PHYSICAL REVIEW D 108, 095012 (2023)

Flavor physics in SU(5) GUT with scalar fields in the 45 representation

Toru Goto,1 Satoshi Mishima ,1,2 and Tetsuo Shindou3
1
Theory Center, IPNS, KEK, Tsukuba 305-0801, Japan
2
Department of Liberal Arts, Saitama Medical University, Moroyama, Saitama 350-0495, Japan
3
Division of Liberal-Arts, Kogakuin University, 2665-1 Nakano-machi, Hachioji, Tokyo 192-0015, Japan

(Received 6 September 2023; accepted 4 October 2023; published 7 November 2023)

We study a realistic SU(5) grand unified model, where a 45 representation of scalar fields is added to the
Georgi-Glashow model in order to realize the gauge coupling unification and the masses and mixing of
quarks and leptons. The gauge coupling unification together with constraints from proton decay implies mass
splittings in scalar representations. We assume that an SU(2) triplet component of the 45 scalar, which is
called S3 leptoquark, has a TeV-scale mass, and color-sextet and color-octet ones have masses of the order of
106 GeV. We calculate one-loop beta functions for Yukawa couplings in the model, and derive the low-
energy values of the S3 Yukawa couplings which are consistent with the grand unification. We provide
predictions for lepton-flavor violation and lepton-flavor-universality violation induced by the S3 leptoquark,
and find that current and future experiments have a chance to find a footprint of our SU(5) model.

DOI: 10.1103/PhysRevD.108.095012

I. INTRODUCTION proposed by Georgi and Glashow [3]. In the minimal SU(5)

GUT model, the right-handed down quarks and the left-
The idea of Grand Unification is an attractive candidate
for the fundamental theory behind the present under- handed lepton doublets are embedded in 5̄ representations
standing of particle physics described by the Standard of SU(5), and the left-handed quark doublets, the right-
Model (SM) [1–6]. Although the SM has been established handed up quarks, and the right-handed charged leptons are
as a successful effective model at the electroweak (EW) embedded in 10 representations. The SM Higgs doublet is
scale by the discovery of the Higgs boson, reaching a embedded in a 5 representation of scalars. In addition, the
minimal model also contains a 24 representation of scalars,
deeper understanding of nature is a desire of particle
which breaks the SU(5) gauge symmetry to the SM ones.
physicists. Interestingly, some properties of the SM sug-
Although the concept of the minimal SU(5) GUT is
gest the existence of a Grand Unified Theory (GUT) as a
beautiful, there are two serious issues that have to be solved
high-energy theory beyond the SM. For example, the
to construct a more realistic model. First, the three gauge
renormalization group (RG) runnings of the gauge cou-
couplings are not unified at a high-energy scale only with
plings in the SM show a unification tendency at a high
the RG runnings in the SM. In the minimal SU(5) model,
scale [4], and the charge quantization of the SM fermions
there is a grand desert between the EW and the GUT scales,
suggests the unification of matter. Once the SM gauge
where there is no new contribution to the RG runnings.
groups for electromagnetic, weak, and strong interactions
Second, the measured values of the masses of the charged
are unified to a GUT gauge symmetry group, quarks and
leptons and the down-type quarks cannot be accommodated
leptons are consequently unified in a single or a few
with the minimal SU(5) GUT, where they originate from a
representations of the GUT group. Various groups, such as
common Yukawa interaction in the GUT Lagrangian.
SU(5), SO(10), E6 , etc., have been considered as the GUT
The first issue on the gauge coupling unification can be
gauge symmetry group [7].
overcome by introducing extra fields in the grand desert,
The SU(5) is the minimal simple group which contains
since such fields modify the RG runnings of the gauge
the SM gauge groups SUð3ÞC × SUð2ÞL × Uð1ÞY . The
couplings. A famous example of this direction is the
minimal version of the GUT model based on the SU(5)
supersymmetric SU(5) GUT model, in which the super-
symmetry, called the minimal SU(5) GUT, was originally
partners of the SM particles as well as the second Higgs
doublet are introduced and the gauge coupling unification
occurs at the scale of the order of 1016 GeV [8–12]. In
Published by the American Physical Society under the terms of nonsupersymmetric SU(5) GUT models, a single or a few
the Creative Commons Attribution 4.0 International license.
Further distribution of this work must maintain attribution to particles in an extra representation of SU(5) are predicted to
the author(s) and the published article’s title, journal citation, lie in the grand desert in order to realize the gauge coupling
and DOI. Funded by SCOAP3. unification [13–30].

2470-0010=2023=108(9)=095012(28) 095012-1 Published by the American Physical Society

GOTO, MISHIMA, and SHINDOU PHYS. REV. D 108, 095012 (2023)

The second issue on the masses of the charged leptons the TeV scale in our model on the phenomenology of flavor
and the down-type quarks can be resolved by introducing a observables, such as leptonic and semileptonic B decays,
45 representation of scalar fields to the minimal SU(5) Bs − B̄s mixing, ϒðnSÞ decays, tau-lepton decays, and
GUT. Similar to the 5 scalar, the 45 scalar couples with Z → μ∓ τ decay. We show that Belle II with 50 ab−1 and
the 10 and the 5̄ fermions since 10 ⊗ 5̄ ¼ 5 ⊕ 45. This LHCb with 300 fb−1 have a chance to find a footprint of
coupling makes modifications in the relation between our SU(5) GUT model.
the charged-lepton and the down-type-quark Yukawa This paper is organized as follows. In Sec. II, we
matrices at the GUT scale through the Georgi-Jarlskog introduce an SU(5) GUT model with a 45 scalar, and
mechanism [31]. explain how it solves the issues in the minimal SU(5) GUT.
We combine the above two ideas on the extensions of In Sec. III, we present and discuss phenomenological
the minimal SU(5) GUT, and construct a concrete implications of our model. Section IV contains our sum-
example of a realistic SU(5) GUT model, where the mary and conclusions. Some technical details are given in
gauge coupling unification and the correct fermion the Appendixes.
masses are realized simultaneously. This kind of model
having the 45 scalar can be found, for example, in II. MODEL
Refs. [13,15,18,21–25,27,28]. In the current study, we
introduce the 45 scalar to reproduce the charged-lepton A. Lagrangian
and the down-type-quark Yukawa matrices correctly, and We consider an SU(5) GUT model, where the SM
make an SU(2) triplet component of the 45 scalar light fermions reside in 10 and 5̄ representations of SU(5),
enough to achieve the gauge coupling unification [18]. denoted by Ψ10i and Ψ5̄i with i ¼ 1, 2, 3 being the
This triplet scalar is called S3 . The Yukawa interactions generation index, and the scalar sector is composed of
between the 45 scalar and the 10 and the 5̄ fermions are one 24, one 5, and one 45-dimensional scalar representa-
given by 10 · 10 · 45 and 10 · 5̄ · 45, where we omit the tion, denoted by Σ, Φ5 , and Φ45 , respectively. The SU(5)-
former by hand to suppress baryon-number-violating symmetric renormalizable Lagrangian is given by
interactions mediated by the light S3 . In addition to the
S3 , we assume that color-sextet and color-octet compo- 1
nents in the 24 and the 45 scalars have masses of the order L ¼ − ðV μν ÞB A ðV μν ÞA B þ iðΨ̄10i ÞAB γ μ Dμ ðΨ10i ÞAB
4
of 106 GeV in order to avoid too rapid proton decay þ iðΨ̄5̄i ÞA γ μ Dμ ðΨ5̄i ÞA þ ½Dμ ΣB A ½Dμ ΣA B 
mediated by the GUT gauge bosons.1 In this case, the
SU(2) triplet scalar has a mass of Oð103 –106 GeVÞ, and þ ½Dμ ðΦ†5 ÞA ½Dμ ðΦ5 ÞA  þ ½Dμ ðΦ†45 ÞCAB ½Dμ ðΦ45 ÞAB
C 
the GUT scale is of Oð1016 –1017 GeVÞ. We do not þ LY − VðΣ; Φ5 ; Φ45 Þ; ð1Þ
consider any mechanisms to generate the mass splittings
in the GUT multiplets and to forbid the 10 · 10 · 45 where V μν is the field strength tensor of the SU(5) gauge
interactions, which are beyond the scope of the current
bosons, A; B; C ¼ 1; …; 5 are SU(5) indices, and LY and
work. Moreover, we do not specify the origin of the
VðΣ; Φ5 ; Φ45 Þ represent the Yukawa interactions and the
nonzero neutrino masses, which are studied in the frame-
scalar potential, respectively. The summation over repeated
work of the SU(5) GUT with the 45 scalar, for example, in
indices is implied. Here the fields Ψ10i , Σ, and Φ45 satisfy
Refs. [21,22,25,27,32–34].
the following relations:
This triplet scalar S3 carries the SM gauge quantum
numbers ð3; 3; −1=3Þ, and has Yukawa couplings to a
lepton and a quark. The conjugate state of S3 , having ðΨ10i ÞAB ¼ −ðΨ10i ÞBA ; ðΣB A Þ ¼ ΣA B ; ΣA A ¼ 0;
ð3̄; 3; 1=3Þ, is often called S3 leptoquark [35,36]. If the ðΦ45 ÞAB
C ¼ −ðΦ45 ÞC ;
BA
ðΦ45 ÞAB
A ¼ 0: ð2Þ
mass of the S3 leptoquark lies at the TeV scale, S3 can affect
various flavor observables. Unlike the phenomenological In general the Yukawa term LY in Eq. (1) consists of the
models where the S3 leptoquark is introduced by hand as, four interactions:
for instance, in Refs. [37–48], the flavor structure of the
Yukawa couplings associated with S3 is constrained by the 1
−LY ¼ ðY U Þ ϵ ðΨ ÞAB ðΦ5 ÞC ðΨ10j ÞDE
measured values of the charged-lepton and the down-type 8 5 ij ABCDE 10i
quark masses. It provides peculiar correlations in the flavor †
observables. We study the impact of the S3 leptoquark at þ ðY D
5 Þij ðΨ10i Þ ðΦ5 ÞA ðΨ5̄j ÞB
AB

1
þ ðY U Þ ϵ ðΨ ÞAB ðΦ45 ÞCD F ðΨ10j Þ
EF
1 4 45 ij ABCDE 10i
For example, one can increase the GUT scale to evade the
constraint from the proton decay by making the ð8; 2; 1=2Þ scalar 1
þ ðY D Þ ðΨ ÞAB ðΦ†45 ÞCAB ðΨ5̄j ÞC þ H:c:; ð3Þ
in the 45 representation light [13,18,30]. 2 45 ij 10i

095012-2
FLAVOR PHYSICS IN SU(5) GUT WITH SCALAR FIELDS IN … PHYS. REV. D 108, 095012 (2023)

where the totally antisymmetric tensor is defined as The explicit expression for the scalar potential
ϵ12345 ¼ 1, and Y U U
5 and Y 45 are symmetric and antisym-
VðΣ; Φ5 ; Φ45 Þ is given in Appendix A.
metric matrices in the generation space, respectively: The SU(5) gauge symmetry is assumed to be broken
down to the SM gauge symmetry SUð3ÞC × SUð2ÞL ×
Uð1ÞY by the vacuum expectation value (VEV) of a SM-
singlet scalar field in Σ: hΣi ¼ v24 diagð2; 2; 2; −3; −3Þ.
5 Þij ¼ ðY 5 Þji ;
ðY U U
ðY U
45 Þij ¼ −ðY 45 Þji :
U
ð4Þ The field Σ is decomposed around the VEV as

0   1
1ffiffiffiffi p1ffiffi ðΣG Þâ
ðΣ Þ
B 8 b̂ þ 2 v 24 −
â p
2 15 1
Σ δâ b̂ 2 β C
ΣA B B
¼@   C ð5Þ
A;
p1ffiffi ðΣ Þα ðΣ3 Þαβ − 3 v24 − 2p1ffiffiffi
ffi Σ1 δα β
2 G b̂ 15

where â; b̂ ¼ 1, 2, 3 and α, β ¼ 1, 2 are SU(3) and SU(2) where the mass eigenstates are denoted with a hat, and
indices, respectively. The spontaneous breaking of SU(5) V CKM is the Cabibbo-Kobayashi-Maskawa (CKM)
typically provides the masses of the scalars Σ1 , Σ3 , and Σ8 matrix in the Particle Data Group (PDG) phase convention
ðÞ
of the order of v24 , while ΣG corresponds to the massless [49,50]. Analogous to the CKM matrix that represents a
would-be Nambu-Goldstone boson, which gives masses to mismatch of the bases in qL , the unitary matrices V QU ,
the gauge bosons associated with the broken symmetries. V QE , and V DL are introduced in Ψ10 and Ψ5̄ as in Eq. (6).
These massive vector bosons are called X bosons.
C. Scalar spectrum and gauge coupling unification
B. Fermions The scalar Φ5 is decomposed to the so-called color triplet
lLi , and
ucRi , dcRi , ecRi ð5Þ
The SM fermions qLi , are Higgs S1 and the SUð2ÞL doublet Hð5Þ :
embedded into the 10 and 5̄ representations as
!  ð5Þâ 
S1
1 ϵâ b̂ ĉ ðV QU Þi k ucRkĉ qâβ
Li ðΦ5 ÞA ¼ ; ð9Þ
ðΨ10i ÞAB ¼ pffiffiffi ; Hð5Þα
2 −qb̂α
Li ϵαβ ðV QE Þi k ecRk
 
ðΨ5̄i ÞA ¼ dcRiâ ϵαβ ðV DL Þi k lβLk ; while the Φ45 consists of the scalars S̃1 , R2 , S3 , S6 , S8 ,
ð45Þ
ð6Þ Hð45Þ , and S1 as

 
where i, k are the generation indices, and the totally 1 1 ð45Þê
antisymmetric tensors are defined as ϵ12 ¼ ϵ12 ¼ 1 and ðΦ45 Þâĉ b̂ ¼ pffiffiffi ϵâ b̂ d̂ ðηa Þĉ d̂ Sa − ϵ S ;
2
6
2 ĉ d̂ ê 1
ϵ123 ¼ ϵ123 ¼ 1. Without loss of generality, one can rotate 1
the basis of Ψ10 and Ψ5̄ as ðΦ45 Þâγ b̂ ¼ pffiffiffi ϵâ b̂ d̂ R2d̂γ ;
2
 
1 1 1 â ð45Þβ
Ψ10 → U 10 Ψ10 ; Ψ5̄ → U5 Ψ5̄ ; ð7Þ ðΦ45 Þâβĉ ¼ pffiffiffi p ffiffiffi ðλ Þ â aβ
a ĉ 8 S þ p ffiffi
ffi δĉ H ;
2 2 2 3
1 αβ
where U10 and U5 are arbitrary unitary matrices in the ðΦ45 Þαβ ĉ ¼ pffiffiffi ϵ S̃1ĉ ;
generation space. By using the degrees of freedom asso- 2
 
ciated with the unitary rotations, we can take the basis α 1 1 α  1 α ð45Þb̂
where the up-type quarks and the charged leptons are in ðΦ45 Þγ ¼ pffiffiffi pffiffiffi ðσ a Þ γ S3 − δγ S1
b̂ b̂
;
2 2 2
their mass eigenstates: pffiffiffi
3
  ðΦ45 Þγ ¼ − pffiffiffi ϵαβ ϵγδ H ð45Þδ ;
αβ ð10Þ
ûLi 2 2
qLi ¼ ; uRi ¼ ûRi ; dRi ¼ d̂Ri ;
ðV CKM Þi j d̂Lj
  where σ a (a ¼ 1, 2, 3) are the Pauli matrices, λa ða ¼
ν̂Li
lLi ¼ ; eRi ¼ êRi ; ð8Þ 1; 2; …; 8Þ the Gell-Mann matrices, and ηa ða ¼ 1; 2; …; 6Þ
êLi the symmetric matrices defined by

095012-3
GOTO, MISHIMA, and SHINDOU PHYS. REV. D 108, 095012 (2023)

TABLE I. The decomposition of the scalar fields Σ, Φ5 , and Φ45 under the SM gauge groups.

Field SU(5) Field SUð3ÞC SUð2ÞL Uð1ÞY Field SU(5) Field SUð3ÞC SUð2ÞL Uð1ÞY
Σ 24 Σ1 1 1 0 Φ45 45 S̃1 3̄ 1 4=3
Σ3 1 3 0 R2 3̄ 2̄ −7=6
ΣG 3 2̄ −5=6 S3 3 3 −1=3
ΣG 3̄ 2 5=6 S6 6̄ 1 −1=3
Σ8 8 1 0 S8 8 2 1=2
Φ5 5 H ð5Þ 1 2 1=2 H ð45Þ 1 2 1=2
ð5Þ
S1 3 1 −1=3 ð45Þ
S1 3 1 −1=3

80 10 10 1 0 1 0 1 0 19
>
> 1 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 >
>
< =
B CB CB C 1 B C 1 B C 1 B C
fη1 ; η2 ; η3 ; η4 ; η5 ; η6 g ¼ @ 0 0 0 A; @ 0 1 0 A; @ 0 0 0 A; pffiffiffi @ 1 0 0 A; pffiffiffi @ 0 0 0 A; pffiffiffi @ 0 0 1 A :
>
> 2 2 2 >
>
: 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 1 0 ;
ð11Þ

The decompositions of Σ, Φ5 , and Φ45 are summarized in At the energy scale above the mass of an additional light
ð45Þ
Table I. Here the scalar H ð45Þ (S1 ) has the same quantum scalar, the scalar contributes to the RG running of the gauge
ð5Þ couplings. The gauge coupling unification is realized if an
numbers under the SM gauge group as H ð5Þ (S1 ).
appropriate set of light scalars is considered. We define
Therefore, they can mix with each other, and we define α3 ðμÞ, α2 ðμÞ, and α1 ðμÞ as
the mass eigenstates H, H0 , HC , and S1 by introducing the
mixing angles θH and θS and the phases δH and δS : gs ðμÞ2 gðμÞ2
α3 ðμÞ ¼ αs ðμÞ ¼ ; α2 ðμÞ ¼ ;
     4π 4π
H cH e−iδH sH Hð5Þ
¼ ; 5 g0 ðμÞ2
H0 −eiδH sH cH H ð45Þ α1 ðμÞ ¼ ; ð13Þ
3 4π
    ð5Þ
!
HC cS e−iδS sS S1 where gs , g, and g0 are the gauge couplings of SUð3ÞC ,
¼ ; ð12Þ
S1 −eiδS sS cS ð45Þ
S1 SUð2ÞL , and Uð1ÞY , respectively, and μ is the renormaliza-
tion scale. Our analysis assumes that the SM gauge
where cH ¼ cosθH , sH ¼ sinθH , cS ¼ cosθS , and sS ¼ sinθS . couplings are unified at the scale M X , i.e., α3 ðMX Þ ¼
The presence of the two doublet scalars allows us to explain α2 ðMX Þ ¼ α1 ðM X Þ ≡ αX ðMX Þ, and all the scalar masses
the masses of the down-type quarks and the charged leptons are not heavier than MX . Then, above the M X scale, all
simultaneously. the scalars contribute to the running as complete SU(5)
Owing to the symmetry breaking of SU(5) to the SM multiplets so that the coupling unification holds above MX .
gauge groups, mass splitting may occur among the scalar We also make an ansatz that the mass of the X boson is
fields embedded in the SU(5) multiplets. At least one equal to the unification scale M X .
SUð2ÞL -doublet scalar has to be light to break the EW Solving the renormalization group equations (RGEs) in
symmetry spontaneously below the TeV scale.2 We assume Appendix C with the unification assumption, we get the
that the scalar H is light and corresponds to the SM Higgs three relations,
 SM 
M X X Bϕgs
doublet. Bgs
−1 −1 MX
It is well-known that the SM gauge couplings do not αX ðM X Þ ¼ α3 ðmZ Þ − log þ log ;
unify only by naive RG running in the SM. The mass 2π mZ ϕ
2π mϕ
 SM 
M X X Bϕg
splitting of the SU(5) scalar multiplets can improve the Bg
−1 −1 MX
situation. We consider the scenario where some of the αX ðM X Þ ¼ α2 ðmZ Þ − log þ log ;
scalar fields, in addition to H, are much lighter than others. 2π mZ ϕ
2π mϕ
 SM ϕ 
3 Bg0 M X X Bg0 MX
2
The EW symmetry breaking can also be driven by the VEV of α−1
X ðM X Þ ¼ α−1
1 ðmZ Þ − log þ log ;
Σ3 below the TeV scale. However, we assume that Σ3 does not 5 2π mZ ϕ
2π mϕ
develop a VEV since it causes a dangerous contribution to the ρ
parameter. ð14Þ

095012-4
FLAVOR PHYSICS IN SU(5) GUT WITH SCALAR FIELDS IN … PHYS. REV. D 108, 095012 (2023)

TABLE II. Input values for the Z-boson mass mZ , the gauge couplings αs ðmZ Þ and α−1 ðmZ Þ, the weak mixing angle sin2 θW ðmZ Þ, the
quark masses, and the CKM parameters sij and δ, taken from Ref. [50]. Other parameters, such as the pole masses of the charged leptons
me , mμ , and mτ , are also taken from Ref. [50].

Parameter Value Parameter Value Parameter Value Parameter Value

mZ 91.1876 GeV mu ð2 GeVÞ 0.00216 GeV md ð2 GeVÞ 0.00467 GeV s12 0.22650
αs ðmZ Þ 0.1179 mc ðmc Þ 1.27 GeV ms ð2 GeVÞ 0.093 GeV s13 0.00361
α−1 ðmZ Þ 127.952 mpole
t
172.76 GeV mb ðmb Þ 4.18 GeV s23 0.04053
sin2 θW ðmZ Þ 0.23121 δ 1.196 rad

where mZ is the Z-boson mass, ϕ is summed over all the addition to S3 , are lighter than the unification scale MX .
ϕ With their contributions to the RGEs of the gauge
relevant scalars, and the coefficients BSM
gi and Bgi are given
−1
in Table IV. Eliminating αX ðMX Þ [51], one can get two couplings, MX can be significantly heavier with keeping
independent equations, the coupling unification. Therefore, we consider a
scenario where the masses of S6 , S8 , and Σ8 are below
2 mH 2 mS 7 mS̃ 4 mR MX . For simplicity, we assume that the other scalar
log C þ log 1 þ log 1 þ log 2
5 mZ 5 mH0 5 mS3 5 mS3 components, except for the SM-like Higgs doublet H,
are as heavy as MX .
9 mS 4 mS mΣ
þ log 6 þ log 8 þ log 8 Let us explain in more detail the masses of the other
5 mS3 5 mS3 mΣ3 scalars embedded in the GUT representations. The mass
¼ 2π½−2α−1 −1 −1
3 ðmZ Þ þ 3α2 ðmZ Þ − α1 ðmZ Þ parameter m2H associated with the SM-like Higgs doublet H
is of the order of the weak scale according to the LHC
≃ 79.8; ð15Þ
measurements, while the other scalars associated with the 5
MX mS mS mS mΣ mΣ and 45 representations obey the mass relation given in
44 log þ 6 log 3 þ log 6 þ 4 log 8 þ log 3 2 8 Eq. (A13). We simply choose that Σ1 , Σ3 , H 0 , HC , S1 , and
mZ mR2 mS̃1 mS̃1 MX
S̃1 have a common mass MX . As a consequence, the mass
¼ 2π½−2α−1 −1 −1
3 ðmZ Þ − 3α2 ðmZ Þ þ 5α1 ðmZ Þ of R2 is determined as m2R2 ≈ ð2 þ 4s2H ÞM 2X =3, where sH is
≃ 1193; ð16Þ the sine of the mixing angle defined in Eq. (12). For
s2H < 1=4, mR2 is lighter than MX .
where the gauge couplings at μ ¼ mZ are evaluated for six
In Fig. 1(a), the contours of MX and mΣ8 are shown in the
active quark flavors [52] with the input values shown in
parameter space of mS3 and mS6 ¼ mS8 . The gauge cou-
Table II.
As a general property, the S3 contribution improves the pling unification favors rather light S3 , which can be as
gauge coupling unification [18]. In the case that only light as a TeV scale. The light gray regions are for
the Higgs boson and S3 are lighter than MX in the scalar MX < 3 × 1015 GeV, which is disfavored by the proton
sector, the gauge coupling unification occurs at MX ∼ decay search as mentioned above. For example, if we take
Oð1014 GeVÞ with mS3 ∼ Oð108 GeVÞ. However, M X is mS3 ¼ 2 TeV, mS6 ¼ mS8 ¼ mΣ8 ¼ 5.2 × 106 GeV, and
severely constrained by proton decay search experiments, cot θH ¼ 50, the gauge coupling unification is realized at
since contributions from the GUT gauge-boson exchange MX ¼ 9.7 × 1016 GeV as shown in Fig. 1(b).
generate the dimension-six operators relevant to the In the phenomenological analysis, we use a benchmark
proton decay. Then the proton lifetime is naively expected scenario with the following mass spectrum:
as [7,53] (1) The masses of the quarks and leptons, the SM gauge
bosons, and the SM-like Higgs boson H are set to be
M 4X consistent with their measurements;
τp ∼ ; ð17Þ (2) S3 has a TeV-scale mass: mS3 ∼ Oð103 GeVÞ;
α2X m5p
(3) S6 , S8 , and Σ8 have intermediate masses, and we set
where mp is the mass of proton, and one finds a naive lower them to an identical scale, i.e., mS6 ¼ mS8 ¼ mΣ8 ≡
bound as MI ∼ Oð106 GeVÞ;
(4) The other particles including the X bosons have
MX > 5 × 1015 GeV; ð18Þ masses of the order of the GUT scale MX .
by using the experimental lower limit on the lifetime
τðp → π 0 eþ Þ > 2.4 × 1034 years [54] and α2X ∼ Oð10−3 Þ. D. Yukawa couplings
In order to avoid the rapid proton decay by making Below the GUT scale, the Yukawa interactions with the
MX much heavier, we assume that S6 , S8 , and Σ8 , in scalars H, S3 , S6 , and S8 are given by

095012-5
GOTO, MISHIMA, and SHINDOU PHYS. REV. D 108, 095012 (2023)

FIG. 1. (a) The contours of MX (solid) and mΣ8 (dashed) in the unit of GeV for realizing the coupling unification in the plane of mS3 and
mS6 ¼ mS8 for cot θH ¼ 50. In the blue shaded region, the gauge coupling unification does not occur by RG running. The light gray region
is disfavored by the proton decay experiments because M X is too small. The green dot-dashed line corresponds to the case with mS6 ¼
mS8 ¼ mΣ8 . (b) RG runnings of the gauge couplings for mS3 ¼ 2 TeV, mS6 ¼ mS8 ¼ mΣ8 ≡ MI ¼ 5.2 × 106 GeV, and cot θH ¼ 50.

ðY QQ Þij
−LY ¼ ðY U Þij ϵαβ ūRâi H α qâβ þ ðY D Þij d̄Râi Hα qâα þ ðY E Þij ēRi Hα lαLj þ 3 −câα a β ab̂ ĉγ
ϵâ b̂ ĉ ϵαβ q̄Li ðσ Þ γ S3 qLj
Lj Lj
2
câγ α a β
ðY QQ
6 Þij
þ ðY QL
3 Þij ϵαβ q̄Li ðσ a Þ γ S3â lLj þ ϵαβ q̄Li ðη Þâ b̂ SA6  qb̂β
câα A
Lj þ ðY 6 Þij d̄Râi ðη Þ S6 uRb̂j
DU A â b̂ A c
2
A â Aα b̂β A â A b̂α
þ ðY UQ
8 Þij ϵαβ ūRâi ðλ Þ b̂ S8 qLj þ ðY 8 Þij d̄Râi ðλ Þ b̂ S8α qLj þ H:c:;
DQ
ð19Þ

where the first three terms lead to the fermion mass terms which can lead to realistic Yukawa matrices at the low
after the Higgs field H acquires a VEV at the EW scale. The energy. Moreover, the GUT-scale matching conditions for
45 scalar plays an essential role in reproducing the masses the other couplings in Eq. (19) read as
of the SM fermions. If the 45 scalar is absent, the Yukawa
matrices must obey a condition Y E ¼ V TQE Y TD V DL at the 1 U 1 U 1 T U
3 ¼ Y 45 ;
Y QQ 6 ¼ − pffiffiffi Y 45 ;
Y QQ 8 ¼ − V QU Y 45 ;
Y UQ
GUT scale. This condition conflicts with the low-energy 2 2 2
values of the masses of the down-type quarks and the 1 1
3 ¼ − pffiffiffi Y 45 V DL ;
Y QL 6 ¼ ðY 45 Þ V QU ;
D
charged leptons. In the current model, this problem is Y DU D T
2 2 2
solved by the presence of the Yukawa coupling Y D 45 .
1
Because the SM-like Higgs field H is a mixture of Hð5Þ 8 ¼ pffiffiffi ðY 45 Þ :
Y DQ D T
ð21Þ
and Hð45Þ as in Eq. (12), the Yukawa matrices Y U , Y D , and 2 2
Y E are given at the GUT scale by
The scalar S3 couples to a quark and a lepton simulta-
neously and thus is a leptoquark. The RGEs for these
 rffiffiffi T couplings are given in Appendix C.
1 T 2 iδH Let us count the physical degrees of freedom in the
Y U ¼ − V QU cH Y 5 þU U
e sH Y 45 ;
2 3 Yukawa sector. In the general case, there are four Yukawa
 T
1 1 −iδH matrices Y U U D D
5 , Y 45 , Y 5 , and Y 45 in the GUT Lagrangian.
Y D ¼ − pffiffiffi cH Y 5 − pffiffiffi e
D D
sH Y 45 ; U U
Since Y 5 and Y 45 are symmetric and antisymmetric
2 2 6
 pffiffiffi  matrices, respectively, the four matrices contain 54 param-
1 T 3 −iδH
Y E ¼ − pffiffiffi V QE cH Y 5 þ pffiffiffi e
D D
sH Y 45 V DL ; ð20Þ eters in total. By the redefinitions of the fermion fields by
2 2 2 U10 and U 5 in Eq. (7), 18 degrees of freedom out of the

095012-6
FLAVOR PHYSICS IN SU(5) GUT WITH SCALAR FIELDS IN … PHYS. REV. D 108, 095012 (2023)

54 can be eliminated. Thus there remain 36 physical 0 1

1 0 0
parameters in the Yukawa matrices. Taking the basis where B QU C
the up-type quarks and the charged leptons are their mass V QU ¼ @ 0 eiα2 0 A: ð25Þ
iαQU
eigenstates, the Yukawa matrices Y U U D D
5 , Y 45 , Y 5 , and Y 45 are
0 0 e 3

written as
The other two matrices V DL and V QE can be parametrized as
1
YU5 ¼− ðV  Ŷ þ Ŷ U V †QU Þ; 0 1
cH QU U 1 0 0
pffiffiffi
3 B QE C
Y 45 ¼ pffiffiffi iδ
U
ðV QU Ŷ U − Ŷ U V †QU Þ; V QE ¼ V CKM @ 0 eiα2 0 AV̂ QE ;
2e sH H
0 0
QE
eiα3
1 0 1 0 1
YD5 ¼ − pffiffiffi ð3V CKM Ŷ D þ V QE Ŷ E V †DL Þ; 1 0 0 eiβ1
DL
0 0
2 2cH B C B C
V DL ¼ @ 0 0 AV̂ DL @ 0
DL
pffiffiffi 0 A;
DL
eiα2 eiβ2
3
Y 45 ¼ −iδH ðV CKM Ŷ D − V QE Ŷ E V †DL Þ; ð22Þ 0 0
DL
0 0
D DL
eiα3 eiβ3
e sH
ð26Þ
where Ŷ U , Ŷ D , and Ŷ E represent diagonal matrices in the
mass basis. It is noted that an overall phase in V QU and where V̂ QE and V̂ DL are the 3 × 3 unitary matrices para-
three phases in V QE (and/or V DL ) can be removed by metrized by three angles and one phase as the CKM matrix,
Uð1ÞB , Uð1Þe , Uð1Þμ , and Uð1Þτ transformations. The right- and V CKM is extracted in V QE .
hand sides of Eq. (22) then contain nine eigenvalues in Ŷ U , We define the coupling Ȳ QL3 in the mass basis of the
Ŷ D , and Ŷ E , three mixing angles and one phase in V CKM , down-type quarks and the charged leptons:
eight parameters in V QU , and fifteen ones in V QE and V DL .
In general, the scalar S3 can have two types of Yukawa pffiffiffi
6
couplings, Y QQ 3 and Y QL
3 , and the combination of these 3 ¼ V CKM Y 3 ¼ −
Ȳ QL T QL
ðŶ D V DL − V̄ QE Ŷ E Þ; ð27Þ
4e−iδH sH
couplings leads to baryon-number-violating dimension-six
operators, which cause too fast proton decay. For example,
the bound from p → π 0 eþ is estimated as where V̄ QE ¼ V †CKM V QE . The mixings in V̄ QE (V DL ) cause
flavor transitions between different generations of the
 
1 −25
mS3 2 down-type quarks (the charged leptons). To suppress
3 Þ12 ðY 3 Þ11 ðV CKM Þ2 j
jðY QQ QL
≲ 10 : ð23Þ dangerous contributions to flavor-changing processes asso-
2 TeV
ciated with the first generation [44,47], such as K → πνν̄
and μ− → e− γ, we assume that V̂ QE and V̂ DL have only the
Because ðY QL
3 Þ11 ∼ yd =sH with yd being the Yukawa mixing between the second and the third generations at the
coupling for down quark, this condition implies a strong
GUT scale:
upper bound on ðY U45 Þ12 :

 2 0 1
mS3 1 0 0
−20 B
45 Þ12 j ≲ 10
jðY U
2 TeV
sH : ð24Þ V̂ QE ¼ @ 0 cos θQE sin θQE C
A;
0 − sin θQE cos θQE
Other components in Y U 0 1
45 also have to be highly suppressed 1 0 0
to avoid the constraints from the proton decay. As B C
V̂ DL ¼ @ 0 cos θDL sin θDL A; ð28Þ
explained in Appendix C, the coupling Y QQ 3 in Eq. (19)
is forbidden in the whole range of the renormalization scale 0 − sin θDL cos θDL
by an accidental global symmetry Uð1ÞB × Uð1ÞL if Y QQ 3 is
once set to be zero at the GUT scale. Therefore, in the where the mixing angles θQE and θDL are varied from 0 to
following, we make an ansatz that Y U 45 ¼ 0 at the π=2. The three Yukawa matrices Y 10 , Y 5 , and Y D 45 are then
GUT scale. determined at the GUT scale by the thirteen input param-
We here show a parametrization for the mixing matrices eters in addition to Ŷ U , Ŷ D , Ŷ E , and V CKM , i.e., the two
V QU , V QE , and V DL . According to the matching condition mixing angles θQE and θDL , the nine phases in V QU , V QE ,
in Eq. (22), the ansatz Y U 45 ¼ 0 at the GUT scale requires and V DL , and the two parameters sH ¼ sin θH and δH in the
that V QU should be a diagonal phase matrix: Higgs sector.

095012-7
GOTO, MISHIMA, and SHINDOU PHYS. REV. D 108, 095012 (2023)

The Y QL
3 term in Eq. (19) is decomposed in terms of the
by inputting V QE and V DL , δH , and sH . The couplings Y U ,
fields in the EW broken phase as follows [36]: Y D , Y E , Y QL DU DQ
3 , Y 6 , and Y 8 are calculated at the GUT scale
pffiffiffi ¯ c QL with Eqs. (20) and (21), and we then perform the RG
LY ¼ −û¯ cL Y QL 1=3
3 êL S3 − 2d̂L Ȳ 3 êL S3
4=3 evolution from the GUT scale to the low scale. The fermion
pffiffiffi masses and the CKM elements at the low scale obtained
þ 2û¯ cL Y QL −2=3
− d̂¯ L Ȳ QL 1=3
c
3 ν̂L S3 3 ν̂L S3 þ H:c:; ð29Þ from this procedure are different from the original values due
to the effects from Y QL DU DQ
3 , Y 6 , and Y 8 . We iterate the RG
where the hatted quark and lepton fields represent the mass
running with the obtained values of Y QL DU DQ
3 , Y 6 , and Y 8
eigenstates as in Eq. (8), and SQ 3 denotes a charge eigenstate together with the original values of the SM fermion masses
with charge Q defined in the matrix form and the CKM elements until the difference in the masses and
0 1 the CKM elements becomes small enough. In this way we
1=3 4=3
p1ffiffi ðS Þ ðS Þ can determine a set of the GUT parameters that are consistent
1 B 2 3 ĉ 3 ĉ C
pffiffiffi ðσ A Þα β ðS3 ÞĉA ¼ @ A: ð30Þ with the low-energy values of the SM fermion masses, the
2 −2=3 1ffiffi 1=3
ðS3 Þĉ − 2 ðS3 Þĉ
p
CKM matrix elements, and the gauge couplings.
We fix the mass of the S3 leptoquark to be mS3 ¼ 2 TeV
to avoid constraints from high-pT searches at the LHC [57].
III. PHENOMENOLOGICAL ANALYSIS In addition, there are the thirteen arbitrary parameters: the
A. Input parameters three mixing angles θQE , θDL , and θH , and the ten phases
αQU
2 , α3 , α2 , α3 , α2 , α3 , β1 , β2 , β3 , and δH . In
QU QE QE DL DL DL DL DL
We study low-energy phenomenology of the SU(5) GUT
model proposed in the last section, where there is an S3 general the Yukawa couplings ðȲ QL 3 Þij in Eq. (27) become
leptoquark with a TeV-scale mass. As explained in Sec. II D larger for a smaller Higgs mixing angle θH . We choose
the S3 leptoquark has the Yukawa couplings with the cot θH ¼ 50 as a benchmark scenario, while the other
left-handed quarks and the left-handed leptons, which lead parameters are varied arbitrarily in their physical domain.
to rich flavor phenomenology at the low-energy scale. The S3 contributions to the flavor observables considered
In particular, the S3 couplings generate processes with below are reduced by taking a heavier mS3 and/or a
lepton-flavor violation (LFV) and lepton-flavor-universal- smaller cot θH.
ity violation (LFUV), while such exotic flavor processes
are severely constrained by experiments. Our aim is to B. Leptoquark couplings
investigate whether current and future flavor experiments The Yukawa couplings ðȲ QL3 Þij of the S3 leptoquark are
have a potential to explore our GUT-inspired scenario. The
constrained by the GUT relation in Eq. (27) to accom-
S3 Yukawa matrix Y QL 3 in our scenario cannot have an modate with the measured masses of the down-type quarks
arbitrary structure unlike that in phenomenological lepto- and the charged leptons at the low-energy scale. The RG
quark models where S3 is introduced by hand. The effects from the GUT scale MX to the S3 mass scale mS3 are
coupling Y QL D
3 originates from Y 45 in the GUT Lagrangian, shown in Fig. 2. It is noted that the magnitudes of the
and Y D45 also contribute to the SM Yukawa couplings Y D couplings typically enhance at the lower scale. In particular,
and Y E as in Eq. (20), which could help to explain the the 22 coupling is increased by about a factor of 2,
observed masses of the down-type quarks and the charged receiving one-loop corrections with the other couplings.
leptons. Thus, nontrivial correlations are expected among Therefore, the inclusion of the RG evolution is essential to
flavor observables where the S3 leptoquark contributes. study low-energy phenomenology associated with the
The parameters in the GUT model, such as the Yukawa 22 coupling, such as b → sμþ μ− processes.
couplings Y U D D
5 , Y 5 , and Y 45 and the mixing matrices V QE According to Fig. 2, the couplings with the second-
and V DL , are constrained by the low-energy values of the generation fermions are typically smaller than those with
SM fermion masses and the CKM matrix elements. We use the third-generation ones:
the fermion masses and the CKM matrix elements listed in
Table II as inputs, and calculate the running masses at the jðȲ QL
3 Þ22 j ≪ jðȲ 3 Þ23 j ∼ jðȲ 3 Þ32 j ≲ jðȲ 3 Þ33 j:
QL QL QL
ð31Þ
EW scale by taking into account QCD corrections for
quarks with RunDec [55,56] and one-loop QED correc- On the other hand, the couplings with the first-generation
tions for charged leptons [52]. The masses and the CKM fermions are negligibly small due to our ignorance of the
matrix elements as well as the gauge couplings at the EW corresponding mixings in Eq. (28).
scale are then evolved up to the GUT scale with the one-
loop RGEs in Appendix C, where the Yukawa couplings C. Matching onto low-energy theory
Y QL DU DQ
3 , Y 6 , and Y 8 are neglected at this stage. At the GUT The gauge couplings and the Yukawa couplings in
scale we calculate the couplings Y U D D
5 , Y 5 , Y 45 with Eq. (22) Eq. (19) at the GUT scale are evolved down to the mass

095012-8
FLAVOR PHYSICS IN SU(5) GUT WITH SCALAR FIELDS IN … PHYS. REV. D 108, 095012 (2023)

FIG. 2. Comparisons of the Yukawa couplings of the S3 leptoquark at the GUT scale μ ¼ M X and at the S3 mass scale μ ¼ mS3 , where
they are identical to each other on the dotted lines.

scale mS3 using the RGEs given in Appendix C, where S6 , where Y QL

3 in the right-hand side is the S3 Yukawa coupling
S8 , and Σ8 are decoupled at the intermediate scale MI . The at the S3 mass scale, obtained from the coupling at the GUT
leptoquark S3 is then decoupled at the scale mS3 , and the scale in Eq. (21) applying the RG evolution. We also define
theory is matched onto the Standard Model Effective Field ð1;3Þ
the coefficients C̄lq in the mass basis of the down-type
Theory (SMEFT). The corresponding tree-level matching
quarks and the charged leptons, and their matching con-
conditions are presented in Refs. [58,59], while the one-
ditions read as
loop ones are calculated in Ref. [60]. In addition, the one-
loop anomalous dimensions in the SMEFT are found in h i h i 3
ð1Þ ð3Þ
Refs. [61–63]. C̄lq ðmS3 Þ ¼ 3 C̄lq ðmS3 Þ ¼ ðȲ QL Þ ðȲ QL Þ ;
ijkl ijkl 4m2S3 3 ki 3 lj
We adopt the dimension-six SMEFT operators in the so-
called Warsaw basis [64], where the Lagrangian in the ð33Þ
SMEFT is given by the sum of the renormalizable SM
3 ¼ V CKM Y 3 .
where Ȳ QL
Lagrangian and termsP with higher-dimensional operators T QL

Oi : LSMEFT ¼ LSM þ i Ci Oi . At the tree level, only the The SMEFT coefficients in Eqs. (32) and (33) are
ð1Þ
semileptonic operators ½Olq ijkl ¼ ðlLi γ μ lLj Þðq̄Lk γ μ qLl Þ evolved down to the EW scale, at which the SMEFT is
ð3Þ matched onto the low-energy effective field theory
and ½Olq ijkl ¼ ðlLi γ μ σ a lLj Þðq̄Lk γ μ σ a qLl Þ are generated
(LEFT) [65] by integrating out the EW gauge bosons,
by integrating out the S3 leptoquark, where the correspond- the Higgs boson, and the top quark. The LEFT operators
ing Feynman diagram above the S3 mass scale is presented used in our phenomenological analysis are listed in
in Fig. 3(a), and that below the S3 mass scale, i.e., in the Eq. (D2). The tree-level matching conditions for the
SMEFT, is in Fig. 3(b). The tree-level matching conditions coefficients Li in the LEFT Lagrangian of Eq. (D1) can
for the semileptonic operators are given by be found in Refs. [65,66], while the one-loop ones are
h i h i 3
calculated in Ref. [67]. Moreover, the RGEs for Li are
ð1Þ ð3Þ
Clq ðmS3 Þ ¼ 3 Clq ðmS3 Þ ¼ ðY QL Þ ðY QL Þ ; calculated at the one-loop level in Refs. [68,69]. We
ijkl ijkl 4m2S3 3 ki 3 lj decompose Li into the sum of SM and new physics
ð32Þ (NP) contributions as Li ¼ LSM i þ Li . In the current
NP

model only the semileptonic operators with the left-handed

fermions are generated through the tree-level matching. For
example, we have the following coefficients at the weak
scale μ ¼ mZ :
h iNP h i h i
ð1Þ ð3Þ
LV;LL
νd ðm Z Þ ¼ C̄ lq ðmZ Þ − C̄ lq ðmZ Þ ; ð34Þ
ijkl ijkl ijkl
h iNP h i h i
ð1Þ ð3Þ
LV;LL
ed ðm Z Þ ¼ C̄ lq ðmZ Þ þ C̄lq ðmZ Þ ; ð35Þ
ijkl ijkl ijkl
h iNP h i
ð3Þ
νedu ðmZ Þ
LV;LL ¼ 2V wk Clq ðmZ Þ ; ð36Þ
FIG. 3. Diagrams for the tree-level matching at the S3 mass ijkl ijwl
scale. Corresponding diagrams (a) in the model above the S3
mass scale, and (b) in the SMEFT below the S3 mass scale. where V wk denotes a CKM matrix element.

095012-9
GOTO, MISHIMA, and SHINDOU PHYS. REV. D 108, 095012 (2023)

In our numerical analysis, we also include one-loop corrections to the matching onto the SMEFT, the RG evolution from
μ ¼ mS3 to μ ¼ mZ , the matching onto the LEFT, and the RG evolution from μ ¼ mZ to the lower energy scale. Let us
consider the LEFT coefficient LV;LL
ed for b → s processes as an example. Solving the RGEs in the leading-logarithmic
approximation, the coefficient LV;LL
ed is given at the bottom scale μ ¼ mb by

h iNP   2   2 
ðȲ QL
3 Þ2i ðȲ 3 Þ3j
QL
α mS3 g2 ð1 − 4c4W Þ mS3 11
LV;LL ðm Þ ¼ 1 − log þ log þ
m2S3 m2b 32π 2 c2W m2Z
b
ed
ij23 2π 6
    
y2t 
ðY QL
3 Þ3i ðY 3 Þ3j
QL

ðY QL
3 Þ3i ðȲ 3 Þ3j
QL
ðȲ QL
3 Þ2i ðY 3 Þ3j
QL
þ 2V ts V tb þ V ts þ V tb I ed ðxt Þ
64π 2 m2S3 m2S3 m2S3
 QL† QL QL† 
3ðN c þ 1Þ ðȲ 3 Ȳ 3 Ȳ 3 Þi2 ðȲ 3 Þ3j ðȲ 3 Þ2i ðȲ 3 Ȳ 3 Ȳ 3 Þ3j
QL QL QL QL† QL
− þ
8 ð4πÞ2 m2S3 ð4πÞ2 m2S3
  2 
5 ðȲ 3 Ȳ 3 Þij ðȲ 3 Ȳ 3 Þ32
QL† QL QL QL†
α ðȲ QL
3 Ȳ 3 Þ32
QL† mS3 19
− − δij log − ; ð37Þ
4 ð4πÞ2 m2S3 6π m2S3 m2b 12

where cW ¼ cos θW is the cosine of the Weinberg angle,  h iNP

π
yt represents the SM Yukawa coupling of the top quark, ½CNP
10A ðμÞij ¼ pffiffiffi − LV;LL
ðμÞ
2GF αV ts V tb
ed
ij23
xt ¼ m2t =m2W with mt and mW being the masses of the top 
h iNP
quark and the W boson, N c ¼ 3 is the number of colors, α
þ LV;LR
de ðμÞ ; ð41Þ
is the electromagnetic coupling, and I ed ðxÞ is the loop 23ij
function defined by
π h i
V;LL NP
 2 L ij ¼ pffiffiffi
½CNP L νd : ð42Þ
mS3 3ðx þ 1Þ x2 − 2x þ 4 2GF αV ts V tb ij23
I ed ðxÞ ¼ − log − þ log x:
m2W 2ðx − 1Þ ðx − 1Þ2
The argument μ is omitted in Eq. (42), since the coefficients
ð38Þ ½CNP
L ij and ½Lνd ij23 have no scale dependence. Let us
V;LL NP

consider the coefficients for the b → sμþ μ− transition.

In Eq. (37), the S3 couplings Y QL and Ȳ QL should be The coefficients ½CNP 9V ðμÞ22 and ½C10A ðμÞ22 in Eqs. (40)
3 3 NP
understood as those evaluated at the S3 mass scale. The and (41) are dominated by the LEFT coefficient
one-loop expressions for the other LEFT coefficients ½LV;LL
ed ðμÞ2223 generated at the tree level, while the con-
relevant to our analysis are given in Appendix D.
tributions from ½LV;LR
de ðμÞ3222 induced at the one-loop level
It is convenient to convert the LEFT coefficients of the
b → s semileptonic operators into the coefficients in the are subdominant. Hence, the approximate relation
weak Hamiltonian [70]: ½CNP
9V ðμÞ22 ≈ −½C10A ðμÞ22 holds in the current model [38].
NP

4G α  h
D. Constraints
HW ¼ − pffiffiFffi V ts V tb ½C9V ij ðŝ¯ L γ μ b̂L Þðê¯ i γ μ êj Þ
2 4π In the current model, the S3 leptoquark has sizable
þ ½C10A ij ðŝ¯ L γ μ b̂L Þðê¯ i γ μ γ 5 êj Þ couplings to quarks and leptons in the second and third
i generations. Strong constraints on the parameter space of
þ ½CL ij ðŝ¯ L γ μ b̂L Þðν̂¯ i γ μ ð1 − γ 5 Þν̂j Þ þ H:c:; ð39Þ the model come from the mass difference of Bs and B̄s
mesons denoted by ΔMs, the branching ratios for the B →
where GF is the Fermi constant, and the NP contributions to K ðÞ νν̄ decays, the LFUV tests in the B → K ðÞ lþ l−
the coefficients at the scale μ are related to the LEFT ones as (l ¼ e, μ) decays, and the branching ratio for the Bs →
μþ μ− decay. NP contributions to ΔM s are generated at the
h i h iNP
π one-loop level, while those to the others are at the tree level.
C9V ðμÞ ¼ pffiffiffi
NP
L V;LL
ðμÞ The current experimental data for these observables are
2GF αV ts V tb
ed
ij ij23
h iNP  summarized in Table III together with other relevant
þ LV;LR observables. For the B → K ðÞ lþ l− decays, we do not
de ðμÞ ; ð40Þ
23ij consider their branching ratios and the angular observables

095012-10
FLAVOR PHYSICS IN SU(5) GUT WITH SCALAR FIELDS IN … PHYS. REV. D 108, 095012 (2023)

TABLE III. Current measurements and future experimental sensitivities of flavor observables. The first column
represents the corresponding transition, and the second column shows the dominant coupling that induces the
transition, where Loop denotes a loop-level transition.

Transition Couplings Observable Current measurement Future sensitivity

þ − RKþ ½0.1; 1.1 þ0.090þ0.029
b → sμ μ ðȲ QL
3 Þ22 ðȲ 3 Þ32
QL
0.994−0.082−0.027
[78,79]
RK0 ½0.1; 1.1 þ0.093þ0.036
0.927−0.087−0.035
[78,79]
RKþ ½1.1; 6.0 þ0.042þ0.022
0.949−0.041−0.022
[78,79] 0.007 [80]
RK0 ½1.1; 6.0 þ0.072þ0.027
1.027−0.068−0.026
[78,79] 0.008 [80]
BðBs → μþ μ− Þ ð3.01  0.35Þ × 10−9 [81] 0.16 × 10−9 [80]
Loop 3 Þ23 ðȲ 3 Þ33
ðȲ QL QL ΔM s ð17.765  0.006Þ ps−1 [81]
b → sνν̄ 3 Þ23 ðȲ 3 Þ33
ðȲ QL QL BðBþ → K þ νν̄Þ < 1.6 × 10−5 (90%) [82] 11 % of SM [83]
BðB0 → K S νν̄Þ < 1.3 × 10−5 (90%)[84]
BðBþ → K þ νν̄Þ < 4.0 × 10−5 (90%) [85] 9.3 % of SM [83]
BðB0 → K 0 νν̄Þ < 1.8 × 10−5 (90%) [84] 9.6 % of SM [83]
b → cτ− ν̄ ðȲ QL
3 Þ23 ðȲ 3 Þ33
QL RðDÞ 0.357  0.029 [86] ð2.0  2.5Þ% [83]
RðD Þ 0.284  0.012 [86] ð1.0  2.0Þ% [83]
b → sτþ τ− ðȲ QL
3 Þ23 ðȲ 3 Þ33
QL BðBs → τþ τ− Þ < 5.2 × 10−3 (90%) [87] 5 × 10−4 [80]
BðBþ → K þ τþ τ− Þ < 2.25 × 10−3 (90%) [88] 2.0 × 10−5 [83]
BðB0 → K 0 τþ τ− Þ < 3.1 × 10−3 (90%) [89] 5.3 × 10−4 [90]
b → sμþ τ− ðȲ QL 
3 Þ23 ðȲ 3 Þ32
QL
BðBs → μ∓ τ Þ < 3.4 × 10−5 (90%) [91] 3 × 10−6 [80]
BðBþ → K þ μ− τþ Þ < 5.9 × 10−6 (90%) [92] 3.3 × 10−6 [83]
BðB0 → K 0 μ− τþ Þ < 1.0 × 10−5 (90%) [93]
b → sμ− τþ ðȲ QL 
3 Þ22 ðȲ 3 Þ33
QL BðBþ → K þ μþ τ− Þ < 2.45 × 10−5 (90%) [92] 3.3 × 10−6 [83]
BðB0 → K 0 μþ τ− Þ < 8.2 × 10−6 (90%) [93]
τ− → μ− s̄s 3 Þ22 ðȲ 3 Þ23
ðȲ QL  QL Bðτ− → μ− ϕÞ < 2.3 × 10−8 (90%) [94] 8.4 × 10−10 [95]
bb̄ → μ τ∓ ðȲ QL
3 Þ32 ðȲ 3 Þ33
QL
Bðϒð1SÞ → μ τ∓ Þ < 2.7 × 10−6 (90%) [96]
Bðϒð2SÞ → μ τ∓ Þ < 3.3 × 10−6 (90%) [97]
Bðϒð3SÞ → μ τ∓ Þ < 3.1 × 10−6 (90%) [97]
Loop ðȲ QL
3 Þ32 ðȲ 3 Þ33
QL Bðτ− → μ− γÞ < 4.2 × 10−8 (90%) [98] 6.9 × 10−9 [95]
Bðτ− → μ− μþ μ− Þ < 2.1 × 10−8 (90%) [99] 3.6 × 10−10 [95]
BðZ → μ∓ τ Þ < 6.5 × 10−6 (95%) [100] Oð10−9 Þ [101]

that exhibit some tensions with the SM [71], since they the one-loop level. Contributions from other coefficients
suffer from hadronic uncertainties [72–77]. with the right-handed quarks are suppressed by the small
For the mass difference ΔMs , we utilize the following quark masses and neglected here. We use the PDG average
formula that is normalized to the SM value: of the measurements for ΔM s [81], which gives a constraint
on the product of the S3 Yukawa couplings ðȲ QL 3 Ȳ 3 Þ32 .
QL†

ΔMs CLL;NP ðmb Þ Because of the hierarchy in the magnitudes of the cou-
¼ 1 þ bs
;
ΔM s
SM loop
RSM plings, the product is dominated by ðȲ QL 
3 Þ23 ðȲ 3 Þ33
QL
pffiffiffi QL  QL 
compared with ðȲ 3 Þ21 ðȲ 3 Þ31 and ðȲ 3 Þ22 ðȲ 3 Þ32 .
QL QL
2
Cbs ðmb Þ ¼ −
LL;NP
LV;LL ðmb Þ NP
2323 ; ð43Þ The product ðȲ QL
3 Ȳ 3 Þ32 is also constrained from the
QL†
4GF ðV tb V ts Þ2 dd
branching ratios for B → K ðÞ νν̄, which are calculated as
where the SM loop contribution RloopSM ¼ ð1.310  0.010Þ ×
−3
10 and the SM prediction ΔMSM þ0.7 −1 BðB → K ðÞ νν̄Þ 1 X jCSM
L δij þ ½CL ij j
NP 2
s ¼ ð18.4−1.2 Þ ps are ¼ ; ð44Þ
2
evaluated in Ref. [45]. Our analysis includes the theoretical BðB → K ðÞ νν̄ÞSM 3 ij jCSM
L j
uncertainty in ΔM SM s , which is much larger than the
2
experimental one. In the current model, the LEFT coef- where the SM coefficient is given by CSM L ¼ −X t =sW
ficient ½LV;LL
dd ðmb Þ2323 , given in Eq. (D10), is generated at
NP
with Xt ¼ 1.469 and s2W ¼ 1 − c2W , and the SM predictions

095012-11
GOTO, MISHIMA, and SHINDOU PHYS. REV. D 108, 095012 (2023)

0 0
FIG. 4. Left: constraints from ΔM s =ΔM SM s and BðB → K νν̄Þ. The gray region represents the predictions which are consistent with
the low-energy values of the gauge couplings and the fermion masses and mixing. The vertical bands in magenta correspond to the
experimental measurements at the one and two sigma ranges, and the horizontal lines are the 90% upper limit at Belle (black dashed line)
and the SM prediction (blue solid line). Right: constraints on Re½ðCNP9V Þ22  and Re½ðC10A Þ22  at the mb scale, where the oblique dotted line
NP

represents Re½ðC9V Þ22  ¼ −Re½ðC10A Þ22 . The magenta region can satisfy the experimental bounds from ΔM s and BðB0 → K 0 νν̄Þ,
NP NP

while the cyan region can satisfy further with RKþ ½1.1; 6.0, RK0 ½1.1; 6.0, and BðBs → μþ μ− Þ. These regions are overlaid on top of the
gray one, which corresponds to that in the left plot.

are BðBþ → K þ νν̄ÞSM ¼ ð3.98  0.43  0.19Þ × 10−6 , experimental searches for the violation of the lepton-flavor-
BðB0 → K 0 νν̄ÞSM ¼ ðτB0 =τBþ ÞBðBþ → K þ νν̄ÞSM , BðB0 → universality (LFU) in b → s semileptonic decays provide
K 0 νν̄ÞSM ¼ ð9.19  0.86  0.50Þ × 10−6 , and BðBþ → severe constraints on our scenario. The LFU ratios RH
K þ νν̄ÞSM ¼ ðτBþ =τB0 ÞBðB0 → K 0 νν̄ÞSM with τBþ and (H ¼ K þ ; K 0 ) are defined by
τB0 being the lifetimes of B mesons [102]. The NP R q2max þ μ− Þ
contribution CNP L is defined by Eq. (42), where the one- q2min
dq2 dBðB→Hμ
dq2
loop expression of the LEFT coefficient ½LV;LL RH ½q2min ; q2max  ¼ R q2 ; ð45Þ
νd ij23 is given
NP þ e− Þ

0 0
max
q2min
dq2 dBðB→He
dq2
in Eq. (D3). We select B → K νν̄ as a representative
of the B → K ðÞ νν̄ processes in our numerical analysis,
where q2min and q2max are given in units of GeV2 . For
where the use of the other processes gives similar results.3
example, approximate formulas for the region of
The upper limit on BðB0 → K 0 νν̄Þ is reported from the
1.1 GeV2 < q2 < 6.0 GeV2 are given in Ref. [104]:
Belle experiment [84], and provides a constraint on

ðȲ QL
3 Þ23 ðȲ 3 Þ33 .
QL
RK ½1.1; 6.0 ≈ 1.00 þ 0.23ReðΔCNP
9V Þ − 0.25ReðΔC10A Þ;
NP
In the left plot of Fig. 4, we present constraints in the
plane of ΔM s =ΔM SM 0 0
s and BðB → K νν̄Þ, where the gray
ð46Þ
region is obtained with the model parameters that are
consistent with the low-energy values of the gauge cou- RK ½1.1; 6.0 ≈ 1.00 þ 0.20ReðΔCNP
9V Þ − 0.27ReðΔC10A Þ;
NP

plings, the fermion masses, and the CKM matrix elements. ð47Þ
Here and hereafter, we take mS3 ¼ 2 TeV and cot θH ¼ 50
as well as the input parameters in Table II. A large portion where ΔCNP 9V ≡ ½C9V ðmb Þ22 − ½C9V ðmb Þ11 and ΔC10A ≡
NP NP NP
of the parameter space is excluded by the measurement of ½CNP
10A ðmb Þ22 − ½C10A ðmb Þ11 . These LFU ratios are calcu-
NP
ΔMs (magenta vertical bands) [81] and by the upper limit lated very accurately in the SM, where the hadronic
for BðB0 → K 0 νν̄Þ (black horizontal dashed line) [84], uncertainty is highly canceled by considering the ratios
where the two bands for ΔM s correspond to the one-sigma [105], and the QED correction provides a positive con-
and two-sigma regions. tribution to the ratios about less than 3% for 1 GeV2 <
Moreover, the measurements for the b → sμþ μ− proc-
q2 < 6 GeV2 [106,107]. The above approximate formulas
esses listed in Table III provide constraints on the product
are derived by neglecting the QED corrections. The
of the Yukawa couplings ðȲ QL 3 Þ22 ðȲ 3 Þ32 . In particular,
QL
theoretical uncertainties are negligible in our study. The
recent measurements at LHCb [79] listed in Table III
3 are compatible with the SM predictions. We adopt only
Very recently the Belle II collaboration has reported the first
evidence of the Bþ → K þ νν̄ decay as BðBþ → K þ νν̄Þ ¼ ð2.4  RK ½1.1; 6.0 and RK ½1.1; 6.0 as constraints, since the ratios
0.5þ0.5
−0.4 Þ × 10
−5
[103]. We do not take into account it in our in the low q2 regions RK ½0.1; 1.1 and RK ½0.1; 1.1 have
analysis. larger experimental uncertainties. In addition, we also

095012-12
FLAVOR PHYSICS IN SU(5) GUT WITH SCALAR FIELDS IN … PHYS. REV. D 108, 095012 (2023)

FIG. 5. Allowed region for the products of the Yukawa couplings of the S3 leptoquark at the S3 mass scale, where the cyan region
shows the parameter points that are consistent with ΔM s , RKþ ½1.1; 6.0, RK0 ½1.1; 6.0, and BðBs → μþ μ− Þ within two sigma and
BðB0 → K 0 νν̄Þ at 90% CL. The cyan regions are overlaid on top of the gray ones, which correspond to those in Fig. 4.

consider the branching ratio for the leptonic decay regions are overlaid on top of the gray one, which
Bs → μþ μ− , which is written simply with the NP contri- corresponds to that in the left plot.
bution to C10A : We also present allowed regions for the products of the
S3 Yukawa couplings at the S3 mass scale in Fig. 5. Here
½CNP 2 the cyan regions show the parameter points that are
10A ðmb Þ22
BðBs → μþ μ− Þ ¼ BðBs → μþ μ− ÞSM 1 þ ; consistent with ΔMs , RKþ ½1.1; 6.0, RK0 ½1.1; 6.0, and
C10A ðmb Þ
SM
BðBs → μþ μ− Þ within two sigma and BðB0 → K 0 νν̄Þ at
ð48Þ 90% CL. It is noted that the cyan regions are overlaid on top
of the gray regions that correspond to those in Fig. 4. The
where the SM values are BðBs → μþ μ− ÞSM ¼ ð3.65  magnitudes of the products in the upper row of Fig. 5
0.23Þ × 10−9 [108] and CSM are smaller than those in the lower row because of the
10A ðmb Þ ¼ −4.2 [109]. It is noted
that a nonvanishing decay width difference ΔΓs of the Bs hierarchy given in Eq. (31). The product ðȲ QL
3 Þ22 ðȲ 3 Þ32
QL

system has to be taken into account when comparing the is highly constrained by RKþ ½1.1; 6.0, RK0 ½1.1; 6.0, and
theoretical value calculated using Eq. (48) with the exper- BðBs → μþ μ− Þ, while ðȲ QL
3 Þ23 ðȲ 3 Þ33 is by ΔM s and
QL
0 0
imental data in Table III, since the time dependence of the BðB → K νν̄Þ. The other products are less constrained by
decay rate is integrated in the experiment [110,111]. This these observables.
gives only a minor effect on our numerical analysis. In the
current model, ½CNP 9V ðmb Þ22 and ½C10A ðmb Þ22 appearing in
NP
E. Predictions
RK ½1.1; 6.0, RK ½1.1; 6.0, and BðBs → μþ μ− Þ are domi- The S3 leptoquark can generate various LFV and LFUV
nated by the LEFT coefficient ½LV;LL ed ðmb Þ2223 , which is with the second- and third-generation fermions. Under
given in terms of the product of the S3 Yukawa couplings the constraints studied in Sec. III D, we here consider
ðȲ QL
3 Þ22 ðȲ 3 Þ32 at the tree level. the following observables: RðDðÞ Þ, BðBs → τþ τ− Þ,
QL

The right plot of Fig. 4 shows constraints on BðBs → μ∓ τ Þ, BðB → K ðÞ μ∓ τ Þ, BðϒðnSÞ → μ τ∓ Þ,
Re½CNP9V ðmb Þ22 and Re½C10A ðmb Þ22 . The magenta region Bðτ− → μ− ϕÞ, Bðτ− → μ− γÞ, Bðτ− → μ− μþ μ− Þ, and
NP

can satisfy the experimental bounds from ΔMs within two BðZ → μ∓ τ Þ. The first six observables receive tree-level
sigma and BðB0 → K 0 νν̄Þ at 90% confidence level (CL), contributions, while the rest are induced at the one-loop
while the cyan region can satisfy further RKþ ½1.1; 6.0, level. Figures 6 and 7 show predictions for these observ-
RK0 ½1.1; 6.0, and BðBs → μþ μ− Þ within two sigma. These ables in the current model. Here we only consider

095012-13
GOTO, MISHIMA, and SHINDOU PHYS. REV. D 108, 095012 (2023)

the resolution of the RðDðÞ Þ anomaly requires an extension

of the model [112,113]. We do not consider such a
possibility in the current paper.
Next, let us consider decay processes involving
b → sτþ τ− transition. The studies of NP contributions to
this transition are found, for example, in Refs. [114,115]. In
the current model, the contributions to the b → sτþ τ−
leptonic and semileptonic decays arise at the tree level

through the product ðȲ QL 3 Þ23 ðȲ 3 Þ33 . As in the case of
QL

Bs → μþ μ− in Eq. (48), the leptonic mode receives NP

contribution to C10A :

½CNP 2
FIG. 6. Predictions for RðDÞ and RðD Þ (denoted by the red 10A ðmb Þ33
BðBs → τþ τ− Þ ¼ BðBs → τþ τ− ÞSM 1 þ ;
points) with the HFLAV average of their experimental measure- C10A ðmb Þ
SM

ments at the levels of one sigma, two sigma, and three sigma ð51Þ
(denoted by the orange ellipses) and the SM values (denoted by
the black cross). Theoretical uncertainties associated with the SM
errors are not included in the predictions. where the SM prediction is BðBs → τþ τ− ÞSM ¼ ð7.73 
0.49Þ × 10−7 [108]. Moreover, the branching ratios of the
semileptonic modes in the large q2 region are calculated
flavor-changing-neutral-current processes except for in Ref. [115]:
RðDðÞ Þ, since the S3 effects on charged-current processes,
such as B0 → DðÞ− μþ ν and Dþ þ
s → μ ν, are not significant. BðB → Kτþ τ− Þ½15;22
In Fig. 6, we present the predictions for the ratios
¼ 10−7 ð1.20 þ 0.15Re½CNP
9V ðmb Þ33
RðDðÞ Þ ¼ BðB0 → DðÞ τþ νÞ=BðB0 → DðÞ lþ νÞ for l ¼
2
e, μ calculated under the constraints from ΔMs =ΔMSM s ,
− 0.42Re½CNP
10A ðmb Þ33 þ 0.02j½C9V ðmb Þ33 j
NP

BðB0 → K 0 νν̄Þ, RKþ ½1.1; 6.0, RK0 ½1.1; 6.0, and þ 0.05j½CNP 2
10A ðmb Þ33 j Þ; ð52Þ
BðBs → μþ μ− Þ. At the tree level RðDðÞ Þ are given by
BðB → K  τþ τ− Þ½15;19
ðÞ ðÞ
RðD Þ ≈ RðD ÞSM ð1 þ 2Re½CNP
V 1 ðmb Þ33 Þ; ð49Þ
¼ 10−7 ð0.98 þ 0.38Re½CNP
9V ðmb Þ33 − 0.14Re½C10A ðmb Þ33
NP

2 2
where we adopt the SM predictions RðDÞSM ¼ 0.298  þ 0.05j½CNP
9V ðmb Þ33 j þ 0.02j½C10A ðmb Þ33 j Þ;
NP
ð53Þ
0.004 and RðD ÞSM ¼ 0.254  0.005 [86]. The coefficient
CNP which are the averages of the charged and the neutral
V 1 is defined through the effective Lagrangian,
modes. The predicted branching ratios in the SM are of
4G Oð10−7 Þ [115]. The branching ratios for these leptonic and
Leff ¼ − pffiffiFffi V cb ðδij þ CNP ¯ μ ¯
V 1 ðmb Þ ij Þðb̂L γ ĉL Þðν̂Li γ μ êLj Þ; semileptonic modes can largely deviate from their SM
2 values. Figure 7(a) shows that BðBs → τþ τ− Þ can be as
1 large as Oð10−5 Þ, which is an order of magnitude smaller
V 1 ðmb Þ 33 ¼ − pffiffiffi νedu ðmb Þ 3332 ;
LV;LL ð50Þ
NP
CNP
2 2GF V cb than the future sensitivity at LHCb with 300 fb−1 [80].
Similarly, the predictions for BðB → K ðÞ τþ τ− Þ in the large
where we use the tree-level result for the LEFT coefficient q2 region can be enhanced by an order of magnitude, but it
½LV;LL
νedu ðmb Þ3332 given in Eq. (36). We keep only the 33
NP
is still much smaller than the future sensitivity at Belle II
component of CNP V 1 in Eq. (49), since the dominant NP with 50 fb−1 [83].
contributions arise in the 23, 32, and 33 ones in the current We also study the LFV processes b → sμþ τ− and b →
model and only the 33 one has an interference with the SM sμ τþ , which are generated through the products of the S3
−
 QL 
contribution. We use the average of the experimental data Yukawa couplings ðȲ QL 3 Þ23 ðȲ 3 Þ32 and ðȲ 3 Þ22 ðȲ 3 Þ33 ,
QL QL

by the Heavy Flavor Averaging Group (HFLAV) [86]. Here respectively. Because of the hierarchy in the magnitudes of
the b → cτ− ν̄ transition is dominated by the contribution the S3 Yukawa couplings presented in Eq. (31) and Fig. 5,

from the product ðȲ QL
3 Þ23 ðȲ 3 Þ33 , which also contributes to
QL
the relation jðȲ QL  QL 
3 Þ23 ðȲ 3 Þ32 j ≫ jðȲ 3 Þ22 ðȲ 3 Þ33 j holds
QL QL

b → sνν̄ and ΔMs . It is known that the S3 contribution that typically. At the LHC experiments, the branching ratio
explains the b → c anomaly is severely constrained by the for the leptonic decay is measured as a sum of the two
b → sνν̄ processes and ΔMs [40]. Consequently, the S3 channels Bs → μ− τþ and Bs → μþ τ− . The corresponding
contribution does not alter RðDðÞ Þ significantly, and thus theoretical formula is given by [116]

095012-14
FLAVOR PHYSICS IN SU(5) GUT WITH SCALAR FIELDS IN … PHYS. REV. D 108, 095012 (2023)

FIG. 7. Predictions on relevant flavor processes, where the colored regions satisfy the experimental bounds from RKþ ½1.1; 6.0,
RK0 ½1.1; 6.0, BðBs → μþ μ− Þ, and ΔM s within two sigma and BðB0 → K 0 νν̄Þ at 90% CL. The red and black dashed lines show the
present upper bound on each processes by LHC experiments and B factories, respectively, and the red and black dotted lines show the
sensitivities expected at the LHCb with 300 fb−1 and the eþ e− experiments (such as the Belle II with 50 ab−1 and FCC-ee), respectively.
The cyan regions in (a), (b), and (c) are excluded by the upper limit on BðBþ → K þ μ− τþ Þ at Belle.

BðBs → μ∓ τ Þ ¼ BðBs → μ− τþ Þ þ BðBs → μþ τ− Þ;

 
τBs f 2Bs mBs m2τ α2 G2F jV ts V tb j2 m2τ 2
¼ 1− 2
64π 3 mBs
 
2 2 2 2
× j½CNP 9V ðm b Þ23 j þ j½C NP
10A ðm b Þ 23 j þ j½CNP
9V ðmb Þ32 j þ j½CNP
10A ðm b Þ32 j ; ð54Þ

095012-15
GOTO, MISHIMA, and SHINDOU PHYS. REV. D 108, 095012 (2023)

where mτ and ττ are the mass and the lifetime of τ lepton; mBs , τBs , and f Bs are the mass, the lifetime, and the decay constant
of Bs meson; and the muon mass is neglected. As shown in Fig. 7(a), the prediction on BðBs → μ∓ τ Þ
can be as large as Oð10−5 Þ, which may be probed by the LHCb measurement with 300 fb−1 [80]. For the semileptonic
channels, approximate numerical formulas are given by [117]
  þ
2 τB
BðBþ → K þ μ− τþ Þ ¼ 10−9 12.5j½CNP 2
9V ðmb Þ32 j þ 12.9j½C10A ðmb Þ32 j
NP
; ð55Þ
τB0
  þ
2 τB
BðBþ → K þ μþ τ− Þ ¼ 10−9 12.5j½CNP 9V ðm Þ
b 23 j2
þ 12.9j½CNP
10A ðm Þ
b 23 j ; ð56Þ
τB0
 
BðB0 → K 0 μ− τþ Þ ¼ 10−9 22.1j½CNP 9V ðm Þ
b 32 j2
þ 20.6j½CNP
10A ðm Þ
b 32 j2
; ð57Þ
 
BðB0 → K 0 μþ τ− Þ ¼ 10−9 22.1j½CNP
9V ðm Þ
b 23 j2
þ 20.6j½CNP
10A ðm Þ
b 23 j2
: ð58Þ

It is noted that Bþ → K þ μ− τþ (B0 → K 0 μ− τþ ) and Bþ → Fig. 7 except for that of BðBs → μ∓ τ Þ. Figure 7(c)
K þ μþ τ− (B0 → K 0 μþ τ− ) receive contributions from shows a strong correlation between BðBs → μ∓ τ Þ and
jðȲ QL  QL 
3 Þ23 ðȲ 3 Þ32 j and jðȲ 3 Þ22 ðȲ 3 Þ33 j, respectively. We
QL QL BðBþ → K þ μ− τþ Þ, since both of them are induced mainly
 þ
here present results on B → K þ μ∓ τ , since future
þ by jðȲ QL
3 Þ23 ðȲ 3 Þ32 j. The current upper limit on BðB →
QL
þ − þ ∓ 
sensitivities at Belle II can be found for these processes K μ τ Þ directly leads to the limit on BðBs → μ τ Þ.
in Ref. [83]. As shown in Fig. 7(b), BðBþ → K þ μ− τþ Þ These correlations among the b → sμþ τ− and b → sμ− τþ
can be large enough to be observed at Belle II with observables can be explored by the combination of the
50 ab−1 , while BðBþ → K þ μþ τ− Þ is out of the reach of Belle II and the LHCb measurements.
Belle II. A part of the parameter space is already excluded Besides, we consider the LFV decays of heavy quarko-
by the current measurement of BðBþ → K þ μ− τþ Þ at nia, ϒðnSÞ → μ∓ τ (n ¼ 1, 2, 3). The branching ratios for
Belle, but it does not alter the other predictions in these processes are given by [118–120]

2 2
1 3mϒðnSÞ Led ðmϒðnSÞ Þ
V;LL
2333
BðϒðnSÞ → μ τ∓ Þ ¼ BðϒðnSÞ → eþ e− Þ ; ð59Þ
SM
2 8πα

where mϒðnSÞ is the mass of ϒðnSÞ, and the charged-lepton masses are neglected. From the bottom-right plot in
2
Fig. 5, we estimate the magnitude of the LEFT coupling as j½LV;LL ed ðmϒðnSÞ Þ2333 j ∼ jðȲ 3 Þ32 ðȲ 3 Þ33 j=mS3 ≲
QL QL

Oð10−8 GeV2 Þ. Therefore, the branching ratios are as large as Oð10−11 Þ, which are too small to be measured at current
and near-future experiments.
Furthermore, the S3 leptoquark contributions also induce LFV decays of tau lepton. At the tree level, the τ− → μ− ϕ decay with
τ → μ− s̄s transition is generated through the S3 exchange. The branching ratio for τ− → μ− ϕ is given by [121]
−

   
f 2ϕ m3τ ττ m2ϕ 2 2m2ϕ 2
Bðτ− → μ− ϕÞ ¼ 1− 2 1þ 2 ed ðmτ Þ
LV;LL 3222 þ LV;LR
ed ðmτ Þ 3222
128π mτ mτ
 
8e  16e2 m2ϕ 2
þ Re Leγ ðmτ Þ 23 ð Led ðmτ Þ
V;LL
3222 þ Led ðmτ Þ
V;LR
3222 þ 2 þ Leγ ðmτ Þ 23 ; ð60Þ
mτ 9m2ϕ m2τ

where mϕ and f ϕ are the mass and the decay constant of ϕ meson, e is the electric charge, and the LEFT coefficients
½LV;LL
ed ðmτ Þ3222 , ½Led ðmτ Þ3222 , and ½Leγ ðmτ Þ23 are given in Eqs. (D5), (D6), and (D9), respectively. In the current model, the
V;LR

branching ratio for τ− → μ− ϕ is not significantly enhanced due to the smallness of the ðȲ QL 3 Þ22 coupling in the tree-level
contribution. As shown in Fig. 7(d), Bðτ− → μ− ϕÞ might be observed at the Belle II experiment [83]. We also consider the loop-
induced LFV processes of tau lepton, τ− → μ− γ and τ− → μ− μþ μ− . The branching ratio for τ− → μ− γ is given by

m3τ ττ 2
Bðτ− → μ− γÞ ¼ ½Leγ ðmτ ÞNP
23 ; ð61Þ
4π

095012-16
FLAVOR PHYSICS IN SU(5) GUT WITH SCALAR FIELDS IN … PHYS. REV. D 108, 095012 (2023)

and that for τ− → μ− μþ μ− can be found, e.g., in Refs. [122,123]:

m5τ ττ 2 2
Bðτ− → μ− μþ μ− Þ ¼ 3
2 LV;LL
ee ðmτ Þ 3222 þ Lee ðmτ Þ 2232 þ Lee ðmτ Þ 3222
V;LL V;LR
1536π
8e  i
þ Re Leγ ðmτ Þ 23 2 LV;LL ee ðmτ Þ 3222 þ 2 Lee ðmτ Þ 2232 þ Lee ðmτ Þ
V;LL V;LR
3222
mτ
 
32e2 m2τ 11 2
þ 2 log 2 − Leγ ðmτ Þ23 : ð62Þ
mτ mμ 4

The LEFT coefficients ½LV;LL ee ðmτ Þ2232 ¼ ½Lee ðmτ Þ3222 ,

V;LL charged leptons are reproduced correctly by the
½LV;LR
ee ðmτ Þ3222 and ½Leγ ðmτ Þ23 , evaluated at the τ mass
NP Georgi-Jarlskog mechanism. In addition, we have shown
scale, are given in Eqs. (D7), (D8), and (D9), respectively. In that the three gauge couplings can be unified through the
the expression of Bðτ− → μ− μþ μ− Þ, contributions from the RG running under the constraint from proton decay, if S3 ,
RL and RR operators are neglected, since LFV occurs S6 , and S8 in Φ45 and Σ8 in the 24-dimensional scalar
dominantly in the left-handed leptons in the current model. representation Σ lie much below the GUT scale. In
The predictions for Bðτ− → μ− γÞ and Bðτ− → μ− μþ μ− Þ are particular, the mass of S3 , which is a scalar leptoquark,
shown in Figs. 7(d) and 7(e). They exhibit a strong correlation can be of the order of TeV.
with each other, but are slightly smaller than the planned The Yukawa couplings of the S3 leptoquark at the low-
sensitivities of Belle II with 50 fb−1 [83]. energy scale is constrained by the matching condition at the
In the current model, the muon anomalous magnetic GUT scale in Eq. (22). In our scenario, the S3 leptoquark
moment (known as the muon g − 2) is generated through couples strongly to the SM fermions in the second and third
 generations, where the magnitudes of the couplings obey
the product ðY QL 3 Þ32 ðY 3 Þ32 via the dipole coupling
QL
the hierarchy shown in Eq. (31) and Fig. 2. In particular, the
½Leγ ðmτ Þ22 . We find that this contribution is too small
NP
coupling ðȲ QL
3 Þ22 is suppressed compared with ðȲ 3 Þ23 ,
QL
to explain the long-standing tension between the measured
value and the SM prediction of the muon g − 2 [124,125]. ðȲ 3 Þ32 , and ðȲ 3 Þ33 . The smallness of ðȲ 3 Þ22 leads to
QL QL QL

The S3 leptoquark also affects W-boson and Z-boson the characteristic patterns of correlations among flavor
couplings with the SM fermions. We evaluate them with the observables.
one-loop expressions in Ref. [126], which include radiative We have investigated flavor phenomenology in this
corrections beyond the leading-logarithmic approximation. realistic GUT scenario with the S3 leptoquark at the TeV
The effects on the W-boson couplings are not significant to scale. We have derived constraints on the S3 Yukawa
be measured at the current and planned future experiments. couplings from ΔMs , BðB → K ðÞ νν̄Þ, RKðÞ , and BðBs →
We here present only the result for BðZ → μ∓ τ Þ, which μþ μ− Þ, where the results are shown in Fig. 5. We have then
is calculated with the formulas given in Appendix E. calculated various decays of B mesons, ϒðnSÞ, tau lepton,
Figure 7(f) shows a strong correlation between BðZ → and Z boson. In the current model, the RðDðÞ Þ anomaly
μ∓ τ Þ and Bðτ− → μ− γÞ. In our scenario, the BðZ → cannot be explained by the S3 contribution due to the strong
μ∓ τ Þ can be as large as Oð10−9 Þ. The present exper- constraints from ΔMs and BðB → K ðÞ νν̄Þ. The LFV
imental bounds are given by the LEP experiment as BðZ → processes Bs → μ∓ τ , Bþ → K þ μ− τþ , and τ− → μ− ϕ
μ∓ τ Þ < 1.2 × 10−5 [127] and the LHC experiment as may be observed at Belle II with 50 ab−1 and LHCb with
BðZ → μ∓ τ Þ < 6.5 × 10−6 [100]. On the other hand, the 300 fb−1 . It is noted that BðBþ → K þ μþ τ− Þ cannot reach
FCC-ee experiment has a sensitivity to Oð10−9 Þ [101]. the future sensitivity at Belle II unlike BðBþ → K þ μ− τþ Þ.
In the case that BðZ → μ∓ τ Þ is enhanced enough, Bðτ− → Therefore, the observation of Bþ → K þ μ− τþ together with
μ− γÞ is also significantly enhanced. the nonobservation of Bþ → K þ μþ τ− is a clear signal of
the current model. On the other hand, it is rather hard to
observe the other processes τ− → μ− γ, τ− → μ− μþ μ− , and
IV. SUMMARY
Z → μ∓ τ , and much more data are needed for their
We have constructed a realistic GUT model which observations.
addresses two serious issues in the minimal SU(5) GUT: In general, it is challenging to probe a GUT model, since
the realization of the gauge coupling unification and that the unification occurs at a very high-energy scale. The
of the flavor structures in the down-type-quark and the proton decay is a direct probe for GUT, but it has not been
charged-lepton sectors. By introducing a 45-dimensional observed yet. We have provided a well-motivated bench-
scalar representation Φ45 to the minimal SU(5) GUT, the mark scenario which may be able to be probed by the
Yukawa matrices of the down-type quarks and the precise measurements of the flavor observables at the

095012-17
GOTO, MISHIMA, and SHINDOU PHYS. REV. D 108, 095012 (2023)

Belle II and LHCb experiments. Besides, the S3 leptoquark ACKNOWLEDGMENTS

at the TeV scale can be directly searched for at the current This work was supported in part by the Japan Society
and future hadron-collider experiments. We thus conclude for the Promotion of Science (JSPS) KAKENHI Grants
that the precise flavor measurements as well as the direct No. 17K05429 (S. M.) and No. 20H00160 (T. S.).
searches for the S3 leptoquark play complementary roles to Computations were partially carried out on the computer
the searches for proton decay in probing our GUT scenario. system (sushiki) at YITP in Kyoto University.

APPENDIX A: SCALAR POTENTIAL AND MASSES

The scalar potential VðΣ; Φ5 ; Φ45 Þ in the SU(5)-symmetric renormalizable Lagrangian in Eq. (1) is given by

VðΣ; Φ5 ; Φ45 Þ ¼ V 24 þ V 5 þ V 45 þ V 24·5 þ V 24·45 þ V 5·45 þ V 24·5·45 ; ðA1Þ

where each term is defined as

ð1Þ ð2Þ
V 24 ¼ m224 trΣ2 þ χ 24 trΣ3 þ λ24 ðtrΣ2 Þ2 þ λ24 trΣ4 ; ðA2Þ

V 5 ¼ m25 Φ†5 Φ5 þ λ5 ðΦ†5 Φ5 Þ2 ; ðA3Þ

ð1Þ ð2Þ
V 45 ¼ m245 ðΦ†45 ÞABC ðΦ45 ÞBC † A BC 2 † A † D
A þ λ45 ½ðΦ45 ÞBC ðΦ45 ÞA  þ λ45 ðΦ45 ÞBC ðΦ45 ÞD ðΦ45 ÞEF ðΦ45 ÞA
BC EF

ð3Þ ð4Þ
þ λ45 ðΦ†45 ÞABC ðΦ45 ÞBF † D † A † D
A ðΦ45 ÞEF ðΦ45 ÞD þ λ45 ðΦ45 ÞBC ðΦ45 ÞF ðΦ45 ÞEA ðΦ45 ÞD
EC BC EF

ð5Þ ð6Þ
þ λ45 ðΦ†45 ÞABC ðΦ†45 ÞBAD ðΦ45 ÞEC FD † A † B
F ðΦ45 ÞE þ λ45 ðΦ45 ÞBC ðΦ45 ÞDE ðΦ45 ÞF ðΦ45 ÞA ;
CD EF
ðA4Þ

V 24·5 ¼ χ 5 Φ†5 ΣΦ5 þ að1Þ ðtrΣ2 ÞΦ†5 Φ5 þ að2Þ Φ†5 Σ2 Φ5 ; ðA5Þ

ð1Þ ð2Þ
V 24·45 ¼ χ 45 ðΦ†45 ÞABC ΣD † A ð1Þ 2 † A
A ðΦ45 ÞD þ χ 45 ðΦ45 ÞBC ΣD ðΦ45 ÞA þ b ðtrΣ ÞðΦ45 ÞBC ðΦ45 ÞA
BC C BD BC

þ bð2Þ ðΦ†45 ÞABC ðΣ2 ÞD ð3Þ † A 2 B DC ð4Þ † A

A ðΦ45 ÞD þ b ðΦ45 ÞBC ðΣ ÞD ðΦ45 ÞA þ b ðΦ45 ÞBC ΣA ΣD ðΦ45 ÞE
BC E B DC

þ bð5Þ ðΦ†45 ÞABC ΣCE ΣBD ðΦ45 ÞDE ð6Þ † A

A þ b ðΦ45 ÞBC ΣA ΣD ðΦ45 ÞE ;
B E DC
ðA6Þ

V 5·45 ¼ cð1Þ ðΦ†45 ÞABC ðΦ45 ÞBC † ð2Þ † † A BC D ð3Þ † † A

A ðΦ5 Φ5 Þ þ c ðΦ5 ÞA ðΦ45 ÞBC ðΦ45 ÞD ðΦ5 Þ þ c ðΦ5 ÞC ðΦ45 ÞA ðΦ45 ÞBD ðΦ5 Þ
BC D

† † † A †
þ ½cð4Þ ðΦ45 ÞBC AD ð5Þ
A ðΦ45 ÞB ðΦ5 ÞC ðΦ5 ÞD þ c ðΦ45 ÞBC ðΦ45 ÞD ðΦ45 ÞA ðΦ5 ÞE
BC DE

þ cð6Þ ðΦ†45 ÞABC ðΦ45 ÞBD †

A ðΦ45 ÞD ðΦ5 ÞE þ H:c:;
CE
ðA7Þ

V 24·5·45 ¼ χ̃ðΦ†5 ÞC ΣAB ðΦ45 ÞBC ð1Þ † 2 A BC ð2Þ †

A þ d ðΦ5 ÞC ðΣ ÞB ðΦ45 ÞA þ d ðΦ5 ÞD ΣC ΣB ðΦ45 ÞA þ H:c:
D A BC
ðA8Þ

The 24-representation scalar Σ gets the VEV as

0 1
2v24 0 0 0 0
B 0 2v24 0 0 0 C
B C
B C
hΣi ¼ B
B 0 0 2v24 0 0 C;
C ðA9Þ
B C
@ 0 0 0 −3v24 0 A
0 0 0 0 −3v24

ð1Þ ð2Þ
when the condition 2m224 þ 4ð30λ24 þ 7λ24 Þv224 − 3χ 24 v24 ¼ 0 is satisfied [128].4 From the potential, the squared masses
of the component fields in the scalars Σ, Φ5 , and Φ45 can be read at the tree level as

4
The minimization of the scalar potential for the 45 representation is studied in Refs. [129–131].

095012-18
FLAVOR PHYSICS IN SU(5) GUT WITH SCALAR FIELDS IN … PHYS. REV. D 108, 095012 (2023)

3 ð2Þ 15 ð2Þ 15
m2Σ1 ¼ −2m224 þ χ 24 v24 ; m2Σ3 ¼ 40λ24 v224 − χ 24 v24 ; m2Σ8 ¼ 10λ24 v224 þ χ 24 v24 ;
2 2 2
0 pffiffi 1
! 2 ð2Þ 2 ffiffi
5p3 ð1Þ ð2Þ 2
2
0 m̃5 þ 3 ã v 24 − ð d̃ þ 3d Þv24
mH B 2 2
 C
R†H 2
RH ¼ @ pffiffi  A;
0 mH 0 − 5pffiffi3
ðd̃ð1Þ
þ 3d Þv ð2Þ 2 2 7 ð2Þ 19 ð3Þ
m̃ þ b̃ þ b̃ þ 8b þ b þ b ð4Þ 17 ð5Þ 75 ð6Þ
v2
2 2 24 45 4 8 4 8 24
0 1
!
m2HC 0 m̃25 − 2ãð2Þ v224 − p5ffiffi2 ðd̃ð1Þ − 2dð2Þ Þv224
B C
R†S RS ¼ @   A;
0 m2S1 − p5ffiffi2 ðd̃ð1Þ
− 2d ð2Þ
Þv224 m̃245 þ 1 ð2Þ
þ − 3 ð3Þ
− 2b þ 17 ð4Þ ð5Þ 25 ð6Þ
v224
2 b̃ 4 b̃ 4 b 2 b
   
ð2Þ ð3Þ 9 ð4Þ ð2Þ ð3Þ 9 ð4Þ
m2S̃ 2
¼ m̃45 þ −2b̃ þ 3b̃ − b þ 8b ð5Þ
v224 ; 2 2
mR2 ¼ m̃45 þ 3b̃ − 2b̃ − b þ 3b ð5Þ
v224 ;
1 2 2
   
1 11
m2S3 ¼ m̃245 þ 3b̃ð2Þ þ b̃ð3Þ þ 3bð4Þ − 7bð5Þ v224 ; m2S6 ¼ m̃245 þ −2b̃ð2Þ − 2b̃ð3Þ þ bð4Þ þ 3bð5Þ v224 ;
2 2
 
ð2Þ 1 ð3Þ 1 ð4Þ
m2S8 2
¼ m̃45 þ −2b̃ þ b̃ þ b − 7b ð5Þ
v224 ; ðA10Þ
2 2

where the following combinations of the parameters are introduced:

 
2 2 ð1Þ ð2Þ 2 2 2 ð1Þ ð2Þ ð3Þ 3 ð4Þ ð5Þ
m̃5 ¼ m5 þ ð30a þ 6a Þv24 ; m̃45 ¼ m45 þ 30b þ 6b þ 6b − b þ b v224 ;
2
ð1Þ ð2Þ
χ5 χ 45 χ 45 χ̃
ãð2Þ ¼ að2Þ − ; b̃ð2Þ ¼ bð2Þ − ; b̃ð3Þ ¼ bð3Þ − ; d̃ð1Þ ¼ dð1Þ − ; ðA11Þ
v24 v24 v24 v24

and the rotation matrices RH and RS are given as in Eq. (12):

   
cH e−iδH sH cS e−iδS sS
RH ¼ ; RS ¼ : ðA12Þ
−eiδH sH cH −eiδS sS cS

The masses of Σ1 , Σ3 , and Σ8 can be freely chosen, since V 24 in Eq. (A8) contains a sufficient number of parameters. On the
other hand, the masses of the other scalars are constrained by the following relation:

−8ðs2H m2H þ c2H m2H0 Þ þ 6ðs2S m2HC þ c2S m2S1 Þ þ 6m2S̃ − 6m2R2 þ 9m2S3 þ 3m2S6 − 10m2S8 ¼ 0: ðA13Þ
1

APPENDIX B: MATCHING CONDITIONS AT THE GUT SCALE

Below the GUT scale, the Yukawa interactions are given in terms of the component fields in Eqs. (6), (9), and (10) as
follows:

−LY ¼ ðY U Þij ϵαβ ūRâi Hα qâβ  âα  α

Lj þ ðY D Þij d̄Râi H α qLj þ ðY E Þij ēRi H α lLj

þ ðY 0U Þij ϵαβ ūRâi H 0α qâβ 0 0 âα 0 0 α

Lj þ ðY D Þij d̄Râi H α qLj þ ðY E Þij ēRi H α lLj

β â c
ðY QQ
C Þij câα b̂ ĉβ
þ ðY QL
C Þij ϵαβ q̄Li H Câ lLj þ ðY C Þij ūRâi H C eRj þ ðY C Þij ϵ
câα UE DU â b̂ ĉ
d̄Râi HCb̂ ucRĉj þ ϵâ b̂ ĉ ϵαβ q̄Li HC qLj
2
ðY QQ Þij
þ ðY 1 Þij ϵαβ q̄Li S1â lLj þ ðY 1 Þij ūRâi S1 eRj þ ðY 1 Þij ϵ d̄Râi S1b̂ uRĉj þ 1
QL câα β UE â c DU â b̂ ĉ c câα b̂ ĉβ
ϵâ b̂ ĉ ϵαβ q̄Li S1 qLj
2
â c
ðỸ UU
1 Þij â b̂ ĉ âα β 
þ ðỸ ED
1 Þ ēRi S̃ d
1 Râj þ ϵ ūRâi S̃1b̂ ucRĉj þ ðY UL
2 Þij ϵαβ ūRâi R2 lLj þ ðY 2 Þij ēRi R2âα qLj
EQ âα
ij
2

095012-19
GOTO, MISHIMA, and SHINDOU PHYS. REV. D 108, 095012 (2023)

câγ α a β
ðY QQ
3 Þij câα a β ab̂ ĉγ
þ ðY QL
3 Þij ϵαβ q̄Li ðσ a Þ γ S3â lLj þ ϵâ b̂ ĉ ϵαβ q̄Li ðσ Þ γ S3 qLj
2
ðY QQ
6 Þij b̂β
þ ðY DU
6 Þ d̄Râi ðη Þ
A â b̂ A c
S u
6 Rb̂j þ ϵαβ q̄Li
câα A
ðη Þâ b̂ SA
6 qLj
ij
2
A â Aα b̂β
þ ðY UQ
8 Þij ϵαβ ūRâi ðλ Þ b̂ S8 qLj þ ðY 8 Þij d̄Râi ðλ Þ b̂ S8α qLj þ H:c:;
DQ A â A b̂α
ðB1Þ

where Y QQ QQ
C and Y 1 are symmetric matrices in the flavor space, while Ỹ UU QQ QQ
1 , Y 3 , and Y 6 are antisymmetric matrices:

C Þ ¼ YC ;
ðY QQ T QQ
ðY QQ
1 Þ ¼ Y1 ;
T QQ
ðỸ UU
1 Þ ¼ −Ỹ 1 ;
T UU
ðY QQ
3 Þ ¼ −Y 3 ;
T QQ
ðY QQ
6 Þ ¼ −Y 6 :
T QQ
ðB2Þ

The Yukawa couplings in Eq. (B1) are matched onto those in Eq. (3) at the tree level as
rffiffiffi !T rffiffiffi !T
1 T 2 iδH 1 T 2
Y U ¼ − V QU cH Y 5 þ U
e sH Y 45 ;U
Y 0U ¼ V QU e −iδH
sH Y 5 −
U
c Y U
;
2 3 2 3 H 45
 T  T
1 1 −iδH 1 1
Y D ¼ − pffiffiffi cH Y D 5 − p ffiffi ffi e s Y
H 45
D ; Y 0D ¼ pffiffiffi eiδH sH Y D 5 þ p ffiffi
ffi c Y
H 45
D ;
2 2 6 2 2 6
pffiffiffi ! pffiffiffi !
1 T 3 −iδH 1 3
Y E ¼ − pffiffiffi V QE cH Y D 5 þ pffiffiffi e sH Y D45 V DL ; Y 0E ¼ pffiffiffi V TQE eiδH sH Y D 5 − pffiffiffi cH Y 45 V DL ;
D
2 2 2 2 2 2
   
1 1 iδS 1 −iδS s Y D þ p 1
Y QL
C ¼ p ffiffi
ffi c Y
S 5
Dþ p ffiffi
ffi e s Y
S 45
D V
DL ; Y QL
1 ¼ p ffiffiffi −e S 5 ffiffiffi c Y D V
S 45 DL ;
2 2 2 2 2 2
1 T  pffiffiffi
−iδS s Y U V
 1 T  iδS pffiffiffi 
C ¼ V QU cS Y 5 − 2e
Y UE 1 ¼ − V QU e sS Y 5 þ 2cS Y 45 V QE ;
U
S 45 QE ; Y UE U U
2 2
 T  T
1 1 iδS 1 −iδS 1
Y DU
C ¼ p ffiffi
ffi −c Y
S 5
D
þ p ffiffiffi e s Y D
S 45 V QU ; Y DU
1 ¼ p ffiffi
ffi e s Y
S 5
D
þ p ffiffiffi c Y
S 45
D
V QU ;
2 2 2 2 2 2
1 1 1 T U 1 T D
C ¼ cS Y 5 ;
Y QQ Y QQ ¼ − eiδS sS Y U 1 ¼ pffiffiffi V QU Y 45 V QU ;
Ỹ UU 1 ¼ V QE Y 45 ;
Ỹ ED
U
1 5;
2 2 2 2
1 T U 1 T D 1 1 D
2 ¼ pffiffiffi V QE Y 45 ;
Y EQ 2 ¼ V QU Y 45 V DL ;
Y UL Y QQ
3 ¼ YU 45 ; 3 ¼ − pffiffiffi Y 45 V DL ;
Y QL
2 2 2 2 2
1 1 1 1
Y QQ
6 ¼ − pffiffiffi Y U45 ; Y DU
6 ¼ ðY D ÞT V QU ; Y UQ
8 ¼ − V TQU Y U 45 ; Y DQ
8 ¼ pffiffiffi ðY D 45 Þ :
T
ðB3Þ
2 2 45 2 2 2

APPENDIX C: RENORMALIZATION GROUP The RGEs of the Yukawa couplings Y ϕψ̄ψ 0 associated with
EQUATIONS
the interaction of the form ½Y ϕψ̄ψ 0 jk ψ̄ j ϕψ 0k are given by
The scale dependence of the gauge couplings is gov-
erned by the RGEs,
d 1
β gi Y ϕψ̄ψ 0 ¼ β ϕ ; ðC3Þ
dgi
¼ ; ðC1Þ d log μ ð4πÞ2 Y ψ̄ψ 0
d log μ ð4πÞ2
where gi ¼ gs , g and g0 , and βgi denotes the corresponding where the one-loop beta functions can generally be written
as [132–135]
beta function. The one-loop contributions to the beta
functions are given by X
  βY ϕ ¼ −3 g2i ½Ci2 ðψÞY ϕψ̄ψ 0 þ Y ϕψ̄ψ 0 Ci2 ðψ 0 Þ
X ϕ ψ̄ψ 0
βgi ¼ Bgi þ
SM
Bgi θðmϕ − μÞ g3i ; ðC2Þ i
1
ϕ
þ ½Y 2 ðψÞY ϕψ̄ψ 0 þ Y ϕψ̄ψ 0 Y 2 ðψ 0 Þ
2
where ϕ ¼ H0 , HC , S1 , S̃1 , R2 , S3 , S6 , S8 , Σ1 , Σ3 , and Σ8 ,
ϕ þ Y ϕψ̄ψ 0 ΘðϕÞ þ 2ΓY ϕ : ðC4Þ
and the coefficients BSM
gi and Bgi are listed in Table IV. ψ̄ψ 0

095012-20
FLAVOR PHYSICS IN SU(5) GUT WITH SCALAR FIELDS IN … PHYS. REV. D 108, 095012 (2023)

ϕ
TABLE IV. BSM
gi and Bgi for the RGEs of the gauge couplings.

BΣgi1 BΣgi3 BΣgi8

0
gi BSM
gi BH
gi BH
gi
C
BSgi1 BgS̃i1 BRgi2 BSgi3 BSgi6 BSgi8
gs −7 0 1=6 1=6 1=6 1=3 1=2 5=6 2 0 0 1=2
g −19=6 1=6 0 0 0 1=2 2 0 4=3 0 1=3 0
g0 41=6 1=6 1=9 1=9 16=9 49=18 1=3 2=9 4=3 0 0 0

Below we list explicit formulas for the Yukawa couplings defined in Eq. (B1)5: Y ϕψ̄ψ 0 ¼ Y U , Y D , Y E , Y 0U , Y 0D , Y 0E , Y QL UE
C , YC ,
QQ QL QQ ED UU EQ QL QQ QQ UQ DQ
Y DU UE DU UL DU
C , Y C , Y 1 , Y 1 , Y 1 , Y 1 , Ỹ 1 , Ỹ 1 , Y 2 , Y 2 , Y 3 , Y 3 , Y 6 , Y 6 , Y 8 , and Y 8 . In the beta functions, the
coupling Y ϕψ̄ψ 0 should be understood as Y ϕψ̄ψ 0 θðmϕ − μÞ by considering the decoupling of heavy particles, and for ϕ ¼
H; H0 ; HC and S1 , the term Y ϕψ̄ψ 0 ΘðϕÞ is replaced as

8
> H0 0
>
> ψ̄ψ 0 ΘðHÞ þ Y ψ̄ψ 0 ΘðH HÞ
YH for ϕ ¼ H;
>
>
>
< YH
0 0  0
for ϕ ¼ H 0 ;
ψ̄ψ 0 ΘðH Þ þ Y ψ̄ψ 0 ΘðH H Þ
H
Y ϕψ̄ψ 0 ΘðϕÞ → ðC5Þ
> YH S1 
> ψ̄ψ 0 ΘðH C Þ þ Y ψ̄ψ 0 ΘðS1 H C Þ for ϕ ¼ H C ;
C
>
>
>
> S1
: Y ψ̄ψ 0 ΘðS1 Þ þ Y H 
ψ̄ψ 0 ΘðH C S1 Þ for ϕ ¼ S1 :
C

(1) Gauge-boson-loop contributions:

X 4 3 1 X 4 4 X 4 1
g2i Ci2 ðqL Þ ¼ g2s þ g2 þ g02 ; g2i Ci2 ðuR Þ ¼ g2s þ g02 ; g2i Ci2 ðdR Þ ¼ g2s þ g02 ;
i
3 4 36 i
3 9 i
3 9
X 3 1 X
g2i Ci2 ðlL Þ ¼ g2 þ g02 ; g2i Ci2 ðeR Þ ¼ g02 ; ðC6Þ
i
4 4 i

where Ci2 ðψ c Þ ¼ Ci2 ðψÞ.

(2) Self-energy contributions to the fermions:

Y 2 ðqL Þ ¼ Y †U Y U þ Y †D Y D þ Y 0† 0 0† 0
U Y U þ Y D Y D þ Y C ðY C Þ þ 2Y C Y C þ Y 1 ðY 1 Þ þ 2Y 1
QL QL T QQ† QQ QL QL T QQ† QQ
Y1
16 UQ† UQ 16 DQ† DQ
þ Y EQ†
2 Y 2 þ 6Y 3
EQ
Y 3 þ 3Y QL
QQ† QQ
3 ðY 3 Þ þ 2Y 6
QL T
Y6 þ
QQ† QQ
Y Y8 þ Y8 Y8 ;
3 8 3
Y 2 ðuR Þ ¼ 2Y U Y †U þ 2Y 0U Y 0†
U þ Y UE UE†
C Y C þ 2ðY C Þ
DU T DU
Y C þ Y UE UE†
1 Y 1 þ 2ðY 1 Þ
DU T DU
Y 1 þ 2Ỹ UU UU†
1 Ỹ 1
32 UQ UQ†
þ 2Y UL UL†
2 Y2 þ 2ðY DU
6 Þ Y6
T DU
þ Y Y ;
3 8 8
32 DQ DQ†
Y 2 ðdR Þ ¼ 2Y D Y †D þ 2Y 0D Y 0†
D þ 2Y C Y C
DU DU†
þ 2Y DU DU†
1 Y1 þ ðỸ ED
1 Þ Ỹ 1
T ED þ 2Y DU Y DU† þ Y Y ;
6 6
3 8 8
Y 2 ðlL Þ ¼ Y †E Y E þ Y 0† 0
E Y E þ 3Y C Y C þ 3Y 1 Y 1 þ 3Y 2 Y 2 þ 9Y 3 Y 3 ;
QL† QL QL† QL UL† UL QL† QL

Y 2 ðeR Þ ¼ 2Y E Y †E þ 2Y 0E Y 0†
E þ 3ðY C Þ Y C
UE T UE
þ 3ðY UE
1 Þ Y1
T UE
þ 3Ỹ ED ED†
1 Ỹ 1 þ 6Y EQ EQ†
2 Y2 ; ðC7Þ

where Y 2 ðψ c Þ ¼ ½Y 2 ðψÞT .

5
The RGEs for the SM, S1 , and S3 Yukawa couplings were recently studied in Refs. [136,137].

095012-21
GOTO, MISHIMA, and SHINDOU PHYS. REV. D 108, 095012 (2023)

(3) Self-energy contributions to the scalars:

ΘðHÞ ¼ trð3Y †U Y U þ 3Y †D Y D þ Y †E Y E Þ; ΘðH0 Þ ¼ trð3Y 0† 0 0† 0 0† 0

U Y U þ 3Y D Y D þ Y E Y E Þ;

ΘðHC Þ ¼ trð2Y QL†

C Y C þ Y C Y C þ 2Y C Y C þ 2Y C Y C Þ;
QL UE† UE DU† DU QQ† QQ

ΘðS1 Þ ¼ trð2Y QL†

1 Y 1 þ Y 1 Y 1 þ 2Y 1
QL UE† UE
Y 1 þ 2Y QQ†
DU† DU
1 1 Þ;
Y QQ
ΘðS̃1 Þ ¼ trðỸ ED†
1 Ỹ 1 þ Ỹ 1
ED
Ỹ 1 Þ;
UU† UU
ΘðR2 Þ ¼ trðY UL†
2 Y 2 þ Y 2 Y 2 Þ;
UL EQ† EQ

ΘðS3 Þ ¼ trð2Y QL†

3 Y 3 þ 2Y 3
QL
Y 3 Þ;
QQ† QQ
ΘðS6 Þ ¼ trðY DU†
6 6 þ Y6
Y DU Y 6 Þ;
QQ† QQ

ΘðS8 Þ ¼ trð2Y DQ†

8 8 þ 2Y 8
Y DQ Y 8 Þ;
UQ† UQ
ΘðH H0 Þ ¼ trð3Y †U Y 0U þ 3Y D Y 0† 0†
D þ Y E Y E Þ;

ΘðS1 HC Þ ¼ trð2Y QL†

1 YC þ Y1 YC
QL UE UE†
þ 2Y DU†
1 C þ 2Y 1 Y C Þ;
Y DU QQ QQ†
ðC8Þ

where Θðϕ Þ ¼ ΘðϕÞ, ΘðH0 HÞ ¼ ½ΘðH H0 Þ , and ΘðHC S1 Þ ¼ ½ΘðS1 HC Þ .
(4) Vertex corrections:

ð0Þ† ð0Þ† ð0Þ ð0Þ ð0Þ

ΓY ð0Þ ¼ −Y U Y D Y D − Y 0U Y D Y 0D − Y UE
C Y E ðY C Þ þ 2ðY C Þ Y D Y C − Y 1 Y E ðY 1 Þ
QL T DU T QQ UE QL T
U

ð0Þ ð0Þ† ð0Þ 16 UQ ð0Þ† DQ

þ 2ðY DU
1 Þ Y D Y 1 − Y 2 Y E Y 2 þ 2ðY 6 Þ Y D Y 6 −
T QQ UL EQ DU T QQ
Y Y Y ;
3 8 D 8
ð0Þ† ð0Þ† ð0Þ ð0Þ ð0Þ 16 DQ ð0Þ† UQ
ΓY ð0Þ ¼ −Y D Y U Y U − Y 0D Y U Y 0U þ 2Y DU
C Y U Y C þ 2Y 1 Y U Y 1 − 2Y 6 Y U Y 6 −
QQ DU QQ DU QQ
Y Y Y ;
D 3 8 U 8
ð0Þ ð0Þ ð0Þ†
ΓY ð0Þ ¼ −3ðY UE
C Þ Y U Y C − 3ðY 1 Þ Y U Y 1 − 3Y 2 Y U Y 2 ;
T QL UE T QL EQ UL
E

ΓY QL ¼ −Y TU Y UE
i Y E − Y 0T UE 0
U Yi Y E − 2Y QQ Y C − 2Y QQ
QQ† QL
C Yi Y 1 þ ðY EQ
QQ† QL
1 Yi 2 Þ Yi
T UE† UL
Y2
i

þ 6Y QQ Y 3 ði ¼ C; 1Þ;
QQ† QL
3 Yi

ΓY UE
i
¼ −2Y U Y QL
i Y TE − 2Y 0U Y QL
i Y 0T
E − 2ðY C Þ Y i Y C − 2ðY DU
DU T DU UE
1 Þ Yi Y 1 − 2Ỹ UU
T DU UE DU†
1 Yi ðỸ ED
1 Þ
T

þ 2Y UL QL†
2 Yi ðY EQ
2 Þ ði ¼ C; 1Þ;
T

ΓY DU
i
¼ 2Y D Y QQ†
i Y TU þ 2Y 0D Y QQ†
i Y 0T
U − YC Yi
DU UE
ðY UE
C Þ − Y1 Yi
T DU UE
ðY UE
1 Þ − ðỸ 1 Þ Y i
T ED T UE† UU
Ỹ 1
16 DQ QQ† UQ T
− Y Y ðY 8 Þ ði ¼ C; 1Þ;
3 8 i
ΓY QQ ¼ Y TD Y DU
i Y U þ Y TU Y DU†
i Y D þ Y 0T DU 0
D Yi Y U þ Y 0T DU† 0
U Yi Y D − Y QL Y C − Y QQ
QL† QQ
C Yi
QL
C Yi ðY QL
C Þ
T
i

− Y QL
1 Yi Y 1 − Y QQ
QL† QQ QL
1 Yi ðY QL
1 Þ − 3Y 3 Y i
T QQ QL
ðY QL
3 Þ þ 3Y 3 Y i
T QL QL† QQ
Y3
8 8 DQ T DU UQ
− ðY UQ ÞT Y DU† 8 − ðY 8 Þ Y i
Y DQ Y 8 ði ¼ C; 1Þ;
3 8 i
3
ΓỸ ED
1
¼ 2ðY UE
C Þ Ỹ 1
T UU†
ðY DU
C Þ þ 2ðY 1 Þ Ỹ 1
T UE T UU†
ðY DU
1 Þ ;
T

ΓỸ UU ¼ −Y UE
C Ỹ 1 Y C þ ðY C Þ Ỹ 1 ðY C Þ − Y 1 Ỹ 1 Y 1 þ ðY 1 Þ Ỹ 1 ðY 1 Þ ;
ED DU DU T ED† UE T UE ED DU DU T ED† UE T
1

0 EQ† 0
ΓY UL
2
¼ −Y U Y EQ†
2 YE − YUY2 YE þ YC Y2 YC þ Y1 Y2 Y1 ;
UE EQ QL UE EQ QL

0 UL† 0
ΓY EQ ¼ −Y E Y UL†
2 Y U − Y E Y 2 Y U þ ðY C Þ Y 2 ðY C Þ þ ðY 1 Þ Y 2 ðY 1 Þ ;
UE T UL QL T UE T UL QL T
2

ΓY QL ¼ 2Y QQ
C Y3 Y C þ 2Y QQ
QQ† QL
1 Y3 Y 1 þ 2Y QQ
QQ† QL QQ† QL
3 Y3 Y3 ;
3

ΓY QQ ¼ −Y QQ
C Y 3 ðY C Þ þ Y C Y 3 Y C − Y 1 Y 3 ðY 1 Þ þ Y 1 Y 3 Y 1 þ Y 3 Y 3 ðY 3 Þ
QL QL T QL QL† QQ QQ QL QL T QL QL† QQ QQ QL QL T
3

þ Y QL QL† QQ
3 Y3 Y3 ;
8 DQ QQ† UQ T
ΓY DU ¼ −2Y D Y QQ†
6 Y TU − 2Y 0D Y QQ†
6 Y 0T
U − Y8 Y6 ðY 8 Þ ;
6 3

095012-22
FLAVOR PHYSICS IN SU(5) GUT WITH SCALAR FIELDS IN … PHYS. REV. D 108, 095012 (2023)

4
ΓY QQ ¼ −Y TU Y DU† Y D þ Y TD Y DU Y U − Y 0T DU† 0
U Y6 Y D þ Y 0T DU 0
D Y6 Y U − ðY UQ ÞT Y DU† Y DQ
6
6 6
3 8 6 8

4
þ ðY DQ ÞT Y DU Y UQ
8 ;
3 8 6

1
ΓY UQ ¼ −Y U Y DQ†
8 Y D − Y 0U Y DQ†
8 Y 0D − ðY DU ÞT Y DQ
8 Y QQ − ðY DU
1 ÞT Y DQ
8 Y QQ
1 þ ðY DU ÞT Y DQ
8 Y QQ
8
C C
2 6 6

2
þ Y UQ Y DQ† Y DQ
8 ;
3 8 8
1
ΓY DQ ¼ −Y D Y UQ†
8 Y U − Y 0D Y UQ†
8 Y 0U − Y DU
C Y8 Y C − Y DU
UQ QQ
1 Y8 Y 1 − Y DU
UQ QQ
Y UQ Y QQ
8 2 6 8 6

2
þ Y DQ Y UQ† Y UQ
8 : ðC9Þ
3 8 8

Notice that the coupling Y QQ APPENDIX D: LEFT LAGRANGIAN

3 is shown to be vanishing in
the whole range of the renormalization scale below the The LEFT Lagrangian is given by
3 ¼ 0 and H C , S1 , and S̃1 decouple at the
GUT scale if Y QQ
X
GUT scale. In fact, the Yukawa interaction Lagrangian in LLEFT ¼ LQCDþQED þ Li Qi ; ðD1Þ
Eq. (B1) becomes invariant under Uð1ÞB and Uð1ÞL i
separately if the terms with HC , S1 , and S̃1 are removed
and if Y QQ
3 is set to vanish. The latter condition is satisfied if where LQCDþQED is the renormalizable QCD and QED
Y 45 ¼ 0 at the GUT scale in the tree-level approximation.
U
Lagrangian with the SM fermions except for top quark. The
The appropriate assignment of the baryon and lepton LEFT operators Qi up to dimension six are classified in
numbers in the above case is listed in Table V. Ref. [65]. The operators relevant to the current study are

QV;LL
νd ijkl ¼ ðν̂¯ Li γ μ ν̂Lj Þðd̂¯ Lk γ μ d̂Ll Þ; QV;LL
ed ijkl ¼ ðê¯ Li γ μ êLj Þðd̂¯ Lk γ μ d̂Ll Þ;
¯ μ ¯
ee ijkl ¼ ðêLi γ êLj ÞðêLk γ μ êLl Þ;
QV;LL ½QV;LR
ee ¼ ðê¯ Li γ μ êLj Þðê¯ Rk γ μ êRl Þ;
ijkl

QV;LL
dd ijkl ¼ ðd̂¯ Li γ μ d̂Lj Þðd̂¯ Lk γ μ d̂Ll Þ; Qeγ ij ¼ ðê¯ Li σ μν êRj ÞFμν ;

QV;LR
ed ijkl ¼ ðê¯ Li γ μ êLj Þðd̂¯ Rk γ μ d̂Rl Þ; QV;LR
de ijkl ¼ ðd̂¯ Li γ μ d̂Lj Þðê¯ Rk γ μ êRl Þ;

QV;LL
νedu ijkl ¼ ðν̂¯ Li γ μ êLj Þðd̂¯ Lk γ μ ûLl Þ; ðD2Þ

where there also exist the Hermitian conjugates of the non- one-loop level. The one-loop matching formulas
self-conjugate operators. from a model with the S1 and S3 leptoquarks to
The Wilson coefficients for these operators are calcu- the SMEFT are listed in Refs. [58–60].
lated as follows. (2) The RG running effects of the SMEFT operators are
(1) The S3 field is integrated out at the S3 mass scale, taken into account. The anomalous dimensions for
and the model is matched onto the SMEFT at the the dimension-six operators in the SMEFT are listed
in Refs. [61–63].
(3) The SMEFT is matched onto the LEFT at the weak
TABLE V. Assignment of the baryon and lepton numbers to the scale. The one-loop matching formulas are listed in
scalars, where the Yukawa interactions in Eq. (B1) are invariant
Refs. [65–67].
under Uð1ÞB and Uð1ÞL separately if HC , S1 , and S̃1 decouple and
(4) The RG effects in the LEFT are taken into account.
Y QQ
3 is set to vanish at the GUT scale.
The corresponding anomalous dimensions are given
R2 S3 S6 S8 in Refs. [68,69].
The coefficients in the LEFT Lagrangian at the relevant
3B −1 1 −2 0 scale for the process under consideration are given in the
L 1 1 0 0
leading-logarithmic approximation by

095012-23
GOTO, MISHIMA, and SHINDOU PHYS. REV. D 108, 095012 (2023)

   2    2 
ðȲ QL
3 Þ2i ðȲ 3 Þ3j
QL
g2 ð1 þ 2c2W Þ mS3 11 3g2 mS3 11
LV;LL
νd
NP
¼ 1þ log þ − log þ
ij23
2m2S3 32π 2 c2W m2Z 6 16π 2 m2W 6
    
y2 ðY Þ3i ðY 3 Þ3j 1  ðY 3 Þ3i ðȲ 3 Þ3j
QL QL QL QL
ðȲ 3 Þ2i ðY 3 Þ3j
QL QL
þ t 2 4V ts V tb 3 þ V þ V I νd ðxt Þ
m2S3 2 ts m2S3 m2S3
tb
64π
 QL† QL QL† 
3ðN c þ 1Þ ðȲ 3 Ȳ 3 Ȳ 3 Þi2 ðȲ 3 Þ3j ðȲ 3 Þ2i ðȲ 3 Ȳ 3 Ȳ 3 Þ3j
QL QL QL QL† QL
− þ
16 ð4πÞ2 m2S3 ð4πÞ2 m2S3
1 ðȲ 3 Ȳ 3 Þij ðȲ QL
3 Ȳ 3 Þ32
QL† QL QL†
− ; ðD3Þ
4 ð4πÞ2 m2S3
  2 
α ðȲ QL
3 Ȳ 3 Þ32
QL† mS3 19
de ðmb Þ23ij
½LV;LR NP
¼ −δij log − ; ðD4Þ
6π m2S3 m2b 12


 2   2 
ðȲ QL
3 Þ23 ðȲ 3 Þ22
QL
α mS3 g2 ð1 − 4c4W Þ mS3 11
ed ðmτ Þ3222 ¼
½LV;LL NP
1 − log þ log þ
m2S3 2π m2τ 32π 2 c2W m2Z 6
 
y2t  ðY QL 
3 Þ33 ðY 3 Þ32
QL QL 
 ðY 3 Þ33 ðȲ 3 Þ22
QL
ðȲ QL 
3 Þ23 ðY 3 Þ32
QL
þ 2V ts V ts þ V ts þ V ts I ed ðxt Þ
64π 2 m2S3 m2S3 m2S3
 

3ðN c þ 1Þ ðȲ QL† 3 Ȳ 3 Ȳ 3 Þ32 ðȲ 3 Þ22
QL QL† QL
ðȲ QL
3 Þ23 ðȲ 3 Ȳ 3 Ȳ 3 Þ22
QL QL† QL
− þ
8 ð4πÞ2 m2S3 ð4πÞ2 m2S3
 QL  QL  2 
5 ðȲ QL†
3 Ȳ 3 Þ32 ðȲ 3 Ȳ 3 Þ22
QL QL QL†
α 2 ðY 3 Þ33 ðY 3 Þ32 mt 3 ðY QL†
3 Y 3 Þ32
QL
− − N c Qd log 2 −
4 ð4πÞ2 m2S3 6π m2S3 mb 4 m2S3
  2 
ðY QL Þ ðY QL Þ mS3
− N c ðI 3dL − Qd s2W Þy2t 3 332 32 32 log − 1 ; ðD5Þ
ð4πÞ mS3 m2t
 QL  QL  2 
α 2 ðY 3 Þ33 ðY 3 Þ32 mt 3 ðY QL†
3 Y 3 Þ32
QL
½LV;LR
ed ðmτ Þ3222
NP
¼ − N c Qd log 2 −
6π m2S3 mb 4 m2S3
QL    2 
2 ðY 3 Þ33 ðY 3 Þ32
QL mS3
2
− N c ð−Qd sW Þyt log −1 ; ðD6Þ
ð4πÞ2 m2S3 m2t

5N c ðY QL†
3 Y 3 Þ32 ðY 3 Y 3 Þ22
QL QL† QL
½LV;LL
ee ðmτ Þ3222 ¼ −
8 ð4πÞ2 m2S3
 QL  QL  2 
α ðY 3 Þ33 ðY 3 Þ32 mt 3 ðY QL†
3 Y 3 Þ32
QL
− N Q Q log 2 −
12π c d e m2S3 mb 4 m2S3
  2 
N ðY QL Þ ðY QL Þ mS3
− c ðI 3eL − Qe s2W Þy2t 3 332 32 32 log − 1 ; ðD7Þ
2 ð4πÞ mS3 m2t
 QL  QL  2 
α ðY 3 Þ33 ðY 3 Þ32 mt 3 ðY QL†
3 Y 3 Þ32
QL
ee ðmτ Þ3222
½LV;LR ¼ − N c Qd Qe log 2 −
6π m2S3 mb 4 m2S3

  2 
2 2 ðY QL
3 Þ33 ðY 3 Þ32
QL mS3
− N c ð−Qe sW Þyt log −1 ; ðD8Þ
ð4πÞ2 m2S3 m2t

eN c mej ðY QL†
3 Y 3 Þij
QL
½Leγ ðmτ ÞNP ¼ ; ðD9Þ
ij
8 ð4πÞ2 m2S3

095012-24
FLAVOR PHYSICS IN SU(5) GUT WITH SCALAR FIELDS IN … PHYS. REV. D 108, 095012 (2023)

APPENDIX E: Z → μ∓ τ 
5 ðȲ QL
3 Ȳ 3 Þ32 ðȲ 3 Ȳ 3 Þ32
QL† QL QL†
½LV;LL
dd ðmb Þ2323
NP
¼− ; ðD10Þ
8 ð4πÞ2 m2S3 The S3 affects the Z-boson effective couplings with
charged leptons which are defined as
where Qd ¼ −1=3, Qe ¼ −1, I 3dL ¼ I 3eL ¼ −1=2, and
e
I νd ðxÞ is the loop function defined by L¼ Z ½ē γ ðge Þ e þ ēRi γ μ ðgeR Þij eRj ; ðE1Þ
sW cW μ Li μ L ij Lj
 2
mS3 3ðx þ 1Þ x2 þ 10x − 8
I νd ðxÞ ¼ − log − þ log x:
m2W 2ðx − 1Þ ðx − 1Þ2 where ðgeL Þij ¼ ge;SM
L δij þ ðgeL ÞNP
ij and ðgR Þij ¼ gR
e e;SM
δij þ
ðD11Þ ðgR Þij with the SM tree-level couplings gL ¼ I 3eL −
e NP e;SM

Qe s2W and ge;SM

R ¼ −Qe s2W . According to Ref. [126] (see
Similar one-loop expressions for the low-energy coeffi- also Refs. [141–143]), the S3 contribution to the left-
cients can also be found in Refs. [138–141]. handed coupling reads as

 
Nc u;SM xt ðxt − 1 − log xt Þ xZ 2
ðgeL ÞNP
ij ¼ ðg u;SM
− g Þ þ Fðx Þ þ Oðx Z ðY 3 Þ3i ðY 3 Þ3j
Þ QL QL
ð4πÞ2 ðxt − 1Þ2
t
L R
12
   X 2
N c xZ 1 ge;SM
þ g u;SM
log x − iπ − þ L
ðY QL
3 Þwi ðY 3 Þwj
QL
3ð4πÞ2
Z
L
6 6 w¼1
   e;SM  X3
2N c xZ d;SM 1 gL
þ 2
g log x Z − iπ − þ ðȲ QL
3 Þwi ðȲ 3 Þwj ;
QL
ðE2Þ
3ð4πÞ L
6 6 w¼1

where xZ ¼ m2Z =m2S3 , xt ¼ m2t =m2S3 , gu;SM

L , gu;SM
R and gd;SM
L are the SM couplings for up-type and down-type quarks defined
analogous to those for charged leptons, and the function FðxÞ is defined as

ðx − 1Þð5x2 − 7x þ 8Þ − 2ðx3 þ 2Þ log x ðx − 1Þðx2 − 5x − 2Þ þ 6x log x

FðxÞ ¼ −gu;SM 4
− gu;SM
L
ðx − 1Þ R
ðx − 1Þ4
ðx − 1Þð−11x2 þ 7x − 2Þ þ 6x3 log x
þ ge;SM : ðE3Þ
L
3ðx − 1Þ4

Using the above effective coupling, the branching ratio for Z → μ∓ τ is given by

GF m3Z
BðZ → μ∓ τ Þ ¼ BðZ → μ− τþ Þ þ BðZ → μþ τ− Þ ¼ pffiffiffi ðjðgeL ÞNP 2 e NP 2
23 j þ jðgL Þ32 j Þ; ðE4Þ
3π 2ΓZ

where ΓZ is the total decay width of Z boson.

[1] J. C. Pati and A. Salam, Phys. Rev. Lett. 31, 661 (1973). [8] P. Langacker, in Prodceedings of the 1st International
[2] J. C. Pati and A. Salam, Phys. Rev. D 10, 275 (1974); 11, Symposium on Particles, Strings and Cosmology (World
703(E) (1975). Scientific, Singapore, 1990), pp. 237–269.
[3] H. Georgi and S. L. Glashow, Phys. Rev. Lett. 32, 438 [9] J. R. Ellis, S. Kelley, and D. V. Nanopoulos, Phys. Lett. B
(1974). 260, 131 (1991).
[4] H. Georgi, H. R. Quinn, and S. Weinberg, Phys. Rev. Lett. [10] U. Amaldi, W. de Boer, and H. Furstenau, Phys. Lett. B
33, 451 (1974). 260, 447 (1991).
[5] H. Georgi, AIP Conf. Proc. 23, 575 (1975). [11] P. Langacker and M.-x. Luo, Phys. Rev. D 44, 817 (1991).
[6] H. Fritzsch and P. Minkowski, Ann. Phys. (N.Y.) 93, 193 [12] C. Giunti, C. W. Kim, and U. W. Lee, Mod. Phys. Lett. A
(1975). 06, 1745 (1991).
[7] P. Langacker, Phys. Rep. 72, 185 (1981). [13] K. S. Babu and E. Ma, Phys. Lett. 144B, 381 (1984).

095012-25
GOTO, MISHIMA, and SHINDOU PHYS. REV. D 108, 095012 (2023)

[14] H. Murayama and T. Yanagida, Mod. Phys. Lett. A 07, 147 [45] L. Di Luzio, M. Kirk, A. Lenz, and T. Rauh, J. High
(1992). Energy Phys. 12 (2019) 009.
[15] A. Giveon, L. J. Hall, and U. Sarid, Phys. Lett. B 271, 138 [46] A. Angelescu, D. Bečirević, D. A. Faroughy, F. Jaffredo,
(1991). and O. Sumensari, Phys. Rev. D 104, 055017 (2021).
[16] I. Dorsner and P. Fileviez Perez, Nucl. Phys. B723, 53 [47] A. Crivellin, D. Müller, and L. Schnell, Phys. Rev. D 103,
(2005). 115023 (2021); 104, 055020(A) (2021).
[17] I. Dorsner, P. Fileviez Perez, and R. Gonzalez Felipe, Nucl. [48] N. Košnik and A. Smolkovič, Phys. Rev. D 104, 115004
Phys. B747, 312 (2006). (2021).
[18] I. Dorsner and P. Fileviez Perez, Phys. Lett. B 642, 248 [49] L.-L. Chau and W.-Y. Keung, Phys. Rev. Lett. 53, 1802
(2006). (1984).
[19] I. Dorsner, P. Fileviez Perez, and G. Rodrigo, Phys. Rev. D [50] P. A. Zyla et al. (Particle Data Group), Prog. Theor. Exp.
75, 125007 (2007). Phys. 2020, 083C01 (2020).
[20] B. Bajc and G. Senjanovic, J. High Energy Phys. 08 (2007) [51] J. Hisano, H. Murayama, and T. Yanagida, Nucl. Phys.
014. B402, 46 (1993).
[21] P. Fileviez Perez, Phys. Lett. B 654, 189 (2007). [52] H. Arason, D. J. Castano, B. Kesthelyi, S. Mikaelian, E. J.
[22] I. Dorsner and I. Mocioiu, Nucl. Phys. B796, 123 (2008). Piard, P. Ramond, and B. D. Wright, Phys. Rev. D 46, 3945
[23] P. Fileviez Perez, H. Iminniyaz, and G. Rodrigo, Phys. (1992).
Rev. D 78, 015013 (2008). [53] P. Nath and P. Fileviez Perez, Phys. Rep. 441, 191 (2007).
[24] I. Dorsner, S. Fajfer, J. F. Kamenik, and N. Kosnik, Phys. [54] A. Takenaka et al. (Super-Kamiokande Collaboration),
Rev. D 81, 055009 (2010). Phys. Rev. D 102, 112011 (2020).
[25] P. Fileviez Perez and C. Murgui, Phys. Rev. D 94, 075014 [55] K. G. Chetyrkin, J. H. Kuhn, and M. Steinhauser, Comput.
(2016). Phys. Commun. 133, 43 (2000).
[26] P. Cox, A. Kusenko, O. Sumensari, and T. T. Yanagida, [56] F. Herren and M. Steinhauser, Comput. Phys. Commun.
J. High Energy Phys. 03 (2017) 035. 224, 333 (2018).
[27] I. Doršner, S. Fajfer, and N. Košnik, Eur. Phys. J. C 77, 417 [57] A. Juste Rozas (on behalf of the ATLAS and CMS
(2017). Collaborations), Report No. ATL-PHYS-SLIDE-2023-
[28] D. Bečirević, I. Doršner, S. Fajfer, D. A. Faroughy, N. 034, 2023.
Košnik, and O. Sumensari, Phys. Rev. D 98, 055003 [58] J. de Blas, M. Chala, M. Perez-Victoria, and J. Santiago,
(2018). J. High Energy Phys. 04 (2015) 078.
[29] J. Schwichtenberg, Eur. Phys. J. C 79, 351 (2019). [59] J. de Blas, J. Criado, M. Perez-Victoria, and J. Santiago,
[30] N. Haba, Y. Mimura, and T. Yamada, Phys. Rev. D 99, J. High Energy Phys. 03 (2018) 109.
075018 (2019). [60] V. Gherardi, D. Marzocca, and E. Venturini, J. High
[31] H. Georgi and C. Jarlskog, Phys. Lett. 86B, 297 (1979). Energy Phys. 07 (2020) 225; 01 (2021) 006.
[32] M. H. Rahat, P. Ramond, and B. Xu, Phys. Rev. D 98, [61] E. E. Jenkins, A. V. Manohar, and M. Trott, J. High Energy
055030 (2018). Phys. 10 (2013) 087.
[33] M. J. Pérez, M. H. Rahat, P. Ramond, A. J. Stuart, and B. [62] E. E. Jenkins, A. V. Manohar, and M. Trott, J. High Energy
Xu, Phys. Rev. D 100, 075008 (2019). Phys. 01 (2014) 035.
[34] M. J. Pérez, M. H. Rahat, P. Ramond, A. J. Stuart, and B. [63] R. Alonso, E. E. Jenkins, A. V. Manohar, and M. Trott,
Xu, Phys. Rev. D 101, 075018 (2020). J. High Energy Phys. 04 (2014) 159.
[35] W. Buchmuller, R. Ruckl, and D. Wyler, Phys. Lett. B 191, [64] B. Grzadkowski, M. Iskrzynski, M. Misiak, and J. Rosiek,
442 (1987); 448, 320(E) (1999). J. High Energy Phys. 10 (2010) 085.
[36] I. Doršner, S. Fajfer, A. Greljo, J. F. Kamenik, and N. [65] E. E. Jenkins, A. V. Manohar, and P. Stoffer, J. High
Košnik, Phys. Rep. 641, 1 (2016). Energy Phys. 03 (2018) 016.
[37] Y. Sakaki, R. Watanabe, M. Tanaka, and A. Tayduganov, [66] J. Aebischer, A. Crivellin, M. Fael, and C. Greub, J. High
Phys. Rev. D 88, 094012 (2013). Energy Phys. 05 (2016) 037.
[38] G. Hiller and M. Schmaltz, Phys. Rev. D 90, 054014 [67] W. Dekens and P. Stoffer, J. High Energy Phys. 10 (2019)
(2014). 197.
[39] A. K. Alok, B. Bhattacharya, D. Kumar, J. Kumar, D. [68] J. Aebischer, M. Fael, C. Greub, and J. Virto, J. High
London, and S. U. Sankar, Phys. Rev. D 96, 015034 Energy Phys. 09 (2017) 158.
(2017). [69] E. E. Jenkins, A. V. Manohar, and P. Stoffer, J. High
[40] I. Doršner, S. Fajfer, D. A. Faroughy, and N. Košnik, Energy Phys. 01 (2018) 084.
J. High Energy Phys. 10 (2017) 188. [70] G. Buchalla, A. J. Buras, and M. E. Lautenbacher, Rev.
[41] L. Di Luzio, M. Kirk, and A. Lenz, Phys. Rev. D 97, Mod. Phys. 68, 1125 (1996).
095035 (2018). [71] N. Gubernari, M. Reboud, D. van Dyk, and J. Virto,
[42] S. Fajfer, N. Košnik, and L. Vale Silva, Eur. Phys. J. C 78, J. High Energy Phys. 09 (2022) 133.
275 (2018). [72] S. Jäger and J. Martin Camalich, J. High Energy Phys. 05
[43] J. Alda, J. Guasch, and S. Penaranda, Eur. Phys. J. C 79, (2013) 043.
588 (2019). [73] J. Lyon and R. Zwicky, arXiv:1406.0566.
[44] R. Mandal and A. Pich, J. High Energy Phys. 12 (2019) [74] S. Descotes-Genon, L. Hofer, J. Matias, and J. Virto,
089. J. High Energy Phys. 12 (2014) 125.

095012-26
FLAVOR PHYSICS IN SU(5) GUT WITH SCALAR FIELDS IN … PHYS. REV. D 108, 095012 (2023)

[75] S. Jäger and J. Martin Camalich, Phys. Rev. D 93, 014028 [106] M. Bordone, G. Isidori, and A. Pattori, Eur. Phys. J. C 76,
(2016). 440 (2016).
[76] M. Ciuchini, M. Fedele, E. Franco, S. Mishima, A. Paul, L. [107] G. Isidori, S. Nabeebaccus, and R. Zwicky, J. High Energy
Silvestrini, and M. Valli, J. High Energy Phys. 06 (2016) Phys. 12 (2020) 104.
116. [108] C. Bobeth, M. Gorbahn, T. Hermann, M. Misiak, E.
[77] M. Ciuchini, M. Fedele, E. Franco, A. Paul, L. Silvestrini, Stamou, and M. Steinhauser, Phys. Rev. Lett. 112,
and M. Valli, Phys. Rev. D 107, 055036 (2023). 101801 (2014).
[78] LHCb Collaboration, Phys. Rev. Lett. 131, 051803 (2023). [109] T. Blake, G. Lanfranchi, and D. M. Straub, Prog. Part.
[79] LHCb Collaboration, Phys. Rev. D 108, 032002 (2023). Nucl. Phys. 92, 50 (2017).
[80] R. Aaij et al. (LHCb Collaboration), arXiv:1808.08865. [110] K. De Bruyn, R. Fleischer, R. Knegjens, P. Koppenburg,
[81] R. L. Workman et al. (Particle Data Group), Prog. Theor. M. Merk, and N. Tuning, Phys. Rev. D 86, 014027 (2012).
Exp. Phys. 2022, 083C01 (2022). [111] K. De Bruyn, R. Fleischer, R. Knegjens, P. Koppenburg,
[82] J. P. Lees et al. (BABAR Collaboration), Phys. Rev. D 87, M. Merk, A. Pellegrino, and N. Tuning, Phys. Rev. Lett.
112005 (2013). 109, 041801 (2012).
[83] W. Altmannshofer et al. (Belle-II Collaboration), Prog. [112] A. Crivellin, D. Müller, and T. Ota, J. High Energy Phys.
Theor. Exp. Phys. 2019, 123C01 (2019); 2020, 029201(E) 09 (2017) 040.
(2020). [113] D. Buttazzo, A. Greljo, G. Isidori, and D. Marzocca,
[84] J. Grygier et al. (Belle Collaboration), Phys. Rev. D 96, J. High Energy Phys. 11 (2017) 044.
091101 (2017); 97, 099902(E) (2018). [114] C. Bobeth and U. Haisch, Acta Phys. Pol. B 44, 127
[85] O. Lutz et al. (Belle Collaboration), Phys. Rev. D 87, (2013).
111103 (2013). [115] B. Capdevila, A. Crivellin, S. Descotes-Genon, L. Hofer,
[86] Y. S. Amhis et al. (Heavy Flavor Averaging Group), Phys. and J. Matias, Phys. Rev. Lett. 120, 181802 (2018).
Rev. D 107, 052008 (2023). [116] A. Dedes, J. Rosiek, and P. Tanedo, Phys. Rev. D 79,
[87] R. Aaij et al. (LHCb Collaboration), Phys. Rev. Lett. 118, 055006 (2009).
251802 (2017). [117] D. Bečirević, O. Sumensari, and R. Zukanovich Funchal,
[88] J. Lees et al. (BABAR Collaboration), Phys. Rev. Lett. 118, Eur. Phys. J. C 76, 134 (2016).
031802 (2017). [118] A. Abada, D. Bečirević, M. Lucente, and O. Sumensari,
[89] T. V. Dong et al. (Belle Collaboration), Phys. Rev. D 108, Phys. Rev. D 91, 113013 (2015).
L011102 (2023). [119] D. E. Hazard and A. A. Petrov, Phys. Rev. D 94, 074023
[90] L. Aggarwal et al. (Belle-II Collaboration), arXiv:2207 (2016).
.06307. [120] L. Calibbi, T. Li, X. Marcano, and M. A. Schmidt, Phys.
[91] R. Aaij et al. (LHCb Collaboration), Phys. Rev. Lett. 123, Rev. D 106, 115039 (2022).
211801 (2019). [121] T. Goto, Y. Okada, and Y. Yamamoto, Phys. Rev. D 83,
[92] S. Watanuki et al. (Belle Collaboration), Phys. Rev. Lett. 053011 (2011).
130, 261802 (2023). [122] Y. Okada, K.-i. Okumura, and Y. Shimizu, Phys. Rev. D
[93] R. Aaij et al. (LHCb Collaboration), J. High Energy Phys. 61, 094001 (2000).
06 (2023) 143. [123] Y. Kuno and Y. Okada, Rev. Mod. Phys. 73, 151 (2001).
[94] N. Tsuzuki et al. (Belle Collaboration), J. High Energy [124] T. Aoyama et al., Phys. Rept. 887, 1 (2020).
Phys. 06 (2023) 118. [125] D. P. Aguillard et al. (Muon g-2 Collaboration), arXiv:
[95] S. Banerjee et al. arXiv:2203.14919. 2308.06230.
[96] S. Patra et al. (Belle Collaboration), J. High Energy Phys. [126] P. Arnan, D. Becirevic, F. Mescia, and O. Sumensari,
05 (2022) 095. J. High Energy Phys. 02 (2019) 109.
[97] J. P. Lees et al. (BABAR Collaboration), Phys. Rev. Lett. [127] P. Abreu et al. (DELPHI Collaboration), Z. Phys. C 73,
104, 151802 (2010). 243 (1997).
[98] A. Abdesselam et al. (Belle Collaboration), J. High Energy [128] A. J. Buras, J. R. Ellis, M. K. Gaillard, and D. V.
Phys. 10 (2021) 19. Nanopoulos, Nucl. Phys. B135, 66 (1978).
[99] K. Hayasaka et al., Phys. Lett. B 687, 139 (2010). [129] P. H. Frampton, S. Nandi, and J. J. G. Scanio, Phys. Lett.
[100] G. Aad et al. (ATLAS Collaboration), Phys. Rev. Lett. 127, 85B, 225 (1979).
271801 (2022). [130] P. Kalyniak and J. N. Ng, Phys. Rev. D 26, 890 (1982).
[101] M. Dam, SciPost Phys. Proc. 1, 041 (2019). [131] P. Eckert, J. M. Gerard, H. Ruegg, and T. Schucker, Phys.
[102] A. J. Buras, J. Girrbach-Noe, C. Niehoff, and D. M. Straub, Lett. 125B, 385 (1983).
J. High Energy Phys. 02 (2015) 184. [132] M. E. Machacek and M. T. Vaughn, Nucl. Phys. B222, 83
[103] E. Ganiev (Belle-II Collaboration), Proceedings of the (1983).
European Physical Society Conference on High Energy [133] M. E. Machacek and M. T. Vaughn, Nucl. Phys. B236, 221
Physics (EPS-HEP), Hamburg, 2023. (1984).
[104] A. Celis, J. Fuentes-Martin, A. Vicente, and J. Virto, Phys. [134] M. E. Machacek and M. T. Vaughn, Nucl. Phys. B249, 70
Rev. D 96, 035026 (2017). (1985).
[105] B. Capdevila, S. Descotes-Genon, L. Hofer, and J. Matias, [135] M.-x. Luo, H.-w. Wang, and Y. Xiao, Phys. Rev. D 67,
J. High Energy Phys. 04 (2017) 016. 065019 (2003).

095012-27
GOTO, MISHIMA, and SHINDOU PHYS. REV. D 108, 095012 (2023)

[136] K. Kowalska, E. M. Sessolo, and Y. Yamamoto, Eur. Phys. [140] M. Bordone, O. Catà, T. Feldmann, and R. Mandal, J. High
J. C 81, 272 (2021). Energy Phys. 03 (2021) 122.
[137] M. Fedele, F. Wuest, and U. Nierste, arXiv:2307 [141] A. Crivellin, C. Greub, D. Müller, and F. Saturnino, J. High
.15117. Energy Phys. 02 (2021) 182.
[138] A. Crivellin, D. Müller, and F. Saturnino, J. High Energy [142] F. Feruglio, P. Paradisi, and A. Pattori, Phys. Rev. Lett.
Phys. 06 (2020) 020. 118, 011801 (2017).
[139] V. Gherardi, D. Marzocca, and E. Venturini, J. High [143] F. Feruglio, P. Paradisi, and A. Pattori, J. High Energy
Energy Phys. 01 (2021) 138. Phys. 09 (2017) 061.

095012-28