Version June 19, 2025

Preprint typeset using LATEX style openjournal v. 09/06/15

ASTRONOMICAL CARDIOLOGY: A SEARCH FOR HEARTBEAT STARS USING GAIA AND TESS

Jowen Callahan,1 , D. M. Rowan 1,2 , C. S. Kochanek 1,2 , K. Z. Stanek 1,2 ,

1 Department
of Astronomy, The Ohio State University, 140 West 18th Avenue, Columbus, OH, 43210, USA
2 Center for Cosmology and Astroparticle Physics, The Ohio State University, 191 W. Woodruff Avenue, Columbus, OH, 43210, USA

Version June 19, 2025

ABSTRACT
Heartbeat stars are a subclass of binary stars with short periods, high eccentricities, and phase-folded light
arXiv:2506.14869v1 [astro-ph.SR] 17 Jun 2025

curves that resemble an electrocardiogram. We start from the Gaia catalogs of spectroscopic binaries and use
TESS photometry to identify 112 new heartbeat star systems. We fit their phase-folded light curves with an
analytic model to measure their orbital periods, eccentricities, inclinations, and arguments of periastron. We
then compare these orbital parameters to the Gaia spectroscopic orbital solution. Our periods and eccentricities
are consistent with the Gaia solutions for 85% of the single-line spectroscopic binaries but only 20% of the
double-line spectroscopic binaries. For the two double-line spectroscopic binary heartbeat stars with consistent
orbits, we combine the TESS phase-folded light curve and the Gaia velocity semi-amplitudes to measure
the stellar masses and radii with PHOEBE. In a statistical analysis of the HB population, we find that non-giant
heartbeat stars have evolved off the main sequence and that their fractional abundance rises rapidly with effective
temperature.
Subject headings: binaries: eclipsing – binaries: spectroscopic

1. INTRODUCTION and FX UMa (Wang et al. 2023). HBs are also discovered
Heartbeat stars (HB) are a relatively rare subclass of de- serendipitously in various TESS star samples (Murphy et al.
tached binary stars with short periods (𝑃 ≲ 100 days) and high 2020; Kochukhov et al. 2021; Sharma et al. 2022; Martín-
eccentricities (𝑒 ≳ 0.2). In these binaries, the stars are tidally Ravelo et al. 2024).
deformed near pericenter, causing periodic brightness varia- There have also been systematic surveys for HBs in TESS. By
tions. HBs are easily identifiable because their phase-folded analyzing the first dozen TESS sectors, Kołaczek-Szymański
light curves resemble an electrocardiogram (Thompson et al. et al. (2021) found 20 new HB systems through the visual
2012), and a simple analytic model (Kumar et al. 1995) can inspection of the phase-folded light curves of spectroscopic
be used to estimate the properties of the orbit. The tides can binaries. Li et al. (2024a,b) doubled the number of HBs found
also produce tidally excited oscillations (TEOs), waning pulsa- by deliberate searches in TESS by cross-matching and visually
tions of the stellar material that is observed between pericenter inspecting the TESS light curves of the stars in the PPM catalog
passages. (Röser & Bastian 1993; Röser et al. 1994; Nesterov et al.
Since the amplitude of the photometric variations in HBs 1995). Recently, Solanki et al. (2025) identified 180 HBs
is typically small, it was only with the Kepler mission that using the TESS full-frame images by applying convolutional
HBs were detected (Welsh et al. 2011). KOI-54 was one neural networks to TESS phase-folded light curves, followed
of the first HBs from Kepler, and it also had TEOs (Welsh by a visual inspection to confirm the candidates.
et al. 2011; Fuller & Lai 2012; Burkart et al. 2012; O’Leary In contrast to previous HB searches, which start from large
& Burkart 2014). However, HBs would not have their name samples of phase-folded light curves, here we start from the
until Thompson et al. (2012) applied the analytic Kumar et al. Gaia catalogs of double-line (SB2) and single-line (SB1) spec-
(1995) model to fit the Kepler phase-folded light curves of 17 troscopic binaries. This has the advantage of starting from a
new HB systems. Hundreds of HBs were identified in Kepler sample of known binaries, and we can then use the SB2 HB
(Hambleton et al. 2013; Kirk et al. 2016; Shporer et al. 2016). systems to estimate the masses and radii of the stars. In Sec-
These systems are all on the main sequence (MS) and have tion §2, we discuss the Gaia data and how we model and
periods of 10 ≲ 𝑃 ≲ 100 days. The Optical Gravitational search their phase-folded TESS light curves for HBs. Sec-
Lensing Experiment (OGLE) has discovered nearly 1000 HBs tions §2.1 and §2.2 describe our search strategies for HBs in
across the upper MS and giant branch (Wrona et al. 2022a,b). the Gaia SB2 and SB1 samples, respectively. In Section §3,
The Transiting Exoplanet Survey Satellite (TESS, Ricker we compare the orbital parameters we measure from the TESS
et al. 2015) provides another opportunity to discover short pe- phase-folded light curves to the Gaia orbital solutions. We
riod (𝑃 < 10 days) HBs. TESS observes the sky in twenty-six also model the SB2 HBs with consistent Gaia/HB orbital pa-
sectors which are continuously observed for roughly twenty- rameters to determine the masses and radii of the systems.
seven days. Like in Kepler, early detections of HBs in TESS Finally, Section §4 discusses the statistics of HBs and future
came as a byproduct of planetary searches (Wheeler & Kip- steps.
ping 2019). TESS photometry has also been used to identify
2. HEARTBEAT SEARCH
or characterize HB variability in known binaries like MA-
CHO 80.7443.1718 (Jayasinghe et al. 2019, 2021; Kołaczek- We start from the 181,529 SB1s and 5,376 SB2s (Gaia
Szymański et al. 2022), KIC 5006817 (Merc et al. 2021), Collaboration et al. 2023) in Gaia DR3. We used phase-
V680 Mon (Paunzen et al. 2021), WR 21a (Barbá et al. 2022), folded light curves from the TESS Quick-Look Pipeline (QLP,
Huang et al. 2020) for all of the Gaia SB1s and SB2s. The
2

Fig. 1.— Examples of heartbeat star phase-folded light curves created using the Kumar et al. (1995) model, varying the orbital eccentricities (e= 0.4, left three
columns and e= 0.6, right three columns), for a range of arguments of periastron (horizontal) and inclinations (vertical). This Figure is based on Figure 5 of
Thompson et al. (2012).

typical Gaia spectroscopic binary has four TESS sectors of

observations. 1 − 3 sin2 𝑖 sin2 (𝜈 + 𝜔)
𝐹 = 𝑍+𝑆 , (1)
(1 − 𝑒 cos 𝐸) 3
where 𝑆 sets the amplitude, 𝑍 is the mean flux, 𝜔 is the
2.1. SB2 Modeling argument of periastron, 𝑖 is the inclination, and 𝑒 is the orbital
Since there are only 5,376 Gaia SB2 targets, we identified eccentricity. The true anomaly is
HB candidates through visual inspection of the TESS phase- 2 tan (𝐸/2)
folded light curves. For all targets, we run a Lomb-Scargle 𝜈= √ √ , (2)
periodogram (LS, Lomb 1976; Scargle 1982) on each TESS tan−1 ( 1 + 𝑒/ 1 − 𝑒)
sector independently. We visually inspected the phase-folded
and the eccentric anomaly
light curves and identified 10 HB candidates. We then use
phase dispersion minimization (PDM, Stellingwerf 1978) to 2𝜋(𝑡 − 𝑡0 )
refine the orbital period. Before modeling the phase-folded 𝐸 − 𝑒 sin 𝐸 = . (3)
light curve, we select a single TESS sector for each target 𝑃
and remove long-term trends in the data, if necessary. For is determined by the period, 𝑃 and the epoch of periastron, 𝑡0 .
example, TIC 265056364 has a continuous decrease in the The shape of the phase-folded light curve varies significantly
flux over the course of the TESS sector, which could be due to with the orbital configuration, as shown in Fig. 1.
variable blended light or systematic trends. We remove these We used Markov Chain Monte Carlo (MCMC) methods as
signals with a linear fit. implemented by Foreman-Mackey et al. (2013) to sample over
We fit the phase-folded light curves using the Kumar et al. the parameters in the model. We ran the MCMC models with
(1995) analytic model. The flux is modeled as fourteen walkers, ten thousand iterations, and a burn-in of one
 3

1.004
TIC = 461247919 1.003 TIC = 316413777 1.008
TIC = 82985825
Residuals Normalized Flux

Residuals Normalized Flux

Residuals Normalized Flux

1.003 Ecc = 0.35 Ecc = 0.20 1.006 Ecc = 0.19
1.002 Period = 8.34 1.002 Period = 5.56 Period = 2.50
1.004
1.001
1.001 1.002
1.000
0.999 1.000
1.000
0.998 0.998
0.997 0.999 0.996
0.996 0.994
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
0.001 0.0005 0.004
0.0000 0.002
0.000 0.000
0.0005
0.001 0.002
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.0010 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
Phase
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
Phase Phase
TIC = 252588526 TIC = 204451102 TIC = 427458382

Residuals Normalized Flux

1.006
Residuals Normalized Flux

Residuals Normalized Flux

1.0015 Ecc = 0.40 1.008
Ecc = 0.48 Ecc = 0.32
Period = 4.24 Period = 9.02 Period = 3.40
1.004 1.006
1.0010
1.004
1.002 1.0005
1.002
1.000 1.0000 1.000
0.998 0.9995 0.998
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.0004 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
0.001 0.001
0.0002
0.000 0.000
0.0000
0.001 0.0002 0.001
0.002 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
Phase Phase Phase
1.008 TIC = 98552498 1.005 TIC = 76094846 TIC = 265056364
Residuals Normalized Flux

Residuals Normalized Flux

Residuals Normalized Flux

Ecc = 0.28
1.006 Period = 3.96 1.004 Ecc = 0.44
1.004 Ecc = 0.24
Period = 9.42 Period = 3.71
1.004 1.003 1.002
1.002 1.002
1.001 1.000
1.000
0.998 1.000
0.998
0.996 0.999
0.994 0.998 0.996
0.992 0.00 0.25 0.50 0.75 0.997
0.004 1.00 1.25 1.50 1.75 2.00 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
0.002 0.002
0.002
0.000
0.000 0.000
0.002
0.002
0.002 0.004
0.004 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
Phase Phase Phase
Fig. 2.— The phase-folded light curves (upper panel) and the model residuals (lower panel) for nine of the double-line spectroscopic heartbeat binaries. The
lines are the Kumar et al. (1995) models. The text boxes show the TESS input catalog (TIC), eccentricity, and period (days) of each target.

TABLE 1
The results of the light curve curve fit, Gaia score, period, and eccentricity, flags for whether the orbital parameters agree, the existence of eclipses, and TEOs
with the orbital harmonic (𝑛). The full table is available in the electronic version of the paper.
TIC Type Score incl (°) 𝜔 (°) P (days) Gaia P (days) P frac diff e Gaia e Gaia e err Match Eclipses TEOs 𝑛
1598625 SB1 0.442 36.130 232.130 3.252 3.241 0.004 0.330 0.404 0.096 ✓ ✗ ✗ −
9175061 SB1 0.476 55.720 141.410 2.900 2.900 0.000 0.160 0.147 0.034 ✓ ✗ ✓ 2
11578775 SB1 0.236 48.110 115.970 5.327 5.331 0.001 0.180 0.180 0.104 ✓ ✗ ✗ −
13933268 SB1 0.374 51.480 308.690 4.129 4.126 0.001 0.110 0.127 0.019 ✓ ✗ ✗ −
15257961 SB1 0.468 38.260 360.000 7.748 7.763 0.002 0.110 0.137 0.053 ✓ ✗ ✗ −
22567490 SB1 0.320 56.800 61.110 6.567 6.554 0.002 0.550 0.250 0.034 ✗ ✓ ✗ −
26412885 SB1 0.474 75.900 85.750 3.490 3.490 0.000 0.160 0.108 0.045 ✓ ✗ ✗ −
26689977 SB1 0.262 40.260 301.160 3.692 3.697 0.001 0.110 0.125 0.053 ✓ ✗ ✗ −
28543727 SB1 0.562 90.000 26.480 2.576 2.579 0.001 0.190 0.130 0.022 ✗ ✗ ✗ −
29451470 SB1 0.497 65.910 343.170 3.011 3.260 0.076 0.110 0.102 0.043 ✓ ✗ ✓ 2
29518898 SB1 0.376 90.000 306.270 4.649 4.628 0.004 0.130 0.163 0.074 ✓ ✗ ✗ −
31096993 SB1 0.219 33.590 356.980 6.019 6.018 0.000 0.160 0.191 0.031 ✓ ✗ ✗ −
42430804 SB1 0.239 45.180 216.590 8.998 0.403 21.314 0.560 0.168 0.060 ✗ ✗ ✗ −
42741990 SB1 0.336 37.350 319.100 5.785 5.817 0.005 0.150 0.094 0.102 ✓ ✗ ✗ −
44787510 SB1 0.720 22.360 360.000 3.553 1.860 0.910 0.470 0.096 0.019 ✗ ✗ ✗ −
61160125 SB1 0.138 48.110 345.120 5.960 5.908 0.009 0.220 0.396 0.100 ✗ ✗ ✗ −
63548665 SB1 0.286 50.370 298.410 4.770 1.124 3.243 0.190 0.057 0.047 ✗ ✗ ✗ −
65315466 SB1 0.554 90.000 226.240 3.196 3.196 0.000 0.190 0.232 0.086 ✓ ✗ ✗ −
73486038 SB1 0.771 57.060 85.460 6.445 6.451 0.001 0.260 0.302 0.048 ✓ ✗ ✗ −
 4
1.0 2.3. The Final Sample
Candidate Region
SB1 Heartbeat Stars Fig. 5 shows the orbital periods and eccentricities of the 112
0.8 SB2 Heartbeat Stars HBs detected in the Gaia SB1 and SB2 samples. Of these 112
HBs, four were found by Solanki et al. (2025) (TIC 252588526,
0.6 TIC 271554988, TIC 344586348, and TIC 370209445). The

N Targets
HB/ line

orbital periods range from 1.51 to 12.17 days, with a median

2 2

101
of 4.36 days. The eccentricities range from 0.10 to 0.57 (TIC
0.4 42430804), with a median of 0.20. All of our targets fall
below the period-eccentricity envelope expected from tidal
0.2 circularization (Mazeh 2008). Fig. 5 also shows the HBs from
Wrona et al. (2022a,b), Li et al. (2023), and Solanki et al.
(2025). As compared to the HBs identified in Wrona et al.
0.00.0 0.2 0.4 0.6 0.8 1.0 (2022a,b), our systems have much shorter periods than their
Orbital Eccentricity median value of 265.3 days. The median eccentricities are
similar (0.24 for OGLE). Our period cut will exclude giants
Fig. 3.— The density distribution of the 𝜒 2 ratio 𝑅 and eccentricity for the
fits to all the Gaia stars (grey background). We defined the selection region while the OGLE magnitude limits favour giants given the
based on the results for the SB2 HBs (dots). The final models of the SB1 distance to the Magellanic Clouds.
HBs (stars) can lie outside the selection region when masking eclipses leads Fig. 6 shows the stars on an extinction-corrected Gaia CMD.
to changes in the model. We use distances from Bailer-Jones et al. (2021) and extinc-
tions from the mwdust three-dimensional dust map (Bovy et al.
2016; Drimmel et al. 2003; Marshall et al. 2006; Green et al.
2019). The candidates have good parallaxes (𝜛/𝜎𝜛 > 10)
hundred iterations. Fig. 2 shows the model fits for nine of
and none are highly extincted (𝐴𝑉 < 2.0 mag), so their CMD
the SB2 HBs and Table 1 reports the median estimates for the
locations should be reliable. As compared to the full sample of
argument of periastron, inclination, eccentricity, and period.
Gaia spectroscopic binaries, the HBs tend to be higher on the
MS (median 𝑀𝐺 = 1.64). Fig. 6 also includes 478 HBs from
OGLE (Wrona et al. 2022a,b), 160 from Kepler (Shporer et al.
2.2. SB1 Modeling 2016; Li et al. 2023; Dimitrov et al. 2017; Kirk et al. 2016),
Since there are 181,529 Gaia SB1 orbital solutions, we and 179 from TESS (Kołaczek-Szymański et al. 2021; Solanki
performed a semi-automated search rather than relying only on et al. 2025).
visual inspection. We only analyzed the 43,599 systems with We used periodograms of the residuals of the Kumar et al.
Gaia orbital periods less than 13 days so that there would be at (1995) models, to search for TEOs. We list the systems with
least two binary orbits in a TESS sector. The TESS light curves periodogram powers larger than a false alarm level of 10−5 at
often have features at the start and end of each sector leading to a simple harmonic 𝑛/𝑃 of the orbit in Table 1. Ten of the 102
significant numbers of false positives for longer periods. For SB1 HB systems show TEOs, Fig. 7 shows three examples.
each light curve sector, we used a LS periodogram to identify Similar to Li et al. (2024b), our TEOs appear at low orbital
periodic variables. We then bin the phase-folded light curves harmonics with a median value of 𝑛 = 3 and a maximum at
into one hour time-bins, and fit the Kumar et al. (1995) analytic 𝑛 = 16 for TIC 173561516. The TEOs can be used to study
model (Eq. 1) to the phase-folded light curve using the Trust the internal structure of these stars.
Region Reflective algorithm. Since the LS periodogram often
returns a simple fraction of the orbital period, we do this for 3. GAIA COMPARISON
𝑃 𝐿𝑆 , 2𝑃 𝐿𝑆 , and 3𝑃 𝐿𝑆, where 𝑃 𝐿𝑆 is the period corresponding
to the maximum power in the periodogram. Fig. 8 compares the Gaia spectroscopic orbit solutions and
For each target, we select the sector and period combination HB fit orbital periods, eccentricities, and arguments of peri-
that minimizes the reduced 𝜒𝐻 2 astron. We were particularly interested in the SB2s because
𝐵 of the Kumar model fit. We agreement between the orbital parameters means that the Gaia
2
then compare this to the 𝜒𝑙𝑖𝑛𝑒 for a linear fit to the phase-folded velocity semi-amplitudes can be combined with the photomet-
light curves. Fig. 3 shows the distribution of 𝑅 = 𝜒𝐻 2 /𝜒 2
𝐵 𝑙𝑖𝑛𝑒 ric variability to measure masses and radii (Rowan et al. 2023).
and orbital eccentricities for the 43,599 systems. We use Unfortunately, only 2 of the 10 SB2 HBs have orbital periods
the HBs identified in the SB2 sample to define the selection and eccentricities that are in good agreement. It is difficult to
region (𝑅 < 0.5 and 𝑒 > 0.15) shown in Fig. 3. This left 7,314 pinpoint why an individual Gaia orbital solution is incorrect,
binaries, which we then visually inspected to select 102 HB since the epoch RVs are not included in Gaia DR3, but it is
candidates. Fig. 4 shows examples of the phase-folded light likely due to the large number of degenerate periods that are
curves with their Kumar model fits, and Table 1 reports the possible when fitting sparse RVs measured over a long time
orbital parameters. baseline to short period orbits. Furthermore, most of the SB2
Some of our HB candidates also have primary or secondary HBs are B/A stars on the upper MS (Fig. 6), and the 846–
eclipses, such as TIC 147307851, TIC 143634957, and TIC 870nm wavelength range of the Gaia RVS instrument was
22567490 in Fig. 4 (they are flagged in Table 1). Since the primarily designed for cool stars (Cropper et al. 2018). This
Kumar model only accounts for tidal distortions, we manually is one reason that the RVs for hot stars only became available
mask the eclipses when necessary and re-fit the phase-folded with DR3 (Blomme et al. 2023).
light curve. In some cases, the revised fit pushes the eccen- The SB1 HBs have considerably more accurate Gaia orbits,
tricity below the 0.15 cutoff used for selection, but we keep with 85% of the orbital periods and eccentricities agreeing
these systems in our final sample. with the light curve solution. Bashi et al. (2022) introduced
a “score” statistic ranging from 0 ≤ 𝑆 ≤ 1 to quantify the
 5
1.008 1.006 TIC = 303109427

Residuals Normalized Flux

TIC = 42430804 1.006 TIC = 153123772 Ecc = 0.42
Residuals Normalized Flux

Residuals Normalized Flux

1.006 Ecc = 0.56 Ecc = 0.55 1.004 Period = 10.94
Period = 9.00 1.004 Period = 11.37
1.004 1.002 1.002

1.002 1.000 1.000

1.000 0.998 0.998
0.998 0.996 0.996
0.996
0.002 0.994 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.002 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
0.002
0.001 0.001
0.000 0.000 0.000
0.001 0.001 0.002
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
Phase
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
Phase Phase
1.006 TIC = 452822117 TIC = 252419878 1.03
TIC = 456278778
Residuals Normalized Flux

Residuals Normalized Flux

1.006

Residuals Normalized Flux

1.004 Ecc = 0.46 Ecc = 0.43 Ecc = 0.45
Period = 7.51 1.004 Period = 3.63 1.02 Period = 8.84
1.002
1.000 1.002 1.01
0.998 1.000
0.996 1.00
0.998
0.994 0.99
0.996
0.992
0.994 0.98 0.00
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
0.001 0.002 0.01
0.000
0.000 0.00
0.001
0.002 0.002 0.01
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
Phase Phase Phase
1.006
1.05 TIC = 44787510 TIC = 271554988 TIC = 337605068
Residuals Normalized Flux

Residuals Normalized Flux

Residuals Normalized Flux

Ecc = 0.47 1.004 Ecc = 0.31 Ecc = 0.29
1.04 Period = 3.55 1.003 Period = 4.38 1.004 Period = 11.04
1.03 1.002 1.002
1.02 1.001
1.01 1.000 1.000
1.00 0.999 0.998
0.99 0.998
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
0.01 0.001 0.001

0.00 0.000 0.000

0.001
0.01 0.001
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
Phase Phase Phase
1.0125
1.003 TIC = 156175661
Residuals Normalized Flux

1.008 TIC = 299295231 TIC = 85177749 1.0100 Ecc = 0.26

Residuals Normalized Flux

Residuals Normalized Flux

Ecc = 0.29 Ecc = 0.28

1.006 Period = 5.93 1.002 Period = 3.90 1.0075 Period = 2.90
1.004 1.0050
1.001 1.0025
1.002
1.0000
1.000 1.000
0.9975
0.998 0.9950
0.999
0.996 0.9925
0.002 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.0010 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.002 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
0.001 0.0005 0.000
0.000 0.0000 0.002
0.001 0.0005 0.004
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
Phase
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
Phase Phase
1.020 1.006
TIC = 147307851 TIC = 143634957 1.0050 TIC = 22567490
Residuals Normalized Flux

Residuals Normalized Flux

Residuals Normalized Flux

1.015 Ecc = 0.32 1.004 Ecc = 0.25 1.0025 Ecc = 0.32

Period = 3.52 Period = 11.63 Period = 6.57
1.010 1.002 1.0000
1.005 1.000 0.9975
1.000 0.9950
0.998
0.9925
0.995 0.996
0.9900
0.990 0.994 0.9875
0.985 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.992
0.002 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.9850 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
0.002 0.002
0.001 0.001
0.000 0.000 0.000
0.002 0.001 0.001
0.002 0.002
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
Phase Phase Phase
Fig. 4.— Examples of the phase folded light curves of single-line spectroscopic heartbeat binaries with the light curves binned by one hour. The model residuals
are shown below each light curve. The text boxes show the TESS input catalog (TIC), eccentricity, and period (days) of each target. The bottom row shows three
SB1s that also have eclipses between the dashed lines. For the final models, we masked the data in these regions.
6

1.0 TIC = 9175061

Residuals Normalized Flux

1.0015
Ecc = 0.16
Gaia Clean Score Stars 1.0010 Period = 2.90 0.30

OGLE HB Stars 1.0005

1.0000
0.25

Kepler HB Stars
0.8 TESS HB Stars
0.9995 0.20

Power
0.9990

SB1 HB Stars 0.9985 0.15

SB2 HB Stars 0.9980 0.00

0.0005
0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.10

0.6 0.0000
0.0005
0.05
Ecc

0.0010 0.00
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
Phase Freq (1/d)
0.4 TIC = 29451470 0.200

Residuals Normalized Flux

1.0075 Ecc = 0.11
1.0050 Period = 3.01 0.175

1.0025 0.150
1.0000

0.2 0.9975
0.125

Power
0.9950 0.100
0.9925
0.075
0.004 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00

0.0 0 1 2 3
0.002
0.000
0.050

0.025

Log Period (days)

0.002
0.004 0.00 0.000
0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
Phase
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
Freq (1/d)
1.0025
TIC = 94860335

Residuals Normalized Flux

Fig. 5.— The distribution of heartbeat stars in period and eccentricity, 1.0020 Ecc = 0.36 0.16

1.0015
Period = 7.51
where the ones discussed here are the stars (SB1) and circles (SB2). The grey 0.14
1.0010
background is the distribution of all the "clean" Gaia SB1s (Bashi et al. 2022). 1.0005
0.12

The triangles are the Wrona et al. (2022a,b), Li et al. (2023), and Solanki et al. 1.0000 0.10

Power
0.9995
(2025) catalogs of OGLE, Kepler, and TESS HBs. The curve is the envelope 0.9990
0.08
outside of which orbits will rapidly circularize (Mazeh 2008). 0.9985 0.06
0.0010 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
0.0005 0.04

2 0.0000
0.0005
0.0010
0.02

0.00
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
Phase Freq (1/d)

0 Fig. 7.— Examples of targets with tidally excited oscillations. The left panel
shows the phase folded binned light curves, the model and the model residual.
The text boxes show the TESS input catalog (TIC), eccentricity, and period
(days) of each target. The right panel shows the periodogram of the residuals
MG [mag]

2 with vertical lines at orbital harmonic frequencies (𝑛/𝑃) and a horizontal line
at a false alarm probability of 10 −5 .

amplitudes to constrain the mass ratio

4
 2
𝜎𝐾1 𝜎𝐾2 2

𝐾1
6 𝑞= , with uncertainty 𝜎𝑞2 = 𝑞 2 + . (4)
𝐾2 𝐾12 𝐾22
0.0 0.5 1.0 1.5 2.0 and the projected semi-minor axis of the companion
BP - RP [mag]
𝐾2 𝑃 𝜎𝐾
Fig. 6.— Colour Magnitude Diagram (CMD) of the Gaia SB1 and SB2 𝑎 2 sin 𝑖 = √ , with uncertainty 𝜎𝑎2 𝑠𝑖𝑛𝑖 = 𝑎 2 sin 𝑖 2 .
2𝜋 1 − 𝑒 2 𝐾2
systems (grey and purple background), the SB1 (stars) and SB2 (dots) HBs
detected here, and the HB systems (triangles) from Wrona et al. (2022a,b), (5)
Shporer et al. (2016), Li et al. (2023), Dimitrov et al. (2017), Kirk et al. (2016), Only the uncertainties in 𝐾1 and 𝐾2 are important. We ini-
Kołaczek-Szymański et al. (2021), and Solanki et al. (2025). For the OGLE tialize the PHOEBE MCMC walkers near the values from the
Magellanic Cloud HBs we used the Skowron et al. (2021) reddening map and
the distance and extinction assumptions from MacLeod & Loeb (2025). The initial conditions, use Gaussian priors on 𝑞 and 𝑎 2 sin 𝑖, and
0.5, 1, 5, 10 Gyr isochrones are solar metallically MIST isochrones (Dotter then run the MCMC minimization with twenty walkers and
2016; Choi et al. 2016; Paxton et al. 2011, 2013, 2015, 2018). five hundred iterations.
Fig. 9 shows the posteriors for TIC 98552498. The mass
accuracy of Gaia SB1 orbital solutions. Of the 87 matching and radius of the primary are (2.61 ± 0.27) 𝑀⊙ and (1.41 ±
SB1 HB systems, 73 have scores below the 𝑆 = 0.587 cutoff 0.40) 𝑅 ⊙ , while the mass and radius of the secondary are (2.53
recommended by Bashi et al. (2022) for the selection of a ± 0.29) 𝑀⊙ and (2.85 ± 0.11) 𝑅 ⊙ . Fig. 10 shows the posteriors
“clean” SB1 sample. Despite this, the score statistic seems to for TIC 76094846. The mass and radius of the primary are
be a limited indicator of the Gaia orbit quality, as almost all of (8.19 ± 1.19) 𝑀⊙ and (5.14 ± 0.18) 𝑅 ⊙ , while the mass and
the targets with incorrect Gaia 𝑃, 𝑒, or 𝜔 also have 𝑆 < 0.587. radius of the secondary are (8.11 ± 1.39) 𝑀⊙ and (4.97 ±
For the two SB2 HB systems with correct Gaia peri- 0.47) 𝑅 ⊙ . Both systems are consistent with masses and radii
ods and eccentricities we used the PHysics Of Eclipsing of massive stars measured from detached eclipsing binaries
BinariEs (PHOEBE) package, a Python package for modeling (Torres et al. 2010).
EB systems (Prša & Zwitter 2005; Prša et al. 2016; Conroy
et al. 2020), to determine the stellar masses and radii. We 4. DISCUSSION
use the period and orbital parameters from the HB models as We used the TESS light curves of Gaia spectroscopic bina-
the initial conditions. We use the Gaia 𝐾1 and 𝐾2 velocity ries to identify one of the largest samples of HBs, finding 112
 7
1.0 HBs among the 186,905 spectroscopic binaries in the Gaia
12 DR3 catalog, of these, 18 show eclipses, and 10 have TEOs.
Out of the ten SB2 HBs, only two have orbital parameters in
0.8
10
HB Period [days]
good agreement with the Gaia values. The problem likely
arises because of the sparse Gaia RVs, the short periods, and

Gaia Score
8 0.587
that most of the systems are hot stars (Fig. 6). For the SB1
HBs, 85% of the orbits agreed with the Gaia estimates.
6 One of the original goals of this project was to find the SB2
HB systems and then model them to determine the component
0.4
masses and radii. The problems with the SB2 solutions meant
4 we could attempt this for only two systems. TIC 98552498
0.2 consists of a primary with mass (2.61 ± 0.27) 𝑀⊙ and radius
2 (1.41 ± 0.40) 𝑅 ⊙ , and a secondary with mass (2.53 ± 0.29)
𝑀⊙ and radius (2.85 ± 0.11) 𝑅 ⊙ . TIC 76094846 consists of
00.0 2.5 5.0 7.5 10.0 12.5 0.0 a primary with mass (8.19 ± 1.19) 𝑀⊙ and radius (5.14 ±
Gaia Period [days] 0.18) 𝑅 ⊙ , and a secondary with mass (8.11 ± 1.39) 𝑀⊙ and
radius (4.97 ± 0.47) 𝑅 ⊙ . To analyze the remaining SB2 HBs,
we require RV measurements from independent sources or
0.8 1.0
the actual Gaia RV measurements (which will be released in
DR4).
0.8 There are two striking features of the distribution of HBs
0.6 along the MS of the CMD shown in Fig. 6. First, the HBs
seem to be more luminous at fixed color than their parent
Gaia Score
samples. Second, there seem to be many fewer HBs on the
HB Ecc

0.587
0.4 lower MS. To quantify these issues, we first applied a cut
removing stars with 𝑀𝐺 < 4(𝐵 𝑃 − 𝑅 𝑃 − 2) to eliminate most
0.4 of the giant branch (Fig. 11). This is imperfect for stars redder
0.2 than roughly 𝐵 𝑃 − 𝑅 𝑃 > 1 mag, particularly for the SB1
systems. Fig. 11 also shows the 0.5 Gyr MIST isochrone
0.2 (the "single" isochrone) and the same isochrone shifted to be
0.75 mag brighter (the "binary" isochrone), the shift expected
0.0 for a binary of stellar twins. These are just to guide the eye
0.0
0.0 0.2 0.4 0.6 0.8 since they are not a good match to the observed upper MS.
Gaia Ecc To examine the first point, we computed the median mag-
nitudes of the samples in 𝐵 𝑃 − 𝑅 𝑃 colour bins with a width
1.0 of 0.1 mag. The medians of the full (or period trimmed) SB1
350 and SB2 samples behave as expected on the upper MS. They
HB Periastron [degrees]

roughly track the isochrones, and the SB2 magnitudes tend

300 0.8 to be a little more luminous since having a more luminous
companion will favor being able to obtain an SB2 solution.
250
Gaia Score

On the lower MS, the scatter of more luminous stars above the
0.587
200 MS but below our cut against giants affects the SB1 sample
more strongly than the SB2 sample because having an evolved
150 0.4 primary makes measuring the velocity of the secondary less
likely. The median absolute magnitudes of the HBs (SB1,
100 SB2, or all HBs) are then more luminous than the SB2 sam-
0.2 ple. This strongly indicates that HBs are not just binaries, but
50 binaries in which the primary has started to evolve. Following
0 0.0
the logic in MacLeod & Loeb (2025), this is a natural conse-
0 100 200 300 quence of HB amplitudes being higher as stellar evolutionary
Gaia Periastron [degrees] time scales become shorter.
To examine the second point, we computed the fraction of
Fig. 8.— Comparisons of the Gaia SB1 orbital parameters and those from
HBs as a function of color. There are two reasons the HB
the HB fits in Table 1. The points are coloured by the Gaia score from fraction might drop as we examine stars lower on the MS.
Bashi et al. (2022), centered on their score criterion of 𝑆 < 0.587 for good First, the redder, lower mass stars have longer and longer
orbital solutions (colour bar). The dashed lines are the range we consider a evolutionary time scales, which will lead to lower amplitudes
reasonable match between the values. Two of our HBs, marked with stars, do (MacLeod & Loeb 2025). Second, stars below the Kraft break
not have scores.
(𝑇𝐾 ≃ 6200 K or 𝐵 𝑃 − 𝑅 𝑃 ≃ 0.67 for the isochrone in Fig. 11,
Kraft 1967; Schatzman 1962) have convective envelopes and
so should have shorter damping time scales. Both of these
effects will lead to lower HB amplitudes. We have some
ability to distinguish the two issues because the highest mass
stars below the Kraft break (Kraft 1967; Schatzman 1962)
can be old enough to have started to evolve off the MS. We
 8

q = 0.98+0.04
0.06

M1 (M ) = 2.63+0.24
0.29
M1 (M )
2.0 2.4 2.8 3.2

R1 (R ) = 1.21+0.60
0.19
2.8 3.0 3.2 0.8 1.2 1.6 2.0
R1 (R )

R2 (R ) = 2.86+0.09
0.12
R2 (R )

e = 0.29+0.00
.0 .4 .6 .2 .8 .4 0 4 8 2 6 .6

0.00
520 52 48 49 49 50 0.28 0.28 0.28 0.29 0.29 2
e

i (°) = 49.51+0.36
0.32
i (°)

0 (°) = 52.03+0.21
0.25
(°)
0.8 .2 .6
51 51
0
8
6
4
2.0
2.4
2.8
3.2
0.8
1.2
1.6
2.0

2.6
2.8
3.0
0.2 .2
0.2 0
0.2 4
0.2 8
0.2 2
4896
.6
.2
.8
.4
.2
.6
.0
.4
0.8
0.9
1.0

8
8
8
9
3

49
49
50
51
51
52
52

q M1 (M ) R1 (R ) R2 (R ) e i (°) 0 (°)
Fig. 9.— The posteriors from the PHOEBE MCMC models of TIC 98552498 for the mass ratio, q, primary mass, primary radius, secondary radius, eccentricity,
inclination, and 𝜔. The projected probability distribution for each parameter and their 1𝜎 error bars (vertical lines) are shown at the top of each column. The
inset shows the light curve and the best PHOEBE model.

computed the fractions of SB1 and SB2 HBs as a function of full SB1 sample for the normalization of the fraction, rather
color again, but along lines parallel to the one labeled "Kraft than just those with periods < 13 days, but the trends are
break" in Fig. 11 in 𝐵 𝑃 − 𝑅 𝑃 = 0.1 mag bins. We report the unchanged. On the upper MS, HBs are not rare - over 1% of
median fractions and 90% confidence uncertainties or upper Gaia SB1 systems bluer than 𝐵 𝑃 − 𝑅 𝑃 < 0.5 mag are HBs.
limits. Starting from a sample of binaries can simply be viewed as a
The results shown in Fig. 12 are striking. The fraction of means of increasing the efficiency of HB searches, particularly
HBs drops very rapidly for redder, cooler stars starting above since many catalogs are dominated by shorter period systems.
the Kraft break (Kraft 1967; Schatzman 1962) and then may There are significant systematic effects in the catalogs which
be flattening for still cooler stars. The SB2 fractions and would need to be taken into account for any statistics as a
limits are in agreement with those for the SB1s. The absolute function of period. This can be done more broadly by selecting
normalizations of the SB1 fractions change when we use the Gaia or APOGEE (e.g., Abdurro’uf et al. 2022) stars with
 9

q = 0.98+0.10
0.07

M1 (M ) = 7.95+1.43
0.95
6 8 10 12
M1 (M )

R1 (R ) = 5.12+0.20
0.16
R1 (R )
4.8 5.2 5.6

R2 (R ) = 4.93+0.51
0.43
4.0 4.8 5.6 6.4
R2 (R )

e = 0.48+0.01
0.01
6.5 8.0 9.5 1.0 7.2 8.0 8.8 9.6 65 80 95
10 10 010(°)11 2 2 i 2(°)2 0.4 0.4 e0.4

i (°) = 28.48+0.60
0.58

0 (°) = 108.18+0.87
0.87
0
5
0
6
8
10
12

4.8
5.2
5.6

4.0
4.8
5.6
6.4

65
80
95

28 2
28 0
29 8
.6

10 5
10 0
11 5
1.0
0.9
1.0
1.2

.
.
.

6.
8.
9.
27

q M1 (M ) R1 (R ) R2 (R ) e i (°) (°)
0.4
0.4
0.4

10

0
Fig. 10.— The posteriors from the PHOEBE MCMC models of TIC 76094846.

significant velocity scatters (e.g., Badenes et al. 2018) rather curves can detect very weak signals.
than requiring an orbit solution. It would be good to find a It is clear that there is considerable physics in characteriz-
solution to the period search limitations of TESS. Evolved HBs ing and understanding the distribution of HBs as a function of
with longer periods may be detectable in the upcoming TESS orbital and stellar properties. What we qualitatively observe
Cycle 8, which includes longer 54 day sectors and “rolled” can largely be explained as the competition between evolution-
sectors where some parts of the sky will be observed over ary and dissipation rates, as discussed by MacLeod & Loeb
consecutive sectors. It might also be possible to search for (2025). High mass stars evolve faster and have less dissipation,
longer periods in TESS and then use heavily binned All-Sky leading to larger numbers of HBs. There are, however, a large
Automated Survey for Supernovae (ASAS-SN, Shappee et al. number of variables (mass, evolutionary state, orbit, etc.),
2014; Kochanek et al. 2017) light curves to verify the periods which means that significantly larger samples will be needed
and characterize the HB signal. While directly searching for to characterize the physical parameter space of HBs fully.
such weak signals in ASAS-SN might well be problematic,
Hon et al. (2025) have shown that binned ASAS-SN light
 10
Screenshot from 2025-06-05 14-52-47.png

10 1

Fraction (NHB/Nnot)
10 2

10 3

Kraft Break
SB1 HB Stars
SB2 HB Stars
10 4 0.0 0.5 1.0 1.5 2.0
BP - RP [mag]
Fig. 11.— Mean magnitudes in 0.1 mag wide colour bins for stars below the 10 1
Kraft Break
"Giant Branch Cutoff". The curves labeled SB1 and SB2 are for all the Gaia
binaries, the SB1 HB and SB2 HB curves are for the heartbeat stars found SB1 HB Stars
here, and the HB stars curve is for all HBs used in Fig. 5 and Fig. 6. An 0.5
Gyr MIST isochrone (Dotter 2016; Choi et al. 2016; Paxton et al. 2011, 2013,

Fraction (NHB/Nnot)
2015, 2018) is shown along with a binary "twin" isochrone a factor of two
more luminous. The Kraft break line crosses the isochrone at 𝑇𝑒 𝑓 𝑓 = 6200𝐾 10 2
(Kraft 1967; Schatzman 1962)

ACKNOWLEDGMENTS
JC thanks Tharindu Jayasinghe and the Time Domain Re-
search Group for their guidance, advice, and support. JC 10 3
also thanks Collin Christy for his assistance through the years.
DMR acknowledges support from the OSU Presidential Fel-
lowship. CSK, KZS, and DMR are supported by NSF grants
AST-2307385 and 2407206. ASAS-SN is funded by Gor- 10 4 0.0 0.5 1.0 1.5 2.0
don and Betty Moore Foundation grants GBMF5490 and
GBMF10501 and the Alfred P. Sloan Foundation grant G- BP - RP [mag]
2021-14192.
Our work presents results from the European Space Agency Fig. 12.— The median fraction of HBs as a function of colour. Here the
space mission Gaia. Gaia data are being processed by the Gaia colour bins are done in slices parallel to the Kraft break line in Fig. 11. The
upper panel is for all SB2 systems and the SB1 systems with 𝑃 < 13 days.
Data Processing and Analysis Consortium (DPAC). Funding The lower panel shows the fraction for all SB1 systems with no period limit.
for the DPAC is provided by national institutions, in particular The uncertainties and limits are at 90% confidence.
the institutions participating in the Gaia MultiLateral Agree-
ment. This paper also includes data collected by the TESS
mission. Funding for the TESS mission is provided by the
NASA’s Science Mission Directorate.
 11

REFERENCES
Abdurro’uf et al., 2022, ApJS, 259, 35 MacLeod M., Loeb A., 2025, arXiv e-prints, p. arXiv:2503.17133
Badenes C., et al., 2018, ApJ, 854, 147 Marshall D. J., Robin A. C., Reylé C., Schultheis M., Picaud S., 2006, A&A,
Bailer-Jones C. A. L., Rybizki J., Fouesneau M., Demleitner M., Andrae R., 453, 635
2021, AJ, 161, 147 Martín-Ravelo P., Gamen R., Arias J. I., Chené A.-N., Barbá R. H., 2024,
Barbá R. H., Gamen R. C., Martín-Ravelo P., Arias J. I., Morrell N. I., 2022, A&A, 690, A306
MNRAS, 516, 1149 Mazeh T., 2008, in Goupil M. J., Zahn J. P., eds, EAS Publications Series
Bashi D., Shahaf S., Mazeh T., Faigler S., Dong S., El-Badry K., Rix H. W., Vol. 29, EAS Publications Series. pp 1–65 (arXiv:0801.0134),
Jorissen A., 2022, MNRAS, 517, 3888 doi:10.1051/eas:0829001
Blomme R., et al., 2023, A&A, 674, A7 Merc J., Kalup C., Rathour R. S., Sánchez Arias J. P., Beck P. G., 2021,
Bovy J., Rix H.-W., Green G. M., Schlafly E. F., Finkbeiner D. P., 2016, ApJ, Contributions of the Astronomical Observatory Skalnate Pleso, 51, 45
818, 130 Murphy S. J., Paunzen E., Bedding T. R., Walczak P., Huber D., 2020,
Burkart J., Quataert E., Arras P., Weinberg N. N., 2012, Monthly Notices of MNRAS, 495, 1888
the Royal Astronomical Society, 421, 983 Nesterov V. V., Kuzmin A. V., Ashimbaeva N. T., Volchkov A. A., Röser S.,
Choi J., Dotter A., Conroy C., Cantiello M., Paxton B., Johnson B. D., 2016, Bastian U., 1995, A&AS, 110, 367
ApJ, 823, 102 O’Leary R. M., Burkart J., 2014, MNRAS, 440, 3036
Conroy K. E., et al., 2020, ApJS, 250, 34 Paunzen E., Hümmerich S., Fedurco M., Bernhard K., Komžík R., Vaňko
Cropper M., et al., 2018, A&A, 616, A5 M., 2021, MNRAS, 504, 3749
Dimitrov D. P., Kjurkchieva D. P., Iliev I. K., 2017, MNRAS, 469, 2089 Paxton B., Bildsten L., Dotter A., Herwig F., Lesaffre P., Timmes F., 2011,
Dotter A., 2016, ApJS, 222, 8 ApJS, 192, 3
Drimmel R., Cabrera-Lavers A., López-Corredoira M., 2003, A&A, 409, 205 Paxton B., et al., 2013, ApJS, 208, 4
Foreman-Mackey D., Hogg D. W., Lang D., Goodman J., 2013, PASP, 125, Paxton B., et al., 2015, ApJS, 220, 15
306 Paxton B., et al., 2018, ApJS, 234, 34
Fuller J., Lai D., 2012, MNRAS, 420, 3126 Prša A., Zwitter T., 2005, ApJ, 628, 426
Gaia Collaboration et al., 2023, A&A, 674, A1 Prša A., et al., 2016, ApJS, 227, 29
Green G. M., Schlafly E., Zucker C., Speagle J. S., Finkbeiner D., 2019, Ricker G. R., et al., 2015, Journal of Astronomical Telescopes, Instruments,
ApJ, 887, 93 and Systems, 1, 014003
Hambleton K., et al., 2013, in Pavlovski K., Tkachenko A., Torres G., eds, Röser S., Bastian U., 1993, Bulletin d’Information du Centre de Donnees
EAS Publications Series Vol. 64, EAS Publications Series. pp 285–294, Stellaires, 42, 11
doi:10.1051/eas/1364039 Röser S., Bastian U., Kuzmin A., 1994, A&AS, 105, 301
Hon M., et al., 2025, ApJ, 984, L3 Rowan D. M., Jayasinghe T., Stanek K. Z., Kochanek C. S., Thompson T. A.,
Huang C. X., et al., 2020, Research Notes of the American Astronomical Shappee B. J., Giles W., 2023, MNRAS, 523, 2641
Society, 4, 204 Scargle J. D., 1982, ApJ, 263, 835
Jayasinghe T., Stanek K. Z., Kochanek C. S., Thompson T. A., Shappee B. J., Schatzman E., 1962, Annales d’Astrophysique, 25, 18
Fausnaugh M., 2019, MNRAS, 489, 4705 Shappee B. J., et al., 2014, ApJ, 788, 48
Jayasinghe T., et al., 2021, MNRAS, 506, 4083 Sharma A. N., Bedding T. R., Saio H., White T. R., 2022, MNRAS, 515, 828
Kirk B., et al., 2016, AJ, 151, 68 Shporer A., et al., 2016, ApJ, 829, 34
Kochanek C. S., et al., 2017, PASP, 129, 104502 Skowron D. M., et al., 2021, ApJS, 252, 23
Kochukhov O., Labadie-Bartz J., Khalack V., Shultz M. E., 2021, MNRAS, Solanki S., et al., 2025, ApJS, 276, 17
506, L40 Stellingwerf R. F., 1978, ApJ, 224, 953
Kołaczek-Szymański P. A., Pigulski A., Michalska G., Moździerski D., Thompson S. E., et al., 2012, ApJ, 753, 86
Różański T., 2021, A&A, 647, A12 Torres G., Andersen J., Giménez A., 2010, A&A Rev., 18, 67
Kołaczek-Szymański P. A., Pigulski A., Wrona M., Ratajczak M., Udalski Wang K., Ren A., Andersen M. F., Grundahl F., Chen T., Pallé P. L., 2023,
A., 2022, A&A, 659, A47 AJ, 166, 42
Kraft R. P., 1967, ApJ, 150, 551 Welsh W. F., et al., 2011, ApJS, 197, 4
Kumar P., Ao C. O., Quataert E. J., 1995, ApJ, 449, 294 Wheeler A., Kipping D., 2019, MNRAS, 485, 5498
Li M.-Y., Qian S.-B., Zhu L.-Y., Liao W.-P., Zhao E.-G., Shi X.-D., Sun Wrona M., et al., 2022a, ApJS, 259, 16
Q.-B., 2023, ApJS, 266, 28 Wrona M., Kołaczek-Szymański P. A., Ratajczak M., Kozłowski S., 2022b,
Li M.-Y., et al., 2024a, arXiv e-prints, p. arXiv:2407.18621 ApJ, 928, 135
Li M.-Y., et al., 2024b, arXiv e-prints, p. arXiv:2408.01019
Lomb N. R., 1976, Ap&SS, 39, 447
arXiv, making the reviewing process simpler for authors and
This paper was built using the Open Journal of Astrophysics referees alike. Learn more at http://astro.theoj.org.
LATEX template. The OJA is a journal which provides fast and
easy peer review for new papers in the astro-ph section of the