tension with R21 at .2.5σ . For a generalized interaction function of this type

. Q = 3 H (αρdm − βρde ), (29.14)

where .α, .β (.α /= β) are the coupling parameters of this interaction function and they
will have the same sign, Planck 2015+BAO gives . H0 = 69.81+1.18 −1.39 km/s/Mpc at 68%
CL, and hence the tension with R21 is reduced down to .2σ. The sign changeable
interaction functions did not get much attention in the literature without any specific
reason, however, they deserve further attention.
In Ref. [66], an interacting scenario between CDM and DE with a constant .wde
other than .−1, has been constrained with the following interaction function
( )
ρde
. Q = 3 H ξρde sin −1 , (29.15)
ρdm

where .ξ is the coupling parameter of the interaction function. The analysis with
Planck 2018 gives . H0 = 72.67+5.43
−8.26 km/s/Mpc at 68% CL and Planck 2018+BAO
gives. H0 = 69.17+1.52
−1.71 km/s/Mpc at 68% CL [66]. Hence, for Planck 2018 and Planck
2018+BAO, the tension on . H0 with respect to R21, is reduced down to .0.05σ , .2σ ,
respectively. That means, for Planck 2018 alone case, the tension is completely
solved. While it is important to mention that the resolution of the tension in this case
is also influenced by the uncertainties.
In Ref. [104], an interacting scenario between a non-cold DM characterized by a
non-null equation of state of DM, .wdm (≥ 0), and a DE component mimicking the
vacuum energy sector, i.e. .wde = −1, has been explored for the interaction function
. Q = 3H ξρde of (29.11). Unlike other interacting scenarios in which DM is assumed
to be pressureless or cold, here, .wdm is considered to be a free-to-vary parameter in
.[0, 1] and the cosmological probes are allowed to decide its fate. This generalized

interacting scenario is labeled as IWDM. We present the results in Table 29.1 and
Fig. 29.1 for different datasets. We notice a mild evidence of .wdm /= 0 for all the
datasets at 68% CL, while at 95% CL, this evidence goes away. In addition, an
evidence of.ξ /= 0 is also found from Planck 2018 and Planck 2018+Lensing datasets
together with higher values of . H0 in both the datasets: . H0 = 70.6+4.3
−2.4 km/s/Mpc
at 68% CL for Planck 2018 and . H0 = 70.3+4.3 −2.8 km/s/Mpc at 68% CL for Planck
2018+Lensing. Therefore, the tension in . H0 with respect to R21 is reduced down to
.0.7σ for both Planck 2018 and Planck 2018+Lensing.
Table 29.1 Summary of the observational constraints on the interacting scenario between a non-cold DM and vacuum considering various observational
datasets. Here, the free parameters of this interacting scenario are as follows: .Ωb h 2 is the physical baryon density; .Ωdm h 2 is the non-cold DM density; .θ MC
is ratio of the sound horizon to the angular diameter distance; .τ refers to the optical depth; .n s is the spectral index; the amplitude of the primordial scalar
perturbations is denoted by. As . Among the derived parameters,. H0 ,. S8 are already defined;.Ωm refers to the density parameter of matter (non-cold DM+baryons)
at present time and .rdrag refers to the comoving size of the sound horizon at .z drag (redshift at which baryon-drag optical depth equals unity). The table has been
taken from Ref. [104]
Parameters Planck 2018 Planck 2018+lensing Planck 2018+BAO Planck 2018+Pantheon Planck 2018+Lens-
ing+BAO+Pantheon

.Ωb h 2 .0.02248+0.00017+0.00035 .0.02249+0.00016+0.00032 +0.00016+0.00032 .0.02247+0.00016+0.00033 +0.00015+0.00031

−0.00017−0.00033 −0.00016−0.00032 .0.02250−0.00016−0.00031 −0.00017−0.00032 .0.02252−0.00015−0.00030

.Ωdm h2 .0.077+0.033 +0.055 < 0.14

.0.082−0.036 +0.023+0.036 .0.109+0.013+0.022 +0.011+0.020
−0.056 < 0.14 .0.104−0.015−0.040 −0.011−0.024 .0.112−0.009−0.020

.100θ MC +0.0024+0.0049
.1.0436−0.0036−0.0046 .1.0433+0.0020+0.0051 +0.0008+0.0024 .1.04147+0.00065+0.0015 .1.04137+0.00054+0.0012
−0.0037−0.0043 .1.0419−0.0014−0.0021 −0.00076−0.0014 −0.00066−0.0012
.τ +0.0075+0.015 .0.0526+0.0072+0.015 .0.0541+0.0077+0.016 .0.0533+0.0075+0.015 .0.0539+0.0070+0.015
.0.0534−0.0074−0.015 −0.0072−0.014 −0.0075−0.015 −0.0074−0.015 −0.0071−0.014
.n s .0.9628+0.0046+0.0093 .0.9636+0.0042+0.0082 .0.9637+0.0042+0.0083 +0.0044+0.0088 +0.0039+0.0078
−0.0046−0.0088 −0.0042−0.0082 −0.0042−0.0083 .0.9623−0.0044−0.0089 .0.9646−0.0039−0.0078
+0.015+0.030 +0.014+0.029 +0.016+0.032 +0.015+0.031 +0.014+0.028
.ln(1010 As ) .3.046−0.016−0.030 .3.043−0.014−0.028 .3.046−0.016−0.031 .3.046−0.015−0.030 .3.044−0.014−0.027

.wdm +0.00053 < 0.0025

.0.00122−0.00097 +0.00047 < 0.0023
.0.00112−0.00091 .0.00115+0.00045
−0.00098 < 0.0025 .0.00123+0.00052
−0.00099 < 0.0031 .0.00108+0.00040
−0.00096 < 0.0023
.ξ +0.14+0.17
.0.11−0.07−0.20 +0.14+0.17
.0.10−0.08−0.19 +0.053+0.12
.0.048−0.064−0.11 +0.036+0.072
.0.036−0.036−0.072 .0.025+0.032+0.063
−0.032−0.063
.Ωm .0.21+0.07+0.19 +0.08+0.18 +0.057+0.10 .0.285+0.033+0.066 .0.289+0.028+0.055
−0.14−0.16 .0.22−0.14−0.17 .0.270−0.046−0.11 −0.032−0.066 −0.028−0.055
. H0 [Km/s/Mpc] .70.6+4.3+5.2 .70.3+4.3+5.4 .68.8+1.3+3.0 +1.0+2.1 +0.82+1.6
−2.4−6.0 −2.8−5.6 −1.6−2.7 .68.2−1.0−2.1 .68.33−0.81−1.5

. S8 .1.01+0.08+0.41 .0.98+0.06+0.40 +0.032+0.13 .0.869+0.028+0.070 .0.849+0.022+0.056

−0.23−0.27 −0.21−0.24 .0.880−0.071−0.11 −0.038−0.066 −0.031−0.052
.rdrag [Mpc] +0.33+0.64
.146.80−0.33−0.65 .146.91+0.29+0.56 +0.31+0.61
.146.90−0.31−0.63 .146.78+0.33+0.64 .146.98+0.27+0.53
−0.29−0.57 −0.33−0.65 −0.27−0.53
29 On the Interacting Dark Energy Scenarios—The Case for Hubble Constant Tension
539
540 S. Pan and W. Yang

Fig. 29.1 One dimensional marginalized posterior distributions of some of the free and derived
parameters and the two dimensional joint contours for the interacting scenario between a non-cold
DM and vacuum have been displayed considering various observational datasets. This figure has
been taken from Ref. [104]

29.3.2 IDE with Dynamical Equation of State of DE

The interacting scenarios in which the DE equation of state is dynamical, are the most
generalized scenarios since such scenarios can recover the interacting scenarios with
constant .wde models as a special case. From both the theoretical and observational
grounds, such scenarios are quite appealing. One can investigate such scenarios
following two approaches—either one can parametrize the DE equation of state in
terms of some dynamical variable, e.g. the scale factor of the universe, see for instance
Refs. [36, 37, 49, 51, 54, 77, 79], or, one can assume that DE is described by a scalar
field .φ with potential .V (φ) where the equation of state of DE is dynamical being
dependent on the scalar field and its potential, see Refs. [116–120]. Surprisingly,
29 On the Interacting Dark Energy Scenarios—The Case for Hubble Constant Tension 541

according to the existing literature, only a few works have been done considering
the parametrized forms for .wde [36, 37, 49, 51, 54, 77, 79] without any particular
reasons, while these scenarios and quite simple and they can lead to higher values
of . H0 . We suspect that one of the possible reasons could be the behaviour of the
interacting scenarios at the level of perturbations which is certainly affected by the
parametrized dynamical DE equation of state and an arbitrary choice of the dynamical
.wde may invite instabilities in the large scale. However, with a suitable choice of a

dynamical .wde , one can cure the instability issues from the picture. In this section
we shall discuss the impact of these scenarios in the context of the . H0 tension.
In Ref. [79], an interacting scenario between CDM and DE has been investigated
with the following interaction function

. Q = 3 H ξ(1 + wde )ρde , (29.16)

where .ξ is the coupling parameter of the interaction model and .wde has a dynamical
nature which can take one of the expressions below

. Model I : wde (a) = w0 a[1 − log(a)], (29.17)

. Model II : wde (a) = w0 a exp(1 − a), (29.18)
. Model III : wde (a) = w0 a[1 + sin(1 − a)], (29.19)
. Model IV : wde (a) = w0 a[1 + arcsin(1 − a)]. (29.20)

where .w0 in every model refers to the present value of .wde . In Table 29.2, we present
the estimated values of . H0 in each interacting scenario for Planck 2015 and Planck
2015+BAO datasets [79]. From Table 29.2, it is clearly seen that for CMB data
alone, the estimated values of . H0 , for all dynamical equation of state parameters
(29.17)—(29.20), are significantly large compared to R21 by S. H0 ES collaboration
[6], however, due to the large error bars, the tension with R21 is reduced down to
less than .1 σ . However, for the Planck 2015+BAO dataset, the situation becomes
quite interesting. As shown in Table 29.2, the . H0 values are very close to its local
estimation by S. H0 ES [6] and the uncertainties on . H0 are significantly decreased
compared yo Planck 2015. Thus, with respect to R21, the tension is reduced down to
.1.1 σ (for Model I), .0.7 σ (for Model II), .0.3 σ (for Model III), .0.1 σ (for Model IV).
We thus conclude that the last two models are quite sound since they can completely
resolve the . H0 tension for both Planck 2018 and Planck 2015+BAO datasets.
In the context of interacting scalar field models, the determination of the cosmo-
logical parameters is highly dependent on the choice of the scalar field potential and
also on the interaction function as well. As argued in Ref. [118], such scalar field
interacting models can also relieve the . H0 tension if a suitable choice of the potential
is made.
However, considering both the scalar field interacting scenarios and the
parametrized .wde interacting scenarios, we believe that both the scenarios are quite
appealing in the light of the . H0 tension, and hence, they need further attention in the
upcoming years.
542 S. Pan and W. Yang

Table 29.2 The table presents the constraints on . H0 for CMB and CMB+BAO datasets where
CMB data have been taken from Planck 2015 considering various interacting scenarios in which
the interaction function has the form (29.16) and the equation of state for DE is dynamical
Equation of state Hubble constant . H0 (km/s/Mpc)
Planck 2015 Planck 2015+BAO
+13 +1.5
.wDE (a) = w0 a[1 − log(a)] .81−14 .71.0−1.5
+14 +1.5
.wDE (a) = w0 a exp(1 − a) .84−7 .71.7−1.7
+12 +1.6
.wDE (a) = .84−5 .73.5−1.7
w0 a[1 + sin(1 − a)]
+14 +1.5
.wDE (a) = .82−17 .72.8−1.8
w0 a[1 + arcsin(1 − a)]

29.3.3 Interacting Scenarios with Time-Dependent Coupling

Parameter

In most of the interaction functions explored in the literature, the coupling parameter
is usually assumed to be constant. However, there is no such guiding principle avail-
able yet which dictates that the coupling parameter of the interaction functions should
be constant instead of dynamical. Even though the interaction scenarios with constant
coupling parameter are relatively easier to handle, but the interacting functions with
time dependent coupling represent more generalized scenarios [113, 114, 121–125].
In the following we shall discuss that one can reconcile the Hubble constant tension
within these generalized frameworks.
Let us focus on an interacting scenario between CDM and DE with constant
equation-of-state .wde and the interaction function has the following form

. Q = 3 H ξ(X ) f (ρdm , ρde ), (29.21)

where. f is any analytic function of.ρdm , ρde ;.ξ(X ) is a dimensionless time-dependent

coupling and .X could be any cosmological variable. We consider .X = a/a0 = a (.a0
is the present value of the scale factor and we set .a0 = 1) and consider the first two
terms of the Taylor series expansion of .ξ(a) around .a0 = 1 as follows

.ξ(a) = ξ0 + ξa (1 − a) , (29.22)

where .ξ0 , .ξa are constants and .ξ0 refers to the present day value of .ξ(a). Now,
depending on the choice of . f (ρdm , ρde ), and the DE equation of state parameter, .wde ,
one could explore the dynamics of the underlying interacting scenario and estimate
the constraints on the key cosmological parameters. We focus on the following two
well known interaction functions of the form
29 On the Interacting Dark Energy Scenarios—The Case for Hubble Constant Tension 543

. Q = 3ξ(X ) H ρde , (29.23)

ρdm ρde
. Q = 3ξ(X ) H , (29.24)
ρdm + ρde

29.3.3.1 The Case for .wde = −1

The simplest interacting scenario is the one when CDM interacts with the vacuum
energy corresponding to the DE equation of state .wde = −1, known as the Inter-
acting Vacuum Scenario (IVS). The interacting scenarios driven by the interaction
functions in Eq. (29.23) [labeled as IVS1], (29.24) [labeled as IVS2] together with
the dynamical coupling .ξ(a) as in (29.22), were found to increase the Hubble con-
stant value compared to the .ΛCDM based Planck’s estimation, see for instance, Ref.
[122]. In particular, for the interaction function . Q = 3H [ξ0 + ξa (1 − a)]ρde of Eq.
(29.23), one finds [122] . H0 = 69.2 ± 2.8 km/sec/Mpc at 68% CL (Planck 2015),
and . H0 = 68.5+1.1
−1.5 km/sec/Mpc at 68% CL (Planck 2015+BAO) [122], and hence,
with respect to R21, the tension is alleviated at .1.3σ and .2.7σ , respectively.
On the other hand, for the interaction function . Q = 3H [ξ0 + ξa (1 − a)] ρρdmdm+ρ
ρde
de
of Eq. (29.24), one finds . H0 = 68.3 ± 3.5 km/sec/Mpc at 68% CL (Planck 2015)
+1.3
[122] and . H0 = 68.0−1.5 km/sec/Mpc at 68% CL (Planck 2015+BAO) [122], and
hence with respect to R21, the tension is alleviated at .1.3σ and .2.9σ , respectively.
Now, referring to Tables 2 and 3 of [122], one can find that even though the
strong evidence for non-null values of .ξ0 and .ξa is not found from both Planck 2015
and Planck 2015+BAO, however, we cannot readily exclude the possibility of their
non-null values. This can be evident from Figs. 29.2 and 29.3 where we show the
evolution of .ξ(z) [.1 + z = a0 /a = 1/a] considering the 68% CL values of .ξ0 and .ξa
estimated from Planck 2015 and Planck 2015+BAO. One can notice from Figs. 29.2
and 29.3 that .ξ(z) has a transition from .ξ(z) > 0 to .ξ(z) < 0 which is prominent in
the right graph of each Figs. 29.2 and 29.3. This is an interesting observation which
tells that for .ξ(z) > 0 [equivalently, . Q(t) > 0], the energy flow occurs from the cold
DM to vacuum sector while for.ξ(z) < 0 [equivalently,. Q(t) < 0] energy flow occurs
from the vacuum energy sector to cold DM. Finally, as the choice of the dynamical
coupling .ξ(a) is not unique, therefore, investigations with various choices for .ξ(a)
will be certainly interesting.

29.3.3.2 The Case for .wde = Constant .(/= −1)

One can extend the interacting scenarios of the previous Sect. 29.3.3.1 by a con-
stant DE equation of state .wde (/= −1) which is allowed to vary in a certain
region. Such extensions can lead to many interesting outcomes and also alleviate
the Hubble constant tension [123]. As reported in Ref. [123], if the DE equation
of state .wde is allowed to vary in the phantom regime (i.e. .wde < −1), then for
the interaction function . Q = 3H [ξ0 + ξa (1 − a)]ρde , the . H0 is constrained to be
+10.1
. H0 = 81.8−11.7 km/sec/Mpc at 68% CL for Planck 2018 together with a mild evi-
544 S. Pan and W. Yang

Fig. 29.2 Evolution of the coupling parameter .ξ(z) [.1 + z = a0 /a = 1/a] of Eq. (29.22) for IVS1
for CMB (left graph) and CMB+BAO (right graph). The red (green) solid curve corresponds to
the best fit curve for CMB (CMB+BAO) with the corresponding 68% CL region. Figure has been
drawn taking the values of (.ξ0 , ξa ) from Table 3 of Ref. [122]

Fig. 29.3 Evolution of the coupling parameter .ξ(z) [.1 + z = a0 /a = 1/a] of (29.22) for IVS2 for
CMB (left graph) and CMB+BAO (right graph). The red (green) solid curve corresponds to the best
fit curve for CMB (CMB+BAO) with the corresponding 68% CL region. Figure has been drawn
taking the values of (.ξ0 , ξa ) from Table 3 of Ref. [122]

dence for .ξa = −0.077+0.064

−0.032 at 68% CL and a strong preference for a phantom DE
+0.494
(.wde − 1.800−0.387 at 68% CL), thus, the tension with R21 is alleviated within.1σ . For
+1.4
Planck 2018+BAO dataset, the same scenario leads to . H0 = 69.4−1.9 km/sec/Mpc at
68% CL together with a strong preference for a phantom DE (.wde = −1.207+0.196 −0.208 at
95% CL) but here the evidence for .ξa goes away. In this case, the tension with R21 is
reduced down to .1.9σ . On the other hand, for the quintessence DE (.wde > −1), the
interaction function . Q = 3H [ξ0 + ξa (1 − a)]ρde can lead to a higher value of the
+4.1
Hubble constant. For Planck 2018 alone, one finds that . H0 = 70.2−3.1 km/sec/Mpc
at 68% CL, and thus the tension with R21 is resolved within .0.7σ . On the other hand,
+1.3
for Planck 2018+BAO, this interaction function leads to . H0 = 68.4−1.4 km/sec/Mpc
at 68% CL which thus reduces the tension with R21 down to .2.7 σ .
For the interaction function. Q = 3H [ξ0 + ξa (1 − a)] ρρdmdm+ρ
ρde
de
of Eq. (29.24), if.wde
is allowed to vary in the phantom regime, then. H0 tension can be alleviated [123]. For
29 On the Interacting Dark Energy Scenarios—The Case for Hubble Constant Tension 545

Planck 2018 alone, this interacting scenario returns . H0 > 72.5 km/sec/Mpc at 95%
CL with a strong evidence of phantom DE (.wde = −1.66652146+0.485 −0.370 at 68% CL)
+1.0
and for Planck 2018+BAO, . H0 = 69.4−1.5 with a mild indication of the phantom DE
+0.067
(.wde = −1.080−0.031 at 68% CL). Therefore, the tension in . H0 with R21 is reduced
down to .2 σ and .2.3 σ , respectively, for Planck 2018 alone and Planck 2018+BAO
dataset.

29.4 Interacting DE in a Non-flat Universe

In almost every works on the interacting scenarios, the curvature of our universe
is usually assumed to be zero. This is usually motivated from some earlier investi-
gations reporting a very small value of the curvature density parameter [126–129],
and, on the other hand, in presence of the curvature of the universe, the analysis
of the underlying cosmological model becomes complicated a bit compared to the
analysis of the models with no curvature. However, with the growing sensitivity in
the astronomical data, as the constraints on the cosmological parameters are getting
improved, therefore, there is no physical motivation to exclude the curvature param-
eter from the analysis. In fact, one can include the curvature parameter in the analysis
of the cosmological models and finally allow the observational data to decide its fate.
Some recent investigations argued that the observational evidences are in favour of
a closed universe scenario [130, 131]. And it was also contended in [130, 131] that
the preference of a closed universe may worsen the . H0 tension. Thus, the question
arises, can interacting scenarios in the context of a non-flat universe mitigate or solve
the . H0 tension? This issue was investigated in Refs. [132, 133].
In Ref. [132], an interaction between a CDM and DE with .wde = −0.999 was
investigated for the interaction function (29.11). As discussed in [132], different
datasets return different values of . H0 . For Planck 2018 alone, we find a lower value
of . H0 with an evidence of a closed universe (.Ωk = −0.036+0.017 −0.013 at 68% CL for
Planck 2018) but no evidence of an interaction, however, from Planck 2018+BAO,
while .Ωk is consistent to the flat universe, but we obtain an evidence of an interaction
+0.31
(.ξ = −0.32−0.09 at 68%) within 68% CL together with . H0 = 69.7+1.2 −1.6 km/s/Mpc at
68% CL which thus alleviates the tension with R21 at .1.9σ .
Further in Ref. [133], an interaction between the CDM and DE with .wde /= −1
has been investigated with the interaction function . Q = 3H ξρde .5 For the large scale
stability of the interacting scenario, the DE equation of state .wde was divided into
two separate regions, namely, .wde < −1 (phantom DE) and .wde > −1 (quintessence
+1.2
DE). We find that for .wde < −1, . H0 = 69.0−1.7 km/s/Mpc at 68% CL for Planck
−0.045
2018+BAO with an evidence of an interaction at 68% CL (.ξ = −0.052−0.025 at 68%
CL), however, the curvature parameter is consistent to a flat universe [133]. Therefore,

5Note that the interaction model (29.11) is same with the interaction function . Q = 3H ξρde . The
presence of the factor .3 has some mathematical importance, but no physics is associated with this
choice.
546 S. Pan and W. Yang

in this case, the tension with R21 is alleviated at .2.3σ for Planck 2018+BAO. For
wde > −1, one leads to. H0 = 68.5 ± 1.4 km/s/Mpc at 68% CL for Planck 2018+BAO
.
[133], and hence, the tension with R21 is reduced down to .2.6 σ .
The interaction functions presented in this section are not the only choices. One
can investigate other kind of interaction functions considering the fact that .wde could
be dynamical. We anticipate that the interacting scenarios in the context of a non-
flat universe are promising and these models need further investigations with the
upcoming observational data from various astronomical surveys.

29.5 Summary and Conclusions

The theme of the present chapter is the theory of non-gravitational interaction

between two heavy dark components of the universe, namely, DM and DE. The
interaction between them is characterized by the transfer or energy between them.
Over the last couple of years, interacting DE-DM scenarios have gained remarkable
attention in the scientific community for many appealing consequences, such as,
alleviation of the cosmic coincidence problem, crossing of the phantom divide line,
and addressing the cosmological tensions. The heart of this theory is the interaction
function . Q(t) that governs the transfer of energy between the dark sectors, and, as a
consequence, this interaction function modifies the evolution equations for DE and
DM at the level of background and perturbations. Thus, it is natural to expect that the
cosmological parameters should be affected due to the presence of this interaction
function. However, there is no such fundamental principle available yet for deriving
the interaction function. Even though some attempts have been made to formulate
the interaction function, but no conclusive statement has been made so far in this
direction. Nevertheless, due to the existing features available in this particular theory,
interacting DE has been proved as a phenomenologically rich cosmological theory.
In this chapter we have discussed a variety of interacting scenarios between DE and
DM characterized by the triplet .(Q(t), wdm , wde ) where . Q(t) has a wide variations;
.wdm , representing the equation of state of the DM sector could be either zero or
some non-negative constant, however, the DE equation of state, .wde , could be either
constant or time dependent. Initially we have focused on the interacting scenarios
assuming that the curvature of the universe is zero. In this set up, we have found that
depending on the nature of .wde and the interaction function . Q(t), the .5σ tension on
. H0 between Planck [5] and S. H0 ES collaboration [6] can be significantly reduced. In
particular, in some models, . H0 tension is reduced down to .< 3σ ; in some scenarios
this tension is reduced down to .< 2σ ; in some scenarios it is reduced down to .< 1σ ;
and in some scenarios, one can completely solve this tension. On the other hand,
assuming that our universe may not be spatially flat, we have discussed the impact
of the interacting scenarios on the . H0 tension.
Based on the discussions presented in this chapter, it is pretty clear that the inter-
acting scenarios between DE and DM are really promising for reconciling the Hubble
constant tension. In addition, we would like to stress that there are many gaps in this
29 On the Interacting Dark Energy Scenarios—The Case for Hubble Constant Tension 547

domain which need to be explored crucially with the upcoming cosmological probes
from several astronomical surveys.

Acknowledgements S. Pan acknowledges the financial support from the Department of Science
and Technology (DST), Govt. of India under the Scheme “Fund for Improvement of S&T Infrastruc-
ture (FIST)” [File No. SR/FST/MS-I/2019/41]. W. Yang acknowledges the support by the National
Natural Science Foundation of China under Grants No. 12175096 and No. 11705079, and Liaoning
Revitalization Talents Program under Grant no. XLYC1907098.

References

1. A.G. Riess et al., Supernova search team. Astron. J. 116, 1009–1038 (1998)
[arXiv:astro-ph/9805201 [astro-ph]]
2. S. Perlmutter et al., Supernova cosmology project. Astrophys. J. 517, 565–586 (1999)
[arXiv:astro-ph/9812133 [astro-ph]]
3. S. Weinberg, Rev. Mod. Phys. 61, 1–23 (1989)
4. I. Zlatev, L.M. Wang, P.J. Steinhardt, Phys. Rev. Lett. 82, 896–899 (1999)
[arXiv:astro-ph/9807002 [astro-ph]]
5. N. Aghanim et al., Planck. Astron. Astrophys. 641, A6 (2020) [erratum: Astron. Astrophys.
652, C4 (2021)] [arXiv:1807.06209 [astro-ph.CO]]
6. A.G. Riess, W. Yuan, L.M. Macri, D. Scolnic, D. Brout, S. Casertano, D.O. Jones,
Y. Murakami, L. Breuval, T.G. Brink et al., Astrophys. J. Lett. 934(1), L7 (2022)
[arXiv:2112.04510 [astro-ph.CO]]
7. A.G. Riess, L. Breuval, W. Yuan, S. Casertano, L.M. Macri, J.B. Bowers, D. Scolnic, T. Cantat-
Gaudin, R.I. Anderson, M.C. Reyes, Astrophys. J. 938(1), 36 (2022) [arXiv:2208.01045
[astro-ph.CO]]
8. M. Asgari, T. Tröster, C. Heymans, H. Hildebrandt, J.L. van den Busch, A.H. Wright,
A. Choi, T. Erben, B. Joachimi, S. Joudaki et al., Astron. Astrophys. 634, A127 (2020)
[arXiv:1910.05336 [astro-ph.CO]]
9. M. Asgari et al., KiDS. Astron. Astrophys. 645, A104 (2021) [arXiv:2007.15633 [astro-
ph.CO]]
10. S. Joudaki, H. Hildebrandt, D. Traykova, N.E. Chisari, C. Heymans, A. Kannawadi, K.
Kuijken, A.H. Wright, M. Asgari, T. Erben et al., Astron. Astrophys. 638, L1 (2020)
[arXiv:1906.09262 [astro-ph.CO]]
11. T.M.C. Abbott et al., DES. Phys. Rev. D 105(2), 023520 (2022) [arXiv:2105.13549 [astro-
ph.CO]]
12. A. Amon et al., DES. Phys. Rev. D 105(2), 023514 (2022) [arXiv:2105.13543 [astro-ph.CO]]
13. L.F. Secco et al., DES. Phys. Rev. D 105(2), 023515 (2022) [arXiv:2105.13544 [astro-ph.CO]]
14. A. Loureiro et al., KiDS and Euclid. Astron. Astrophys. 665, A56 (2022) [arXiv:2110.06947
[astro-ph.CO]]
15. C. Heymans, T. Tröster, M. Asgari, C. Blake, H. Hildebrandt, B. Joachimi, K. Kuijken,
C.A. Lin, A.G. Sánchez, J.L. van den Busch et al., Astron. Astrophys. 646, A140 (2021)
[arXiv:2007.15632 [astro-ph.CO]]
16. T.M.C. Abbott et al., DES. Phys. Rev. D 102(2), 023509 (2020) [arXiv:2002.11124 [astro-
ph.CO]]
17. O.H.E. Philcox, M.M. Ivanov, Phys. Rev. D 105(4), 043517 (2022) [arXiv:2112.04515 [astro-
ph.CO]]
18. E. Di Valentino, O. Mena, S. Pan, L. Visinelli, W. Yang, A. Melchiorri, D.F. Mota, A.G. Riess,
J. Silk, Class. Quant. Grav. 38(15), 153001 (2021) [arXiv:2103.01183 [astro-ph.CO]]
548 S. Pan and W. Yang

19. L. Perivolaropoulos, F. Skara, New Astron. Rev. 95, 101659 (2022) [arXiv:2105.05208 [astro-
ph.CO]]
20. E. Abdalla, G. Franco Abellán, A. Aboubrahim, A. Agnello, O. Akarsu, Y. Akrami,
G. Alestas, D. Aloni, L. Amendola, L.A. Anchordoqui et al., JHEAp 34, 49–211 (2022)
[arXiv:2203.06142 [astro-ph.CO]]
21. M. Kamionkowski, A.G. Riess [arXiv:2211.04492 [astro-ph.CO]]
22. C. Wetterich, Astron. Astrophys. 301, 321–328 (1995) [arXiv:hep-th/9408025 [hep-th]]
23. L. Amendola, Coupled quintessence. Phys. Rev. D 62, 043511 (2000) [astro-ph/9908023]
24. L.P. Chimento, A.S. Jakubi, D. Pavon, W. Zimdahl, Phys. Rev. D 67, 083513 (2003)
[arXiv:astro-ph/0303145 [astro-ph]]
25. R.G. Cai, A. Wang, JCAP 0503, 002 (2005) [hep-th/0411025]
26. D. Pavón, W. Zimdahl, Phys. Lett. B 628, 206 (2005) [arXiv:gr-qc/0505020]
27. S. del Campo, R. Herrera, G. Olivares, D. Pavon, Phys. Rev. D 74, 023501 (2006)
[arXiv:astro-ph/0606520 [astro-ph]]
28. M.S. Berger, H. Shojaei, Phys. Rev. D 73, 083528 (2006) [arXiv:gr-qc/0601086 [gr-qc]]
29. S. del Campo, R. Herrera, D. Pavón, Phys. Rev. D 78, 021302 (2008) [arXiv:0806.2116
[astro-ph]]
30. S. del Campo, R. Herrera, D. Pavón, JCAP 0901, 020 (2009) [arXiv:0812.2210 [gr-qc]]
31. J.D. Barrow, T. Clifton, Phys. Rev. D 73, 103520 (2006) [gr-qc/0604063]
32. L. Amendola, G. Camargo Campos, R. Rosenfeld, Phys. Rev. D 75, 083506 (2007) [astro-
ph/0610806]
33. T. Koivisto, D.F. Mota, Phys. Rev. D 75, 023518 (2007) [hep-th/0609155]
34. J.H. He, B. Wang, JCAP 0806, 010 (2008) [arXiv:0801.4233 [astro-ph]]
35. M.B. Gavela, D. Hernandez, L. Lopez Honorez, O. Mena, S. Rigolin, JCAP 0907, 034 (2009)
[arXiv:0901.1611 [astro-ph.CO]]
36. E. Majerotto, J. Väliviita, R. Maartens, Mon. Not. Roy. Astron. Soc. 402, 2344 (2010)
[arXiv:0907.4981 [astro-ph.CO]]
37. J. Valiviita, R. Maartens, E. Majerotto, Mon. Not. Roy. Astron. Soc. 402, 2355–2368 (2010)
[arXiv:0907.4987 [astro-ph.CO]]
38. T. Clemson, K. Koyama, G.B. Zhao, R. Maartens, J. Väliviita, Phys. Rev. D 85, 043007 (2012)
[arXiv:1109.6234 [astro-ph.CO]]
39. V. Salvatelli, N. Said, M. Bruni, A. Melchiorri, D. Wands, Phys. Rev. Lett. 113(18), 181301
(2014) [arXiv:1406.7297 [astro-ph.CO]]
40. W. Yang, L. Xu, JCAP 1408, 034 (2014) [arXiv:1401.5177 [astro-ph.CO]]
41. W. Yang, L. Xu, Phys. Rev. D 89(8), 083517 (2014) [arXiv:1401.1286 [astro-ph.CO]]
42. Y. Wang, D. Wands, G.B. Zhao, L. Xu, Phys. Rev. D 90(2), 023502 (2014). https://doi.org/
10.1103/PhysRevD.90.023502 [arXiv:1404.5706 [astro-ph.CO]]
43. V. Faraoni, J.B. Dent, E.N. Saridakis, Phys. Rev. D 90(6), 063510 (2014) [arXiv:1405.7288
[gr-qc]]
44. S. Pan, S. Bhattacharya, S. Chakraborty, Mon. Not. Roy. Astron. Soc. 452(3), 3038 (2015)
[arXiv:1210.0396 [gr-qc]]
45. W. Yang, L. Xu, Phys. Rev. D 90(8), 083532 (2014) [arXiv:1409.5533 [astro-ph.CO]]
46. Y.H. Li, J.F. Zhang, X. Zhang, Phys. Rev. D 93(2), 023002 (2016) [arXiv:1506.06349 [astro-
ph.CO]]
47. R.C. Nunes, S. Pan, E.N. Saridakis, Phys. Rev. D 94(2), 023508 (2016) [arXiv:1605.01712
[astro-ph.CO]]
48. W. Yang, H. Li, Y. Wu, J. Lu, JCAP 1610(10), 007 (2016) [arXiv:1608.07039 [astro-ph.CO]]
49. S. Pan, G.S. Sharov, Mon. Not. Roy. Astron. Soc. 472(4), 4736 (2017) [arXiv:1609.02287
[gr-qc]]
50. A. Mukherjee, N. Banerjee, Class. Quant. Grav. 34(3), 035016 (2017) [arXiv:1610.04419
[astro-ph.CO]]
51. G.S. Sharov, S. Bhattacharya, S. Pan, R.C. Nunes, S. Chakraborty, Mon. Not. Roy. Astron.
Soc. 466(3), 3497 (2017) [arXiv:1701.00780 [gr-qc]]
29 On the Interacting Dark Energy Scenarios—The Case for Hubble Constant Tension 549

52. R.Y. Guo, Y.H. Li, J.F. Zhang, X. Zhang, JCAP 1705(05), 040 (2017) [arXiv:1702.04189
[astro-ph.CO]]
53. R.G. Cai, N. Tamanini, T. Yang, JCAP 1705(05), 031 (2017) [arXiv:1703.07323 [astro-
ph.CO]]
54. W. Yang, N. Banerjee, S. Pan, Phys. Rev. D 95(12), 123527 (2017) [arXiv:1705.09278 [astro-
ph.CO]]
55. W. Yang, S. Pan, D.F. Mota, Phys. Rev. D 96(12), 123508 (2017) [arXiv:1709.00006 [astro-
ph.CO]]
56. W. Yang, S. Pan, J.D. Barrow, Phys. Rev. D 97(4), 043529 (2018) [arXiv:1706.04953 [astro-
ph.CO]]
57. S. Pan, A. Mukherjee, N. Banerjee, Mon. Not. Roy. Astron. Soc. 477(1), 1189 (2018)
[arXiv:1710.03725 [astro-ph.CO]]
58. W. Yang, S. Pan, A. Paliathanasis, Mon. Not. Roy. Astron. Soc. 482(1), 1007 (2019)
[arXiv:1804.08558 [gr-qc]]
59. W. Yang, S. Pan, L. Xu, D.F. Mota, Mon. Not. Roy. Astron. Soc. 482(2), 1858 (2019)
[arXiv:1804.08455 [astro-ph.CO]]
60. A. Paliathanasis, S. Pan, W. Yang, Int. J. Mod. Phys. D 28(12), 1950161 (2019)
[arXiv:1903.02370 [gr-qc]]
61. S. Pan, W. Yang, C. Singha, E.N. Saridakis, Phys. Rev. D 100(8), 083539 (2019)
[arXiv:1903.10969 [astro-ph.CO]]
62. W. Yang, S. Pan, E. Di Valentino, B. Wang, A. Wang, JCAP 05, 050 (2020) [arXiv:1904.11980
[astro-ph.CO]]
63. W. Yang, S. Vagnozzi, E. Di Valentino, R.C. Nunes, S. Pan, D.F. Mota, JCAP 1907, 037
(2019) [arXiv:1905.08286 [astro-ph.CO]]
64. W. Yang, S. Pan, R.C. Nunes, D.F. Mota, JCAP 04, 008 (2020). https://doi.org/10.1088/1475-
7516/2020/04/008 [arXiv:1910.08821 [astro-ph.CO]]
65. W. Yang, E. Di Valentino, S. Pan, S. Basilakos, A. Paliathanasis, Phys. Rev. D 102(6), 063503
(2020). https://doi.org/10.1103/PhysRevD.102.063503 [arXiv:2001.04307 [astro-ph.CO]]
66. S. Pan, W. Yang, A. Paliathanasis, Mon. Not. Roy. Astron. Soc. 493(3), 3114–3131 (2020).
https://doi.org/10.1093/mnras/staa213[arXiv:2002.03408 [astro-ph.CO]]
67. Y.L. Bolotin, A. Kostenko, O.A. Lemets, D.A. Yerokhin, Int. J. Mod. Phys. D 24(03), 1530007
(2015) [arXiv:1310.0085 [astro-ph.CO]]
68. B. Wang, E. Abdalla, F. Atrio-Barandela, D. Pavón, Rept. Prog. Phys. 79(9), 096901 (2016)
[arXiv:1603.08299 [astro-ph.CO]]
69. B. Wang, Y.G. Gong, E. Abdalla, Phys. Lett. B 624, 141 (2005) [hep-th/0506069]
70. S. Das, P.S. Corasaniti, J. Khoury, Phys. Rev. D 73, 083509 (2006) [astro-ph/0510628]
71. H.M. Sadjadi, M. Honardoost, Phys. Lett. B 647, 231 (2007) [gr-qc/0609076]
72. S. Pan, S. Chakraborty, Int. J. Mod. Phys. D 23(11), 1450092 (2014) [arXiv:1410.8281 [gr-qc]]
73. S. Kumar, R.C. Nunes, Phys. Rev. D 94(12), 123511 (2016) [arXiv:1608.02454 [astro-ph.CO]]
74. S. Kumar, R.C. Nunes, Phys. Rev. D 96(10), 103511 (2017) [arXiv:1702.02143 [astro-ph.CO]]
75. E. Di Valentino, A. Melchiorri, O. Mena, Phys. Rev. D 96(4), 043503 (2017)
[arXiv:1704.08342 [astro-ph.CO]]
76. W. Yang, S. Pan, E. Di Valentino, R.C. Nunes, S. Vagnozzi, D.F. Mota, JCAP 1809(09), 019
(2018) [arXiv:1805.08252 [astro-ph.CO]]
77. W. Yang, A. Mukherjee, E. Di Valentino, S. Pan, Phys. Rev. D 98(12), 123527 (2018)
[arXiv:1809.06883 [astro-ph.CO]]
78. S. Kumar, R.C. Nunes, S.K. Yadav, Eur. Phys. J. C 79(7), 576 (2019) [arXiv:1903.04865
[astro-ph.CO]]
79. S. Pan, W. Yang, E. Di Valentino, E.N. Saridakis, S. Chakraborty, Phys. Rev. D 100(10),
103520 (2019) arXiv:1907.07540 [astro-ph.CO]
80. E. Di Valentino, A. Melchiorri, O. Mena, S. Vagnozzi, Phys. Dark Univ. 30, 100666 (2020)
[arXiv:1908.04281 [astro-ph.CO]]
81. E. Di Valentino, A. Melchiorri, O. Mena, S. Vagnozzi, Phys. Rev. D 101(6), 063502 (2020)
[arXiv:1910.09853 [astro-ph.CO]]
550 S. Pan and W. Yang

82. Y. Zhai, W. Giarè, C. van de Bruck, E. Di Valentino, O. Mena, R.C. Nunes, JCAP 07, 032
(2023) [arXiv:2303.08201 [astro-ph.CO]]
83. A. Pourtsidou, T. Tram, Phys. Rev. D 94(4), 043518 (2016) [arXiv:1604.04222 [astro-ph.CO]]
84. R. An, C. Feng, B. Wang, JCAP 1802(02), 038 (2018) [arXiv:1711.06799 [astro-ph.CO]]
85. J. Gleyzes, D. Langlois, M. Mancarella, F. Vernizzi, JCAP 1508, 054 (2015)
[arXiv:1504.05481 [astro-ph.CO]]
86. C.G. Böehmer, N. Tamanini, M. Wright, Phys. Rev. D 91(12), 123002 (2015)
[arXiv:1501.06540 [gr-qc]]
87. C.G. Böehmer, N. Tamanini, M. Wright, Phys. Rev. D 91(12), 123003 (2015)
[arXiv:1502.04030 [gr-qc]]
88. C. van de Bruck, J. Morrice, JCAP 1504, 036 (2015) [arXiv:1501.03073 [gr-qc]]
89. L. Xiao, R. An, L. Zhang, B. Yue, Y. Xu, B. Wang, Phys. Rev. D 99(2), 023528 (2019)
[arXiv:1807.05541 [astro-ph.CO]]
90. G. D’Amico, T. Hamill, Nemanja Kaloper, Phys. Rev. D 94, 103526 (2016) [arXiv:1605.00996
[hep-th]]
91. M.C.D. Marsh, Phys. Rev. Lett. 118, 011302 (2017) [arXiv:1606.01538 [astro-ph.CO]]
92. R. Kase, S. Tsujikawa, Phys. Rev. D 101(6), 063511 (2020) [arXiv:1910.02699 [gr-qc]]
93. S. Alexander, M. Cortês, A.R. Liddle, J. Magueijo, R. Sims, L. Smolin, Phys. Rev. D 100(8),
083507 (2019) [arXiv:1905.10382 [gr-qc]]
94. S. Pan, G.S. Sharov, W. Yang, Phys. Rev. D 101(10), 103533 (2020) [arXiv:2001.03120
[astro-ph.CO]]
95. J. Valiviita, E. Majerotto, R. Maartens, JCAP 07, 020 (2008) [arXiv:0804.0232 [astro-ph]]
96. S. Pan, J. de Haro, W. Yang, J. Amorós, Phys. Rev. D 101(12), 123506 (2020)
[arXiv:2001.09885 [gr-qc]]
97. W. Yang, S. Pan, L. Aresté Saló, J. de Haro, Phys. Rev. D 103(8), 083520 (2021)
[arXiv:2104.04505 [astro-ph.CO]]
98. C.M. Muller, Phys. Rev. D 71, 047302 (2005) [arXiv:astro-ph/0410621 [astro-ph]]
99. S. Kumar, L. Xu, Phys. Lett. B 737, 244–247 (2014) [arXiv:1207.5582 [gr-qc]]
100. C. Armendariz-Picon, J.T. Neelakanta, JCAP 03, 049 (2014) [arXiv:1309.6971 [astro-ph.CO]]
101. M. Kopp, C. Skordis, D.B. Thomas, S. Ilić, Phys. Rev. Lett. 120(22), 221102 (2018)
[arXiv:1802.09541 [astro-ph.CO]]
102. S. Ilić, M. Kopp, C. Skordis, D.B. Thomas, Phys. Rev. D 104(4), 043520 (2021)
[arXiv:2004.09572 [astro-ph.CO]]
103. K. Naidoo, M. Jaber, W.A. Hellwing, M. Bilicki [arXiv:2209.08102 [astro-ph.CO]]
104. S. Pan, W. Yang, E. Di Valentino, D.F. Mota, J. Silk, JCAP 07, 064 (2023) [arXiv:2211.11047
[astro-ph.CO]]
105. E. Di Valentino, A. Melchiorri, J. Silk, Phys. Lett. B 761, 242–246 (2016) [arXiv:1606.00634
[astro-ph.CO]]
106. V.F. Mukhanov, H.A. Feldman, R.H. Brandenberger, Phys. Rept. 215, 203–333 (1992)
107. C.P. Ma, E. Bertschinger, Astrophys. J. 455, 7–25 (1995) [arXiv:astro-ph/9506072 [astro-ph]]
108. K.A. Malik, D. Wands, Phys. Rept. 475, 1–51 (2009) [arXiv:0809.4944 [astro-ph]]
109. S.K. Choi et al., ACT. JCAP 12, 045 (2020) [arXiv:2007.07289 [astro-ph.CO]]
110. S. Aiola et al., ACT. JCAP 12, 047 (2020) [arXiv:2007.07288 [astro-ph.CO]]
111. H. Wei, Nucl. Phys. B 845, 381–392 (2011). https://doi.org/10.1016/j.nuclphysb.2010.12.010
[arXiv:1008.4968 [gr-qc]]
112. H. Wei, Commun. Theor. Phys. 56, 972–980 (2011) [arXiv:1010.1074 [gr-qc]]
113. Y.H. Li, X. Zhang, Eur. Phys. J. C 71, 1700 (2011) [arXiv:1103.3185 [astro-ph.CO]]
114. J.J. Guo, J.F. Zhang, Y.H. Li, D.Z. He, X. Zhang, Sci. China Phys. Mech. Astron. 61(3),
030011 (2018) [arXiv:1710.03068 [astro-ph.CO]]
115. F. Arevalo, A. Cid, L.P. Chimento, P. Mella, Eur. Phys. J. C 79(4), 355 (2019)
[arXiv:1901.04300 [gr-qc]]
116. E.M. Teixeira, R. Daniel, N. Frusciante, C. van de Bruck [arXiv:2309.06544 [astro-ph.CO]]
117. A.Gómez-Valent, Z. Zheng, L. Amendola, C. Wetterich, V. Pettorino, Phys. Rev. D 106(10),
103522 (2022) [arXiv:2207.14487 [astro-ph.CO]]
29 On the Interacting Dark Energy Scenarios—The Case for Hubble Constant Tension 551

118. A. Gómez-Valent, V. Pettorino, L. Amendola, Phys. Rev. D 101(12), 123513 (2020)

[arXiv:2004.00610 [astro-ph.CO]]
119. C. Van De Bruck, J. Mifsud, Phys. Rev. D 97(2), 023506 (2018) [arXiv:1709.04882 [astro-
ph.CO]]
120. C. van de Bruck, J. Mifsud, J. Morrice, Phys. Rev. D 95(4), 043513 (2017) [arXiv:1609.09855
[astro-ph.CO]]
121. Y. Wang, G.B. Zhao, Astrophys. J. 869(1), 26 (2018) [arXiv:1805.11210 [astro-ph.CO]]
122. W. Yang, O. Mena, S. Pan, E. Di Valentino, Phys. Rev. D 100, 083509 (2019)
[arXiv:1906.11697 [astro-ph.CO]]
123. W. Yang, E. Di Valentino, O. Mena, S. Pan, Phys. Rev. D 102(2), 023535 (2020)
[arXiv:2003.12552 [astro-ph.CO]]
124. Y.H. Yao, X.H. Meng [arXiv:2207.05955 [astro-ph.CO]]
125. W. Yang, S. Pan, O. Mena, E. Di Valentino [arXiv:2209.14816 [astro-ph.CO]]
126. E. Gaztanaga, R. Miquel, E. Sanchez, Phys. Rev. Lett. 103, 091302 (2009) [arXiv:0808.1921
[astro-ph]]
127. M.J. Mortonson, Phys. Rev. D 80, 123504 (2009) [arXiv:0908.0346 [astro-ph.CO]]
128. S.H. Suyu, T. Treu, S. Hilbert, A. Sonnenfeld, M.W. Auger, R.D. Blandford, T. Collett,
F. Courbin, C.D. Fassnacht, L.V.E. Koopmans et al., Astrophys. J. Lett. 788, L35 (2014)
[arXiv:1306.4732 [astro-ph.CO]]
129. B. L’Huillier, A. Shafieloo, JCAP 01, 015 (2017) [arXiv:1606.06832 [astro-ph.CO]]
130. E. Di Valentino, A. Melchiorri, J. Silk, Nature Astron. 4(2), 196–203 (2019)
[arXiv:1911.02087 [astro-ph.CO]]
131. W. Handley, Phys. Rev. D 103(4), L041301 (2021) [arXiv:1908.09139 [astro-ph.CO]]
132. E. Di Valentino, A. Melchiorri, O. Mena, S. Pan, W. Yang, Mon. Not. Roy. Astron. Soc.
502(1), L23–L28 (2021) [arXiv:2011.00283 [astro-ph.CO]]
133. W. Yang, S. Pan, E. Di Valentino, O. Mena, A. Melchiorri, JCAP 10, 008 (2021)
[arXiv:2101.03129 [astro-ph.CO]]
Chapter 30
Modified Gravity

Emmanuel N. Saridakis

30.1 Introduction

As we discussed in the Introduction of this book, there are many directions that one
can follow in order to alleviate the .H0 tension. In particular, since .θs = DrsA , where .rs
is the sound horizon and .DA is the angular diameter distance, one could try to alter
either.rs or.DA or both. Solutions affecting.rs are usually called “early-time” solutions,
while solutions affecting .DA are usually called “late-time” solutions. Incorporating
gravitational modifications [1] is probably the only framework in the .H0 -literature
which is so broad that could be used in all directions, namely one could have early-
time modified-gravity solutions, late-time modified-gravity solutions, or modified-
gravity solutions that alter both regimes. Since early modified-gravity scenarios will
be investigated in the next chapter, in the present chapter we will focus on modified-
gravity solutions that alter the late-time evolution of the Universe.
The effect of modified gravity on the late-time universe evolution is two-fold. The
first is that it induces new terms in the Friedmann equations, which can collectively be
absorbed into an effective dark-energy sector. The second is that it typically leads to a
modified Newton’s constant (actually this is the first thing one expects that modified
gravity should do). Hence, in every cosmology governed by a modified theory of
gravity one typically obtains Friedmann equations of the form [1]

8π G eff ( eff
)
H2 =
. ρm + ρDE , (30.1)
3 ( )
eff
.Ḣ = −4π G eff ρm + ρDE + pm + pDE ,
eff
(30.2)

E. N. Saridakis (B)
National Observatory of Athens, Lofos Nymfon 11852, Athens, Greece
e-mail: msaridak@noa.gr
CAS Key Laboratory for Researches in Galaxies and Cosmology, Department of Astronomy,
University of Science and Technology of China, Hefei, Anhui 230026, People’s Republic of China
© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 553
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_30
554 E. N. Saridakis

eff eff
where .ρDE and .pDE are respectively the effective dark energy density and pressure,
and .G eff the effective Newton’s constant, all depending on the parameters of the
theory. Hence, qualitatively, we deduce that in order to alleviate the .H0 tension in
this framework, i.e. obtain a higher .H0 than standard lore predicts, we have two
ways [2, 3]. (i) One could either try to obtain smaller effective Newton’s constant,
since “weaker” gravity is reasonable to induce faster expansion. (ii) One could try to
obtain suitable modified-gravity-oriented extract terms in the effective dark-energy
sector, which could lead to faster expansion, e.g. obtain and effective dark-energy
equation-of-state parameter .wDE := pDE /ρDE lying in the phantom regime.
In this chapter we will present three broad classes of modified gravity that can
fulfill the above qualitative requirements in the correct quantitative way and allevi-
ate the .H0 tension. The first is Horndeski gravity [4] (or equivalently Generalized
Galileon theories [5, 6]), the second is .f (T ) gravity [7], and the third is .f (R) gravity
[8].

30.2 Horndeski/Generalized Galileon Theories

In this section we focus on Horndeski gravity [4], which coincides with generalized
Galileon theory [5, 6]. The most general Lagrangian with curvature terms and with
one extra scalar degree of freedom .φ, giving rise to second-order field equations, is
[4, 9, 10]
∑ 5
.L = Li , (30.3)
i=2

with

L2 .= K(φ, X ), (30.4)
L3 .= −G 3 (φ, X )□φ, (30.5)
L4 .= G 4 (φ, X ) R + G 4,X [(□φ)2 − (∇μ ∇ν φ) (∇ μ ∇ ν φ)], (30.6)
1
L5 .= G 5 (φ, X ) G μν (∇ μ ∇ ν φ) − G 5,X [(□φ)3 − 3(□φ) (∇μ ∇ν φ) (∇ μ ∇ ν φ)
6
+ 2(∇ μ ∇α φ) (∇ α ∇β φ) (∇ β ∇μ φ)] . (30.7)

As usual,.R is the Ricci scalar,.G μν is the Einstein tensor, the functions .K and.G i (.i =
3, 4, 5) depend on .φ and its kinetic energy .X = −∂ μ φ∂μ φ/2, and .G i,X := ∂G i /∂X ,
. G i,φ := ∂G i /∂φ. Focusing on Friedmann-Robertson-Walker (FRW) geometry with
metric .ds2 = −dt 2 + a2 (t) δij dxi dxj , and adding the matter Lagrangian .Lm corre-
sponding to a perfect fluid with energy density .ρm and pressure .pm , and performing
variation we obtain the two generalized Friedmann equations:
30 Modified Gravity 555

. 2XK,X − K + 6X φ̇H G 3,X − 2X G 3,φ − 6H 2 G 4 + 24H 2 X (G 4,X + X G 4,XX )

−12HX φ̇G 4,φX − 6H φ̇G 4,φ + 2H 3 X φ̇(5G 5,X + 2X G 5,XX )
−6H 2 X (3G 5,φ + 2X G 5,φX ) = −ρm , (30.8)

. K − 2X (G 3,φ + φ̈G 3,X ) + 2(3H 2 + 2Ḣ )G 4 − 8Ḣ X G 4,X − 12H 2 X G 4,X

−4H Ẋ G 4,X − 8HX Ẋ G 4,XX + 2(φ̈ + 2H φ̇)G 4,φ + 4X G 4,φφ
+4X (φ̈ − 2H φ̇)G 4,φX − 4H 2 X 2 φ̈G 5,XX − 2X (2H 3 φ̇ + 2H Ḣ φ̇ + 3H 2 φ̈)G 5,X
+4HX (Ẋ − HX )G 5,φX + 4HX φ̇G 5,φφ
+2[2(Ḣ X + H Ẋ ) + 3H 2 X ]G 5,φ = −pm , (30.9)

with dots marking derivatives with respect to .t. Moreover, variation with respect to
φ(t) gives
.
1 d 3
. (a J ) = Pφ , (30.10)
a3 dt
with

J :=. φ̇K,X + 6HX G 3,X − 2φ̇G 3,φ − 12HX G 4,φX + 6H 2 φ̇(G 4,X + 2X G 4,XX )
+2H 3 X (3G 5,X + 2X G 5,XX ) + 6H 2 φ̇(G 5,φ + X G 5,φX ), (30.11)

. Pφ := K,φ − 2X (G 3,φφ + φ̈G 3,φX ) + 6(2H 2 + Ḣ )G 4,φ

+6H (Ẋ + 2HX )G 4,φX − 6H 2 X G 5,φφ + 2H 3 X φ̇G 5,φX . (30.12)

Lastly, as usual we consider the matter conservation equation .ρ̇m + 3H (ρm + pm ) =

0.
In the following subsections we will present various sub-cases of Horndeski the-
ories that can alleviate .H0 tension. Since Horndeski theory recovers .ɅCDM cos-
mology for .G 4 = 1/(16π G), .K = −2Ʌ = const, and .G 3 = G 5 = 0, our strategy
is to introduce deviations which are negligible at high redshifts, where the CMB
structure is formed, but become significant at low redshifts, in which local Hubble
measurements take place.

30.2.1 Horndeski Theories with Non-minimal Derivative

Coupling

One can first examine one interesting subclass of Horndeski gravity that contains
the .G 5 term, which is called “non-minimal derivative coupling”. In particular, one
556 E. N. Saridakis

Fig. 30.1 Left Graph: The normalized .H (z)/(1 + z)3/2 in units of km/s/Mpc as a function of the
redshift, for .ɅCDM cosmology (black–solid) and for Model I with .V0 = 0.08 and with .G 5 (X ) =
ξ X 2 , for.ξ = 1.5 (green–dotted),.ξ = 1.3 (red–dashed) and.ξ = 1 (blue–dashed-dotted), in.H0 units.
We have imposed .Ωm0 ≈ 0.31. Right graph: The corresponding normalized effective Newton’s
G
constant . Geff given in (30.15) as a function of the redshift. The figure is from [11]

can consider models with .G 4 = 1/(16π G) and .G 3 = 0, which is the case in .ɅCDM
cosmology, and impose a simple scalar-field potential and standard kinetic term,
namely.K = −V (φ) + X [11]. Additionally, since.G 5 affects the friction terms of the
scalar field [12–14], one chooses the .G 5 term to depend only on .X , i.e. .G 5 (φ, X ) =
G 5 (X ) [11]. Inserting into (30.8), (30.9) gives the effective dark energy density and
pressure [11]

ρDE
. = 2X − K + 2H X φ̇(5G 5,X + 2X G 5,XX ),
3
(30.13)
( 3 )
. = K − 2X G 5,X 2H φ̇ + 2H Ḣ φ̇ + 3H φ̈ − 4H X φ̈G 5,XX .
2 2 2
pDE (30.14)

One can choose . G 5 (X ) suitably in order for .H (z) to coincide with

√
.HɅCDM (z) := H0 Ωm0 (1 + z)3 + 1 − Ωm0 at .z = zCMB ≈ 1100, namely
.H (z → zCMB ) ≈ HɅCDM (z → zCMB ), but satisfy .H (z → 0) > HɅCDM (z → 0).
For simplicity, we focus on dust matter (i.e. .pm = 0), and without loss of generality
we consider .K = −V0 φ + X .

• Model I: .G 5 (X ) = ξ X 2
A first model is the one with .G 5 (X ) = ξ X 2 , i.e where .G 5 has a quadratic depen-
dence on the field’s kinetic energy. In the left graph of Fig. 30.1 we depict the
normalized .H (z)/(1 + z)3/2 for various choices of .ξ . As we observe, the scenario
coincides with .ɅCDM one at high and intermediate redshifts, while at small red-
shifts it leads to higher values. In particular, the present-day value .H0 depends on
the model parameter .ξ , and it can be around .H0 ≈ 74 km/s/Mpc for .ξ = 1.3 (note
that since .ξ has dimensions of .[M ]−9 and since .H0 ≈ 10−61 in Planck units, this
implies that .ξ 1/9 ∼ 1040 GeV.−1 ). As we see, .H0 tension can be alleviated at 3.σ if
.1.2 < ξ < 1.7. The mechanism behind the alleviation is the decreased effective
Newton’s constant, brought about in turn by the friction term. Since in Horndeski
theories we have [15, 16]
30 Modified Gravity 557

G eff 1( )−1
. = G 4 − 2X G 4,X + X G 5,φ − φ̇HX G 5,X , (30.15)
G 2
one can easily see that in the present Model I this exhibits a decrease as can be
observed in the right graph of Fig. 30.1. Thus, although .wDE in this scenario does
not evolve into the phantom regime, the decrease in.G eff is adequate to alleviate the
tension. Finally, one can check that the scenario at hand is stable at the perturbative
level too [11].
• Model II: .G 5 (X ) = λX 4
In the case.G 5 (X ) = λX 4 , one obtains a similar behavior with that of Fig. 30.1, and
one can obtain .H0 ≈ 74 km/s/Mpc for .λ = 1 in .H0 units (since .λ has dimensions
of .[M ]−17 we acquire .λ1/17 ∼ 1030 GeV.−1 ), and the tension can be alleviated at
3.σ if .0.5 < λ < 1.2 in .H0 units. Similarly to the previous model, the mechanism
behind the tension alleviation is the decreased .G eff at intermediate redshifts [11].
In summary, the above particular sub-class of Horndeski/generalized Galileon
gravity can alleviate the .H0 tension due to the effect of the kinetic-energy-dependent
. G 5 term on decreasing . G eff . Nevertheless, note that models with . G 5 / = 0 need some
fine-tuning in order to obtain gravitational wave speed equal to the speed of light
[11].

30.2.2 Shift-Symmetric Horndeski Theories

Let us now focus on shift-symmetric Horndeski theories, namely models that are
invariant under the transformation .φ → φ + c where .c is constant. The cubic-order
shift-symmetric action reads as [17]
{ [ 2 ]
√ MPl
S=
. d4 x −g R + G 2 (X ) + G 3 (X )□φ , (30.16)
2

where .MPl is the reduced Planck mass. In this case we obtain the effective dark
density and pressure as

. ρDE = −G 2 − 2φ̇ 2 G 2,X + 6 H φ̇ 3 G 3,X , (30.17)

. pDE = G 2 − 2φ̈ φ̇ G 3,X .
2
(30.18)

This subclass of Horndeski gravity can also alleviate the .H0 tension. In particular,
one can choose:

• Galileon Ghost Condensate (GGC)

One can consider the Galileon Ghost Condensate (GGC) model, having [18, 19]

. G 2 (X ) = a1 X + a2 X 2 , G 3 (X ) = 3a3 X , (30.19)
558 E. N. Saridakis

with .a1 , .a2 , .a3 constants. This model deviates from the Cubic Galileon scenario
[20] due to the term .a2 X 2 , which is responsible for altering the late-time cosmic
evolution, since it prevents the Universe from approaching the tracker solution.
We mention here that the mechanism behind the .H0 tension alleviation is the
fact that dark-energy equation-of-state parameter can lie in the phantom regime
(.wDE < −1) without exhibiting ghosts [19, 21]. Note also that this model
suppresses the large-scale CMB temperature anisotropy comparing to .ɅCDM
paradigm. These features imply statistical preference of GGC over .ɅCDM
when data from ( CMB, )BAO, −1 SNIa, and RSDs are used [21]. In summary, one
obtains .H0 = 69.3+3.6 −3.0 km s Mpc−1 at 95% confidence level with best fit
−1
Mpc−1 [21], consistent within 2.σ with the direct measurements.
bf
.H0 = 70 km s
• Generalized Cubic Covariant Galileon (GCCG):
A different shift-symmetric model is the Generalized Cubic Covariant Galileon
(GCCG) one [9], having
4(1−p) 1−4p3
. G 2 (X ) = −c2 M2 (−X /2)p , G 3 (X ) = −c3 M3 (−X /2)p3 , (30.20)

with .M2 , .M3 constants with dimension .[M ], and where .c2 , .c3 , .p, .p3 are dimen-
sionless constants (one can also re-define them as .q = p3 − p + 21 and .s = p/q).
The GCCG model is an extension of the Covariant Galileon (.s = 2, .q = 1/2), and
similarly to the latter it also accepts a tracker solution with .H φ̇ 2q = constant for
+8 −1
.q > 0. In such a case one acquires .H0 = 72−5 km s Mpc−1 at 95% CL, finding
a consistency with Cepheids measurements at 1.σ [22], while it is also compat-
ible with the estimation obtained with the tip of the red giant branch in LMC,
−1
.H0 = (72.4 ± 2) km s Mpc−1 [23]. Nevertheless, note that for the GCCG model
if the tension between CMB data and low redshift .H0 measurements disappears,
another tension between the latter and BAO data appears [22, 24].

30.2.3 Brans-Dicke Theories

Brans-Dicke (BD) theory [25] was the first scalar-tensor theory that appeared in the
literature, which later on was discovered that it falls within Horndeski class. It is
characterized by the action
{ [ ( ) ] {
√ 1 ωBD μν √
. S= dx4 −g ϕR − g ∂μ ϕ∂ν ϕ − ρɅ + dx4 −gLm ,
16π ϕ
(30.21)
where .ρɅ is the usual vacuum energy and .ωBD (or equivalently .∈BD := 1/ωBD ) is
a dimensionless parameter. Note that one can generalize the original BD theory by
introducing a potential .V (ϕ) for the scalar field. Moreover, the Klein-Gordon equa-
tion is .□ϕ = 8π GT /(3 + 2ωBD ), with .T the trace of the total energy-momentum
tensor. As expected, General Relativity is recovered by considering the simultaneous
30 Modified Gravity 559

Fig. 30.2 The .1σ and .2σ

confidence level regions in
the (.σ8 –.H0 )-plane for the
GR-.ɅCDM, type-II RRVM
and BD-.ɅCDM (see text),
together with the data points
used in [27]. The type-II
RRVM and BD-.ɅCDM are
able to alleviate the .H0
tension without worsening
the .σ8 one. .H0 is expressed
in units of . km s−1 Mpc−1 .
The figure is from [28]

limits .ϕ → 1 and .∈BD → 0, namely by suppressing the dynamics of .ϕ and match-

ing its constant value with the Newton’s constant .G, and actually the first condition
is needed to obtain GR.+Ʌ, while the second condition guarantees that BD theory
recovers the weak-field regime of GR, but not for higher orders in the parametrization
of post-Newtonian expansion [26].
Let us call BD-.ɅCDM the Brans-Dicke model with the usual vacuum energy,
a theory that has been recently confronted with a large string of cosmological data
[28], such as SNIa+BAO+.H (z)+CMB+LSS, where .H (z) may include or not the
S.H0 ES prior on .H0 . Focusing only on the results obtained including the latter, in
Fig. 30.2 we depict the results of the fittings. As we observe, BD-.ɅCDM scenario
is capable of reducing the .H0 (and .σ8 ) tension to .∼ 1.5 σ level (when the high-.l
polarization and lensing data from Planck 2018 are not considered) [28]. This is
similar to the case of type-II Ricci-Running-Vacuum models (RRVM), which the
BD-.ɅCDM model mimics [27]. Additionally, we mention that, as confirmed by
various information criteria, the model is favoured comparing to standard .ɅCDM
scenario [27, 28]. One should add that in order not to spoil the correct fit of the CMB
temperature data the model needs to accommodate values of the spectral index .ns
larger than those of the GR-.ɅCDM, nevertheless this does not cause problems since
the dynamics of .ϕ can compensate the corresponding changes in the matter power
spectrum [28]. Finally, note that the.H0 tension is alleviated for.|∈BD | ≾ O(10−3 ), and
since this parameter has an impact on the LSS the compatibility is possible provided
a screening mechanism is invoked.

30.2.4 Minimal Scalar-Tensor Theories

Let us now focus on a different subclass of Horndeski modified gravity that can
correspond to extended versions of Brans-Dicke teories. In particular, one considers
the action is [29, 30]
560 E. N. Saridakis
{ [ ]
√ F(φ) g μν
.S= d4 x −g R− ∂μ φ∂ν φ − V (φ) + Lm , (30.22)
2 2

where .F(φ) is a generic function of .φ. In such cases one can easily attain values for
H0 that are larger than the .ɅCDM one, due to a degeneracy between the coupling to
.
the Ricci curvature and .H0 , and is connected to the rolling of the scalar field which
regulates the gravitational strength.
In the case where .F(φ) = ξ φ 2 , with .ξ > 0 the coupling, namely of an effectively
massless scalar field at rest in the radiation era and with adiabatic initial condition for
the scalar field perturbations [31], the current constraints (on the Hubble
) parameter
from Planck 2018 data at 68% confidence level are .H0 = 69.6+0.8 −1.7 km s
−1
Mpc−1
[32]. When BAO data(from BOSS) DR12 are added, the constraint at 68% confidence
level becomes .H0 = 68.78+0.53 −0.78 km s
−1
Mpc−1 , while once a Gaussian likelihood
based on the determination of the Hubble constant from the S.H0 ES team is also
included one obtains .H0 = (70.1 ± 0.8) km s−1 Mpc−1 . Note that these constraints
have been obtained by fixing the value of the current value of the scalar field .φ0 :=
φ(z = 0) to
1 2F0 + 4F0,φ 2
. = G, (30.23)
8π F0 2F0 + 3F0,φ
2

where .F0 := F(φ0 ) and .F0,φ = ∂F/∂φ|φ=φ0 , in order to guarantee that the current
effective Newton’s constant corresponds to the standard one [33] (see [34–36] for
studies of these scenarios without imposing such condition).

30.3 .f (T) Gravity

In this section we will show how the .H0 tension can be alleviated in a different
class of gravitational modification, namely in the framework of torsional gravity.
In torsional formulation one uses as dynamical variables the vierbeins fields, which
form an orthonormal basis at a manifold point. In a coordinate basis they are related
to the metric through .gμν (x) = ηAB , eμA (x) , eνB (x), with Greek and Latin indices
respectively used for the coordinate and tangent space. One introduces the Weitzen-
böck connection .Wνμ λ
:= eAλ ∂μ eνA [37], and thus the corresponding torsion tensor
is .T μν := Wνμ − Wμν = eAλ (∂μ eνA − ∂ν eμA ). The torsion tensor incorporates all the
λ λ λ

information of the gravitational field, and the torsion scalar arises from its contraction
as
1 ρμν 1
.T := T Tρμν + T ρμν Tνμρ − Tρμρ T νμν . (30.24)
4 2
This forms the Lagrangian of teleparallel gravity (similarly to general relativity where
the Lagrangian is the Ricci scalar), and since variation in terms of the vierbeins
gives the same equations with general relativity, the constructed theory was named
teleparallel equivalent of general relativity (TEGR).
30 Modified Gravity 561

In its most simple formulation, it is possible to consider.f (T ) extensions of telepar-

allel gravity, intended as corrections to TEGR where only the torsion scalar .T is
considered. The action takes the form of
{
1 [ ]
.S = d4 x e T + f (T ) + Sm , (30.25)
16π G

where .f (T ) is a generic function of the torsion scalar .T , .Sm is the action of matter
√
fields, and .e = det(eμi ) = −g.
For a flat FLRW background, it is possible to derive the relation between the
torsion .T and the Hubble parameter, .T = 6H 2 . The sign of such a torsion element
depends on the signature of the metric, which is identified by the vierbien fields
.eμ = diag(−1, a, a, a). Assuming that matter sector is described by a perfect fluid
a

with energy density .ρ and pressure .p, the field equations give

8π G f (T ) TfT
H2 =
. ρm − + (30.26)
3 6 3
4π G (ρm + pm )
.Ḣ = − . (30.27)
1 + fT + 2 TfTT

Moreover, the set of equations is closed with the equation of continuity for the matter
sector .ρ̇m + 3H (ρm + pm ) = 0. Equations (30.26), (30.27) can be rewritten in terms
of the effective energy density .ρT and pressure .pT arising from .f (T ),

MPl2 [ ]
ρ
. f (T ) = 2 TfT − f (T ) (30.28)
2
[ ]
MPl2 f (T ) − TfT + 2 T 2 fTT
p
. f (T ) = , (30.29)
2 1 + fT + 2 TfTT

and define the effective torsion equation-of-state

pf (T ) f (T ) − TfT + 2 T 2 fTT
w :=
. =[ ][ ]. (30.30)
ρf (T ) 1 + fT + 2 TfTT 2TfT − f (T )

These effective models are hence responsible for the accelerated phases of the early
or/and late Universe [7].
In order to study the background evolution, we can rewrite the first FLRW equa-
tion, making explicit the form of the torsional energy density [38–40]
[ ]
H (a)2 −3 −4 1 '
. := E(a) = Ωm a + Ωr a + (−f + 2Tf ) ,
2
(30.31)
H0 T0

where we considered the relation .T = 6H 2 and with .T0 the present value of .T .
Let us now try to suitably extract an .f (T ) form that could alleviate .H0 tension. As
we mentioned in the Introduction, in order to achieve this, one needs either a positive
562 E. N. Saridakis

correction to the first Friedmann equation at late times that could yield an increase
in .H0 compared to the .ɅCDM scenario, or a weaker gravity, quantified by a smaller
. G eff . In the case of .f (T ) gravity the effective gravitational coupling is given by [38]

( )−1
∂f (T )
. G eff = G 1+ . (30.32)
∂T

According to the Effective Field Theory approach to .f (T ) gravity [41–43] one

can choose .f (T ) in order to obtain a Hubble function evolution of the form
/
d (z) d 2 (z)
.H (z) = − + + HɅCDM (z),
2
(30.33)
4 16
√
where.HɅCDM (z) := H0 Ωm (1 + z)3 + ΩɅ is the Hubble rate in.ɅCDM cosmology
after recombination, .Ωm is the present matter density parameter, and primes denote
derivatives with respect to .z. The function .d (z) can be selected to be .d < 0 in order
to have .H (z → zCMB ) ≈ HɅCDM (z → zCMB ), therefore the .H0 tension is solved. One
should choose .|d (z)| < H (z), and thus, since .H (z) decreases for smaller .z, the devi-
ation from .ɅCDM will be significant only at low redshifts. Additionally, since .G eff
becomes smaller than .G during the growth of matter perturbations, the .σ8 tension is
also solved.
We follow Ref. [42] and we consider the ansatz.f (T ) = −[6H02 (1 − Ωm ) + F(T )],
where.F(T ) describes the deviation from GR (mind the difference in the conventions).
Under these assumptions, the first Friedmann equation becomes

F ' (z)
T (z) + 2
. T (z) − F(z) = 6HɅCDM
2
(z) . (30.34)
T ' (z)

In order to solve the .H0 tension, we need .T (0) = 6H02 , with .H0 =
74.03 km s−1 Mpc−1 following the results with local measurements [44], while in
the early era of .z ≿ 1100 we require the Universe expansion to evolve as in .ɅCDM,
namely .H (z ≿ 1100) ≃ HɅCDM (z ≿ 1100). This implies .F(z)|z≿1100 ≃ cT 1/2 (z)
(the value .c = 0 corresponds to standard GR, while for .c /= 0 we obtain .ɅCDM
too).
By solving Eq. (30.34) with initial and boundary conditions at .z ∼ 0 and
.z ∼ 1100, we can find the .f (T ) form that can alleviate .H0 tension. In particular,

we find that we can fit the numerical solutions very efficiently through .F(T ) ≈
375.47 [T /(6H02 )]−1.65 , dubbed Model-1, and .F(T ) ≈ 375.47 [T /(6H02 )]−1.65 +
25T 1/2 , dubbed Model-2. We examine .G eff , and as expected we find that at high
redshifts in both models .G eff becomes .G, recovering the .ɅCDM paradigm. We
have checked that both models can easily pass the BBN constraints, which demand
.|G eff /G − 1| ≤ 0.2 [45], as well as the ones from the Solar System, which demand
' −3 −1
.|G
eff (z = 0)/G| ≤ 10 h and .|G eff "(z = 0)/G| ≤ 105 h−2 [46].
30 Modified Gravity 563

Fig. 30.3 Evolution of the Hubble parameter .H (z) in the two .f (T ) models (purple solid line) and
in .ɅCDM cosmology (black dashed line). The red point (with error bars) represents the latest data
from extragalactic Cepheid-based local measurement of .H0 [44]. The figure is from [42]

In Fig. 30.3 we present the evolution of .H (z), for the two .f (T ) models, and
we compare them with .ɅCDM. We stress that the .H0 tension can be alleviated
as .H (z) remains statistically consistent with all CMB and CC measurements at all
redshifts. We remind the reader that the two .f (T ) models differing merely by a term
.∝ T
1/2
, which does not affect the background as explained before, are degenerate
at the background level. However, at the perturbation level, the two models behave
differently as their gravitational coupling .G eff differs.
In summary, we conclude that the class of .f (T ) gravity described by

f (T ) = −2Ʌ/MPl2 + αT β ,
. (30.35)

where only two out of the three parameters .Ʌ, .α, and .β are independent (the third one
is eliminated using .Ωm0 ), can alleviate .H0 tension with suitable parameter choices.
And as we mentioned, the mechanism behind this alleviation is the fact that the
effective Newton’s constant.G eff in.f (T ) gravity can be smaller that.G at intermediate
times, thus leading to faster expansion. Moreover, these type of .f (T ) gravity models
could also be examined through galaxy-galaxy lensing effects [47], strong lensing
effects around black holes [48], and gravitational-wave experiments [49].
For completeness, we proceed to the investigation of different.f (T ) models too, i.e
models that are efficient in successfully passing the confrontation with observations
[38, 50].
1. The power-law model [38, 50] (hereafter .f1 CDM model), in which

.f (T ) = α(−T )b , (30.36)
564 E. N. Saridakis

Fig. 30.4 The likelihood

contours for the three .f (T )
models described in the text,
in the .H0 -.Ωm plane, for
CMB+lensing+BAO+R19+Pth+DES
data. The grey region
indicate the latest constrain
on Hubble constant of Riess,
i.e .H0 = .74.03 ± 1.42
km/s/Mpc [44]. The figure is
from [51]

ΩF0
where.α = (6H02 )1−b 2b−1 and hence the only free parameter is.b. Thus, for.b = 0
the model .f1 CDM reduces to .ɅCDM cosmology, i.e. .T + f (T ) = T − 2Ʌ, with
.Ʌ = 3ΩF0 H0 and .ΩF0 = ΩɅ0 .
2

2. The square-root exponential model (hereafter .f2 CDM) [38, 50]

√
f (T ) = αT0 (1 − e−p
.
T /T0
), (30.37)
ΩF0
with .α = 1−(1+p)e −p . This model reduces to .ɅCDM paradigm for .p → +∞, and

thus we can replace .p by .p = 1/b which tends to 1 for .b → 0+ .

3. The exponential model (hereafter .f3 CDM) [38, 50]:

f (T ) = αT0 (1 − e−pT /T0 ),

. (30.38)
ΩF0 +
with .α = 1−(1+2p)e −p , where using .p = 1/b implies that for .b → 0 the .f3 CDM
model reduces to .ɅCDM one.

In Fig. 30.4 we present the likelihood contours in the case of the aforementioned
three models. As we observe, the power-law, .f1 CDM model of (30.36), gives .H0 =
(73.85 ± 1.05) km s−1 Mpc−1 [51] (see also [52] for an analysis with several other
combinations of large-scale data). Hence, this allows to significantly relax the Hubble
constant tension, nevertheless it worsens .σ8 tension [53–55] since the correlation of
the two parameters does not seem to be removed.
Finally, we close this section mentioning that one can have other classes of tor-
sional gravity, arising from different extensions of TEGR, that can also alleviate the
.H0 tension, such as .f (T , B) models [56, 57] , where .B is the boundary term that
quantifies the difference of (Levi-Civita) Ricci scalar .R and (Weitzenböck) torsion
scalar .T [58].
30 Modified Gravity 565

30.4 .f (R) Gravity

In this section we investigate the possible alleviation of the .H0 tension in the frame-
work of another class of modified gravity, namely .f (R) gravity [8]. In .f (R) gravita-
tional theories one extends the Einstein-Hilbert action to
{
√ f (R)
.S = d 4 x −g . (30.39)
16π G

Following the metric formulation, variation of the action (30.39) with respect to the
metric .gμν leads to the field equations

1 (m)
F G. μν = − gμν (FR − f ) + ∇μ ∇ν F − gμν □F + 8π G Tμν , (30.40)
2
with .G μν = Rμν − (1/2) gμν R the Einstein tensor, .∇μ the covariant derivative, .□ :=
g μν ∇μ ∇ν , and where we have defined .F(R) := f,R = df (R)/dR. Additionally, .Tμν (m)

is the energy-momentum tensor for the matter sector.

Applying the above theory in a flat FRW metric, we obtain the modified Friedmann
equations as

H 2 = 8π G (ρm + ρDE ) ,
. (30.41)
. − 2Ḣ = 8π G (ρm + pm + ρDE + PDE ) , (30.42)

where we have considered that the matter sector corresponds to a perfect fluid with
energy density .ρm and pressure .pm , while the effective dark energy sector is charac-
terized by
[ ]
1 1
. ρDE := (FR − f ) − 3H Ḟ + 3 (1 − F) H ,
2
(30.43)
8π G 2
1 [ 1 ( )]
. pDE := − (FR − f ) + F̈ + 2H Ḟ − (1 − F) 2Ḣ + 3H 2 ,(30.44)
8π G 2
and thus we can introduce the effective dark-energy equation-of-state parameter as
.w := p(DE /ρDE . Finally,
) note that in flat FRW geometry one obtains the useful relation
.R = 6 2H + Ḣ .
2

Amongst the various specific models, we consider the exponential one, where
[59–61]: [ ]
( R )
.f (R) = R − 2Ʌ 1 − exp −β . (30.45)
2Ʌ

This model recovers .ɅCDM cosmology at the limit .β → ∞. Additionally, note that
at large redshifts the model also comes close to .ɅCDM scenario, since the curvature
becomes large, namely .R >> Ʌ/β.
566 E. N. Saridakis

Fig. 30.5 Likelihood contours for the exponential .f (R) gravity model (30.45), at .1σ , .2σ and .3σ
confidence levels, for SNe Ia + CC + .HBAO + CMB observational data. The figure is from [61]

Figure 30.5 depicts the likelihood contours at .1σ , .2σ and .3σ confidence levels
for SNe Ia + CC + .HBAO + CMB data, in the .H0 − Ω0m and .H0 − β planes, while
the absolute maximums are described by stars, circles, etc. [61]. The right panels
depict the same contour plots but with additional details, e.g. re-scaled along the .Ω0m
axis. As we observe, the .H0 tension can be partially alleviated in the framework of
exponential .f (R) gravity.
We close this section by mentioning that, as we saw, .H0 can only be partially
alleviated in exponential .f (R) gravity, and things are similar in the case of other
viable models of the literature such as the Hu-Sawicki one [62].

c1 RHS (R/RHS )p
f (R) = R −
. , (30.46)
c2 (R/RHS )p + 1
30 Modified Gravity 567

where .c1 , .c2 , .RHS , .p > 0 are the model parameters, or the Starobinsky one [63]
[ ( )−n ]
R2
f (R) = R − λRS
. 1− 1+ 2 , (30.47)
RS

with .λ(> 0), .RS , .n > 0 the model parameters, or the Tsujikawa model [64]
( )
R
.f (R) = R − μRT tanh , (30.48)
RT

with .μ(> 0) and .RT (> 0) the two positive parameters. The reason for this is that in
f (R) gravity the effective Newton’s constant is [65]
.

2
k fRR
G eff 1 1 + 4 a2 fR
. = , (30.49)
G fR 1 + 3 k 22 fRR
a fR

where note that we have a dependence on the wave-number .k. On the other hand,
perturbation analysis reveals that in order for an .f (R) model to be stable one should
have .f,R > 0 for .R ≥ R0 , with .R0 the present value of the Ricci scalar, in order to
avoid a ghost state, and .f,RR > 0 for .R ≥ R0 , in order to avoid the scalar-field degree
of freedom to become tachyonic [8]. Additionally, one should have .0 < Rff,R,RR (r) < 1
at .r = − Rff ,R = −2, in order to have the presence and stability of a late-time de Sitter
solution [8]. One can see that .f (R) choices that satisfy the above requirements typi-
G
cally lead to . Geff > 1 and thus operate in the opposite direction of the .H0 alleviation.
Nevertheless, such models can still alleviate the tension if they lead to strong phan-
tom behavior (according to the requirements of [2, 3]), however it is clear that the
full .H0 -tension alleviation is hard to be obtained since viable and pathologies-free
.f (R) models induce effects that simultaneously increase and decrease .H0 .

30.5 Bi-scalar Modified Theories of Gravity

As a last modified gravity model that can alleviate the .H0 tension we consider a class
of bi-scalar theories, characterized by the action [66, 67]
{
√ ( )
S=
. d 4 −g f R, (∇R)2 , □R , (30.50)

with .(∇R)2 = g μν ∇μ R∇ν R. In the following we focus on models with

.f (R, (∇R) , □R) = K ((R, (∇R) ) + G (R, (∇R) )□R. One can rewrite the above
2 2 2

action by converting the Lagrangian using double Lagrange multipliers, resulting to

actions of multi-scalar fields coupled minimally to gravity. In order to achieve it, one
568 E. N. Saridakis

fixes the dependence of .f on .□R = β. In particular,√the .χ and .φ fields are introduced

1 − 23 χ
through the conformal transformations .gμν = 2 e ĝμν , .ϕ := fβ , and they enter
in a specific combination in a way that the final form of the action is equivalent to the
original higher-derivative gravitational action. Doing so, and performing variation,
one results in the Friedmann equations [66, 67]

1
. H2 = (ρDE + ρm ) (30.51)
3
. 2Ḣ + 3H 2 = −(pDE + pm ), (30.52)

with the effective dark energy and pressure defined as

1 2 1 −2√ 2 χ 2 2 [ (√ ) ]
ρDE :=. χ̇ − e 3 K − φ̇ φ̇ 6χ̇ − 9H − 3φ̈ GB
2 4 [ 3 ]
1 −√ 2 χ φ ( )
+ e 3 Ḃφ̇GB + + φ̇ 2 Gφ − 2KB , (30.53)
2 2

( )
1 2 1 −2√ 2 χ 1 −√ 2 χ φ
pDE := χ̇ + . e 3 K + e 3 Ḃφ̇GB + φ̇ 2 Gφ − , (30.54)
2 4 2 2
√2
with .K = K (φ, B) and .G = G (φ, B), with .B = 2e 3 χ g μν ∇μ φ∇ν φ, and where
for simplicity we set the Planck mass to 1. Finally, one can define the effective
dark-energy equation-of-state parameter as .wDE := pDE /ρDE .
In the following we desire to find specific models which can coincide with.ɅCDM
cosmology at CMB redhsifts, while at low-redhsifts to deviate from it, giving rise to
higher .H0 .
• Model I: .K (φ, B) = 21 φ − ζ2 B and G (φ, B) = 0
In this case we solve the cosmological equations numerically and in the left
graph of Fig. 30.6 we present the normalised combination .H (z)/(1 + z)3/2 as
a function of the redshift for .ɅCDM cosmology, and for Model I for different
values of .ζ . We find that the present value of .H0 depends on the model parameter
.ζ as expected, and for .ζ = −10 the present value of the Hubble parameter is
around .H0 ≈ 74km/s/Mpc, which is consistent with the direct measurement of
the present Hubble parameter (note that in natural units this corresponds to a
typical value .ζ 1/4 ∼ −10−19 GeV.−1 ). Additionally, in the right graph of Fig. 30.6
we plot the evolution of .wDE (z), and as we can see it lies in the phantom regime
(.wDE < −1) most of the time. Hence, the mechanism behind .H0 alleviation in the
present model is phantom behavior, which is known to be one of the mechanisms
since it induces faster expansion [2, 3]. Notice the difference between the
single-field, Horndeski theories analyzed in Sect. 30.2.1 above, in which the
alleviation mechanism was the smaller .G eff .
30 Modified Gravity 569

Fig. 30.6 Left Graph: The normalized .H (z)/(1 + z)3/2 in units of km/s/Mpc as a function of
the redshift, for .ɅCDM cosmology (blue dotted line) and for Model I .ζ = −12 (solid blue line),
for .ζ = −10 (solid black line), and for .ζ = −8 (solid red line), in Planck units. We have imposed
.Ωm0 ≈ 0.31. Right graph: The corresponding effective dark-energy equation-of-state parameter
.wDE as a function of the redshift, for .ζ = −10 in Planck units. The figure is form [68]

• Model II: .K (φ, B) = 21 φ and G (f , B) = xiB,

In this case we repeat the same steps as in the previous model and we find that
the present Hubble value .H0 depends on the model parameter .ξ . Specifically, for
.ξ = −10 it is around .H0 ≈ 74km/s/Mpc (in natural units .ξ ∼ −10 corresponds
to a typical value .ξ 1/8 ∼ −10−19 GeV.−1 ). Similarly to the previous case, the
mechanism behind .H0 -tension alleviation is the phantom behavior.

30.6 Conclusions

Modified gravity is one of the best frameworks that can alleviate .H0 tension, and
simultaneously other tensions such as the .σ8 one, due to the huge freedom one has
to construct models. In particular, modified gravity can alter either the early-time
Universe evolution, in which case one obtains early-time solutions to the .H0 tension
which are considered in the next chapter, or it can alter the late-time evolution, and
thus it can lead to late-time solutions, or both.
Gravitational modifications affect the late-time evolution of the universe through
the new terms they bring in the Friedmann equations, namely in the effective dark-
energy sector, as well as through the effective Newton’s constant they induce. If these
effects lead to weaker gravity (smaller .G eff ) at suitable redshifts, or to more repulsive
effective dark-energy (for instance exhibiting phantom behavior) then they can cause
faster expansion comparing to .ɅCDM paradigm, and thus lead to an increased .H0
value.
We presented .H0 -tension alleviations based on various subclasses of Horn-
deski/Generalized Galileon theories, such as Horndeski gravity with non-minimal
derivative coupling, shift-symmetric Horndeski theories, Brans-Dicke theories, and
570 E. N. Saridakis

minimal scalar-tensor theories. Then we presented alleviations based on torsional

modifications of gravity, and in particular of .f (T ) gravity. Finally we presented
solutions based on .f (R) gravity and bi-scalar modified theories of gravity. Such
capabilities of gravitational modifications may be added in their other known advan-
tages, and act as additional indication that modified gravity might be needed for the
description of Nature.

Acknowledgements The author acknowledges the contribution of the LISA CosWG, and the
COST Actions CA18108 “Quantum Gravity Phenomenology in the multi-messenger approach”
and CA21136 “Addressing observational tensions in cosmology with systematics and fundamental
physics (CosmoVerse)”.

References

1. CANTATA Collaboration, E.N. Saridakis et al., Modified gravity and cosmology: an update
by the CANTATA network. arXiv:2105.12582 [gr-qc]
2. L. Heisenberg, H. Villarrubia-Rojo, J. Zosso, Simultaneously solving the .H0 and .σ8 tensions
with late dark energy. Phys. Dark Univ. Series 39, 101163 (2023). (arXiv:2201.11623 [astro-
ph.CO])
3. L. Heisenberg, H. Villarrubia-Rojo, J. Zosso, Can late-time extensions solve the .H0 and .σ8
tensions? Phys. Rev. D Series 106(4), 043503 (2022). (arXiv:2202.01202 [astro-ph.CO])
4. G.W. Horndeski, Second-order scalar-tensor field equations in a four-dimensional space. Int.
J. Theor. Phys. Series 10, 363–384 (1974)
5. A. De Felice, S. Tsujikawa, Generalized Galileon cosmology. Phys. Rev. D Series 84, 124029
(2011). (arXiv:1008.4236 [hep-th])
6. C. Deffayet, X. Gao, D.A. Steer, G. Zahariade, From k-essence to generalised Galileons. Phys.
Rev. D Series 84, 064039 (2011). (arXiv:1103.3260 [hep-th])
7. Y.-F. Cai, S. Capozziello, M. De Laurentis, E.N. Saridakis, f(T) teleparallel gravity and cos-
mology. Rept. Prog. Phys. Series 79(10), 106901 (2016). (arXiv:1511.07586 [gr-qc])
8. A. De Felice, S. Tsujikawa, f(R) theories. Living Rev. Rel. Series 13, 3 (2010).
(arXiv:1002.4928 [gr-qc])
9. A. De Felice, S. Tsujikawa, Conditions for the cosmological viability of the most general
scalar-tensor theories and their applications to extended Galileon dark energy models. JCAP
series 02, 007 (2012). (arXiv:1110.3878 [gr-qc])
10. T. Kobayashi, M. Yamaguchi, J. Yokoyama, Generalized G-inflation: inflation with the
most general second-order field equations. Prog. Theor. Phys. Series 126, 511–529 (2011).
(arXiv:1105.5723 [hep-th])
11. M. Petronikolou, S. Basilakos, E.N. Saridakis, Alleviating .H0 tension in Horndeski gravity.
arXiv:2110.01338 [gr-qc]
12. E.N. Saridakis, S.V. Sushkov, Quintessence and phantom cosmology with non-minimal deriva-
tive coupling. Phys. Rev. D Series 81, 083510 (2010). (arXiv:1002.3478 [gr-qc])
13. G. Koutsoumbas, K. Ntrekis, E. Papantonopoulos, E.N. Saridakis, Unification of dark matter—
dark energy in generalized Galileon theories. JCAP Series 02, 003 (2018). (arXiv:1704.08640
[gr-qc])
14. S. Karydas, E. Papantonopoulos, E.N. Saridakis, Successful Higgs inflation from com-
bined nonminimal and derivative couplings. Phys. Rev. D Series 104(2), 023530 (2021).
(arXiv:2102.08450 [gr-qc])
15. E. Bellini, I. Sawicki, Maximal freedom at minimum cost: linear large-scale structure in general
modifications of gravity. JCAP Series 07, 050 (2014). (arXiv:1404.3713 [astro-ph.CO])
30 Modified Gravity 571

16. S. Peirone, K. Koyama, L. Pogosian, M. Raveri, A. Silvestri, Large-scale structure phenomenol-

ogy of viable Horndeski theories. Phys. Rev. D Series 97(4), 043519 (2018). (arXiv:1712.00444
[astro-ph.CO])
17. T. Kobayashi, Horndeski theory and beyond: a review. Rept. Prog. Phys. Series 82(8), 086901
(2019). (arXiv:1901.07183 [gr-qc])
18. C. Deffayet, O. Pujolas, I. Sawicki, A. Vikman, Imperfect dark energy from kinetic gravity
braiding. JCAP Series 10, 026 (2010). (arXiv:1008.0048 [hep-th])
19. R. Kase, S. Tsujikawa, Dark energy scenario consistent with GW170817 in theories beyond
Horndeski gravity. Phys. Rev. D Series 97(10), 103501 (2018). (arXiv:1802.02728 [gr-qc])
20. C. Deffayet, G. Esposito-Farese, A. Vikman, Covariant Galileon. Phys. Rev. D Series 79,
084003 (2009). (arXiv:0901.1314 [hep-th])
21. S. Peirone, G. Benevento, N. Frusciante, S. Tsujikawa, Cosmological data favor Galileon ghost
condensate over.ɅCDM. Phys. Rev. D Series 100(6), 063540 (2019). (arXiv:1905.05166 [astro-
ph.CO])
22. N. Frusciante, S. Peirone, L. Atayde, A. De Felice, Phenomenology of the generalized cubic
covariant Galileon model and cosmological bounds. Phys. Rev. D Series 101(6), 064001 (2020).
(arXiv:1912.07586 [astro-ph.CO])
23. W. Yuan, A.G. Riess, L.M. Macri, S. Casertano, D. Scolnic, Consistent calibration of the tip of
the red giant branch in the large Magellanic cloud on the Hubble space telescope photometric
system and a re-determination of the Hubble constant. Astrophys. J. Series 886, 61 (2019).
(arXiv:1908.00993 [astro-ph.GA])
24. J. Renk, M. Zumalacárregui, F. Montanari, A. Barreira, Galileon gravity in light of ISW, CMB,
BAO and H.0 data. JCAP Series 10, 020 (2017). (arXiv:1707.02263 [astro-ph.CO])
25. C.H. Brans, R.H. Dicke, Mach’s principle and a relativistic theory of gravitation. Phys. Rev
Series 124, 925 (1961)
26. V. Faraoni, J. Côté, A. Giusti, Do solar system experiments constrain scalar–tensor gravity?
Eur. Phys. J. C Series 80(2), 132 (2020). (arXiv:1906.05957 [gr-qc])
27. J. Solà Peracaula, A. Gómez-Valent, J. de Cruz Pérez, C. Moreno-Pulido, Running vacuum
against the .H0 and .σ8 tensions. EPL 134(1), 19001 (2021) arXiv:2102.12758 [astro-ph.CO]
28. J. Solà Peracaula, A. Gómez-Valent, J. de Cruz Pérez, C. Moreno-Pulido, Brans–Dicke cos-
mology with a .Ʌ-term: a possible solution to .ɅCDM tensions. Class. Quant. Grav. 37(24),
245003 (2020) arXiv:2006.04273 [astro-ph.CO]
29. S. Capozziello, R. de Ritis, Noether’s symmetries and exact solutions in flat nonminimally
coupled cosmological models. Class. Quant. Grav. Series 11, 107–117 (1994)
30. Y.-F. Cai, E.N. Saridakis, M.R. Setare, J.-Q. Xia, Quintom cosmology: theoretical implications
and observations. Phys. Rept. Series 493, 1–60 (2010). (arXiv:0909.2776 [hep-th])
31. D. Paoletti, M. Braglia, F. Finelli, M. Ballardini, C. Umiltà, Isocurvature fluctuations in the
effective Newton’s constant. Phys. Dark Univ. Series 25, 100307 (2019). (arXiv:1809.03201
[astro-ph.CO])
32. M. Ballardini, M. Braglia, F. Finelli, D. Paoletti, A.A. Starobinsky, C. Umiltà, Scalar-tensor
theories of gravity, neutrino physics, and the .H0 tension. JCAP Series 10, 044 (2020).
(arXiv:2004.14349 [astro-ph.CO])
33. B. Boisseau, G. Esposito-Farese, D. Polarski, A.A. Starobinsky, Reconstruction of a scalar
tensor theory of gravity in an accelerating universe. Phys. Rev. Lett. Series 85, 2236 (2000).
(arXiv:gr-qc/0001066)
34. Ö. Akarsu, N. Katırcı, N. Özdemir, J.A. Vázquez, Anisotropic massive Brans-Dicke grav-
ity extension of the standard .ɅCDM model. Eur. Phys. J. C Series 80(1), 32 (2020).
(arXiv:1903.06679 [gr-qc])
35. S. Joudaki, P.G. Ferreira, N.A. Lima, H.A. Winther, Testing gravity on cosmic scales:
a case study of Jordan-Brans-Dicke theory. Phys. Rev. D Series 105(4), 043522 (2022).
(arXiv:2010.15278 [astro-ph.CO])
36. M. Ballardini, F. Finelli, D. Sapone, Cosmological constraints on Newton’s gravitational con-
stant. arXiv:2111.09168 [astro-ph.CO]
37. R. Weitzenböock, Invariantentheorie (Noordhoff, Gronningen, 1923)
572 E. N. Saridakis

38. S. Nesseris, S. Basilakos, E.N. Saridakis, L. Perivolaropoulos, Viable .f (T ) models are practi-
cally indistinguishable from.ɅCDM. Phys. Rev. D Series 88, 103010 (2013). (arXiv:1308.6142
[astro-ph.CO])
39. R.C. Nunes, Structure formation in .f (T ) gravity and a solution for .H0 tension. JCAP Series
05, 052 (2018). (arXiv:1802.02281 [gr-qc])
40. R.C. Nunes, S. Pan, E.N. Saridakis, New observational constraints on f(T) gravity from cosmic
chronometers. JCAP Series 08, 011 (2016). (arXiv:1606.04359 [gr-qc])
41. C. Li, Y. Cai, Y.-F. Cai, E.N. Saridakis, The effective field theory approach of teleparallel
gravity, .f (T ) gravity and beyond. JCAP Series 10, 001 (2018). (arXiv:1803.09818 [gr-qc])
42. S.-F. Yan, P. Zhang, J.-W. Chen, X.-Z. Zhang, Y.-F. Cai, E.N. Saridakis, Interpreting cosmolog-
ical tensions from the effective field theory of torsional gravity. Phys. Rev. D Series 101(12),
121301 (2020). (arXiv:1909.06388 [astro-ph.CO])
43. X. Ren, S.-F. Yan, Y. Zhao, Y.-F. Cai, E.N. Saridakis, Gaussian processes and effective field
theory of .f (T ) gravity under the .H0 tension. arXiv:2203.01926 [astro-ph.CO]
44. A.G. Riess, S. Casertano, W. Yuan, L.M. Macri, D. Scolnic, Large Magellanic cloud
Cepheid standards provide a 1% foundation for the determination of the Hubble constant
and stronger evidence for physics beyond .ɅCDM. Astrophys. J. Series 876(1), 85 (2019).
(arXiv:1903.07603 [astro-ph.CO])
45. C.J. Copi, A.N. Davis, L.M. Krauss, A New nucleosynthesis constraint on the variation of G.
Phys. Rev. Lett. Series 92, 171301 (2004). (arXiv:astro-ph/0311334)
46. S. Nesseris, L. Perivolaropoulos, The limits of extended quintessence. Phys. Rev. D Series 75,
023517 (2007). (arXiv:astro-ph/0611238)
47. Z. Chen, W. Luo, Y.-F. Cai, E.N. Saridakis, New test on general relativity and .f (T ) torsional
gravity from galaxy-galaxy weak lensing surveys. Phys. Rev. D Series 102(10), 104044 (2020).
(arXiv:1907.12225 [astro-ph.CO])
48. S.-F. Yan, C. Li, L. Xue, X. Ren, Y.-F. Cai, D.A. Easson, Y.-F. Yuan, H. Zhao, Testing the
equivalence principle via the shadow of black holes. Phys. Rev. Res. Series 2(2), 023164
(2020). (arXiv:1912.12629 [astro-ph.CO])
49. Y.-F. Cai, C. Li, E.N. Saridakis, L. Xue, .f (T ) gravity after GW170817 and GRB170817A.
Phys. Rev. D Series 97(10), 103513 (2018). (arXiv:1801.05827 [gr-qc])
50. S. Basilakos, S. Nesseris, F.K. Anagnostopoulos, E.N. Saridakis, Updated constraints on .f (T )
models using direct and indirect measurements of the Hubble parameter. JCAP Series 08, 008
(2018). (arXiv:1803.09278 [astro-ph.CO])
51. M. Benetti, S. Capozziello, G. Lambiase, Updating constraints on f(T) teleparallel cosmology
and the consistency with Big Bang nucleosynthesis. Mon. Not. Roy. Astron. Soc. Series 500(2),
1795–1805 (2020). (arXiv:2006.15335 [astro-ph.CO])
52. R. Briffa, C. Escamilla-Rivera, J.L. Said, J. Mifsud, N.L. Pullicino, Impact of .H0 priors on
.f (T ) late time cosmology arXiv:2108.03853 [astro-ph.CO]
53. N. MacCrann, J. Zuntz, S. Bridle, B. Jain, M.R. Becker, Cosmic discordance: are Planck CMB
and CFHTLenS weak lensing measurements out of tune? Mon. Not. Roy. Astron. Soc. Series
451(3), 2877–2888 (2015). (arXiv:1408.4742 [astro-ph.CO])
54. R.A. Battye, T. Charnock, A. Moss, Tension between the power spectrum of density per-
turbations measured on large and small scales. Phys. Rev. D Series 91(10), 103508 (2015).
(arXiv:1409.2769 [astro-ph.CO])
55. M. Benetti, L.L. Graef, J.S. Alcaniz, The .H0 and .σ8 tensions and the scale invariant spectrum.
JCAP Series 07, 066 (2018). (arXiv:1712.00677 [astro-ph.CO])
56. S. Bahamonde, C.G. Böhmer, M. Wright, Modified teleparallel theories of gravity. Phys. Rev.
D Series 92(10), 104042 (2015). (arXiv:1508.05120 [gr-qc])
57. G. Farrugia, J. Levi Said, V. Gakis, E.N. Saridakis, Gravitational waves in modified teleparallel
theories. Phys. Rev. D 97(12), 124064 (2018) arXiv:1804.07365 [gr-qc]
58. C. Escamilla-Rivera, J. Levi Said, Cosmological viable models in .f (T , B) theory as solutions
to the .H0 tension. Class. Quant. Grav. 37(16), 165002 (2020) arXiv:1909.10328 [gr-qc]
59. G. Cognola, E. Elizalde, S. Nojiri, S.D. Odintsov, L. Sebastiani, S. Zerbini, A class of viable
modified f(R) gravities describing inflation and the onset of accelerated expansion. Phys. Rev.
D Series 77, 046009 (2008). (arXiv:0712.4017 [hep-th])
30 Modified Gravity 573

60. R.C. Nunes, S. Pan, E.N. Saridakis, E.M.C. Abreu, New observational constraints on .f (R)
gravity from cosmic chronometers. JCAP Series 01, 005 (2017). (arXiv:1610.07518 [astro-
ph.CO])
61. S.D. Odintsov, D. Sáez-Chillón Gómez, G.S. Sharov, Analyzing the.H0 tension in.F(R) gravity
models. Nucl. Phys. B 966, 115377 (2021) arXiv:2011.03957 [gr-qc]
62. W. Hu, I. Sawicki, Models of f(R) cosmic acceleration that evade solar-system tests. Phys. Rev.
D Series 76, 064004 (2007). (arXiv:0705.1158 [astro-ph])
63. A.A. Starobinsky, Disappearing cosmological constant in f(R) gravity. JETP Lett. Series 86,
157–163 (2007). (arXiv:0706.2041 [astro-ph])
64. S. Tsujikawa, Observational signatures of .f (R) dark energy models that satisfy cosmological
and local gravity constraints. Phys. Rev. D Series 77, 023507 (2008). (arXiv:0709.1391 [astro-
ph])
65. S. Basilakos, S. Nesseris, L. Perivolaropoulos, Observational constraints on viable f(R)
parametrizations with geometrical and dynamical probes. Phys. Rev. D Series 87(12), 123529
(2013). (arXiv:1302.6051 [astro-ph.CO])
66. A. Naruko, D. Yoshida, S. Mukohyama, Gravitational scalar–tensor theory. Class. Quant. Grav.
33(9), 09LT01 (2016) arXiv:1512.06977 [gr-qc]
67. E.N. Saridakis, M. Tsoukalas, Cosmology in new gravitational scalar-tensor theories. Phys.
Rev. D Series 93(12), 124032 (2016). (arXiv:1601.06734 [gr-qc])
68. S. Banerjee, M. Petronikolou, E.N. Saridakis, Alleviating the.H0 tension with new gravitational
scalar tensor theories. Phys. Rev. D Series 108(2), 024012 (2023). (arXiv:2209.02426 [gr-qc])
Chapter 31
Early-Time Modified Gravity
and the Hubble Tension

Matteo Braglia

31.1 Introduction

As of the current writing of this review, the value of the Hubble parameter obtained
through distance ladder calibration of type Ia supernovae (SN Ia) from the Pan-
theon+ compilation [1] by the S. H0 ES collaboration [2] stands at . H0 = 73.0 ±
1.0 km s−1 Mpc−1 . In contrast, the one inferred by the cosmic microwave background
(CMB) measurements from the Planck DR3 [3] is. H0 = 67.36 ± 0.54 km s−1 Mpc−1 .
This discrepancy reveals a 5.σ tension between the two datasets. Importantly, the ten-
sion does not appear unique to the two datasets; rather, numerous studies suggest that
it may be a generic tension between indirect, or early time, and direct, or late time,
measurements of the Hubble parameter [4–9]. This has stimulated intense efforts in
the community, on two fronts. On one hand, it is crucial to try and identify possible
systematic effects that could plague such measurements. On the other, we should be
open to the possibility that the . H0 tension genuinely hints at new Physics beyond
the .ΛCDM standard cosmological model. Within this review, we embrace the latter
perspective, focusing on a particular form of novel Physics potentially indicated by
the tension: Early time Modified Gravity (EMG).
Cosmological tests of deviations from General Relativity (GR) have garnered con-
siderable attention (see e.g. the reviews [10, 11]), forming a crucial aspect of many
large cosmological collaborations. Throughout scientific history, modified gravity
effects have been invoked to account for observations that defy explanation within the
standard .ΛCDM framework. Consequently, it is unsurprising that various research
groups have recently turned to modified gravity as a potential resolution to the Hub-
ble tension. Such deviations from GR can manifest from the early universe to the
present day. Attempts to address the . H0 tension have been made at both low and high

M. Braglia (B)
Center for Cosmology and Particle Physics, New York University, 726 Broadway, New York, NY
10003, USA
e-mail: mb9289@nyu.edu
© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 575
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_31
576 M. Braglia

redshifts, involving modifications to gravity around the times of matter-dark energy

or radiation-matter equality, respectively. Our focus lies on the latter scenario.
The primary objective of this review is to provide an accessible explanation of how
early-time modifications to gravity impact the comoving size of the sound horizon at
baryon drag (.rs ), creating a degeneracy with the . H0 value inferred from CMB data.
Essentially, a modification to gravity can be understood as injecting energy into the
cosmic fluid, thereby altering the Hubble parameter during the early universe [12].
This mechanism has been extensively explored in the literature as a potential solution
to the Hubble tension, utilizing various phenomenological parameterizations [13–16]
and models where both the background and the perturbed cosmological dynamics is
fully worked out. The latter include Jordan-Brans-Dicke (JBD) models [17–23], mod-
els with nonminimal coupled scalar fields [22, 24–27] (sometimes also denoted as
extended quintessence [28, 29]), Galileon-like models within Horndeski theory [30,
31], or the so-called early modified gravity (EMG) model presented in Refs. [32,
33],1 but the list is far from comprehensive.
In this pursuit, we delve into how modifications to gravity during the early uni-
verse can either expand or contract the comoving sound horizon (.rs ). While the core
mechanism is universal across these models, the specific embodiment of the grav-
ity modification holds paramount importance. Modern data possess the sensitivity
required to scrutinize the cosmological evolution of different model dynamics, thus
making model-building a central aspect.
Our review does not aim for exhaustive coverage of each individual model. Instead,
we thoroughly examine the EMG model introduced in Refs. [32, 33], illustrating
the aforementioned features. Within this model, a self-interacting scalar field non-
minimally coupled to gravity evolves from the radiation era, driven by its coupling
to non-relativistic matter. Subsequently, the scalar field enters a phase of coherent
oscillations due to a quartic term in its potential, causing its energy density to dilute
faster than non-relativistic matter after recombination. The reduction in scalar field
amplitude naturally suppresses gravity modifications at late redshifts, aligning excel-
lently with laboratory and Solar System tests of GR without necessitating screening
mechanisms.
The review is structured as follows. In Sect. 31.2, we introduce the formalism of
Scalar-Tensor (ST) theories, and discuss experimental constraints on deviations from
GR. Then, in Sect. 31.3, we we delve into the EMG dynamics and explain the features
that make it an appealing candidate to solve the . H0 tension. Finally we present some
of the latest cosmological constraints and summarize their status in Sect. 31.4.

1 Here, we are just adopting the nomenclature used in the literature. In fact, all the models listed

posses some kind of non-minimal coupling (NMC) to gravity, as well as produce an early time
modification to it.
31 Early-Time Modified Gravity and the Hubble Tension 577

31.2 Scalar-Tensor Theories and Experimental Constraints

In this paper, we will be dealing with ST theories where a scalar field .σ is non-
minimally coupled to the Ricci scalar as follows:
{ [ ]
√ F(σ ) g μν
. S= d x −g
4
R− ∂μ σ ∂ν σ − Λ − V (σ ) + Lm , (31.1)
2 2

where .Lm is the matter Lagrangian, .V is the scalar field potential and .Λ is the
cosmological constant. The field equations derived from the variation of the action
with respect to the field .σ and the metric are the Klein-Gordon (KG) and Friedmann
equations respectively and are given by

. σ̈ + 3H σ̇ + Vσ = 21 Fσ R, (31.2)
σ̇ 2
. 3H F = ρ +
2
2
+ V − 3H Ḟ ≡ ρ + ρscf . (31.3)

It is clear that the model and its cosmological dynamics are then uniquely specified
by the non-minimal coupling . F(σ ), the potential .V (σ ) and the initial conditions on
.σ deep in the radiation era. The form of . F(σ ) is arbitrary, as long as it satisfy . F > 0
and . F (2 F + 3 Fσ2 ) > 0, required to avoid negative kinetic terms in the quadratic
action of tensor and scalar field perturbations respectively [34].
Comparing to the Einstein-Hilbert action, we immediately note that the Planck
mass . Mpl2 ≡ 1/(8π G) is now a functional of .σ , and therefore can vary with time
according to the scalar field evolution. Specifically, one can define a cosmolog-
ical Newton constant .G N = 1/(8π F) and an effective Newton constant as [34,
35] .G eff = (2F + 4Fσ2 )/(2F + 3Fσ2 )(8π F). The latter is the quantity governing the
interactions between test bodies.
Deviations from general relativity (GR) can also be parameterized by expand-
ing the metric in powers of the gravitational potential .ϕ in the weak field regime
as .ds 2 = −(1 + 2ϕ − 2βPN ϕ2 )dt 2 + (1 − 2γPN ϕ)dxi dx i , where we encapsulated
modifications to gravity in the functions .γPN and .βPN . They are usually denoted as
Post-Newtonian (PN) parameters [36] and, for the Lagrangian in Eq. (31.1), are given
by .γPN = 1 − F,σ2 /(F + 2F,σ2 ) and .βPN = 1 + dγdσPN F F,σ /(8F + 12F,σ2 ). For GR we
have . Fσ = 0 so that both .γPN and .βPN reduce to 1.
Before we discuss constraints on this class of theories from large scale cosmo-
logical data, let us use quickly summarize other experimental constraints on the
quantities we have just introduced.
Big Bang Nucleosynthesis (BBN) constraints. The time at which weak and
nuclear processes freeze out is extremely sensitive to the expansion rate . H , which
is in turn proportional to the cosmological Newton constant, see e.g. Eq. (31.3) (and
also Eq. (31.4) in the next Section), and is therefore tightly constrained by BBN
observations. The most up-to-date constraints on the variation of the gravitational
constant .G N are .ΔG N = 0.02 ± 0.06 at .95% CL based on Ref. [37]. Let us note that
this constraint is only valid at very early times, long before recombination.
578 M. Braglia

Laboratory constraints. Constraints on the deviation from GR at later times can

be obtained through laboratory experiments. Modern techniques used to constrain
. G eff include for example atom interferometry and extremely precise torsion balances.
Based on such measurements, the current value recommended by CODATA is .G =
6.67430(15) × 10−11 N.·m2 · kg−2 , with a standard uncertainty of .22 ppm. Different
experiments are in tension on the 4th figure after the comma, so there is currently an
uncertainty (in units where .G = 1) of .10−4 [36].
Solar System constraints. Finally, solar system tests set stringent bounds on the
PN parameters at late times [36]. The tightest constraint .γPN − 1 = (2.1 ± 2.3) ×
10−5 was derived in Ref. [38] using the Doppler tracking of the Cassini spacecraft.
Observations of the anomalous 43 arcseconds perihelion shift of Mercury’s orbit
are affected by a combination of .γPN and .βPN . Breaking this degeneracy using the
bounds on .γPN quoted above, Ref. [36] derived .βPN − 1 = (4.1 ± 7.8) × 10−5 .
Hence, it becomes evident that although deviations from GR are tightly con-
strained at late times and within the scales of our Solar System, the constraints are
notably less stringent when high redshifts. As we will demonstrate, these limitations
are lenient to the extent that they permit significant dynamic variations in early times,
which could potentially yield intriguing implications for the . H0 tension.

31.3 Dynamics of Early Modified Gravity

Let us delve into the dynamics of the EMG. For the moment being, let us refrain from
making any assumptions on the function . F(σ ). We will only require that the scalar
field potential has a global minimum at .σ = 0, and its second derivative is much
smaller than the Hubble rate at very high redshift, before matter-radiation equality. If
the latter condition is not met, the scalar field becomes dynamical at very early times
and settles into its minimum long before recombination where the visibility function
has no support. As a result there would be no imprints on the CMB anisotropies.
Additionally, we will assume that the scalar field never dominates the total energy
budget of the Universe. In other words, its effective energy density defined through
.ρscf in Eq. (31.3) is always much smaller than .ρcrit .

With these assumptions in mind, the evolution of .σ can be understood intuitively

by examining the KG Eq. (31.2). At very early times.z ≫ z eq , the Universe is radiation
dominated so that the Ricci scalar is extremely small, and the source term on the
right hand side of (31.2) is negligible. Simultaneously, since the field is very light by
assumption, the source term.Vσ is also very small and the KG equation reduces to.σ̈ +
3H σ̇ ≅ 0. The Hubble friction thus keeps the scalar field frozen to its initial condition
.σi . This contributes to a shift in the cosmological Newton constant as . G N (z ≫
z eq )/G N (z = 0) /= 1. Furthermore, neglecting, the term.3H Ḟ in Eq. (31.3), the NMC
induces a relative displacement of the Hubble rate, compared to the standard .ΛCDM
evolution, as given by:
31 Early-Time Modified Gravity and the Hubble Tension 579

Fig. 31.1 For some illustrative parameters, we plot the evolution of the scalar field [top-left], the
relative variation of . H with respect to .ΛCDM [top-right], the relative variation of the cosmological
Newton constant [bottom-left] and the post-Newtonian parameter .γPN [bottom-right]

H EMG − H CDM Mpl2 − F(σi )

. (z ≫ z eq ) ∼ , (31.4)
H CDM 2
as depicted in the top-right panel of Fig. 31.1. Equation (31.4) is crucial in compre-
hending the potential of ST models in addressing the . H0 tension. These models have
the ability to decrease (.ξ > 0) or increase (.ξ < 0) the Hubble rate at early times
with a{ corresponding increase/decrease in the comoving size of the sound horizon
∞
.r s =
z dec dz cs (z)/H (z) respectively, which is highly degenerate with . H0 [12].
As we approach matter-radiation equality, the Ricci scalar is no longer negligible,
causing the source term .−Fσ R on the right hand side of the KG equation to grad-
ually activate. This drives .σ to larger/smaller values for positive/negative couplings
respectively. At some critical redshift, that we denote by .z c , the second derivative of
the potential becomes comparable to the Hubble rate, and the potential source term
can no longer be neglected, eventually dominating the dynamics of the scalar field.
Thus far, we have presented the general feature of the class of theories in Eq. (31.1).
To gain a deeper understanding, let us now specify the functional form for the NMC
and the potential. As illustrative example, we consider the EMG model introduced
in Ref. [32], corresponding to the choice
580 M. Braglia

. F(σ ) = Mpl2 + ξ σ 2 , (31.5)

. V (σ ) = λ σ /4,
4
(31.6)

where .ξ and .λ are dimensionless constants2 . It is noteworthy that the minimally

coupled limit, i.e. the .ξ = 0 limit, of this model corresponds to the RnR EDE model
introduced in Ref. [39]. To illustrate the cosmological evolution of our model, we fix
the initial conditions on the scalar fields to3 .σi = 0.375, σ̇i = 0, and, for a handful
of representative combinations of .ξ and .λ, numerically integrate the background KG
and Friedmann equations using the Einstein-Boltzmann code hi_class [40–43].
The results are displayed in Fig. 31.1. We now turn to comment each cases separately.
Massless case (.λ = 0), positive coupling (.ξ > 0). This case, represented by the
purple line in Fig. 31.1, was examined in Refs. [22, 24, 25]. As we approach .z eq the
scalar field begins to thaw, and grows until today, leading to a deviation of .G N and
the PN parameters from their observed value. This results in a departure from the
late-time.ΛCDM standard cosmological evolution and fails to adhere to Solar System
constraints. In principle, we could get around the former problem by promoting . Mpl
in Eq. (31.5) to a free parameter (see e.g. Ref. [14, 24]), while the second issue must
be tackled by incorporating a screening mechanism. We will not discuss this case
further in the following.
Massless case (.λ = 0), negative coupling (.ξ < 0). This case was studied in
Refs. [24–27], and is more interesting due to several reasons. Because the potential
is absent, the scalar field is entirely driven by the coupling to non- relativistic matter
around the radiation-matter equality era as before. However, unlike the positive cou-
pling case, it decreases towards .σ = 0 and its energy density dilutes away as fast or
faster than radiation, reducing .rs in a similar way to extra relativistc degrees of free-
dom [44–46], at least for what concerns the background evolution [25, 26]. Notably,
unlike popular EDE models, which requires fine tuning in both the initial condition
.σi and the value of the potential .λ (see also next bullet point), here the scalar field

starts to roll in a natural way after radiation-matter equality, driven by its NMC to
gravity.4 Another positive feature of the model, as can be seen from Fig. 31.1, is the
consistency with late time Solar System constraints, unlike most models of modified
gravity that necessitate screening mechanisms [11].
General case (.λ /= 0). This is the most general case, where the potential inter-
venes around the critical redshift .z c at which the second derivative of the potential
becomes comparable to the Hubble rate. This makes the evolution of .σ resemble
the one of EDE models (.ξ = 0 case in the figure) where the scalar fields undergoes
damped oscillations around the minimum at .σ = 0 and is diluted much faster than
radiation, and injects a large and sharp (in time) amount of energy into the cosmic

2 For convenience, we define .V0 through .λ = 102V0 /(3.516 × 10109 ) as in Ref. [32].
3 The condition .σ̇i = 0 is not restricting, as any non-zero .σ˙i would be immediately damped by
Hubble friction.
4 As discussed in Refs. [24, 25] and further emphasized in Ref. [27], the invariance of the theory

under conformal transformations could also justify setting.ξ = −1/6, leading to a minimal extension
of the .ΛCDM with only 1 extra parameter.
31 Early-Time Modified Gravity and the Hubble Tension 581

fluid. However, the NMC produces some notable differences. First of all, as already
discussed, the scalar field starts to roll, driven by the NMC, before the potential
kicks in. Furthermore, not only do we have a sharp energy injection around .z c , but
also a constant constribution to .ΔH/H , as discussed above. Interestingly, the faster
dilution results in a more efficient decay of .σ , suppressing even more .G N , eff and the
PN parameters at late times.
Finally, let us stress that the models considered above provide only a glimpse into
early modifications to gravity. The interplay between the NMC to gravity and the
comoving sound horizon, and thus . H0 , can be observed in various other models [19–
23, 30, 31, 47, 48]. However, unilke the model detailed in this Section, not all of
these alternatives are consistent with late time Solar System or laboratory constraints.

31.4 Current Constraints on EMG and Relation to the . H0

Tension

In this chapter, we delve into the cosmological constraints that shed light on the
modified gravity theories introduced earlier. We will start by offering a detailed
account of the constraints applied to the Early Time Modified Gravity (EMG) model,
and subsequently, we will juxtapose the inferred value of . H0 with other models
discussed in Sects. 31.1–31.3.
Cosmological Constraints on the EMG Model. We present the cosmologi-
cal constraints on the EMG model using a variety of datasets. The triangle plot
in Fig. 31.2 provides a comprehensive view of these constraints. Notably, due to
the intrinsic degeneracy between the non-minimal coupling .ξ σi2 , which quantifies
.ΔH/H prior to recombination (see Eq. (31.4)) at early times, and the comoving
sound horizon .rs , the Hubble parameter derived from Planck data assumes a value of
+0.83
. H0 = 68.58−1.3 , km, s
−1
, Mpc−1 at the .95% confidence level (CL) [33]. This value
is higher than the . H0 value inferred from the standard .ΛCDM model. However, it is
important to note that Planck can only establish an upper limit on the .ξ σi2 so that,
no definitive detection of modifications to gravity can be made solely from Planck
data.
An intriguing development arises when ACT data is included. Although this inclu-
sion does not alter the posterior distribution of . H0 , it does lead to the detection of
a modification to gravity at a .95% CL, specifically .ξ σi2 = 0.033 ± 0.014. As high-
lighted in Ref. [33], taking both the data fit and the additional model parameters into
account, the EMG model becomes favored over the .ΛCDM model at a significance
of .≥ 2σ using the Aikike criterion. This preference is prominently driven by a fea-
ture observed in the Temperature-Temperature (TT) angular power spectra of both
Planck and ACT around .l ∼ 750, which the EMG model adeptly fits. Interestingly,
while ACT hints at other popular approaches aimed at resolving the . H0 tension, such
as Axion Early Dark Energy (EDE) models [57, 58], the addition of SPT data seems
to contradict those models, in contrast with the EMG model which is still favored by
data.
582 M. Braglia

Fig. 31.2 1D and 2D posterior distributions (68% and 95% C.L.) for . H0 and the EMG model
parameters. The data used include TT, TE, EE data from Planck DR3 [49], ACT DR4 [50] and
SPT [51], lensing potential reconstructed from Planck DR3 [52], measurements of the BAO and
redshift space distortion ‘. f σ8 ’ from the CMASS and LOWZ galaxy samples of BOSS DR12 at
.z = 0.38, 0.51, and 0.61 [53] and the BAO measurements from 6dFGS at.z = 0.106 and SDSS DR7
at .z = 0.15 [54, 55], information about the luminosity distance to over 1550 SN Ia in the redshift
range .0.001 < z < 2.3 [56], and finally the SH0ES determination of the SN Ia magnitude . Mb =
−19.253 ± 0.027 [2] modeled as a Gaussian likelihood. See Ref. [33] for a detailed description

The triangle plots in the figure highlight that incorporating external (non-CMB)
datasets minimally impacts the constraints on model parameters. Only when a Gaus-
sian prior on the Supernova Type Ia (SN Ia) magnitude (. Mb ) is included, a significant
shift is observed due to the correlation between the EMG energy injection and . H0 .
Furthermore, we quantify the residual tension with the SH0ES collaboration.
Utilizing the . Q DMAP tension metric introduced in [6, 59], we observe a substantial
reduction in the tension on . H0 , from .6.4σ in the .ΛCDM model to .3.3σ in the EMG
31 Early-Time Modified Gravity and the Hubble Tension 583

Fig. 31.3 Constraints on . H0 from P18+BAO data (error-bars) and mean values of . H0 (black stars)
for some models involving early times modifications to gravity. Constraints are taken from the
references cited in the legend. The black stars come from analyses involving priors on . H0 from the
SH0ES collaboration. We warn the reader that different models were analyzed with priors on . H0
that reflect the time of publication of the papers cited in the figure

model.5 While the EMG model does not outright resolve the tension, it does play a
role in alleviating it.
The results are intriguing, prompting further investigation and testing of the
model’s robustness against forthcoming cosmological data releases. This is particu-
larly crucial given the current marginal significance of the detection.
Model Comparison. We depict the estimated . H0 values for the models discussed
in this review in Fig. 31.3, and we compare them with the EMG model. As access
to constraints from CMB data alone for all models was not available, we present
the estimated values for the Planck+BAO data combination, which generally yields
slightly lower values of. H0 than the Planck-only data, as BAO are highly constraining.
Evidently, models that introduce modifications to gravity before or around recombi-
nation consistently predict higher values of . H0 in comparison to the .ΛCDM model,
because of the degeneracies outlined in Sect. 31.3.
Additionally, we incorporate information from the SH0ES measurement in the
form of a . H0 prior or an . Mb prior, represented by the black stars in the figure.
However, we must caution the reader about the differing publication dates of the cited
papers, leading to distinct priors for each model. Consequently, a direct comparison
becomes challenging. Nevertheless, it is evident that the EMG model, as introduced
in Ref. [32], exhibits promising potential in addressing the . H0 tension.
In summary, this review has delved into the cosmological constraints of the EMG
model and compared its predictions for . H0 with other early-time modified gravity

5This estimate of the tension in .ΛCDM, quoted in Ref. [33], is larger than what is reported by the
SH0ES collaboration [2] due to the inclusion of new CMB data and the use of the full Pantheon+
dataset [56].
584 M. Braglia

models. While the EMG model does not fully resolve the tension, the mechanism
addressing the . H0 tension explained in this review offers valuable insights and lays
the basis for studying modified gravity models that can finally resolve the . H0 ten-
sion. Furthermore, we note that modifying the gravitational strength certainly impact
the efficiency of the clustering of matter perturbations as well, with non-trivial con-
sequences on the growth of structures. This could potentially offer a lever arm on
addressing the so-called . S8 tension (see e.g. [60] for a review), which the models
described here fail to address (but do not worsen either).
In conclusion, this review has delved extensively into the cosmological constraints
surrounding the EMG model, offering a thorough understanding of its implica-
tions. By comparing the EMG model’s predictions for . H0 with those of other early-
time modified gravity models, we gain valuable insights into the intricate interplay
between gravity modifications and cosmological observations. Although the EMG
model does not entirely solve the. H0 tension, the mechanism elucidated in this review
lays a foundational understanding for exploring modified gravity models that possess
the potential to ultimately resolve this long-standing discrepancy.

References

1. A. Carr, T.M. Davis, D. Scolnic, D. Scolnic, K. Said, D. Brout, E.R. Peterson, R.

Kessler, Publ. Astron. Soc. Austral. 39, e046 (2022). https://doi.org/10.1017/pasa.2022.41
[arXiv:2112.01471 [astro-ph.CO]]
2. A.G. Riess, W. Yuan, L.M. Macri, D. Scolnic, D. Brout, S. Casertano, D.O. Jones, Y. Murakami,
L. Breuval, T.G. Brink et al., Astrophys. J. Lett. 934(1), L7 (2022). https://doi.org/10.3847/
2041-8213/ac5c5b [arXiv:2112.04510 [astro-ph.CO]]
3. N. Aghanim et al., [Planck], Astron. Astrophys. 641, A6 (2020) [erratum: Astron. Astrophys.
652, C4 (2021)]. https://doi.org/10.1051/0004-6361/201833910[arXiv:1807.06209 [astro-
ph.CO]]
4. E. Di Valentino, L.A. Anchordoqui, O. Akarsu, Y. Ali-Haimoud, L. Amendola, N. Arendse, M.
Asgari, M. Ballardini, S. Basilakos, E. Battistelli et al., Astropart. Phys. 131, 102605 (2021).
https://doi.org/10.1016/j.astropartphys.2021.102605 [arXiv:2008.11284 [astro-ph.CO]]
5. E. Di Valentino, O. Mena, S. Pan, L. Visinelli, W. Yang, A. Melchiorri, D.F. Mota, A.G. Riess,
J. Silk, Class. Quant. Grav. 38(15), 153001 (2021). https://doi.org/10.1088/1361-6382/ac086d
[arXiv:2103.01183 [astro-ph.CO]]
6. N. Schöneberg, G. Franco Abellán, A. Pérez Sánchez, S.J. Witte, V. Poulin, J. Les-
gourgues, Phys. Rept. 984, 1-55 (2022). https://doi.org/10.1016/j.physrep.2022.07.001
[arXiv:2107.10291 [astro-ph.CO]]
7. P. Shah, P. Lemos, O. Lahav, Astron. Astrophys. Rev. 29(1), 9 (2021). https://doi.org/10.1007/
s00159-021-00137-4 [arXiv:2109.01161 [astro-ph.CO]]
8. E. Abdalla, G. Franco Abellán, A. Aboubrahim, A. Agnello, O. Akarsu, Y. Akrami, G. Alestas,
D. Aloni, L. Amendola, L.A. Anchordoqui et al., JHEAp 34, 49-211 (2022).https://doi.org/10.
1016/j.jheap.2022.04.002 [arXiv:2203.06142 [astro-ph.CO]]
9. L. Verde, T. Treu, A.G. Riess, Nature Astron. 3, 891. https://doi.org/10.1038/s41550-019-
0902-0 [arXiv:1907.10625 [astro-ph.CO]]
10. T. Clifton, P.G. Ferreira, A. Padilla, C. Skordis, Phys. Rept. 513, 1-189 (2012). https://doi.org/
10.1016/j.physrep.2012.01.001 [arXiv:1106.2476 [astro-ph.CO]]
11. K. Koyama, Rept. Prog. Phys. 79(4), 046902 (2016). https://doi.org/10.1088/0034-4885/79/
4/046902 [arXiv:1504.04623 [astro-ph.CO]]
31 Early-Time Modified Gravity and the Hubble Tension 585

12. L. Knox, M. Millea, Phys. Rev. D 101(4), 043533 (2020). https://doi.org/10.1103/PhysRevD.

101.043533 [arXiv:1908.03663 [astro-ph.CO]]
13. M.X. Lin, M. Raveri, W. Hu, Phys. Rev. D 99(4), 043514 (2019). https://doi.org/10.1103/
PhysRevD.99.043514 [arXiv:1810.02333 [astro-ph.CO]]
14. M. Ballardini, F. Finelli, D. Sapone, JCAP 06(06), 004 (2022). https://doi.org/10.1088/1475-
7516/2022/06/004 [arXiv:2111.09168 [astro-ph.CO]]
15. Z. Sakr, D. Sapone, JCAP 03(03), 034 (2022). https://doi.org/10.1088/1475-7516/2022/03/
034 [arXiv:2112.14173 [astro-ph.CO]]
16. G. Benevento, J.A. Kable, G.E. Addison, C.L. Bennett, Astrophys. J. 935(2), 156 (2022).
https://doi.org/10.3847/1538-4357/ac80fd [arXiv:2202.09356 [astro-ph.CO]]
17. P. Jordan, Nature 164, 637–640 (1949). https://doi.org/10.1038/164637a0
18. C. Brans, R.H. Dicke, Phys. Rev. 124, 925–935 (1961). https://doi.org/10.1103/PhysRev.124.
925
19. C. Umiltà, M. Ballardini, F. Finelli, D. Paoletti, JCAP 08, 017 (2015). https://doi.org/10.1088/
1475-7516/2015/08/017 [arXiv:1507.00718 [astro-ph.CO]]
20. M. Ballardini, F. Finelli, C. Umiltà, D. Paoletti, JCAP 05, 067 (2016). https://doi.org/10.1088/
1475-7516/2016/05/067 [arXiv:1601.03387 [astro-ph.CO]]
21. J. Solà Peracaula, A. Gomez-Valent, J. de Cruz Pérez, C. Moreno-Pulido, Astrophys. J.
Lett. 886(1), L6 (2019). https://doi.org/10.3847/2041-8213/ab53e9[arXiv:1909.02554 [astro-
ph.CO]]
22. M. Ballardini, M. Braglia, F. Finelli, D. Paoletti, A.A. Starobinsky, C. Umiltà, JCAP 10, 044
(2020). https://doi.org/10.1088/1475-7516/2020/10/044 [arXiv:2004.14349 [astro-ph.CO]]
23. M. Ballardini, A.G. Ferrari, F. Finelli, JCAP 04, 029 (2023). https://doi.org/10.1088/1475-
7516/2023/04/029 [arXiv:2302.05291 [astro-ph.CO]]
24. M. Rossi, M. Ballardini, M. Braglia, F. Finelli, D. Paoletti, A.A. Starobinsky,
C. Umiltà, Phys. Rev. D 100(10), 103524 (2019). https://doi.org/10.1103/PhysRevD.100.
103524[arXiv:1906.10218 [astro-ph.CO]]
25. M. Braglia, M. Ballardini, W.T. Emond, F. Finelli, A.E. Gumrukcuoglu, K. Koyama,
D. Paoletti, Phys. Rev. D 102(2), 023529 (2020). https://doi.org/10.1103/PhysRevD.102.
023529[arXiv:2004.11161 [astro-ph.CO]]
26. G. Ballesteros, A. Notari, F. Rompineve, JCAP 11, 024 (2020). https://doi.org/10.1088/1475-
7516/2020/11/024 [arXiv:2004.05049 [astro-ph.CO]]
27. T. Adi, E.D. Kovetz, Phys. Rev. D 103(2), 023530 (2021). https://doi.org/10.1103/PhysRevD.
103.023530 [arXiv:2011.13853 [astro-ph.CO]]
28. F. Perrotta, C. Baccigalupi, S. Matarrese, Phys. Rev. D 61, 023507 (1999). https://doi.org/10.
1103/PhysRevD.61.023507 [arXiv:astro-ph/9906066 [astro-ph]]
29. N. Bartolo, M. Pietroni, Phys. Rev. D 61, 023518 (2000). https://doi.org/10.1103/PhysRevD.
61.023518 [arXiv:hep-ph/9908521 [hep-ph]]
30. M. Zumalacarregui, Phys. Rev. D 102(2), 023523 (2020). https://doi.org/10.1103/PhysRevD.
102.023523 [arXiv:2003.06396 [astro-ph.CO]]
31. A.G. Ferrari, M. Ballardini, F. Finelli, D. Paoletti, N. Mauri, [arXiv:2307.02987 [astro-ph.CO]]
32. M. Braglia, M. Ballardini, F. Finelli, K. Koyama, Phys. Rev. D 103(4), 043528 (2021). https://
doi.org/10.1103/PhysRevD.103.043528 [arXiv:2011.12934 [astro-ph.CO]]
33. G. Franco Abellán, M. Braglia, M. Ballardini, F. Finelli, V. Poulin, [arXiv:2308.12345 [astro-
ph.CO]]
34. R. Gannouji, D. Polarski, A. Ranquet, A.A. Starobinsky, JCAP 09, 016 (2006). https://doi.org/
10.1088/1475-7516/2006/09/016 [arXiv:astro-ph/0606287 [astro-ph]]
35. B. Boisseau, G. Esposito-Farese, D. Polarski, A.A. Starobinsky, Phys. Rev. Lett. 85, 2236
(2000). https://doi.org/10.1103/PhysRevLett.85.2236 [arXiv:gr-qc/0001066 [gr-qc]]
36. C.M. Will, Living Rev. Rel. 17, 4 (2014). https://doi.org/10.12942/lrr-2014-4
[arXiv:1403.7377 [gr-qc]]
37. J. Alvey, N. Sabti, M. Escudero, M. Fairbairn, Eur. Phys. J. C 80(2), 148 (2020). https://doi.
org/10.1140/epjc/s10052-020-7727-y [arXiv:1910.10730 [astro-ph.CO]]
586 M. Braglia

38. B. Bertotti, L. Iess, P. Tortora, Nature 425, 374–376 (2003). https://doi.org/10.1038/

nature01997
39. P. Agrawal, F.Y. Cyr-Racine, D. Pinner, L. Randall, [arXiv:1904.01016 [astro-ph.CO]]
40. J. Lesgourgues, [arXiv:1104.2932 [astro-ph.IM]]
41. D. Blas, J. Lesgourgues, T. Tram, JCAP 07, 034 (2011). https://doi.org/10.1088/1475-7516/
2011/07/034 [arXiv:1104.2933 [astro-ph.CO]]
42. M. Zumalacárregui, E. Bellini, I. Sawicki, J. Lesgourgues, P.G. Ferreira, JCAP 08, 019 (2017).
https://doi.org/10.1088/1475-7516/2017/08/019 [arXiv:1605.06102 [astro-ph.CO]]
43. E. Bellini, I. Sawicki, M. Zumalacárregui, JCAP 02, 008 (2020). https://doi.org/10.1088/1475-
7516/2020/02/008 [arXiv:1909.01828 [astro-ph.CO]]
44. A.G. Riess, L. Macri, S. Casertano, H. Lampeitl, H.C. Ferguson, A.V. Filippenko, S.W. Jha,
W. Li, R. Chornock, Astrophys. J. 730, 119 (2011) [erratum: Astrophys. J. 732, 129 (2011)].
https://doi.org/10.1088/0004-637X/732/2/129 [arXiv:1103.2976 [astro-ph.CO]]
45. M. Wyman, D.H. Rudd, R.A. Vanderveld, W. Hu, Phys. Rev. Lett. 112, no.5, 051302 (2014).
https://doi.org/10.1103/PhysRevLett.112.051302 [arXiv:1307.7715 [astro-ph.CO]]
46. J.L. Bernal, L. Verde, A.G. Riess, JCAP 10, 019 (2016). https://doi.org/10.1088/1475-7516/
2016/10/019 [arXiv:1607.05617 [astro-ph.CO]]
47. J. Solà Peracaula, A. Gómez-Valent, J. de Cruz Pérez, C. Moreno-Pulido, Class. Quant. Grav.
37(24), 245003 (2020). https://doi.org/10.1088/1361-6382/abbc43 [arXiv:2006.04273 [astro-
ph.CO]]
48. S. Joudaki, P.G. Ferreira, N.A. Lima, H.A. Winther, Phys. Rev. D 105(4), 043522 (2022).
https://doi.org/10.1103/PhysRevD.105.043522 [arXiv:2010.15278 [astro-ph.CO]]
49. N. Aghanim et al., [Planck], Astron. Astrophys. 641, A5 (2020). https://doi.org/10.1051/0004-
6361/201936386 [arXiv:1907.12875 [astro-ph.CO]]
50. S.K. Choi et al., [ACT], JCAP 12, 045 (2020). https://doi.org/10.1088/1475-7516/2020/12/
045 [arXiv:2007.07289 [astro-ph.CO]]
51. L. Balkenhol et al., [SPT-3G], Phys. Rev. D 108(2,) 023510 (2023). https://doi.org/10.1103/
PhysRevD.108.023510 [arXiv:2212.05642 [astro-ph.CO]]
52. N. Aghanim et al., [Planck], Astron. Astrophys. 641, A8 (2020). https://doi.org/10.1051/0004-
6361/201833886 [arXiv:1807.06210 [astro-ph.CO]]
53. S. Alam et al., [BOSS], Mon. Not. Roy. Astron. Soc. 470(3), 2617–2652 (2017). https://doi.
org/10.1093/mnras/stx721 [arXiv:1607.03155 [astro-ph.CO]]
54. F. Beutler, C. Blake, M. Colless, D.H. Jones, L. Staveley-Smith, L. Campbell, Q. Parker,
W. Saunders, F. Watson, Mon. Not. Roy. Astron. Soc. 416, 3017–3032 (2011). https://doi.org/
10.1111/j.1365-2966.2011.19250.x [arXiv:1106.3366 [astro-ph.CO]]
55. A.J. Ross, L. Samushia, C. Howlett, W.J. Percival, A. Burden, M. Manera, Mon. Not. Roy.
Astron. Soc. 449(1), 835-847 (2015). https://doi.org/10.1093/mnras/stv154 [arXiv:1409.3242
[astro-ph.CO]]
56. D. Brout, D. Scolnic, B. Popovic, A.G. Riess, J. Zuntz, R. Kessler, A. Carr, T.M. Davis,
S. Hinton, D. Jones et al., Astrophys. J. 938(2), 110 (2022). https://doi.org/10.3847/1538-
4357/ac8e04 [arXiv:2202.04077 [astro-ph.CO]]
57. V. Poulin, T.L. Smith, T. Karwal, M. Kamionkowski, Phys. Rev. Lett. 122(22), 221301 (2019).
https://doi.org/10.1103/PhysRevLett.122.221301 [arXiv:1811.04083 [astro-ph.CO]]
58. T.L. Smith, M. Lucca, V. Poulin, G.F. Abellan, L. Balkenhol, K. Benabed, S. Galli, R. Mur-
gia, Phys. Rev. D 106(4), 043526 (2022). https://doi.org/10.1103/PhysRevD.106.043526
[arXiv:2202.09379 [astro-ph.CO]]
59. M. Raveri, W. Hu, Phys. Rev. D 99(4), 043506 (2019). https://doi.org/10.1103/PhysRevD.99.
043506 [arXiv:1806.04649 [astro-ph.CO]]
60. E. Di Valentino, L.A. Anchordoqui, Ö. Akarsu, Y. Ali-Haimoud, L. Amendola, N. Arendse, M.
Asgari, M. Ballardini, S. Basilakos, E. Battistelli et al., Astropart. Phys. 131, 102604 (2021).
https://doi.org/10.1016/j.astropartphys.2021.102604 [arXiv:2008.11285 [astro-ph.CO]]
Chapter 32
Primordial Magnetic Fields
and the Hubble Tension

Karsten Jedamzik and Levon Pogosian

32.1 Introduction

The statistical significance of the Hubble tension, currently just over .5σ , is primarily
driven by the difference between the S. H0 ES measurement of . H0 = 73.04 ± 1.04
km/s/Mpc using Cepheid calibrated supernova [1] and the . H0 = 67.36 ± 0.54
km/s/Mpc obtained from fitting the .∧CDM model to the Planck CMB data [2].
Other independent measurements tend to re-enforce the tension [3], with a clear
trend of all measurements that do not rely on a model of recombination giving a
higher . H0 , in the .69-.73 km/s/Mpc range [4–9], and estimates that use the standard
treatment of recombination giving. H0 of around 67–68 km/s/Mpc [10–12]. This trend
points to a missing ingredient in the standard description of the universe prior and/or
during recombination—something that would help reduce the sound horizon at last
scattering, .r⋆ . One such ingredient could be a stochastic primordial magnetic field
(PMF) embedded in the plasma and generating inhomogeneities in the baryon distri-
bution, or baryon clumping [13, 14]. The inhomogeneities make the recombination
complete faster [13, 15], resulting in a smaller .r⋆ .
At the time of writing this Chapter, it remains an open question whether PMFs will
play a key role in resolving the Hubble tension. Preliminary investigations [16–20],
based on a toy-model of baryon clumping [13], have shown that cosmological data is
broadly compatible with the presence of PMFs, and they could help relieve the tension
to a certain extent. Moreover, the required strength of the PMF appears to be of just the
right order of magnitude to help explain all observed galactic, cluster and intergalactic
magnetic fields. Needless to say that, if confirmed, this would be a truly exciting

K. Jedamzik (B)
Laboratoire de Univers et Particules de Montpellier, UMR5299-CNRS, Universite de Montpellier,
34095 Montpellier, France
e-mail: karsten.jedamzik@umontpellier.fr
L. Pogosian (B)
Department of Physics, Simon Fraser University, Burnaby, BC V5A 1S6, Canada
e-mail: levon@sfu.ca
© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 587
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_32
588 K. Jedamzik and L. Pogosian

development, solving two puzzles at the same time. Achieving a higher level of
certainty requires obtaining the exact ionization history from compressible magneto-
hydrodynamics (MHD) simulations including the effects of radiation transport that
are currently underway.
In what follows, we start by providing an overview of cosmological mag-
netic fields, their possible primordial origin and evolution, and their observational
signatures. We then describe the physics of recombination in the presence of mag-
netic fields, introducing the simple model of baryon clumping used in the existing
studies. After presenting the current observational status of this model, we discuss
the current state of the MHD simulations and prospects for the future.

32.2 Primordial Magnetic Fields and Their Observational

Signatures

The PMF hypothesis was first proposed as a possible explanation of the origin of
galactic magnetic fields [21, 22]. All galaxies, irrespective of their type or age, appear
to contain a magnetic field of .∼micro-Gauss (.μG) strength that is coherent over the
extent of the galaxy. Clusters of galaxies are also known to have magnetic fields of
similar strength. The origin of the galactic and cluster fields is not entirely under-
stood [23–25]. It may well be that initially small fields of strength of .∼ 10−20 –.10−18
G excited by the Biermann battery operating at first structure formation are subse-
quently amplified by the small-scale and/or large-scale dynamo to reach approximate
equipartition of magnetic energy with the turbulent energy in collapsed structures
(see [26] for a review on astrophysical dynamos). Interestingly, .μG strength fields
are observed in high redshift galaxies that would be too young to have gone through
the number of revolutions necessary for the large-scale dynamo to work [27]. Still,
there is a possibility that the supernova explosions in protogalaxies provided the mag-
netic seed fields that were later amplified by compression, shearing and stochastic
motions [28, 29]. An alternative explanation is that magnetic fields emerged from the
evolution of the early Universe, having been generated either during cosmic phase
transitions (e.g. the electroweak transition) [30] or during inflation [31, 32] (see also
[23, 25, 33, 34]) for reviews).
The pre-recombination universe was fully ionized and, if a PMF was ever gen-
erated, dissipation due to magnetic diffusion would be negligible. Dissipation due
to the kinetic viscosity damping fluid motions induced by the PMF on small scales
would, however, significantly drain the energy of the PMF [35, 36] over the many
e-folds of cosmic expansion between the very early Universe and the recombination
era. Notwithstanding this fact, if only the smallest fraction .∼ 10−10 of a PMF initially
in equipartition with the radiation would survive, it could be sufficient to eliminate
the need for dynamo amplification altogether. As discussed below, it could also leave
observable signatures in the CMB.
32 Primordial Magnetic Fields and the Hubble Tension 589

While PMFs have been a subject of continuous study over many decades, there
were commonly considered to be unnecessary because less exotic astrophysical
explanations could not be ruled out. The interest in PMFs was renewed when evi-
dence emerged of magnetic fields in .∼Mpc size voids in the intergalactic space.
Observations of TeV blazars by Hess and Fermi [37–39] have led to the surprising
conclusion that magnetic fields exist in the extra-galactic medium between galax-
ies with a very large volume filling factor [40]. TeV gamma-rays emitted by the
blazar may pair-produce on the diffuse extra-galactic background star light. The
resulting electron-positron pairs will subsequently Compton scatter on CMB pho-
tons converting them to GeV gamma-rays. Observations of blazars have not found
these GeV photons with the flux expected. Though other less understood explana-
tions exist [41], the most straightforward explanation is that the electron-positron
pairs were deflected by magnetic fields out of the light cone. A lower limit on the
field strength of . B ≳ 3 × 10−15 G (assuming a coherence scale of .∼kpc) could be
inferred. While adding to the case for PMFs, it is not ruled out that such fields could
be the result of outflows from galaxies, filling essentially all space with magnetic
fields. The synchrotron emission from a few mega-parsec (Mpc) long ridge connect-
ing two merging clusters of galaxies is another observation that is well-matched by
simulations based on the PMF hypothesis [42].
The question of the origin of the magnetic fields in the universe is difficult to
settle without the help of further observations. It is hoped that blazar gamma-ray
observations by the future Cherenkov Telescope Array mission [43] may raise the
lower limit to the .∼3–10 pico-gauss (pG) range [44]. However, the only way to
be certain of the primordial origin of the observed fields is to find their signatures
in the CMB. Since the PMF survives Silk damping [35, 36], magnetosonic modes
on large scales, .∼1–10 Mpc, may lead to additional power in the high . tail of
the CMB temperature anisotropy spectrum [45–65]. Dissipation of magnetic fields
before recombination may lead to spectral distortions [66–70], and after recombina-
tion may lead to an increase of the optical depth [63, 68, 70–73]. The PMF may lead to
additional polarization anisotropies due to Faraday rotation [47–49, 65, 74–82] The
non-Gaussianity of the PMF may lead to a signal in the bi-spectrum or tri-spectrum
of the anisotropies [63, 83–93]. All of the above effects have led to upper limits in
the nano-gauss (nG) range. However, as we elaborate in the following subsection,
CMB appears to be most sensitive to the generation of small-scale baryonic density
fluctuations before recombination, with detectable PMF strength in the 0.01–0.1 nG
range [14]. Moreover, as shown in [16], the baryon clumping before recombination
may help relieve the Hubble tension.

32.3 Recombination with Primordial Magnetic Fields

A statistically homogeneous and isotropic PMF is characterized by a power spectrum

expected to be a smooth function of the Fourier number.k. On large scales, the universe
is a highly conducting plasma with the magnetic field “frozen-in” and diluting with
590 K. Jedamzik and L. Pogosian

the expansion as . B = B0 /a 2 , where . B0 is the present day (comoving) strength and .a

is the scale factor normalized to unity today. On smaller scales, the PMF generates
perturbations in the plasma and dissipates its strength. PMFs generated causally in
phase transitions generically [94] have a Batchelor spectrum [95] that monotonically
increases with .k until a peak at a certain small integral scale [96], beyond which it
decays. Inflationary magnetogenesis, on the other hand, can produce a range of
spectral shapes depending on the particular model (see e.g. [33] for a review), with
the simplest models [31, 32] predicting a scale-invariant spectrum. CMB constraints
are often quoted in terms of the comoving field strength smoothed on the 1 Mpc scale,
. B1Mpc , which, in many cases, is not a representative measure of the PMF. The more

relevant quantity for the baryon clumping effect is the effective comoving strength,
. Beff , quantifying the average magnetic total energy density. For scale-invariant fields,
the two measures are essentially the same. For causal fields, however, the difference
is quite dramatic, with . Beff ≫ B1Mpc . Unless specified otherwise, the field strengths
quoted in this paper are . Beff .
Unlike most of the evolution of the magnetized plasma in the early universe, when
it is an incompressible fluid, MHD becomes compressible before recombination for
scales well below the photon mean free path .lγ ∼ 1 Mpc1 . Photons are free-steaming
over .∼ kpc scales and their sole effect is to introduce a drag force on peculiar motions
in the CMB rest frame characterized by a linear drag coefficient.α. The Euler equation
for the baryon velocity field takes the form of

∂v ∇ρ 1
. + (v · ∇)v + cs2 = −αv − B × (∇ × B) (32.1)
∂t ρ 4πρ
−5
where the speed of sound, .cs ≈√ 2 × 10 c, is very small on scales well-below .lγ ,
while being much larger, .cs = 1/3c, on length scales larger than .lγ . The first term
on the right-hand side (RHS) is the photon drag, while the last term is the Lorentz
force due to the PMF, .B × (∇ × B) = ∇ B 2 /2 − (B · ∇)B, which pushes baryons
into regions of lower magnetic energy density. This force is opposed by the baryon-
photon fluid pressure characterized by .cs2 . On scales well-below .lγ , the plasma is free
to compress until the baryonic pressure gradients backreact.
Using the continuity equation,

∂ρ ( )
. + ∇ ρv = 0 , (32.2)
∂t
one can estimate the amplitude of the generated density fluctuations in a back-of-
the-envelope estimate [13]. Imagine a stochastic magnetic field at uniform baryon
density and negligible velocities initially. These are very realistic initial conditions,
since at earlier times the evolving .lγ is smaller than the magnetic fluctuations scale
. L, such that the plasma is incompressible, and photon diffusion suppresses any fluid
motion. At redshifts .z ∼ 104 –.106 , depending on the scale, the photon mean free

1 All cited length scales are comoving.

32 Primordial Magnetic Fields and the Hubble Tension 591

path .lγ starts exceeding the scale . L. At this point Eq. (32.1) applies and there is a
significant drop in .cs . Initially, only the terms on the RHS of Eq. (32.1) are important
√
and the plasma obtains a terminal velocity .v ≃ c2A /(αL), where .c A = B/ 4πρ is
the Alfven speed of the plasma. Using Eq. (32.2), it is straigthforward to show
that the density fluctuations grow as .δρ/ρ ≃ vt/L ≃ c2A t/αL 2 with time .t. These
density fluctuations become larger with time until either the pressure forces become
important in counteracting further compression, or the source magnetic stress term
decays. The former happens when the last term on the LHS of Eq. (32.1),.(cs2 /L)δρ/ρ,
is of the order of the magnetic force term .c2A /L. That is, density fluctuations cannot
become larger than.δρ/ρ ≲ (c A /cs )2 . It has been shown by numerical simulation [97]
that even in the highly dragged viscous MHD regime magnetic fields decay, albeit
slower than during turbulent MHD. Magnetic fields excite fluid motions which in turn
get dissipated by the photon drag. It was found that a magnetic structure decays when
the Eddy turn-over rate .v/L ≃ c2A /αL 2 equals the Hubble rate . H ≃ 1/t. Entering
this into the expression for .δρ/ρ one finds that the average density fluctuation cannot
exceed unity by much. This leads to the following estimate for the generated density
fluctuations: [ ( )2 ]
δρ cA
. ≃ min 1, . (32.3)
ρ cs

Since, at recombination,
( )
B
c = 4.34km/s
. A (32.4)
0.03nG

and .cs = 6.33km/s, even fairly weak fields may generate order-unity density
fluctuations on small scales. This simple estimate has subsequently been confirmed
by full numerical simulations [14].
As the Universe expands, density fluctuations on ever larger scales are generated
and subsequently decay. Causally generated PMFs loose a significant fraction of their
power with each e-fold as the peak of the Batchelor spectrum keeps moving to larger
scales and weaker fields according to . B(L) = B0 (L/L 0 )−5/2 . Scale-invariant fields,
on the other hand, are largely unaffected by the evolution. As discussed above, that
magnetic fields and the associated density fluctuations decay when the Eddy turnover
time equals the Hubble time. This can be used to derive a correlation between the
magnetic field strength and the scale of the smallest magnetic structure that survives
until just before recombination [97]:
( )
L
. Brec ≲ 80 pG , (32.5)
kpc

where approximate equality is attained for fields which do undergo dissipation in

the first place. For fields which are so weak and/or are on such large scales that
the Eddy turnover time never reaches the Hubble time, no dissipation ever occurs
and no correlation can be derived (hence the inequality sign). Thus, quite generally,
592 K. Jedamzik and L. Pogosian

Table 32.1 Different measures (in nG) of a PMF that has an effective comoving magnetic strength
of . Beff = 0.05 nG at the redshift of last-scattering .z ⋆ . All the measures are the same for scale-
invariant fields, but differ significantly for a causally generated PMF with a Batchelor spectrum
pre−collapse
. Beff at .z = 10
Spectrum . Beff at .z ⋆ . B1Mpc at .z ⋆
6 . Beff

Scale-invariant 0.05 0.05 0.05 0.05

Batchelor 0.05 .10
−9 5 0.008

for sufficiently strong fields, . B ∼ 10 − 100 pG, inhomogeneities are generated on

.∼0.1–1 kpc scales shortly before recombination. Further dissipation occurs across
recombination as a result of the sudden drop in the free electron density. This leads
to a factor of .6.2 drop in the strength of the causal PMF, while, again, a scale-
invariant field strength is not affected. After recombination, the PMF evolution is
almost entirely halted or at least significantly slowed down as recent numerical
simulations show [98]. The post-recombination field strength is the one relevant for
the subsequent structure formation and is commonly referred to as “pre-collapse”. For
causal fields, the resultant pre-collapse correlation is given by . B0 ≃ 5 pG(L/kpc)2 .
As the differences between various measures of the PMF are model-dependent and
can be a source of confusion, we provide a glossary in Table 32.1. It has been shown
by detailed simulations [99, 100] that, irrespective of their coherence length, pre-
collapse magnetic fields of .∼5 pG lead to final post-collapse cluster fields of .∼ .μG
with Faraday rotation measures in good agreement with observations.
The magnetically induced inhomogeneities are on scales much too small (i.e.
. ∼ 10 –.10 ) to be observed directly as a contribution to CMB anisotropies. How-
6 7

ever, the baryon clumping has a prominent effect on the rate of recombination,
speeding it up and resulting in a smaller sound horizon at photon-baryon decou-
pling, .r⋆ . The speed up of recombination follows from the recombination rate being
proportional to the square of the electron density .n e . Since, generally, .⟨n 2e ⟩ > ⟨n e ⟩2
in a clumpy universe, where .⟨⟩ denotes spatial average, the recombination rate is
enhanced compared to .∧CDM. A smaller .r⋆ raises the value of . H0 inferred from
the very accurately measured angular size of the sound horizon, helping reduce the
Hubble tension.

32.4 Inhomogeneous Recombination

The evolution of the cosmic electron density depends on the full details of the prob-
ability distribution function (PDF) . p(∆) to find baryons at density .∆⟨ρ⟩, where .⟨ρ⟩
is the average baryonic density and .∆ is an enhancement factor. The shape of the

2 It is noted that there is no conflict between the statement that the magnitude of . Beff diminishes
by a factor .6.2 during recombination, on the one hand, and Eq. 32.5 and this relation, on the other
hand, as the coherence length . L is increasing during the dissipation.
32 Primordial Magnetic Fields and the Hubble Tension 593

baryon density PDF is currently unknown, and neither is its evolution before recom-
bination, which could be substantial, particularly for blue Batchelor spectra. In the
absence of knowledge of the PDF, [14] and [16] utilized a simple three zone model
(see Sect. 32.4.1) to describe the distribution of baryons. In the three-zone model,
the amplitude of baryon inhomogeneity is controlled by the clumping factor
( )
⟨ρ 2 ⟩ − ⟨ρ⟩2
.b≡ (32.6)
⟨ρ⟩2

which sets the second moment of the PDF, while the higher order moments are
effectively determined by the choice of the three-zone model parameters. As we
show next, the higher moments, and the third moment in particular, play a very
important role in determining the enhancement of the recombination rate.
The evolution of the electron number density around recombination is given by

dn e ( hνα )
. + 3 H n e = −C αe n 2e − βn H0 e− T , (32.7)
dt
where .n e is the free electron density, .n H is the total matter number density (protons
and hydrogen), and.n H0 = n H − n e is the density of neutral hydrogen. For simplicity,
to illustrate the importance of the higher moment, we consider a universe with only
hydrogen (i.e. no helium). Since Eq. (32.7) is quadratic in .n e , and since .⟨n 2e ⟩ > ⟨n e ⟩2
in inhomogeneous universes, we expect the recombination rate to be enhanced when
inhomogeneities exist. It would be tempting to solve the above equation by the
introduction of a clumping factor given by the underlying density distribution, i.e.
.b = (⟨n H ⟩ − ⟨n H ⟩ )/⟨n H ⟩ , such that the recombination term would be replaced
2 2 2

by .αe (1 + b)⟨n e ⟩ and the effect of inhomogeneities would be entirely described

2

by the first two moments of the distribution, i.e. the average density .⟨n H ⟩ and the
variance .⟨n 2H ⟩. However, such a procedure would be incorrect as it would assume that
.⟨n e ⟩ ∼ ⟨n H ⟩. This is not the case if different density regions have different ionization
2 2

fractions .χe = n e /n H . Rather, as shown below, the average ionization depends on

all the moments of the underlying probability distribution of the density fluctuations
. P(n H ).
Consider detailed balance, i.e. set the RHS of Eq. (32.7) to be approximately zero.
This assumption holds fairly during the middle of recombination because both the
recombination and the photoionisation rates are much larger than the Hubble rate,
but is increasingly inaccurate towards the end of recombination. The equilibrium
ionization fraction can then be derived as
βe 1 hνα βe 1 − hνα
χ2 =
. e (n H − n e )e− T = e T (1 − χe ). (32.8)
αe n H
2 αe n H

Let us now allow for inhomogeneities in the total baryon density by defining

n
. H = ∆⟨n H ⟩, (32.9)
594 K. Jedamzik and L. Pogosian

where .⟨...⟩ denotes the ensemble average, which we will assume to be the same as
the average over any sufficiently large volume. Then we can write

βe 1 hνα A
χ2 =
. e e− T (1 − χe ) = (1 − χe ), (32.10)
αe ∆⟨n H ⟩ ∆

where we have defined

βe 1 − hνα
. A≡ e T . (32.11)
αe ⟨n H ⟩

From (32.10), we can write the local ionized fraction as

( )
A 4∆ 1/2 A
χ (∆) =
. e 1+ − . (32.12)
2∆ A 2∆

In the homogeneous case, .∆ = 1, we have

( )
A 4 1/2 A
χ 0 (∆ = 1) =
. e 1+ − . (32.13)
2 A 2

Note that in the limit . A ≫ 1 we have a full ionisation, i.e. .χe → 1, while in the
opposite limit, . A ≪ 1, we have .χe → 0.
For consistency, the inhomogeneity parameter .∆(x) (x are space coordinates)
must fulfill ʃ
1
. dV ∆(x) = 1 (32.14)
Vtot

where the integration is over the total spatial volume .Vtot . For simplicity, as assumed
earlier, we will take averages over this volume to be the same as ensemble averages
and the above constraint can be simply written as .⟨∆⟩ = 1. The average electron
density is given by
.⟨n e ⟩ = ⟨χe (∆)∆⟨n H ⟩⟩ , (32.15)

and the average ionzation fraction is

/
⟨n e ⟩ A 4∆
.⟨χe ⟩ ≡ = ⟨ 1+ − 1⟩. (32.16)
⟨n H ⟩ 2 A

Note that this is not the same as the average of Eq. (32.10), where .χe = n e /n H .
Namely, .⟨n e ⟩/⟨n H ⟩ /= ⟨n e /n H ⟩. The relevant quantity for us is actually .⟨n e ⟩, which
is what appears in the CMB calculations, hence the relevant quantity is the one in
Eq. (32.16).
One can now show that, for arbitrary distribution functions . P(n H ) or . P̃(∆) ful-
filling the integral constraint
32 Primordial Magnetic Fields and the Hubble Tension 595
ʃ
⟨∆⟩ ≡
. d∆∆ P̃(∆) = 1, (32.17)

the .⟨χe ⟩ of Eq. (32.16) is smaller than .χe0 . Namely, that

/ /
⟨ 4∆ ⟩ 4
. 1+ < 1+ . (32.18)
A A

This follows from Jensen’s inequality [101] which states that the expectation value
⟨g(X )⟩ of any concave function
.
√ .g(X ) of a random variable . X is smaller than .g(⟨X ⟩).
In our case, the function . 1 + 4∆/A is certainly concave and hence we have our
result.
For smaller density perturbations one can also see this without using Jensen’s theo-
rem by introducing.δ ≡ ∆ − 1, such that.⟨δ⟩ = 0, and performing a Taylor expansion
under the assumption that .δ < 1. Then,
/ /
⟨ 4 4δ ⟩ 4
. 1+ + ≈ 1+ − γ2 ⟨δ 2 ⟩ + γ3 ⟨δ 3 ⟩ − ..., (32.19)
A A A

where all .γi > 0. One can see that the leading second order term is negative (the
function is concave), as well as all even order moment terms, while the odd moments
come with a positive signs. This means, e.g., that PDFs with a large third moment
could reduce the impact on the ionized fraction which is a trend we have observed
while experimenting numerically with different three-zone models. Ultimately, the
full redshift evolution of the baryon PDF and the associated ionized fraction will be
determined exactly from MHD simulations that are currently in progress.

32.4.1 The Three-Zone Model

In the absence of the baryon density PDF derived from MHD simulations, when dis-
cussing the current observational constraints on the PMF-enhanced recombination,
we will use the simple three-zone model introduced in [13]. The model is described
by the density parameters .∆i and volume fractions . f Vi in each zone. Baryon densities
in the individual zones are simply given by .n ib = ⟨n b ⟩∆i . Parameters .∆i and . f Vi have
to fulfil the following constraints:

∑
3 ∑
3 ∑
3
. f Vi = 1, f Vi Δi = 1, f Vi ∆i2 = 1 + b , (32.20)
i=1 i=1 i=1

i.e. the total volume fraction is one and the three-zone model has average density.⟨n b ⟩
and clumping factor .b. This leads to three constraints for six free parameters . f Vi , ∆i ,
such that one may choose three parameters freely. To obtain the average ionization
596 K. Jedamzik and L. Pogosian

fraction .⟨χe ⟩, one computes the ionization fraction in each of the zones and take the
average, i.e.
∑3
.⟨χe ⟩ = f Vi Δi χei . (32.21)
i=1

In [13], the parameters were chosen to be. f V2 = 1/3,.∆1 = 0.1 and.∆2 = 1, which
we will subsequently refer to as the M1 model. In [16], for comparison, we introduced
a second model, M2, that had . f V2 = 1/3, .∆1 = 0.3 and .∆2 = 1. It is not clear at this
point if either of these models provides a good representation of the actual baryon
PDF, but we note that M2 has a larger third moment and, for reasons discussed earlier,
results in a lesser reduction of the average ionized fraction for the same value of the
clumping factor .b.

32.5 The Impact of Baryon Clumping on CMB Spectra

The discussion in this and the subsequent Sections is largely based on the results
obtained in [19], where the impact of clumping on the CMB spectra and, in par-
ticular, their effect on CMB polarization were studied. As previously discussed,
inhomogeneous recombination completes sooner, which shifts the peak of the vis-
ibility function to an earlier epoch. This lowers .r⋆ , defined as the comoving sound
horizon at the peak of the visibility function, which has the effect of shifting the
acoustic peaks in CMB spectra to smaller angular scales. One can compensate for
the shift with a larger value of . H0 , which may help in relieving the Hubble tension.
In addition to the shift of the peaks, earlier recombination means that CMB polar-
ization is produced at an earlier epoch and, as a consequence, has a larger overall
amplitude. This is because the value of the speed of sound .cs (on scales where the
baryon-photon fluid is tightly coupled) was larger at earlier times, and the amplitude
of polarization is set by the quadrupole of temperature anisotropy, which is derived
from the dipole, which is set by the time derivative of the monopole, which in turn
is proportional to .cs [102].
Clumping also broadens the visibility function, which is a consequence of the over-
dense baryon pockets recombining earlier and underdense baryon pockets recom-
bining later. This broadening tends to further enhance polarization, because of the
period of time during which polarization can be generated is longer.
The two effects, the shift of the peak and the broadening of the visibility function,
are illustrated in Fig. 32.1, which compares the visibility functions in the ∧CDM
model, with an M1 model with .b = 2 with all other parameters kept the same, and
with the M1 model (∧CDM+b ) that best fits the Planck CMB data combined with
the S. H0 ES prior on . H0 . The broadening effect is apparent from the lower peak,
since the visibility function is normalized to integrate to unity. We note that, while
the general trends in the visibility function are common to all clumping models,
32 Primordial Magnetic Fields and the Hubble Tension 597

Fig. 32.1 Impact of baryon clumping on the CMB visibility function. We compare the visibil-
ity functions in the Planck best-fit ∧CDM model (solid red line), an M1 model with .b = 2 and
all cosmological parameters set to the best-fit ∧CDM (solid orange line), and the ∧CDM+b
Planck+S. H0 ES best-fit M1 model (dotted blue line). One can see that clumping shifts the peak
of the visibility function to earlier times, and increases its width. Figure reproduced from [19]

the quantitative details are dependent on the shape and the evolution of the baryon
density PDF.
Another significant effect of clumping on CMB spectra comes from a modifica-
tion of the Silk damping scale .r D . Here, there are three competing effects. Firstly,
.r D decreases due to an overall earlier completion of recombination. This decrease,
however, could be negated by the fact that an earlier broad helium recombination
in clumping models results in a smaller electron density available at the later stages
of recombination, and due to the broadening of the visibility function. Since much
of the Silk damping occurs right at recombination, where the visibility function is
of order unity, the details of the visibility function play an important role. The first
effect, due the overall shift to higher redshifts, would reduce the Silk damping effect
on CMB spectra, as it pushes the onset of the damping tail to higher multipoles
. . The second and third effects, however, can also be important, and the balance
between them is model-dependent and varies with the clumping factor. In the best-fit
M1 model, we find that there is less Silk damping compared to ∧CDM. But, in the
best-fit M2 model (see Sect. 32.4.1), the Silk damping is virtually identical to that
in the best-fit ∧CDM. Also, in M1, at (observationally disallowed) high values of .b,
the Silk damping is actually enhanced. Additional discussion of the evolution of the
damping scale as a function of .b in a different clumping models can be found in [18].
The evolution of magnetically induced clumping is unknown at present and is a sig-
nificant source of uncertainty in observational constraints on the PMF. For example,
if clumping was stronger at .z ∼ 3000 than at .z ∼ 1200, the helium recombination
could be the dominant effect, inducing more Silk damping.
The reduction in Silk damping present in the M1 model is illustrated in Fig. 32.2,
which compares the CMB spectra in the Planck best-fit .∧CDM to those in the
598 K. Jedamzik and L. Pogosian

Fig. 32.2 Impact of baryon clumping on CMB power spectra. Shown is the relative difference
between the CMB spectra in the Planck best-fit .∧CDM model and the M1 models best-fit to
Planck+S. H0 ES (blue dashed line) and Planck+BAO+SN+S. H0 ES (orange solid line). In the case
of TE, which is positively or negatively correlated depending on . , we avoid divisions by zero by
computing .(C − C ∧C D M )/C , where .C is the absolute value of .C ∧C D M convolved with a
ref ref

Gaussian of width .σ = 100 centred at . . The plot illustrates the importance of small-scale CMB
anisotropies for constraining baryon clumping. Figure reproduced from [19]

Planck+S. H0 ES and Planck+BAO+SN+S. H0 ES best-fit M1 models. The

enhancements in the temperature (TT) and E-mode polarization (EE) spectra at
high . is due to the smaller .r D . The same enhancement is also seen in the
temperature-polarization cross-correlation (TE), which is negative due to the anti-
correlation between T and E at high . . This illustrates the fact that high resolu-
tion CMB measurements will be a key discriminant in constraining inhomogeneous
recombination models.3

32.6 Relieving the Hubble Tension with PMFs

The extent to which PMFs can help relieve the Hubble tension is limited by the fact
that reducing the sound horizon .r⋆ , by itself, can at best raise the value of . H0 to
.∼ 70 km/s/Mpc without causing new tensions between CMB, BAO and the large-

scale structure clustering. In what follows, we first reproduce the general argument
illustrating this point, originally made in [103], before describing the current obser-
vational status of the M1 baryon clumping model.

3 Figure 32.2 also shows that at . ≲ 20 the polarization is reduced, which is due to the lower best
fit value of the optical depth .τ .
32 Primordial Magnetic Fields and the Hubble Tension 599

32.6.1 Why Reducing . r⋆ Helps, But Can Not Fully Resolve

the Hubble Tension by Itself

The acoustic peaks in CMB anisotropy spectra provide a very accurate measurement
of the angular size of the sound horizon at recombination,
r⋆
θ =
. ⋆ , (32.22)
D(z ⋆ )

where . D(z ⋆ ) is the comoving distance from a present day observer to the last scat-
tering surface, where .z ⋆ is the redshift of the peak of the ʃvisibility function. In
∞
a given model, .r⋆ and . D(z ⋆ ) can be determined from .r⋆ = z⋆ cs (z)dz/H (z) and
ʃ z⋆
. D(z ⋆ ) =
0 c dz/H (z), where .cs (z) is the sound speed of the photon-baryon fluid,
. H (z) is the redshift-dependent cosmological expansion rate and .c is the speed of
light. To complete the prescription, one also needs to determine .z ⋆ using a model of
recombination.
The sound waves responsible for the acoustic peaks in the CMB spectra are also
imprinted in the galaxy power spectra as Baryon Acoustic Oscillations (BAO) with a
minor difference in this standard ruler. Rather than .r⋆ , which is the sound horizon at
photon decoupling, the scale imprinted in the BAO is the sound horizon at the baryon
decoupling .rd , also known as the “cosmic drag” epoch when the photon drag on
baryons becomes unimportant. The latter takes place at a slightly lower redshift than
recombination, so that .rd ≈ 1.02r⋆ in ∧CDM. The proportionality factor between
the two sound horizons does not change appreciably in alternative recombination
scenarios.
While the difference between.r⋆ and.rd is small, the difference in redshifts at which
the acoustic features in the CMB and BAO are observed is significant, with the latter
measured at .0 ≲ z ≲ 2.5 accessible to galaxy redshift surveys. The angular scale of
the BAO feature measured using galaxy correlations in the transverse direction to
the line of sight is
rd
.θ⊥ (z obs ) ≡ ,
BAO
(32.23)
D(z obs )

where .z obs is the redshift of the correlated galaxies.

Let us now consider Eqs. (32.22) and (32.23) while remaining agnostic about
the particular model that determines the sound horizon. Namely, let us treat .r⋆ as
an independent parameter and assume .rd = 1.02r⋆ . Let us also assume that after
recombination the expansion of the universe is well-described by the ∧CDM model
and, for simplicity, ignore the contribution of radiation to the distance integrals. D(z ⋆ )
and . D(z obs ). Then, for the CMB acoustic feature, we can write
(ʃ )−1
z⋆
r⋆ dz
.θ⋆ = √ , (32.24)
1/2
2998 Mpc 0 ωm (1 + z)3 + h 2 /ωm − 1
600 K. Jedamzik and L. Pogosian

Fig. 32.3 A plot illustrating that achieving a full agreement between CMB, BAO and S. H0 ES
through a reduction of .rd requires a higher value of .ωm . Shown are the CMB lines of degeneracy
between the sound horizon .rd and the Hubble constant . H0 at three different values of .ωm : .0.143,
.0.155 and .0.167. Also shown are the .68% and .95% CL bands derived from the combination
of all current BAO data marginalized over .ωm , the .∧CDM based bounds from Planck and the
determination of the Hubble constant by S. H0 ES. Figure reproduced from [103]

where .ωm = Ωm h 2 is the matter density today, .Ωm is the fractional matter density,
.h is . H0 in units of 100 km/s/Mpc, and .2998 Mpc = c/100km/s/Mpc. An analogous

equation for the BAO is obtained by replacing .(r⋆ , θ⋆ , z ⋆ ) with .(rd , θ⊥BAO , z obs ).
For a given .ωm , and with the precisely measured .θ⋆ , Eq. (32.24) defines a line
in the .rd -. H0 plane4 . Similarly, a BAO measurement at each different redshift also
defines a respective line in the.rd -. H0 plane. However, the slopes of the CMB and BAO
lines are very different due to the vast difference in redshifts at which the standard
ruler is observed, .z ⋆ ≈ 1100 for CMB vs .z obs ∼ 1 for BAO. This is illustrated in
Fig. 32.3 that shows the .rd − H0 degeneracy line for CMB at three different values
of .ωm , as well as the .68% and .95% confidence level (CL) regions derived from the
combination of all presently available BAO observations while treating .rd as as an
independent parameter and marginalizing over.ωm (see [104] for details). Also shown
are the ∧CDM constraint from Planck (in red) and the S. H0 ES determination of . H0
(the grey band). On can see from Fig. 32.3 that BAO is consistent with either Planck
or S. H0 ES, depending on the value of .rd , but the latter two are in clear disagreement
with each other for the Planck-∧CDM-preferred .ωm ≈ 0.143.
One can reconcile Planck with S. H0 ES by reducing .rd and “moving up” the CMB
.r d − H0 degeneracy line. However, doing so at a fixed .ωm would quickly move the
values of .rd and . H0 out of the purple band in Fig. 32.3 creating a tension with BAO.

4 From here on we will use .rd to represent the acoustic feature in both the CMB and BAO.
32 Primordial Magnetic Fields and the Hubble Tension 601

Fig. 32.4 The 68 and 95% confidence level bounds on . S8 and .Ωm . Shown are the constraints
derived by fitting the .∧CDM model to a joint dataset of DES and SN and to Planck, along with
the contours for Model 2 and Model 3 corresponding to the .ωm = 0.155 and .ωm = 0.167 .rd − H0
lines in Fig. 32.3. Figure reproduced from [103]

In order to reconcile CMB, BAO and S. H0 ES at the same time, one needs to reduce
r and increase the matter density, as illustrated in the Figure. In particular, a full
. d

resolution of the tension would require .ωm ≈ 0.167, which is substantially higher
than the best-fit ∧CDM value. A larger matter density results in more clustering
quantified by the . S8 parameter. It is defined as . S8 ≡ σ8 (Ωm /0.3)0.5 , where .σ8 is the
matter clustering amplitude on the scale of .8 h −1 Mpc. A larger . S8 would exacerbate
the already existing mild tension between its Planck best-fit value and that measured
from weak gravitational lensing of galaxies [105].
The reason why simply reducing .r⋆ would not fully solve the Hubble tension is
seen in Fig. 32.4. The figure shows the .68% and .95% CL constraints on . S8 and .Ωm
by DES supplemented by the Pantheon SN sample [106]. Also shown is the Planck
∧CDM best fit, corresponding to.ωm = 0.143, as well as the two cases corresponding
to .ωm = 0.155 (Model 2) and .ωm = 0.167 (Model 3). The constraints for the latter
two were obtained by simultaneously fitting BAO and CMB acoustic peaks at the
corresponding fixed values of .ωm . One can see that when attempting to bring CMB,
BAO and S. H0 ES in agreement with each other by reducing .rd , one increases the . S8
tension.
Thus, any proposed solution to the Hubble tension that amounts to reducing .rd
without other significant changes to ∧CDM would not be able to raise the CMB-
extracted value of . H0 above .70 km/s/Mpc. Baryon clumping belongs to that category
of models, hence we do not expect it to fully resolve the Hubble tension. But it could
still help relieve it to a statistically acceptable level.
602 K. Jedamzik and L. Pogosian

32.6.2 Observational Status of the M1 Baryon Clumping

Model

Without knowing the detailed ionized fraction evolution obtained from MHD simula-
tions with a PMF (see Sect. 32.7), the M1 three-zone model [13] was used in [17–20,
103] for exploring the ability of baryon clumping to resolve the Hubble tension.
Below we summarize the current observational constraints of the M1 model.
Figure 32.5 shows the marginalized posteriors for the clumping factor .b and . H0
from Planck, Planck+S. H0 ES and Planck+S. H0 ES+BAO+SN, where we used the
eBOSS DR16 BAO compilation from [12] and SN stands for the Pantheon supernovae
sample [106]. We see that Planck by itself shows no preference for clumping, with.b <
0.47 at 95% CL and only a small increase in the best-fit . H0 . Adding the S. H0 ES prior
(. H0 ) to Planck gives .b = 0.48 ± 0.19 and . H0 = 70.32 ± 0.85 km/s/Mpc. Adding
+0.15
the BAO and SN data results in .b = 0.40−0.19 and . H0 = 69.68 ± 0.67 km/s/Mpc, a
reduction due to the general difficulty in maintaining the agreement between CMB
and BAO while decreasing the sound horizon, as discussed in the previous subsection.
Additional constraints on baryon clumping can be derived by adding the high-
resolution CMB temperature and polarization data from the Atacama Cosmology
Telescope fourth data release (ACT DR4) [108] and the South Pole Telescope Third
Generation (SPT-3G) 2018 data [20, 109]. As discussed in Sect. 32.5 and demon-
strated in Fig. 32.2, baryon clumping impacts the Silk damping tail of the CMB
spectrum, making the spectra at multipoles . > 2000 a powerful probe of baryon

Fig. 32.5 Constraints on the

clumping factor .b and . H0
derived from Planck (lime),
the combination of Planck
with the S. H0 ES prior on . H0
(red), and the combination of
the Planck, S. H0 ES, BAO
and SN data (violet). With
the . H0 prior, there is a
preference for a nonzero
clumping at .∼3.σ level
32 Primordial Magnetic Fields and the Hubble Tension 603

Fig. 32.6 Joint constraints

in the . S8 -.Ωm plane from
Planck+BAO+SN in ∧CDM
(red), and in the M1 model
fit to Planck+BAO+SN with
the S. H0 ES-based . H0 prior
(violate). The ∧CDM based
constraints from the DES Y1
data [107] are shown in grey.
There is a slight relief of the
tension in M1 due to .Ωm h 2
remaining largely the same
as in the best-fit ∧CDM
model, while .h is increased

clumping. Combining Planck and ACT yields .b < 0.34 at 95% CL [17, 19], while
Planck+SPT gives .b < 0.38 at 95% CL [20] in the M1 model.
Many models that resolve the Hubble tension tend to make the . S8 tension worse,
mostly due to the reasons described in the previous subsection. In contrast, the baryon
clumping generally reduces the value of . S8 deduced from CMB, thus also helping
with another tension. The primary reason for the lower . S8 (and .Ωm ) values is that the
best fit.Ωm h 2 value in the clumping model is largely the same as in ∧CDM, while.h is
increased. Figure 32.6 compares the . S8 -.Ωm joint posteriors in the Planck+BAO+SN
best-fit ∧CDM model to those in the M1 ∧CDM+b , together with the DES Year 1
contours.
Based on preliminary results from MHD simulations described in the next section,
it is clear that one should not attribute too much weight to observational con-
straints derived from the three-zone model. However, for general reasons discussed
in Sect. 32.6.1, it is clear that, even in the best-case scenario, baryon clumping could
at best raise the CMB-based value of the Hubble constant to . H0 ∼70 km/s/Mpc.
This may or may not turn out to be enough but, even if the . H0 tension was not fully
relieved, a detection of clumping would be highly interesting by itself, as it would
provide (indirect) evidence of the PMF. If no clumping is detected, it would provide
the tightest constraint on the PMF strength.
604 K. Jedamzik and L. Pogosian

32.7 What to Expect from MHD Simulations

The three-zone model may capture much of the physics of baryon clumping due to a
PMF before recombination. It is not, however, sufficiently accurate for comparison to
current and future high precision CMB data as subtle details are neglected or simply
assumed. The M1 three-zone model, employed in earlier sections, was originally
introduced in [13] to derive the preliminary stringent limit on the PMF strength from
the (older) CMB data at that time. It assumed a particular baryon PDF and neglected
its evolution. Furthermore, at the time of publication of [13], not all of the effects
that impact the ionization history in the presence of PMFs were known. Since then,
it became apparent that a complete exhaustive analysis, employing numerical MHD
simulations and Monte Carlo methods for radiation transfer is required, which is
currently underway.
Figure 32.7 shows results of a recent numerical MHD simulation [98]. The left
panel shows the projected magnetic energy density and the right panel shows the
projected baryon overdensity .ρ/⟨ρ⟩ from a .2563 simulation of a .(24 kpc)3 volume
permeated by a stochastic non-helical magnetic field at redshift .z = 1000. The initial
conditions were a homogeneous baryon density, vanishing peculiar velocities, and
a field of root-mean-square amplitude . Brms = 0.53 nG (corresponding to .c A,rms =
12cs ) at .z = 4500. One can see that substantial baryon density fluctuations .δρ/ρ ∼ 1
on .∼kpc scales are generated by redshift .z = 1000. Most of the volume is occupied
by underdense regions, whereas substantial overdensities exist in small pockets. A
small volume fraction of .∼ 10−3 has overdensities above .10, that are not visible
in the figure as it shows the projected density. In comparison, the magnetic field
energy density is much more diffuse. The maximum clumping factor .b ≈ 1.5 is
attained around .z ≈ 1000. The corresponding baryon PDF is very skew-positive, i.e.
has a substantial positive third moment, and is significantly different than the PDF

Fig. 32.7 Projected magnetic energy density (left panel) and baryon overdensity (right panel) in
a MHD simulation of a stochastic non-helical magnetic field with a Batchelor spectrum at redshift
.z = 1000. The initial conditions, set at .z = 4500, were uniform baryon density, vanishing peculiar
motions and a magnetic field with a root-mean-square amplitude of .0.53 nG, or .c A,rms = 12cs
32 Primordial Magnetic Fields and the Hubble Tension 605

assumed in M1. At redshift.z = 1000, the magnetic field has dissipated to. Brms = 0.2
nG, whereas the “final” field is . Brms = 0.044 nG at .z ≈ 10.
The study in [98] established several new insights:

• Even knowing the full evolution of the baryon PDF is not sufficient for an accurate
computation of the ionization history. Rather, the density evolution of each fluid
element needs to be known.
• The clumping factor receives large contributions from rare very high density
regions. These, however, do not contribute much to the bulk of the recombina-
tion. Therefore, the clumping factor .b is not a useful parameter for gauging the
effect on the ionization history.
• A fraction of Lyman-.α photons emitted during the recombination process locally
may actually travel to other regions with different local conditions. The resultant
cosmic average ionization is significantly changed due to this Lyman-.√α transport.
• The drop in the speed of sound during recombination by a factor of .1/ 2, due the
diminishing electron pressure, leads to further baryon compression by the PMF
during recombination.

The last two points are illustrated in Fig. 32.8, showing that all known effects have
to be taken into account for a precision computation of the ionization history.
Ultimately, results of such numerical simulations should be used for precision
tests against current and future CMB data. A meaningful comparison faces diffi-
culties due to the required immense CPU time. In theory, given the PMF spectrum
and helicity, for each magnetic field strength and cosmological parameters such as

0
Relative average ionization fraction

-0.05

-0.1

-0.15

-0.2

-0.25
600 800 1000 1200 1400 1600
redshift

Fig. 32.8 Relative change in the average ionization fraction.∆X e compared to ∧CDM for numerical
MHD simulations with a all known effects taken into account, b neglecting the Lyman-.α transport
and, c neglecting the reduction of the speed of sound during recombination, as labeled. A relatively
strong magnetic field with scale-invariant spectrum was assumed. The prediction of the M1 model
with .b = 0.5 is shown for comparison d
606 K. Jedamzik and L. Pogosian

baryon density, dark matter density, etc., an independent large simulation should be
performed. For the large number of cosmological parameters sampled in such tests,
it would be impossible. Fortunately, the effects on . X e from changing the cosmo-
logical parameters is much smaller than that of the PMF parameters of appreciable
amplitude, such that they may be added linearly to a good approximation. Given
a particular set of cosmological parameters, numerical simulations still have to be
performed for different initial PMF strengths. This can be done for a discrete set of
strengths with subsequent interpolation. The numerical simulation have to be of a
significant box size to cover a sufficient dynamical range to simulate all the scales
which produce density fluctuations and dissipate afterwards between helium- and
hydrogen- recombination, and for the particular realization of the stochastic field
(i.e. the choice of the random number) to result in a small variance in the results.
This points to fairly large simulations and requires extensive amounts of CPU time.

32.8 Summary

PMFs are often categorized as “exotic physics” because, on one hand, there is no firm
theoretical prediction of their existence within standard models of particle physics
and cosmology, and also because the observational evidence supporting the PMF
hypothesis, while increasingly compelling, is still indirect and can, at least in prin-
ciple, be explained via less exotic astrophysical processes. However, PMFs are a
rather mature field of research sustained for over half-a-century by a small but capa-
ble community of researchers. The reason for this longevity is not (only) due to
PMFs often playing the role of a “ghost fairy” (using the term much liked by a cer-
tain famous cosmologist) coming to the rescue whenever there is a new unexplained
phenomenon. Rather, most of the researchers working on the PMFs are inspired by
the fact that their existence in the early universe is inevitable. It is not a question of
whether they were generated at some level, e.g. during the known phase transitions,
but whether they would be of sufficient strength to make a detectable impact. If there
is even a remote chance of detecting them, it is worth trying, as they would provide
an invaluable insight and a new window into the physics of the early universe.
It is the above context that, in our view, distinguishes the PMF proposal for
relieving the Hubble tension from others. It is not a “fairy” explanation invented to
solve the puzzle of the day, but rather an observation that if the primordial universe
was magnetized at a level sufficient to be relevant to the magnetic fields we see
in galaxies and clusters, it could also help with alleviating the . H0 problem. This
is a fully falsifiable proposal—there is no new physics to invent when it comes to
working out the details of recombination in the presence of a PMF. The required
MHD simulations are challenging but doable, and there will be a firm prediction in
a reasonable time that can then be tested against the data.
Like other models that aim to relieve the Hubble tension by lowering the sound
horizon at recombination, baryon clumping can not raise the value of . H0 beyond
.∼70 km/s/Mpc, which is still over .2σ lower than the S. H0 ES measurement. However,
32 Primordial Magnetic Fields and the Hubble Tension 607

even if PMFs do not fully resolve the Hubble tension, finding evidence for them in
the CMB would be a major discovery in its own right. Alternatively, if not detected,
accounting for their impact on recombination when analyzing data from future CMB
experiments, such as Simons Observatory and CMB-S4, would lead to the tightest
constraints on a PMF.

Acknowledgements We thank Tom Abel, Lennart Balkenhol, Silvia Galli and Gong-Bo Zhao for
collaborations and discussions. This research was enabled in part by support provided by the BC
DRI Group and the Digital Research Alliance of Canada (alliancecan.ca), and the National Sciences
and Engineering Research Council of Canada.

References

1. A.G. Riess et al., A comprehensive measurement of the local value of the Hubble constant with
1 km s.?1 Mpc.?1 uncertainty from the Hubble space telescope and the S. H0 ES team. Astrophys.
J. Lett. 934, L7 (2022). https://doi.org/10.3847/2041-8213/ac5c5b arxiv.org/abs/2112.04510
2. Planck Collaboration, N. Aghanim et al., Planck 2018 results. VI. Cosmologi-
cal parameters. A&A 641, A6 (2020). https://doi.org/10.1051/0004-6361/201833910
arxiv.org/abs/1807.06209
3. E. Abdalla et al., Cosmology intertwined: a review of the particle physics, astrophysics, and
cosmology associated with the cosmological tensions and anomalies. JHEAp Series 34, 49
(2022)
4. D. Pesce et al., The Megamaser cosmology project. XIII. Combined Hubble con-
stant constraints. Astrophys. J. 891, L1 (2020). https://doi.org/10.3847/2041-8213/ab75f0
arxiv.org/abs/2001.09213
5. K.C. Wong et al., H0LiCOW XIII. A 2.4% measurement of . H0 from lensed quasars: .5.3σ
tension between early and late-Universe probes. arxiv.org/abs/1907.04869
6. DES collaboration, STRIDES: a 3.9 percent measurement of the Hubble constant from the
strong lens system DES J0408-5354. Mon. Not. Roy. Astron. Soc. 494, 6072 (2020). https://
doi.org/10.1093/mnras/staa828 arxiv.org/abs/1910.06306
7. D. Harvey, A 4% measurement of. H0 using the cumulative distribution of strong-lensing time
delays in doubly-imaged quasars. arxiv.org/abs/2011.09488
8. M. Millon et al., TDCOSMO. I. An exploration of systematic uncertainties in the inference
of . H0 from time-delay cosmography. Astron. Astrophys. 639, A101 (2020). https://doi.org/
10.1051/0004-6361/201937351 arxiv.org/abs/1912.08027
9. W.L. Freedman, B.F. Madore, T. Hoyt, I.S. Jang, R. Beaton, M.G. Lee et al., Calibration of
the tip of the red giant branch (TRGB). arxiv.org/abs/2002.01550
10. M.M. Ivanov, M. Simonović, M. Zaldarriaga, Cosmological parameters from the BOSS galaxy
power spectrum. JCAP Series 05, 042 (2020)
11. shape ACT collaboration, The Atacama cosmology telescope: DR4 maps and cosmological
parameters. JCAP Series 12, 047 (2020)
12. eBOSS collaboration, The completed SDSS-IV extended Baryon oscillation spectroscopic
survey: cosmological implications from two decades of spectroscopic surveys at the Apache
point observatory. arxiv.org/abs/2007.08991
13. K. Jedamzik, T. Abel, Small-scale primordial magnetic fields and anisotropies in the cosmic
microwave background radiation. JCAP Series 1310, 050 (2013)
14. K. Jedamzik, A. Saveliev, Stringent limit on primordial magnetic fields from the cosmic
microwave background radiation. Phys. Rev. Lett. Series 123, 021301 (2019)
15. P.J.E. Peebles, Principles of physical cosmology (1994)
608 K. Jedamzik and L. Pogosian

16. K. Jedamzik, L. Pogosian, Relieving the Hubble tension with primordial magnetic fields.
Phys. Rev. Lett. Series 125, 181302 (2020)
17. L. Thiele, Y. Guan, J.C. Hill, A. Kosowsky, D.N. Spergel, Can small-scale baryon
inhomogeneities resolve the Hubble tension? An investigation with ACT DR4.
arxiv.org/abs/2105.03003
18. M. Rashkovetskyi, J.B. Muñoz, D.J. Eisenstein, C. Dvorkin, Small-scale clumping at recom-
bination and the Hubble tension. arXiv e-prints (2021) arXiv:2108.02747
19. S. Galli, L. Pogosian, K. Jedamzik, L. Balkenhol, Consistency of Planck, ACT, and SPT
constraints on magnetically assisted recombination and forecasts for future experiments. Phys.
Rev. D series 105, 023513 (2022)
20. SPT-3G collaboration, A measurement of the CMB temperature power spectrum and con-
straints on cosmology from the SPT-3G 2018 TT/TE/EE data set. arxiv.org/abs/2212.05642
21. F. Hoyle, XI solvay congress. Brussells (1958)
22. Y. Zeldovich, Magnetic model of the universe. J. Exptl. Theoret. Phys. (U.S.S.R.) 48, 986
(1965)
23. L.M. Widrow, Origin of galactic and extragalactic magnetic fields. Rev. Mod. Phys. Series
74, 775 (2002)
24. L.M. Widrow, D. Ryu, D.R.G. Schleicher, K. Subramanian, C.G. Tsagas, R.A. Treumann,
The first magnetic fields. Space Sci. Rev. Series 166, 37 (2012)
25. T. Vachaspati, Progress on cosmological magnetic fields. arxiv.org/abs/2010.10525
26. A. Brandenburg, K. Subramanian, Astrophysical magnetic fields and nonlinear dynamo the-
ory. Phys. Rept. Series 417, 1 (2005)
27. R.M. Athreya, V.K. Kapahi, P.J. McCarthy, W. van Breugel, Large rotation measures in radio
galaxies at Z > 2. Astron. Astrophys. Series 329, 809 (1998)
28. A.M. Beck, K. Dolag, H. Lesch, P.P. Kronberg, Strong magnetic fields and large rotation
measures in protogalaxies by supernova seeding. Mon. Not. Roy. Astron. Soc. Series 435,
3575 (2013)
29. D. Seifried, R. Banerjee, D. Schleicher, Supernova explosions in magnetized, primordial dark
matter haloes. Mon. Not. Roy. Astron. Soc. Series 440, 24 (2014)
30. T. Vachaspati, Magnetic fields from cosmological phase transitions. Phys. Lett. Series B265,
258 (1991). https://doi.org/10.1016/0370-2693(91)90051-Q
31. M.S. Turner, L.M. Widrow, Inflation produced, large scale magnetic fields. Phys. Rev. Series
D37, 2743 (1988). https://doi.org/10.1103/PhysRevD.37.2743
32. B. Ratra, Cosmological “seed” magnetic field from inflation. Astrophys. J. Series 391, L1
(1992)
33. R. Durrer, A. Neronov, Cosmological magnetic fields: their generation, evolution and obser-
vation. Astron. Astrophys. Rev. Series 21, 62 (2013)
34. K. Subramanian, The origin, evolution and signatures of primordial magnetic fields. Rept.
Prog. Phys. Series 79, 076901 (2016)
35. K. Jedamzik, V. Katalinic, A.V. Olinto, Damping of cosmic magnetic fields. Phys. Rev. D
Series 57, 3264 (1998)
36. K. Subramanian, J.D. Barrow, Magnetohydrodynamics in the early universe and the damping
of noninear Alfven waves. Phys. Rev. Series D58, 083502 (1998)
37. A. Neronov, I. Vovk, Evidence for strong extragalactic magnetic fields from Fermi observa-
tions of TeV blazars. Sci. Series 328, 73 (2010)
38. F. Tavecchio, G. Ghisellini, L. Foschini, G. Bonnoli, G. Ghirlanda, P. Coppi, The intergalactic
magnetic field constrained by Fermi/LAT observations of the TeV blazar 1ES 0229+200.
Mon. Not. Roy. Astron. Soc. Series 406, L70 (2010)
39. A. Taylor, I. Vovk, A. Neronov, Extragalactic magnetic fields constraints from simul-
taneous GeV-TeV observations of blazars. Astron. Astrophys. 529, A144 (2011)
[arxiv.org/abs/1101.0932]
40. K. Dolag, M. Kachelriess, S. Ostapchenko, R. Tomas, Lower limit on the strength and filling
factor of extragalactic magnetic fields. Astrophys. J. Lett. Series 727, L4 (2011)
32 Primordial Magnetic Fields and the Hubble Tension 609

41. A.E. Broderick, P. Chang, C. Pfrommer, The cosmological impact of luminous TeV Blazars
I: implications of plasma instabilities for the intergalactic magnetic field and extragalactic
gamma-ray background. Astrophys. J. Series 752, 22 (2012)
42. F. Govoni, E. Orrù, A. Bonafede, M. Iacobelli, R. Paladino, F. Vazza et al., A radio ridge
connecting two galaxy clusters in a filament of the cosmic web. Sci. Series 364, 981–984
(2019)
43. CTA Consortium collaboration, B.S. Acharya et al., Science with the Cherenkov telescope
array. WSP 11 (2018). https://doi.org/10.1142/10986 [arxiv.org/abs/1709.07997]
44. A. Korochkin, O. Kalashev, A. Neronov, D. Semikoz, Sensitivity reach of gamma-ray mea-
surements for strong cosmological magnetic fields. Astrophys. J. Series 906, 116 (2021)
45. K. Subramanian, J.D. Barrow, Microwave background signals from tangled magnetic fields.
Phys. Rev. Lett. Series 81, 3575 (1998)
46. K. Subramanian, J.D. Barrow, Small-scale microwave background anisotropies due to tangled
primordial magnetic fields. Mon. Not. Roy. Astron. Soc. Series 335, L57 (2002)
47. A. Mack, T. Kahniashvili, A. Kosowsky, Microwave background signatures of a primordial
stochastic magnetic field. Phys. Rev. Series D65, 123004 (2002)
48. A. Lewis, Observable primordial vector modes. Phys. Rev. Series D70, 043518 (2004)
49. T. Kahniashvili, B. Ratra, Effects of cosmological magnetic helicity on the cosmic microwave
background. Phys. Rev. Series D71, 103006 (2005)
50. G. Chen, P. Mukherjee, T. Kahniashvili, B. Ratra, Y. Wang, Looking for cosmological
Alfven waves in WMAP data. Astrophys. J. 611, 655 (2004). https://doi.org/10.1086/422213
[arxiv.org/abs/astro-ph/0403695]
51. A. Lewis, CMB anisotropies from primordial inhomogeneous magnetic fields.
Phys. Rev. D70, 043011 (2004). https://doi.org/10.1103/PhysRevD.70.043011
[arxiv.org/abs/astro-ph/0406096]
52. H. Tashiro, N. Sugiyama, R. Banerjee, Nonlinear evolution of cosmic magnetic fields and
cosmic microwave background anisotropies. Phys. Rev. Series D73, 023002 (2006)
53. D. Yamazaki, K. Ichiki, T. Kajino, G.J. Mathews, Constraints on the evolution of the primordial
magnetic field from the small scale CMB angular anisotropy. Astrophys. J. Series 646, 719
(2006)
54. M. Giovannini, Entropy perturbations and large-scale magnetic fields. Class. Quant. Grav.
Series 23, 4991 (2006)
55. T. Kahniashvili, B. Ratra, CMB anisotropies due to cosmological magnetosonic waves. Phys.
Rev. Series D75, 023002 (2007)
56. M. Giovannini, K.E. Kunze, Magnetized CMB observables: a dedicated numerical approach.
Phys. Rev. Series D77, 063003 (2008)
57. D.G. Yamazaki, K. Ichiki, T. Kajino, G.J. Mathews, New constraints on the primordial mag-
netic field. Phys. Rev. Series D81, 023008 (2010)
58. D. Paoletti, F. Finelli, CMB constraints on a stochastic background of primordial magnetic
fields. Phys. Rev. Series D83, 123533 (2011)
59. J.R. Shaw, A. Lewis, Constraining primordial magnetism. Phys. Rev. Series D86, 043510
(2012)
60. K.E. Kunze, CMB anisotropies in the presence of a stochastic magnetic field. Phys. Rev.
Series D83, 023006 (2011)
61. D. Paoletti, F. Finelli, Constraints on a stochastic background of primordial magnetic fields
with WMAP and south pole telescope data. Phys. Lett. Series B726, 45 (2013)
62. M. Ballardini, F. Finelli, D. Paoletti, CMB anisotropies generated by a stochastic background
of primordial magnetic fields with non-zero helicity. JCAP Series 1510, 031 (2015)
63. Planck collaboration, Planck 2015 results. XIX. Constraints on primordial magnetic
fields. Astron. Astrophys. 594, A19 (2016). https://doi.org/10.1051/0004-6361/201525821
[arxiv.org/abs/1502.01594]
64. D.R. Sutton, C. Feng, C.L. Reichardt, Current and future constraints on primordial magnetic
fields. Astrophys. J. Series 846, 164 (2017)
610 K. Jedamzik and L. Pogosian

65. A. Zucca, Y. Li, L. Pogosian, Constraints on primordial magnetic fields from Planck combined
with the south pole telescope CMB B-mode polarization measurements. Phys. Rev. Series
D95, 063506 (2017)
66. K. Jedamzik, V. Katalinic, A.V. Olinto, A Limit on primordial small scale magnetic fields
from CMB distortions. Phys. Rev. Lett. Series 85, 700 (2000)
67. A. Zizzo, C. Burigana, On the effect of cyclotron emission on the spectral distortions of the
cosmic microwave background. New Astron. Series 11, 1 (2005)
68. K.E. Kunze, E. Komatsu, Constraining primordial magnetic fields with distortions of the black-
body spectrum of the cosmic microwave background: pre- and post-decoupling contributions.
JCAP Series 1401, 009 (2014)
69. J. Ganc, M.S. Sloth, Probing correlations of early magnetic fields using mu-distortion. JCAP
Series 1408, 018 (2014)
70. D. Paoletti, J. Chluba, F. Finelli, J.A. Rubino-Martin, Improved CMB anisotropy constraints
on primordial magnetic fields from the post-recombination ionization history. Mon. Not. Roy.
Astron. Soc. Series 484, 185 (2019)
71. S.K. Sethi, K. Subramanian, Primordial magnetic fields in the post-recombination era and
early reionization. Mon. Not. Roy. Astron. Soc. Series 356, 778 (2005)
72. K.E. Kunze, E. Komatsu, Constraints on primordial magnetic fields from the optical depth of
the cosmic microwave background. JCAP 1506, 027 (2015). https://doi.org/10.1088/1475-
7516/2015/06/027 [arxiv.org/abs/1501.00142]
73. J. Chluba, D. Paoletti, F. Finelli, J.-A. Rubiño-Martín, Effect of primordial magnetic fields
on the ionization history. Mon. Not. Roy. Astron. Soc. Series 451, 2244 (2015)
74. R. Durrer, P.G. Ferreira, T. Kahniashvili, Tensor microwave anisotropies from a stochastic
magnetic field. Phys. Rev. Series D61, 043001 (2000)
75. T.R. Seshadri, K. Subramanian, CMBR polarization signals from tangled magnetic fields.
Phys. Rev. Lett. Series 87, 101301 (2001)
76. K. Subramanian, T.R. Seshadri, J. Barrow, Small-scale CMB polarization anisotropies due to
tangled primordial magnetic fields. Mon. Not. Roy. Astron. Soc. Series 344, L31 (2003)
77. S. Mollerach, D. Harari, S. Matarrese, CMB polarization from secondary vector and tensor
modes. Phys. Rev. Series D69, 063002 (2004)
78. C. Scoccola, D. Harari, S. Mollerach, B polarization of the CMB from Faraday rotation. Phys.
Rev. Series D70, 063003 (2004)
79. A. Kosowsky, T. Kahniashvili, G. Lavrelashvili, B. Ratra, Faraday rotation of the cosmic
microwave background polarization by a stochastic magnetic field. Phys. Rev. Series D71,
043006 (2005)
80. L. Pogosian, T. Vachaspati, A. Yadav, Primordial magnetism in CMB B-modes. Can. J. Phys.
Series 91, 451 (2013)
81. T. Kahniashvili, Y. Maravin, G. Lavrelashvili, A. Kosowsky, Primordial magnetic helicity
constraints from WMAP nine-year data. Phys. Rev. Series D90, 083004 (2014)
82. L. Pogosian, A. Zucca, Searching for primordial magnetic fields with CMB B-modes. Class.
Quant. Grav. Series 35, 124004 (2018)
83. I. Brown, R. Crittenden, Non-Gaussianity from cosmic magnetic fields. Phys. Rev. Series
D72, 063002 (2005)
84. T. Seshadri, K. Subramanian, CMB bispectrum from primordial magnetic fields on large
angular scales. Phys. Rev. Lett. Series 103, 081303 (2009)
85. C. Caprini, F. Finelli, D. Paoletti, A. Riotto, The cosmic microwave background temperature
bispectrum from scalar perturbations induced by primordial magnetic fields. JCAP Series
0906, 021 (2009)
86. R.-G. Cai, B. Hu, H.-B. Zhang, Acoustic signatures in the cosmic microwave background
bispectrum from primordial magnetic fields. JCAP Series 1008, 025 (2010)
87. P. Trivedi, K. Subramanian, T.R. Seshadri, Primordial magnetic field limits from cosmic
microwave background bispectrum of magnetic passive scalar modes. Phys. Rev. Series D82,
123006 (2010)
88. I.A. Brown, Intrinsic bispectra of cosmic magnetic fields. Astrophys. J. Series 733, 83 (2011)
32 Primordial Magnetic Fields and the Hubble Tension 611

89. M. Shiraishi, D. Nitta, S. Yokoyama, K. Ichiki, K. Takahashi, Cosmic microwave background

bispectrum of vector modes induced from primordial magnetic fields. Phys. Rev. Series D82,
121302 (2010)
90. M. Shiraishi, D. Nitta, S. Yokoyama, K. Ichiki, K. Takahashi, Cosmic microwave background
bispectrum of tensor passive modes induced from primordial magnetic fields. Phys. Rev.
Series D83, 123003 (2011)
91. P. Trivedi, T.R. Seshadri, K. Subramanian, Cosmic microwave background trispectrum and
primordial magnetic field limits. Phys. Rev. Lett. Series 108, 231301 (2012)
92. M. Shiraishi, T. Sekiguchi, First observational constraints on tensor non-Gaussianity sourced
by primordial magnetic fields from cosmic microwave background. Phys. Rev. Series D90,
103002 (2014)
93. P. Trivedi, K. Subramanian, T.R. Seshadri, Primordial magnetic field limits from the CMB
trispectrum: scalar modes and Planck constraints. Phys. Rev. Series D89, 043523 (2014)
94. R. Durrer, C. Caprini, Primordial magnetic fields and causality. JCAP Series 11, 010 (2003)
95. G.K. Batchelor, Small-scale variation of convected quantities like temperature in turbulent
fluid. Part 1. General discussion and the case of small conductivity. J. Fluid Mech. 5, 113
(1959). https://doi.org/10.1017/S002211205900009X
96. K. Jedamzik, G. Sigl, The evolution of the large-scale tail of primordial magnetic fields. Phys.
Rev. Series D83, 103005 (2011)
97. R. Banerjee, K. Jedamzik, The evolution of cosmic magnetic fields: from the very early
universe, to recombination, to the present. Phys. Rev. D Series 70, 123003 (2004)
98. K. Jedamzik, T. Abel, Cosmological recombination in the presence of primordial magnetic
fields (in preparation) (2023)
99. K. Dolag, M. Bartelmann, H. LeschAstron, Astrophys. Series 348, 351 (1999)
100. K. Dolag, M. Bartelmann, H. LeschAstron, Astrophys. Series 387, 383 (2002)
101. J.L.W.V. Jensen, Sur les fonctions convexes et les inégalités entre les valeurs moyennes. Acta
Mathematica Series 30, 175 (1906). https://doi.org/10.1007/BF02418571
102. M. Zaldarriaga, D.D. Harari, Analytic approach to the polarization of the cosmic microwave
background in flat and open universes. Phys. Rev. D Series 52, 3276 (1995)
103. K. Jedamzik, L. Pogosian, G.-B. Zhao, Why reducing the cosmic sound horizon can not fully
resolve the Hubble tension. arxiv.org/abs/2010.04158
104. L. Pogosian, G.-B. Zhao, K. Jedamzik, Recombination-independent determination of the
sound horizon and the Hubble constant from BAO. Astrophys. J. Lett. Series 904, L17 (2020)
105. Kilo-Degree Survey, DES collaboration, DES Y3 + KiDS-1000: consistent cosmology com-
bining cosmic shear surveys. arxiv.org/abs/2305.17173
106. D. Scolnic et al., The complete light-curve sample of spectroscopically confirmed SNe Ia from
Pan-STARRS1 and cosmological constraints from the combined Pantheon sample. Astrophys.
J. Series 859, 101 (2018)
107. shape DES collaboration, Dark energy survey year 1 results: cosmological constraints from
galaxy clustering and weak lensing. Phys. Rev. Series D98, 043526 (2018)
108. shape ACT collaboration, The Atacama cosmology telescope: a measurement of the cosmic
microwave background power spectra at 98 and 150 GHz. JCAP Series 12, 045 (2020)
109. SPT-3G collaboration, Measurements of the E-mode polarization and temperature-E-mode
correlation of the CMB from SPT-3G 2018 data. arxiv.org/abs/2101.01684
Chapter 33
Varying Fundamental Constants Meet
Hubble

Jens Chluba and Luke Hart

33.1 Why Varying Fundamental Constants?

The laws of nature depend on fundamental constants (FCs) such as Newton’s con-
stant, .G, the speed of light, .c, and elementary charge of an electron, .e, to name just a
few. The values of these constants have been determined experimentally, but gener-
ally should emerge directly from the underlying theory. As such, there is no reason to
assume that the values of the FCs determined locally simply translate to other parts
of the cosmos or to other eras in cosmic history (see, [59, 91, 92] for broad review).
Studies of fundamental constants and their possible temporal and spatial variations
are thus of utmost importance, and could provide a glimpse at physics beyond the
standard model, possibly shedding light on the presence of additional scalar fields
and their couplings to the standard sector.
Of the many fundamental constants, the fine-structure constant, .αEM , and electron
rest mass, .m e , are the most interesting to CMB studies (e.g., [6–8, 47, 60, 62, 72,
80]). This is because these constants play crucial roles in the way photons and baryons
interact. Most importantly, they affect important atomic transition rates, which in turn
control the cosmological recombination process and hence the Thomson visibility
function (defining the last scattering surface) that is so crucial to the formation of
the CMB temperature and polarization anisotropies [42, 43, 57, 65, 89].
Using Planck 2013 data, the values of .αEM and .m e around recombination were
proven to coincide with those obtained in the lab to within .≅ 0.4% for .αEM and
.≅ 1% − 6% for .m e [67]. These limits are .≅ 2 − 3 orders of magnitude weaker than

constraints obtained from other ‘local’ measurements [11, 12, 53, 73]; however,
the CMB places limits during very different phases in the history of the Universe,

J. Chluba (B) · L. Hart (B)

Jodrell Bank Centre for Astrophysics, Alan Turing Building, University of Manchester,
Manchester M13 9PL, UK
e-mail: jens.chluba@manchester.ac.uk
L. Hart
e-mail: luke.hart@tneigroup.com
L. Hart
TNEI Services Ltd., Bainbridge House, 88 London Road, Manchester M1 2PW, UK

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 613
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_33
614 J. Chluba and L. Hart

centered around the time of last scattering some .380, 000 years after the Big Bang,
thereby complementing these low-redshift measurements. In addition, CMB mea-
surements can be used to probe spatial variations of the FCs at cosmological distances
[86], opening yet another avenue for exploration.
With the Planck 2013 results in mind, no significant surprises were expected
from the analysis of improved CMB data of the Planck 2015 and 2018 releases.
However, it turns out that when considering models with varying .m e , the geometric
degeneracy becomes significant and can accommodate shifts in the value of the
Hubble parameter when multiple probes are combined [36]. The same geometric
freedom is not encountered when varying .αEM due to the modified dependence of
the visibility function on this parameter. This finding has spurred an increased interest
in studying VFCs in this context, with scenarios that allow for varying .m e [e.g., [41,
54, 83, 90] for additional extensions] ranking high in model comparisons [79]. Since
VFCs can be caused by the presence of scalar fields (e.g., [9, 61]), a natural question
is whether the same scalar fields could also be causing effects relating to EDE (see
Chap. 22), indicating a ’two sides of the same coin’ interplay. It will therefore be
extremely important to ask how different measurements can be combined to shed
light on the physical origin of the Hubble tension.
In this chapter we explain how .αEM and .m e enter the cosmological recombination
problem and calculation of the CMB power spectra. We then briefly recap some of
the constraints, highlighting important findings, before moving on to a discussion
of their role in the Hubble tension. Modifications to the recombination process also
affect the cosmological recombination radiation (CRR) [15, 88], implying that direct
insight into the underlying physics could be gained by future measurements of CMB
spectral distortions (e.g., [14]). We highlight how this avenue may even allow dis-
tinguishing models of EDE, PMFs and VFCs and possibly identify modifications to
the recombination history as the main cause of the tension [38, 58].

33.2 Effects of VFCs on the Recombination Process

The ionization history of the Universe is one of the crucial theoretical ingredients
in the computations of the CMB temperature and polarization anisotropies [42, 65,
89]. The first computations of this transition from the fully-ionized plasma to a
neutral medium were carried out in the late 60 s, recognizing the important role of
Lyman-.α transport and the 2 s-1 s two-photon decay of Hydrogen [64, 94]. These
early calculations reached a precision of .≅20–30% in the free electron fraction,
. X e = Ne /NH , around the maximum of the Thomson scattering visibility function at
redshift .z ≅ 1100. Here . Ne is the free electron number density and . NH denotes the
total number density of hydrogen nuclei in the Universe. However, with the advent
of precision CMB data from WMAP and Planck it became important to improve the
modeling of the cosmological recombination process [18, 42, 57]. Initially, this led to
the development of Recfast [81], which reached .≅1–3% precision for Hydrogen
and neutral Helium recombination. However, for the analysis of data from Planck,
33 Varying Fundamental Constants Meet Hubble 615

this precision was still insufficient, and many detailed atomic physics and radiative
transfer effects had to be accounted for (see, [31, 33, 74, 88, 93] for overview),
leading to changes at the level of several standard deviations in particular for the
value of the spectral index of scalar perturbations, .n S [74, 84]. This necessitated the
development of the highly-flexible and accurate recombination codes CosmoRec
[22] and HyRec [5], which ensured that for the analysis of Planck none of the
standard parameters were biased at a significant level [66, 68].
This short recap highlights the crucial role of the recombination history in the
computations of the CMB anisotropies using standard Boltzmann solvers such as
CAMB [56] and CLASS [55], and conversely, the impressive precision and sensitivity
of the current measurements to subtle modifications in the ionization history. The
statements assume standard physics during the recombination era at .z ≅ 102 − 104 .
In particular, it is assumed that the atomic transition rates for Hydrogen and Helium
are the same as those inferred in the lab.
From standard textbook atomic physics, it is well-understood how .αEM and .m e
affect the energy levels and transition probabilities of Hydrogen and Helium (e.g., [10,
25]). It is immediately clear that varying .αEM and .m e inevitably create changes to the
cosmological ionization history and hence CMB observables. Most importantly, the
energy levels of hydrogen and helium depend on these constants as . E i ∝ αEM 2
me,
which directly affects the recombination redshift. In addition, the atomic bound-
bound transition rates and photoionization/recombination rates are altered when
varying .αEM and .m e . Lastly, the interactions of photons and electrons through Comp-
ton and resonance scattering modify the radiative transfer physics, which control the
dynamics of recombination (e.g., [4, 16, 19, 21, 40, 50, 75]).
In an effective three-level atom approach [64, 82, 94], the individual dependencies
can be summarized as (e.g., [34, 47, 67, 80])

σT ∝ αEM
2
m −2
e A2γ ∝ αEM
8
m e PS A1γ ∝ αEM
6
m 3e
. −2 −2 −1 (33.1)
αrec ∝ αEM m e βphot ∝ αEM m e Teff ∝ αEM m e .
2 5

Here, .σT denotes the Thomson scattering cross section; . A2γ is the two-photon decay
rate of the second shell; .αrec and .βphot are the effective recombination and photoion-
ization rates, respectively; .Teff is the effective temperature at which .αrec and .βphot
need to be evaluated (see explanation below); . PS A1γ denotes the effective dipole
transition rate for the main resonances (e.g., Lyman-.α), which is reduced by the
Sobolev escape probability, . PS ≤ 1 [82, 87] with respect to the vacuum rate, . A1γ .
For a more detailed account of how the transition rates depend on the fundamental
constants we refer to CosmoSpec [15] and the manual of HyRec [5].
The scalings of .σT , . A2γ and . PS A1γ directly follow from their explicit dependence
on .αEM and .m e . For .αrec and .βphot , only the renormalisations of the transition rates is
reflected, again stemming from their explicit dependencies on.αEM and.m e (e.g., [48]).
However, these rates also depend on the ratio of the electron and photon temperatures
to the ionization threshold. This leads to an additional dependence on .αEM and .m e ,
which can be captured by evaluating these rates at a rescaled temperature, with a
scaling indicated through .Teff . Overall, this leads to the effective dependence .αrec ∝
616 J. Chluba and L. Hart

3.44 −1.28
αEM me around hydrogen recombination [15]. The required photoionization rate,
.βphot , is obtained using the detailed balance relation. Slightly different overall scalings
for .αrec and .βphot were used in [67], but the associated effect on the recombination
history were found to be sub-dominant and limited to .z ≤ 800 [34].
For neutral helium, non-hydrogenic effects (e.g., fine-structure transitions, singlet-
triplet couplings) become relevant [25, 26]. However, the corrections should be sub-
dominant and are usually neglected. A detailed discussion of changes to the escape
probabilities during helium recombination can be found in [38]. In [34], the changes
to helium and hydrogen recombination were furthermore treated separately.

33.2.1 Ionization History Modifications Due to Variation

of .αEM

Given the above ingredients, one can now answer the question about how various
effects propagate to the ionization history. A detailed study that also directly demon-
strated the validity of simpler three-level approximations against CosmoRec was
carried out in [34]. The overall effect of varying .αEM is illustrated in Fig. 33.1,
assuming a constant variation parametrized as .αEM = αEM, 0 (1 + Δα/α), where
.αEM, 0 = 1/137 denotes the standard value. Increasing the fine structure constant
shifts the moment of recombination toward higher redshifts. This agrees with the
results found earlier in [8, 47, 72] and can intuitively be understood in the following
manner: .Δα/α > 0 increases the transition energies between different atomic lev-
els and the continuum. This increases the energy threshold at which recombination
occurs, hence increasing the recombination redshift, an effect that is basically cap-
tured by an effective temperature rescaling in the evaluation of the photoionization
and recombination rates (see below).
The relative changes to the ionization history, .ΔX e / X e , for the different quan-
tities in Eq. (33.1) are illustrated in Fig. 33.2. We chose a value for .Δα/α = 10−3 ,
which leads to a percent-level effect on . X e . As expected, the biggest effect appears
after rescaling the temperature for the evaluation of the photonionization and recom-
bination rates. More explicitly, this can be understood when considering the net
recombination rate to the second shell, which can be written as1

ΔRcon = Ne Np αrec − N2 βphot = αrec [Ne Np − g(Tγ )N2 ],

.

where .g(Tγ ) ∝ Tγ 3/2 e−hν2c /kTγ with continuum threshold energy, . E 2c = hν2c . Here,
the exponential factor (.↔ Boltzmann factor) is most important, leading to an expo-
nential effect once we replace .Tγ ' (z) = Tγ (z) × (αEM /αEM, 0 )−2 (m e /m e, 0 )−1 .
The second largest individual effect is due to the rescaling of the two-photon decay
rate, . A2γ . This is expected since .αEM appears in a high power, . A2γ ∝ αEM
8
, and also
because the 2 s-1 s two-photon channel plays such a crucial role for the recombination

1 In full equilibrium, .ΔRcon = 0.

33 Varying Fundamental Constants Meet Hubble 617

Fig. 33.1 Ionization histories for different values of .αEM . The dominant effect is caused by modi-
fications of the ionization threshold, which implies that for increased .αEM recombination finishes
earlier. The curves were computed using the Recfast++ module of CosmoRec [22]. The figure
was taken from [34]

Fig. 33.2 The relative changes in the ionization history for .Δα/α = 10−3 with respect to the
standard case caused by different effects. Recfast++ was used for the computations. The rescaling
of temperature (.↔ mainly affecting the Boltzmann factors) yields .ΔX e / X e ≅ −2.7%, dominating
the total contributions, which peaks with .≅ −3.1% at .z ≅ 1000. Note that the modification due to
.σT has been scaled by a factor of ten to make it visible. The figure was taken from [34]
618 J. Chluba and L. Hart

dynamics [18, 64, 94], allowing .≅ 58% of all hydrogen atoms to become neutral
through this route [17].
The normalizations of the recombination and photoionization rates (blue/dashed
line in Fig. 33.2) give rise to a net delay of .ΔX e / X e ≅ 0.3% at .z ≅ 1000, which
partially cancels the acceleration due to. A2γ . This is due to the stronger scaling of.βphot
with .αEM than .αrec . At low redshifts (.z ≤ 750), recombination is again accelerated,
indicating that a higher fraction of recombination events occurs, as the importance of
photoionization ceases. The correction related to the Lyman-.α channel is found to be
.≅ 3.3 times smaller than for the two-photon channel, yielding .ΔX e / X e ≅ −0.15%
at .z ≅ 1000. Finally, the effect of rescaling .σT are very small and only becomes
noticeable at low redshifts. At these redshifts, the matter and radiation temperature
begins to depart, giving .Te < Tγ . For larger .αEM , this departure is delayed, such that
. Te stays longer close to . Tγ . Hotter electrons recombine less efficiently, so that a slight

delay of recombination appears (cf., Fig. 33.2).

33.2.2 Ionization History Modifications Due to Variation

of .me

We now briefly illustrate the changes caused by the effective electron mass, with
the variation parametrized as .m e = m e, 0 (1 + Δm e /m e ) with respect to the standard
value .m e, 0 . Inspecting the scalings of Eq. (33.1), one expects the overall effect to
be smaller than for .αEM , but otherwise very comparable. For example, the effect of
temperature rescaling should be roughly half as large. Similarly, the effect due to
rescaling. A2γ should be roughly.8 times smaller, and so on. This is in good agreement
with our findings (cf. Fig. 33.3).
The net effect on . X e is about 2.5 times smaller than for .αEM around .z ≅ 1000.
This suggests that the CMB constraint on .m e is weakened by a similar factor. How-
ever, adding the rescaling of the Thomson cross section for the computation of the
visibility function strongly enhances the geometric degeneracy for .m e , such that the
CMB only constraint on .m e is .≥20 times weaker than for .αEM (see Sect. 33.3). In
addition, a small difference related to the renormalizations of the photoionization and
recombination rates (blue/dashed line) appears. For .Δm e /m e > 0, the photoioniza-
tion rate is increased and the recombination rate is reduced for these contributions [cf.
Eq. (33.1)]. Both effects delay recombination (see Fig. 33.3). Thus, around .z ≅ 103
the net effect is slightly larger than for .αEM . In contrast to .αEM , at late times no
net acceleration of recombination occurs. These effects slightly modify the overall
redshift dependence of the total . X e change, in addition to lowering the effect in the
freeze-out tail. At the level .Δm e /m e ≅ 1%, additional higher order terms become
important, allowing one to break the near degeneracy between .αEM and .m e in joint
analyses (see also [67]). At the level of the current CMB only constraints, this aspect
indeed is of relevance (Sect. 33.3.2).
33 Varying Fundamental Constants Meet Hubble 619

Fig. 33.3 Same as in Fig. 33.2 but for .Δm e /m e = 10−3 . The effective temperature rescaling again
dominates the total change. Around.z ≅ 1000, the total effect is.≅ 2.5 times smaller than for.Δα/α =
10−3 mainly due to the weaker dependence of the level energy on .m e . The figure was taken from
[34]

33.2.3 Comparison with CosmoRec and Generalized VFC

Models

We highlight that [34] also explicitly demonstrated that a full treatment of the problem
using the advanced recombination code CosmoRec yields results that are very sim-
ilar to those from a simpler three-level treatment with correction function approach
(e.g., [85] fordetails) to mimic the recombination physics corrections. This also
allowed [34] to perform calculations with explicit time-dependence of .αEM and .m e ,
initially focusing on a simple power-law redshift dependence
( )p
1+z
α
. EM (z) = αEM (z 0 ) , (33.2)
1100

for .αEM and similarly for .m e . Using a principal component analysis (PCA) [37],
which extended the previous recombination perturbation framework developed in
[29, 30, 35], general VFC variations around the recombination era were further
studied. This showed that time-dependence can in fact be independently constrained
already with existing data and also led to various generalized limits on VFCs during
recombination (e.g., [54, 90]), which we will briefly highlight below.
620 J. Chluba and L. Hart

33.3 CMB Anisotropy Constraints

33.3.1 Propagating the Effects to the CMB Anisotropies

The temperature and polarization power spectra of the CMB depend on the dynamics
of recombination through the ionization history, which defines the Thomson visibility
function and last scattering surface (e.g., [44, 65, 89]). Therefore, variations of .αEM
and .m e can leave a direct imprint on the CMB power spectra. A detailed description
of changes to the visibility function and various additional illustrations also for time-
dependent VFCs can be found in [34, 37]. Here, we highlight the changes to the
CMB temperature power spectra for variations of .αEM and .m e , noting that those in
polarization show very similar features. The power spectra were computed using
CAMB [56] for the standard cosmology [66].
To propagate the effect of VFCs to the CMB anisotropies, two changes are
required. The standard recombination history has to be replaced as explained in
the previous section. In addition, the Thomson scattering rate has be to updated in
the Boltzmann code using the modified Thomson cross section. For the latter, two
approaches are possible, one based on directly modifying the Boltzmann code, the
other on mimicking the effect by rescaling the ionization history. Both seem to deliver
consistent final results [34]. However, we stress how important it is to include the
changes to the Thomson scattering rate, as without this effect .αEM and .m e variations
essentially exhibit a very similar phenomenology [34, 36].
In Fig. 33.4, we illustrate the effect of .αEM on the CMB temperature power spec-
trum. Two main features are visible: Firstly, the peaks of the power spectrum are

Fig. 33.4 The CMB temperature power spectra for different values of .αEM . This shows that as the
fine structure constant increases, the anisotropies shift toward smaller scales and higher amplitudes.
The figure was taken from [34]
33 Varying Fundamental Constants Meet Hubble 621

Fig. 33.5 Comparison of the CMB .T T power spectrum deviations when varying .αEM , .m e , .T0 and
We chose .Δα/α = 10−3 , .Δm e /m e = 2 × 10−3 , .ΔT /T = −10−2 and . p = 5 × 10−3 (simul-
. p.
taneously for .αEM and .m e ) to obtain effects at a similar level. Notice the small extra tilt when
comparing the case for .αEM with .m e , which helps when constraining .αEM . The figure was taken
from [34]

shifted to smaller scales (larger .l) when .Δα/α > 0. This happens because ear-
lier recombination moves the last scattering surface towards higher redshifts, which
decreases the sound horizon and increases the angular diameter distance to recom-
bination [8, 47]. Secondly, for .Δα/α > 0, the peak amplitudes are enhanced. This
is mainly because earlier recombination suppresses the effect of photon diffusion
damping on the anisotropies [8, 47]. For variations of .m e , very similar responses are
found, but with an amplitude that is reduced by a factor of .≅2–3 [34].
For small .Δα/α and .Δm e /m e , we illustrate the relative change of the tem-
perature power spectrum in Fig. 33.5. The effect on the peak positions is more
noticeable than the small overall tilt caused by changes related to diffusion damp-
ing. As expected, the changes to the CMB .T T power spectra, .ΔCl /Cl (Δα/α)
and .ΔCl /Cl (Δm e /m e ), become almost indistinguishable when using .Δm e /m e ≈
(2 − 3) Δα/α. This presents a quasi-degeneracy between the two parameters and also
suggests that naively the analysis for .Δα/α could be sufficient to estimate the errors
for a corresponding analysis of .m e . However, when constraining .m e , an enhanced
geometric degeneracy, because of the differing effect of .σT , inflates the error to the
percent level [34, 67]. In this case, higher order terms become important and the near
degeneracy is broken. When also adding information from BAO, the error on .m e is
strongly reduced, and a simple scaling of the errors, .σ (Δm e /m e ) ≅ 3σ (Δα/α), is
recovered [34].
622 J. Chluba and L. Hart

In Fig. 33.5, we also illustrate the effect of varying the average CMB temperature,
T , and a time-dependent model for both .αEM and .m e with a phenomenological
. 0
power-law as in Eq. (33.2) around a pivot redshift of .z = 1100. Changes in the CMB
monopole temperature show a different response pattern than VFCs, making these
two effects principally distinguishable. This is because varying the CMB temperature
affects the ionization history (at leading order like VFCs), but without changing the
Thomson scattering rate. In addition, .T0 modifies the matter-radiation equality and
therefore has a separate overall effect. Similarly, we conclude that the power-law
index of the time-dependent model can be independently constrained [34].

33.3.2 Constraints from Planck

Now that we have developed a detailed understanding about how the CMB
anisotropies are affected by changes of .αEM and .m e , we can directly consider some
of the existing constraints from Planck. Early constraints were derived using Planck
2013 data in [67]. Aside from additional data (e.g., Planck polarization) and improve-
ments in the understanding of systematics and calibration, the later analysis of the
Planck 2015 data yielded similar constraints [34], although with slightly improved
errors.2 Here we highlight the latest Planck 2018 results, which were obtained in
[36]. For details we refer the interested reader to that paper.
In Table 33.1, we summarize some of the Planck 2018 constraints on .αEM and .m e ,
including the 2018 baseline Planck data, with low-.l and high-.l data for temperature
and . E-mode polarisation power spectra, along with the lensing data from the same
release [69, 70]. The addition of .αEM marginally affects the values and errors of
the six standard parameters, yielding .αEM /αEM,0 = 1.0005 ± 0.0024. In contrast,
varying .m e has a strong effect especially on . H0 , shifting it to extremely low values
and allowing for a low value of the electron rest mass, .m e /m e, 0 = 0.888 ± 0.059.
This is caused by the small differences in the way .m e affects the CMB power spectra,
with a crucial role played by adding changes to .σT [34, 36]. As we discuss below, this
large geometric degeneracy is one of the key ingredients for alleviating the Hubble
tension when combined with supernova data.
In Fig. 33.6, we show the two-dimensional posteriors for the constraint on simul-
taneously varying values of .αEM and .m e . For comparison, the contours relating to
the 2015 analysis and also combinations with baryon acoustic oscillation (BAO, [3])
data are illustrated. The constraints for Planck alone exhibit very wide posteriors,
with slight differences in the centroids of the 2015 and 2018 constraints. We note
that a narrower prior of . H0 > 40 km s−1 Mpc−1 (conforming with the initial CMB
analysis of Planck [67]) affects the posterior for the 2015 data, shrinking it in the
.m e -direction. This further highlights the significant differences in the role of .αEM

and .m e on the CMB anisotropies.

2 For a detailed discussion of the effects of various analysis choices we refer to [34].
33 Varying Fundamental Constants Meet Hubble 623

Table 33.1 Marginalised values of the fine structure constant ) .αEM and
( and effective electron mass
.m eusing the Planck 2018 data alone. A very wide prior . H0 > 20 km s−1 Mpc−1 for . H0 was
used to avoid biasing the marginalised .m e posterior, which affected some of the results of the 2013
analysis [34, 36]
Parameter Planck 2018 Planck 2018 Planck 2018
+ varying .αEM + varying .m e
+0.0012
.Ωb h ± 0.00015 ± 0.00015
2 .0.02237 .0.02236 .0.0199−0.0014

.Ωc h .0.1199 ± 0.0012 .0.1201 ± 0.0014 ± 0.0076

2 .0.1058

.100θ MC .1.04088 ± 0.00031 .1.0416 ± 0.0034 .0.958 ± 0.045

.τ .0.0542 ± 0.0074 .0.0540 ± 0.0075 .0.0512 ± 0.0077

s) .3.044 ± 0.014 .3.043 ± 0.015 .3.029 ± 0.017

.ln(10
10 A

.n s .0.9649 ± 0.0041 .0.9637 ± 0.0070 .0.9640 ± 0.0040

.αEM /αEM,0 .− .1.0005 ± 0.0024 .−

.m e /m e ,0 .− .− .0.888 ± 0.059
+9
. H0 [km s−1 Mpc−1 ] .67.36 ± 0.54 .67.56 ± 0.99 .46−10

Fig. 33.6 Posterior contours

between .αEM and .m e for the
Planck 2015 and 2018 data
along with BAO
contributions. Note that the
dashed contour shows the
2015 contour but with a
tighter prior on . H0 ∈
{40, 100} km s−1 Mpc−1 to
conform with the previous
Planck 2013 analysis. The
figure was taken from [36]

The introduction of BAO data significantly tightens the constraints on .m e and we

can also observe a small drift in the centeral value of.αEM . The obtained value changes
from .αEM /αEM, 0 = 0.9989 ± 0.0026 for the 2015 data to .αEM /αEM, 0 = 1.0010 ±
0.0024 with 2018 data, both with BAO included and when simultaneously varying
.m e . By contrast, there is no drift for .m e , with .m e /m e, 0 = 1.0056 ± 0.0080 chang-
ing to .m e /m e, 0 = 1.0054 ± 0.0080 for Planck 2015+BAO and Planck 2018+BAO,
respectively. These findings highlight a strong level of agreement of Planck with
BAO data, which has been emphasized on many occasions (e.g., [68]).
624 J. Chluba and L. Hart

33.4 Alleviating the Hubble Tension with VFCs

The attentive reader will already have noticed the route forward to alleviating the
Hubble tension through varying the electron rest mass. This possibility was first
noticed in [36], where the constraints on VFCs from different data combinations
were studied, also adding supernova/Cepheid (R19, [71]) data.
A summary of the constraints is shown in Fig. 33.7 with particular focus on the
interplay with . H0 . For .αEM , we can notice broadly consistent constraints across all
data combinations with only a small shift in the value of . H0 towards R19 in the
combined constraints, indicating a resistance in terms of geometric freedom. For .m e ,
two effects are found. First, as pointed out above, when adding BAO data, the error
on .m e is significantly reduced in comparison to the CMB only constraints, bringing
. H0 into agreement with the only CMB inference. Second, when also adding R19,

a non-standard value of .m e (at .≅ 3.5σ significance) is traded for a reduction of the

Hubble tension. This begs the question whether .m e could indeed play a role in this
problem. A model comparison study carried out in [79] indeed indicated that a simple
variation of.m e provides a good contender in this respect, although not all issues could
be resolved.

33.4.1 Adding Curvature

As Fig. 33.7 clearly shows, a constant variation of .m e does not fully solve the Hubble
tension, albeit reducing it below 2 .σ . As one possible extension, [36] also studied
solutions with power-law VFC time-dependence, but found this to not further improve
matters. A little later, [83] studied cosmologies with non-zero curvature in addition
to variations of .m e , finding these models to solve the Hubble tension. Indeed, this
possibility was the winning finalist in the . H0 -Olympics model comparison exercise
[79]. However, allowing for non-zero curvature does open yet another non-standard
direction in cosmology (in addition to accepting varying .m e ), with strong resistance
in terms of standard inflationary predictions [24]. One could feel inclined to cling on
to zero curvature cosmologies given the great successes of the inflationary paradigm,
but ultimately observations and careful analysis will have to decide.

33.4.2 Time-Dependent VFCs Models

Most of the results presented above assumed constant (i.e., time-independent/single-

valued) changes to the values of FCs in the early Universe. As highlighted in [34],
explicitly time-dependent VFCs can in principle be constrained independently even
with existing CMB data, given that the CMB responses are distinct (see Fig. 33.5).
This idea was later generalized by applying a principal component analysis (PCA) to
33 Varying Fundamental Constants Meet Hubble 625

Fig. 33.7 Constraints on the

fundamental constants (left)
using various combinations
of Planck data together with
their . H0 values and errors
(right). Top: results from the
fine structure constant .αEM .
Bottom: similar results but
from the effective electron
mass .m e . Here, we have
redacted the constraint for
. H0 from CMB data only
because the error bars are so
large. For the .m e MCMC
analysis, we have widened
the prior on the Hubble
constant such that
. H0 > 20 km s
−1 Mpc−1 .

Figure is from [36]

possible time-dependent VFC perturbations around the standard value in the hopes
to further diminish the Hubble tension [37]. The idea is very simple: given the
(CMB) observables, it is difficult to limit a specific change in the values of FCs
at a given redshift. However, certain correlated variations/perturbations (i.e., VFC
modes) across wider ranges of redshifts can indeed be analysized and constrained.
To construct these VFC eigenmodes, one can use the Fisher information matrix to
assess the observability. This yields a ranked system of possible VFC modes that
can be best constrained by the data. Similar approaches were applied to studies of
perturbations in the reionization [63] and recombination [29, 30, 35] histories.
626 J. Chluba and L. Hart

The first few .αEM and .m e eigenmodes are shown in Fig. 33.8. In a perturbative
sense, most of the information on VFCs can be gleaned from around the maximum
of the Thomson visiblity at .z ≅ 1100. This is reflected in the localization of the VFC
modes around this redshift. Applying these VFC modes to the Planck 2018 data
no significant detection of non-zero mode amplitudes was found [37]. Similarly,
when combining with BAO and R19, no additional improvement with respect to the

Fig. 33.8 The first three VFC principal components for .αEM (top) and{.m e (bottom) as constructed
using Planck 2018 data. The eigenmodes are all normalised such that . |E i2 (z)|dz = 1. The maxi-
mum of the .⌃CDM Thomson visibility function has been marked as vertical dotted line. Figure is
taken from [37]
33 Varying Fundamental Constants Meet Hubble 627

Hubble tension over a constant variation was identified [37]. However, to exclude
the possibility of a more general time-dependent VFC history, this would require
further investigations of time-dependent models that interface between the standard
candle, low-redshift era and the surface of last scattering.
The PCA approach inherently assumes a perturbative variation of the FCs. This
assumption need not hold, and in addition higher order eigenmodes that individually
fall below the detection threshold (due to their more rapid redshift-variablity) could
together allow for more general modifications (see, [30]). In [54], an assessment of
which time-dependent change to .αEM and .m e would be required to reduce the Hub-
ble tension was carried out. It was demonstrated that more general time-dependent
variations of .αEM and .m e around recombination can solve the Hubble tension (and
even reduce the . S8 tension) when applied to CMB data alone. However, once BAO
and supernova data is added, full solutions to the Hubble tension evade a perturba-
tive treatment, although extension of the framework to the non-perturbative regime
can be done [54]. This points towards the possibility that more general theoretical
models could indeed help to restore consistency between various probes, although
more work is certainly needed.

33.5 New Insights from the Cosmological Recombination

Radiation

In the previous sections, we have primarily focused on illustrating the role of VFCs,
and in particular of .m e , in possible solutions to the Hubble tension. In this, one key
ingredient is the associated modification to the cosmological recombination history,
although it seems clear that a simple perturbative change does not provide a sufficient
degree of freedom to fully resolve the tension. However, stepping back from a VFC
driven solution, one important question is how we could tell if indeed a (strongly)
modified recombination history is responsible for the Hubble tension? Furthermore,
solutions based on PMFs (see Chap. 32) are also mainly successful because of their
modifications to the average recombination history. It is thus important to ask how
could we distinguish different options?
One important avenue forward is to use more accurate measurements with upcom-
ing CMB experiments such as The Simons Observatory [2] and CMB-S4 [1] to
improve the constraints on VFC (see, [36, 37] for some projections). However, ulti-
mately access to new observables will be required.
The cosmological recombination processes is also associated with the emission of
photons from the Hydrogen and Helium plasma (e.g., [13, 17, 27, 28, 49, 75–77]).
The cosmological recombination radiation (CRR) can now be accurately computed
using CosmoSpec [15], which was extended to also incorporate the effect of VFCs
[38]. This CMB distortion signal may become observable [39] with future CMB
spectrometers akin to PIXIE [14, 51, 52], opening the exciting possibility to directly
study the dynamics of the recombination process [20, 88].
628 J. Chluba and L. Hart

In Fig. 33.9, we illustrate the effect of varying .αEM and .m e on the CRR. Firstly,
one can clearly see that for the chosen parameters there is a noticeable effect on the
amplitude and position of the CRR features. For .αEM , the modifications are more
pronounced, which in part is due to a near degeneracy of .m e with modifications to the
time of recombination [38]. Nevertheless, it is clear that precision measurements of
the average CMB spectrum could principally identify these modifications in compar-

Fig. 33.9 The total impact on the recombination lines from the variations for.ΔαEM /αEM, 0 = ±0.1
(top). The total variations due to changes in.Δm e /m e, 0 = ±0.2 are shown for comparison (bottom).
Figure taken from [38]
33 Varying Fundamental Constants Meet Hubble 629

Fig. 33.10 Comparison for an EDE model that has different slopes for .z c = 5000 across the
different atomic species. Hydrogen (top) is compared against the Helium I and II (middle, bottom)
lines, respectively. For better clarity, this has been plotted with . f ede = 0.8. Figure taken from [38]

ison to the standard .⌃CDM prediction, thereby shedding new light on the physical
cause of the problem.
However, we would like to highlight that these novel opportunities are not limited
to searches for VFCs. Also EDE models (Chap. 22) indirectly affect the shape of the
CRR [38]. In Fig. 33.10, we illustrate the effect of EDE on the CRR. Importantly,
depending on the parameter choices (see, [38] for additional details), the redshifting
between the Hydrogen and Helium recombination lines can be modified in addition to
when the lines are created. This leads to principally observable features in the CRR,
would a sufficient spectral sensitivity be reached. Finally, small-scale perturbations
in the baryonic density, as possibly induced by PMFs (Chap. 32), can leave an imprint
to the CRR that could be used to identify this process (see Fig. 33.11). Physically, the
modifications are simply related to the fact that the average CRR is no longer given
by the CRR of the average parameters, as non-linear (non-perturbative) corrections
to the recombination process remain. Overall, all these example illustrate that future
CMB spectral distortion measurements could in principle allow us to directly test the
underlying recombination process. From the scientific point of view it would actually
be crucial to investigate the recombination history directly, thereby eliminating one
of the remaining theoretical ingredients of CMB cosmology. However, it is also
630 J. Chluba and L. Hart

Fig. 33.11 Recombination histories (upper panel) and CRR spectra (lower panel) for the PMF
models similar to those consider in [32, 45, 46]. The individual spectra for various values of the
baryon density enhancement factor are shown as gray lines (the. Fb = 1 case corresponds to ⌃CDM).
The averages are instead displayed in blue and red, and show significant second order contributions,
manifesting in smearing of the lines and shifts in their position
33 Varying Fundamental Constants Meet Hubble 631

clear that due to the presence of foregrounds the required spectral sensitivity and
coverage are still quite futuristic [23, 39, 78]. In addition, a control of systematics
and removal of foregrounds will have to be performed to unprecedented precision.
These challenges will remain to be solved by generations of cosmologists to come.

33.6 Conclusions

In this chapter, we illustrated the role of VFCs in the Hubble tension. Previous studies
have shown that varying .m e could indeed offer viable solutions, although no VFC
scenario could fully resolve all aspects of the tension. It is however important to note
that additional studies in the non-perturbative regime (i.e., significant changes with
explicit time-dependence) could further improve the consistency. In addition, so far
no attempt has been made to simultaneously treat VFCs and EDE, although both
could principally originate from the same scalar field. With the advent of improved
cosmological data, these lines of research might become very interesting.
One obvious question remains: how can we ultimately distinguish between various
solutions to the Hubble tension. As highlighted in Sect. 33.5, EDE, PMFs and VFCs
can in principle also be (directly) constrained through detailed measurements of
the CRR. This guaranteed .⌃CDM signal may become observable in the future with
advanced CMB spectrometers and would open a way to confront our understanding of
the cosmological recombination process with direct observational evidence. Should
the Hubble tension not be resolved in the next decades, this future probe might be
one of the important avenues towards a final primordial test.

Acknowledgements This work was supported by the ERC Consolidator Grant CMBSPEC
(No. 725456). JC was furthermore supported by the Royal Society as a Royal Society Univer-
sity Research Fellow at the University of Manchester, UK (No. URF/R/191023). Part of this work
was carried out in the stimulating and peaceful environment offered at the Aspen Center for Physics,
which is supported by the National Science Foundation grant PHY-2210452.

References

1. K.N. Abazajian, P. Adshead, Z. Ahmed, S.W. Allen, D. Alonso, K.S. Arnold, C. Baccigalupi,
J.G. Bartlett, N. Battaglia, B.A. Benson, C.A. Bischoff, J. Borrill, V. Buza, E. Calabrese,
R. Caldwell, J.E. Carlstrom, C.L. Chang, T.M. Crawford, F.-Y. Cyr-Racine, F. De Bernardis,
T. de Haan, S. di Serego Alighieri, J. Dunkley, C. Dvorkin, J. Errard, G. Fabbian, S. Feeney,
S. Ferraro, J.P. Filippini, R. Flauger, G.M. Fuller, V. Gluscevic, D. Green, D. Grin, E. Grohs,
J.W. Henning, J.C. Hill, R. Hlozek, G. Holder, W. Holzapfel, W. Hu, K.M. Huffenberger,
R. Keskitalo, L. Knox, A. Kosowsky, J. Kovac, E.D. Kovetz, C.-L. Kuo, A. Kusaka, M. Le Jeune,
A.T. Lee, M. Lilley, M. Loverde, M.S. Madhavacheril, A. Mantz, D.J.E. Marsh, J. McMahon,
P.D. Meerburg, J. Meyers, A.D. Miller, J.B. Munoz, H.N. Nguyen, M.D. Niemack, M. Peloso,
J. Peloton, L. Pogosian, C. Pryke, M. Raveri, C.L. Reichardt, G. Rocha, A. Rotti, E. Schaan,
M.M. Schmittfull, D. Scott, N. Sehgal, S. Shandera, B.D. Sherwin, T.L. Smith, L. Sorbo, G.D.
632 J. Chluba and L. Hart

Starkman, K.T. Story, A. van Engelen, J.D. Vieira, S. Watson, N. Whitehorn, W.L. Kimmy Wu.
CMB-S4 Science Book, 1st edn. (2016) ArXiv:1610.0274
2. P. Ade, J. Aguirre, Z. Ahmed, S. Aiola, A. Ali, D. Alonso, M.A. Alvarez, K. Arnold, P. Ash-
ton, J. Austermann, H. Awan, C. Baccigalupi, T. Baildon, D. Barron, N. Battaglia, R. Battye,
E. Baxter, A. Bazarko, J.A. Beall, R. Bean, D. Beck, S. Beckman, B. Beringue, F. Bianchini,
S. Boada, D. Boettger, J.R. Bond, J. Borrill, M.L. Brown, S.M. Bruno, S. Bryan, E. Calabrese,
V. Calafut, P. Calisse, J. Carron, A. Challinor, G. Chesmore, Y. Chinone, J. Chluba, H.-M.S.
Cho, S. Choi, G. Coppi, N.F. Cothard, K. Coughlin, D. Crichton, K.D. Crowley, K.T. Crowley,
A. Cukierman, J.M. D’Ewart, R. Dünner, T. de Haan, M. Devlin, S. Dicker, J. Didier, M. Dobbs,
B. Dober, C.J. Duell, S. Duff, A. Duivenvoorden, J. Dunkley, J. Dusatko, J. Errard, G. Fabbian,
S. Feeney, S. Ferraro, P. Fluxà, K. Freese, J.C. Frisch, A. Frolov, G. Fuller, B. Fuzia, N. Gal-
itzki, P.A. Gallardo, J. Tomas Galvez Ghersi, J. Gao, E. Gawiser, M. Gerbino, V. Gluscevic,
N. Goeckner-Wald, J. Golec, S. Gordon, M. Gralla, D. Green, A. Grigorian, J. Groh, C. Groppi,
Y. Guan, J.E. Gudmundsson, D. Han, P. Hargrave, M. Hasegawa, M. Hasselfield, M. Hattori,
V. Haynes, M. Hazumi, Y. He, E. Healy, S.W. Henderson, C. Hervias-Caimapo, C.A. Hill, J.C.
Hill, G. Hilton, M. Hilton, A.D. Hincks, G. Hinshaw, R. Hložek, S. Ho, S.-P.P. Ho, L. Howe,
Z. Huang, J. Hubmayr, K. Huffenberger, J.P. Hughes, A. Ijjas, M. Ikape, K. Irwin, A.H. Jaffe,
B. Jain, O. Jeong, D. Kaneko, E.D. Karpel, N. Katayama, B. Keating, S.S. Kernasovskiy,
R. Keskitalo, T. Kisner, K. Kiuchi, J. Klein, K. Knowles, B. Koopman, A. Kosowsky, N. Krach-
malnicoff, S.E. Kuenstner, C.-L. Kuo, A. Kusaka, J. Lashner, A. Lee, E. Lee, D. Leon, J.S.Y.
Leung, A. Lewis, Y. Li, Z. Li, M. Limon, E. Linder, C. Lopez-Caraballo, T. Louis, L. Lowry,
M. Lungu, M. Madhavacheril, D. Mak, F. Maldonado, H. Mani, B. Mates, F. Matsuda, L. Mau-
rin, P. Mauskopf, A. May, N. McCallum, C. McKenney, J. McMahon, P.D. Meerburg, J. Mey-
ers, A. Miller, M. Mirmelstein, K. Moodley, M. Munchmeyer, C. Munson, S. Naess, F. Nati,
M. Navaroli, L. Newburgh, H.N. Nguyen, M. Niemack, H. Nishino, J. Orlowski-Scherer,
L. Page, B. Partridge, J. Peloton, F. Perrotta, L. Piccirillo, G. Pisano, D. Poletti, R. Puddu,
G. Puglisi, C. Raum, C.L. Reichardt, M. Remazeilles, Y. Rephaeli, D. Riechers, F. Rojas,
A. Roy, S. Sadeh, Y. Sakurai, M. Salatino, M. Sathyanarayana Rao, E. Schaan, M. Schmittfull,
N. Sehgal, J. Seibert, U. Seljak, B. Sherwin, M. Shimon, C. Sierra, J. Sievers, P. Sikhosana,
M. Silva-Feaver, S.M. Simon, A. Sinclair, P. Siritanasak, K. Smith, S.R. Smith, D. Spergel,
S.T. Staggs, G. Stein, J.R. Stevens, R. Stompor, A. Suzuki, O. Tajima, S. Takakura, G. Teply,
D.B. Thomas, B. Thorne, R. Thornton, H. Trac, C. Tsai, C. Tucker, J. Ullom, S. Vagnozzi,
A. van Engelen, J. Van Lanen, D.D. Van Winkle, E.M. Vavagiakis, C. Vergès, M. Vissers,
K. Wagoner, S. Walker, J. Ward, B. Westbrook, N. Whitehorn, J. Williams, J. Williams, E.J.
Wollack, Z. Xu, B. Yu, C. Yu, F. Zago, H. Zhang, N. Zhu, Simons observatory collaboration.
The Simons observatory: science goals and forecasts. JCAP 2019(2), 056 (2019)
3. S. Alam, F.D. Albareti, C.A. Prieto, F. Anders, S.F. Anderson, T. Anderton, B.H. Andrews, E.
Armengaud, A. Aubourg, S. Bailey et al., The eleventh and twelfth data releases of the Sloan
digital sky survey: Final data from SDSS-iii. The Astrophys. J. Suppl. Series 219(1), 12 (2015)
4. Y. Ali-Haïmoud, D. Grin, C.M. Hirata. Radiative transfer effects in primordial hydrogen recom-
bination. Phys. Rev. D 82(12), 123502 (2010)
5. Y. Ali-Haïmoud, C.M. Hirata. HyRec: a fast and highly accurate primordial hydrogen and
helium recombination code. Phys. Rev. D 83(4), 043513 (2011)
6. P.P. Avelino, S. Esposito, G. Mangano, C.J.A.P. Martins, A. Melchiorri, G. Miele, O. Pisanti,
G. Rocha, P.T.P. Viana, Early-universe constraints on a time-varying fine structure constant.
Phys. Rev. D 64(10), 103505 (2001)
7. P.P. Avelino, C.J.A.P. Martins, G. Rocha, P. Viana, Looking for a varying .α in the cosmic
microwave background. Phys. Rev. D 62(12), 123508 (2000)
8. R.A. Battye, R. Crittenden, J. Weller, Cosmic concordance and the fine structure constant.
Phys. Rev. D 63(4), 043505 (2001)
9. J.D. Bekenstein, Fine-structure constant: is it really a constant? Phys. Rev. D 25, 1527–1539
(1982)
10. H.A. Bethe, E.E. Salpeter. Quantum Mechanics of One- and Two-Electron Atoms (1957)
33 Varying Fundamental Constants Meet Hubble 633

11. S. Bize, S.A. Diddams, U. Tanaka, C.E. Tanner, W.H. Oskay, R.E. Drullinger, T.E. Parker,
T.P. Heavner, S.R. Jefferts, L. Hollberg, W.M. Itano, J.C. Bergquist, Testing the stability of
fundamental constants with the .199 Hg+ single-ion optical clock. Phys. Rev. Lett. 90, 150802
(2003)
12. P. Bonifacio, H. Rahmani, J.B. Whitmore, M. Wendt, M. Centurion, P. Molaro, R. Srianand,
M.T. Murphy, P. Petitjean, I.I. Agafonova, S. D’Odorico, T.M. Evans, S.A. Levshakov, S.
Lopez, C.J.A.P. Martins, D. Reimers, G. Vladilo, Fundamental constants and high-resolution
spectroscopy. Astronomische Nachrichten 335, 83 (2014)
13. M.S. Burgin, Hydrogen subordinate line emission at the epoch of cosmological recombination.
Astron. Rep. 47, 709–716 (2003)
14. J. Chluba, M.H. Abitbol, N. Aghanim, Y. Ali-Haïmoud, M. Alvarez, K. Basu, B. Bolliet, C.
Burigana, P. de Bernardis, J. Delabrouille, E. Dimastrogiovanni, F. Finelli, D. Fixsen, L. Hart,
C. Hernández-Monteagudo, J.C. Hill, A. Kogut, K. Kohri, J. Lesgourgues, B. Maffei, J. Mather,
S. Mukherjee, S.P. Patil, A. Ravenni, M. Remazeilles, A. Rotti, J.A. Rubiño-Martin, J. Silk,
R.A. Sunyaev, E.R. Switzer, New horizons in cosmology with spectral distortions of the cosmic
microwave background. Exp. Astron. 51(3), 1515–1554 (2021)
15. J. Chluba, Y. Ali-Haïmoud, COSMOSPEC: fast and detailed computation of the cosmological
recombination radiation from hydrogen and helium. MNRAS 456, 3494–3508 (2016)
16. J. Chluba, J. Fung, E.R. Switzer, Radiative transfer effects during primordial helium recombi-
nation. MNRAS 423, 3227–3242 (2012)
17. J. Chluba, R.A. Sunyaev, Free-bound emission from cosmological hydrogen recombination.
A&A 458, L29–L32 (2006)
18. J. Chluba, R.A. Sunyaev, Induced two-photon decay of the 2s level and the rate of cosmological
hydrogen recombination. A&A 446, 39–42 (2006)
19. J. Chluba, R.A. Sunyaev, Cosmological hydrogen recombination: lyn line feedback and con-
tinuum escape. A&A 475, 109–114 (2007)
20. J. Chluba, R.A. Sunyaev, Is there a need and another way to measure the cosmic microwave
background temperature more accurately? A&A 478, L27–L30 (2008)
21. J. Chluba, R.A. Sunyaev, Time-dependent corrections to the Ly .α escape probability during
cosmological recombination. A&A 496, 619–635 (2009)
22. J. Chluba, R.M. Thomas, Towards a complete treatment of the cosmological recombination
problem. MNRAS 412, 748–764 (2011)
23. V. Desjacques, J. Chluba, J. Silk, F. de Bernardis, O. Doré, Detecting the cosmological recom-
bination signal from space. MNRAS 451, 4460–4470 (2015)
24. E. Di Dio, F. Montanari, A. Raccanelli, R. Durrer, M. Kamionkowski, J. Lesgourgues, Curvature
constraints from large scale structure. JCAP 2016(6), 013 (2016)
25. G.W.F. Drake. Springer Handbook of Atomic, Molecular, and Optical Physics (Springer, 2006)
26. G.W.F. Drake, D.C. Morton, A multiplet table for neutral Helium (.4 He I) with transition rates.
ApJS 170, 251–260 (2007)
27. V.K. Dubrovich. Hydrogen recombination lines of cosmological origin. Soviet Astron. Lett. 1,
196 (1975)
28. V.K. Dubrovich, V.A. Stolyarov. Fossil radio lines of hydrogen in the cosmic background
radiation at decimeter and meter wavelengths. A&A 302, 635 (1995)
29. M. Farhang, J.R. Bond, J. Chluba, Semi blind Eigen analyses of recombination histories using
cosmic microwave background data. ApJ 752, 88 (2012)
30. M. Farhang, J.R. Bond, J. Chluba, E.R. Switzer, Constraints on perturbations to the recom-
bination history from measurements of the cosmic microwave background damping tail. ApJ
764, 137 (2013)
31. W.A. Fendt, J. Chluba, J.A. Rubiño-Martín, B.D. Wandelt, RICO: a new approach for fast and
accurate representation of the cosmological recombination history. ApJS 181, 627–638 (2009)
32. S. Galli, L. Pogosian, K. Jedamzik, L. Balkenhol, Consistency of Planck, ACT, and SPT
constraints on magnetically assisted recombination and forecasts for future experiments. Phys.
Rev. D 105(2), 023513 (2022)
634 J. Chluba and L. Hart

33. S.C.O. Glover, J. Chluba, S.R. Furlanetto, J.R. Pritchard, D.W. Savin, Chapter three—atomic,
molecular, and optical physics in the early universe: from recombination to reionization. Adv.
Atomic Mol. Opt. Phys. 63, 135–270 (2014)
34. L. Hart, J. Chluba, New constraints on time-dependent variations of fundamental constants
using Planck data. MNRAS 474(2), 1850–1861 (2018)
35. L. Hart, J. Chluba, Improved model-independent constraints on the recombination era and
development of a direct projection method. MNRAS 495(4), 4210–4226 (2020)
36. L. Hart, J. Chluba, Updated fundamental constant constraints from Planck 2018 data and
possible relations to the Hubble tension. MNRAS 493(3), 3255–3263 (2020)
37. L. Hart, J. Chluba, Varying fundamental constants principal component analysis: additional
hints about the Hubble tension. MNRAS 510(2), 2206–2227 (2022)
38. L. Hart, J. Chluba, Using the cosmological recombination radiation to probe early dark energy
and fundamental constant variations. MNRAS 519(3), 3664–3680 (2023)
39. L. Hart, A. Rotti, J. Chluba, Sensitivity forecasts for the cosmological recombination radiation
in the presence of foregrounds. MNRAS 497(4), 4535–4548 (2020)
40. C.M. Hirata. Two-photon transitions in primordial hydrogen recombination. Phys. Rev. D
78(2), 023001 (2008)
41. K. Hoshiya, Y. Toda, Electron mass variation from dark sector interactions and compatibility
with cosmological observations. Phys. Rev. D 107(4), 043505 (2023)
42. W. Hu, D. Scott, N. Sugiyama, M. White, Effect of physical assumptions on the calculation of
microwave background anisotropies. Phys. Rev. D 52, 5498–5515 (1995)
43. W. Hu, N. Sugiyama, Anisotropies in the cosmic microwave background: an analytic approach.
ApJ 444, 489–506 (1995)
44. W. Hu, N. Sugiyama, Small-scale cosmological perturbations: an analytic approach. ApJ 471,
542 (1996)
45. K. Jedamzik, T. Abel, Small-scale primordial magnetic fields and anisotropies in the cosmic
microwave background radiation. JCAP 10, 050 (2013)
46. K. Jedamzik, L. Pogosian, Relieving the Hubble tension with primordial magnetic fields. Phys.
Rev. Lett. 125(18), 181302 (2020)
47. M. Kaplinghat, R.J. Scherrer, M.S. Turner. Constraining variations in the fine-structure constant
with the cosmic microwave background. Phys. Rev. D 60(2), 023516 (1999)
48. W.J. Karza, R. Latter, Electron radiative transitions in a Coulomb field. ApJS 6, 167 (1961)
49. E.E. Kholupenko, A.V. Ivanchik, D.A. Varshalovich, Gravitation Cosmol. 11, 161–165 (2005)
50. E.E. Kholupenko, A.V. Ivanchik, D.A. Varshalovich, Rapid HeII-HeI recombination and radi-
ation arising from this process. MNRAS 378, L39–L43 (2007)
51. A. Kogut, D.J. Fixsen, Foreground Bias from parametric models of Far-IR dust emission. ApJ
826, 101 (2016)
52. A. Kogut, D.J. Fixsen, D.T. Chuss, J. Dotson, E.Dwek, M. Halpern, G.F. Hinshaw, S.M. Meyer,
S.H. Moseley, M.D. Seiffert, D.N. Spergel, E.J. Wollack. The Primordial Inflation Explorer
(PIXIE): a nulling polarimeter for cosmic microwave background observations. JCAP 7, 25
(2011)
53. S.M. Kotuš, M.T. Murphy, R.F. Carswell, High-precision limit on variation in the fine-structure
constant from a single quasar absorption system. Monthly Notices of the Royal Astron. Soc.
464(3), 3679 (2017)
54. N. Lee, Y. Ali-Haïmoud, N. Schöneberg, V. Poulin, What it takes to solve the Hubble tension
through modifications of cosmological recombination. Phys. Rev. Lett. 130(16), 161003 (2023)
55. J. Lesgourgues. The cosmic linear anisotropy solving system (CLASS) I: overview (2011).
hyperimagehttp://arxiv.org/abs/1104.2932ArXiv:1104.2932
56. A. Lewis, A. Challinor, A. Lasenby, Efficient computation of cosmic microwave background
anisotropies in closed Friedmann-Robertson-Walker models. ApJ 538, 473–476 (2000)
57. A. Lewis, J. Weller, R. Battye, The cosmic microwave background and the ionization history
of the Universe. MNRAS 373, 561–570 (2006)
58. M. Lucca, J. Chluba, A. Rotti. CRRfast: an emulator for the cosmological recombination radia-
tion with effects from inhomogeneous recombination. arXiv e-prints (2023). arXiv:2306.08085
33 Varying Fundamental Constants Meet Hubble 635

59. C.J.A.P. Martins, The status of varying constants: a review of the physics, searches and impli-
cations. Rep. Progr. Phys. 80(12), 126902 (2017)
60. C.J.A.P. Martins, A. Melchiorri, G. Rocha, R. Trotta, P.P. Avelino, P.T.P. Viana, WMAP con-
straints on varying .α and the promise of reionization. Phys. Lett. B 585, 29–34 (2004)
61. C.J.A.P. Martins, P.E. Vielzeuf, M. Martinelli, E. Calabrese, S. Pandolfi, Evolution of the
fine-structure constant in runaway dilaton models. Phys. Lett. B 743, 377–382 (2015)
62. E. Menegoni, M. Archidiacono, E. Calabrese, S. Galli, C.J.A.P. Martins, A. Melchiorri, Fine
structure constant and the CMB damping scale. Phys. Rev. D 85(10), 107301 (2012)
63. M.J. Mortonson, C. Dvorkin, H.V. Peiris, W. Hu, CMB polarization features from inflation
versus reionization. Phys. Rev. D 79, 103519 (2009)
64. P.J.E. Peebles, Recombination of the primeval plasma. ApJ 153, 1 (1968)
65. P.J.E. Peebles, J.T. Yu, Primeval adiabatic perturbation in an expanding universe. ApJ 162, 815
(1970)
66. Planck Collaboration, P.A.R. Ade, N. Aghanim, M. Arnaud, M. Ashdown, J. Aumont, C. Bacci-
galupi, A.J. Banday, R.B. Barreiro, J.G. Bartlett et al., Planck 2015 results. XIII. Cosmological
parameters (2015). ArXiv:1502.01589
67. Planck Collaboration, P.A.R. Ade, N. Aghanim, M. Arnaud, M. Ashdown, J. Aumont, C. Bac-
cigalupi, A.J. Banday, R.B. Barreiro, E. Battaner, K. Benabed, A. Benoit-Lévy, J.-P. Bernard,
M. Bersanelli, P. Bielewicz, J.R. Bond, J. Borrill, F.R. Bouchet, C. Burigana, R.C. Butler,
E. Calabrese, A. Chamballu, H.C. Chiang, P.R. Christensen, D.L. Clements, L.P.L. Colombo,
F. Couchot, A. Curto, F. Cuttaia, L. Danese, R.D. Davies, R. J. Davis, P. de Bernardis, A. de
Rosa, G. de Zotti, J. Delabrouille, J.M. Diego, H. Dole, O. Doré, X. Dupac, T. A. Enßlin,
H.K. Eriksen, O. Fabre, F. Finelli, O. Forni, M. Frailis, E. Franceschi, S. Galeotta, S. Galli,
K. Ganga, M. Giard, J. González-Nuevo, K.M. Górski, A. Gregorio, A. Gruppuso, F.K. Hansen,
D. Hanson, D.L. Harrison, S. Henrot-Versillé, C. Hernández-Monteagudo, D. Herranz, S.R.
Hildebrandt, E. Hivon, M. Hobson, W.A. Holmes, A. Hornstrup, W. Hovest, K.M. Huffen-
berger, A.H. Jaffe, W.C. Jones, E. Keihänen, R. Keskitalo, R. Kneissl, J. Knoche, M. Kunz,
H. Kurki-Suonio, J.-M. Lamarre, A. Lasenby, C.R. Lawrence, R. Leonardi, J. Lesgourgues,
M. Liguori, P.B. Lilje, M. Linden-Vørnle, M. López-Caniego, P.M. Lubin, J.F. Macías-Pérez,
N. Mandolesi, M. Maris, P.G. Martin, E. Martínez-González, S. Masi, S. Matarrese, P. Mazzotta,
P.R. Meinhold, A. Melchiorri, L. Mendes, E. Menegoni, A. Mennella, M. Migliaccio, M.-A.
Miville-Deschênes, A. Moneti, L. Montier, G. Morgante, A. Moss, D. Munshi, J.A. Murphy,
P. Naselsky, F. Nati, P. Natoli, H.U. Nørgaard-Nielsen, F. Noviello, D. Novikov, I. Novikov, C.A.
Oxborrow, L. Pagano, F. Pajot, D. Paoletti, F. Pasian, G. Patanchon, O. Perdereau, L. Perotto,
F. Perrotta, F. Piacentini, M. Piat, E. Pierpaoli, D. Pietrobon, S. Plaszczynski, E. Pointecouteau,
G. Polenta, N. Ponthieu, L. Popa, G.W. Pratt, S. Prunet, J.P. Rachen, R. Rebolo, M. Reinecke,
M. Remazeilles, C. Renault, S. Ricciardi, I. Ristorcelli, G. Rocha, G. Roudier, B. Rusholme,
M. Sandri, G. Savini, D. Scott, L.D. Spencer, V. Stolyarov, R. Sudiwala, D. Sutton, A.-S.
Suur-Uski, J.-F. Sygnet, J.A. Tauber, D. Tavagnacco, L. Terenzi, L. Toffolatti, M. Tomasi,
M. Tristram, M. Tucci, J.-P. Uzan, L. Valenziano, J. Valiviita, B. Van Tent, P. Vielva, F. Villa,
L.A. Wade, D. Yvon, A. Zacchei, A. Zonca, Planck intermediate results. XXIV. Constraints
on variations in fundamental constants. A&A 580, A22 (2015)
68. Planck Collaboration, N. Aghanim, Y. Akrami, M. Ashdown, J. Aumont, C. Baccigalupi,
M. Ballardini, A.J. Banday, R.B. Barreiro, N. Bartolo, S. Basak, R. Battye, K. Benabed, J.P.
Bernard, M. Bersanelli, P. Bielewicz, J.J. Bock, J.R. Bond, J. Borrill, F.R. Bouchet, F. Boulanger,
M. Bucher, C. Burigana, R.C. Butler, E. Calabrese, J.F. Cardoso, J. Carron, A. Challinor,
H.C. Chiang, J. Chluba, L.P.L. Colombo, C. Combet, D. Contreras, B.P. Crill, F. Cuttaia,
P. de Bernardis, G. de Zotti, J. Delabrouille, J.M. Delouis, E. Di Valentino, J.M. Diego, O. Doré,
M. Douspis, A. Ducout, X. Dupac, S. Dusini, G. Efstathiou, F. Elsner, T.A. Enßlin, H.K.
Eriksen, Y. Fantaye, M. Farhang, J. Fergusson, R. Fernandez-Cobos, F. Finelli, F. Forastieri,
M. Frailis, E. Franceschi, A. Frolov, S. Galeotta, S. Galli, K. Ganga, R.T. Génova-Santos,
M. Gerbino, T. Ghosh, J. González-Nuevo, K.M. Górski, S. Gratton, A. Gruppuso, J.E. Gud-
mundsson, J. Hamann, W. Handley, D. Herranz, E. Hivon, Z. Huang, A.H. Jaffe, W.C. Jones,
636 J. Chluba and L. Hart

A. Karakci, E. Keihänen, R. Keskitalo, K. Kiiveri, J. Kim, T.S. Kisner, L. Knox, N. Krach-

malnicoff, M. Kunz, H. Kurki-Suonio, G. Lagache, J.M. Lamarre, A. Lasenby, M. Lattanzi,
C.R. Lawrence, M. Le Jeune, P. Lemos, J. Lesgourgues, F. Levrier, A. Lewis, M. Liguori,
P.B. Lilje, M. Lilley, V. Lindholm, M. López-Caniego, P.M. Lubin, Y.Z. Ma, J.F. Macías-
Pérez, G. Maggio, D. Maino, N. Mandolesi, A. Mangilli, A. Marcos-Caballero, M. Maris,
P.G. Martin, M. Martinelli, E. Martínez-González, S. Matarrese, N. Mauri, J.D. McEwen, P.R.
Meinhold, A. Melchiorri, A. Mennella, M. Migliaccio, M. Millea, S. Mitra, M.A. Miville-
Deschênes, D. Molinari, L. Montier, G. Morgante, A. Moss, P. Natoli, H.U. Nørgaard-Nielsen,
L. Pagano, D. Paoletti, B. Partridge, G. Patanchon, H.V. Peiris, F. Perrotta, V. Pettorino, F. Pia-
centini, L. Polastri, G. Polenta, J.L. Puget, J.P. Rachen, M. Reinecke, M. Remazeilles, A. Renzi,
G. Rocha, C. Rosset, G. Roudier, J.A. Rubiño-Martín, B. Ruiz-Granados, L. Salvati, M. San-
dri, M. Savelainen, D. Scott, E.P.S. Shellard, C. Sirignano, G. Sirri, L.D. Spencer, R. Sunyaev,
A.S. Suur-Uski, J.A. Tauber, D. Tavagnacco, M. Tenti, L. Toffolatti, M. Tomasi, T. Trombetti,
L. Valenziano, J. Valiviita, B. Van Tent, L. Vibert, P. Vielva, F. Villa, N. Vittorio, B.D. Wan-
delt, I.K. Wehus, M. White, S.D.M. White, A. Zacchei, A. Zonca. Planck 2018 results. VI.
Cosmological parameters (2018). ArXiv:1807.06209
69. Planck Collaboration, N. Aghanim, Y. Akrami, M. Ashdown, J. Aumont, C. Baccigalupi,
M. Ballardini, A.J. Banday, R.B. Barreiro, N. Bartolo, S. Basak, K. Benabed, J.P. Bernard,
M. Bersanelli, P. Bielewicz, J.J. Bock, J.R. Bond, J. Borrill, F.R. Bouchet, F. Boulanger,
M. Bucher, C. Burigana, R.C. Butler, E. Calabrese, J.F. Cardoso, J. Carron, B. Casaponsa,
A. Challinor, H.C. Chiang, L.P.L. Colombo, C. Combet, B.P. Crill, F. Cuttaia, P. de Bernardis,
A. de Rosa, G. de Zotti, J. Delabrouille, J.M. Delouis, E. Di Valentino, J.M. Diego, O. Doré,
M. Douspis, A. Ducout, X. Dupac, S. Dusini, G. Efstathiou, F. Elsner, T.A. Enßlin, H.K. Eriksen,
Y. Fantaye, R. Fernand ez-Cobos, F. Finelli, M. Frailis, A.A. Fraisse, E. Franceschi, A. Frolov,
S. Galeotta, S. Galli, K. Ganga, R.T. Génova-Santos, M. Gerbino, T. Ghosh, Y. Giraud-Héraud,
J. González-Nuevo, K.M. Górski, S. Gratton, A. Gruppuso, J.E. Gudmundsson, J. Hamann,
W. Handley, F.K. Hansen, D. Herranz, E. Hivon, Z. Huang, A.H. Jaffe, W.C. Jones, E. Keihänen,
R. Keskitalo, K. Kiiveri, J. Kim, T.S. Kisner, N. Krachmalnicoff, M. Kunz, H. Kurki-Suonio,
G. Lagache, J.M. Lamarre, A. Lasenby, M. Lattanzi, C.R. Lawrence, M. Le Jeune, F. Levrier,
A. Lewis, M. Liguori, P.B. Lilje, M. Lilley, V. Lindholm, M. López-Caniego, P.M. Lubin, Y.Z.
Ma, J.F. Macías-Pérez, G. Maggio, D. Maino, N. Mandolesi, A. Mangilli, A. Marcos-Caballero,
M. Maris, P.G. Martin, E. Martínez-González, S. Matarrese, N. Mauri, J.D. McEwen, P.R.
Meinhold, A. Melchiorri, A. Mennella, M. Migliaccio, M. Millea, M.A. Miville-Deschênes,
D. Molinari, A. Moneti, L. Montier, G. Morgante, A. Moss, P. Natoli, H.U. Nørgaard-Nielsen,
L. Pagano, D. Paoletti, B. Partridge, G. Patanchon, H.V. Peiris, F. Perrotta, V. Pettorino,
F. Piacentini, G. Polenta, J.L. Puget, J.P. Rachen, M. Reinecke, M. Remazeilles, A. Renzi,
G. Rocha, C. Rosset, G. Roudier, J.A. Rubiño-Martín, B. Ruiz-Granados, L. Salvati, M. San-
dri, M. Savelainen, D. Scott, E.P.S. Shellard, C. Sirignano, G. Sirri, L.D. Spencer, R. Sunyaev,
A.S. Suur-Uski, J.A. Tauber, D. Tavagnacco, M. Tenti, L. Toffolatti, M. Tomasi, T. Trombetti,
J. Valiviita, B. Van Tent, P. Vielva, F. Villa, N. Vittorio, B.D. Wandelt, I.K. Wehus, A. Zacchei,
A. Zonca. Planck 2018 results. V. CMB power spectra and likelihoods (2019). arXiv e-prints
arXiv:1907.12875
70. Planck Collaboration, N. Aghanim, Y. Akrami, M. Ashdown, J. Aumont, C. Baccigalupi,
M. Ballardini, A.J. Banday, R.B. Barreiro, N. Bartolo, S. Basak, K. Benabed, J.P. Bernard,
M. Bersanelli, P. Bielewicz, J.J. Bock, J.R. Bond, J. Borrill, F.R. Bouchet, F. Boulanger,
M. Bucher, C. Burigana, E. Calabrese, J.F. Cardoso, J. Carron, A. Challinor, H.C. Chiang, L.P.L.
Colombo, C. Combet, B.P. Crill, F. Cuttaia, P. de Bernardis, G. de Zotti, J. Delabrouille, E. Di
Valentino, J.M. Diego, O. Doré, M. Douspis, A. Ducout, X. Dupac, G. Efstathiou, F. Elsner, T.A.
Enßlin, H.K. Eriksen, Y. Fantaye, R. Fernandez-Cobos, F. Forastieri, M. Frailis, A.A. Fraisse,
E. Franceschi, A. Frolov, S. Galeotta, S. Galli, K. Ganga, R.T. Génova-Santos, M. Gerbino,
T. Ghosh, J. González-Nuevo, K.M. Górski, S. Gratton, A. Gruppuso, J.E. Gudmundsson,
J. Hamann, W. Hand ley, F.K. Hansen, D. Herranz, E. Hivon, Z. Huang, A.H. Jaffe, W.C.
Jones, A. Karakci, E. Keihänen, R. Keskitalo, K. Kiiveri, J. Kim, L. Knox, N. Krachmalnicoff,
M. Kunz, H. Kurki-Suonio, G. Lagache, J.M. Lamarre, A. Lasenby, M. Lattanzi, C.R. Lawrence,
33 Varying Fundamental Constants Meet Hubble 637

M. Le Jeune, F. Levrier, A. Lewis, M. Liguori, P.B. Lilje, V. Lindholm, M. López-Caniego,

P.M. Lubin, Y.Z. Ma, J.F. Macías-Pérez, G. Maggio, D. Maino, N. Mandolesi, A. Mangilli,
A. Marcos-Caballero, M. Maris, P.G. Martin, E. Martínez-González, S. Matarrese, N. Mauri,
J.D. McEwen, A. Melchiorri, A. Mennella, M. Migliaccio, M.A. Miville-Deschênes, D. Moli-
nari, A. Moneti, L. Montier, G. Morgante, A. Moss, P. Natoli, L. Pagano, D. Paoletti, B. Par-
tridge, G. Patanchon, F. Perrotta, V. Pettorino, F. Piacentini, L. Polastri, G. Polenta, J.L. Puget,
J.P. Rachen, M. Reinecke, M. Remazeilles, A. Renzi, G. Rocha, C. Rosset, G. Roudier, J.A.
Rubiño-Martín, B. Ruiz-Granados, L. Salvati, M. Sandri, M. Savelainen, D. Scott, C. Sirignano,
R. Sunyaev, A.S. Suur-Uski, J.A. Tauber, D. Tavagnacco, M. Tenti, L. Toffolatti, M. Tomasi,
T. Trombetti, J. Valiviita, B. Van Tent, P. Vielva, F. Villa, N. Vittorio, B.D. Wandelt, I.K. Wehus,
M. White, S.D.M. White, A. Zacchei, A. Zonca. Planck 2018 results. VIII. Gravitational lensing
(2018). arXiv e-prints arXiv:1807.06210
71. A.G. Riess, S. Casertano, W. Yuan, L.M. Macri, D. Scolnic, Large Magellanic cloud Cepheid
standards provide a 1% foundation for the determination of the Hubble constant and stronger
evidence for physics beyond .⌃CDM. Astrophys. J. 876(1), 85 (2019)
72. G. Rocha, R. Trotta, C.J.A.P. Martins, A. Melchiorri, P.P. Avelino, R. Bean, P.T.P. Viana,
Measuring .α in the early Universe: cosmic microwave background polarization, re-ionization
and the Fisher matrix analysis. MNRAS 352, 20–38 (2004)
73. T. Rosenband, D.B. Hume, P.O. Schmidt, C.W. Chou, A. Brusch, L. Lorini, W.H. Oskay, R.E.
Drullinger, T.M. Fortier, J.E. Stalnaker, S.A. Diddams, W.C. Swann, N.R. Newbury, W.M.
Itano, D.J. Wineland, J.C. Bergquist, Frequency ratio of al+ and hg+ single-ion optical clocks;
metrology at the 17th decimal place. Science 319(5871), 1808–1812 (2008)
74. J.A. Rubiño-Martín, J. Chluba, W.A. Fendt, B.D. Wandelt, Estimating the impact of recombi-
nation uncertainties on the cosmological parameter constraints from cosmic microwave back-
ground experiments. MNRAS 403, 439–452 (2010)
75. J.A. Rubiño-Martín, J. Chluba, R.A. Sunyaev, Lines in the cosmic microwave background
spectrum from the epoch of cosmological helium recombination. A&A 485, 377–393 (2008)
76. J.A. Rubiño-Martín, C. Hernández-Monteagudo, R.A. Sunyaev, The imprint of cosmological
hydrogen recombination lines on the power spectrum of the CMB. A&A 438, 461–473 (2005)
77. G.B. Rybicki, I.P. dell’Antonio, The time development of a resonance line in the expanding
universe. ApJ 427, 603–617 (1994)
78. M. Sathyanarayana Rao, R. Subrahmanyan, N. Udaya Shankar, J. Chluba, On the detection of
spectral ripples from the recombination epoch. ApJ 810, 3 (2015)
79. N. Schöneberg, G.F. Abellán, A.P. Sánchez, S.J. Witte, V. Poulin, J. Lesgourgues, The H.0
Olympics: a fair ranking of proposed models. Phys. Rep. 984, 1–55 (2022)
80. C.G. Scóccola, S.J. Landau, H. Vucetich. WMAP 5-year constraints on time variation of alpha
and m_e. Memorie della Societ Astronomica Italiana 80, 814 (2009)
81. S. Seager, D.D. Sasselov, D. Scott, A new calculation of the recombination epoch. APJL 523,
L1–L5 (1999)
82. S. Seager, D.D. Sasselov, D. Scott, How exactly did the universe become neutral? ApJS 128,
407–430 (2000)
83. T. Sekiguchi, T. Takahashi, Early recombination as a solution to the H.0 tension. Phys. Rev. D
103(8), 083507 (2021)
84. J.R. Shaw, J. Chluba, Precise cosmological parameter estimation using COSMOREC. MNRAS
415, 1343–1354 (2011)
85. L.D. Shaw, D. Nagai, S. Bhattacharya, E.T. Lau, Impact of cluster physics on the Sunyaev-
Zel’dovich power spectrum. ApJ 725, 1452–1465 (2010)
86. T.L. Smith, D. Grin, D. Robinson, D. Qi, Probing spatial variation of the fine-structure constant
using the CMB. Phys. Rev. D 99(4), 043531 (2019)
87. V.V. Sobolev, Moving Envelopes of Stars (Harvard University Press, Cambridge, 1960), p.1960
88. R.A. Sunyaev, J. Chluba, Signals from the epoch of cosmological recombination (Karl
Schwarzschild Award Lecture 2008). Astronomische Nachrichten 330, 657 (2009)
89. R.A. Sunyaev, Y.B. Zeldovich, Small-scale fluctuations of relic radiation. ApSS 7, 3 (1970)
638 J. Chluba and L. Hart

90. H. Tohfa, J. Crump, E. Baker, L. Hart, D. Grin, M. Brosius, J. Chluba. A cosmic

microwave background search for fine-structure constant evolution (2023). arXiv e-prints
arXiv:2307.06768
91. J.-P. Uzan, The fundamental constants and their variation: observational and theoretical status.
Rev. Modern Phys. 75, 403–455 (2003)
92. J.-P. Uzan, Varying constants, gravitation and cosmology. Living Rev. Relativity 14, 2 (2011)
93. W.Y. Wong, A. Moss, D. Scott, How well do we understand cosmological recombination?
MNRAS 386, 1023–1028 (2008)
94. Y.B. Zeldovich, V.G. Kurt, R.A. Syunyaev, Recombination of hydrogen in the hot model of
the universe. Zhurnal Eksperimental noi i Teoreticheskoi Fiziki 55, 278–286 (1968)
Chapter 34
Anomalies and Tensions in Cosmology
and a Primordial Solution

Dhiraj Kumar Hazra and Arman Shafieloo

34.1 Introduction

Cosmological observations at scales probed by the Cosmic Microwave Background

(CMB) observations (GigaParsecs to about a hundred kiloParsecs) are well explained
by the 6-parameter .⌃CDM model with power-law form of primordial spectrum.
Different CMB observations from space-based missions and ground based surveys
broadly agree with each other apart from moderate tensions appearing in different
releases. With Planck [1] temperature and polarization data, the parameters of the
model are constrained to an unprecedented precision. The.⌃CDM also agrees with the
large scale structure (LSS), lensing, supernovae and other cosmological observations.
However, over the last few years, within this concordance picture of cosmology,
differences in the parameters estimated from different observations within the same
standard model aggravated to become statistically significant. At present, the dis-
agreement in the estimation of Hubble parameter between the CMB and the S. H0 ES
distance ladder measurements [2] is the most crucial tension in cosmology. Along
similar times, mismatch in growth parameter between measurements from lensing
surveys such as KiDS [3], DES [4] and Planck CMB have emerged as another impor-

D. K. Hazra (B)
The Institute of Mathematical Sciences, HBNI, CIT Campus, Chennai 600113, India
e-mail: dhiraj@imsc.res.in
Homi Bhabha National Institute, Training School Complex, Anushakti Nagar, Mumbai 400085,
India
INAF/OAS Bologna, Osservatorio di Astrofisica e Scienza dello Spazio, Area della ricerca
CNR-INAF, via Gobetti 101, 40129 Bologna, Italy
A. Shafieloo
Korea Astronomy and Space Science Institute, Daejeon 34055, Korea
e-mail: shafieloo@kasi.re.kr
University of Science and Technology, Daejeon 34113, Korea

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 639
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_34
640 D. K. Hazra and A. Shafieloo

tant tension. Within Planck CMB temperature data too, the preference for an ad-hoc
lensing parameter or spatial curvature indicates certain anomalies within the data.
Significance of these tensions do not leave much room for discussions related to
random fluctuations. At this stage, the scientific community is divided into two parts
on the origin of these tensions, namely systematic uncertainties and new physics. Over
the last couple of years, the extensive tests on the systematics have been performed
by various groups which did not result any significant reduction of these tensions.
Therefore, new physics is now considered as the most probable solution to these
tensions. Most of the new physics developed in this aspect changes the distance
scales by either changing the dark energy history [5, 6]. Majority of the solutions till
now has attempted to address these tensions separately with different models.
In this chapter we will discuss the possibility of these anomalies and tensions being
related. We will explore a particular solution to these in the spectrum of primordial
quantum fluctuations and demonstrate the possible origin from a scalar field model.

34.2 Tensions and Anomalies: A Few Key Points

In this section, in order to distinguish the terms ‘anomalies’ and ‘tensions’, we provide
two simple pictorial descriptions.
Based on a model (built with well established theories + certain supporting
assumptions) we match our predictions to the observations and obtain bounds on
the model parameters. These bounds refer to the best fit/mean predictions of the
model parameters and statistical uncertainties (standard deviation) associated with
them. For a model, consistent with the data from an observation, it is expected that
the mean model prediction to the observation should be within the reach of a few
standard deviations from the observed data. When that does not happen, we term this
mismatch as an anomaly in the data as shown in the top plot of Fig. 34.1.
Now, thanks to the different satellite and ground based probes operating at different
frequencies, we have access to several observing windows to the Universe. As light
travels at a finite speed, through these windows we can witness different stages of
the Universe. Now remember that the theoretical model we discussed as the .⌃CDM
model with power-law form of primordial spectrum (hereafter baseline), provides
the origin and evolution of the Universe entirely. Therefore, it is possible to test the
model against all observations. We expect that the bounds on the model parameters
from different observations lie within the distance of a few standard deviations.
Tension between the datasets arise when the distance becomes very large. The bottom
plot of Fig. 34.1 describes tensions between observations in the determination of
parameter . A.
At this point, let’s be more specific about the model. The standard model is
based on Einstein’s theory of general relativity (GR). It assumes a homogeneous
and isotropic Universe at large scales (supported by observations). Ordinary matter
(standard particles with mass), radiation (photon) and unknown type of matter (dark
matter, interacts gravitationally) and unknown type of energy (dark energy with neg-
34 Anomalies and Tensions in Cosmology and a Primordial Solution 641

Fig. 34.1 [Top] Example of anomaly in a dataset. [Bottom] Example of tension between datasets

ative pressure) are the ingredients of this Universe. Dark matter and dark energy are
two very crucial ingredients that we must take seriously as they claim about 95%
of the entire energy budget. Moreover we assume the negative pressure of the dark
energy follows an equation of state given by pressure = .−energy density. This partic-
ular dark energy is known as the Cosmological Constant, .⌃. With these constituents
of the Universe, the background part of the standard model is known as .⌃Cold Dark
Matter (.⌃CDM) model. We referred to this as the background part, since we have
not yet discussed the fluctuations (the structures that we observe in the Universe).
These constituents dictate how a homogeneous and isotropic Universe should evolve
from an early stage till today. Solving Einstein’s GR, we understand that at very early
stages the Universe expanded with deceleration governed by radiation and thereafter
by matter. At present we are in a phase of dark energy driven acceleration. We also
assume a flat Universe in the standard model (consistent with several observations).
642 D. K. Hazra and A. Shafieloo

Now, let’s understand the fluctuations. The distribution of structures that we observe
in the sky such as the galaxies, clusters etc. follows certain power spectrum. It tells us
how the structures are correlated at different length scales. This power spectrum has
evolved since its initial form through the evolution of the Universe and non-linear
gravitational collapse. Mathematically we can think of this evolution as a differential
equation. The differential equation can be solved with initial conditions. Therefore
the natural question is: what was the initial power spectrum of the fluctuations? The
seed of these structures has a quantum origin and we need to know the spectrum of
these initial (primordial) quantum fluctuations. Without going into the details we can
assume the spectrum to have a simple form that can be described with an amplitude
and a tilt (power law). So across the length scales if we plot the spectrum it would
look like a tilted line (as of now, we do not have any statistically compelling reason to
go beyond a tilted line). If we combine these background and initial condition—a flat
.⌃CDM model with power law form of primordial spectrum is known as the standard
model of the Universe which is used as the baseline of all our analyses. There-
fore an anomaly or tension would mean inconsistencies with respect to the baseline
predictions. This baseline model is parametrized with 6 parameters—4 describing
background and 2 describing initial conditions (amplitude and tilt). The 4 background
parameters are baryon and cold dark matter densities, today’s Hubble parameter (a
measure of how two points in the Universe are receding from each other per unit
time per unit distance) and the optical depth (free electron fraction integrated along
the line of sight).
We will now be more specific about the observations in question. We will discuss
three types of observations:

• The Cosmic Microwave Background (CMB) anisotropy observation from

Planck satellite [1].
• Today’s Hubble parameter measurement from S. H0 ES [2].
• Large Scale Structure (LSS) surveys from Kilo-Degree Survey (KiDS [3])
and from the Dark Energy Survey (DES [4]).

Planck satellite mapped the CMB sky that captured the anisotropies in photon
temperature and polarization to an unprecedented accuracy upto a very small scales
corresponding to .0.07◦ (this is the resolution for temperature map, polarization maps
has lower resolution). The power spectrum of the temperature anisotropy is plotted
in Fig. 1 of [1]. With the data, the best fit baseline is plotted in green and at the bottom
panel the data, residual to the best fit are also plotted to show the deviations. Here,
we notice an oscillatory pattern between angles .l = 800 − 2000 from the residual
plot. At larger angular scales we also see other patterns but here we will not talk
about those here as they are not directly relevant to the tensions in this study.
These residual patterns can be generated if we take the map of CMB fluctuations
and smooth it in a particular way, which smears the original signals to some extent.
This smearing effect is similar to the one produced by weak gravitational lensing of
34 Anomalies and Tensions in Cosmology and a Primordial Solution 643

the CMB. One can show that if an unphysical lensing amplitude (. Alens ) parameter
is introduced, and when it is more than 1 (in standard model, this must be 1), this
excess can be explained. In Planck analysis [1] the significance of this parameter
is reported to be nearly 3 standard deviations. This implies that inclusion of this
parameter rejects the theoretical consistency of . A L = 1. Since this is an unphysical
parameter, we turn to other physical sources to explain this excess smearing. If we
relax the assumption that our Universe is flat, then also we can reject the standard
model by better addressing the data and a closed Universe picture emerges (see, [1]).
However, the closed Universe picture is also unacceptable since it invites unavoidable
problems with initial conditions, and certain crucial cosmological parameters in this
picture become inconsistent with all other cosmological measurements [7].
Now let is discuss the tensions between the cosmological observations w.r.t. the
standard model. We turn to matter dominated epoch that follows the radiation dom-
inated the Universe. During this epoch, the gravitational attraction brought the par-
ticles together and bound them to form structures. A structure (say, galaxy) that
is closer to us, distorts the image of a background structure through gravitational
lensing. Observations of these structures, lensing and their correlations also help us
understand the Universe by constraining the model of the Universe. Here we will only
talk about a particular parameter called . S8 which is a combination of matter density
and the .σ8 which is related to the amplitude of the matter power spectrum (primor-
dial power spectrum multiplied by a transfer function). It refers to the root mean
square fluctuations of matter density within a spherical volume. The radius of that
sphere is indicated by the number 8 in .h −1 Mpc units. Surveys like the Kilo-Degree
Survey and The Dark Energy Survey have observed galaxies over areas of 1500
and 5000 square degrees. Millions of galaxy observations help us in understanding
gravitational lensing and galaxy clustering, which, in turn, put bounds on the model
parameters. Hence, we get bounds on .σ8 and matter density or equivalently on . S8 .
The bounds obtained from KiDS and DES (Dark Energy Survey Year 3 Results are
completely consistent. However, the bound derived from Planck CMB data shows
an inconsistency at a level of 2–3 standard deviations, a tension of moderate level
between CMB and LSS observations.
The most significant tension in cosmology, however, is noticed between the direct
and indirect determinations of the Hubble parameter. In the indirect determination,
CMB is the most precise observation which determines. H0 = 67.4 ± 0.5 km/s/Mpc.
This .0.7% determination of Hubble parameter comes from CMB temperature, polar-
ization and lensing data in Planck analysis [1]. On the other hand, local Hubble
measurements from the Cepheid and Supernovae observations, where S. H0 ES col-
laboration [2] reports a the constraint . H0 = 73.04 ± 1.04 km/s/Mpc which shows
a .5σ tension with the CMB estimated constraint. In the prologue of this book “The
Hubble Tension, Book Prologue: A Perspective Along Several Axes”, plots the ten-
sion between the direct and indirect measurements of . H0 .
644 D. K. Hazra and A. Shafieloo

34.3 The Idea of Primordial Solution

In this chapter, since we discuss the search for a common solution to the anomalies
and tensions in cosmology, we keep this section for the historical development of
primordial solution.
The attempt to capture CMB anomalies through features in primordial spectrum is
not new and several models are available in the literature. Models have been designed
to capture outliers at different scales in the angular power spectra of temperature
and polarization [8–19]. While there are many interesting models, it will be fair to
comment that we have not yet been able to find any statistically significant (more
than .2σ ) feature candidate from the CMB data.
Apart from forward modelling of features in the spectrum, direct model indepen-
dent reconstruction techniques have been useful to identify the location and type
of features [20–22]. In this context, using Planck data, Modified Richardson-Lucy
(MRL) method was introduced to reconstruct the primordial spectrum assuming the
baseline best-fit background cosmology [23, 24]. In Sect. 3.2 of [24], degeneracy
between lensing effects and primordial features was explored in different Planck
spectra of temperature anisotropy. It was found that certain oscillations can mimic
the lensing signal to some extent. In Fig. 34.2 we quote the relevant plots from the
publication that shows a damped oscillations in the primordial spectrum, peaking
around 0.1/Mpc scales can generate limited lensing-like signals in the angular power
spectrum.
Later, the role of primordial features in alleviating . H0 , S8 tensions was explored
in [25, 26]. In [25] MRL was used to find features in the primordial spectrum that
can address the CMB data as good as the baseline, but with the mean . H0 of S. H0 ES
measurements. In particular, here the reconstruction of power spectrum assumes a
background cosmology that is not the best fit baseline model but a higher . H0 (keep-
ing other parameters fixed). The reconstructed power spectrum contained oscillatory
signals as demonstrated in Fig. 34.3 (top). It is needless to mention that this compli-
cated form of power spectrum is difficult to parameterize and therefore, it is difficult
to realize this in a model on inflation or bouncing cosmology. However, it is inter-
esting to note that the power spectrum contains damped oscillations peaking around
.0.1/Mpc, similar to what was obtained in [24]. When this spectrum is used as the

primordial power spectrum instead of power law in fitting CMB data, we find sub-
stantial increase in the Hubble parameter and decrease in the matter density (bottom
left plot). At the same time, we notice a significant decrease in the disagreement in
the estimation of . S8 (see, .Ωm − σ8 plot in the bottom right).
Therefore, although the realization of the reconstructed spectrum seems implau-
sible from theoretical modelling, it can be appreciated that certain features in the
spectrum can indeed act as a common solution to the anomalies (by mimicking the
lensing signal to certain extent) and tensions (similar primordial signals being able
to resolve . H0 and . S8 tensions). In the next section we discuss the reconstruction of
One Spectrum that combined the efforts of [24] and [25].
34 Anomalies and Tensions in Cosmology and a Primordial Solution 645

Fig. 34.2 Form of primordial power spectrum (top) that mimics the effect of lensing (bottom) to
certain extent

34.4 One Spectrum reconstruction

In order to find a spectrum that simultaneously resolves the anomalies in the CMB
data and the tensions between datasets, we use MRL reconstruction. The methodol-
ogy has been elaborated in [27]. Here we outline certain key points.
The reconstruction is based on the temperature anisotropy power spectrum only,
since the anomalies and tensions are not existent with the polarization data. The
646 D. K. Hazra and A. Shafieloo

Fig. 34.3 [Top] Form of primordial power spectrum that prefers higher . H0 in CMB. [Bottom left]:
Shifts in matter density and Hubble parameter when the reconstructed power spectrum is used.
[Bottom right]: Shifts in the matter density and .σ8 . The size of the contours in case of reconstructed
power spectrum fit reduces as we use a fixed power spectrum and do not marginalize over the
spectral tilt

angular power spectrum of temperature anisotropy (.ClTT ) is a convolution of transfer

function (.G TT
lk ) dependent on the background parameters and the primordial power
spectrum (.Pk ) as:
∑
kmax
.Cl = Glk Pk
TT
(34.1)
k=kmin

The Richardson-Lucy deconvolution algorithm works iteratively and given an

initial power spectrum .Pki it calculates the .i + 1’th spectrum as:
34 Anomalies and Tensions in Cosmology and a Primordial Solution 647

Fig. 34.4 One Spectrum: The reconstructed power spectrum that mimics excess smoothing

lmax
( )
∑ Cldata − Cl
theory(i)
.Pk
i+1
− Pki = Pki × G˜lk theory(i)
(34.2)
l=lmin Cl

This algorithm converges the maximum likelihood solution and it was been
demonstrated in earlier publication [22] that this algorithm, independent of the initial
choice of spectrum deconvolves to the same resulting spectrum.
In obtaining One Spectrum, the choice of .Cldata was of crucial importance. In all
earlier works [22–24], .Cldata was chosen to be the actual data from CMB. There,
after reconstruction, post processing was necessary to identify the possible features
in the power spectrum. In [27], however, the method aims for a smooth primordial
spectrum apriori, and therefore, instead of using the data as in [25], we use the
best fit angular spectrum from lensing amplitude extension of the baseline (base-
line+. Alens ). The main idea behind this is to project the excess smoothing found in the
extended model onto the primordial spectrum. Note that in Eq. 34.2, the .Cldata refers
to unlensed spectrum. Here, to obtain the best fit unlensed spectrum, the lensing
contribution corresponding to . Alens = 1 was subtracted so that the excess lensing
contribution remains in the unlensed spectrum. Next the deconvolution reconstructs
the One Spectrum, containing the primordial oscillations that mimics the excess
smoothing in CMB temperature anisotropy spectrum, as plotted in Fig. 34.4.
Interestingly, compared to [25], One Spectrum appears to be much simpler and
can be parametrized easily. The reconstructed spectrum, allowing an amplitude shift
can be compared with the Planck temperature and polarization data. To begin with,
we present the analysis with. Alens and.ΩK separately against Planck temperature data.
Since the spectrum is designed to mimic lensing effect, we expect that the spectrum
can bring back cosmic concordance by solving lensing anomaly. Figure 34.5 plots
648 D. K. Hazra and A. Shafieloo

Fig. 34.5 [Top] Analysis with . Alens extension. Note that when One Spectrum is used, the lensing
amplitude becomes consistent to 1, with significant improvement in fit compared to the Standard
Model (baseline). [Bottom] Analysis including curvature parameter variation. Reconstructed spec-
trum brings back flat Universe .ΩK = 0, also with significant improvement in fit to the data

the results of such analyses. Note that both . Alens = 1 and .ΩK = 0 comes within
1σ bounds from the mean, when reconstructed One Spectrum is used instead of
.
power law primordial spectrum. Moreover, we notice the distribution of .χ 2 moves
significantly towards the left of standard model.
Analyses including Planck CMB polarization are detailed in [27]. Since the feature
is dominantly located at small scales, we do not expect polarization data to constrain
it better, given their low signal-to-noise ratio.
Apart from removing anomalies within the CMB data, we find that One Spectrum
√ matter density (.Ωm ). Due to lower .Ωm , the spectrum
prefers a higher . H0 and lower
also prefer lower . S8 = σ8 Ωm /0.3. In Fig. 34.6 we plot the shifts in the parameters.
The oscillations in One Spectrum provide the appropriate excess smoothing in the
data compared to the baseline standard model. This in turn, requires certain shifts in
the background parameter space. In fitting Planck temperature data (P18TT: denotes
Plik high-.l TT + low-.l TT + simall-low-E likelihood), we find, . H0 = 69.21 ± 0.50
and . S8 = 0.786 ± 0.015, while for Planck temperature+polarization data (P18TP:
denotes Plik high-.l TTTEEE + low-.l TT + simall-low-E likelihood), they shift to
34 Anomalies and Tensions in Cosmology and a Primordial Solution 649

Fig. 34.6 Shifts in the parameters . H0 (left) and . S8 (right) when One Spectrum is used instead of
power law primordial spectrum

H0 = 69.01 ± 0.44 and . S8 = 0.798 ± 0.013. Similar to the analysis in [25], here
.

too the constraints are tighter than the power-law baseline model, since the form of
the spectrum used to fit here is fixed and in power law form the tilt is allowed to
vary. Therefore, it should not be concluded from this analysis that the model is more
constraining than the baseline model. Rather, we should note that One Spectrum,
while providing a better fit to the data, prefers . Alens = 1 and .ΩK = 0 and a high . H0
and lower . S8 , which was our main goal of the reconstruction.

34.5 Parametrizing One Spectrum

In this section, we discuss the parametric form of the spectrum that captures the main
features of One Spectrum. To assess the significance of the oscillations, parametriza-
tion is necessary. A four parameter model of primordial spectrum can be expressed
as, [ ]
α1 sin (ω(k − k0 ))
.PRestricted (k) = PPower Law (k) 1 + , (34.3)
1 + β(k − k0 )4

with .α1 as amplitude, .k0 as position, .ω as frequency and .β as damping of the oscil-
lations. This power spectrum can now used instead of numerical One Spectrum for
the parameter estimation and the overall spectral amplitude and tilt can be varied
through .PPower Law .
In this parameter estimation, we will focus on two numbers, the significance of
.α1 . The significance refers to the percentage with which the marginalized posterior
distribution of .α1 rejects .α = 0. Simply, it denotes the significance of preference for
the oscillations compared to a featureless spectrum. Secondly, we also note the Bayes’
factor, or the ratio of the evidences between the model with and without oscillatory
features. In our convention, a positive logarithm of Bayes’ factor indicates preference
650 D. K. Hazra and A. Shafieloo

Table 34.1 Logarithm of the Bayes’ factor w.r.t. to the baseline model (first row) and the signifi-
cance of oscillations (second row)
Model/Data P18TT P18TT+HST P18TP P18TP+HST
Parametrized .−0.58 ± 0.52 .3.4 ± 0.53 .−0.01 ± 0.54 .1.46 ± 0.55
model
85% 99.9% 95% 99.5%

for the oscillatory model. In Table 34.1 we list the logarithm of Bayes’ factor (in first
row) and the evidences (in second row) for different dataset combinations. Only
with CMB data we notice 1-2.σ preference for oscillations. The significance increase
when Hubble prior from S. H0 ES is used.
When HST prior is used, while . H0 posterior is expected to shift towards higher
values, we also find . S8 moves towards lower values, making the model more consis-
tent to the lensing and clustering surveys.
The marginalized posterior distributions of the parameters describing the oscilla-
tions are plotted in Fig. 34.7. The dashed lines represent the CMB only results while
the solid lines are obtained upon using the HST prior. In CMB only analyses, when
polarization is added we get marginal improvement in significance for the features,
indicating a better fit to both temperature and polarization data.
This parametrized power spectrum has been tested with several other combina-
tions of datasets. Small scale CMB temperature and polarization data from the Ata-
cama Cosmology Telescope are in agreement with the oscillations. Clustering and
lensing data from the DES also support the lower of. S8 preference of this parametriza-
tion. Summary statistics from the Baryon Acoustic Oscillations data and distance

Fig. 34.7 Significance of the feature parameters. 2 .× 2 plots on the left side shows results with
CMB temperature while the plots on the right side demonstrates the effect of adding polarization
data
34 Anomalies and Tensions in Cosmology and a Primordial Solution 651

Fig. 34.8 Correlation function (left) and the matter power spectrum (right) obtained from the best
fit primordial power spectrum parametrized in Eq. 34.3

modulus from Supernovae data jointly decrease the significance of these features as
they prefer higher matter density.
Since the spectrum exhibits oscillations at a particular frequency (300 Mpc as
can be understood from Fig. 34.7), the correlation function has a bump and a dip
around that length scales as shown in Fig. 34.8. In this figure we also plot the matter
power spectrum. With ongoing and future large scale structure observations such
as Euclid [28], DESI [29] and LSST [30], significance of these oscillations can be
tested jointly with CMB.
We close this section by summarizing that the features in the primordial spectrum
while fitting the CMB data, prefers changes in the cosmological parameters to cer-
tain extent compared to the baseline mean values. Therefore, it brings CMB closer
to the other observations. Given the degeneracy between the background and the
power spectrum parameters, we also observe an increase in the spectral tilt. Since,
we have access to the small scale CMB data from ACT, joint estimations are use-
ful for consistency check with this increased spectral tilt. The ACT data prefers a
higher spectral tilt even in the baseline parametrization. Therefore, the analysis with
P18TP+ACT+DES data, the spectral tilt is found to be .0.9742 ± 0.0035. This shift
in spectral tilt is crucial to model inflationary potentials that we discuss in the next
section.

34.6 A Model of Inflation

The simple parametrization of One Spectrum opens up the window to explore an

inflationary solution. We model the modification in inflationary dynamics over the
standard slow roll inflationary models. This work has been discussed in the letter [31].
Note that for a potential, the spectral tilt (.n s ) and tensor-to-scalar ratio (.r ) in models
652 D. K. Hazra and A. Shafieloo

of inflation are usually fixed (for example, in quadratic model, we find .n s ∼ 0.96
and .r ∼ 0.16). Variations in these two observables are possible only within limited
windows for a fixed potential form. Theoretically, they are also correlated within an
inflationary model. In the last section we mentioned the importance of change in .n s .
Therefore the restrictions imposed by the choice potential form make the potential
parametrization unsuitable for the parameter estimation with combined datasets.
In order to allow free variations in the .n s − r alongside other inflationary param-
eters, we use a baseline slow-roll parametrization given by,

∈ baseline (N ) = ∈1 exp [∈2 (N − N∗ )]

. H (34.4)

Here, .∈1 , ∈2 are free parameters and . N∗ is chosen to be 20 e-folds. Note that we
impose the pivot scale .k = 0.05/Mpc leaves the Hubble radius 50 e-folds before the
end of inflation which directly relates . N∗ with the pivot scale. This parametrization
is equivalent to the power law parametrization in the power spectrum as the scalar
power spectrum amplitude, tilt and the tensor-to-scalar ratio can be varied through
the variation of . Hi2 /∈1 (. Hi represents the Hubble parameter at the onset of inflation),
.∈1 and .∈2 . Therefore, it would be fair to say that this slow roll parametrization allows

us to marginalize over all minimal slow roll potential parametrizations that generate
power law primordial spectrum without running and higher derivatives of spectral
index. The power spectrum parameters can be approximated as,

n ≅ 1 − 2∈1 − ∈2 ,
. s

r ≅ 16∈1 (34.5)

Note, however, we do not use slow roll approximations to solve for the primordial
spectrum and BINGO [32] is used to solve the Mukhanov-Sasaki equations with
Bunch-Davies initial condition.
Using Eq. 34.4 as the baseline model, we add modifications as,
( )
α cos [ω(N − N0 )]
∈ (N ) = ∈ H
. H
baseline
(N ) 1 + , (34.6)
1 + β(N − N0 )2

which introduces oscillatory phase (both local and non-local) in the slow roll dynam-
ics with amplitude.α and frequency .ω. The scalar field rolling slowly down the poten-
tial temporarily fast rolls around . N = N0 . A non-zero .β parameter helps to damp the
oscillation around that e-folds. The power spectrum resulting from such inflationary
dynamics produces an envelope of oscillations in the order of sinusoidal (.sin(ω̃k)),
sinusoidal logarithmic (.sin(ω̃ log k)) and sinusoidal forms.
We compare this inflationary dynamics with combinations of Planck, BICEP-
KECK18 (BK18) and S. H0 ES data. BK18 data is directly relevant here as in this case
we are constraining inflationary dynamics where BK18 data, jointly with Planck
provides an upper bound on .∈1 through the upper bounds on the B-mode anisotropy
spectrum. The Planck temperature and polarization data has been used in the com-
34 Anomalies and Tensions in Cosmology and a Primordial Solution 653

Fig. 34.9 Best fit primordial power spectra to different data combinations

binations of TT, TEEE, TTTEEE (TP). Lensing likelihood has also been used in a
separate analysis with TP.
Markov Chain Monte Carlo through CosmoMC [33] has been used for parameter
estimation and BOBYQA minimization [34] starting from MCMC minimum has
been used to find the best fit for each data combination. Selected best fits are plotted
in Fig. 34.9. All the best fit to different power spectra are consistent to each other
in frequency and phases and they significantly overlap with the One Spectrum. The
best fit feature model spectrum also prefers a higher spectral tilt than baseline as can
be noticed in all plots where temperature auto-correlation data has been used.
In Fig. 34.10 the best fit residuals of feature models compared to the baseline
best fit are plotted in blue. In red, the best fit baseline+. A L residuals are plotted. The
oscillations in the primordial spectrum mimic the excess smoothing effect obtained
from ad-hoc lensing amplitude extension (red). Interestingly, in the polarization
spectrum, these two models exhibit out of phase property. Therefore, with future small
scale CMB polarization observations (projected CMB-S4 error-bands are plotted in
EE and TE panels) it will be possible to distinguish them.
The posteriors of the inflationary dynamics are plotted in Fig. 34.11. Here the 1-3.σ
bounds on the evolution of the slow roll parameters (top panel), Hubble parameter
(bottom left) and the reconstructed potential (bottom right) are plotted. We highlight
only the fractional change by factoring out the baseline model. The potential is
obtained using .V (φ[N ]) = 3MPl 2
H [N ]2 (1 − ∈ H [N ]/3) as a function of e-folds. The
relevant oscillations can be noticed for 0.5 e-folds.
The numerically reconstructed potential can be well approximated by the follow-
ing functional form,

ΔV (φ) α cos[ωφ (φ − φ0 )]
. = , (34.7)
Vbaseline (φ) 1 + β(φ − φ0 )2

where .Vbaseline (φ) represents a baseline potential satisfying strict slow roll inflation.
A damped sinusoidal oscillation in the potential introduces a temporary fast roll
phase which, in turn, creates wiggles in the primordial spectrum with characteristic
oscillations similar to One Spectrum.
654 D. K. Hazra and A. Shafieloo

Fig. 34.10 Power spectrum best fit, residual from the baseline best fit

Fig. 34.11 Constraints on inflationary dynamics obtained using Planck (temperature + polarization)
+ BK18 + S. H0 ES (S21) data
34 Anomalies and Tensions in Cosmology and a Primordial Solution 655

34.7 Summary and Outlook

Anomalies and tensions within and between cosmological datasets are not new find-
ings. In several occasions, apart from statistical fluctuations, systematic uncertainties
were found to be responsible for these. However, in certain cases, anomalies and ten-
sions were found to be indicative of new physics beyond the standard model. The
articles presented in this chapter build the possibility of a non-standard inflationary
dynamics as a common answer to the anomalies and tensions. At the time of writing
this chapter, the envelope of oscillations in the primordial spectrum is the only such
common solution. While compared to other models this primordial solution does not
completely resolve the Hubble tension, importantly, it shifts the cosmological param-
eters in directions that reduces the tensions in major parameters. Moreover, the shifts
are obtained by fitting only to the CMB data, with noticeable improvement in fit com-
pared to the baseline, resulting in 2.σ support for the oscillatory features. Addition of
other datasets such as DES or S. H0 ES data increases the extents of the shifts further.
The oscillations being generated in the inflationary epoch do not interfere with the
expansion histories in the radiation, matter and dark energy dominated epochs and
therefore do not degrade the constraints on parameters compared to the baseline. The
uncertainties in the estimation of parameters (such as . H0 , n s , S8 ) increase by 10%
compared to their baseline estimations. This is an important characteristic to note in
the model building exercise where the discordance is reduced due to the shift in the
posterior volume and not by the enlargement of posterior volume by significantly
reducing the predictability.
Given the ongoing and future observations, observables calculated from this pri-
mordial solution can soon be tested. With the present CMB three-point statistics [35],
primordial bispectrum can be compared. With the small scale CMB data, the small
scale oscillations can be probed in the polarization. The matter power spectrum and
the correlation function can be tested with upcoming large scale structure data. Finally
with future ground based CMB polarization observations, the primordial signal can
be distinguished from other possible effects.

Acknowledgements DKH acknowledge support through the India-Italy mobility program

INT/Italy/P-39/2022 (ER). D.K.H. is supported by CEFIPRA Grant No. 6704-4.

References

1. N. Aghanim et al., Planck 2018 results. VI. Cosmological parameters. Astron. Astrophys. 641,
A6 (2020) [Erratum: Astron.Astrophys. 652, C4 (2021)]
2. Adam G. Riess et al., A comprehensive measurement of the local value of the Hubble constant
with 1 .km s −1 M pc−1 uncertainty from the Hubble space telescope and the S. H0 ES team.
Astrophys. J. Lett. 934(1), L7 (2022)
3. Marika Asgari et al., KiDS-1000 cosmology: cosmic shear constraints and comparison between
two point statistics. Astron. Astrophys. 645, A104 (2021)
656 D. K. Hazra and A. Shafieloo

4. L.F. Secco et al., Dark energy survey year 3 results: cosmology from cosmic shear and robust-
ness to modeling uncertainty. Phys. Rev. D 105(2), 023515 (2022)
5. Xiaolei Li, Arman Shafieloo, A simple phenomenological emergent dark energy model can
resolve the Hubble tension. Astrophys. J. Lett. 883(1), L3 (2019)
6. Vivian Poulin, Tristan L. Smith, Tanvi Karwal, Marc Kamionkowski, Early dark energy can
resolve the Hubble tension. Phys. Rev. Lett. 122(22), 221301 (2019)
7. Eleonora Di Valentino, Alessandro Melchiorri, Joseph Silk, Planck evidence for a closed Uni-
verse and a possible crisis for cosmology. Nature Astron. 4(2), 196–203 (2019)
8. Xingang Chen, Richard Easther, Eugene A. Lim, Generation and characterization of large
non-Gaussianities in single field inflation. JCAP 04, 010 (2008)
9. Raphael Flauger, Liam McAllister, Enrico Pajer, Alexander Westphal, Xu. Gang, Oscillations
in the CMB from axion monodromy inflation. JCAP 06, 009 (2010)
10. Xingang Chen, Folded resonant non-Gaussianity in general single field inflation. JCAP 12, 003
(2010)
11. M. Aich, D.K. Hazra, L. Sriramkumar, T. Souradeep, Oscillations in the inflaton potential:
complete numerical treatment and comparison with the recent and forthcoming CMB datasets.
Phys. Rev. D 87, 083526 (2013)
12. Jennifer A. Adams, Bevan Cresswell, Richard Easther, Inflationary perturbations from a poten-
tial with a step. Phys. Rev. D 64, 123514 (2001)
13. D.K. Hazra, M. Aich, R.K. Jain, L. Sriramkumar, T. Souradeep, Primordial features due to a
step in the inflaton potential. JCAP 10, 008 (2010)
14. D.K. Hazra, A. Shafieloo, G.F. Smoot, A.A. Starobinsky, Wiggly whipped inflation. JCAP
1408, 048 (2014)
15. Jens Chluba, Jan Hamann, Subodh P. Patil, Features and new physical scales in primordial
observables: theory and observation. Int. J. Mod. Phys. D 24(10), 1530023 (2015)
16. M. Ballardini, F. Finelli, C. Fedeli, L. Moscardini, Probing primordial features with future
galaxy surveys. JCAP 10, 041 (2016). [Erratum: JCAP 04, E01 (2018)]
17. D.K. Hazra, D. Paoletti, M. Ballardini, F. Finelli, A. Shafieloo, G.F. Smoot, A.A. Starobin-
sky, Probing features in inflaton potential and reionization history with future CMB space
observations. JCAP 02, 017 (2018)
18. D.K. Hazra, D. Paoletti, I. Debono, A. Shafieloo, G.F. Smoot, A.A. Starobinsky, Inflation story:
slow-roll and beyond. JCAP 12(12), 038 (2021)
19. Akhil Antony, Shweta Jain, Effect of damped oscillations in the inflationary potential. Eur.
Phys. J. C 82(8), 687 (2022)
20. William Hadley Richardson, Bayesian-based iterative method of image restoration.∗. J. Opt.
Soc. Am. 62(1), 55–59 (1972)
21. L.B. Lucy, An iterative technique for the rectification of observed distributions. Astron. J. 79,
745 (1974)
22. Arman Shafieloo, Tarun Souradeep, Primordial power spectrum from WMAP. Phys. Rev. D
70, 043523 (2004)
23. D.K. Hazra, A. Shafieloo, T. Souradeep, Primordial power spectrum: a complete analysis with
the WMAP nine-year data. JCAP 07, 031 (2013)
24. D.K. Hazra, A. Shafieloo, T. Souradeep, Primordial power spectrum from Planck. JCAP 11,
011 (2014)
25. D.K. Hazra, A. Shafieloo, T. Souradeep, Parameter discordance in Planck CMB and low-
redshift measurements: projection in the primordial power spectrum. JCAP 04, 036 (2019)
26. MiaoXin Liu, Zhiqi Huang, Band-limited features in the primordial power spectrum do not
resolve the Hubble tension. Astrophys. J. 897, 166 (2020)
27. D.K. Hazra, A. Antony, A. Shafieloo, One spectrum to cure them all: signature from early
universe solves major anomalies and tensions in cosmology. JCAP 08(08), 063 (2022)
28. G.D. Racca, R. Laureijs, L. Stagnaro, J.-C. Salvignol, J.L. Alvarez, G.S. Criado, L.G. Venancio,
A. Short, P. Strada, T. Bönke, C. Colombo, A. Calvi, E. Maiorano, O. Piersanti, S. Prezelus,
P. Rosato, J. Pinel, H. Rozemeijer, V. Lesna, P. Musi, M. Sias, A. Anselmi, V. Cazaubiel, L.
Vaillon, Y. Mellier, J. Amiaux, M. Berthé, M. Sauvage, R. Azzollini, M. Cropper, S. Pottinger,
34 Anomalies and Tensions in Cosmology and a Primordial Solution 657

K. Jahnke, A. Ealet, T. Maciaszek, F. Pasian, A. Zacchei, R. Scaramella, J. Hoar, R. Kohley,

R. Vavrek, A. Rudolph, M. Schmidt, The Euclid mission design. 9904, 99040O (2016)
29. A. Aghamousa et al., The DESI experiment part I: science, targeting, and survey design, 10
(2016)
30. Željko Ivezić et al., LSST: from science drivers to reference design and anticipated data prod-
ucts. Astrophys. J. 873(2), 111 (2019)
31. A. Antony, F. Finelli, D.K. Hazra, A. Shafieloo, Discordances in cosmology and the violation
of slow-roll inflationary dynamics. Phys. Rev. Lett. 130(11), 111001 (2023)
32. D.K. Hazra, L. Sriramkumar, J. Martin, BINGO: a code for the efficient computation of the
scalar bi-spectrum. JCAP 05, 026 (2013)
33. Antony Lewis, Sarah Bridle, Cosmological parameters from CMB and other data: a Monte
Carlo approach. Phys. Rev. D 66, 103511 (2002)
34. M. Powell, The bobyqa algorithm for bound constrained optimization without derivatives.
Technical Report, Department of Applied Mathematics and Theoretical Physics, 01 (2009)
35. Y. Akrami et al., Planck 2018 results. IX. Constraints on primordial non-Gaussianity. Astron.
Astrophys. 641, A9 (2020)
 Part V
Complications with Theoretical Solutions
Chapter 35
The Tension in the Absolute Magnitude
of Type Ia Supernovae

David Camarena and Valerio Marra

35.1 Introduction

In the study of cosmology, one of the key quantities of interest is the Hubble constant,
denoted as . H0 . It represents the current rate of expansion of the universe and plays
a crucial role in determining its age and evolution. Over the years, various methods
have been employed to measure . H0 , including observations of the cosmic microwave
background (CMB), baryon acoustic oscillations (BAO), and the use of astronomical
distance indicators such as Type Ia supernovae and Cepheids variables.
In recent years, the tension between different measurements of . H0 has garnered
significant attention. While the Planck satellite mission [1], which analyzed the CMB,
reported a value of . H0 around 67 km/s/Mpc, independent measurements using the
local distance ladder technique, such as the S. H0 ES collaboration [2], have consis-
tently obtained higher values, closer to 73 km/s/Mpc. This discrepancy, significant
at the 5.σ level and often referred to as the “Hubble tension,” has led to considerable
interest in exploring cosmological models that can reconcile these conflicting results.
Here, we aim to understand the implications of the local calibration of the lumi-
nosity of Type Ia supernovae. After reviewing the basic background in Sect. 35.2, in
Sect. 35.3 we will start with the 46-dimensional analysis by S. H0 ES and extend it to an
arbitrary cosmographic model. This will allow us to obtain a determination of local
. H0 that only assumes the validity of the standard flat .ɅCDM model, and another one
that only assumes the FLRW metric, that is, large-scale homogeneity and isotropy.

D. Camarena (B)
Department of Physics and Astronomy, University of New Mexico, Albuquerque, NM 87106,
USA
e-mail: dacato115@gmail.com
V. Marra
Núcleo de Astrofísica e Cosmologia & Departamento de Física, Universidade Federal do Espírito
Santo, 29075-910 Vitória, ES, Brazil
e-mail: valerio.marra@me.com

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 661
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_35
662 D. Camarena and V. Marra

We proceed in Sect. 35.4 to highlight the importance of considering the calibration

of the supernova magnitude . M B during cosmological inference, emphasizing the
necessity to test the consistency of a given cosmological model with respect to . M B .
Finally, in Sect. 35.5 we discuss the .“M B tension,” the fact that the supernova lumi-
nosity either gets calibrated by CMB and BAO observations or by the first two rungs
of the cosmic distance ladder, providing inconsistent constraints. Our conclusions
are presented in Sect. 35.6.

35.2 Curved .wCDM Cosmography

Given the luminosity distance .d L , the apparent magnitude is given by:

[ ]
d L (z)
m B (z) = 5 log10
. + 25 + M B , (35.1)
1Mpc

and the distance modulus by:

μ(z) = m B (z) − M B ,
. (35.2)

where . M B is the absolute magnitude of Type Ia supernovae. The luminosity distance

is a prediction of a cosmological model and, assuming the FLRW metric, is:
[ ʃ z ]
c −1/2 1/2 d z̄
d (z) = (1 + z)
. L Ωk0 sinh Ωk0 , (35.3)
H0 0 E(z̄)

where . H = ȧ(t)/a(t) describes the expansion history of the universe and .a(t) is the
scale factor. For a curved .wCDM model it is:

H 2 (z)
. = E 2 (z) = Ωm0 (1 + z)3 + Ωde0 (1 + z)3(1+w) + Ωk0 (1 + z)2 , (35.4)
H02

where we neglected radiation as it is inconsequential at redshifts lower than 100. In the

previous equation the subscript 0 denotes the present-day value of the corresponding
quantity, .Ωm0 is the density parameter for cold dark matter (zero pressure), .Ωde0 the
one for a dark energy fluid with equation of state parameter .w, and .Ωk0 = −kc2 /H02
represents the spatial-curvature contribution to the Friedmann equation (.a(t0 ) = 1).
It is .Ωm0 + Ωde0 + Ωk0 = 1.
Alternatively, using a cosmographic approach one has [3]:
[ ]
cz 1 − q0 1 − q0 − 3q02 + j0 − Ωk0 2
d (z) =
. L 1+ z− z + O(z 3 ) , (35.5)
H0 2 6
35 The Tension in the Absolute Magnitude of Type Ia Supernovae 663

where the Hubble constant . H0 , the deceleration parameter .q0 and the jerk parameter
j are defined, respectively, according to:
. 0

| | ... |
ȧ(t) || −ä(t) || a (t) ||
. H0 = , q0 = 2 , j0 = 3 . (35.6)
a(t) |t0 H (t)a(t) |t0 H (t)a(t) |t0

Cosmography is a model-independent approach in the sense that it does not assume

a specific model as it is based on the Taylor expansion of the scale factor. However,
this does not mean that its parameters do not contain cosmological information. For
example, in the case of curved .wCDM cosmologies, .q0 and . j0 are given by:

Ωm0 1 + 3w flat 1 3 Ʌ 3 fid

q =
. 0 + Ωde0 = + w(1 − Ωm0 ) = Ωm0 − 1 = −0.55 , (35.7)
2 2 2 2 2
9 flat 9 Ʌ
. j0 = Ωde0 + Ωm0 + w(1 + w)Ωde0 = 1 + w(1 + w)(1 − Ωm0 ) = 1 , (35.8)
2 2
where in the last equality of (35.7) the concordance value of .Ωm0 = 0.3 was adopted.
Finally, when determining the luminosity distance, one should adopt both the
peculiar velocity-corrected CMB-frame redshift .z (also called Hubble diagram red-
shift) and the heliocentric redshift .z hel so that the luminosity distance is given by [4]:

1 + z hel
d (z, z hel ) =
. L d L (z) . (35.9)
1+z

35.3 Local Determination of . H0

The S. H0 ES collaboration measures the Hubble constant in the redshift range.0.023 <
z < 0.15, where the minimum redshift is large enough in order to reduce the impact of
cosmic variance by the local large-scale structure [5, 6] and the maximum redshift
is small enough to provide a late-time local measurement of the Hubble constant
and reduce the impact of cosmology in its determination. Ideally, one would want
a cosmology-independent local constraint. S. H0 ES traditionally fits the value of . H0
in (35.5), but fixes the values of the higher-order parameters to their fiducial values
of .q0fid = −0.55 and . j0fid = 1, using information from observations beyond the local
universe, like higher-.z supernovae.
In [7, 8] we obtained a local determination of . H0 without fixing the deceleration
parameter .q0 . Here, we will improve upon our previous results by extending the
analysis of [2] using their latest data.1 Specifically, we adopt the .χ 2 statistic:
( )T ( )
χ 2 (θ, q0 , j0 ) = y(q0 , j0 ) − Lθ C y−1 y(q0 , j0 ) − Lθ ,
. (35.10)

1 https://github.com/PantheonPlusSH0ES/DataRelease.
664 D. Camarena and V. Marra

where . y is the 3492-dimensional data vector, .θ is the 46-d parameter vector consid-
ered by S. H0 ES, . L is the 46.×3492 equation matrix, so that the model is . Lθ , and .C y is
the 3492.×3492 covariance matrix. The last elements of . y are about the 277 Hubble
flow supernovae (nhf) that are used to determine . H0 :

y nhf = m B,i − 5 log10 H0 d L − 25

. i (35.11)
[ ]
1 − q0 1 − q0 − 3q0 + j0 − Ωk0 2
2
= m B,i − 5 log10 cz i 1 + zi − z i − 25 ,
2 6

and in Eq. (35.10) we let the components . yinhf be functions of .q0 and . j0 .2
The baseline analysis by S. H0 ES adopts .χ 2 (θ, q0fid , j0fid ), so that . y is constant and
the model is linear. In this case the best-fit model and its covariance matrix are:
( )−1
θ (q0 , j0 ) = L T C y−1 L L T C y−1 y(q0 , j0 ) ,
. bf (35.12)
( T −1 )−1
.C θ = L C y L , (35.13)

from which we can note that the best fits depend on .q0 and . j0 , but their covariance
matrix does not. Using the properties of multi-variate Gaussian distributions (the
model is linear), it is then straightforward to plot the marginalized constraints on any
combination of parameters by considering the eigenvectors and eigenvalues of the
covariance matrix. This is equivalent to performing a Bayesian analysis with wide
flat priors, as done in [2]. In Fig. 35.1 we show the marginalized constraints on . M B
and . H0 .
If we now do not fix .q0 and . j0 , the model used in Eq. (35.10) becomes nonlinear
and a full Bayesian analysis is necessary. One possible approach is to approximate
the posterior as a Gaussian distribution via its Fisher approximation, . Fα,i j = 21 ∂α∂ i χ∂α j ,
2 2

where .α = {θ, q0 , j0 } is the 48-d parameter vector, and .Cα = Fα−1 is the covariance
matrix on the parameters (the inverse of the Fisher matrix).3 However, the posterior
will likely show a non-Gaussian character in the. j0 direction as the latter is expected to
be poorly constrained by supernovae at .z < 0.15. Consequently one should perform
a computationally demanding 48-d Bayesian analysis.
Fortunately, we can exploit the fact that the model is linear in all parameters but
.q0 and . j0 to substantially speed up the estimation of the posterior. We can indeed

marginalize analytically on all the parameters we are not interested in:

( ) −1
χ2
. marg (M B , H0 , q0 , j0 ) = χ 2 θbf (q0 , j0 ), q0 , j0 + V C M B ,H0
V, (35.14)
. V = {M B , H0 } − {M Bbf (q0 , j0 ), H0bf (q0 , j0 )} , (35.15)

where . M Bbf and . H0bf comes from the corresponding entries of .θbf (q0 , j0 ) [2], and
.C M B ,H0 is obtained by removing all the rows and columns of .C θ expect the ones

2 We set .Ωk0 = 0, but the latter is degenerated with . j0 , see discussion below.
Note that . 21 ∂θ∂ i χ∂θ j exactly gives .Cθ−1 as, in this case, the model is linear.
2 2
3
35 The Tension in the Absolute Magnitude of Type Ia Supernovae 665

Fig. 35.1 Marginalized 68.3 and 95.4% credible level constraints on the supernova Ia absolute
magnitude. M B , the Hubble constant. H0 and the deceleration parameter.q0 from the 48-dimensional
analysis of Eq. (35.14) using the latest data from the S. H0 ES collaboration, that is, the first two rungs
of the distance ladder together with the 277 Hubble flow supernovae in .0.023 < z < 0.15. Blue
shows the S. H0 ES baseline analysis, red the analysis that only assumes the flat .ɅCDM model and
promote .q0 (or, equivalently, .Ωm0 ) to a free parameter, and the empty contours show the analysis
that only assumes the FLRW metric, that is, large-scale homogeneity and isotropy (both .q0 and . j0
are free, the latter is marginalized over)

relative to . M B and . H0 , as this operation is equivalent to marginalizing the posterior

over all the other linear parameters.
In Fig. 35.1 and Table 35.1 we compare the baseline results by S. H0 ES with our
results. It is interesting to note that the analysis relative to .q0 free and . j0 = 1 provides
a very competitive constraint that is independent of the particular .ɅCDM model as it
assumes just flat FLRW metric and cosmological constant. It features a 1.5% uncer-
tainty, as compared with the 1.4% uncertainty of the baseline analysis. In particular,
marginalizing over .q0 is equivalent to marginalizing over .Ωm0 . Indeed, their relation
666 D. Camarena and V. Marra

Table 35.1 Marginalized constraints (mode and 68% uncertainty) for the three cosmographic
analyses of Fig. 35.1. The first line corresponds to the baseline analysis by S. H0 ES [2], providing
the same constraints. The second analysis gives a constraint that is marginalized over any flat.ɅCDM
model. The last analysis only assumes the FLRW metric. Following [2], an additional systematic
error of 0.3 km/s/Mpc was added to the uncertainty of . H0
Case . M B (mag) . H0 (km/s/Mpc) .q0

.q0 = −0.55, . j0 = 1 .−19.253 ± 0.029 .73.04 ± 1.04 (1.4%) –

.q0 free, . j0 = 1 .−19.253 ± 0.029 .73.14 ± 1.10 (1.5%) .−0.61 ± 0.18
.q0 and . j0 free .−19.252 ± 0.029 .74.56 ± 1.61 (2.2%) .−2.0 ± 1.2

is linear, so that the same wide flat prior is adopted, and the relative difference of
.ɅCDM with respect to the corresponding cosmographic model, see Eq. (35.7), is
−6
.≈ 10 at .z = 0.15.
Next, by adopting an improper flat prior for . j0 , one obtains a measurement that
only assumes the FLRW metric, that is, large-scale homogeneity and isotropy. It is
interesting to note that in Eq. (35.5) the degenerate combination . j0 − Ωk0 appears
so that by marginalizing over . j0 , we are effectively marginalizing also over spatial
curvature. The data poorly constrain this combination, giving . j0 − Ωk0 = 28+39−28 .
Finally, we found that the deceleration parameter is in good agreement with the
value relative to the standard model, contrary to what we found in [7, 8] with the
Pantheon and Supercal datasets. The difference is likely due to the improved modeling
of uncertainty and redshift corrections in the Pantheon+ dataset [9].

35.4 Cosmological Inference with the Local Prior on . M B

The latest combined data release by the S. H0 ES and Pantheon+ collaborations sig-
nificantly improved the robustness and consistency of cosmological analyses that
include a local determination of the Hubble constant. Now, the distance calibration
from 37 Cepheids is provided for the 42 supernovae in the Cepheid host galaxies,4
together with the combined supernova and Cepheid covariance matrix [9]. This pro-
vides an absolute calibration for the supernova Ia absolute magnitude . M B as the
likelihood is now given by:
{
m B,i − M B − μ(z i ) if Hubble diagram supernova,
. Vi =
m B,i − M B − μceph,i if supernova in Cepheid host,
( )−1
χsne
2
= V T ∑sne + ∑ceph V , (35.16)

4 Some galaxies hosted more than one supernova. Note also that some supernovae have been
observed by more than one telescope so that in Pantheon+ there are actually 47 supernova entries
that are calibrated by Cepheid distances.
35 The Tension in the Absolute Magnitude of Type Ia Supernovae 667

where .∑ceph has nonzero entries only for the supernovae in Cepheid hosts. The
“Hubble diagram” (HD) supernovae are the 1466 supernovae—corresponding to
1580 measurements—that are used to constrain cosmological models but not to
calibrate . M B . These have a redshift .z > 0.01, again to remove the cosmic variance
from the local structure. The 42 supernovae (47 light curves) in Cepheid host galaxies
are instead used to calibrate . M B , whose calibration is then propagated to the HD
supernovae. This breaks the degeneracy between . M B and the .5 log10 H0 term inside
the distance modulus .μ(z), see Eq. (35.2).
The above methodology is ideal as it includes the correlations between Cepheid
distances and calibrating and HD supernovae, ensuring an accurate representative-
ness of the cosmic distance ladder in cosmological inference [7, 8, 10–12]. Indeed,
if one just uses HD supernovae, together with a local prior on . H0 , one will (i) double
count the 277 low-redshift supernova measurements that were used to obtain the prior
on . H0 , (ii) assume the validity of cosmography, in particular fixing the deceleration
parameter to the standard model value of .q0 = −0.55, (iii) not include in the analysis
the fact that . M B is constrained by the local calibration, an information which would
otherwise be neglected in the analysis, biasing both model selection and parameter
constraints [8].
The approach of Eq. (35.16) allows to carry out a variety of analyses. For example,
one can only consider the 238 HD supernovae (but 277 supernova measurements)
that are used to determine the Hubble constant and determine. H0 according to various
fiducial models, that is, varying .q0 and . j0 in the case of cosmography or varying .Ωm0
and .w in the case of .wCDM models. This approach is equivalent to the one of the
previous Section, which only uses the quantities . L, . y and .C from S. H0 ES. Here, we
preferred to adopt the methodology of the previous Section because we noticed that
the Pantheon+ dataset produces a value of . H0 which is .0.5 km/s/Mpc higher than
the S. H0 ES baseline value when considering only the 277 low-redshift supernova
measurements that are used to determine the Hubble constant.5
The most interesting use of Eq. (35.16) is the possibility of constraining the Hubble
constant conditionally to a particular cosmological model, bypassing completely the
use of cosmography. In this case the tension in the high- and low-redshift constraint
on . H0 is traduced into a tension between the local calibration of . M B and the one
induced by the CMB via BAO observations. This highlights the important of using
the new Pantheon+ and S. H0 ES likelihood.
In order to offer an easy way to implement the absolute calibration of . M B , here
we apply the demarginalization method introduced in [7] to obtain an effective prior
on . M B . As target for the demarginalization procedure, we consider both the con-
straint . H0 = 73.6 ± 1.1 km/s/Mpc for the flat .ɅCDM model from [9] and a value
.0.5 km/s/Mpc lower, because of the discrepancy mentioned above when using the
Pantheon+ & S. H0 ES likelihood.6 We use the 1580 HD supernova entries and obtain:

5 We contacted the authors of [2] and [9]; according to Adam Riess there may be an inconsistency
between the production and application of the covariance matrix of Cepheid/SN Ia host distances
in the Pantheon+ fitting code. This issue is under investigation.
6 This approach was suggested by Adam Riess.
668 D. Camarena and V. Marra

Fig. 35.2 Comparison between the analysis that uses the Pantheon+ & S. H0 ES likelihood and the
one that uses the Pantheon+ likelihood with the Gaussian prior on . M B from Eq. (35.17), for the flat
.ɅCDM model (left) and the flat .wCDM model (right)

. M B = −19.243 ± 0.032 mag for H0 = 73.6 ± 1.1km/s/Mpc , (35.17)

. M B = −19.258 ± 0.032 mag for H0 = 73.1 ± 1.1km/s/Mpc . (35.18)

As a cross-check, using Eq. (35.17) together with the Pantheon+ likelihood, we

obtain a good agreement with the results from [9]:7

. H0 = 73.6 ± 1.1 Ωm0 = 0.333 ± 0.018 (ΛCDM) ,

H0 = 73.5 ± 1.1 Ωm0 = 0.306+0.063
−0.076 w = −0.88 ± 0.15 (wCDM) .

Figure 35.2 shows a full comparison between the analysis that uses the Pantheon+
& S. H0 ES likelihood and the one that uses the Pantheon+ likelihood with the Gaussian
prior on . M B . Constraints and correlations are reproduced accurately. Furthermore,
the use of the Gaussian prior has the advantage that one can marginalize analytically
over . M B , see [8] for details.

35.5 . M B Tension Between S. H0 ES and Planck

Several theoretical models have been proposed to resolve the discrepancy between
the CMB determination of the Hubble constant by Planck [1] and the local mea-
surement by S. H0 ES [2]. These models modify the .ɅCDM cosmology either in the
late or early universe. Early-time modifications of the standard cosmological model
typically accommodate a larger . H0 value by reducing the sound horizon scale .rs

7 Ref. [9] adopted a prior .Ωm0 > 0.1, affecting their results in that parameter space region.
35 The Tension in the Absolute Magnitude of Type Ia Supernovae 669

Table 35.2 The . M B tension. Top: marginalized constraints adopting the .ɅCDM model. The first
line corresponds to the results by [9], providing the same constraints. The second line shows the
constraints that are obtained when including the full CMB likelihood together with the latest BAO
measurements instead of the local prior on Cepheid host galaxies by S. H0 ES. Bottom: as above but
for the Pantheon dataset [17] and adopting the model-independent (inverse) distance ladder of [16]
.ɅCDM analysis . M B (mag) . H0 (km/s/Mpc) .Ωm0

Pantheon+ & S. H0 ES .−19.244 ± 0.029 .73.55 ± 1.02 .(±0.3) .0.332 ± 0.018

22
Pantheon+ & CMB + .−19.438 ± 0.007 .67.39 ± 0.19 .0.315 ± 0.002
BAO
Tension .6.5σ .5.7σ –
Model-independent .MB (mag) . H0 (km/s/Mpc) .q0
analysis
Pantheon & S. H0 ES 21 .−19.244 ± 0.037 .74.30 ± 1.45 .−0.91 ± 0.22
[8]
Pantheon & CMB + .−19.401 ± 0.027 .69.71 ± 1.28 .−1.09 ± 0.29
BAO [16]
Tension .3.4σ .2.4σ –

during recombination, whereas late-time modifications typically increase the Hub-

ble rate at low redshifts [13–15]. While no compelling evidence for physics beyond
the .ɅCDM model has been found, late-time modifications of the standard paradigm
face an additional problem: the . M B tension.
The local determination of the Hubble constant provided by the S. H0 ES collabo-
ration relies on a set of astrophysical distances that are used to calibrate the absolute
magnitude of supernovae. In [16] we built a model-independent version of the inverse
distance ladder, which uses BAO distances and a prior on .rs , coming from the Planck
2018 data release [1], to effectively calibrate the supernovae. Our analysis demon-
strated that the absolute magnitude obtained from the cosmic distance ladder [7] is
in tension with the underlying calibration on . M B provided by the combination of
CMB + BAO, see Table 35.2. Since the latter constitutes a standard ruler that relies
on the value adopted for .rs , the discrepancy in . M B suggests that, independent of
the physics, models that solely change the Hubble flow, while keeping a sound hori-
zon distance consistent with the one inferred from CMB under the assumption of a
.ɅCDM cosmology, will fail at explaining the discrepancy between the early and late
times.
Indeed, the inability of late-time models to adequately explain the Hubble tension
has been clearly demonstrated by several analyses [10, 11], including the analy-
sis presented in [8]. This particular analysis employs a toy model, referred to as
the ‘hockey-stick’ model, wherein the dark energy component rapidly undergoes
a phantom transitions at low redshift. This illustrates that the . M B tension must be
addressed in order to resolve the Hubble discrepancy effectively. As shown in [8],
when analyzed with a prior on . H0 together with CMB and BAO data, the hockey-
stick model, .hsCDM, provides a value of the Hubble constant consistent with the
670 D. Camarena and V. Marra

local observation. Nevertheless, the underlying calibration of . M B that is obtained in

the analysis still is in tension with the calibration inferred from the cosmic distance
ladder, showing that .hsCDM does not really restore the agreement between early and
late times distances. This issue, and spurious conclusions, can be avoided if instead
a prior on . M B is assumed.8 Indeed, the analysis of the .hsCDM model under the
assumption of this prior explicitly shows the failure of the model to alleviate the . H0
(. M B ) tension. Similar conclusions hold for void models, which try to explain the
local higher expansion rate by placing the observer inside a large void. As far as the
observed luminosity distance-redshift relation is concerned, these models are very
similar to the .hsCDM model and are ruled out for the very same reasons [18].
A category of models that potentially resolve the . M B tension encompasses those
that feature a transition in supernova brightness at the extremely low redshifts of
the first two rungs—specifically, before entering the Hubble flow with supernovae
in the range .0.023 < z < 0.15. This transition could be realized by a sudden shift
in the effective gravitational constant by approximately 10% around 50–150 million
years ago, consequently altering the Chandrasekhar mass and thereby the supernova
luminosity [19, 20]. While this proposal conveniently resolves the Hubble tension, it
also yields a series of testable predictions that could either support or disprove it [21].
Physical mechanisms that might trigger such an ultra-late gravitational transition
include a first-order scalar-tensor theory phase transition from an early false vacuum,
corresponding to the measured value of the cosmological constant, to a new vacuum
with lower or zero vacuum energy [22].
We update the current significance of the . M B tension using the latest data and
adopting the fiducial.ɅCDM model. Using BAO data from BOSS DR12 [23], eBOSS
ELG [24], eBOSS LRG [25, 26], SDSS MGS [27], and WiggleZ [28], high-.
TT+TE+EE, low-. TT, and low-. EE CMB data from Planck 2018 [1],9 and the
latest Pantheon+ & S. H0 ES release, we constrain the .ɅCDM model considering the
cosmological equivalents to the cosmic distance ladder (Pantheon+ & S. H0 ES data)
and the inverse distance ladder (Pantheon+ in combination with BAO and CMB).
The results of this analysis are presented in Fig. 35.3 and Table 35.2. We used Cos-
moSIS and CAMB [29, 30] to carry out the analyses. We find that the tension on . M B
exceeds the 6.σ level.

35.6 Conclusions

In this study, we have examined three distinct yet interconnected topics. Firstly,
we extended the analysis conducted by the S. H0 ES collaboration to determine the
local value of the Hubble constant. We broadened their 46-dimensional analysis to
cosmography with arbitrary values of the deceleration and jerk parameters.

8 Alternatively, one should build a likelihood containing all the rungs of the cosmic distance lad-
der [12]. For a more detailed discussion, see Sect. 35.4.
9 For simplicity, we used the compressed version of high-. CMB data.
35 The Tension in the Absolute Magnitude of Type Ia Supernovae 671

Fig. 35.3 Marginalized

constraints from the
cosmological equivalents to
the cosmic distance ladder
(Pantheon+ & S. H0 ES data)
and the inverse distance
ladder (Pantheon+ in
combination with BAO and
CMB) when assuming the
.ɅCDM model as fiducial
model of the universe. This
figures shows how the
tension on the Hubble
constant can be interpreted
as a tension in the calibration
of the supernova brightness

We found that the latest Pantheon+ & S. H0 ES data provides a competitive

constraint that solely assumes the validity of the standard flat .ɅCDM model:

. 73.14 ± 1.10 (1.5%) km/s/Mpc,

and a model-independent constraint that only assumes the FLRW metric:

74.56 ± 1.61 (2.2%) km/s/Mpc.

.

Next, we underscored the importance of considering the calibration of the super-

nova magnitude . M B by the first two rungs of the distance ladder when conducting
cosmological inference. Indeed, . M B is a crucial astrophysical parameter and should
not be integrated out, but examined to understand the consistency of a given cosmo-
logical model.
This discussion led us to the issue of the .“M B tension.” The supernova luminosity
either gets calibrated by CMB and BAO observations or by the first two rungs of the
cosmic distance ladder.

We found that, assuming the standard flat .ɅCDM model, the two constraints
on . M B are in tension at the 6.5.σ level.
672 D. Camarena and V. Marra

The discrepancy in. M B suggests that, independent of the physics involved, models
that solely change the Hubble flow, while maintaining a sound horizon distance
consistent with the one inferred from CMB, fail to explain the discrepancy between
the early- and late-time measurements of . H0 .

Acknowledgements We are grateful to Adam Riess for his valuable feedback regarding the S. H0 ES
data release. DC thanks the Robert E. Young Origins of the Universe Chair fund for its generous
support. VM thanks CNPq (Brazil, 307969/2022-3) and FAPES (Brazil, TO 365/2022, 712/2022,
976/2022, 1020/2022, 1081/2022) for partial financial support. The authors acknowledge the use
of the computational resources of the Sci-Com Lab of the Department of Physics–UFES.

References

1. Planck Collaboration, N. Aghanim et al., Planck 2018 results. VI. Cosmological param-
eters. Astron. Astrophys. 641, A6 (2020). https://doi.org/10.1051/0004-6361/201833910
arXiv:1807.06209 [astro-ph.CO]
2. A.G. Riess et al., A comprehensive measurement of the local value of the Hubble constant with 1
km s.−1 Mpc.−1 uncertainty from the Hubble space telescope and the S. H0 ES team. Astrophys. J.
Lett. 934, (1), L7 (2022). https://doi.org/10.3847/2041-8213/ac5c5b arXiv:2112.04510 [astro-
ph.CO]
3. J.-Q. Xia, V. Vitagliano, S. Liberati, M. Viel, Cosmography beyond standard candles
and rulers. Phys. Rev. D 85, 043520 (2012). https://doi.org/10.1103/PhysRevD.85.043520
arXiv:1103.0378 [astro-ph.CO]
4. J. Calcino, T. Davis, The need for accurate redshifts in supernova cosmology. JCAP 01, 038
(2017). https://doi.org/10.1088/1475-7516/2017/01/038 arXiv:1610.07695 [astro-ph.CO]
5. V. Marra, L. Amendola, I. Sawicki, W. Valkenburg, Cosmic variance and the measurement
of the local Hubble parameter. Phys. Rev. Lett. 110, 241305 (2013). https://doi.org/10.1103/
PhysRevLett.110.241305 arXiv:1303.3121 [astro-ph.CO]
6. D. Camarena, V. Marra, The impact of the cosmic variance on . H0 on cosmological
analyses. Phys. Rev. D98, 023537 (2018). https://doi.org/10.1103/PhysRevD.98.023537
arXiv:1805.09900 [astro-ph.CO]
7. D. Camarena, V. Marra, Local determination of the Hubble constant and the deceleration
parameter. Phys. Rev. Res. 2(1), 013028 (2020). https://doi.org/10.1103/PhysRevResearch.2.
013028 arXiv:1906.11814 [astro-ph.CO]
8. D. Camarena, V. Marra, On the use of the local prior on the absolute magnitude of Type Ia
supernovae in cosmological inference. Mon. Not. Roy. Astron. Soc. 504, 5164–5171 (2021).
https://doi.org/10.1093/mnras/stab1200 arXiv:2101.08641 [astro-ph.CO]
9. D. Brout et al., The Pantheon+ analysis: cosmological constraints. Astrophys. J. 938(2), 110
(2022). https://doi.org/10.3847/1538-4357/ac8e04 arXiv:2202.04077 [astro-ph.CO]
10. G. Benevento, W. Hu, M. Raveri, Can late dark energy transitions raise the Hubble con-
stant? Phys. Rev. D 101(10), 103517 (2020). https://doi.org/10.1103/PhysRevD.101.103517
arXiv:2002.11707 [astro-ph.CO]
11. G. Efstathiou, To . H0 or not to . H0 ?. Mon. Not. Roy. Astron. Soc. 505(3), 3866–3872 (2021).
https://doi.org/10.1093/mnras/stab1588 arXiv:2103.08723 [astro-ph.CO]
12. K.L. Greene, F.-Y. Cyr-Racine, Hubble distancing: focusing on distance measurements
in cosmology. JCAP 06(06), 002 (2022). https://doi.org/10.1088/1475-7516/2022/06/002
arXiv:2112.11567 [astro-ph.CO]
13. L. Knox, M. Millea, Hubble constant hunter’s guide. Phys. Rev. D 101(4), 043533 (2020).
https://doi.org/10.1103/PhysRevD.101.043533 arXiv:1908.03663 [astro-ph.CO]
35 The Tension in the Absolute Magnitude of Type Ia Supernovae 673

14. N. Schöneberg, G. Franco Abellán, A. Pérez Sánchez, S. J. Witte, V. Poulin, J. Lesgourgues,

The . H0 Olympics: a fair ranking of proposed models. Phys. Rept. 984, 1–55 (2022). https://
doi.org/10.1016/j.physrep.2022.07.001 arXiv:2107.10291 [astro-ph.CO]
15. P. Shah, P. Lemos, O. Lahav, A buyer’s guide to the Hubble constant. Astron. Astrophys.
Rev. 29(1), 9 (2021). https://doi.org/10.1007/s00159-021-00137-4 arXiv:2109.01161 [astro-
ph.CO]
16. D. Camarena, V. Marra, A new method to build the (inverse) distance ladder. Mon.
Not. Roy. Astron. Soc. 495, 2630–2644 (2020). https://doi.org/10.1093/mnras/staa770
arXiv:1910.14125 [astro-ph.CO]
17. D. Scolnic et al., The complete light-curve sample of spectroscopically confirmed SNe Ia from
Pan-STARRS1 and cosmological constraints from the combined Pantheon sample. Astrophys.
J. 859(2), 101 (2018). https://doi.org/10.3847/1538-4357/aab9bb arXiv:1710.00845 [astro-
ph.CO]
18. D. Camarena, V. Marra, Z. Sakr, C. Clarkson, A void in the Hubble tension? The end of the line
for the Hubble bubble. Class. Quant. Grav. 39(18), 184001 (2022). https://doi.org/10.1088/
1361-6382/ac8635 arXiv:2205.05422 [astro-ph.CO]
19. V. Marra, L. Perivolaropoulos, Rapid transition of Geff at zt.≃0.01 as a possible solution of the
Hubble and growth tensions. Phys. Rev. D 104(2), L021303 (2021). https://doi.org/10.1103/
PhysRevD.104.L021303 arXiv:2102.06012 [astro-ph.CO]
20. G. Alestas, D. Camarena, E. Di Valentino, L. Kazantzidis, V. Marra, S. Nesseris,
L. Perivolaropoulos, Late-transition versus smooth. H (z)-deformation models for the resolution
of the Hubble crisis. Phys. Rev. D 105(6), 063538 (2022). https://doi.org/10.1103/PhysRevD.
105.063538 arXiv:2110.04336 [astro-ph.CO]
21. L. Perivolaropoulos, Is the Hubble crisis connected with the extinction of dinosaurs? Universe
8(5), 263 (2022). https://doi.org/10.3390/universe8050263 arXiv:2201.08997 [astro-ph.EP]
22. E. Abdalla et al., Cosmology intertwined: a review of the particle physics, astrophysics, and cos-
mology associated with the cosmological tensions and anomalies. JHEAp 34, 49–211 (2022).
https://doi.org/10.1016/j.jheap.2022.04.002 arXiv:2203.06142 [astro-ph.CO]
23. BOSS Collaboration, S. Alam et al., The clustering of galaxies in the completed SDSS-III
Baryon oscillation spectroscopic survey: cosmological analysis of the DR12 galaxy sample.
Mon. Not. Roy. Astron. Soc. 470(3), 2617–2652 (2017). https://doi.org/10.1093/mnras/stx721
arXiv:1607.03155 [astro-ph.CO]
24. A. de Mattia et al., The completed SDSS-IV extended Baryon oscillation spectroscopic survey:
measurement of the BAO and growth rate of structure of the emission line galaxy sample from
the anisotropic power spectrum between redshift 0.6 and 1.1. Mon. Not. Roy. Astron. Soc.
501(4), 5616–5645 (2021). https://doi.org/10.1093/mnras/staa3891 arXiv:2007.09008 [astro-
ph.CO]
25. J.E. Bautista et al., The completed SDSS-IV extended Baryon oscillation spectroscopic survey:
measurement of the BAO and growth rate of structure of the luminous red galaxy sample from
the anisotropic correlation function between redshifts 0.6 and 1. Mon. Not. Roy. Astron. Soc.
500(1), 736–762 (2020). https://doi.org/10.1093/mnras/staa2800 arXiv:2007.08993 [astro-
ph.CO]
26. H. Gil-Marin et al., The completed SDSS-IV extended Baryon oscillation spectroscopic survey:
measurement of the BAO and growth rate of structure of the luminous red galaxy sample from
the anisotropic power spectrum between redshifts 0.6 and 1.0. Mon. Not. Roy. Astron. Soc.
498(2), 2492–2531 (2020). https://doi.org/10.1093/mnras/staa2455 arXiv:2007.08994 [astro-
ph.CO]
27. A.J. Ross, L. Samushia, C. Howlett, W.J. Percival, A. Burden, M. Manera, The clustering of the
SDSS DR7 main Galaxy sample—I. A 4 per cent distance measure at.z = 0.15. Mon. Not. Roy.
Astron. Soc. 449(1), 835–847 (2015). https://doi.org/10.1093/mnras/stv154 arXiv:1409.3242
[astro-ph.CO]
28. E.A. Kazin et al., The WiggleZ dark energy survey: improved distance measurements to z =
1 with reconstruction of the baryonic acoustic feature. Mon. Not. Roy. Astron. Soc. 441(4),
3524–3542 (2014). https://doi.org/10.1093/mnras/stu778 arXiv:1401.0358 [astro-ph.CO]
674 D. Camarena and V. Marra

29. A. Lewis, A. Challinor, A. Lasenby, Efficient computation of CMB anisotropies in

closed FRW models. Astrophys. J. 538, 473–476 (2000). https://doi.org/10.1086/309179
arXiv:astro-ph/9911177 [astro-ph]
30. C. Howlett, A. Lewis, A. Hall, A. Challinor, CMB power spectrum parameter degeneracies
in the era of precision cosmology. JCAP 04, 027 (2012). https://doi.org/10.1088/1475-7516/
2012/04/027 arXiv:1201.3654 [astro-ph.CO]
Chapter 36
CMB Anomalies and the Hubble Tension

William Giarè

36.1 CMB Anomalies: A Brief Multi-experiment Overview

The most recent measurements of the Cosmic Microwave Background temperature

and polarization anisotropies released by the Planck Collaboration [4] have achieved
sub-percent accuracy in extracting most of the cosmological parameters [5]. While
this is undoubtedly a blessing, it also represents a challenge: as precision improves,
any deviations or anomalies may become more statistically significant and potentially
point to tensions in our current understanding of the Universe [3, 20, 45].
One notable example is the higher lensing amplitude at about .2.8σ observed in
the Planck data [24, 25, 43]. Since more lensing is expected with more Cold Dark
Matter (CDM), the lensing anomaly immediately recast a preference for a closed
Universe, which in turn helps to explain other large-scale anomalies in the data, such
as the deficit of amplitude in the quadrupole and octupole modes. Consequently, the
final Planck indication for a closed Universe becomes very significant, reaching the
level of 3.4 .σ [23, 30].
Excitingly, the lensing and curvature anomalies, as well as any other observed
discrepancies, may indicate the presence of new physics beyond the standard.⌃CDM
model of cosmology. However, it is also entirely plausible that anomalies simply
reflect observational systematic errors and/or statistical fluctuations in the data. One
effective way to discriminate between the two possibilities is to compare the results
of multiple CMB measurements: any persistent deviations across independent probes
may strongly hint at the presence of new physics. Fortunately, this is now possible
thanks to the observations of the CMB angular power spectra recently released by
the Atacama Cosmology Telescope (ACT) [6, 17] and the South Pole Telescope
(SPT) [9, 12, 27]. While these experiments may not be as constraining as the Planck

W. Giarè (B)
School of Mathematics and Statistics, University of Sheffield, Hounsfield Road, Sheffield S3
7RH, UK
e-mail: w.giare@sheffield.ac.uk
© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024 675
E. Di Valentino and D. Brout (eds.), The Hubble Constant Tension, Springer Series
in Astrophysics and Cosmology, https://doi.org/10.1007/978-981-99-0177-7_36
676 W. Giarè

satellite, they are still able to produce cosmological bounds that are accurate enough
for independent tests of the Planck results [6–9].
Interestingly, both the ACT and SPT data have provided full support for the
inflationary prediction of a spatially flat Universe and for a lensing amplitude con-
sistent with .⌃CDM [6, 8, 9]. This suggests that both the curvature and lensing
anomalies may be explained in terms of statistical fluctuations or systematic errors.
However, the same datasets have revealed other relevant deviations from the standard
cosmological model. For instance, the the Data Release 4 of the Atacama Cosmol-
ogy Telescope shows preference for a unitary spectral index of primordial perturba-
tions (.n s = 1.009 ± 0.015) [6, 29], in agreement with the phenomenological scale-
invariant spectrum proposed by Harrison, Zel’dovich, and Peebles [33, 44, 51], and
in tension with Planck with a significance of .99.3% Confidence Level (CL). Other
mild yet relevant anomalies in the ACT data concern the value of the effective number
of relativistic degrees of freedom in the early Universe (. Neff = 2.46 ± 0.26) [6, 21],
in disagreement at more than 2 .σ with the expected value of . Neff = 3.04 predicted
by the Standard Model of particle physics for three active neutrinos [11]; a .∼ 2σ
preference for a positive running of the scalar spectral index of inflationary pertur-
bations (.αs ≡ dn s /d log k = 0.058 ± 0.028) [6, 28, 29]; and a .∼ 3σ indication in
favor of early Dark Energy [34, 46].
In turn, these deviations are not supported by Planck, which instead shows .∼ 8σ
evidence for .n s /= 1, a good agreement with the predictions of the standard model for
. Neff , and no evidence for Early Dark Energy [5, 34, 46]. Therefore, also in this case,
the lack of cross-validation between independent CMB measurements may suggest
statistical fluctuations and/or as of yet unknown systematic errors.
Regardless of whether these anomalies arise from observational systematic
effects, statistical fluctuations, or new physics, if we aim to achieve precision cosmol-
ogy and test fundamental physics, it is important to acquire a better comprehension
of the collective agreement between independent experiments and to have a clear
understanding of the limitations and uncertainties underlying both current data and
the cosmological model. This becomes a crucial need in relation to the. H0 -tension [13,
49] as many proposed solutions1 call for a new paradigm shift in cosmology, while
relying almost entirely on the resilience of such observations.
In this chapter, we address this problem and discuss the global consistency of
independent CMB experiments. In Sect. 36.2, we investigate potential reasons for
discrepancies by exploring both the limitations inherent in current data (Sect. 36.2.1)
and the uncertainties of the standard cosmological model (Sect. 36.2.2). We eventu-
ally argue that anomalies seem to indicate a potential mismatch between the early
and late Universe (Subsect. 36.2.3). In Sect. 36.3, we discuss the implications for the
. H0 -tension and its most common solutions.

1 See, e.g., Refs. [3, 26, 36, 39, 40, 46] for recent reviews.
36 CMB Anomalies and the Hubble Tension 677

36.2 Global Consistency of CMB Experiments

Based on the overview provided in the previous section, it is evident that what makes
CMB anomalies difficult to interpret individually is that different experiments often
point in discordant directions, and none of the most relevant deviations can be cross-
validated through independent probes. For this reason, accurate statistical methods
have been developed to quantify the global consistency between independent experi-
ments under a given model of cosmology, for instance by starting from the constraints
on the parameter space of the model itself [15, 16, 22, 31, 41, 42].
Assuming a standard .⌃CDM cosmology, the data release 4 of the Atacama Cos-
mology Telescope can be evaluated in mild-to-moderate tension with Planck and
the South Pole Telescope at a gaussian equivalent level of 2.6 .σ and 2.4 .σ , respec-
tively [22, 32]. Depending on one’s level of tolerance, these controversial tensions
ranging between 2 and 3 .σ may be taken more or less seriously. If our objective is to
accomplish “precision cosmology” and effectively assess fundamental physics, it is
reasonable to conclude that they warrant further investigations. However, it remains
crucial to understand the causes of these differences and shed light on their poten-
tial implications for the numerous proposed resolutions of the . H0 -tension that are
emerging from these observations.

36.2.1 The Limitations of Current Data

Assuming a baseline .⌃CDM cosmology, the main source of tension between ACT
and Planck arises from the measurements of the scalar spectral index, .n s , with Planck
reporting a value of .n s = 0.9649 ± 0.0044 [5] and ACT producing a value of .n s =
1.008 ± 0.015 [6]; in tension at approximately 2.7 .σ . Moreover, the discrepancy
between the two experiments extends to the measurement of the baryon energy
density with Planck and ACT giving respectively .Ωb h 2 = 0.02236 ± 0.00015 [5]
and .Ωb h 2 = 0.02153 ± 0.00030 [6]; in tension at about 2.5 .σ .
If we believe these differences to emerge from limitations in the data, then a logical
step is to identify which (missing) part of the dataset is responsible for the discrepancy
and eventually discard (include) it in the cosmological parameter inference analyses.
Several possible explanations have been attempted to explain the disagreement.

• Temperate data: One suggested explanation could be attributed to the limited

amount of ACT data at low multipoles in the TT Spectrum where ACT has a
minimum sensitivity in multipole of .l = 600 [6]. In absence of data around the
first two acoustic peaks, a strong degeneracy exists between .Ωb h 2 and .n s as a
lower value of the former can be mimicked by a larger value of the latter, tilting
the spectrum in the opposite direction [6]. Typically, the lack of information is
filled by using measurements from Planck or a third independent experiment such
as WMAP [10, 35] in the multipole range .l ≤ 600. While this helps in diluting
the tension, it is worth noting that there is no a-priori reason to combine ACT with
678 W. Giarè

low-.l data from another experiment: despite the degeneracy, accurate data at the
scales probed by ACT are enough to achieve precise constraints on both .n s and
.Ωb h . For instance, considering the Planck data only in the multipole range in
2

common with ACT, the final results still indicate .n s /= 1 at more than 4 .σ [29].
This cast doubts on the explanation of the mismatch in terms of lack of information
on the first acoustic peaks as one would expect a reasonable agreement between
two experiments measuring an overlapping range of multipoles.
• Polarization data: Another possibility is that the limited amount of ACT low
multipoles polarization data (which cover only multipoles .l ≥ 350 in TE and EE)
is responsible for the discrepancy. However, incorporating low-multipole polar-
ization data from the Planck measurements of E-modes at .2 ≤ l ≤ 30 fails to
explain these anomalies [29]. Nonetheless, it was also suggested that an over-
all TE calibration could resolve the mismatch [6]. Neglecting any information
from ACT polarization measurements (TE EE) and combining ACT temperature
anisotropies (TT) with SPT polarization data (TE EE) the tension is in fact reduced
below .2σ , restoring the agreement between the experiments for .Ωb h 2 [29]. How-
ever, the same combination of ACT and SPT data still favors a value of .n s around
unity [29], indicating that ACT polarization may only partially contribute to the
tension, without fully reconciling it.
• Non-CMB data: Yet another possible avenue to increase the data constraining
power is to break the geometrical degeneracy among cosmological parameters by
exploiting CMB-independent data, such as astrophysical measurements of Baryon
Acoustic Oscillation (BAO) and Redshift Space Distortion (RSD) [18, 19], as well
as galaxy clustering and cosmic shear observations [1, 2]. While astrophysical data
have been found to alleviate several tensions and restore the parameter values to
those predicted by .⌃CDM, this is not the case. Including them does not shift the
results significantly but leads to tighter errors, ultimately increasing the tension [6,
29].

36.2.2 The Unknowns of the Cosmological Model

Although observational systematic errors and/or statistical fluctuations in the data

may provide a simple explanation for the discrepancies between different CMB
experiments, we cannot discount the possibility that part of these anomalies are
rooted in the limitations of the standard cosmological model to accurately explain
precise observations at very different angular scales.
From a theoretical standpoint, many ingredients of .⌃CDM lack solid understand-
ing in fundamental physics. The nature of Dark Matter and Dark Energy remains
unknown in the context of the Standard Model of fundamental interactions as well
as the physical origin of the inflationary epoch is also unclear. To account for all
these unknowns, the .⌃CDM model adopts a principle of simplicity and incorporates
these components into a streamlined theoretical framework consisting of just six
36 CMB Anomalies and the Hubble Tension 679

Fig. 36.1 Constraints on cosmological parameters in various minimal extensions to.⌃CDM derived
from the analysis of the Planck 2018 [5] (grey) and ACT-DR4 [6] (red) temperature and polarization
data. The vertical dashed black lines indicate the reference values of parameters in.⌃CDM. The hor-
izontal magenta band in the 2D-plots involving the Hubble constant shows the direct measurements
obtained by the S. H0 ES Collaboration (. H0 = 73 ± 1 km/s/Mpc) [47]

free parameters. However, the value of cosmological parameters inferred from CMB
measurements clearly depends on the global parameterization adopted for the back-
ground cosmology and—ultimately—on the model’s assumptions. For this reason, a
possibility usually explored when finding anomalies in the cosmological parameter
values, is to extend the baseline model and study how the overall tension changes.
Figure 36.1 provides a comprehensive summary of the constraints on cosmological
parameters derived from Planck and ACT in various minimal extensions to .⌃CDM,
including those involving the anomalous parameters discussed so far. Directing our
680 W. Giarè

attention to the enigmatic foundations of the standard cosmological model, a careful

inspection on the figure can yield valuable insights into the uncertainties underlying
their theoretical parameterization and potentially reveal any missing ingredients in
the theory able to account for the observed discrepancies.

• Dark Energy: Within the standard cosmological model, the late-time accelerated
expansion of the Universe is attributed to a cosmological constant term in the
Einstein equations, resulting in a constant energy-density component with a neg-
ative equation of state, .w = −1. Relaxing this assumption we can use the Planck
and ACT data to directly constrain .w and test its consistency with the cosmo-
logical constant value. Interestingly, Planck suggests a phantom equation of state
(.w < −1) at about .2σ , while ACT is consistent with .w = −1, albeit with large
uncertainties that cannot exclude phantom or quintessential (.w > −1) behaviors,
see Fig. 36.1. In any case, late-time modifications can hardly reconcile the differ-
ence between the two experiments, as it mainly arises from a mismatch of the early
Universe.
• Dark Matter and Neutrinos: As the tension between Planck and ACT arises pri-
mary from a mismatch in the values of early-time parameters, interesting exten-
sions are those associated with the early Universe content, such as neutrinos, Dark
Matter and ∑Dark radiation. For instance, relaxing the value of total neutrino mass
(fixed at . m ν ∼ 0.06 eV within .⌃CDM) produces a shift in the value of .n s
measured by ACT towards the Planck result. However, such a shift can be asso-
ciated with the ACT’s peculiar preference for higher values of neutrino mass,
which contradicts the tight constraints placed by the Planck data. A second an
interesting avenue involves the radiation energy-density in the early Universe. In
the case of three active neutrinos, the radiation energy density can be calculated
quite accurately within the standard model of particle physics and parameterized
in terms of the effective number of relativistic degrees of freedom, . Neff = 3.044.
By relaxing this theoretical constraint and leaving . Neff as a free parameter to be
inferred by data, we can use the ACT and Planck observations to directly verify
the consistency with the Standard Model’s predictions as well as to test any the-
oretical extensions involving additional (thermal) degrees of freedom in the early
Universe. From Fig. 36.1 we see that Planck is in good agreement with the Stan-
dard Model predictions, while the ACT’s result (. Neff = 2.46 ± 0.26) is anomalous
and in disagreement with the expected value for three active neutrinos by about
2.5 .σ . Interestingly, when the energy density in the early Universe is significantly
lower than the expected value, the disagreement between the two experiments gets
diluted by the simultaneous effect of widening the uncertainties and shifting the
values of .n s towards that measured by Planck. The interpretation of this result
highly depends on our attitude regarding the reasons for the disagreement. If we
attribute the observed discrepancy between ACT and Planck to new physics, this
may suggest a potential role for the neutrino sector (and generally for new physics
at early times) in accounting for at least part of this difference. On the other hand,
if we rely primarily on the strength of the theoretical predictions, this may indicate
that observational systematic effects are introducing biases into the determination
36 CMB Anomalies and the Hubble Tension 681

of . Neff , raising concern on the results obtained when considering new physics at
early times.
• Inflation: One of the most common results from single-field models of inflation is
the production of nearly scale-invariant perturbations that can be well described by
a power-law form of the adiabatic scalar components. In the simplest scenarios, any
residual scale dependence is parameterized solely in terms of the scalar spectral
index .n s , where .n s = 1 corresponds to exact scale-invariance. However, additional
parameterizations that incorporate higher-order terms in the power-law expansion
should also be considered. In this regard, another important clue that can be inferred
from Fig. 36.1 is the tendency for the value of .n s predicted by ACT to decrease
significantly when extending the inflationary sector of the theory by accounting
also for the possibility of a running in the scalar spectral index .αs = dn s /d log k.
In this case, the difference on the spectral index .n s is solved, but the disagreement
maps into a strong controversy in the values of .αs : while Planck prefers a slightly
negative running (.αs = −0.0119 ± 0.0079), ACT favors a positive running (.αs =
0.058 ± 0.028).

36.2.3 How Many Early-Late Universe Mismatches Are

There?

To draw some general conclusions from the precedent analysis and gain a fresh
perspective on the CMB anomalies, we want to argue that the global tension emerging
between ACT and Planck always maps into mismatches between the early and late
Universe.
Considering all the minimal extensions in Fig. 36.1, we can notice that the Planck
anomalies always involve parameters associated with the local Universe such as the
lensing amplitude, the spacetime geometry, and the dark energy equation of state.
Instead, the ACT anomalies always involve parameters related to the early Universe
such as the baryon energy density, the spectral index, its running, and a free-to-vary
neutrino sector. Therefore, the picture emerging is the following one: Planck shows
mild-to-moderate deviations from .⌃CDM when extending the late time sector of the
theory, remaining consistent for extensions involving modifications at early-times. In
contrast, the Atacama Cosmology Telescope agrees pretty well with .⌃CDM at late-
times, but shows anomalies in the early Universe parameters. However, considering
the large experimental uncertainties obtained when extending the late-time sector
of the theory, the difference between the two probes remains mostly driven by a
mismatch in the early Universe.
682 W. Giarè

36.3 Implications for the . H0 -Tension

Any determination of the Hubble constant inferred from observations of the cosmic
microwave background necessarily depends on the cosmological model. For this
reason, several alternative models have been proposed to address the widely known
tension between the direct value of . H0 measured by the S. H0 ES collaboration using
luminosity distances of Type Ia supernovae [47] (. H0 = 73 ± 1 Km/s/Mpc) and the
indirect value inferred by the Planck satellite [5] (. H0 = 67.4 ± 0.5 Km/s/Mpc).
If we intend to explain the mismatch in terms of new physics (and clarify the
potential role of the CMB anomalies in this matter), it is important to understand what
aspects of the theory are involved in the game when inferring . H0 from observations
of the CMB. It is well known that measuring the positions of the acoustic peaks in the
CMB angular power spectrum, we can determine the angular size of the sound horizon
(.θs ) which is the cosmological parameter most accurately measured by Planck [5].
On the other hand, we can fix the baryon density .Ωb h 2 from the relative heights of
even and odd peaks, while we can obtain the cold dark matter density .Ωc h 2 from the
absolute heights of acoustic peaks. Once these parameters are known, we can pick
a model of the early Universe (i.e., pre-recombination) and calculate the physical
dimension of sound horizon (.rs ). Given .rs and .θs , we can easily infer the angular
diameter distance from the CMB, . D A (zls ) = rs /θs (with .zls ≅ 1100). Finally, by
assuming a cosmological model for the subsequent late-time expansion (i.e., post-
recombination), we eventually infer the value of . H0 from . D A (zls ).
Therefore, any theoretical attempts aimed at solving the . H0 -tension should fall
into one or both these categories:

• Late time solutions: The first possibility is to modify the late time expansion
history and change the value of the Hubble constant inferred from the angular
diameter distance . D A (zls ). This is typically done by introducing new late-time
physics and/or extending the dark sector of the theory. Based on Fig. 36.1, one
might expect these solutions to be preferred by data, given the significant room
left by the CMB observations for new physics at late-times. However, they should
be approached with some caution as late-time modifications often leads to a strong
geometric degeneracy among parameters which needs to be broken in order to
genuinely test new physics. In practice, this can be done by incorporating local
Universe observations, such as baryon acoustic oscillations [18, 19] and/or super-
novae measurements [14, 48]. They often clean away any late-time anomalies and
restore the parameter values to .⌃CDM. Therefore, contrary to what one might
expect from CMB data, there is very little room to accommodate new physics
at late-times. However, it is worth considering that the inclusion of certain local
probes becomes a subject of debate when testing theoretical models that signif-
icantly deviate from .⌃CDM. For instance, this objection is sometimes raised
against BAO measurements, which assume a .⌃CDM-like cosmology as an ansatz
from the outset, but are in tension with CMB in many extended theoretical frame-
works [23]. Therefore, taking a conservative approach, we cannot discount the
possibility that (part of) the . H0 -tension may be alleviated by late-time solutions.
36 CMB Anomalies and the Hubble Tension 683

Fig. 36.2 Constraints on the early Universe parameters and their relation with. H0 (in the color-bar)
as obtained by the ACT-DR4 data release [6]. The black empty contours define the region of the
parameter space preferred by the Planck 2018 data [5]

In any case, it is unlikely that the tension between ACT and Planck will have a
significant impact on these solutions since these experiments primarily disagree
at early times, as seen in Sect. 36.2.
• Early time solutions: The second possibility is to consider new physics in early
Universe to increase the physical size of the sound horizon .rs . This is typically
done by incorporating new light degrees of freedom before recombination. How-
ever, these solutions come with their own caveats. For instance, many indications of
this kind of new early-time physics arise when combining multiple CMB measure-
ments at different angular scales (such as Planck and ACT), without finding clear
cross-validation when these experiments are considered separately. Therefore, this
may cast doubt on whether these indications genuinely reflect new physics or if
they emerge from the controversial tension between these probes, as also seen in
684 W. Giarè

Sect. 36.2.2 for many simpler minimal early-Universe extensions. To explore this
issue further, we remain agnostic and refer to Fig. 36.2. Assuming a background
.⌃CDM cosmology, the figure depicts the constraints on the early Universe param-
eters of interest for early-time solutions, showing a consistent shift between ACT
and Planck. Notably, ACT allows for greater flexibility in accommodating higher
values of the sound horizon, while Planck peaks where ACT prefers very low
values of . H0 . This means that increasing . H0 requires moving towards the region
of the parameter space where the disagreement between the two probes becomes
more significant. In addition, the sound horizon, the spectral index, and the Hubble
constant are all positively correlated: increasing .rs to increase . H0 will naturally
push .n s towards higher values. This is why many potential solutions to the . H0
tension suggest a prominent role of .n s , often pointing towards values of .n s ∼ 1 as
measured by ACT, bringing the tension around this parameter back on the table,
see, e.g., Refs. [37, 38, 50]. However, one might argue that relaxing the early
Universe sector of the theory will change the correlation among parameters. As
seen in Sect. 36.2.2, considering a free-to-vary . Neff can partially accommodate the
disagreement between ACT and Planck. While this is true, the results in Fig. 36.1
show a significant shift of . Neff towards smaller values, which is in the opposite
direction of what we need to solve the . H0 tension. This also raises concerns about
some proposed solutions whose underlying philosophy consists of incorporating
large additional degrees of freedom as they appear to be disfavored in the first
place.

References

1. T.M.C. Abbott et al., Dark energy survey year 1 results: cosmological constraints from galaxy
clustering and weak lensing. Phys. Rev. D 98(4), 043526 (2018)
2. T.M.C. Abbott et al., Dark energy survey year 3 results: cosmological constraints from galaxy
clustering and weak lensing. Phys. Rev. D 105(2), 023520 (2022)
3. E. Abdalla et al., Cosmology intertwined: a review of the particle physics, astrophysics, and
cosmology associated with the cosmological tensions and anomalies. JHEAp 34, 49–211 (2022)
4. N. Aghanim et al., Planck 2018 results. I. Overview and the cosmological legacy of Planck.
Astron. Astrophys. 641, A1 (2020)
5. N. Aghanim et al., Planck 2018 results. VI. Cosmological parameters. Astron. Astrophys. 641,
A6 (2020). [Erratum: Astron.Astrophys. 652, C4 (2021)]
6. S. Aiola et al., The Atacama cosmology telescope: DR4 maps and cosmological parameters.
JCAP 12, 047 (2020)
7. K. Aylor et al., A comparison of cosmological parameters determined from CMB temperature
power spectra from the south pole telescope and the Planck satellite. Astrophys. J. 850(1), 101
(2017)
8. L. Balkenhol et al., Constraints on.⌃CDM extensions from the SPT-3G 2018 EE and TE power
spectra. Phys. Rev. D 104(8), 083509 (2021)
9. L. Balkenhol et al., A measurement of the CMB temperature power spectrum and constraints
on cosmology from the SPT-3G 2018 TT/TE/EE Data Set
10. C.L. Bennett et al., Nine-year Wilkinson microwave anisotropy probe (WMAP) observations:
final maps and results. Astrophys. J. Suppl. 208, 20 (2013)
36 CMB Anomalies and the Hubble Tension 685

11. J.J. Bennett, G. Buldgen, P.F. De Salas, M. Drewes, S. Gariazzo, S. Pastor, Y.Y.Y. Wong,
Towards a precision calculation of . Neff in the standard model II: neutrino decoupling in the
presence of flavour oscillations and finite-temperature QED. JCAP 04, 073 (2021)
12. B.A. Benson et al., SPT-3G: a next-generation cosmic microwave background polarization
experiment on the south pole telescope. Proc. SPIE Int. Soc. Opt. Eng. 9153, 91531P (2014)
13. J.L. Bernal, L. Verde, A.G. Riess, The trouble with . H0 . JCAP 10, 019 (2016)
14. D. Brout et al., The Pantheon+ analysis: cosmological constraints. Astrophys. J. 938(2), 110
(2022)
15. R. Calderón, A. Shafieloo, D.K. Hazra, W. Sohn, On the consistency of .⌃CDM with CMB
measurements in light of the latest Planck. ACT, and SPT data
16. T. Charnock, R.A. Battye, A. Moss, Planck data versus large scale structure: methods to quantify
discordance. Phys. Rev. D 95(12), 123535 (2017)
17. S.K. Choi et al., The Atacama cosmology telescope: a measurement of the cosmic microwave
background power spectra at 98 and 150 GHz. JCAP 12, 045 (2020)
18. K.S. Dawson et al., The Baryon oscillation spectroscopic survey of SDSS-III. Astron. J. 145,
10 (2013)
19. K.S. Dawson et al., The SDSS-IV extended Baryon oscillation spectroscopic survey: overview
and early data. Astron. J. 151, 44 (2016)
20. E. Di Valentino, Challenges of the standard cosmological model. Universe 8(8), 399 (2022)
21. E. Di Valentino, W. Giarè, A. Melchiorri, J. Silk, Health checkup test of the standard cosmo-
logical model in view of recent cosmic microwave background anisotropies experiments. Phys.
Rev. D 106(10), 103506 (2022)
22. E. Di Valentino, W. Giarè, A. Melchiorri, J. Silk, Quantifying the global ‘CMB tension’ between
the Atacama cosmology telescope and the Planck satellite in extended models of cosmology.
Mon. Not. Roy. Astron. Soc. 520(1), 210–215 (2023)
23. E. Di Valentino, A. Melchiorri, J. Silk, Planck evidence for a closed Universe and a possible
crisis for cosmology. Nature Astron. 4(2), 196–203 (2019)
24. E. Di Valentino, A. Melchiorri, J. Silk, Cosmological constraints in extended parameter space
from the Planck 2018 legacy release. JCAP 01, 013 (2020)
25. E. Di Valentino, A. Melchiorri, J. Silk, Investigating cosmic discordance. Astrophys. J. Lett.
908(1), L9 (2021)
26. E. Di Valentino, O. Mena, S. Pan, L. Visinelli, W. Yang, A. Melchiorri, D.F. Mota, A.G. Riess,
J. Silk, In the realm of the Hubble tension—a review of solutions. Class. Quant. Grav. 38(15),
153001 (2021)
27. D. Dutcher et al., Measurements of the E-mode polarization and temperature-E-mode correla-
tion of the CMB from SPT-3G 2018 data. Phys. Rev. D 104(2), 022003 (2021)
28. M. Forconi, W. Giarè, E. Di Valentino, A. Melchiorri, Cosmological constraints on slow roll
inflation: an update. Phys. Rev. D 104(10), 103528 (2021)
29. W. Giarè, F. Renzi, O. Mena, E. Di Valentino, A. Melchiorri, Is the Harrison-Zel’dovich spec-
trum coming back? ACT preference for .n s ∼ 1 and its discordance with Planck
30. W. Handley, Curvature tension: evidence for a closed universe. Phys. Rev. D 103(4), L041301
(2021)
31. W. Handley, P. Lemos, Quantifying tensions in cosmological parameters: interpreting the DES
evidence ratio. Phys. Rev. D 100(4), 043504 (2019)
32. W. Handley, P. Lemos, Quantifying the global parameter tensions between ACT. SPT and
Planck. Phys. Rev. D 103(6), 063529 (2021)
33. E.R. Harrison, Fluctuations at the threshold of classical cosmology. Phys. Rev. D 1, 2726–2730
(1970)
34. J.C. Hill et al., Atacama cosmology telescope: constraints on prerecombination early dark
energy. Phys. Rev. D 105(12), 123536 (2022)
35. G. Hinshaw et al., Nine-Year Wilkinson microwave anisotropy probe (WMAP) observations:
cosmological parameter results. Astrophys. J. Suppl. 208, 19 (2013)
36. K. Jedamzik, L. Pogosian, G.-B. Zhao, Why reducing the cosmic sound horizon alone can not
fully resolve the Hubble tension. Commun. in Phys. 4, 123 (2021)
686 W. Giarè

37. J.-Q. Jiang, Y.-S. Piao, Toward early dark energy and ns=1 with Planck, ACT, and SPT obser-
vations. Phys. Rev. D 105(10), 103514 (2022)
38. J.-Q. Jiang, G. Ye, Y.-S. Piao, Return of Harrison-Zeldovich spectrum in light of recent cos-
mological tensions
39. M. Kamionkowski, A.G. Riess, The Hubble tension and early dark energy
40. L. Knox, M. Millea, Hubble constant hunter’s guide. Phys. Rev. D 101(4), 043533 (2020)
41. A. La Posta, U. Natale, E. Calabrese, X. Garrido, T. Louis, Assessing the consistency between
CMB temperature and polarization measurements with application to Planck, ACT, and SPT
data. Phys. Rev. D 107(2), 023510 (2023)
42. W. Lin, M. Ishak, A Bayesian interpretation of inconsistency measures in cosmology. JCAP
05, 009 (2021)
43. P. Motloch, W. Hu, Tensions between direct measurements of the lens power spectrum from
Planck data. Phys. Rev. D 97(10), 103536 (2018)
44. P.J.E. Peebles, J.T. Yu, Primeval adiabatic perturbation in an expanding universe. Astrophys.
J. 162, 815–836 (1970)
45. L. Perivolaropoulos, F. Skara, Challenges for .⌃CDM: an update, 5 (2021)
46. V. Poulin, T.L. Smith, T. Karwal, The ups and downs of early dark energy solutions to the
Hubble tension: a review of models, hints and constraints circa (2023)
47. A.G. Riess et al., A comprehensive measurement of the local value of the Hubble constant with
1 km s.−1 Mpc.−1 uncertainty from the Hubble space telescope and the S. H0 ES team. Astrophys.
J. Lett. 934(1), L7 (2022)
48. D. Scolnic et al., The Pantheon+ analysis: the full data set and light-curve release. Astrophys.
J. 938(2), 113 (2022)
49. L. Verde, T. Treu, A.G. Riess, Tensions between the early and the late universe. Nature Astron.
3, 891 (2019)
50. G. Ye, J.-Q. Jiang, Y.-S. Piao, Toward inflation with ns=1 in light of the Hubble tension and
implications for primordial gravitational waves. Phys. Rev. D 106(10), 103528 (2022)
51. Y.B. Zeldovich, A hypothesis, unifying the structure and the entropy of the universe. Mon. Not.
Roy. Astron. Soc. 160, 1P-3P (1972)