Gravitational Waves from Pati-Salam Dynamics

We show that it is possible to use gravitational wave detectors to observe the occurrence of a first order phase transition in Pati-Salam extensions of the Standard Model. We show that the peak frequency of the expected gravitational wave signals ranges within $0.1-10$ Hz. We find amusing that the next generation of gravity waves detectors are able to explore time honoured extensions of the Standard Model occurring at energy scales inaccessible by present and future particle physics accelerators.


I. INTRODUCTION
The idea of using gravitational wave as a complementary approach to explore particle physics started some time ago e.g. [1][2][3]. However, the bulk of the research concentrated, so far, on the electroweak phase transition which is typically in the detection region of the LISA gravitational wave detector as nicely summarised in [4,5]. Of special interest is the possible detection of gravity waves originated in Grand Unified Theories (GUT). The interest arises also because typically the new physics energy scale of GUTs is beyond the reach of the existing and even future 100 TeV colliders.
A prerequisite to start even discussing gravitational wave detection is that the underlying theory must undergo a strong first order phase transition at some point during the evolution of the Universe. Additionally the higher is the energy scale of the first order phase transition the higher will be the peak frequency of the gravitational wave that needs to be detected. Inevitably, the upper frequency limit of the existing and planned gravitational wave detectors (roughly at order of 10 3 Hz) provides an upper bound on the detectable energy scale (roughly at 10 4 − 10 5 TeV). In this sense, among the different types of GUTs, only the two semi-simple GUTs: Pati-Salam model [6] and Trinification model [7] satisfy this criterion. In this work, we will focus mainly on the gravitational wave signatures of the minimal Pati-Salam model. Our investigation differs from the one in [8] in which an alternative model of Pati-Salam model was considered. In that work the authors employed a rather involved matter content that featured, however, a simpler first order phase transition structure 1 . 1 In the work of [8], the authors try to realize the gauge coupling unification and symmetry breaking to an intermediate step (leftright model) first and thus their scalar sectors are overall more complicated. However, their first order phase transition occurs only when SU (4) is breaking while in our case both SU (4) and SU (2) R breaks and thus their analysis of the first order phase transition is simpler and fewer couplings are involved.
The Pati-Salam model of matter field unification [6] is a time-honoured example in which one can address the hypercharge triviality issue by embedding it in an asymptotically free theory. From a phenomenological standpoint it can be commended because it does not induce fast proton decay, and it can even be extended to provide a stable proton [9] while automatically providing a rationale for the existence of right handed neutrinos (see more details in a recent nice review [10]).
So far, asymptotic freedom has been the well traveled route to resolve the triviality problem. An alternative route is that in which the UV theory acquires an interacting fixed point, before gravity sets in, de facto saving itself from the presence of a cutoff. This unexplored route was opened when the first safe gauge-Yukawa theory was discovered in [11].
To achieve a safe theory with a small number of colours we employ large number of matter fields techniques [12,13]. The first phenomenological applications of the large N f limit appeared in [14] where it was first explored whether the SM augmented by a large number of vector-like fermions can have an ultra-violet fixedpoint in all couplings. The full treatment appeared in [15] and further generalized in [16]. It was found in [15] and later on proved in [16] that while the non-abelian gauge couplings, Higgs quartic and Yukawa coupling can exhibit a safe fixed point, the hypercharge remains troublesome. In fact, for abelian theories the fermion mass anomalous dimension diverges at the alleged fixed point [17] suggesting that a safe extension of the SM, like the asymptotically free counterpart, is best obtained by embedding the SM in a non-abelian gauge structure. The first non-abelian safe PS and Trinification embeddings were put forward in [18,19]. However, in the minimal models, only one generation of SM fermions can be modelled, since all the Yukawa couplings are determined by the same UV fixed point value with no resulting hierarchy at low energy. Yukawa hierarchies among three generations of SM fermions are discussed in [20].
In this work, we will start by investigating gravitational wave signatures emerging in Pati-Salam extensions of the Standard Model embedded in an asymptotically safe scenario. We use these predictions as an initial seed value to study the first order phase transition and gravitational wave signatures. Later, we will depart from the safety scenario and will explore a more general parameter space. Therefore, our work of studying the phase transition and gravitational wave generation is very general and applies to both safe and non-safe embeddings of the Pati-Salam model.
We discover that the next-generation gravity waves detectors are able to explore time honoured extensions of the Standard Model occurring at energy scales inaccessible by present and future particle physics colliders. More precisely we show that the peak frequency of the expected gravitational wave signals ranges within 0.1 − 10 Hz.
The paper is organised as follows. In Section II we introduce the Pati-Salam model while in Section III we compute the finite temperature corrections to the relevant part of the potential of the theory. The order of the phase transition as well as gravitational waves generation and detection are studied in Section IV. The predictions for the gravity waves signals stemming from the model parameters are presented in Section V. We conclude in Section VI. In the appendix we provide some detailed computations.

II. INTRODUCING THE PATI-SALAM MODEL
We first briefly review the Pati-Salam embedding of the SM suggested in [18].
Consider the time-honored PS gauge symmetry group G PS [6] with gauge couplings g 4 , g L and g R , respectively. Here the gauge group SU (4) ⊃ SU (3) C ⊗ U (1) B−L , where SU (3) C denotes the SM QCD gauge group. The SM quark and lepton fields are unified into the G PS irreducible representations where i = 1, 2, 3 is a flavor index. In order to induce the breaking of G PS to the SM gauge group, we introduce a scalar field φ R which transforms as the fermion multiplet ψ R , that is, φ R ∼ (4, 1, 2): where the neutral component φ 0 R takes a non-zero vev, We also introduce an additional (complex) scalar field Φ ∼ (1, 2, 2), with Gauge Yukawa Scalar SU (4) : g4 ψ L/R : y, yc φR : λR1, λR2 SU (2)L : gL NL : yν portal: λRΦ 1 , λRΦ 2 , λRΦ 3 SU (2)R : gR F : yF Φ : λ1, λ2, λ3, λ4 which is responsible of the breaking of the EW symmetry. The most general Yukawa Lagrangian for the matter fields ψ L/R is: (5) where y and y c are the Yukawa couplings for the third generation only. Note that the Yukawa couplings of the first two generations can be generated through the clockwork mechanism [20].
In the case of a self-conjugate bi-doublet field Φ ≡ Φ c , one obtains degenerate masses at tree-level, namely In order to separate the neutrino and top masses in Eq. (6), we implement the seesaw mechanism [21][22][23][24] by adding a new chiral fermion singlet N L ∼ (1, 1, 1), which has Yukawa interaction (see e.g. [18,25] for more details) In order to split the mass of top, bottom and tau lepton in Eq. (6), we introduce a new vector-like fermion F ∼ (10, 1, 1) with mass M F and Yukawa interactions (see e.g. [18,25] for more details): All the field contents and couplings are summarized in Tab. I.

A. Tree Level Effective Potential of Pati-Salam Model
The relevant terms in the tree level effective potential can be written as: It is important to note that we do not include any explicit mass terms in the tree level potential. The symmetry breaking in this work is induced by Coleman-Weinberg mechanism.
If we write out φ R explictly as : where we choose the symmetry breaking direction of φ R and thus all field components except the φ R15 direction are zero. As mentioned above, φ R triggers breaking Out of sixteen scalar fields, there are nine Goldstone bosons and seven physical bosons. Therefore, eight gauge bosons of SU (4) (corresponding to QCD gluons) and one gauge field from SU (4) ⊗ SU (2) R (which is simply U (1) Y , a linear combination of the U (1) B−L from SU (4) and U (1) R from SU (2) R , with Y = 2I R + B − L) remain massless. The other nine gauge bosons of SU (4) ⊗ SU (2) R (six leptoquark, two right boson W ± R and one Z ) become massive. With Eq.(9), we can construct the mass matrix of the scalar fields and obtain sixteen tree level mass eigenvalues. These mass eigenvalues can be divided into nine Goldstone bosons with a mass M 2 Gold = v 2 (λ R1 + λ R2 ) and seven physical Higgses, one out of which has a mass of M 2 Higgs1 = 3v 2 (λ R1 + λ R2 ) and six other Higgses with a mass M 2 Higgs2 = v 2 λ R1 .

B. Loop Level Effective Potential of Pati-Salam Model
In this section, we will discuss the one loop contributions to the effective potential from scalar, gauge fields and fermions. The general formula is well known and can be written as: where the sum runs over the bosons (+) and fermions (−) and n i counts the internal degrees of freedom (d.o.f.) of each species i. The symbols m i , µ and C i correspond respectively to the tree level mass terms, renormalization scale and constant (equal to 5/6 for gauge bosons and 3/2 for scalars and fermions in Minimal Subtraction Scheme). We define the background field as ρ. In the following, we write out the scalars, gauge fields and fermions contribution explicitly. The Higgs fields contributions (7 d.o.f.) to the one loop effective potential V Higgs are: The Goldstone contributions (9 d.o.f.) to the one-loop effective potential V Gold are: The lepto-quark contributions from SU (4) gauge fields (6 lepto-quark ×3 polarization=18 d.o.f.) to the one-loop effective potential are: where the tree level lepto-quark mass is given by to the one loop effective potential are: where the tree level W R mass is given by M 2 where the tree level Z mass is given by M 2 The neutrino singlet contribution (4 d.o.f. of Dirac Fermion) to the one loop effective potential is: where the tree level neutrino singlet mass is given by with a mass term M 2 F = 1 2 y 2 F v 2 . All in all, the total one-loop effective potential is:

C. Finite Temperature Effective Potential of Pati-Salam Model
The one loop finite temperature effective potential has the following general form where +n i (−n i ) corresponds to bosons (fermions). We can further write the thermal integral in the form of the polynomials which can significantly simplify the calculations. We focus on the integral part of Eq. (20) and define: where we have used a ≡ m 2 i /T 2 . For high temperature expansions (m i /T 1), the thermal integral can be expanded respectively for bosons and fermions as: where c B and c F are respectively c B = 3/2 − 2γ E + 2 log (4π) and c F = 3/2−2γ E +2 log (π) and γ E ≈ 0.5772. For low temperature expansions (m i /T 1), the thermal integral for both bosons and fermions can be expanded as 2 : (23) To include the information for both high temperature and low temperature, we need to have an expression to connect the above two expressions Eq. (22) and Eq. (23). We find: Thus, we have the finite temperature effective potential (without ring contributions so far) as: 2 Note that there are typos in the expressions of low energy expansion in [26].

D. Ring Contribution to the Effective Potential of Pati-Salam Model
The general formula for the ring contributions can be written as: where π i (0) denotes the corresponding thermal mass contributions to the species i from the relevant bosonic d.o.f. j (in the outside rings of the daisy diagram). To consider the ring diagram contributions to the Higgs field, for example, π Higgs should include all the scalar field (thermal mass) contributions denoted as π Higgs1 Higgs , π Higgs2 Higgs , π Gold Higgs as well as the gauge field contributions. For thermal mass contributions to the scalar field from the gauge and scalar fields (i.e. scalar field in the big central ring of the Daisy diagram), we have the following general formula for the contributions of different species j in the outside ring of the daisy diagram i.e. π j scalar (0) = 1 12 Thus, we obtain the thermal mass from the two Higgs fields and Goldstone fields respectively as: Similarly, the scalar thermal mass contributions from the gauge fields are obtained in the following: To obtain the total thermal mass contributions to the Higgs field, we need to include all the above thermal masses i.e. Eq.(28), Eq. (29) and we have: Note that for each scalar field d.o.f. (either the Higgs or Goldstone bosons), it receives the same ring diagram contributions j π j i . Thus, by using Eq. (26) and Eq. (30), we obtain the total ring contributions to the scalar fields in the Pati-Salam model are: Now we consider the case where the gauge fields are in the central ring of the Daisy diagram. We have the following general formulas to calculate the gauge, scalar and fermion fields contributions to the gauge thermal masses for both abelian and Non-abelian cases. For abelian case, we have: where L denotes the longitudinal thermal mass since it can be shown that that the transverse thermal mass is suppressed and Y S , Y F correspond respectively to the hypercharge of relevant scalar and fermion fields. For nonabelian case, we have: where t 2 (R S ) , t 2 (R F ) corresponds respectively to the Dynkin indices of the scalar and fermion representations, We obtain the total thermal mass contributions to the lepto-quark, W ± R , and Z are: When computing ring contributions for gauge fields, we use the original basis instead of the mass eigenstates. Thus, both m 2 i (ρ) and i π j i (0) are rewritten as matrices M 2 (ρ) and Π (0) respectively rather than eigenvalues as in the above scalar case. Eq. (27) can be correspondingly modified as: where we include all contributions to the gauge rings and take into account only the massive gauge bosons for the big rings. The Π (0) is a diagonal 10-by-10 matrix with the entries of (i, i) being 5g 2 4 T 2 /3 and entries of (j, j) being 4g 2 R T 2 /3 for i = (1, . . . , 7) and j = (8,9,10). In contrast, M 2 (ρ) is a nearly-diagonal symmetric 10by-10 matrix with the first six diagonal elements being g 2 4 ρ 2 /4, the seventh being 3g 2 4 ρ 2 /8, and the last three diagonal being g 2 R ρ 2 /4, plus two off-diagonal elements: [M 2 (ρ)] 7, 10 =[M 2 (ρ)] 10, 7 = 3/32g R g 4 ρ 2 .

E. Complete Finite Temperature Potential
Now we are ready to write out the total finite temperature effective potential of the Pati-Salam model. It can be written as:  TABLE II. This table summarizes the sample coupling solutions at the Pati-Salam symmetry breaking scale. We did not include λ1, λ2, λ3, λ4, λRΦ 1 , λRΦ 2 , λRΦ 3 , y, yc since they are irrelevant in studying the finite temperature effective potential. Note that this set of solutions is obtained from a safe UV fixed point.

IV. FIRST ORDER PHASE TRANSITION AND GRAVITATIONAL WAVE
In this section, we will discuss the order of the possible early time Pati-Salam phase transition and the impact on possible gravitational wave signals.

A. Strong First Order Phase Transition
Here we focus on showing that a strong first order phase transition can occur at around the Pati-Salam symmetry breaking scale with a sample coupling solutions shown in Tab. II. We did not include all the couplings in the table since the remaining couplings are irrelevant in the analysis of our effective potential. We further note that the sample solutions in Tab. II are the ones leading to an asymptotically safe extension of the Pati-Salam model. However, we will show that the occurrence of a first order phase transition is not limited to this set of specific values of the couplings.
The finite temperature effective potential Eq. (36) is shown in Fig. 1. Here we have set the renormalization scale µ at 5000 TeV that is reasonable as the lower bound on the Pati-Salam physics scale is at 2000 TeV or so, derived from the upper limit Br (K L → µ ± e ∓ ) < 4.7 × 10 −12 [27]. We have also chosen the temperature T to match the critical temperature i.e. T = T c = 2680 TeV at which the potential has degenerate minima.
A positive non-trivial (away from the origin) minimum occurs for φ R ∼ 8400 TeV and it is denoted as φ Rc and thus φ Rc /T c ∼ 3.13 > 1. This shows that the associated phase transition is a strong first order one.

B. Connection between First Order Phase Transition and Coleman-Weinberg Symmetry Breaking
We noticed that a strong first order phase transition occurs when spontaneous symmetry breaking happens via the Coleman-Weinberg mechanism. This is in line with the results and expectations of [28]. Of course, in other models first order phase transitions can still occur when symmetry breaking is generated via a hard negative mass square in the potential [29].
Around the finite temperature transition the Coleman-Weinberg values of the couplings reported in Tab. II  To gain insight it is interesting to show the symmetry breaking phenomenon via the stream plot provided in Fig. 3. The green line consisting of two symmetry breaking lines (λ R1 +λ R2 = 0 for λ R2 < 0 and λ R2 /2+λ R1 = 0 for λ R2 > 0) divides the plot into two phases. The right hand side of the green line corresponding to the vacuum stable phase while the left side is related to the symmetry breaking phase. In our convention the arrows point towards the infrared. The two dots correspond respectively to a saddle point (the red one) and to an UV fixed point in both couplings. The bare couplings are meant to be fixed at some high energy scale on the right hand side of the plot. A glance at the plot shows that the only consistent way to radiatively cross the green line is by initiating the flow in the bottom right corner of the plot. One might be tempted to cross it from left to right by starting near the black dot. However this scenario would lead to an unstable potential at high energies and therefore is discarded. Focussing on the bottom right corner there is a special asymptotically safe trajectory emanating from the red dot. On that trajectory the theory will avoid a Landau pole and can be considered fundamental (up to gravity) in the deep ultraviolet. Another point is that the trajectory leads to a predictive infrared physics. We are also pleased to see that there is a wider region of UV bare couplings values that lead to a Coleman-Weinberg phenomenon beyond the asymptotically safe limit.

C. Bubble nucleation
The time is ripe to discuss bubble nucleation within our model. We will provide a brief review of the method and apply it to our case.
The general picture is that as the universe cools down, a second minimum, away from the origin, develops below a critical temperature. This triggers the tunnelling from the false vacuum, at the origin, to the stable vacuum below the critical temperature. Assuming the transition to be first order, the tunnelling rate per unit volume Γ (T ) from the metastable (false) vacuum to the stable one is suppressed by the three dimensional Euclidean action S 3 (T ) and we have [30]: The Euclidean action has the form: where we use the difference of the potential F (ρ, T ) ≡ V (ρ, T ) − V (0, T ) to adjust the "datum point" of the potential at zero. The bubble configuration (instanton solution) is give by solving the following equation of motion of the action in Eq. (38): with the associated boundary conditions: To find the solutions we use the so called overshooting and under shooting method. We also used the numerical package, CosmoTransitions [31] to cross-check our results. For T = 2200 TeV the bubble profile is shown in Fig. 4. We can insert the bubble profile ρ (r, T ) into the Euclidean action Eq. (38) and thus S 3 will be dependent on T only. The next step is to obtain the nucleation temperature which is defined as the temperature at which the rate of bubble nucleation per Hubble volume and time is approximately one. This means: where H is the Hubble constant. By using Eq.(37), we obtain: where m pl is the Planck mass. By solving Eq. (42) numerically, we find the nucleation temperature T n is around 1260 TeV. The inverse duration of the phase transition β relative to the Hubble rate H * at the nucleation temperature T n is given by: We numerically obtain β/H * 183. Next, we will calculate another important parameter α which is the ratio of the latent heat released by the phase transition normalized against the radiation density: where v Tn is the vacuum expectation value of the finite temperature effective potential at the nucleation temperature, and g * (=150) is the relativistic d.o.f. in the universe. We find α Tn ≡ α(T = T n ) = 0.217.

D. Gravitational Waves
We are now have all the instruments to address the generation and potential observation of gravitational waves stemming from the Pati-Salam early times phase transition.
For the reader's benefit we provide a brief review of the ingredients needed to discuss the acoustic gravitational waves signals by following Ref. [32]. The discussion about collision dynamics of scalar field shells and turbulence can be found in [32] and their effects can be safely neglected in light of being sub-leading.
The power spectrum of the acoustic gravitational wave is given by: where the adiabatic index Γ AI = ω/ 4/3. ω and denote respectively the volume-averaged enthalpy and energy density respectively. U f is a measure of the rootmean-square (rms) fluid velocity and is given by: where κ f is the efficiency parameter and it is well approximated by when v ω (wall speed) → 1. The spectral shape S sw (f ) is given by: with peak frequency f sw approximated by: T n 100 GeV g * 100 1 6 (49) with z p a simulation-derived factor that is of order 10, and following [33] we take it to be 6.9.
By substituting α Tn and β/H * from Eq. (43) and Eq. (44) into the above power spectrum formula for the acoustic gravitational wave Eq. (45), we plot the curves of energy density against frequency (solid lines and the sample solution in Tab. II is in red) in Fig. 5 where the coupling solutions in Tab. IV are used. We have also included the future bounds (dashed lines) coming from planned gravitational wave detection experiments such as LIGO Voyager [34,35], LISA [4], TianQing 3 [36], BBO [37,38], Einstein Telescope (ET) [39,40] and Cosmic Explorer (CE) [34]. They are shown respectively in blue, cyan, orange, purple, green and magenta in Fig. 5. Interestingly, we find the predicted acoustic gravitational wave signal predicted to be within the detection region of LIGO Voyager which is planned to be operational around 2027-2028.

V. PATI-SALAM DRIVEN GRAVITY WAVES
We are now in a position to analyse in more detail the parameter space of bare couplings leading to observable λ1 λ2 λ3 λ4 λRΦ 1 λRΦ 2 λR1 λR2 y, yc yF yν 0.13 0.01 0.03 0.05 0.10 0.01 0.34 -0.29 0.53 0 0. 67   TABLE III. This table summarizes the UV fixed point solution for NF = 13 involving the bubble diagram contributions in the Yukawa and quartic RG beta functions. yF is asymptotically free and thus is zero at the fixed point.
gravitational waves within the Pati-Salam grand unified framework.
For convenience we start with the asymptotically safe Pati-Salam scenario that has helped us quickly identify the relevant parameter space for the occurrence of a strong first order phase transition.

A. Asymptotically Safe Case
In this section, we discuss an asymptotically safe embedding of the Pati-Salam framework by adding a large number of vector-like fields into the theory. In this limit we will argue for the existence of an UV fixed point which solves the triviality problem while yielding a highly predictive theory at lower energies.
Without further ado we introduce N F pairs of vectorlike fermions charged under the fundamental representation of the Pati-Salam gauge group Eq. (1) with the following charge assignments: For simplicity, we assume that these new vector-like fermions appear at the Pati-Salam symmetry breaking scale.
Employing the large N F beta functions reported in appendix A we can compute the RG flow connecting the UV fixed point (red dot in Fig. 3) and the the SM in the infrared. For each N F 1 input, we obtain a set of UV fixed point solutions. Follow the RG flow starting from the determined UV fixed point to the electroweak scale, we can check whether it matches onto the SM.
At the PS symmetry breaking scale, we need to use matching conditions for both the gauge couplings and scalar quartic couplings. In particular, after PS symmetry breaking, the scalar bi-doublet should match the conventional two Higgs doublet model (we implement the beta functions of the two Higgs doublet model provided in [41]). We have searched the full parameter space in the range of N F ∈ (10, 200) and find that N F = 13 with the UV fixed point solutions shown in Tab. III agree best with the low energy data (both the Higgs mass and the top Yukawa coupling at the electroweak scale). We note that y F is asymptotically free for all viable solutions. We have therefore provided a UV safe completion of the SM 4 .
The sample solutions in Tab. II are already the asymptotically safe solutions corresponding to N F = 13. This set is particularly interesting because: • It corresponds to a possible UV safe fixed point rendering (up to gravity) our Pati-Salam model UV complete.
• The Pati-Salam symmetry is dynamically broken through the Coleman-Weinberg mechanism below 10000 TeV (see Fig. 2) without adding any mass terms 5 .
• Below 2680 TeV a strong first order phase transition occurs and at the nucleation temperature T n = 1260 TeV gravitational wave signals can be generated. These are within the reach of the planned LIGO Voyager experiment detection region (see Fig. 5) as well as the detection regions envisioned for the Einstein Telescope (ET), Cosmic Explorer (CE) and Big Bang Observer (BBO).
We show the results as the red solid curve in both Fig. 5 and Fig. 6).

B. Beyond the safe scenario
Here, we will go beyond the safe scenario by exploring a more general parameter space able to generate testable gravitational wave signals.
We observe that the gauge couplings g 4 , g R , g L are fixed by the Standard Model once the Pati-Salam symmetry breaking scale is chosen. In addition, when varying the quartic couplings we must ensure the presence of the Standard Model Higgs with its 125 GeV mass at the electroweak scale. We therefore vary only the Yukawa couplings y F , y ν and the two quartic couplings λ R1 , λ R2 to satisfy this constraint.
Scanning the Yukawa coupling parameter space, we discover that when increasing either y F or y ν (see black row of Tab. IV), the dimensionless energy density of the gravitational wave signal increases accordingly and the peak frequency will shift slightly to the left. This is clear when comparing the black curve with the red (safe) curve in Fig. 5.
When scanning the quartic couplings parameter space, we find that the gravitational waves signal also depends on λ R1 + λ R2 . Varying λ R1 , λ R2 with fixed λ R1 + λ R2 , the dimensionless energy density and the peak of the frequency of the gravitational wave signals are roughly this analysis is still valid because the associated trajectories are valid for any energy scale sufficiently close to the would-be UV fixed point due to the nature of the precise results of the large N f expansion away from the fixed point. 5 This result does not depend on the existence of the fixed point but it is a welcome prediction.  fixed. When increasing λ R1 + λ R2 (see Brown and Grey row of Tab. V) the dimensionless energy density of the gravitational wave signal decreases accordingly and the peak frequency shifts significantly to the left with respect to the safe scenario. This can be seen from Fig. 6. Thus, differently from the safe scenario where the peak frequency is roughly around 10Hz, going beyond the safe scenario allows for a peak of frequency ranging between 0.1 and 10Hz.

VI. CONCLUSIONS
We investigated the gravitational wave signatures stemming from the Pati-Salam model by identifying the parameter space of its couplings supporting a strong first order phase transition.
We started the analysis by employing a safe version of the Pati-Salam extension of the Standard Model and then quickly generalised to more generic situations. We find that a Coleman-Weinberg spontaneous breaking of the symmetry triggers a first order phase transition that can be observed via the next generation of gravitational wave detectors such as LIGO Voyager, the Einstein Telescope (ET) and the Cosmic Explorer (CE).
Beyond the safe scenario we notice that the Yukawa couplings y ν , y F affect mostly the gravitational wave energy density while the combination of quartic couplings λ R1 + λ R2 shifts its peak frequency.
Concluding, we discover that the peak frequency of the gravitational wave signals stemming from the Pati-Salam model ranges within 0.1 − 10 Hz. Our results lead to the exciting news that the next generation of gravity waves detectors will be able to explore important extensions of the Standard Model appearing not at the electroweak scale but at much higher energy scales not accessible through present and future particle physics accelerators. The beauty of large-N F beta function is by noticing that a subset of the Feynman diagrams (denoted as bubble chain) can be summed up into a closed form at 1/N F order. Thus, all the higher order information up to 1/N F order is encoded in the summation functions denoted as F 1 (A), H 1 (A), H 0 (A) below. It also deserves to note that these summation functions possess the pole structures: To the leading 1/N F order, the higher order (ho) contributions to the general RG functions of the gauge couplings were computed in [16], while for the simple gauge groups in [13,42] and for the abelian in [12].
Here we summarize the results. The ho contributions to dα i /d log µ (in the semi-simple case) are: with the functions H 1i and the t'Hooft couplings A i where I 1 (x) and I 2 (x) are: The Dynkin indices are T R = 1/2 (N ci ) for the fundamental (adjoint) representation while d R k ψ denotes the dimension of the fermion representation.
The RG functions of the (semi-simple) gauge couplings are: