Constraining dark matter from strong phase

transitions in a U (1)Lµ−Lτ model: Implications for

neutrino masses and muon g − 2

Sandhya Choubey,a,b Sarif Khan,c Marco Merchanda,b and Sampsa Vihonena,b

a
KTH Royal Institute of Technology, Department of Physics, SE-10691 Stockholm, Sweden
b
The Oskar Klein Centre for Cosmoparticle Physics, AlbaNova University Centre, SE-10691 Stock-
holm, Sweden
arXiv:2406.16460v1 [hep-ph] 24 Jun 2024

c
Institut für Theoretische Physik, Georg-August-Universität Göttingen, Friedrich-Hund-Platz 1,
37077 Göttingen, Germany
E-mail: [email protected], [email protected],
[email protected], [email protected]

Abstract: In this paper, we study a non-minimal gauged U (1)Lµ −Lτ model, where we
add two complex singlet scalars, three right-handed Majorana neutrinos (RHN), and a
vector-like dark fermion to the Standard Model (SM), all non-trivially charged under the
extra gauge symmetry. The model offers an easy resolution to the muon (g − 2) anomaly,
which fixes the scale of spontaneous symmetry breaking. Furthermore, the two-zero minor
structure in the RHN mass matrix provides successful predictions for neutrino oscillation
parameters, including the Dirac phase. The extended scalar sector can easily induce first-
order phase transitions. We identify all possible phase transition patterns in the three-
dimensional field space. We quantify the associated gravitational waves from the sound
wave source and demonstrate that the signatures can be observed in future space-based
experiments. We find that strong first-order phase transitions require large values of scalar
quartic couplings which constrain the scalar dark matter (DM) relic density to a maximum
of 10−2 and 10−5 when we consider the DM direct detection bound. Nonetheless, the
model successfully explains the DM relic density via contribution from the vector-like dark
fermion. We show the allowed range of the model parameters that can address all the
beyond SM issues targeted in this study.
Contents

1 Introduction 1

2 The model 3
2.1 Tree level scalar potential 5

3 The muon (g − 2) anomaly 6

4 Neutrino mass 7

5 Finite temperature 10
5.1 Effective potential 11
5.2 Phase transitions 11

6 Gravitational waves 13

7 Dark matter 15
7.1 Dark matter bounds 17
7.2 Allowed parameter regions from dark matter 19

8 Gravitational waves and dark matter 24

9 Conclusions 27

A Neutrino parameters 29

B Field dependent masses 31

B.1 Daisy resummation 32

1 Introduction

The standard model of particle physics (SM) is arguably the most successful theory to
describe the physics of elementary particles. Despite its agreement with experimental ob-
servations [1], the SM alone is unable to account for several well-established properties of
the Universe. The deficiencies of the SM include (i) the observed neutrino masses and
their oscillation parameters, (ii) the particle identity of the dark matter (DM) and (iii) a
consistent explanation of the present day baryon asymmetry. In order to accommodate the
aforementioned issues simultaneously, one must consider beyond the SM (BSM) theories
that may include new fundamental interactions, particles and symmetries.
One of the simplest extensions to the SM can be obtained by introducing an Abelian
gauge symmetry, under which the second and third lepton generations are non-trivially

–1–
charged. This extension is commonly referred to as the U (1)Lµ −Lτ extension [2–4]. Under
the new gauge symmetry, leptonic particles of flavor µ have charge +1 while those of flavor
τ have charge −1, hence the name Lµ − Lτ . For each lepton generation, we include a
corresponding Majorana right-handed neutrino (RHN), which is a singlet under the SM. In
this way the smallness of the neutrino masses is explained via the type-I seesaw mechanism.
The U (1)Lµ −Lτ gauge symmetry is free from anomalies and is assumed to be spontaneously
broken by the vacuum expectation value (vev) of a complex singlet scalar, providing a mass
to the associated Z ′ gauge boson. These assumptions give rise to a very particular RHN
neutrino mass structure, which is commonly known as the two-zero minor structure [5–9]
since the diagonal 22 and 33 entries vanish due to the extra symmetry. The model with
this particle content has been dubbed the minimal gauged U (1)Lµ −Lτ in Refs. [7–9] and
has been studied in the context of the baryon asymmetry via leptogenesis [10, 11], particle
DM phenomenology [12–14] and (g − 2)µ anomaly [15–18], to name the most common.
Particularly interesting is the case in which more than one deficiency of the SM can be
simultaneously accounted for in the same scenario. This is the case for the model introduced
in Refs. [13, 14], where the minimal gauged U (1)Lµ −Lτ was responsible for providing a DM
particle candidate consistent with experimental constraints, neutrino masses and neutrino
oscillation parameters. The model was also able to solve the (g − 2)µ anomaly [19, 20]. On
the other hand, incorporating viable leptogenesis while maintaining the correct (g−2)µ value
is unfeasible, as the latter requires spontaneous symmetry breaking at the scale O(1 − 100)
GeV while the former needs very large RHN masses1 , i.e., Mi ≳ O(109 ) GeV [23].
The discovery of gravitational waves (GWs) by the LIGO/Virgo/KAGRA Collabora-
tion [24–30] from inspiral binary systems and the potential detection by other experiments
such as the Pulsar Timing Array (PTA) collaborations [31, 32] and the future Laser Inter-
ferometer Space Antenna (LISA) [33] have sparked a new gold rush in the characterization
of all possible signals. Of notable relevance for particle physics are those coming from early
universe first-order phase transitions (PTs) which are tightly connected to the structure
of the scalar potential and highly influenced by the size of any Higgs-portal interactions.
Furthermore, first-order PTs can provide the required out-of-equilibrium condition neces-
sary for electroweak (EW) baryogenesis [34]. It is well known that strong PTs prefer rather
large Higgs-portal interactions as they strengthen the potential barrier between two poten-
tial minima [35–38]. In contrast, a successful explanation of the DM in the minimal gauged
U (1)Lµ −Lτ model [13, 14] requires a sufficiently small Higgs-portal interaction as otherwise
direct detection experiments [39] rule out the model. The incompatibility between particle
DM and the appearance of strong PTs in the minimal scenario has prompted us to consider
a slight modification of the original theory.
In the present paper we study a non-minimal gauged U (1)Lµ −Lτ by adding to the
minimal scenario an extra complex singlet scalar with vanishing vev and a vector-like
fermion. These extra fields are singlets under the SM gauge group and their non-zero
charges under U (1)Lµ −Lτ are parameterized by arbitrary constants qx and qψ , respectively.

1
The compatibility of leptogenesis with the (g − 2)µ anomaly can be attained by invoking a non-minimal
version of the U (1)Lµ −Lτ model with two complex scalars with non-zero vevs, see Refs. [21, 22].

–2–
This choice is not only generic but it also makes it possible to regulate the direct couplings
between the extra fields and the Z ′ gauge boson, which control the production of DM via
the freeze-out mechanism.
As one would expect, the model achieves the solution to the (g − 2)µ anomaly and
successfully predicts the neutrino oscillation parameters. In this context, we undertake
a major revision of the neutrino sector by using the most recent NuFit data [40] and
provide the scale of the fundamental parameters for which the predicted neutrino oscillation
parameters fall within their experimental limits.
The extended scalar sector of our model warrants a careful investigation of its finite
temperature potential. We undertake such endeavour keeping track of all the dynamical
fields which renders the effective potential a 3-dimensional function in field-space. We then
perform a systematic scan of the free parameters and find the regions for which strong
PTs can be produced. Along the way we examine the multiple transition patterns that
can arise and investigate useful correlations between the fundamental scalar couplings and
the strength of the PT. After having uncovered the relevant parameter space of PTs we
assess the predicted GW signals coming from the sound wave source. This allows to further
restrict the parameter space by considering only benchmark points which predict detectable
signals at LISA.
As the last step of our work, we study the two-component DM phenomenology in which
the dark fermion mass and U (1)Lµ −Lτ charge are the only remaining free parameters. This
allows us to easily find solutions which abide by DM relic abundance and direct detection
bounds.
In summary, our extended gauged U (1)Lµ −Lτ scenario can account for: (i) the (g − 2)µ
anomaly, (ii) neutrino oscillation observables, (iii) strong PTs and associated GWs and
provides (iv) a two-component particle DM consistent with all current limits.
The paper is structured as follows. In section 2, we introduce the model and set the
notation writing down the general Lagrangian and the free parameters in the model. In
section 3, we review the solution to the (g − 2)µ anomaly. We then proceed to review
the status of neutrino oscillation parameters in section 4, discussing the numerical fitting
procedure in detail. In section 5, we present the one-loop effective potential along with
the finite temperature contributions, showing the parameter space which is consistent with
first-order PTs and classify all the possible patterns of field-space dynamics. Section 6
is dedicated to examine the associated GW signals. It is shown that that the predicted
GW signals can be probed in the future GW experiments. In section 7 we investigate
DM and review the most stringent constraints and provide the viable parameter space.
Following that, in section 8, we provide the viable parameter space that is consistent with
all observations and display the most interesting correlations in Lagrangian parameters.
We finally provide our conclusions and outlook in section 9.

2 The model

In this section we present the details of our non-minimal gauged U (1)Lµ −Lτ model. For
completeness, the particle content and charges are presented in Table (1). Notice that the

–3–
 Gauge Baryonic Fields Vector-like Lepton Fields Scalar Fields
Group (QiL , uiR , diR ) ψL , ψR (LeL , eR , NRe ) (LµL , µR , NRµ ) (LτL , τR , NRτ ) Φ Φ′ ΦDM
U (1)Lµ −Lτ 0 qψ 0 1 −1 0 1 qDM

Table 1. Particle content and their corresponding charges under the gauge group U (1)Lµ −Lτ .

only SM particles with non-vanishing charges are the µ and τ charged leptons. The model
includes three RHN Majorana neutrinos NRi (i = e, µ, τ ) and a vector-like fermion ψ which
is one of the DM components. The scalar sector contains the SM Higgs doublet Φ and
two extra complex scalars which we denote by Φ′ and ΦDM . The former is responsible for
spontaneous symmetry breaking of the gauged U (1)Lµ −Lτ and provides the mass to the
associated Z ′ boson, while the latter does not acquire a vev and comprises the second DM
component. It is worthwhile to stress that all new degrees of freedom are neutral under the
SM gauge group.
The complete Lagrangian for the model can be expressed as

L = LSM + Lscalar + Lψ + LZ ′ + LN , (2.1)

where LSM denotes the full SM Lagrangian but does not include the scalar sector. The
scalar sector is given by the second term, Lscalar , and will be introduced in the next subsec-
tion. The third term contains the vector-like dark fermion, namely Lψ = ψ̄ (iγ µ Dµ − mψ ) ψ,
where the covariant derivative contains the non-trivial charges:

Dµ X ≡ ∂µ − igZ ′ Q(X)Zµ′ X, (2.2)

where X = ψ, NRe , NRµ , NRτ , Φ′ , ΦDM and Q(X) = (qψ , 0, 1, −1, 1, qDM ) are the respective
charges. Here gZ′ is the new gauge coupling constant. We keep the charge Q(X) as a
free parameter for the fermion and scalar DM defined as qψ and qDM , respectively. The
physical mass of the U (1)Lµ −Lτ gauge boson after the spontaneous symmetry breaking
takes the form,
mZ ′ = gZ ′ v ′ , (2.3)
where v ′ is the vev of Φ′ . The dynamics of the extra gauge boson is encoded in its kinetic
αβ
term via LZ ′ = 14 Fµτ Fµτ αβ . In this work we ignore any possible kinetic mixing with the
photon. The effects of the kinetic mixing have been studied elsewhere [41–43].
Following the particle charge assignments from table (1), the neutrino sector can be
written as
X i 1 1
LN = N̄i γ µ Dµ Ni − Mee N¯ec Ne − Mµτ (N¯µc Nτ + N¯τc Nµ )
2 2 2
i=e, µ, τ
1 1
− heµ (N¯ec Nµ + N¯µc Ne )Φ′† − heτ (N¯ec Nτ + N¯τc Ne )Φ′
2X 2
− yi L̄i Φ̃Ni + h.c. (2.4)
i=e, µ, τ

This piece of the Lagrangian gives rise to the celebrated two-zero minor structure [5–9].
We will review the neutrino sector, in full detail, in section 4.

–4–
2.1 Tree level scalar potential
The scalar sector of our model is characterized by its tree-level potential, which we write
in the most generic form:

V0 (Φ, Φ′ , ΦDM ) =VSM (Φ) + µ2DM Φ†DM ΦDM + λDM (Φ†DM ΦDM )2
+ µ2ϕ′ Φ′† Φ′ + λϕ′ (Φ′† Φ′ )2
+ λDϕ (Φ†DM ΦDM )(Φ† Φ) + λDϕ′ (Φ†DM ΦDM )(Φ′† Φ′ )
+ λϕϕ′ (Φ′† Φ′ )(Φ† Φ), (2.5)

where VSM (Φ) = −µ2ϕ Φ† Φ + λϕ (Φ† Φ)2 is the SM Higgs potential.

Since the Higgs doublet Φ and the U (1)Lµ −Lτ singlet Φ′ get vev at zero temperature,
we write them as
 
G+  ′
v + ϕ ′+i G ′

′ ϕ
Φ =  v + ϕ + i Gϕ  , Φ = √ , (2.6)
√ 2
2
where G± , Gϕ and Gϕ′ are the Goldstone bosons which are absorbed into the longitudi-
nal components of the W ± , Z and Z ′ gauge bosons, respectively. The DM candidate is
contained in the extra U (1)Lµ −Lτ complex singlet which we parameterize as

χeiα
ΦDM = √ , (2.7)
2
where α is an arbitrary phase. This choice is to emphasize that after spontaneous symmetry
breaking there are two scalar degrees of freedom with degenerate mass, mDM .
The initial total of 9 free parameters in (2.5) are reduced to only 7 by means of the phe-
nomenological constraints v = 246 GeV and mϕ = 125 GeV. We choose as free parameters
the set
Scalar sector parameters = {θ, v ′ , mϕ′ , mDM , λDϕ , λDϕ′ , λDM }. (2.8)
Making use of the minimization conditions and mass matrix diagonalization, the dependent
parameters are the mass squared parameters,
1
µ2ϕ = (2v 2 λϕ + v ′2 λϕϕ′ ),
2
1
µϕ′ = (−2v ′2 λϕ′ − v 2 λϕϕ′ ),
2 (2.9)
2
1
µ2DM = m2DM − (v 2 λDϕ + v ′2 λDϕ′ ),
2
and also the quartic couplings
1
λϕ = (m2 cos2 θ + m2ϕ′ sin2 θ) ,
2v 2 ϕ
1
λ ϕ′ = ′2 (m2ϕ sin2 θ + m2ϕ′ cos2 θ) , (2.10)
2v
1
λϕϕ′ = ′ ((m2ϕ′ − m2ϕ ) sin θ cos θ) .
vv

–5–
 Before continuing to the next section, we provide a full counting of the free model
parameters. As we have shown, the scalar sector provides us with 7 parameters, which are
listed in Eqn. (2.8). The neutrino sector is described by the Yukawa couplings in Eqn. (2.4),
which gives us

neutrino sector parameters = {Mee , Mµτ , heµ , heτ , ye , yµ , yτ }. (2.11)

Additionally, one needs to set the vector-like fermion charge qψ and mass mψ , the new
gauge coupling constant gZ ′ and the DM scalar charge qx . Therefore, the model is fully
specified with a total of 18 free parameters. As we proceed to the following sections, each
phenomenological observable will fix the range of a subset of the free model parameters. We
will demonstrate that the full parameter space can by narrowed down by the aforementioned
considerations.

3 The muon (g − 2) anomaly

The (g − 2)µ value can be calculated very accurately within the SM. The SM prediction
for (g − 2)µ is aµ (SM) = 116591810(43) × 10−11 (0.37 ppm) [20] and it includes corrections
from electrodynamics, hadronic vacuum polarization, hadronic light by light as well as EW
processes. At the same time, advancements in the experimental methodologies have led to
a very precise measurement of (g − 2)µ . The most notable contributions have been carried
out at Brookhaven (BNL E821) [44], CERN [45–48] and Fermilab [19]. The measured value
differs from the SM prediction by [19, 20]

∆aµ = aµ (exp) − aµ (SM) = (251 ± 59) × 10−11 , (3.1)

which corresponds to 4.2 σ confidence level (CL) discrepancy.

It is noteworthy to highlight that recent progress in the lattice computation deviates
from the experimental measurements of the hadronic vacuum polarization by 2.1σ, which
diminishes the significance of the ∆aµ discrepancy to 1.5σ [49]. Additionally, the new
measurements from the CMD-3 experiments suggest a less-significant disagreement between
the experimental observations and the theoretical predictions, albeit with a remaining 2.4σ
discrepancy [50]. The new measurements from CMD-3 introduce a conflicting perspective
with prior measurements of e+ e− → π + π − conducted by both the same and different
experiments [51–54].
In the present work, we calculate the contribution to (g − 2)µ in the model. The
contribution arises from the one-loop diagram mediated by the extra gauge boson, giving

gZ2 ′ 1 2x(1 − x)2

Z
∆aµ = 2 dx , (3.2)
8π 0 (1 − x)2 + rx
m2µ
where r = m2Z ′
and mµ is the muon mass. Taking mZ ′ = 100 MeV and gZ ′ = 9 × 10−4 ,
for example, leads to the deviation ∆aµ = 226 × 10−11 , which is within the experimental
range specified in Eqn. (3.2). The allowed region in the gZ ′ − mZ ′ is discussed in detail, in
the context of DM, in section 7. It is shown that the model can explain the muon (g − 2)µ

–6–
anomaly in the case the anomaly survives the future measurements on (g − 2)µ . In contrast,
the exclusion of (g − 2)µ can also be accommodated in the present work as a large region
of the parameter space predicts a negligible contribution to (g − 2)µ . We further note that
the resolution to the (g − 2)µ can be potentially ruled out by a high precision observation
of the luminosity from white dwarfs cooling proceeding from Z ′ decays into neutrinos [41].
However, this bound can be evaded by introducing tree-level kinetic mixing which cancels
the kinetic mixing generated by the loop diagrams.

4 Neutrino mass

The neutrino masses in this model arise from the Lagrangian presented in Eqn. (2.4). The
model has two kinds of mass matrices, namely the Majorana mass matrix consisting of
right-handed neutrinos and the Dirac mass matrix consisting of both left- and right-handed
neutrinos. The Majorana mass matrix is given by
v′ v′
 
M ee √ heµ √ heτ

 2 2 

 
 ′ 
 v
(4.1)
iξ

MR =   √2 heµ 0 Mµτ e  ,
 
 
 
 v′ 
√ heτ Mµτ e iξ 0
2
whereas we write the Dirac mass matrix, in diagonal form without loss of generality
 
ye 0 0
 
 
MD =  0 yµ 0 . (4.2)
 
 
 
0 0 yτ

On the other hand, the full neutrino mass matrix reads

!
0 MD
M= T M . (4.3)
MD R

The neutrino mass matrix in Eqn. (4.3) can be diagonalised in different limits of MD
and MR , for example, the pseudo-Dirac limit MD ≫ MR and the seesaw limit MD ≪ MR .
In the present work, we focus on the seesaw limit MD ≪ MR , in which case the eigenvalues
take the form

mν ≃ −MD MR−1 MD
T
, (4.4)
mN ≃ MR , (4.5)

where mν represents the light neutrino mass matrix and mN represents the heavy neutrino
mass matrix, respectively.

–7–
 The light neutrino mass matrix mν in Eqn. (4.4) can be diagonalised to obtain the
three neutrino mass eigenstates m1 , m2 and m3 ,

mνd = diag(m1 , m2 , m3 ) = UP M N S mν UPT M N S , (4.6)

where UP M N S is the well-known Pontecorvo-Maki-Nakagawa-Sakata matrix,

 
c12 c13 s12 c13 s13 e−iδCP
UP M N S = −s12 c23 − c12 s23 s13 eiδCP c12 c23 − s12 s23 s13 eiδCP s23 c13  , (4.7)
 
s12 s23 − c12 c23 s13 eiδCP −c12 s23 − s12 c23 s13 eiδCP c23 c13

and cij = cos θij , sij = sin θij for i, j = 1, 2, 3. The neutrino mixing angles θ12 , θ13 and
θ23 can be derived directly from UP M N S , as shown in appendix A, whereas the CP phase
δCP is obtained from the Jarlskog invariant. The remaining neutrino mixing parameters
∆m221 and ∆m231 are determined from m1 , m2 and m3 as ∆m2ij ≡ m2i − m2j . Without
loss of generality, we limit our study to the normal ordering of the neutrino masses, where
m1 < m2 < m3 .
We adopt the following numerical approach for the neutrino sector in this model. We
generate random values for the model parameters Mee , Mµτ , heµ , heτ , ξ, fe , fµ and fτ and
use χ2 minimization to locate model parameter values that satisfy the 3σ CL constraints
for the neutrino oscillation parameters θ12 , θ13 , ∆m221 and ∆m231 . In order to show the
full predictivity of the model, θ23 and δCP are allowed to have values beyond their 3σ CL
ranges. The 3σ CL ranges are adopted from NuFit 5.2 [40], focusing on the values that do
not account for the atmospheric neutrino data from SuperKamiokande. The χ2 function
that is to be minimized is given as
test − θ true )2
(θ12 (θtest − θtrue )2 (θ23
test − θ true )2 (δ test − δ true )2
χ2 = 2
12
+ 13 2 13 + 2
23
+ CP 2 CP
σθ12 σθ13 σθ23 σδCP
(4.8)
((∆m221 )test − (∆m221 )true )2 ((∆m231 )test − (∆m231 )true )2
+ 2 + 2 ,
σ∆m 2 σ∆m 2
21 31

where θ12
true , θ true , θ true , δ true , (∆m2 )true and (∆m2 )true correspond to the central values
13 23 CP 21 31
of the neutrino oscillation parameters and σθ12 , σθ13 , σθ23 , σδCP , σ∆m221 and σ∆m231 are their
associated standard deviations. To ensure that the available parameter space for θ23 and
δCP is covered comprehensively, we accept χ2 values as large as 500. At the same time, we
ensure that the experimental constraints are satisfied for all neutrino oscillation parameters
except θ23 and δCP . With this approach, we scan model parameters for v ′ = 111 GeV
to obtain 10,000 points that satisfy the neutrino constraints by 3σ CL. We then exclude
the data points that are not in agreement with radiative muon decay constraints from
MEG [55, 56], leaving 9551 valid data points.
Fig. 1 displays the scatter plot for the neutrino oscillation parameters θ23 and δCP
with five different cuts for the sum of light neutrino masses i mi . The experimental
P

constraints are shown for θ23 and δCP at 3σ CL with the grey rectangles, which include
both the constraints that take into account the SuperKamiokande (SK) atmospheric data
and those that do not. The data points are marked by the different colors to show whether

–8–
Figure 1. Predicted values for neutrino mixing parameters θ23 and δCP . Each point satisfies the
NuFit 5.2 constraints at 3σ CL for the neutrino mixing parameters that are not shown in the figure.

they satisfy the cuts i mi < 0.12 eV (brown markers), 0.23 eV (blue markers), 0.51 eV
P

(green markers), 0.83 eV (magenta markers) and 1.00 eV (red markers), respectively. The
displayed i mi cuts are chosen according to the different cosmological data sets2 discussed
P

in Ref. [58]. Fig. 1 indicates that a small number of points satisfy both the most stringent
bound on neutrino masses, i mi < 0.12 eV, and the constraints on neutrino oscillation
P

parameters when the SK data is not taken into account. When we consider a higher limit on
the sum, i mi < 0.23 eV, a greater number of points that satisfy the constraints becomes
P

available. We see that in this case the model predicts θ23 and δCP values that satisfy the
constraints without SK data as well.
An interesting result in the predictions for θ23 and δCP is that the atmospheric mixing
angle θ23 can attain values near the maximal mixing value 45◦ when i mi is sufficiently
P

large. This behaviour is further illustrated in Fig. 2, where the lightest neutrino mass m1
is presented as function θ23 (left panel) and δCP (right panel). In both panels, we have
excluded the data points that are not consistent with the neutrino oscillation parameter
constraints at 3σ CL. In this regard, the cuts are performed for the five neutrino oscillation
parameters that are not displayed on the horizontal axis. We have again used the experi-
mental constraints that do not account the atmospheric data from SK. The corresponding
3σ CL ranges are also shown for θ23 and δCP parameters. The results show that θ23 = 45◦
is achievable in this model when m1 is very large. In a similar manner, maximally violating
CP phases δCP = 90◦ and 270◦ are reachable when m1 ∼ 1 eV or above. However, both
2 P
The cosmological bound on the sum of neutrino masses can go as low as i mi < 0.072 eV, as was
found in the recent analysis of CMB and DESI [57]. The most stringent constraints apply to the ΛCDM
scenario where no new physics beyond SM is included. With the introduction of new physics, the upper
P
bound on i mi can be relaxed.

–9–
Figure 2. Correlation between the lightest neutrino mass m1 and neutrino mixing parameters θ23
and δCP . Each data point satisfies that NuFit 5.2 constraints for the remaining neutrino mixing
parameters at 3σ CL while assuming normal hierarchy.

cases are excluded by the cosmological bounds on i mi .

P

We finally checked that our results are in agreement with the sum rules that were
discussed in Ref. [7]. All of the data points shown in Figures 1 and 2 are found to be
consistent with the sum rules by 3σ CL.

5 Finite temperature

In this section we introduce the 1-loop effective potential at zero and finite temperature.
We then proceed to discuss the appearance of first order phase transitions and the multiple
vacuum tunneling patterns in field space.
In the preceding sections we fixed the gauge sector and neutrino parameters with the
g − 2 and neutrino oscillation data. The remaining free parameters are the dark vector-like
fermion mass and charge mψ , qψ , respectively, and also the parameters in the tree-level
scalar potential, Eqn. (2.5). For the tree-level scalar potential, we perform a scan with the
following uniform ranges:

mϕ /2 < mϕ′ , mDM < 500 GeV,

−3 < λDϕ , λDϕ′ < 3,

−0.1 < θ < 0.1.

−103 ≤ qDM ≤ 103 . (5.1)

In our numerical scans, we verify that the quartic couplings respect the positivity and
perturbativity bounds. For the positivity of the potential, we employ the formulae from

– 10 –
 √
the appendix of Ref. [13]. We later apply the perturbativity bound |qDM gZ ′ | ≤ 4π when
we study DM in section 7.
Choosing the extra scalar masses above the Higgs threshold allows us to avoid the
stringent constraints coming from Higgs invisible constraints. The portal couplings are
chosen to be in the natural range, while the mixing angle can easily satisfy the constraints
from the Higgs signal strengths measurements. For simplicity, we fix λDM = 1 in our scans.
We have verified that our conclusions remain qualitatively the same for different values of
λDM . The vector-like fermion properties are found by requiring the correct DM abundance
and detection constraints. The corresponding results are also presented in section 7.

5.1 Effective potential

The zero temperature 1-loop correction is given by the Coleman-Weinberg potential. In the
on-shell scheme, the potential can be written as

m2i (ϕ, ϕ′ , χ) 3
   
X di
VCW (ϕ, ϕ′ , χ) = (−1)Fi m 4
i (ϕ, ϕ′
, χ) log 2 − + 2m 2
i (ϕ, ϕ′
, χ)m2
0i ,
64π 2 m 0i 2
i
(5.2)
where the index i runs over all particles contributing to the potential with Fi = 0 (1) for
bosons (fermions), di is the number of degrees of freedom of the particle species, mi (h, H, χ)
is the field-dependent mass of particle i and m0i is the mass at zero temperature, respec-
tively.
At finite temperature one must add the contribution [59]

T4 X mi (ϕ, ϕ′ , χ)
 
′
VT (ϕ, ϕ , χ, T ) = 2 di J∓ , (5.3)
2π T
i

where di are the number of degrees of freedom and the J∓ functions are defined as
Z ∞  √ 
2 2
J∓ (x) = ± dyy 2 log 1 ∓ e− y +x , (5.4)
0

and the upper (lower) sign is for bosons (fermions).

The finite temperature effective potential is given by the superposition of Eqns. (2.5),
(5.2) and (5.3). It can be written as

Vef f (ϕ, ϕ′ , χ, T ) = V0 (ϕ, ϕ′ , χ) + VCW (ϕ, ϕ′ , χ) + VT (ϕ, ϕ′ , χ, T ). (5.5)

5.2 Phase transitions

The non-trivial multi-field potential can easily feature potential barriers that could give
rise to temperature fluctuations. These fluctuations can then transition the system from
unstable local minima towards the absolute minimum of the free energy. In the theory of
vacuum tunneling at finite temperature [60–64], the transition probability between such
states is given by
S3 (T ) 3/2 4 −S3 (T )/T
 
Γ3 (T ) = T e , (5.6)
2πT

– 11 –
where S3 denotes the O(3) symmetric Euclidean action of the theory, which are obtained
by integrating the equation of motion for the scalars. The phase transition proceeds via
bubble nucleation which the associated nucleation temperature Tn reached when the decay
rate in Eqn. (5.6) becomes comparable to the Hubble volume. This happens when Γ3 ≈ H 4 ,
where the Hubble scale is written as
ρR ∆Vef f (ϕ, ϕ′ , χ, T )
H2 = 2 + 2 , (5.7)
3MPl 3MPl

with ρR = π 2 g∗ T 4 /30, the energy density in radiation with g∗ number of degrees of freedom
[65]. Furthermore, we take into account the energy density contribution from the scalar
potential, which can be especially significant for strong phase transitions where a large
amount of latent heat is released. The latent heat released normalized to the radiation
energy density is encoded by

T ∂Vef f (ϕ, ϕ′ , χ, T )
 
1 ′
α(T ) ≡ ∆Vef f (ϕ, ϕ , χ, T ) − ∆ . (5.8)
ρR 4 ∂T

In the above formulae, the potential energy difference is between the local and global
minima, i.e., ∆X ≡ Xlocal − Xglobal . The process of bubble nucleation is also characterized
by an inverse time scale or inverse time duration β. This can be written as
 
β d S3 (T )
≡T . (5.9)
H dT T T →Tn

Another crucial property of the nucleation process is the velocity at which the bubble wall
interface expands and whether the bubble wall runs away at the speed of light or reaches a
steady state velocity. Since we consider a polynomial potential with tree-level mass terms
in the present work, the amount of supercooling is very limited and the walls are expected
to reach a steady state. Furthermore, since the only particles beyond the SM that can gain
or lose mass during the transition are the U (1)Lµ −Lτ gauge boson and the RHN Majorana
fermions, the source of friction coming from their distribution functions is also small. We
have verified that the changes in the masses of these particles are negligible. In this regard,
we take a conservative approach and use the analytic formula [66, 67]
q q
 ∆V for ∆V
< vJ (α) ,
αρR
vw = q αρR (5.10)
∆V
1 for αρR ≥ v J (α) ,
√
2
where vJ = √1 1+ 3α +2α
3 1+α is the Chapman-Jouguet velocity.

direction All ϕ − ϕ′ ϕ−χ ϕ′ − χ ϕ ϕ′ χ

label 0 1 2 3 4 5 6

Table 2. All possible patterns of a PT in this model. The direction row indicates which fields
underwent changes in their vev. For example, the transitions with label 0 have all fields changing
their vev while those with the label 4 only the Higgs field vev changes.

– 12 –
 As the symmetry of the theory is restored at sufficiently high temperature, the vacuum
expectation value of the fields must evolve so that the symmetry is spontaneously broken
as the universe cools. Different values of the parameters of the theory can lead to phase
transitions with very distinct dynamical evolution of the fields. We classify all of the
possibilities in table 2.
The results of the parameter scan reveal that a strong first order phase transition can
only occur for labels 0 and 2. This symmetry breaking pattern corresponds to the condition

∆ϕ ̸= 0 & ∆χ ̸= 0 & (∆ϕ′ = 0 || ∆ϕ′ ̸= 0) , (5.11)

which means that during the phase transition the change in vacuum expectation value of
the h and χ fields is always different from zero whilst the extra singlet state ϕ′ may or may
not change its vev. Other patterns were also obtained, but they suffer from a very weak
transition strength and fall below the crossover regime, α ≈ 10−3 . Below this regime, we
expect that a lattice study would reveal a different nature of the transition. Moreover, we
find that each PT pattern is in a one-to-one correspondence with a single benchmark point
each, thus there are no multi-step PTs.
From the data, we inspect the Spearman rank correlation coefficient ρ between the
strength and the model parameters. The result is that the DM mass-squared, µ2DM , has
the largest correlation with ρ = −0.7. We display the results in Fig. 3, showing the variation
of α with µ2DM /v 2 .
We do not find any PT with label 6 in the scan, meaning that the transitions happening
solely across the DM direction are ruled out. At the same time, the transitions in the
Higgs direction only, which are relevant for EW baryogenesis, fall in the crossover region.
However, we stress that transitions with labels 0 and 2 could also be used in the context of
EW baryogenesis, but we leave this topic for another study.

6 Gravitational waves

In this section, we estimate the GW spectrum from bubble dynamics in a first-order phase
transition. Several dedicated numerical simulations have been performed in the literature
in which the hydrodynamic interactions between the plasma and the accelerating bubble
walls are solved on a lattice computation, see Refs. [68–73]. The qualitative results of these
lattice simulations can be encoded in a template function that describes the GW strain
amplitude as a function of the frequency. In the present model, assuming that the potential
does not exhibit a significant amount of supercooling, we expect the dominant contribution
to be from the sound wave source. We use the sound wave source template from Ref. [74],
which reads
ΩGW = Ωp S(f ), (6.1)

where the peak amplitude is obtained as

 2  −1
−5 1/3 0.6κα β
Ωp = 1.64 × 10 (100/g∗ ) 1/3
(8π) max(vw , cs ) min(H∗ τsh , 1). (6.2)
1+α H∗

– 13 –
Figure 3. Strength of the PT as a function of DM mass-squared parameter normalized to the EW
scale. The color coding of the scatter plot corresponds to the PT patterns of table 2. The gray area
in the figure marks the region with crossover transitions.

Figure 4. Gravitational wave spectrum for the benchmark points with the strongest phase transi-
tions. The color label of each curve corresponds to each type of transition as displayed in table 2.
The prospective experimental sensitivities of various detectors are also shown.

Here the subscript ∗ indicates that quantities are evaluated at the production time. In our
case, this corresponds to the temperature at which the bubbles of the new phase percolate
Tp ≈ Tn . For the sound speed in the plasma, we take c2s = 1/3. The efficiency factor κ is
taken from the seminal work [75]. The estimated decay time into turbulence is encoded in

– 14 –
 q
the variable H∗ τsh = 43 1+α 1/3 max(v , c ) (β/H )−1 .
0.6α (8π) w s ∗
The spectral function adopts the form of a double-broken power law
 3  2 !−1  4 !−1
f f f
S(f ) = N 1+ 1+ , (6.3)
f1 f1 f2
R∞
where the normalization factor is found by requiring −∞ d log f S(f ) = 1, while the fre-
quency breaks are obtained as
 g 1/6  T 5 max(vw , cs )
 
−6 ∗ ∗ β 1 f2
f1 = 1.13 × 10 , = . (6.4)
100 100GeV H∗ max(vw , cs ) f1 2 |vw − cs |
We display the associated GW signals in Fig. 4 and we superimpose the projected sensi-
tivities of LISA and AEDGE [76]. We only display the predictions coming from benchmarks
with very strong transitions strengths. These correspond to transition patterns of labels 0
and 2.

7 Dark matter

As we have discussed before, the U (1)Lµ −Lτ gauge symmetry takes an important role in
explaining the neutrino oscillation parameters and the muon (g − 2) by its implications
on the structure of the neutrino mass and by the presence of the extra gauge boson. In
this section, we examine another important effect in determining the DM relic density.
Before discussing this further, we first comment on the need for the two-component DM. In
principle, one can have a single-component scalar DM and we can solve DM relic density
bound from the Planck data. However, as we have shown in the GW section above, a strong
first-order PT would require high values of λDϕ and λDϕ′ , which are already in tension with
the direct detection data obtained by the LUX-ZEPLIN collaboration [39]. One method to
overcome this potential constraint from the direct detection of scalar DM is to introduce a
second DM candidate, a dark vector-like fermion ψ. The fermionic DM component would
be present in the Universe in higher amounts and reduce the effective direct detection cross-
Ω
section for the scalar DM by a multiplicative factor ΩΦ ΦDM+Ωψ . The detailed description of
DM
the model has been addressed in the model, see section 2. To study the two-component
DM scenario, we have implemented our model file in FeynRules [77] and then generated
the CalcHEP [78] files to use in micrOMEGAs [79]. Moreover, we compare our output from
micrOMEGAs with our analytical estimates. In this section, we first discuss the analytical
estimate and then proceed to present our results, where the correlations are shown among
the different model parameters generated using micrOMEGAs.
The abundance of DM can be obtained after solving the following Boltzmann equation,
dYA S(x) ⟨σv⟩
YA2 − YA,eq
2
(7.1)


=− ,
dx xH(x)
m
where A = ψ, ΦDM , x = ψ,DMT , and T is the temperature of the Universe. The entropy
(S(x)) and the Hubble parameter (H(x)) can be expressed as
2π 2 π  gρ 1/2 m2A −2
S(x) = gs m3A x−3 , H(x) = x , (7.2)
45 3 10 MP l

– 15 –
where mA = mψ , mDM is the DM mass, gs and gρ are the relativistic d.o.f associated
with the entropy and the matter. Although gs and gρ slightly vary during the evolu-
tion of the Universe, we  consider them to be equal to the effective d.o.f. defined as
√
g∗ = √ggsρ 1 + 13 gTs dg
dT
s
[80]. In the above equation, we can solve in a two-step pro-
cess by determining the freeze-out temperature Tf = m xf . The freeze-out temperature
A

represents the temperature of the Universe during its evolution when the DM interaction
strength falls below the Hubble expansion rate which can be determined as

nA ⟨σv⟩AA = H at x = xf , (7.3)

where the number density of DM A is nA = gπA2 m3A x−3/2 e−x , gA is internal d.o.f of particle
A, and the effective cross-section times velocity in the non-relativistic limit can be expressed
as ⟨σv⟩ = aA + 6bxA where aA and bA are the s−wave and p−wave annihilation modes. After
applying the aforementioned expression, we obtain the freeze-out temperature by solving
the following equation,

⟨σv⟩
 
1 m 
A 1 1 h g∗ i
x = 23.64 + ln x + ln + ln gA + ln − ln . (7.4)
2 100 2 2.6 × 10−9 GeV2 2 106

After solving Eqn. (7.4), we can determine the freeze-out temperature and thereby
solve the Boltzmann equation as given in Eqn. (7.1). By calculating the evolution of the
Boltzmann equation from x = xf to x = ∞, we can estimate the amount of the co-moving
number density as

YA, eq (xf )
YA = h q i   , (7.5)
2π 10 1 6bA
1 + YA, eq 15 gρ gs mA MPl xf aA + xf κ

where κ = 1/2 in case the particle and anti-particle are not the same. Otherwise κ = 1 [80].
The factors aA and bA can be calculated for the ΦDM and ψ DM candidates. We
present here only the DM annihilation to the extra gauge bosons, which is the dominant
mode in the present work. The result reads as follows,
4
gZ4 ′ qDM 4
gZ4 ′ m2Z ′ qDM
aΦDM ≃ −
,
π 8m2DM π 16m4DM

11gZ4 ′ qDM
4 3gZ4 ′ m2Z ′ qDM
4
bΦDM ≃ − +
,
π 192m2DM π 128m4DM

gZ4 ′ qψ4 gZ4 ′ qψ4 m2Z ′ 3gZ4 ′ qψ4 m4Z ′

aψ ≃ − − ,
16m2ψ π 32m4ψ π 128m6ψ π
3gZ4 ′ qψ4 29gZ4 ′ qψ4 m2Z ′ 23gZ4 ′ qψ4 m4Z ′
bψ ≃ + + . (7.6)
128m2ψ π 768m4ψ π 1024m6ψ π

Once we obtain the co-moving number density YA , we can determine the total DM density,
m 
A
X X
ΩDM h2 = ΩA h2 = 2.755 × 108 YA . (7.7)
GeV
A=ΦDM ,ψ A=ΦDM ,ψ

– 16 –
 We have checked that the analytical result obtained using the above expressions and
micrOMEGAS match each other with less than 10 % discrepancy3 . In our study, we have
used micrOMEGAS to generate the DM relic density and its associated observables. In this
respect, we have varied the model parameters within the following ranges,
mϕ
≤ mϕ′ [GeV] ≤ 2000 , 10−4 ≤ mZ ′ [GeV] ≤ 1 , 1 ≤ mψ [GeV] ≤ 1000 ,
2
10−6 ≤ gZ ′ ≤ 10−1 , 10−4 ≤ sin θ ≤ 0.5 , 1 ≤ qDM ≤ 105 ,
µ2
1 ≤ qψ ≤ 105 , 10−2 ≤ λDϕ ≤ 3 , 10−2 ≤ λDϕ′ ≤ 3 , −2 ≤ DM < 4. (7.8)
v2
It is worth mentioning that we have chosen the gZ ′ and mZ ′ values to explain the muon
(g − 2), which requires small gZ ′ values, i.e., gZ ′ ∼ 10−3 . Therefore, in order to have a
dominant contribution of scalar and fermion DM to the gauge boson, we have considered
relatively larger values for their U (1)Lµ −Lτ charges, qDM and qψ .

7.1 Dark matter bounds

Below are listed the points that satisfy the bounds associated with the DM.

• DM relic density bound: In the present work, we have considered the Planck
bound on the total DM relic density coming from the fermion and scalar sectors.
In particular, we have considered 7σ variation in DM relic density as shown below
[81, 82],

ΩDM h2 = (Ωψ + ΩΦDM ) h2 = 0.1200 ± 0.0084 . (7.9)

We consider a wide range for DM relic density around its central value ΩDM h2 = 0.12
to get a large number of points in reasonable time while satisfying the DM relic density.
Furthermore, if we consider the narrower 1σ range in DM relic density as given in the
Planck paper [81, 82] instead, we would not expect any change in our conclusion. In
such case, a slight reduction in the allowed parameter space in different planes would
be expected. The main diagrams determining the relic densities of ϕDM and ψ DM
candidates are shown in Fig. 5.

• DM Direct detection: In this work, we have considered direct detection bound

on the scalar DM and no bound applied to the fermion DM due to its inert nature.
Scalar DM can interact with the nucleon of the detector as shown in Fig. 6. The spin-
independent direct detection (SIDD) cross-section for the scalar DM can be estimated
in the non-relativistic limit as

gΦ† Φ ϕ cos θ gΦ† Φ ϕ′ sin θ 2

!
µ2red fN
2 M2
σSI = N DM DM
− DM DM
, (7.10)
4πm2DM v 2 m2ϕ m2ϕ′
 
3 1 K1 (x)
The discrepancy further decreases once we multiply ⟨σv⟩ by 2
1+ K2 (x)
, where Ki (x) is the modified
Bessel function of the second kind [80].

– 17 –
 ΦDM Z′ ΦDM
φ, φ′

φ, φ′
φDM
Φ†DM Z′ Φ†DM
φ, φ′

ΦDM φ, φ′ , Z ′ ψ
Z′

Φ†DM φ, φ′ , Z ′ ψ̄
Z′

Figure 5. Main annihilating diagrams for the DM candidates, ΦDM and ψ, that determine their
relic density.

where µred = mmDM +MN is the reduced mass between DM and nucleon, MN is the
DM MN

nucleon mass, fN ∼ 0.3 [83] is the nucleon form factor and the DM couplings with
the Higgs bosons taking the following form,

gΦ† Φ ϕ′ = λDϕ′ v ′ cos θ + λDϕ v sin θ ,

 
DM DM

gΦ† Φ ϕ = λDϕ′ v ′ sin θ − λDϕ v cos θ . (7.11)

 
DM DM

In the present work, we have considered the most recent bound coming from the
measurement by the LUX-ZEPLIN collaboration [39]. All the points which will be
shown in the later plots are obtained after satisfying the LUX-ZEPLIN data. Since
we have a component DM scenario, one of the components can be small and the other
component can be large. Therefore, while comparing with the LUX-ZEPLIN data
we have used an effective SIDD cross-section fΦDM σSI , where fΦDM is the fraction of
scalar DM present in the Universe.

• DM Indirect detection: In the present work, the scalar DM can also annihilate to
the SM particle through Higgses mediator process as Φ†DM ΦDM → f¯f, V V where f is
SM fermions and V is the SM gauge bosons. One can estimate the indirect detection
cross-section by using the following relation,
Z ∞ √ 
1 2

√ s
⟨σv⟩ΦDM = 4 2 mDM

σΦ† Φ →f f¯ s − 4mDM sK1 ds .
8mDM T K2 T 4m2DM DM DM T
(7.12)

In the present work, DM mainly annihilates to the extra gauge boson Z ′ and the
Higgses ϕ, ϕ′ . Hence, DM annihilation to the SM visible sector is suppressed. It
also does not effectively take part in the DM relic density calculation. Therefore, it

– 18 –
 ΦDM ΦDM ΦDM f

φ, φ′ , Z ′
φ, φ′
Φ†DM
N N f¯

ψ µ− , τ −

Z′

ψ̄ µ+ , τ +

Figure 6. DM direct detection and indirect detection for the scalar and fermionic DM candidates.

is hard to detect DM at the present with the indirect detection techniques because
of the suppressed direct annihilation. Although there will be a possibility to detect
DM by indirect detection from the cascade decays of the final states gauge boson and
Higgses, which we expect to be suppressed. The diagrams associated with the indirect
detection of ϕDM and ψ are shown in Fig. 6.

• Collider bounds:
We have considered the relevant bound which may appear from the different collider
searches. We have taken into account the relevant bounds that can appear for the
SM and BSM Higgs by using the HiggsSignal [84] and HiggsBounds [85]
packages. The experimental constraints that are important for the present work are
Higgs signal strength, its decay to the invisible sector and BSM Higgs searches at the
collider through its decay to the visible sector. We have used the in-built HiggsBound
and HiggsSignal packages in micrOMEGAs by externally feeding with the appropriate
model file with the relevant couplings.

7.2 Allowed parameter regions from dark matter

We are now are ready to present the scatter plots that display the correlation between
relevant parameters. We have ensured that the results satisfy all of the experimental con-
straints described in section 7.1. We obtain a very nice correlation between the parameters
and the observables, as we will describe below in detail. The main contributing diagrams
which determine the ϕDM and ψ DM relic densities are shown in Fig. 5.
In Fig. 7, we show the scatter plots for the mZ ′ − gZ ′ plane where the colour variation
corresponds to the different values of λϕ′ , qψ , mDM and λDϕ′ from the top left to the
bottom right figures, respectively. While generating the plots, we have checked that all
constraints relevant for DM are satisfied as discussed in Section 7.1. In the top left panel,

– 19 –
Figure 7. Scatter plots in the mZ ′ − gZ ′ plane after satisfying the DM relic density bound. The
colorbars in the four plots represent quartic coupling λϕ′ , charge of the extra fermion qDM , scalar
DM mass mDM and quartic coupling λDϕ′ . We have shown the bounds coming from the neutrino
trident experiment CCFR, BBN and future access by the muon collider, M3 and NA64µ .

we show the colour variation in the quartic coupling λϕ′ . As can be seen in Eq. (2.10), the
2
gZ
quartic coupling varies with the vev of singlet scalar as λϕ′ ∝ 1
v ′2
= ′
m2Z ′
. One can clearly
see the colour variation follows the aforementioned equation. For example, if one fixes gZ ′
and increases the value of mZ ′ , then a decrement in the λϕ′ appears. Similarly, if one fixed
mZ ′ and increased the value of gZ ′ in increments, then one would see an increment in the
quartic coupling. There are no points in this panel that would involve high values of gZ ′
and low values of mZ ′ due to the pertubativity bound on the quartic coupling λϕ′ < 4π,

– 20 –
Figure 8. Left (light) panel exhibits the scatter plot in the mψ − g ′ qψ (mDM − g ′ qDM ) plane. In
the left panel (right panel), the colorbar shows the amount of fermionic (scalar) DM contribution
compared to the total amount of DM in percentage.

as can be seen above the magenta points. Furthermore, we also have no points below
the green points, which can be better understood by looking at the lower panel. In the
figure, red patch explains the muon (g − 2) anomaly. We can see that mZ ′ > 0.1 GeV is
already in conflict with the neutrino trident bound obtained from CCFR experiment [86],
which is depicted with the blue region. The light brick region shows the values that are
ruled out by BBN [87], whereas the dashed black line, dashed red line and dashed light
blue line represent the future sensitivities of the muon collider [88], NA64µ [89] and M3
experiments [90], respectively. In the upper right panel, we show the same scatter plot but
with the colour variation corresponding to the U (1)Lµ −Lτ charge qψ . It can be seen that
for qψ ∼ 100, the model can satisfy both the DM and muon (g − 2) constraints at the same
time, whilst the smaller values of qψ are ruled out by the CCFR bound and the larger values
correspond to the the smaller contribution to the (g − 2)µ anomaly. In the bottom left and
bottom right panels, it can be understood that the absence of points below the magenta
points in the bottom left panel and green points in the bottom right panel. This mainly
results from the presence of the upper bound on the scalar DM, which we have considered in
the present work to be mDM < 2000 GeV. Therefore, the aforementioned effect comes from
our choice of the model parameters. As we have shown before, m2DM ∝ λDϕ′ v ′2 , meaning
that for the values of v ′ obtained in those regions one needs λDϕ′ < 0.01. This is the reason
why we did not obtain any points in the lower right corners in the panels. This can be seen
in the bottom right panel in particular, where green points corresponding to λDϕ′ > 0.01
can be found.
In the left panel of Fig. 8, we have shown the scatter plot in the mψ − gZ ′ qψ plane.

– 21 –
Figure 9. Left and right panels show the scatter plots in the qψ − g ′ and qDM − g ′ planes,
respectively. In the left (right) panel, the color variation shows the amount of fermionic (scalar)
DM contribution with respect to the total DM density in percentage.

m
Eqns. (7.5), (7.6) and (7.7) let us infer that DM relic density varies as Ωψ h2 ∝ (qψ g ψ′ )4 .
Z
To get the correct value of DM relic density, one needs to look at the sharp magenta line
which gives 100% contribution to the DM density from ψ. If one looks at the higher values
of qψ gZ ′ , there is a lower contribution from ψ in the total DM contribution. The upper
√
bound on the qψ gZ ′ comes from the perturbativity bound, which is given as qψ gZ ′ < 4π.
In the right panel, if we take a fixed value of qψ gZ ′ and go towards the right end for higher
values DM mass mψ , we can then see a higher contribution from from ψ DM as depicted by
the magenta points. These points are in agreement with the analytical expression shown
before. So far the behaviour of DM is mainly controlled by the ψ̄ψ → Z ′ Z ′ . As one crosses
the mψ > 60 GeV region, however, one then has also lower values of gZ′ ′ qψ . This is due
to the fact that for those points one has a new annihilation mode of DM. In other words,
ψ̄ψ → Φ†DM ΦDM opens up and has the dominant contribution. The aforementioned points
are also affected by the qDM values. In the right panel, we have shown a similar plot for the
scalar DM candidate. We can thus draw the same conclusion for the present figure as well,
but we can also see that very lower values of gZ ′ qDM are allowed, which was not possible
for the left panel. This is because those lower values do not contribute to DM relic density
significantly instead Φ†DM ΦDM → ϕϕ, ϕ′ ϕ′ , ϕϕ′ , f f¯ annihilation modes are active which are
mainly governed by the λDϕ , λDϕ′ quartic couplings.
In the left panel and right panel of Fig. 9, we show the scatter plots in the qψ − gZ ′ and
qDM − gZ ′ planes, respectively. For a better understanding of the parameter dependence on
the DM relic density, we also show the fraction of the fermion DM present as percentage of
the total DM contribution in the left panel and for the scalar DM in the right panel. In the

– 22 –
Figure 10. left and right panels show the scatter plots in the λDϕ −fΦDM σSI and mΦDM −fΦDM σSI
planes, respectively. The color variation in the left panel represents the amount of scalar DM in
percentage compared to the total amount of DM density, whereas the right panel shows the variation
of different values gZ ′ qDM .

left panel, we can see that qψ and gZ ′ are anti-correlated, which means that for a particular
value of qψ gZ ′ , we get ψ contribution in the total DM relic density which does not evade the
Planck relic density bound. In the left panel, we can see for a fixed value of qψ , increasing
the value of gZ ′ would lead to a lower contribution of ψ DM in the total DM density. This
can be clearly seen from the colour variation in the left panel. This behaviour can be easily
understood from the analytical estimate, which we have presented in Eq. (7.6). On the
other hand, the region above the green points is ruled out as a result of the perturbativity
bound on gZ ′ qψ , whereas the region below the deep magenta points is ruled out due to
overproduction of ψ DM. In the right panel, we have shown a scatter plot in the qDM − gZ ′
plane and colour variation shows the fraction of scalar DM in percentage. Again, we find
the region above the green points to be ruled out because of the perturbativity bound
√
gZ ′ qDM < 4π, but we do not have any constraints from below. This happens since the
lower region is dominated by the DM annihilation modes like Φ†DM ΦDM → ϕϕ, ϕ′ ϕ′ , ϕϕ′ , f f¯.
In Fig. 10, we show the scatter plots in the λDϕ −fΦDM σSI and mDM −fΦDM σSI planes.
The colorbar in the left panel describes the fraction of scalar DM present in percentage,
while the right panel represents the different values of gZ ′ qDM . In the left panel, it can
be seen that if we increase the value of the quartic coupling λDϕ , then more annihilation
occurs and there are more green points in the plot. It is also found that low values of
effective direct detection spin-independent cross-section lead to a low abundance of scalar
DM. This is evident from the colour variation in the left panel. On the other hand, in the
right panel we show the scatter plot in the mDM − fΦDM σSI plane. In this case, all points
are obtained after passing through the LUX-ZEPLIN direct detection bound. It can be seen

– 23 –
in the right panel that the points around the ν−floor have both higher and lower values
of gZ ′ qDM depending on the values of the quartic couplings λDϕ , λDϕ′ . We can also see
the points which correspond to the lower values of the SIDD cross-section (deep magenta
points) represent a very small fraction of scalar DM. We can finally see there is a fraction
in the allowed region below the ν−floor which will be hard to probe by the current set-up
of the direct detection experiments.

8 Gravitational waves and dark matter

This section shows the correlations between GW and DM via the thermodynamic parame-
ters of the PTs after implementing the bounds coming from DM relic density, collider and
direct detection. As we will see in the plots a large portion of the parameter spaces are ruled
out after implementing the collider and dark sector bounds. In showing the correlations,
we have varied the model parameters in the range shown in Eqn. (5.1). Moreover, we have
√
kept gZ ′ and mZ ′ fixed at 10−3 and 0.1 GeV, respectively, and require |gZ ′ qDM | ≤ 4π.
In the left upper panel of Fig. 11, we have shown the scatter plot in the α − β/H
plane where the colour variation implies the variation of scalar DM relic density. We can
see anti-correlation between α and β/H. This is expected, as the larger values of α mean
a larger separation between the two vacua, resulting in a larger transition time between
the two vacua and vice versa. The larger values of α also mean higher emissions of latent
heat during the transition which increases the Hubble parameter. In the plots, we have
shown the points that give strong first-order PTs along the 0, 2, 3 directions as indicated
in Table 2. As can be seen from the lower panels of Fig. 11, one needs higher values of λDϕ
to obtain the strong first-order PTs prompting more annihilations for the scalar DM. We
get at most ΩΦDM h2 = 0.01. This is not a problem because we also have fermionic vector
DM in the present work which can give us the rest of the contribution in the total DM relic
density. This is seen in the upper right panel plot where we have shown the variation of the
fermionic DM relic density variation with its mass for three different values of the charge
qψ . From the plot, we can see that as we increase the value of qψ then we have a lower
contribution from ψ in the DM relic density for the same value ψ mass, mψ . Therefore,
by tuning the qψ values we can easily satisfy the total DM relic density given by Planck.
In the lower panel, the left and right plots show the allowed points in the mDM − λDϕ
plane after demanding strong first-order PTs. The colour variations both in the left panel
and right panel show the values of α and β/H. The values of these two parameters are
anti-correlated, as exhibited in the colour variation. The lower panels of Fig. 11 indicate
that one needs high values of the quartic coupling λDϕ to have the strong first-order PTs.
These values of λDϕ can also be constrained by the DM direct detection as will be shown
later in this section.
In Fig. 12, we show the scatter plots in the mΦDM −fΦDM σSI plane after demanding the
strong first-order PTs. In both panels, we can see through the colour variation the values
of α, β/H distributed over the allowed region. As we discussed in context of the previous
figure, one needs large values of λDϕ to achieve strong first-order PTs, which results in a
large portion of parameter space being ruled out by the direct detection (DD) experiment

– 24 –
Figure 11. Upper: Left panel shows the scatter plot in the α − β/H plane where the colour
variation shows the scalar DM relic density. Right panel shows the variation of the fermion DM
relic density with its mass for three different values of U (1)Lµ −Lτ charges of the fermionic DM.
Lower: let and right show the scatter plots in the mDM − λDϕ plane where the colour variation
shows different values of α (β/H) in the left panel (right panel).

LUX-ZEPLIN. We will see in the next figure that the exclusion of a large portion of the
parameter space by the DD does not hamper the potential of having all the allowed range
of α and β/H for the strong first-order PTs.
In the left panel and right panel of Fig. 13, we have shown scatter plots in the mϕ′ −sin θ
and α − β/H planes after imposing a selection of bounds from strong first-order PTs, DM
relic density, collider and DM direct detection constraints. In both panels, the green points
are obtained by implementing strong first-order PTs and upper bound on DM relic density
i.e. ΩΦDM h2 ≤ 0.12, whereas the cyan points are obtained by additionally applying the
collider bounds, which mainly come from the Higgs signal strength and Higgs branching to

– 25 –
Figure 12. Scatter plots in the mΦDM − fΦDM σSI plane where the colour bar in the left (right)
shows the variation of α (β/H).

Figure 13. Left panel shows the scatter plot in the mϕ′ − sin θ plane and RP shows the scatter
plot in the α − β/H plane. Three different colours represent different bounds coming from strong
first-order PTs, DM relic density, collider and direct detection on their subsequent implementation
on the allowed data.

invisible decay. Finally, the blue points are obtained by additionally applying the DM direct
detection bound. In the case of collider bounds associated with the SM Higgs and additional
Higgs have been implemented using the HiggsSignal and HiggsBound, respectively. In the
left panel, we can see that when the collider bounds are applied, the values sin θ > 0.02 are
ruled out due to the SM Higgs decaying to the additional gauge boson and further decaying

– 26 –
to neutrinos. In Ref. [91], the ATLAS collaboration has displayed the allowed branching
ratio for the SM Higgs invisible decay, Brϕ→inv < 10.7 %. Adopting this value, we have also
checked analytically that,

Γϕ→Z ′ Z ′
≤ 10.7% ⇒ sin θ < 0.02 , (8.1)
Γϕ

where Γϕ = 4.1+5.0
−4.0 MeV [92] and
s !
m3ϕ sin2 θ 4m2Z ′ 4m2Z ′ 12m4Z ′
Γϕ→Z ′ Z ′ = 1− 1− + . (8.2)
32πv ′2 m2ϕ m2ϕ m4ϕ

Moreover, when we apply the DM DD bound then we have fewer allowed spaces represented
by the blue points. Those blue points represent a very suppressed fraction of the scalar DM
relic density in the range ΩΦDM h2 ≃ 10−5 − 10−7 , as otherwise the points would be ruled
out by the DM SIDD bound. This can easily be understood from Eqns. (7.10) and (7.11),
which imply that DM SIDD is linearly proportional to λDϕ and suppression can only appear
in the case the model has a very small fraction of the scalar DM, fΦDM . On the other hand,
in the right panel, we can see that even after implementing the stronger bound from the
DD, we can have relatively all the allowed range for α and β/H which we obtain after
implementing the strong first-order PTs bound as well. Therefore, we may conclude that
severe bounds from the collider and direct detection experiments can reduce the allowed
parameter space but do not alter GW signal strength significantly.

9 Conclusions

The possibility of gauging the U (1)Lµ −Lτ symmetry as an extension of the Standard Model
is a well-motivated endeavour and has recently become a subject of increasing interest.
In this work, we investigated a non-minimal version of this model, featuring two complex
singlet scalars, three Majorana right-handed neutrinos and a vector-like fermion.
Like in its original version, the model offers a simple resolution to the long-standing
puzzle of the muon (g − 2) anomaly via the presence of an additional one-loop diagram
mediated by the Z ′ gauge boson. At the same time, the RHNs provide a viable type-I
seesaw mechanism, and the choice of U (1)Lµ −Lτ charge for one of the extra complex scalars
induces a two-zero minor structure along the diagonal elements of the RHN neutrino mass
matrix. We tested the viability of the neutrino oscillation parameters predicted in this
model in light of the recent data from the NuFit collaboration and found that the model
can successfully account for all experimental observables. Additionally, we confirmed that
maximal θ23 = π/4 as well as maximal CP phase δCP are possible in this model only for
large values of the sum of neutrino masses, leading to a potential conflict with cosmological
data. The present model can also explain the 1 σ allowed range for θ23 , with/without
atmospheric data representing the lower/higher octant values without violating any other
constraints. Moreover, we determined the δCP using the Jarlskog variable and found values
within the allowed range provided by the NuFit collaboration.

– 27 –
 It is well known that the interaction of an additional scalar with the SM Higgs doublet
can easily trigger a potential barrier and render the EW phase transition strongly first-order.
Thus, it is perhaps not surprising that our model, with its three-dimensional potential, offers
a vast array of possibilities in terms of symmetry-breaking patterns. We studied the finite
temperature evolution of the one-loop effective potential, taking into account all three field-
space dimensions consisting of the SM Higgs doublet and two additional singlet scalars,
including the DM candidate. Our analysis showed that the strongest phase transitions
occur in a well-defined range of the DM squared-mass parameter corresponding to values
µ2DM ≥ −v 2 and that the Lagrangian parameter µ2DM has the strongest correlation with
the phase transition latent heat. We have chosen µ2DM values such that we always get
mϕDM > 0 by following Eqn. 2.9.
We then studied dark matter in association with GW and discussed the associated
phenomenology. First, we focused on the new annihilation mode of scalar DM to the
additional gauge bosons. This can be achieved by considering large values of the U (1)Lµ −Lτ
gauge charge of the scalar DM. Small values of the charge can also serve the purpose
if we do not aim to explain the muon (g − 2). This part of the DM study does not
depend on the quartic couplings λDϕ and λDϕ′ , but this is not true when we focus on GW.
Demanding the strong first-order PTs requires large values of the quartic couplings, which
affect our scalar DM, from its abundance to the bound from direct detection. We found
that the parameters associated with the scalar DM corresponding to the strong first-order
PTs give us negligible contributions to the DM relic density, i.e., ΩDM h2 < 0.01. Moreover,
when we apply the DM direct detection bound on scalar DM, the scalar DM contribution
further diminishes to ΩDM h2 ∼ 10−5 − 10−7 . Therefore, to obtain the remainder of DM
contribution, we introduce another vector fermion DM, which can easily fill the rest of
the DM density without altering the other phenomenologies. The additional fermion DM
mainly annihilates to the additional gauge boson Z ′ , and its contribution fully depends on
the U (1)Lµ −Lτ charge values taken for the fermion DM. We also found that sin θ > 0.02 is
ruled out mainly due to the tight constraint on the SM Higgs invisible decay width.
In summary, our present model can simultaneously explain several beyond SM prob-
lems, namely muon (g − 2), neutrino mass, dark matter, and first-order phase transition. In
the future, we will further explore the relationship between GW and baryon asymmetry in
the canonical leptogenesis mechanism, as well as the Dirac CP phase in the present setup.

Acknowledgements

The computations were enabled by resources provided by the National Academic Infras-
tructure for Supercomputing in Sweden (NAISS) at KTH partially funded by the Swedish
Research Council through grant agreement no. 2022-06725. This work used the Scientific
Compute Cluster at GWDG, the joint data center of Max Planck Society for the Advance-
ment of Science (MPG) and University of Göttingen. This project has received funding
/support from the European Union’s Horizon 2020 research and innovation programme
under the Marie Skłodowska-Curie grant agreement No 860881-HIDDeN.

– 28 –
A Neutrino parameters

The results obtained in this work have been checked against the NuFit 5.2 dataset [40].
In this section, the constrained values of the neutrino oscillation parameters are presented
for the normal ordering of neutrino masses. When the atmospheric neutrino data from Su-
perKamiokande is taken into account, the best-fit values and associated 1σ CL uncertainties
are given by
sin2 θ12 = 0.303+0.012
−0.012 , (A.1)
−5
∆m221 = 7.41+0.21
−0.20 × 10 eV2 (A.2)
sin2 θ23 = 0.451+0.019
−0.016 , (A.3)
−3
∆m231 = 2.507+0.026
−0.027 × 10 eV2 (A.4)
−2
sin2 θ13 = 2.225+0.056
−0.059 × 10 , (A.5)
◦
δCP = 232+36
−26 . (A.6)
When the atmospheric neutrino data from SuperKamiokande is not taken into account, the
best-fit values for the neutrino mixing parameters are instead

sin2 θ12 = 0.303+0.012

−0.011 , (A.7)
−5
∆m221 = 7.41+0.21
−0.20 × 10 eV2 (A.8)
sin2 θ23 = 0.572+0.018
−0.023 , (A.9)
−3
∆m231 = 2.511+0.028
−0.027 × 10 eV2 (A.10)
−2
sin2 θ13 = 2.203+0.056
−0.059 × 10 , (A.11)
◦
δCP = 197+42
−25 . (A.12)
The three mixing angles θ12 , θ13 and θ23 can also be extracted from the PMNS matrix
U as follows [93]:
sin2 θ13 = |U13 |2 , (A.13)
|U23 |2
sin2 θ23 = , (A.14)
1 − |U13 |2
|U12 |2
sin2 θ12 = . (A.15)
1 − |U13 |2
To calculate the CP phase δCP , we use the Jarlskog invariant, which is given by
∗ ∗
(A.16)
 
JCP ≡ Im Uαi Uαj Uβi Uβj
max
≡ JCP sin δCP = c12 s12 c23 s23 c213 s13 sin δCP , (A.17)

where cij and sij stand for cos θij and sin θij , respectively. The CP phase can therefore be
solved from the equation
∗ U∗ U
 
Im Uαi Uαj βi βj
sin δCP = max . (A.18)
JCP

– 29 –
In order to determine the correct geometrical phase for δCP , one must also obtain an
analytical expression for cos δCP . This can be achieved by examining the real part of the
matrix product Re Uαi Uαj βi βj , which results in
∗ U∗ U
 

∗ U∗ U 2 2 2 2
 
Re Uαi Uαj βi βj + s12 s13 c13 s13
cos δCP = max . (A.19)
JCP

The CP phase δCP can now be determined by comparing the values of sin δCP and cos δCP
in Eqns. (A.18) and (A.19). If sin δCP > 0 and cos δCP > 0, δCP can be obtained directly
from sin δCP . If sin δCP > 0 and cos δCP < 0, the correct CP phase is obtained with
transformation δCP → π − δCP . If sin δCP < 0 and cos δCP < 0, the phase is given as
δCP → δCP + π instead. Finally, if sin δCP < 0 and cos δCP > 0, then δCP → −δCP .
The charged lepton flavor violation constraints from MEG [55] are taken into account
by calculating the elements of the active-sterile neutrino mixing matrix UνN . Following
Ref. [56], we obtain the mixing between the light and heavy neutrinos from MD and MR
matrices:
X X
Ue2 = |(UνN )ei |2 = mD MR−1 . (A.20)

i i

We take the MEG limit on the radiative muon decay, Br(µ →) ≲ 7.5 × 10−13 , and calculate
the upper bound for Ue . Any data points not satisfying the MEG limit are then excluded
in the numerical analysis, as discussed in section 4. The model parameters values obtained
in the neutrino analysis fall within the following ranges:

Mee ∈ [−3.4, 88.3] × 107 ,

Mµτ ∈ [−1.2, 167.0] × 106 ,

ξ ∈ [−7.3, 58.5] × 104 ,

heµ ∈ [−1.3, 34.7] × 109 ,

heτ ∈ [−1.7, 9.2] × 109 ,

ye ∈ [−10.5, 7.6] × 10−4 ,

yµ ∈ [−15.6, 9.5],

yτ ∈ [−1.1, 3.8]. (A.21)

All of the data points satisfy the experimental constraints specified in Eqns. (A.7)-(A.11).
It should be noted that while some of the values considered for yµ and yτ are large, the
majority of the values obtained for these two parameters are close to zero. It is also found
that the values presented for ξ are multiples of π, which indicates that the off-diagonal
terms Mµτ eiξ in MR are always real.

– 30 –
B Field dependent masses

At finite temperature, all the radial modes from each scalar can mix and the mass matrix
can be written as

∂ 2 V0
M2 ≡
∂ϕi ∂ϕj
 
1
2
(χ2 λDϕ +6ϕ2 λϕ +ϕ′2 λϕϕ′ −2µ2ϕ ) ϕϕ′ λϕϕ′ ϕχλDϕ
= ϕϕ′ λϕϕ′ 1
(χ2 λDϕ′ +6ϕ′2 λϕ′ +ϕ2 λhϕ′ )+µ2ϕ′ ϕ′ χλDϕ′
 
2 
1
ϕχλDϕ ϕ′ χλDϕ′ 2
(ϕ2 λDϕ +ϕ′2 λDϕ′ +6χ2 λDM )+µ2DM

(B.1)

The field dependent masses for the scalars are obtained from the eigenvalues of Eqn. (B.1).
Since closed-form expressions are difficult to obtain or they are very lengthy, we choose not
to write them down here. Instead, they are calculated numerically.
On the other hand, the field dependent masses of the gauge bosons are easy to obtain,
namely

g2 2 g 2 + g ′2 2
m2W (ϕ) = ϕ , m2Z (ϕ) = ϕ , m2Z ′ (ϕ′ , χ) = gZ2 (qDM
2
χ2 + ϕ′2 ). (B.2)
4 4

The contribution from the quark sector is dominated by the top quark since it has the
largest Yukawa coupling, i.e.,
y2
m2t (ϕ) = t ϕ2 . (B.3)
2
We neglect the contribution from other quarks in this work.
The case of the right-handed neutrinos requires more attention as the field dependent
mass matrix is given by [13]

ϕ′ ϕ′
 
√ heµ √ heτ
 Mee 2 2 
 
 
 ′ 
′
 ϕ
(B.4)
iξ

 √2 heµ
MR (ϕ ) =  0 Mµτ e  .
 
 
 
 ϕ′ 
√ heτ Mµτ eiξ 0
2

One can multiply the matrix in Eqn. (B.4) by its Hermitian conjugate and digonalize the
result. Therefore,

U † MR† (ϕ′ )MR (ϕ′ )U = diag m2N1 (ϕ′ ), m2N2 (ϕ′ ), m2N3 (ϕ′ ) , (B.5)

where each neutrino eigenstate has 4 degrees of freedom. We do not include the contribution
from the left-handed neutrinos as their masses are tiny.

– 31 –
B.1 Daisy resummation
The thermally-corrected masses of the scalars are the eigenvalues of the thermally-corrected
Hessian matrix  
Πϕ (T )
M2 → M2 +  Πϕ′ (T ) . (B.6)
 
Πχ (T )
This corresponds to a resummation of a large number of daisy diagrams which enhance the
perturbative control at large temperatures [94, 95].
The thermal masses in this model are found to be
g ′2 λϕ λϕϕ′
 2
λDϕ yt2

3g
Πϕ = + + + + + T2 , (B.7)
16 16 2 12 12 4
!
gZ2 λ ϕ′ λϕϕ′ λDϕ′ h2eµ + h2eτ
Π ϕ′ = + + + + T2 (B.8)
4 3 6 12 12
and
2
gZ2 qDM

λDϕ λDϕ′

λDM
Πχ = + + + T2 . (B.9)
4 3 6 12
The longitudinal polarization states of vector bosons receive temperature corrections
as
m2gauge → m2gauge + Πgauge (T ). (B.10)

For the SM W bosons, this reads [96]

11 2 2
ΠW = g T (B.11)
6
and the eigenvalues of the mass matrix
!
1 2 2 11 2 2
4g ϕ + 6 g T − 41 gg ′ ϕ2
2
MZ/γ = (B.12)
− 14 gg ′ ϕ2 1 ′2 2 11 ′2 2
4g ϕ + 6 g T

determines the finite temperature mass of the Z boson and the photon. Similarly, the
longitudinal component of the Z ′ acquires the correction
1
ΠZ ′ = gZ2 qDM
2
+ 4 T 2. (B.13)


6

References

[1] Particle Data Group Collaboration, R. L. Workman et al., Review of Particle Physics,
PTEP 2022 (2022) 083C01.
[2] R. Foot, New Physics From Electric Charge Quantization?, Mod. Phys. Lett. A 6 (1991)
527–530.
[3] X. G. He, G. C. Joshi, H. Lew, and R. R. Volkas, NEW Z-prime PHENOMENOLOGY,
Phys. Rev. D 43 (1991) 22–24.

– 32 –
 [4] X.-G. He, G. C. Joshi, H. Lew, and R. R. Volkas, Simplest Z-prime model, Phys. Rev. D 44
(1991) 2118–2132.
[5] T. Araki, J. Heeck, and J. Kubo, Vanishing Minors in the Neutrino Mass Matrix from
Abelian Gauge Symmetries, JHEP 07 (2012) 083, [arXiv:1203.4951].
[6] A. Crivellin, G. D’Ambrosio, and J. Heeck, Addressing the LHC flavor anomalies with
horizontal gauge symmetries, Phys. Rev. D 91 (2015), no. 7 075006, [arXiv:1503.03477].
[7] K. Asai, K. Hamaguchi, and N. Nagata, Predictions for the neutrino parameters in the
minimal gauged U(1)Lµ −Lτ model, Eur. Phys. J. C 77 (2017), no. 11 763,
[arXiv:1705.00419].
[8] K. Asai, K. Hamaguchi, N. Nagata, S.-Y. Tseng, and K. Tsumura, Minimal Gauged
U(1)Lα −Lβ Models Driven into a Corner, Phys. Rev. D 99 (2019), no. 5 055029,
[arXiv:1811.07571].
[9] K. Asai, Predictions for the neutrino parameters in the minimal model extended by linear
combination of U(1)Le −Lµ , U(1)Lµ −Lτ and U(1)B−L gauge symmetries, Eur. Phys. J. C 80
(2020), no. 2 76, [arXiv:1907.04042].
[10] K. Asai, K. Hamaguchi, N. Nagata, and S.-Y. Tseng, Leptogenesis in the minimal gauged
U(1)Lµ −Lτ model and the sign of the cosmological baryon asymmetry, JCAP 11 (2020) 013,
[arXiv:2005.01039].
[11] A. Granelli, K. Hamaguchi, N. Nagata, M. E. Ramirez-Quezada, and J. Wada, Thermal
leptogenesis in the minimal gauged U(1)Lµ −Lτ model, JHEP 09 (2023) 079,
[arXiv:2305.18100].
[12] S. Baek and P. Ko, Phenomenology of U(1)(L(mu)-L(tau)) charged dark matter at PAMELA
and colliders, JCAP 10 (2009) 011, [arXiv:0811.1646].
[13] A. Biswas, S. Choubey, and S. Khan, Neutrino Mass, Dark Matter and Anomalous Magnetic
Moment of Muon in a U (1)Lµ −Lτ Model, JHEP 09 (2016) 147, [arXiv:1608.04194].
[14] A. Biswas, S. Choubey, and S. Khan, FIMP and Muon (g − 2) in a U(1)Lµ −Lτ Model, JHEP
02 (2017) 123, [arXiv:1612.03067].
[15] J. Heeck and W. Rodejohann, Gauged Lµ − Lτ Symmetry at the Electroweak Scale, Phys.
Rev. D 84 (2011) 075007, [arXiv:1107.5238].
[16] S. Baek, N. G. Deshpande, X. G. He, and P. Ko, Muon anomalous g-2 and gauged L(muon) -
L(tau) models, Phys. Rev. D 64 (2001) 055006, [hep-ph/0104141].
[17] E. Ma, D. P. Roy, and S. Roy, Gauged L(mu) - L(tau) with large muon anomalous magnetic
moment and the bimaximal mixing of neutrinos, Phys. Lett. B 525 (2002) 101–106,
[hep-ph/0110146].
[18] D. Borah, M. Dutta, S. Mahapatra, and N. Sahu, Muon (g−2) and XENON1T excess with
boosted dark matter in Lµ-Lτ model, Phys. Lett. B 820 (2021) 136577, [arXiv:2104.05656].
[19] Muon g-2 Collaboration, B. Abi et al., Measurement of the Positive Muon Anomalous
Magnetic Moment to 0.46 ppm, Phys. Rev. Lett. 126 (2021), no. 14 141801,
[arXiv:2104.03281].
[20] T. Aoyama et al., The anomalous magnetic moment of the muon in the Standard Model,
Phys. Rept. 887 (2020) 1–166, [arXiv:2006.04822].

– 33 –
[21] S. Eijima, M. Ibe, and K. Murai, Muon g − 2 and non-thermal leptogenesis in U(1)Lµ −Lτ
model, JHEP 05 (2023) 010, [arXiv:2303.09751].
[22] D. Borah, A. Dasgupta, and D. Mahanta, TeV scale resonant leptogenesis with Lµ-Lτ gauge
symmetry in light of the muon g-2, Phys. Rev. D 104 (2021), no. 7 075006,
[arXiv:2106.14410].
[23] S. Davidson and A. Ibarra, A Lower bound on the right-handed neutrino mass from
leptogenesis, Phys. Lett. B 535 (2002) 25–32, [hep-ph/0202239].
[24] LIGO Scientific, Virgo Collaboration, B. P. Abbott et al., Observation of Gravitational
Waves from a Binary Black Hole Merger, Phys. Rev. Lett. 116 (2016), no. 6 061102,
[arXiv:1602.03837].
[25] LIGO Scientific, Virgo Collaboration, B. P. Abbott et al., GW151226: Observation of
Gravitational Waves from a 22-Solar-Mass Binary Black Hole Coalescence, Phys. Rev. Lett.
116 (2016), no. 24 241103, [arXiv:1606.04855].
[26] LIGO Scientific, VIRGO Collaboration, B. P. Abbott et al., GW170104: Observation of
a 50-Solar-Mass Binary Black Hole Coalescence at Redshift 0.2, Phys. Rev. Lett. 118 (2017),
no. 22 221101, [arXiv:1706.01812]. [Erratum: Phys.Rev.Lett. 121, 129901 (2018)].
[27] LIGO Scientific, Virgo Collaboration, B. . P. . Abbott et al., GW170608: Observation of
a 19-solar-mass Binary Black Hole Coalescence, Astrophys. J. Lett. 851 (2017) L35,
[arXiv:1711.05578].
[28] LIGO Scientific, Virgo Collaboration, B. P. Abbott et al., GW170814: A Three-Detector
Observation of Gravitational Waves from a Binary Black Hole Coalescence, Phys. Rev. Lett.
119 (2017), no. 14 141101, [arXiv:1709.09660].
[29] LIGO Scientific, Virgo Collaboration, B. P. Abbott et al., GWTC-1: A
Gravitational-Wave Transient Catalog of Compact Binary Mergers Observed by LIGO and
Virgo during the First and Second Observing Runs, Phys. Rev. X 9 (2019), no. 3 031040,
[arXiv:1811.12907].
[30] LIGO Scientific, Virgo Collaboration, R. Abbott et al., GWTC-2: Compact Binary
Coalescences Observed by LIGO and Virgo During the First Half of the Third Observing
Run, Phys. Rev. X 11 (2021) 021053, [arXiv:2010.14527].
[31] NANOGrav Collaboration, G. Agazie et al., The NANOGrav 15 yr Data Set: Evidence for
a Gravitational-wave Background, Astrophys. J. Lett. 951 (2023), no. 1 L8,
[arXiv:2306.16213].
[32] EPTA Collaboration, J. Antoniadis et al., The second data release from the European Pulsar
Timing Array - I. The dataset and timing analysis, Astron. Astrophys. 678 (2023) A48,
[arXiv:2306.16224].
[33] LISA Collaboration, P. Amaro-Seoane et al., Laser Interferometer Space Antenna,
arXiv:1702.00786.
[34] D. E. Morrissey and M. J. Ramsey-Musolf, Electroweak baryogenesis, New J. Phys. 14 (2012)
125003, [arXiv:1206.2942].
[35] M. Lewicki, M. Merchand, L. Sagunski, P. Schicho, and D. Schmitt, Impact of theoretical
uncertainties on model parameter reconstruction from GW signals sourced by cosmological
phase transitions, arXiv:2403.03769.

– 34 –
[36] D. Curtin, P. Meade, and C.-T. Yu, Testing Electroweak Baryogenesis with Future Colliders,
JHEP 11 (2014) 127, [arXiv:1409.0005].
[37] A. Beniwal, M. Lewicki, J. D. Wells, M. White, and A. G. Williams, Gravitational wave,
collider and dark matter signals from a scalar singlet electroweak baryogenesis, JHEP 08
(2017) 108, [arXiv:1702.06124].
[38] A. Beniwal, M. Lewicki, M. White, and A. G. Williams, Gravitational waves and electroweak
baryogenesis in a global study of the extended scalar singlet model, JHEP 02 (2019) 183,
[arXiv:1810.02380].
[39] LZ Collaboration, J. Aalbers et al., First Dark Matter Search Results from the LUX-ZEPLIN
(LZ) Experiment, Phys. Rev. Lett. 131 (2023), no. 4 041002, [arXiv:2207.03764].
[40] I. Esteban, M. C. Gonzalez-Garcia, M. Maltoni, T. Schwetz, and A. Zhou, The fate of hints:
updated global analysis of three-flavor neutrino oscillations, JHEP 09 (2020) 178,
[arXiv:2007.14792].
[41] P. Foldenauer and J. Hoefken Zink, How to rule out (g − 2)µ in U (1)Lµ −Lτ with White
Dwarf Cooling, arXiv:2405.00094.
[42] M. Escudero, D. Hooper, G. Krnjaic, and M. Pierre, Cosmology with A Very Light Lµ − Lτ
Gauge Boson, JHEP 03 (2019) 071, [arXiv:1901.02010].
[43] D. Croon, G. Elor, R. K. Leane, and S. D. McDermott, Supernova Muons: New Constraints
on Z’ Bosons, Axions and ALPs, JHEP 01 (2021) 107, [arXiv:2006.13942].
[44] Muon g-2 Collaboration, G. W. Bennett et al., Final Report of the Muon E821 Anomalous
Magnetic Moment Measurement at BNL, Phys. Rev. D 73 (2006) 072003, [hep-ex/0602035].
[45] G. Charpak, F. J. M. Farley, and R. L. Garwin, A New Measurement of the Anomalous
Magnetic Moment of the Muon, Phys. Lett. 1 (1962) 16.
[46] G. Charpak, P. J. M. Farley, E. L. Garwin, T. Muller, J. C. Sens, and A. Zichichi, The
anomalous magnetic moment of the muon, Nuovo Cim. 37 (1965) 1241–1363.
[47] F. Combley and E. Picasso, The Muon (G-2) Precession Experiments: Past, Present and
Future, Phys. Rept. 14 (1974) 1.
[48] CERN-Mainz-Daresbury Collaboration, J. Bailey et al., Final Report on the CERN Muon
Storage Ring Including the Anomalous Magnetic Moment and the Electric Dipole Moment of
the Muon, and a Direct Test of Relativistic Time Dilation, Nucl. Phys. B 150 (1979) 1–75.
[49] S. Borsanyi et al., Leading hadronic contribution to the muon magnetic moment from lattice
QCD, Nature 593 (2021), no. 7857 51–55, [arXiv:2002.12347].
[50] CMD-3 Collaboration, F. V. Ignatov et al., Measurement of the e+ e− → π + π − cross section
from threshold to 1.2 GeV with the CMD-3 detector, arXiv:2302.08834.
[51] M. Davier, A. Hoecker, B. Malaescu, and Z. Zhang, Reevaluation of the hadronic vacuum
polarisation contributions to the Standard Model predictions of the muon g − 2 and α(m2Z )
using newest hadronic cross-section data, Eur. Phys. J. C 77 (2017), no. 12 827,
[arXiv:1706.09436].
[52] A. Keshavarzi, D. Nomura, and T. Teubner, Muon g − 2 and α(MZ2 ): a new data-based
analysis, Phys. Rev. D 97 (2018), no. 11 114025, [arXiv:1802.02995].
[53] G. Colangelo, M. Hoferichter, and P. Stoffer, Two-pion contribution to hadronic vacuum
polarization, JHEP 02 (2019) 006, [arXiv:1810.00007].

– 35 –
[54] M. Davier, A. Hoecker, B. Malaescu, and Z. Zhang, A new evaluation of the hadronic
vacuum polarisation contributions to the muon anomalous magnetic moment and to α(m2Z ),
Eur. Phys. J. C 80 (2020), no. 3 241, [arXiv:1908.00921]. [Erratum: Eur.Phys.J.C 80, 410
(2020)].
[55] MEG II Collaboration, K. Afanaciev et al., A search for µ+ → e+ γ with the first dataset of
the MEG II experiment, Eur. Phys. J. C 84 (2024), no. 3 216, [arXiv:2310.12614].
[56] S. Morisi, Heavy neutral lepton search and µ → eγ constraints in case of type-I seesaw,
arXiv:2403.00983.
[57] DESI Collaboration, A. G. Adame et al., DESI 2024 VI: Cosmological Constraints from the
Measurements of Baryon Acoustic Oscillations, arXiv:2404.03002.
[58] S. Roy Choudhury and S. Choubey, Updated Bounds on Sum of Neutrino Masses in Various
Cosmological Scenarios, JCAP 09 (2018) 017, [arXiv:1806.10832].
[59] L. Dolan and R. Jackiw, Symmetry Behavior at Finite Temperature, Phys. Rev. D 9 (1974)
3320–3341.
[60] S. R. Coleman, The Fate of the False Vacuum. 1. Semiclassical Theory, Phys. Rev. D 15
(1977) 2929–2936. [Erratum: Phys.Rev.D 16, 1248 (1977)].
[61] C. G. Callan, Jr. and S. R. Coleman, The Fate of the False Vacuum. 2. First Quantum
Corrections, Phys. Rev. D 16 (1977) 1762–1768.
[62] S. R. Coleman and F. De Luccia, Gravitational Effects on and of Vacuum Decay, Phys. Rev.
D 21 (1980) 3305.
[63] A. D. Linde, Fate of the False Vacuum at Finite Temperature: Theory and Applications,
Phys. Lett. B 100 (1981) 37–40.
[64] A. D. Linde, Decay of the False Vacuum at Finite Temperature, Nucl. Phys. B 216 (1983)
421. [Erratum: Nucl.Phys.B 223, 544 (1983)].
[65] K. Saikawa and S. Shirai, Primordial gravitational waves, precisely: The role of
thermodynamics in the Standard Model, JCAP 05 (2018) 035, [arXiv:1803.01038].
[66] M. Lewicki, M. Merchand, and M. Zych, Electroweak bubble wall expansion: gravitational
waves and baryogenesis in Standard Model-like thermal plasma, JHEP 02 (2022) 017,
[arXiv:2111.02393].
[67] J. Ellis, M. Lewicki, M. Merchand, J. M. No, and M. Zych, The scalar singlet extension of
the Standard Model: gravitational waves versus baryogenesis, JHEP 01 (2023) 093,
[arXiv:2210.16305].
[68] M. Hindmarsh, S. J. Huber, K. Rummukainen, and D. J. Weir, Gravitational waves from the
sound of a first order phase transition, Phys. Rev. Lett. 112 (2014) 041301,
[arXiv:1304.2433].
[69] M. Hindmarsh, S. J. Huber, K. Rummukainen, and D. J. Weir, Numerical simulations of
acoustically generated gravitational waves at a first order phase transition, Phys. Rev. D 92
(2015), no. 12 123009, [arXiv:1504.03291].
[70] M. Hindmarsh, S. J. Huber, K. Rummukainen, and D. J. Weir, Shape of the acoustic
gravitational wave power spectrum from a first order phase transition, Phys. Rev. D 96
(2017), no. 10 103520, [arXiv:1704.05871]. [Erratum: Phys.Rev.D 101, 089902 (2020)].

– 36 –
[71] D. Cutting, M. Hindmarsh, and D. J. Weir, Vorticity, kinetic energy, and suppressed
gravitational wave production in strong first order phase transitions, Phys. Rev. Lett. 125
(2020), no. 2 021302, [arXiv:1906.00480].
[72] M. Hindmarsh and M. Hijazi, Gravitational waves from first order cosmological phase
transitions in the Sound Shell Model, JCAP 12 (2019) 062, [arXiv:1909.10040].
[73] M. Hindmarsh, Sound shell model for acoustic gravitational wave production at a first-order
phase transition in the early Universe, Phys. Rev. Lett. 120 (2018), no. 7 071301,
[arXiv:1608.04735].
[74] LISA Cosmology Working Group Collaboration, C. Caprini, R. Jinno, M. Lewicki,
E. Madge, M. Merchand, G. Nardini, M. Pieroni, A. Roper Pol, and V. Vaskonen,
Gravitational waves from first-order phase transitions in LISA: reconstruction pipeline and
physics interpretation, arXiv:2403.03723.
[75] J. R. Espinosa, T. Konstandin, J. M. No, and G. Servant, Energy Budget of Cosmological
First-order Phase Transitions, JCAP 06 (2010) 028, [arXiv:1004.4187].
[76] AEDGE Collaboration, Y. A. El-Neaj et al., AEDGE: Atomic Experiment for Dark Matter
and Gravity Exploration in Space, EPJ Quant. Technol. 7 (2020) 6, [arXiv:1908.00802].
[77] A. Alloul, N. D. Christensen, C. Degrande, C. Duhr, and B. Fuks, FeynRules 2.0 - A
complete toolbox for tree-level phenomenology, Comput. Phys. Commun. 185 (2014)
2250–2300, [arXiv:1310.1921].
[78] A. Belyaev, N. D. Christensen, and A. Pukhov, CalcHEP 3.4 for collider physics within and
beyond the Standard Model, Comput. Phys. Commun. 184 (2013) 1729–1769,
[arXiv:1207.6082].
[79] G. Belanger, F. Boudjema, A. Pukhov, and A. Semenov, MicrOMEGAs: A Program for
calculating the relic density in the MSSM, Comput. Phys. Commun. 149 (2002) 103–120,
[hep-ph/0112278].
[80] P. Gondolo and G. Gelmini, Cosmic abundances of stable particles: Improved analysis, Nucl.
Phys. B 360 (1991) 145–179.
[81] Planck Collaboration, P. A. R. Ade et al., Planck 2015 results. XIII. Cosmological
parameters, Astron. Astrophys. 594 (2016) A13, [arXiv:1502.01589].
[82] Planck Collaboration, N. Aghanim et al., Planck 2018 results. VI. Cosmological parameters,
Astron. Astrophys. 641 (2020) A6, [arXiv:1807.06209]. [Erratum: Astron.Astrophys. 652,
C4 (2021)].
[83] P. Junnarkar and A. Walker-Loud, Scalar strange content of the nucleon from lattice QCD,
Phys. Rev. D 87 (2013) 114510, [arXiv:1301.1114].
[84] P. Bechtle, S. Heinemeyer, O. Stρal, T. Stefaniak, and G. Weiglein, HiggsSignals:
Confronting arbitrary Higgs sectors with measurements at the Tevatron and the LHC, Eur.
Phys. J. C 74 (2014), no. 2 2711, [arXiv:1305.1933].
[85] P. Bechtle, S. Heinemeyer, O. Stal, T. Stefaniak, and G. Weiglein, Applying Exclusion
Likelihoods from LHC Searches to Extended Higgs Sectors, Eur. Phys. J. C 75 (2015), no. 9
421, [arXiv:1507.06706].
[86] W. Altmannshofer, S. Gori, M. Pospelov, and I. Yavin, Neutrino Trident Production: A
Powerful Probe of New Physics with Neutrino Beams, Phys. Rev. Lett. 113 (2014) 091801,
[arXiv:1406.2332].

– 37 –
[87] B. Ahlgren, T. Ohlsson, and S. Zhou, Comment on “Is Dark Matter with Long-Range
Interactions a Solution to All Small-Scale Problems of Λ Cold Dark Matter Cosmology?”,
Phys. Rev. Lett. 111 (2013), no. 19 199001, [arXiv:1309.0991].
[88] Muon Collider Collaboration, J. de Blas et al., The physics case of a 3 TeV muon collider
stage, arXiv:2203.07261.
[89] S. N. Gninenko, N. V. Krasnikov, and V. A. Matveev, Muon g-2 and searches for a new
leptophobic sub-GeV dark boson in a missing-energy experiment at CERN, Phys. Rev. D 91
(2015) 095015, [arXiv:1412.1400].
[90] Y. Kahn, G. Krnjaic, N. Tran, and A. Whitbeck, M3 : a new muon missing momentum
experiment to probe (g−2)µ and dark matter at Fermilab, JHEP 09 (2018) 153,
[arXiv:1804.03144].
[91] ATLAS Collaboration, G. Aad et al., Combination of searches for invisible decays of the
Higgs boson using 139 fb−1 of proton-proton collision data at s=13 TeV collected with the
ATLAS experiment, Phys. Lett. B 842 (2023) 137963, [arXiv:2301.10731].
[92] CMS Collaboration, A. M. Sirunyan et al., Measurements of the Higgs boson width and
anomalous HV V couplings from on-shell and off-shell production in the four-lepton final
state, Phys. Rev. D 99 (2019), no. 11 112003, [arXiv:1901.00174].
[93] Particle Data Group Collaboration, M. Tanabashi et al., Review of Particle Physics,
Phys. Rev. D 98 (2018), no. 3 030001.
[94] R. R. Parwani, Resummation in a hot scalar field theory, Phys. Rev. D 45 (1992) 4695,
[hep-ph/9204216]. [Erratum: Phys.Rev.D 48, 5965 (1993)].
[95] D. Curtin, P. Meade, and H. Ramani, Thermal Resummation and Phase Transitions, Eur.
Phys. J. C 78 (2018), no. 9 787, [arXiv:1612.00466].
[96] A. Katz and M. Perelstein, Higgs Couplings and Electroweak Phase Transition, JHEP 07
(2014) 108, [arXiv:1401.1827].

– 38 –