Arxiv:1909.09740V2 [Hep-Ph] 2 May 2020 Neutrinos

Gravitational footprints of massive neutrinos and lepton number breaking

Andrea Addazi,1, 2, ∗ Antonino Marcianò,1, † António P. Morais,3, ‡ Roman Pasechnik,4, § Rahul Srivastava,5, ¶ and José W. F. Valle5, ∗∗

1Department of Physics & Center for Field Theory and Particle Physics, Fudan University, 200433 Shanghai, China 2College of Physics, Sichuan University, Chengdu, 610065, China 3Departamento de Física, Universidade de Aveiro and CIDMA, Campus de Santiago, 3810-183 Aveiro, Portugal, EU 4Department of Astronomy and Theoretical Physics, Lund University, 221 00 Lund, Sweden, EU 5AHEP Group, Institut de Física Corpuscular – C.S.I.C./Universitat de València, Parc Cientíﬁc de Paterna. C/ Catedrático José Beltrán, 2 E-46980 Paterna (Valencia) - SPAIN

We investigate the production of primordial Gravitational Waves (GWs) arising from First Order Phase Transitions (FOPTs) associated to neutrino mass generation in the context of type-I and inverse seesaw schemes. We examine both “high-scale” as well as “low-scale” variants, with either explicit or spontaneously broken lepton number symmetry U(1)L in the neutrino sector. In the latter case, a pseudo-Goldstone majoron-like boson may provide a candidate for cosmological dark matter. We find that schemes with softly-broken U(1)L and with single Higgs-doublet scalar sector lead to either no FOPTs or too weak FOPTs, precluding the detectability of GWs in present or near future measurements. Nevertheless, we found that, in the majoron-like seesaw scheme with spontaneously broken U(1)L at finite temperatures, one can have strong FOPTs and non-trivial primordial GW spectra which can fall well within the frequency and amplitude sensitivity of upcoming experiments, including LISA, BBO and u-DECIGO. However, GWs observability clashes with invisible Higgs decay constraints from the LHC. A simple and consistent fix is to assume the majoron-like mass to lie above the Higgs-decay kinematical threshold. We also found that the majoron-like variant of the low-scale seesaw mechanism implies a different GW spectrum than the one expected in the high-scale seesaw. This feature will be testable in future experiments. Our analysis shows that GWs can provide a new and complementary portal to test the neutrino mass generation mechanism.

I. INTRODUCTION and neutrinoless double beta decay (0ν2β) searches, neutrino masses can be probed through Charged Lepton Fla- vor Violation (CLFV) experiments at the high intensity Non-zero neutrino masses constitute one of the most ro- and/or high energy frontier [21–23]. Moreover, neutrino bust evidences for new physics [1–3]. Despite great efforts mass generation can leave signatures at high-energy col- over the last two decades to underpin the origin of neu- liders like the Large Hadron Collider (LHC) [24–27]. trino mass, the basic underlying mechanism remains as elusive as ever. Small neutrino masses can be generated The detection of Gravitational Waves (GWs) by the in many ways, both for Majorana [4,5] and Dirac [6–11] LIGO team has opened an entirely novel method to probe arXiv:1909.09740v2 [hep-ph] 2 May 2020 neutrinos. Here, we focus on the various variants of the the underlying new physics associated to neutrino mass popular type-I seesaw mechanism for Majorana neutri- generation. It was advocated that the spectrum of pri- nos [12–15]. We consider both high- and low-scale [16–20] mordial GWs, potentially measurable at the currently realizations, with explicit or spontaneous lepton number planned GW interferometers, may represent an impor- violation, in which SU(3)c SU(2)L U(1)Y singlet neu- tant cutting-edge probe for new physics. This follows ⊗ ⊗ trinos act as neutrino mass mediators. Besides oscillation from the fact that these interferometers can be sensitive enough to measure the echoes of the possible First Order Phase Transitions (FOPTs), which might have happened in the past cosmological history [28]. ∗ [email protected] † [email protected] In this letter, we focus on possible gravitational foot- ‡ [email protected] § [email protected] prints of the various variants of the popular type-I and ¶ rahulsri@ific.uv.es inverse seesaw mechanisms for Majorana neutrinos. The ∗∗ valle@ific.uv.es relevant part of the minimal type-I seesaw Lagrangian is 2 given by majoron variants of minimal type-I and inverse seesaw Type−I c c c one should replace = Yν LHν¯ + Mν ν + h.c. (1) LYuk T Here, L = (ν, l) are the SM lepton doublets, H is the M Yσ vσ/√2 , µ Yσ vσ/√2 (5) → → SM Higgs doublet, νc are the three SM singlet “right- handed” neutrinos. The 3 3 matrices Yν and M are the in Eq. (17) and (3), respectively. An additional attractive × Yukawa coupling and the νc mass matrix, respectively. feature of majoron models is the existence of a pseudo Due to the Pauli principle the latter is symmetric. No- Nambu-Goldstone boson commonly dubbed as majoron. tice that, for brevity, we omit family indices through- The latter carries odd charge under Z2 thus providing out this letter. Notice also that the mass term explic- a good [39–41], and testable [42, 43] dark matter candidate. In the standard majoron seesaw schemes, the itly breaks the lepton number symmetry U(1)L to its Z2 subgroup. The electroweak (EW) symmetry is broken by majoron mass is considered to be small, of order keV, for the vacuum expectation value (vev) of the Higgs field, i.e. it to be a suitable warm dark matter candidate. How- H = vh/√2, generating the light neutrino masses ever, the current stringent constraints on invisible Higgs h i 2 boson decay modes [26, 27, 44, 45] put severe limita- Type−I vh T −1 mν = Yν M Yν . (2) tion on how large h σσ coupling could be. On the 2 other hand, the strength→ of the cosmological phase tran- The lightness of the left-handed neutrinos is then as- sition is expected to be strongly correlated with the size cribed to the heaviness of the “right-handed” isosinglet of the Higgs-majoron quartic coupling. It remains an partners e.g. for Y (1),M (1014) GeV, one gets ∼ O ∼ O open question whether it is possible to reconcile the cur- mν (0.1) eV. ∼ O rent LHC bounds on the invisible Higgs decays in the in- Another popular realization of this idea is the “low-scale” verse seesaw scenario featuring a keV-scale majoron dark variant, in which two gauge singlet fermions νc and S are matter with the existence of strong EW FOPTs yielding added sequentially to the SM particle content [16–20]. potentially observable GWs signals. The template of these schemes has exact conservation of lepton number and, as a result, strictly massless neutri- In this work, we consider the case where the lepton num- nos. Yet flavor is violated to a potentially large degree, ber symmetry is broken also explicitly, but softly. This subjected only to constraints from weak interaction pre- way the majoron can pick up a mass. If this is larger cision observables, such as universality tests [29–33]. To than a half of the Standard Model (SM) Higgs boson this template one adds a small seed of lepton number vi- mass, mh/2 62.5 GeV, the invisible Higgs decays will ' olation, leading to nonzero neutrino mass. One example be kinematically forbidden, avoiding the stringent con- is the so-called “inverse seesaw” mechanism, where the straints on the Higgs-majoron quartic coupling. In such smallness of the neutrino mass is linked to the breaking scenario the heavy majoron-like state can provide a vi- able candidate for cold dark matter (CDM). The same of the lepton number symmetry U(1)L to its Z2 subgroup, through the so called µ-term. The relevant part type of scalar CDM scenarios with exact Z2 parity have of Lagrangian in this case is given by been broadly studied in singlet extensions of the SM [46– 56], which share very similar properties with the scalar Inverse c c = Yν LHν¯ + Mν S + µSS + h.c., (3) LYuk sector of the model under consideration. where µ is also a 3 3 symmetric matrix. The light neutrino mass is then× given by 2 Inverse vh T T −1 −1 m = Y M µM Yν . (4) ν 2 ν II. SCALAR SECTOR Note that small neutrino masses are “protected”, since mν 0 as the lepton number symmetry gets restored by → having µ 0 [16–20]. In this case there can be sizable The scalar sector of the majoron inverse seesaw model unitarity violation→ in neutrino propagation [34–36]. has been extensively studied in the literature, including the perturbative unitarity and stability of the scalar po- For both high- and low-scale seesaw, one can have spon- tential, as well as the electroweak precision tests and the taneous breaking of , leading to the so-called U(1)L Z2 bounds on Higgs-majoron couplings [27, 57]. The scalar majoron variants of the seesaw→ [17, 37, 38]. This is ac- potential is written as follows complished by adding the SM singlet scalar σ, which carries two units of lepton number charge. Then σ vσ 2 † † 2 2 † † 2 h i ≡ 0(Φ, σ) = µΦΦ Φ + λΦ(Φ Φ) + µσσ σ + λσ(σ σ) spontaneously breaks U(1)L 2, leading to a dynami- V Z † † 1 2 2 → +λΦσΦ Φσ σ + µ σ + h.c. , (6) cal explanation of the small neutrino masses. To get the 2 b 3 with Φ and σ given by model parameters. In the numerical analysis of GW spectra below we do not enforce the majoron mass mA to be 0 1 G + iG 1 small (e.g. at a keV scale) and treat it as a free parameter Φ = , σ = (φσ + σR + iσI ) , √2 φh + h + iη √2 instead. (7)

0 where h, η, G, G , σR, σI are real scalars. These latter III. GRAVITATIONAL WAVES FROM FOPTS fields represent quantum fluctuations about the classical mean-fields φα = φh, φσ . In the zero-temperature { } limit, the mean-fields approach the corresponding vevs, In order to characterize the features of the GWs origi- i.e. φh,σ(T = 0) vh,σ, where vh = 246 GeV is nating from FOPTs in seesaw schemes, we calculate the ≡ the SM Higgs vev. Besides, one of the physical CP- strenght of the phase transition α at the bubble nucle- even scalar states, with a small or no mixing with σR, ation temperature Tn typically defined through the trace is identified with the SM-like Higgs boson with mass, anomaly as [58, 59] mh1 mh = 125 GeV. The last (soft) term appearing in ≡ 1 h Tn ∂Vi ∂Vf i Eq. (6) implements the explicit breakdown U(1)L Z2, → α = Vi Vf , (12) and hence provides a pseudo-Goldstone mass to the imag- ργ − − 4 ∂T − ∂T inary part of the field σ known as majoron 1. I with

In what follows, we discuss further implications of the 2 π 4 majoron seesaw scenario in both versions of the scalar ργ = g∗ T (13) 30 n sector, with explicit (vanishing vσ) and with sponta- 2 neous (vσ = 0) lepton-number U(1)L Z2 symmetry being the energy density of the radiation medium at the 6 → breaking, for physics of cosmological EW FOPTs and ex- bubble nucleation epoch found in terms of the number amine the associated GWs spectra. In the first version, of relativistic d.o.f.’s. g∗ 106.75 [60–63]. Above, Vi ' no mixing occurs so that and Vf are the values of the effective potential in the 2 symmetric and broken phases just before and after the 2 2 2 2 2 λσhvh m = 2λhv , m = µ + µ + , (8) h1 h h2 σ b 2 transition takes place, respectively. Another key quantity 2 to calculate the GW spectrum is the inverse time-scale 2 2 2 λσhvh mA = µσ µb + , (9) of the phase transition, which, in units of the Hubble − 2 β parameter H, reads as for the SM Higgs boson, CP-even and CP-odd (majoron) scalars, respectively. In the second version, the physical ! β ∂ Sˆ3 CP-even states acquire masses, = Tn , (14) H ∂T T 2 2 Tn 2 2 2 λσvσ λhvh m = λhv + λσv − , (10) h1,h2 h σ ∓ cos 2θ where Sˆ3 is the Euclidean action. In this work, we do in terms of h-σR mixing angle θ, while the majoron gets not consider the case of runaway bubbles and use the a pseudo-Goldstone mass, formalism outlined in Ref. [63] to calculate the spectrum m2 m2 = 2µ2 , µ2 < 0 . (11) of primordial GWs. A ≡ σI − b b While in the latter case, a very light majoron (compared For the case of non-runaway nucleated bubbles, the inten- to the EW scale vh) would imply setting an equally small sity of the GW radiation grows with the ratio ∆vn/Tn, µb parameter, in the former case, Eq. (9), this limit re- where lies on a very strong fine tuning between µσ, µb and λσh ∆vφ = vf vi , φ = h, σ (15) n | φ − φ| defines the difference between the VEVs of the initial 1 One could also add other other explicit breaking terms such as (metastable) and final (stable) phases at the bubble nu- † σΦ Φ but here we stick to the simplest possibility of mass terms. cleation temperature Tn. The quantity ∆vn/Tn also of- 2 More precisely, here we refer to a mixture of explicit and spon- fers a measure of the strength of the phase transition taneous U(1)L breakings in the scalar sector of the model since alternative to α. However, while the latter is of com- the soft scalar mass term µb 6= 0 is always present in the scalar potential (6) to provide a non-zero mass to the physical majoron mon use in the context of GWs, the former is typically state. This should not be confused with the neutrino sector where referred to in the context of electroweak baryogenesis. In U(1)L is always spontaneously broken by a small or large vσ vev this work, we will consider both quantities on the same in the considered majoron versions of the seesaw schemes. footing. 4

From the discussion in Ref. [59, 64], it follows that bubble sizable FOPTs at the quantum level since the fermion wall collisions do not provide an efficient way of produc- thermal loop contributions generate highly suppressed ing GWs in the models of interest to us here. As a result, terms to the effective potential in the high-temperature GWs originate mainly from two sources: expansion. Indeed, we find relatively weak FOPTs for many points in parameter space in this case and the cor- I. Magnetohydrodynamic (MHD) turbulence; 2 peak responding GW “intensity” parameter h ΩGW lies far be- II. Sound shock waves (SW) of the early Universe low the sensitivity of any conceivable experiment3. plasma, generated by the bubble’s violent expansion. In the majoron inverse seesaw scenarios, i.e. when the scalar sector of the SM is extended by incorporating an These contributions arise over transient times in the early additional complex SM-singlet scalar state, the situation Universe and get subsequently “redshifted” by the ex- changes dramatically. Indeed, it is well known that the pansion. To a present observer this appears as a cosmic presence of additional SM scalar singlets significantly en- gravitational stochastic background. Intuitively, one ex- hances the strength and multiplicity of the FOPTs, and pects that from any of these leading order contributions, in some cases leads to potentially detectable GW spectra a high wall velocity is necessary to generate detectable (see e.g. Refs. [68–72]). The presence of an additional GWs. In our numerical analysis, performed with the help classical field φσ = 0 coupled substantially to the Higgs of the CosmoTransitions package [65], we have consid- boson strongly affects6 the shape of the effective potential ered supersonic detonations such that the bubble wall at non-zero temperature T allowing for a richer pattern of velocity maximizes the GW peak amplitude and is above EW FOPTs. Note, in the considered majoron seesaw this the Chapman-Jouguet velocity defined as happens in both variants with explicit and spontaneous q U(1)L breaking at T = 0 discussed above in Sec.II. 1 2 2 vJ = cs + α + α , (16) 1 + α 3 The values of the coupling constants used in our numerical analysis satisfy the conservative bounds provided by with √1 being the speed of sound. Besides, in cs = 3 tree-level perturbativity, λi < 4π and Yi < √4π, for our results the SW contribution dominates the peak fre- the quartic and Yukawa couplings,| | respectively| | 4. Since quency and the peak amplitude. Furthermore, the state the tree-level potential also receives both quantum and of the art expressions derived in Ref. [63] do not account finite temperature corrections, we only considered values for MHD-turbulence effects due to large theoretical uncertainties. Therefore, in the remainder of this work, 9 we will not take such effects into consideration. Note, 10− LISA for certain parameter configurations one also expects se- 0.1325 12 quential phase transition patterns leading to multi-peak 10− BBO GWs spectra studied for the first time in Refs. [66, 67]. 0.1300 peak GW

15 DECIGO-corr σh

Ω 10− λ 2

h 0.1275 u-DECIGO A. Seesaw-induced GWs spectra 10 18 − 0.1250

u-DECIGO-corr 21 To investigate the eventual occurrence of phase transi- 10− 6 4 2 0 10− 10− 10− 10 tions, the standard way is to incorporate into the ef- fpeak [Hz] fective potential the tree-level zero temperature compo- nents, the Coleman-Weinberg corrections, the full one- FIG. 1: The GW spectrum as a function of λσh for the case loop finite-temperature corrections, as well as the Daisy of spontaneously broken U(1)L symmetry. No solutions consistent re-summation. It is worth noticing that within the type-I with the LHC bound on invisibly-decaying Higgs were found. Other model parameters are fixed as: m = 1 keV, m = 591 GeV, seesaw mechanism with explicitly broken lepton number, A h2 vσ = 858 GeV, Yσ,1 = 1.20, and Yσ,2 = 1.66. no FOPTs are obtained. The heavy isosinglet neutrinos practically decouple at the EW scale, and do not alter the nature of the EW phase transition. In contrast, in the non-majoron inverse seesaw mechanism (i.e. without 3 We emphasize that variations of the Yukawa coupling Yν in the adding a SM singlet scalar) the singlet neutrinos lie closer range 1 ÷ 10 cannot allow the detectability of the GW signal in to the EW scale, and can have a sizable coupling to the the type-I seesaw scenario. Higgs boson. However, even by adding a large number of 4 For a more involved analysis of the perturbativity constraints in singlet neutrino species it is impossible to generate any a scalar extension of the SM, see e.g. Ref. [73] 5 within the ranges λi < 5 and Yi < 3.5, even more U(1)L breaking scenario at zero temperature, with van- | | | | conservative bounds in a zero-temperature theory. ishing vσ providing a tiny value of µ in the inverse seesaw mechanism according to Eq. (5). We find a large num- Due to the current LHC constraints on invisible Higgs ber of points with strong FOPTs that generate the GW decays [26, 27, 44, 45] one has a bound λ < 0.03 the σh peak-amplitudes well within the projected LISA sensitiv- Higgs-majoron quartic coupling in the case of∼ light keV- ity, with typical values of Y below unity. scale majoron. Under this assumption in our numerical σ scan we did not find any point with an EW FOPT that 3 is strong enough for a potential observability of the re- LISA 10 sulting GW spectra. This is illustrated in Fig1. 10− 2 There is a strong correlation of the peak-amplitude with BBO 14 peak GW λσh value such that requiring the latter to be very small 10− σ Y Ω DECIGO-corr 2 makes the GW signals well below the reach of LISA or h u-DECIGO 18 1 even the planned BBO and DECIGO missions. Note, this 10− is the case for both considered versions of the majoron

u-DECIGO-corr inverse seesaw model, with explicit and spontaneous lep- 22 10− 6 4 2 0 ton number symmetry breaking in the scalar sector at 10− 10− 10− 10 T = 0. Therefore, we conclude that the standard keV- fpeak [Hz] scale (warm) majoron dark matter scenario associated with the inverse seesaw mechanism cannot be probed by FIG. 2: The GW spectrum as a function of the Yukawa Yσ cou- the GW astrophysics in the current simplest formulation. pling in the case of softly-broken U(1)L symmetry (i.e. vσ = 0). Order one variation of Yσ correspond to several order of magnitude For this reason, from now on we only consider the case variations in the GW power spectrum. Other model parameters are of heavy majoron kinematically closing the mA > mh/2 fixed as λσ = 0.37, λσh = 1.07, M = 239.4 GeV, mh2 = 154.6 GeV invisible Higgs decay channel and thus enabling us to and mA =369.9 GeV. consider larger values of λσh that ensure the existence of the strong FOPTs in the model under consideration. As is typical in models with several scalars, due to the presence of two classical fields φh, φσ in the effective potential at finite temperatures,{ besides} a plenty of single-step FOPTs one also finds double-step and, B. Inverse seesaw with majoron: small vσ case in some rare cases, even triple-step FOPTs for a given parameter space point and at well-separated nucleation Let us now consider the case of a genuine inverse seesaw temperatures. One could naturally expect the presence with majoron and very small singlet VEV vσ, effectively of several peaks in the corresponding GW spectrum as- generating an equally small µSS term in the Lagrangian sociated with each FOPTs in such a chain of transitions. (3). As discussed above, this mechanism offers a dynami- Notice also that the analyses of GW spectra for the cal explanation for light neutrino masses, whose scale can multi-step FOPT scenarios involving EW phase transi- be attributed to the (tiny) scale of spontaneous breaking tions require particular care, as discussed for instance in of the lepton-number U(1)L symmetry in the neutrino Ref. [78]. sector (while being softly-broken in the scalar sector). Multi-peak configurations in the GW spectrum occur In order to understand the role of heavy neutrino in the very frequently in the inverse seesaw with majoron. This generation of the GW spectra in the majoron inverse see- fact is further highlighted in Fig.3. Indeed, the double- saw scenario, we study the sensitivity of the GW peak- peak feature of the GW spectrum is a generic prediction 2 peak amplitude h ΩGW originated by the EW FOPTs with of our model, that can arise for many parameter choices, respect to the variation of the Yukawa coupling Yσ. As as shown in Fig. 3a. Configurations with larger peak shown in Fig.2 an order one variation in the Yukawa cou- multiplicities are also possible, as seen in Fig. 3b, where pling reflects into violent variations by several orders of the color denotes the peak number, 1 (blue), 2 (cyan) magnitude in the GWs spectrum, with all the other pa- or 3 (red). Such a rich structure of the GW spectrum rameters fixed. The results are shown together with the is favoured for relatively large quartic couplings involv- projected sensitivities expected in LISA, and the planned ing σ. From Fig. 3b we also see that the GW spectra u-DECIGO and BBO missions [62, 63, 74, 75]. We have with three peaks are rarer than single or double GW- taken the u-DECIGO sensitivity curves from Ref. [76], peak spectra. We also note that a significant fraction whereas the sensitivities of other experiments are taken of the single peak cases are potentially testable at LISA from Ref. [77]. Here, for simplicity we have used the soft and BBO. However, finding a well-resolved double- or 6

double-peak signatures, these scenarios may suﬀer from 13 10− LISA larger underlined theoretical uncertainties so we decided BBO not to discuss them here. 16 DECIGO-corr 10− u-DECIGO 7 10− peak GW

Ω 19 u-DECIGO-corr 9 2 10− 10− h 11 10− 22 BBO 10− 13 LISA 10− DECIGO-corr

GW 15

25 Ω 10− 10− 3 2 1 0 1 2 10− 10− 10− 10 10 h 17 10− fpeak [Hz] u-DECIGO 19 10− (a) Selected double-peak scenarios within the LISA and BBO (T , ∆vh, ∆vσ) = (122.0, 23.0, 0.0), (92.0, 23.0, 0.0) GeV 21 n n n sensitivity ranges. The two ends of each line represent the h σ u-DECIGO-corr 10− (Tn, ∆vn, ∆vn) = (77.0, 235.0, 114.0), (183.0, 235.0, 114.0) GeV h σ location of the peaks of the double-peak GW spectrum. The (Tn, ∆vn, ∆vn) = (293.0, 0.0, 2941.0) GeV 23 two maxima in each double-peak GW spectra are joined by a 10− 5 4 3 2 1 0 10− 10− 10− 10− 10− 10 straight line, in order to easily identify the peaks associated f [Hz] with each other. FIG. 4: Inverse-seesaw-with-majoron benchmark GW spectra in 9 LISA 10− peak 1 the scenario with soft U(1)L symmetry breaking (i.e. with vanish- ingly small vσ(T = 0) → 0). The green curve represents the case BBO peak 2 with single-step FOPT, while the blue and red curves correspond 16 DECIGO-corr peak 3 10− u-DECIGO to double-step FOPTs whose characteristics are given in TablesI andII. For double-step transitions, the (Tn, vn) pairs in each peak u-DECIGO-corr peak GW h f i 23 are ordered from low to high frequencies. Here, ∆vn = |v − v | Ω h h

2 10− σ f i

h and ∆vn = |vσ − vσ| at a given nucleation temperature T = Tn.

30 10− i i f f Peak Id Tn vh, vσ → vh, vσ α β/H 37 Green 1 293 (0, 0) → (0, 2941) 0.5 4.9 10− 7 4 1 2 5 10− 10− 10− 10 10 Red 1 183 (0, 114) → (235, 0) 7.7 × 10−4 7.2 × 104 fpeak [Hz] Red 2 77 (0, 114) → (235, 0) 0.1 231 Blue 1 122 (193, 0) → (216, 0) 1.4 × 10−2 3.1 × 103 (b) Scatter plot showing the number of peaks for given model −3 4 parameter choices. Notice the appearance of double- and even Blue 2 92 (193, 0) → (216, 0) 9.4 × 10 3.0 × 10 triple-peak features. TABLE I: Phase transition parameters for the three curves in Fig.4. In “peak Id” column, the numbering of multi-step scenarios FIG. 3: The multi-peak feature arising from diﬀerent phase tran- is ordered from high to low nucleation temperature Tn, given in sitions in the cosmological history of the Universe is very generic i units of GeV. The vevs before the transition (vh,σ) and after the in the inverse seesaw with majoron. f transition (vh,σ) are given in units of GeV. triple-peak feature where more than a single peak could Curve mh2 mA λσh λσ M Yσ be observable in a measurement appears to be very chal- Green 236 708 1.7 5 × 10−3 380 2 lenging. No single point with such a feature has been Red 192 970 2.3 1.5 93 0.1 found by our scans. As seen in Fig. 3a in most cases both Blue 325 169 4 2.7 158 0.1 peak-amplitudes occur below the projected sensitivities of any planned measurements rendering the observability TABLE II: Model parameters for the three curves in Fig.4. Mass of the corresponding scenarios very remote. In a subset parameters are given in units of GeV. of cases, only the largest peak has been found in the sensitivity ranges of the planned BBO and u-DECIGO mis- In order to understand the characteristic features of the sions, while the second peak typically lies outside of the GW spectra in the majoron inverse seesaw, in Fig.4 we reach of future detectors. This is a direct consequence of show the GW energy density spectrum obtained for dis- the not-so-large quartic couplings that are restricted to tinct nucleation temperatures Tn. Here, we have depicted be smaller than ﬁve. While larger values of the quartic three benchmark scenarios (for the case of vanishing vσ) couplings could generate well-separate and measurable originating from the single-step FOPT (green line) as well 7

7 10− as double-step FOPTs i.e. with two consecutive strong h σ (Tn, ∆vn, ∆vn) = (53.0, 119.0, 786.0), (134.0, 33.0, 3.0) GeV h σ 9 (Tn, ∆vn, ∆vn) = (82.0, 166.0, 693.0), (149.0, 0.0, 353.0) GeV phase transitions (red and blue lines). The correspond- 10− h σ (Tn, ∆vn, ∆vn) = (53.0, 115.0, 786.0), (136.0, 32.0, 2.0) GeV ing values for the model parameters are given in Tab.I 10 11 and Tab.II. − BBO 13 LISA 10− DECIGO-corr The green curve represents a single-peak scenario with a GW 15

Ω 10− 2

single very strong U(1)L phase transition, ∆vσ(Tn)/Tn h ' 17 10− 10, and softly-broken lepton symmetry. It features a very u-DECIGO

19 strong FOPT but not EW one. In fact, this peak is 10− a probe of U(1)L breaking at ﬁnite temperatures while 21 10 u-DECIGO-corr the associated EWPT is very weak or even of second − 23 10− 5 4 3 2 1 0 order. This is possible due to relatively large values of a 10− 10− 10− 10− 10− 10 f [Hz] quartic (portal) coupling, λσh 2, and majoron-neutrino ' Yukawa coupling, Yσ 2, which make the m/T -ratio ' (a) The expected GW spectra. sizable. Hence, the cubic (m/T )3 terms in the thermal expansion can produce a potential barrier between both LISA 12 vacua, inducing this type of transitions. 10− BBO

On another hand, the other two benchmark points, which 16 DECIGO-corr 10− u-DECIGO

have the double-peak feature, have strong EWPT for peak GW

Ω u-DECIGO-corr both peaks in the red curve and for the smaller, almost 2 20

h 10− invisible, peak of the blue curve. Besides this, the blue 24 curve has no U(1)L breaking in any of the minima of the 10− eﬀective potential at the corresponding nucleation tem- 28 peratures whereas the red curve exhibits a strong U(1)L 10− 4 2 0 2 10− 10− 10 10 breaking for the higher peak and a weak U(1)L breaking fpeak [Hz] for the hidden peak. Similarly to the single-peak scenario, the observable peaks in the red and blue curves (b) Scatter plot showing typical double-peak scenarios. are generated due to large λσh and λσ couplings. Note, the single-peak case (green line) lies well within the LISA FIG. 5: Gravitational footprints of “fake” low-scale seesaw with range [62, 63, 74], while only the largest peaks in the majoron. In both plots we take vσ ∼ O(100) GeV – O(1) TeV. double-step FOPT scenarios are within the planned sensitivity range of the BBO (red line) and u-DECIGO-corr C. Type-I seesaw with majoron: large v (blue line) measurements [75], respectively. σ

i i f f Peak Id Tn vh, vσ → vh, vσ α β/H To further illustrate the importance of future GW mea- Green 1 136 (0, 921) → (32, 919) 9.4 × 10−5 1.2 × 106 surements for probing neutrino mass generation mecha- Green 2 53 (245, 786) → (360, 0) 0.5 378 nisms let us consider a type-I seesaw with majoron where Red 1 134 (0, 922) → (33, 919) 10−4 1.1 × 106 the heavy neutrino mass scale M is generated via a large Red 2 53 (245, 786) → (364, 0) 0.5 612 majoron vev vσ spontaneously breaking the lepton num- Blue 1 149 (0, 353) → (0, 0) 1.5 × 10−3 1.0 × 105 ber symmetry. Blue 2 82 (205, 693) → (40, 0) 0.03 4.9 × 103 As mentioned above, it is clear that, if Yν (1), then 14 ∼ O 14 TABLE III: FOPT parameters for the three curves in Fig. 5a. M = Yσvσ/√2 (10 ) GeV, hence vσ (10 ) GeV i f ∼ O ∼ O The vevs before (vh,σ) and after (vh,σ) the phase transition are for Yσ (1). In this limit, all the new particles can be given in units of GeV. integrated∼ O out for processes occurring at the EW scale5, leading to no-FOPT solutions. −6 Curve m m λ λ λ θ v Y Y However, one can take Yν (10 ), corresponding to h2 A h σh σ σ σ,1 σ,2 ∼ O Green 203 188 0.14 0.03 0.03 0.26 790 0.07 1.58 M = Yσvσ/√2 (100) GeV. The majoron and neu- ∼ O Red 206 188 0.14 0.03 0.03 0.26 790 0.08 1.59 trino ﬁelds do not decouple at the EW scale in this case, Blue 205 188 0.14 −0.02 0.03 −0.18 790 0.08 1.59

TABLE IV: Model parameters for the three curves in Fig. 5a. 5 Masses and vσ are given in units of GeV. All the physical majoron couplings are highly suppressed yielding no effect on EW scale physics. 8 and can still lead to strong FOPTs – hence to poten- This highly motivates future experimental proposals, in- tially observable primordial GW signals. One sees that cluding LISA, u-DECIGO and BBO missions, accessing this “fake” low-scale seesaw scenario requires tiny values to the mHz frontier, as an indirect and complementary of the neutrino “Dirac” Yukawa couplings Yν to fit the probe of neutrino mass generation, providing an impor- small neutrino masses in the presence of relatively large tant information on the electroweak phase transition. U(1) symmetry breaking scale v that can be placed not L σ While “genuine” low-scale seesaw would also predict large too far from the EW scale. Such “fake” low-scale seesaw charged lepton flavor violation [29–33], and unitarity vi- contrasts with the “genuine” low-scale seesaw considered olation in neutrino propagation [34–36], these features in Sec.IIIB, which does not require this restriction. are absent in the “fake” low-scale seesaw. This way one Here, for an easy one-to-one comparison with the “gen- can distinguish the two schemes in high-intensity/energy uine” inverse seesaw scenario studied above, we will con- frontier setups. Here we have shown that “fake” and “gen- sider a type-I majoron seesaw model with six heavy neu- uine” schemes may also have potentially distinct grav- trinos. This way we preserve the number of fermionic itational footprints. We saw explicitly that they can degrees of freedom entering the thermal corrections by produce different gravitational-wave spectra, testable in considering the following Lagrangian upcoming gravitational-wave experiments. As we stand Type−I ¯ c c c right now, the new portal provided by the gravitational- Yuk = Yν,iLHνi + Yσ,iσνi νi + h.c. (17) L wave physics in the multi-messenger era may contribute c where ν and S in the inverse seesaw scenario are replaced to shed light on the mystery of neutrino mass generation. c c by ν1 and ν2 in this extended type-I seesaw variant, and the off-diagonal terms in the heavy neutrino mass matrix and “Dirac” Yukawa couplings Yν,i are assumed to be negligibly small. As our main result, we find that, in ACKNOWLEDGMENTS a large region of parameter space, both “fake” and “genuine” low-scale seesaw + majoron lead to the possibility of strong FOPTs. The corresponding GW spectra in the Useful discussions and correspondence with Zurab “fake” seesaw obtained for vσ (100) GeV – (1) TeV Berezhiani, Yifu Cai and Nico Yunes are gratefully ac- ∼ O O are shown in Fig.5. Parameter values of this model as- knowledged. A.P.M wants to thank Marek Lewicki and sociated to Fig. 5a are given in Tab.IV and Tab.III. Bogumiła Świe˙zewska for insightful discussions about bubble wall collision contributions to the spectrum of GW. A.A. and A.M. wish to acknowledge support by IV. CONCLUSION the NSFC, through grant No. 11875113, the Shanghai Municipality, through grant No. KBH1512299, and by Fudan University, through grant No. JJH1512105. J.B. To conclude, we analysed the most popular implementa- acknowledges his partial support by the NSFC, through tions of the type-I seesaw mechanism for neutrino mass the grants No. 11375153 and 11675145. A.A. and A.M. generation. We studied both the cases of explicit and would like to thank IFIC for hospitality during the prepa- spontaneous breakdown of the lepton number symmetry ration of this work. R.P. is supported in part by the in the neutrino sector. The second, “dynamical” symme- Swedish Research Council grants, contract numbers 621- try breaking implies the majoron field. We have found 2013-4287 and 2016-05996, as well as by the European that various scenarios lead to different patterns of phase Research Council (ERC) under the European Union’s transitions. We showed that explicit lepton number vi- Horizon 2020 research and innovation programme (grant olation in the neutrino sector cannot induce any strong agreement No 668679). The work of R.P. was also sup- electroweak phase transition. Therefore, it does not lead ported in part by the Ministry of Education, Youth and to any gravitational-wave background signal testable by Sports of the Czech Republic, project LT17018. The next-generation satellite interferometers. work of A.P.M. has been performed in the framework of The case when neutrino masses emerge from a dynam- COST Action CA16201 “Unraveling new physics at the ical mechanism in which lepton number violation hap- LHC through the precision frontier” (PARTICLEFACE). pens spontaneously leads to much clearer gravitational A.P.M. is supported by Fundação para a Ciência e a Tec- footprints. Within such majoron seesaw case, we found nologia (FCT), within project UID/MAT/04106/2019 that both the standard type-I seesaw (taken at a low (CIDMA) and by national funds (OE), through FCT, scale) and the “genuine” low-scale type-I seesaw (like I.P., in the scope of the framework contract foreseen in the inverse seesaw) predict a strong gravitational wave the numbers 4, 5 and 6 of the article 23, of the Decree- signal, testable in the 0.1 100 mHz frequency range. Law 57/2016, of August 29, changed by Law 57/2017, − 9 of July 19. A.P.M. is also supported by the Enabling PAR/31000/2017. R.V. and J.W.F.V. are supported by Green E-science for the Square Kilometer Array Research the Spanish grants SEV-2014-0398 and FPA2017-85216- Infrastructure (ENGAGESKA), POCI-01-0145-FEDER- P (AEI/FEDER, UE), PROMETEO/2018/165 (Gener- 022217, and by the project From Higgs Phenomenology to alitat Valenciana) and the Spanish Red Consolider Mul- the Unification of Fundamental Interactions, PTDC/FIS- tiDark FPA2017-90566-REDC.

[1] T. Kajita, “Nobel Lecture: Discovery of atmospheric Models,” vol. D34, p. 1642. 1986. neutrino oscillations,” Rev.Mod.Phys. 88 (2016) 030501. [17] M. Gonzalez-Garcia and J. W. F. Valle, “Fast Decaying [2] A. B. McDonald, “Nobel Lecture: The Sudbury Neutrinos and Observable Flavor Violation in a New Neutrino Observatory: Observation of flavor change for Class of Majoron Models,” Phys.Lett. B216 (1989) solar neutrinos,” Rev.Mod.Phys. 88 (2016) 030502. 360–366. [3] J. W. F. Valle and J. C. Romao, Neutrinos in high [18] E. K. Akhmedov et al., “Left-right symmetry breaking energy and astroparticle physics. John Wiley & Sons in NJL approach,” Phys.Lett. B368 270–280, (2015). www.wiley.com/buy/9783527411979. arXiv:hep-ph/9507275 [hep-ph]. [4] S. Weinberg, “Varieties of baryon and lepton [19] E. K. Akhmedov et al., “Dynamical left-right symmetry nonconservation,” Phys. Rev. D22 1694. breaking,” Phys.Rev. D53 2752–2780, [5] S. M. Boucenna, S. Morisi, and J. W. F. Valle, “The arXiv:hep-ph/9509255 [hep-ph]. low-scale approach to neutrino masses,” Adv.High [20] M. Malinsky, J. Romao, and J. W. F. Valle, “Novel Energy Phys. 2014 (2014) 831598, arXiv:1404.3751 supersymmetric SO(10) seesaw mechanism,” [hep-ph]. Phys.Rev.Lett. 95 161801, arXiv:hep-ph/0506296 [6] E. Ma and R. Srivastava, “Dirac or inverse seesaw [hep-ph]. neutrino masses with B − L gauge symmetry and S3 [21] S. Das et al., “Heavy Neutrinos and Lepton Flavour flavor symmetry,” Phys. Lett. B741 (2015) 217–222, Violation in Left-Right Symmetric Models at the LHC,” arXiv:1411.5042 [hep-ph]. Phys.Rev. D86 055006, arXiv:1206.0256 [hep-ph]. [7] E. Ma, N. Pollard, R. Srivastava, and M. Zakeri, “Gauge [22] F. F. Deppisch, N. Desai, and J. W. F. Valle, “Is B − L Model with Residual Z3 Symmetry,” Phys. Lett. charged lepton flavor violation a high energy B750 (2015) 135–138, arXiv:1507.03943 [hep-ph]. phenomenon?,” Phys.Rev. D89 (2014) 051302, [8] S. Centelles Chuliá et al., “Dirac Neutrinos and Dark arXiv:1308.6789 [hep-ph]. Matter Stability from Lepton Quarticity,” Phys. Lett. [23] F. F. Deppisch, P. Bhupal Dev, and A. Pilaftsis, B767 (2017) 209–213, arXiv:1606.04543 [hep-ph]. “Neutrinos and Collider Physics,” New J.Phys. 17 [9] S. Centelles Chuliá, R. Srivastava, and J. W. F. Valle, (2015) 075019, arXiv:1502.06541 [hep-ph]. “Seesaw roadmap to neutrino mass and dark matter,” [24] A. S. Joshipura and J. W. F. Valle, “Invisible higgs Phys. Lett. B781 (2018) 122–128, arXiv:1802.05722 decays and neutrino physics,” Nuclear Physics B 397 [hep-ph]. no. 1, (1993) 105 – 122. http://www.sciencedirect. [10] S. Centelles Chuliá, R. Srivastava, and J. W. F. Valle, com/science/article/pii/055032139390337O. “Seesaw Dirac neutrino mass through dimension-six [25] M. A. Diaz et al., “Seesaw Majoron model of neutrino operators,” Phys. Rev. D98 no. 3, (2018) 035009, mass and novel signals in Higgs boson production at arXiv:1804.03181 [hep-ph]. LEP,” Nucl. Phys. B527 (1998) 44–60, [11] C. Bonilla, S. Centelles Chuliá, R. Cepedello, arXiv:hep-ph/9803362 [hep-ph]. E. Peinado, and R. Srivastava, “Dark matter stability [26] C. Bonilla, J. C. Romao, and J. W. F. Valle, and Dirac neutrinos using only Standard Model “Electroweak breaking and neutrino mass: invisible symmetries,” arXiv:1812.01599 [hep-ph]. Higgs decays at the LHC (type II seesaw),” New J. [12] P. Minkowski, “µ → eγ at a rate of one out of 1-billion Phys. 18 no. 3, (2016) 033033, arXiv:1511.07351 muon decays?,” Phys. Lett. B67 (1977) 421. [hep-ph]. [13] T. Yanagida, “Horizontal symmetry and masses of [27] C. Bonilla, J. W. F. Valle, and J. C. Romao, “Neutrino neutrinos,” Workshop on the baryon number of the mass and invisible Higgs decays at the LHC,” Phys. Universe and unified theories, O. Sawada and A. Rev. D91 no. 11, (2015) 113015, arXiv:1502.01649 Sugamoto, eds. (1979) 95. [hep-ph]. [14] R. N. Mohapatra and G. Senjanovic, “Neutrino mass [28] A. Kosowsky, M. S. Turner, and R. Watkins, and spontaneous parity violation,” Phys. Rev. Lett. 44 “Gravitational waves from first order cosmological phase (1980) 912. transitions,” Phys. Rev. Lett. 69 (1992) 2026–2029. [15] J. Schechter and J. W. F. Valle, “Neutrino Masses in [29] J. Bernabeu et al., “Lepton Flavor Nonconservation at SU(2) x U(1) Theories,” Phys.Rev. D22 (1980) 2227. High-Energies in a Superstring Inspired Standard [16] R. Mohapatra and J. W. F. Valle, “Neutrino Mass and Model,” Phys.Lett. B187 (1987) 303. Baryon Number Nonconservation in Superstring [30] G. C. Branco, M. N. Rebelo, and J. W. F. Valle, 10

“Leptonic CP Violation With Massless Neutrinos,” scalar,” Nucl. Phys. B 619 (2001) 709–728, Phys. Lett. B225 (1989) 385–392. arXiv:hep-ph/0011335. [31] N. Rius and J. W. F. Valle, “Leptonic CP Violating [48] M. Gonderinger, H. Lim, and M. J. Ramsey-Musolf, Asymmetries in Z0 Decays,” Phys. Lett. B246 (1990) “Complex Scalar Singlet Dark Matter: Vacuum 249–255. Stability and Phenomenology,” Phys. Rev. D 86 (2012) [32] F. Deppisch and J. W. F. Valle, “Enhanced lepton 043511, arXiv:1202.1316 [hep-ph]. flavor violation in the supersymmetric inverse seesaw [49] J. M. Cline, K. Kainulainen, P. Scott, and C. Weniger, model,” Phys. Rev. D72 (2005) 036001, “Update on scalar singlet dark matter,” Phys. Rev. D arXiv:hep-ph/0406040 [hep-ph]. 88 (2013) 055025, arXiv:1306.4710 [hep-ph]. [33] S. Antusch, C. Biggio, E. Fernandez-Martinez, [Erratum: Phys.Rev.D 92, 039906 (2015)]. M. Gavela, and J. Lopez-Pavon, “Unitarity of the [50] E. Gabrielli, M. Heikinheimo, K. Kannike, A. Racioppi, Leptonic Mixing Matrix,” JHEP 0610 (2006) 084. M. Raidal, and C. Spethmann, “Towards Completing [34] F. Escrihuela et al., “On the description of nonunitary the Standard Model: Vacuum Stability, EWSB and neutrino mixing,” Phys.Rev. D92 (2015) 053009, Dark Matter,” Phys. Rev. D 89 no. 1, (2014) 015017, arXiv:1503.08879 [hep-ph]. arXiv:1309.6632 [hep-ph]. [35] D. Forero et al., “Lepton flavor violation and [51] N. Khan and S. Rakshit, “Study of electroweak vacuum non-unitary lepton mixing in low-scale type-I seesaw,” metastability with a singlet scalar dark matter,” Phys. JHEP 1109 (2011) 142, arXiv:1107.6009 [hep-ph]. Rev. D 90 no. 11, (2014) 113008, arXiv:1407.6015 [36] O. Miranda, M. Tortola, and J. W. F. Valle, “New [hep-ph]. ambiguity in probing CP violation in neutrino [52] K. Ghorbani and P. H. Ghorbani, “A Simultaneous oscillations,” Phys.Rev.Lett. 117 (2016) 061804, Study of Dark Matter and Phase Transition: arXiv:1604.05690 [hep-ph]. Two-Scalar Scenario,” JHEP 12 (2019) 077, [37] Y. Chikashige, R. N. Mohapatra, and R. D. Peccei, “Are arXiv:1906.01823 [hep-ph]. There Real Goldstone Bosons Associated with Broken [53] C. Cosme, J. G. Rosa, and O. Bertolami, “Can dark Lepton Number?,” Phys. Lett. 98B (1981) 265–268. matter drive electroweak symmetry breaking?,” [38] J. Schechter and J. W. F. Valle, “Neutrino Decay and arXiv:1811.08908 [hep-ph]. Spontaneous Violation of Lepton Number,” Phys. Rev. [54] D. Azevedo, M. Duch, B. Grzadkowski, D. Huang, D25 (1982) 774. M. Iglicki, and R. Santos, “Testing scalar versus vector [39] V. Berezinsky and J. W. F. Valle, “The KeV majoron as dark matter,” Phys. Rev. D 99 no. 1, (2019) 015017, a dark matter particle,” Phys. Lett. B318 (1993) arXiv:1808.01598 [hep-ph]. 360–366, arXiv:hep-ph/9309214 [hep-ph]. [55] B. Grzadkowski and D. Huang, “Spontaneous [40] M. Lattanzi and J. W. F. Valle, “Decaying warm dark CP -Violating Electroweak Baryogenesis and Dark matter and neutrino masses,” Phys. Rev. Lett. 99 Matter from a Complex Singlet Scalar,” JHEP 08 (2007) 121301, arXiv:0705.2406 [astro-ph]. (2018) 135, arXiv:1807.06987 [hep-ph]. [41] J.-L. Kuo et al., “Decaying warm dark matter and [56] C.-W. Chiang, M. J. Ramsey-Musolf, and E. Senaha, structure formation,” JCAP 1812 no. 12, (2018) 026, “Standard Model with a Complex Scalar Singlet: arXiv:1803.05650 [astro-ph.CO]. Cosmological Implications and Theoretical [42] M. Lattanzi et al., “Updated CMB, X- and gamma-ray Considerations,” Phys. Rev. D 97 no. 1, (2018) 015005, constraints on Majoron dark matter,” Phys.Rev. D88 arXiv:1707.09960 [hep-ph]. 063528, arXiv:1303.4685 [astro-ph.HE]. [57] T. Brune and H. Päs, “Massive Majorons and [43] F. Bazzocchi et al., “X-ray photons from late-decaying constraints on the Majoron-neutrino coupling,” majoron dark matter,” JCAP 0808 (2008) 013, Phys.Rev. D99 (2019) 096005, arXiv:1808.08158 arXiv:0805.2372 [astro-ph]. [hep-ph]. [44] ATLAS Collaboration, M. Aaboud et al., “Search for [58] M. Hindmarsh, S. J. Huber, K. Rummukainen, and invisible Higgs boson decays in vector boson fusion at D. J. Weir, “Numerical simulations of acoustically √ s = 13 TeV with the ATLAS detector,” Phys. Lett. generated gravitational waves at a first order phase B793 (2019) 499–519, arXiv:1809.06682 [hep-ex]. transition,” Phys. Rev. D92 no. 12, (2015) 123009, [45] CMS Collaboration, A. M. Sirunyan et al., “Search for arXiv:1504.03291 [astro-ph.CO]. invisible decays of a Higgs boson produced through [59] M. Hindmarsh, S. J. Huber, K. Rummukainen, and √ vector boson fusion in proton-proton collisions at s = D. J. Weir, “Shape of the acoustic gravitational wave 13 TeV,” Phys. Lett. B793 (2019) 520–551, power spectrum from a first order phase transition,” arXiv:1809.05937 [hep-ex]. Phys. Rev. D96 no. 10, (2017) 103520, [46] R. Costa, A. P. Morais, M. O. P. Sampaio, and arXiv:1704.05871 [astro-ph.CO]. R. Santos, “Two-loop stability of a complex singlet [60] C. Grojean and G. Servant, “Gravitational Waves from extended Standard Model,” Phys. Rev. D 92 (2015) Phase Transitions at the Electroweak Scale and 025024, arXiv:1411.4048 [hep-ph]. Beyond,” Phys. Rev. D75 (2007) 043507, [47] C. Burgess, M. Pospelov, and T. ter Veldhuis, “The arXiv:hep-ph/0607107 [hep-ph]. Minimal model of nonbaryonic dark matter: A Singlet [61] L. Leitao and A. Megevand, “Gravitational waves from 11

a very strong electroweak phase transition,” JCAP 1605 [astro-ph.IM]. no. 05, (2016) 037, arXiv:1512.08962 [astro-ph.CO]. [75] H. Kudoh, A. Taruya, T. Hiramatsu, and Y. Himemoto, [62] C. Caprini et al., “Science with the space-based “Detecting a gravitational-wave background with interferometer eLISA. II: Gravitational waves from next-generation space interferometers,” Phys. Rev. D73 cosmological phase transitions,” JCAP 1604 no. 04, (2006) 064006, arXiv:gr-qc/0511145 [gr-qc]. (2016) 001, arXiv:1512.06239 [astro-ph.CO]. [76] K. Nakayama and J. Yokoyama, “Gravitational Wave [63] C. Caprini et al., “Detecting gravitational waves from Background and Non-Gaussianity as a Probe of the cosmological phase transitions with LISA: an update,” Curvaton Scenario,” JCAP 1001 (2010) 010, JCAP 2003 no. 03, (2020) 024, arXiv:1910.13125 arXiv:0910.0715 [astro-ph.CO]. [astro-ph.CO]. [77] C. J. Moore, R. H. Cole, and C. P. L. Berry, [64] J. Ellis, M. Lewicki, J. M. No, and V. Vaskonen, “Gravitational-wave sensitivity curves,” Class. Quant. “Gravitational wave energy budget in strongly Grav. 32 no. 1, (2015) 015014, arXiv:1408.0740 supercooled phase transitions,” JCAP 1906 no. 06, [gr-qc]. (2019) 024, arXiv:1903.09642 [hep-ph]. [78] J. Ellis, M. Lewicki, and J. M. No, “On the Maximal [65] C. L. Wainwright, “CosmoTransitions: Computing Strength of a First-Order Electroweak Phase Transition Cosmological Phase Transition Temperatures and and its Gravitational Wave Signal,” arXiv:1809.08242 Bubble Profiles with Multiple Fields,” Comput. Phys. [hep-ph]. Commun. 183 (2012) 2006–2013, arXiv:1109.4189 [hep-ph]. [66] T. Vieu, A. P. Morais, and R. Pasechnik, “Multi-peaked signatures of primordial gravitational waves from multi-step electroweak phase transition,” arXiv:1802.10109 [hep-ph]. [67] A. P. Morais and R. Pasechnik, “Probing multi-step electroweak phase transition with multi-peaked primordial gravitational waves spectra,” arXiv:1910.00717 [hep-ph]. [68] K. Hashino, R. Jinno, M. Kakizaki, S. Kanemura, T. Takahashi, and M. Takimoto, “Selecting models of first-order phase transitions using the synergy between collider and gravitational-wave experiments,” Phys. Rev. D99 no. 7, (2019) 075011, arXiv:1809.04994 [hep-ph]. [69] A. Alves, T. Ghosh, H.-K. Guo, K. Sinha, and D. Vagie, “Collider and Gravitational Wave Complementarity in Exploring the Singlet Extension of the Standard Model,” JHEP 04 (2019) 052, arXiv:1812.09333 [hep-ph]. [70] G. Kurup and M. Perelstein, “Dynamics of Electroweak Phase Transition In Singlet-Scalar Extension of the Standard Model,” Phys. Rev. D96 no. 1, (2017) 015036, arXiv:1704.03381 [hep-ph]. [71] K. Hashino, M. Kakizaki, S. Kanemura, P. Ko, and T. Matsui, “Gravitational waves and Higgs boson couplings for exploring first order phase transition in the model with a singlet scalar field,” Phys. Lett. B766 (2017) 49–54, arXiv:1609.00297 [hep-ph]. [72] M. Kakizaki, S. Kanemura, and T. Matsui, “Gravitational waves as a probe of extended scalar sectors with the first order electroweak phase transition,” Phys. Rev. D92 no. 11, (2015) 115007, arXiv:1509.08394 [hep-ph]. [73] S. Moretti and K. Yagyu, “Constraints on Parameter Space from Perturbative Unitarity in Models with Three Scalar Doublets,” Phys. Rev. D91 (2015) 055022, arXiv:1501.06544 [hep-ph]. [74] LISA Collaboration, P. Amaro-Seoane et al., “Laser Interferometer Space Antenna,” arXiv:1702.00786