Arxiv:1612.00632V2 [Gr-Qc] 5 Dec 2016 Similarly to Inﬂation Simple Realizations of the Mat- Ref

Dark matter and baryogenesis in the Fermi-bounce curvaton mechanism

Andrea Addazi,1, ∗ Stephon Alexander,2, † Yi-Fu Cai,3, ‡ and Antonino Marcian`o4, § 1Dipartimento di Fisica, Universit`adi L’Aquila, 67010 Coppito AQ, Italy LNGS, Laboratori Nazionali del Gran Sasso, 67010 Assergi AQ, Italy 2Department of Physics, Brown University, Providence, RI, 02912, USA 3CAS Key Laboratory for Research in Galaxies and Cosmology, Department of Astronomy, University of Science and Technology of China, Chinese Academy of Sciences, Hefei, Anhui 230026, China 4Department of Physics & Center for Field Theory and Particle Physics, Fudan University, 200433 Shanghai, China We elaborate on a toy-model of matter bounce, in which the matter content is constituted by two fermion species endowed with four fermion interaction term. We describe the curvaton mechanism that is forth generated, and then argue that one of the two fermionic species may realize baryogenesis, while the other (lighter) one is compatible with constrains on extra hot dark matter particles.

I. INTRODUCTION ns = 0.968 ± 0.006 (65%) [44], and a tensor-to-scalar ratio r significantly larger than the value allowed by the The matter bounce scenario is an alternative to inflation observational bound r < 0.12 (95%) [45]. that fulfills the same observational constraints as the lat- Nonetheless, there are few instantiations of the matter ter, but carries definite novel predictions about CMB ob- bounce scenario that predict a slight red tilt in the spec- servables to be measured in forthcoming experiments. At trum of scalar perturbations and fulfill the constraints this regard the matter bounce scenario is distinguishable on the tensor-to-scalar ratio [13, 47, 48]. The mecha- from inflation. Scale-invariant perturbations are gener- nisms that are usually considered at this purpose hinge ated in a contracting cosmology, which is then thought on the inclusion of additional matter fields [12, 87], on the to be connected to the current phase of expansion of the choice of a matter field that has a small sound speed (so universe thanks the emergence of a non-singular bounce to enhance the amplitude of vacuum fluctuations) [48], in the dynamics. This is the theoretical peculiarity of and finally on the suppression of the tensor-to-scalar ra- matter bounce models with respect to the inflationary tio during the bounce that is explained accounting to ones, as the cosmological singularity si solved, and com- quantum gravity effects [49]. pleteness of geodesics is restored. Here we will follow a different theoretical perspective, Cosmological perturbations are also dealt with pecu- closer to the intuition developed in particle-physics. In- liarly in each one of the frameworks. In inflation the deed, we intend not to deploy exotic matter fields, or different dynamical evolutions of the causal horizon and matter fields that have not been observed yet in terres- Hubble horizon are at the origin of the generation of trial experiments, and not to resort to quantum gravity scale-invariant Fourier modes that reenter the horizon. effects, extending our framework up to Planck scale. In a In the matter bounce is during the phase of matter- more conservative fashion we rather consider here matter dominated contraction that Fourier modes of the co- fields that belong to the standard model (SM) of particle moving curvature perturbation become scale-invariant. physics, and that correspond to the simplest and most For a detailed introduction to the generation of scale- conservative extensions of it, so encode dark matter in invariant perturbations we refer to [42], while for a re- the picture we will develop. And following the particle- cent review on the status of matter bounce cosmologies physics intuition that to a definite energy scale will corre- we refer to [43]. spond definite physical degrees freedom, we assume as in arXiv:1612.00632v2 [gr-qc] 5 Dec 2016 Similarly to inflation simple realizations of the mat- Ref. [13] that both the energy scale and the matter con- ter bounce scenario have been developed that deploy tent of the universe during its contracting phase are com- scalar matter fields, whose potentials are chosen ad hoc parable to the one of the present universe, which bring so to reproduce a vanishing pressure during the matter- us to consider the importance of dark matter during the dominated phase of contraction of the universe [87]. Dif- pre-bounce matter phase contraction of the Universe. ferently than inflation, observations allow to rule out the We wish to remark that recently bouncing cosmologies matter bounce scenario with a single scalar field [46]. involving dark matter (and dark energy) have received Indeed single scalar field matter bounce models predict much attention in the literature, and that distinctive and an exactly scale-invariant spectrum, while the actual ob- falsifiable predictions on CMB observables have been de- served one has a slight red tilt with a spectral index of rived that will be tested in the near future [13–20] (for a recent review see also [21]). With respect to this vast literature the gist of our proposal relies on the deployment ∗ [email protected] † stephon˙[email protected] of fermionic matter fields. ‡ [email protected] Specifically, we develop here a toy-model in which both § [email protected] matter and dark matter are described by fermionic fields, the dynamics of which is governed by the Dirac action rections to the effective equation of motion of the grav- on curved space time, and a four fermion interaction itational field. It is then worth to mention that quan- term. The latter term is actually due to the resolution tum theories of gravity, as well as effective models in- of the torsional components of the gravitational connec- spired by the problem of quantum gravity, have driven tion with respect to fermionic bilinears, and must be ac- many authors efforts in this sector. At this purpose, a counted for in the first order formalism. We then imple- characterization of the bouncing mechanisms inspired by ment a curvaton mechanism, in which the fermion field loop quantum gravity and its cosmological applications with lighter mass is responsible for the generation of al- — loop quantum cosmology — has been outlined in de- most scale-invariant curvature perturbation modes, and tailed analyses [47, 89]. the heavy mass field drives the dynamics of the back- On the other side, there exists a flourishing literature ground. We then argue that while the light fermion field that takes into account bouncing models from the point can be assumed to be a neutrino, the heavy fermion field of view of string theory, for a complete review of which can be related to the sterile neutrino, and hence by decay- we refer to Refs. [61, 62, 71] as preliminary introductions. ing the lighter neutrinos can accommodate baryogenesis The so called Hoˇrava-Lifshitz proposal can also achieve a through leptogenesis. bouncing phase for early time cosmology, as emphasized We start in Sec. II by differentiating our approach in [54], while the contiguity to the bouncing scenario of from the many other ones present within the literature. f(R) and Gauss-Bonnet theories can be read out respec- In Sec. III we then review the instantiation of the mat- tively from Refs. [102] and [103]. ter bounce mechanism that deploys one fermionic field, Nevertheless, the bouncing scenario does not neces- which from now on we will call Fermi bounce cosmol- sarily require (quantum) gravitational corrections to the ogy. In Sec. IV we review the curvaton mechanism for a energy density, but in stead a vast literature is deploying Fermi bounce cosmology that accounts for two fermionic fields that violate the null energy condition in order to species. In Sec. V we deepen the phenomenological con- achieve the bounce. Among many examples that can be sequences that can be derived for CMB observables, and pointed out, we may cite the ghost condensate scenario comment on the falsifiability of this scenario with respect [70], the so called Fermi bounce mechanism [77, 78] and to introduction of dark matter. In Sec. VI we study the the Lee-Wick theory [66]. Because of their peculiarity of application of this curvaton model to leptogenesis, and resulting from known theories of particle physics, which comment on the phenomenological constraints that can have been corroborated on the flat gravitational back- be inferred from data. Finally, in Sec. VII we spell some ground by means of high-energy terrestrial experiments, outlooks and conclusions. we will focus in the next section on the Fermi bounce models.

II. THE MATTER BOUNCE SCENARIO III. ONE FIELD FERMI-BOUNCE COSMOLOGY It is now days common knowledge that FLRW metrics suffer from singularities in all the curvature invariants. The action for the matter-gravity sector under scrutiny It was already remarked by Hawking and Penrose [24] results from the sum of the gravitational Einstein-Hilbert that the initial singularity is unavoidable if space-time action, further endowed with a topological term àla is described by General Relativity and matter undergoes Holst, plus a non-minimal covariant Dirac action. Fol- null energy conditions (NEC). Many non-singular bounc- lowing previous literature [11], we may refer to this the- ing cosmologies have been hitherto developed in order to ory as the Einstein-Cartan-Holst-Sciama-Kibble theory solve the Big-Bang singularity issue, but at the cost of (ECHSK). In the first order formalism, when gravity is dismissing some of the assumptions behind the Hawking- coupled to fermion fields, we must allow for a torsion- Penrose theorem, most notably NEC. ful part of the spin-connection. Thus the ECHSK one Bouncing mechanisms can be implemented within is necessarily a theory of gravity with torsionful con- frameworks very different from one another. A complete nection [22]. Nonetheless, a second order approach is review, comprehensive of all the bouncing models devel- always possible [23], which adopts a torsionless (Levi- oped hitherto, would be too long to be drawn in this Civita) connectionω ˜[e], and thus differs from the former paper, turning far away from our current purpose of fo- first order treatment in that a four fermion interaction cusing on a model of bounce cosmology that accounts for term emerges. Notice however that the action written, dark matter and only involves fermionic matter fields. no matter if cast in terms of a torsionful of Levi-Civita Nonetheless, before focusing on fermionic matter bounce connection, is invariant under diffeomorphisms and local models and their instantiations able to encode dark mat- Lorentz transformations. ter, we wish to briefly survey the vast scenario offered From now on we will focus on the ECHSK theory, within the literature, and enlighten some paradigmatic which in the first order formalism reads cases that have received much attention. Z The bouncing behavior of the universe at early time 1 4 µ ν IJ KL SHolst = d x |e| eI eJ P KLFµν (ω) , can be indeed reconstructed from high-energy theory cor- 2κ M

2 in which — for instance in Ref. [29] it was shown that for a suit-

IJ IJ IL J able choice of the parameters’ space region of the theory Fµν (ω) = dω + ω ∧ ωL black hole may never form. It is worth to express a comment on the most common is the field-strength of ωIJ , the Lorentz spin-connection, objection against this type of models, which concerns the κ = 8πG is the square of the reduced Planck length, N eventual appearance of instabilities. For instance, it has and been shown in Ref. [6] that some scalar fields’ actions that 1 P IJ = δ[I δJ] − IJ violate NEC might hold ghost and/or tachyonic instabil- KL K L 2γ KL ities, which naturally suggests similar issues might arise in the Fermi bounce context. Nonetheless, despite the involves the Levi-Civita symbol and the Barbero– IJKL analysis of Ref. [6] has been performed under very gen- Immirzi parameter γ. The Dirac action reads S = Dirac eral assumptions, still it relies heavily on the effective La- 1 R d4x|e|L , in which 2 Dirac grangian being second or higher order in the space-time 1h ı i derivatives, so the same conclusions can not be easily ex- L = ψγI eµ 1 − γ ı∇ ψ − mψψ + h.c., Dirac 2 I α 5 µ tended to any generic action, nor to a fermionic action, which is not quadratic in the canonical momenta. Fur- α ∈ R representing the non-minimal coupling parameter. ther work is needed to show whether for the latter system The crucial observation [30, 32] is that the torsionful the linearity in the canonical momenta prevents from sta- part of the spin-connection can be integrated out of the bility issues. For example, in context of Galileon models, ECHSK action via the Cartan equation (see e.g. [29]), several examples in which the Null energy Condition can which is in stead recovered varying the ECHSK action be consistently violated, without any instabilities, was IJ with respect to the spin-connection ω . One finally found in Refs. [7–9]. finds that the total action STot = SEH + SDirac + SInt Nonetheless, we spell here a simple argument in fa- in which the Einstein-Hilbert action SEH and the Dirac vor of stability that deploys mean field approximation. action SDirac are cast in terms of metric compatible vari- It is not difficult to show (for a detailed discussion, we ables, and an additional term SInt is present, which en- refer the reader to [10]) that the four fermion inter- codes a four fermion interaction potential. Specifically, action potential can be recast as a redefinition of the the Einstein-Hilbert action, in terms of the metric com- mass. As hint, we may consider to Fierz decompose the IJ patible variablesω ˜(e) , reads four fermion potential, and then focus only on the lower Z energy-channel of the decomposition, which entails for 1 4 µ ν IJ 3/2 SEH = d x|e|eI eJ Rµν , the densitized field χ = a ψ, from which we can con- 2κ M struct bilinear perturbations, √ while the Dirac action SDirac on curved space-time, once I µ γ e ı∇e µ − m − 2ξκ −g hχχi χ = 0 , the covariant derivative with respect to the torsionless I connection has been denoted with ∇e µ, recasts as the brackets denoting the expectation value of the background fermion bilinear in the mean field approximation. Z 1 4 I µ At every energy-scale lower than the energy scale of the SDirac = d x|e| ψγ eI ı∇e µψ − mψψ + h.c.. 2 M bounce, the effective mass will remain positive, or at most — as it happens at the bounce — it will vanish. This sug- The interaction four fermion potential casts gests there should not be any issue of instabilities that Z can be originated on the perturbed fields. 4 L M SInt =−ξκ d x|e| J5 J5 ηLM , Any successful theory of the early universe must be M able to reproduce the observed nearly scale invariant in which the we have used the definition of the axial cur- spectrum of adiabatic fluctuations in the CMBR. Scale L L invariance has been investigated hitherto within the rent J5 = ψγ5γ ψ, and introduced a function ξ of the real parameters α and γ, nameley framework of bouncing models with a contracting phase, such as Ekpyrotic [25], String Gas [26] and Pre-Big-Bang 3 γ2 2 1 scenarios [27]. On the other hand, for a number of these ξ := 1 + − . 16γ2 + 1 αγ α2 models it has proven difficult to obtain adiabatic scale invariant fluctuations in the contracting phase, mainly due Notice that the sign of ξ is crucial while discussing the to issues in resolving the singularity or mode matching physical applications in cosmology of the ECHSK action. between contracting and expanding phases [27]. For instance, a positive value of ξ corresponds to a cos- Nonetheless, Brandenberger and Finelli, and indepen- mological Fermi-liquids scenario in which the repulsive dently Wands [27, 28], have shown that a scale invariant potential is sustaining an accelerated phase of expansion power spectrum can be generated in a matter dominated of the Universe [31]. Conversely, a negative value of ξ is contracting universe, proving the existence of some “du- providing a violation of NEC, and hence is determining a ality” between the scale invariant power spectrum gener- bounce in cosmological [77, 78] or astrophysical scenarios ated in the inflationary epoch and a contracting matter

3 dominated phase. During this latter phase, gauge in- scopic states that represent coherent states in group the- variant perturbations that cross the Hubble-scale turn oretical meaning — coherent fermionic states are SU(2) out to be scale invariant because of the time-behavior coherent states, known in condensed matter ad BCS of the scale factor — in the cosmological time, a(t) ∼ states of superconductivity. We refer for this discussion (−t)2/3. It is also worth mentioning that at non-singular to the work developed in Ref. [41]. bounces, scale-invariant modes are matched to scale- We close this section emphasizing that the advantage invariant modes in the expanding phase. of the Fermi bounce mechanism mainly relies on the fact A very seminal investigation on the role of fermion that it does not require the existence of any fundamental fields in cosmology was reported in Ref. [33], in which a scalar field not observed through terrestrial experiments wide class of generic potentials of the Dirac field’s scalar in order to drive the space-time background evolution. bilinear was considered, and a detailed scrutiny of differ- The fermionic field added to the gravitational action is ent cosmological scenarios was made available. The first sufficient to account both for the matter bounce scenario analyses of the Fermi bounce mechanism then trace back and the generation of nearly scale-invariant scalar per- to Refs. [34, 38], in which the authors realized that a tor- turbations. sion induced four-fermion interaction might yield a nonsingular bounce. Further developments include the study of Ref. [96], deepening within the framework of a torsion- IV. TWO FIELDS CURVATON MECHANISM free theory, the role of a parity-violating four fermion self-interaction term. Production of scale-invariant scalar In this section we review the curvaton mechanism for curvature perturbations has been finally investigated in Fermi bounce cosmologies, which was first studied in Ref. [77], for the case of one fermion species, and then Ref. [78]. We will deploy a similar strategy than the one extended in Ref. [78] to case of the curvaton mechanism. outlined in Ref. [78] for the analysis of the curvature per- By considering a non-minimal coupling in the Dirac turbations. Nonetheless, we are aware that, in order to action (see e.g. Refs [35–37]) and a topological term prove cosmological perturbations of fermionic fields to be for the torsionful components of the spin-connection non vanishing at the linear order, the procedure first de- ωIJ (see for instance Ref. [30]), the inspection within scribed in Ref. [41] must be implemented. We will move Refs. [77, 78] has considerably enlarged the parameters’s then consistently, along the lines drawn in Ref. [41]. We space of the fermionic theories previously examined in remark anyway that the manipulation of the perturbed view of a bounce. This has allowed not only for a four fermionic bilinear that we perform here will give at the fermion interaction which is regulated by the parameters end very similar results as in Ref. [78]. There are only of the theory via the ξ function, but also for the emer- few differences that concern the observable quantities for gence of a scale-invariant power-spectrum. It is indeed CMBR, namely the scalar power spectrum and the ten- thanks to the presence of a torsion background that the sor to scala ratio parameter r, but this are not significant topological term within the Holst action turns from a experimentally, and are totally due to the four fermion surface term into a contribution to the four fermion in- interaction between the two fermionic species that we are teraction term. While is this latter term that entails an taking into account here. almost scale-invariant power-spectrum of gravitational Differently than from the approach within Ref. [78], in scalar perturbations. which four fermion interaction terms were added for each Notice however that the four-fermion density modi- one of the fermionic species considered by simply follow- fies the Friedman equations to have a negative energy ing a phenomenological recipe, the four fermion terms density that redshifts like ∼ a(t)6, thus the issues with we focus on here follow directly from the ECHSK ac- anisotropies are not yet solved in this scenario, when tion. As a consequence, when two fermionic species are we only take into account the tree-level contributions to taken into account a novel four fermion interaction term the energy density. Nonetheless, quantum corrections between the two species arises. We will then show that to the effective action of fermion fields may provide an once a mass hierarchy between the two species is con- “ekpyrotic-like” contribution that redshifts faster that sidered, the curvaton mechanism is again realized: the ∼ a(t)6, and is then able to wash out anisotropies when spacetime background evolution encode a bounce, and a the universe approach the non-singular bounce [40]. scale invariant scalar power spectrum is generated. Finally, we should also mention an important issue, with relevant observational consequences for the observation of power-spectra and cross-correlations function A. The ECHSK action and the background of the CMBR. This has to deal with the semiclassical dynamics limit of fermion fields, and the appropriate way of deal- ing with objects that fulfill the Pauli exclusion principle. We may start directly from the action for gravity and Dirac fields indeed become physical observable that sat- Dirac fermions in which torsion has been integrated out, isfy micro-causality only when they form bilinear that namely belong to the Clifford algebra. We shall then always deal with these combinations of fields, while adopting macro- S = SGR + Sψ + Sχ + SInt , (1)

4 where again the Einstein-Hilbert action is written using Using the Fierz identities, an evaluating on the coher- IJ IJ mixed-indices Riemann tensor Rµν = Fµν [ωe(e)], i.e. ent states the prodcut of the fermionic bilenears, the ﬁrst Friedmann equation can be cast, accounting for the con- Z 1 4 µ ν IJ tributions due to the two fermionic species, as it follows SGR = d x|e|eI eJ Rµν , (2) 2κ M 2 2 2 κ mψnψ + mχnχ κ (nψ + nχ) H = 3 + ξ 6 , (11) the Dirac action Sψ on curved space-time casts 3 a 3 a in which the double product of fermionic densities in the Z 1 4 I µ last term now accounts for the interaction between the Sψ = d x|e| ψγ eI ı∇e µψ − mψψψ + h.c., (3) 2 M two fermionic species. The scale factor of the metric is easily determined to and ﬁnally the interaction terms of the theory read be

Z 1 4 L M L M L M 3κ(m n +m n ) ξκ (n + n )2 3 SInt =−ξκ d x|e| Jψ Jψ + 2Jψ Jχ + Jχ Jχ ηLM , a= ψ ψ χ χ (t − t )2 − ψ χ , M 4 0 (m n + m n ) (4) ψ ψ χ χ (12) which only involve the axial vector currents J and J of ψ χ and its value in t , when the bounce takes place, imme- the ψ and χ fermionic species, but in the three possible 0 diately follows combinations. 1 1 2 3 2 3 ξκ (nψ + nχ) ξκ (nψ + nχ) For the two fermionic species the Dirac Lagrangians, a0 = − ' − . (13) provided with interactions, respectively read mψnψ + mχnχ mψnψ

Tot 1 I µ L = ψγ e ı∇e µψ − mψψψ + h.c. ψ 2 I B. The cosmological perturbations L K K −ξκ Jψ (Jψ + Jχ ) ηLK , (5) Cosmological perturbations can be studied in the flat and gauge, in which the curvature perturbation variable results to be proportional to the perturbation of the energy Tot 1 I µ L = χγ e ı∇e µχ − mχχχ + h.c. density of the system χ 2 I −ξκ J L(J K + J K ) η , (6) δρ χ χ ψ LK ζ = . (14) ρ + p with the energy-momentum tensors Perturbations of the energy densities of the fermionic ψ 1 I Tot species are linear in the perturbations of the fermionic Tµν = ψγI e(µı∇e ν)ψ + h.c. − gµν Lψ , (7) 4 bilinear. In [41] it has been shown that linear perturba- and tions of the fermionic bilinear are non-vanishing. Thus also the curvature perturbation variable defined in (14), χ 1 I Tot linear by definition, is non-vanishing for fermion fields. T = χγI e ı∇e ν)χ + h.c. − gµν L . (8) µν 4 (µ χ We remind that within the formalism introduced in [41], given a generic operator O in the spinorial inter- The background dynamics of the fermionic bilinears must nal space, the n-th infinitesimal order expansion of the be solved along the lines of [41]. Nonetheless, conclu- expectation value (on a quantum macroscopic coherent sions are here quite similar to what was found in [78]. states) of the fermionic bilinear ψOψ is defined by the On the states of semiclassicality (respectively) for the ψ- expansion fermion field, namely the coherent state |α i, and for the ψ χ-fermion field, namely the coherent state |α i, we can n 0 0 χ δ (hψOψiα ) ≡ n! hα |ψOψ|α i , (15) ψ ψ ψ n easily recover (see e.g. Ref. [41]) that on shell O(δαψ )

¯ nψ nχ in which the perturbation of the modes distribution func- hψψiαψ = 3 , hχχ¯ iαχ = 3 , (9) 0 a a tion αψ ' αψ + δαψ + ... has been considered. For simplicity of notation, we will remove the subscript αψ, and in which the fermionic densities arise from the integration denote perturbations of fermionic bilinears on the coher- of the modes’ distributions of the coherent states, i.e. ent space simply as δhψOψi. Z Z If we now take into account the two fermionic species 2 2 nψ = dµ(k)|αψ(k)| , nχ = dµ(k)|αχ(k)| , with diﬀerent values of the bare mass, we will ﬁnd that the two main contributions to the variation of the energy (10) densities read dµ(k) denoting the appropriate relativistic measure on the Fourier modes space. δρ = mχδhχ χi + mψδhψ ψi + ..., (16)

5 having neglected contributions suppressed by ξκ. On the fermion term interaction with the ψ field, can be then other hand, similarly to (11) the denominator of (14) found once the equations of motion for the field are solved becomes γI eµı∇ − m − 2ξκhχχi + hψψi χ = 0 , (23) m n + m n (n + n )2 I e µ χ p + ρ = χ χ ψ ψ + 2ξκ χ ψ . (17) a3 a6 in which we have used the mean field approximation Not astonishingly, at the zeroth order in ξκ a similar re- for the terms arising from the four fermion interactions. sult as in Ref. [78] is obtained. For a values of mχ <> mχ, which entails suppression at super- horizon scale of the perturbations due to the ψ field Following the recipe implemented in Refs. [77, 78], we — wavenumber of perturbations of ψ-bilinears are can find solutions for the spinorial components of the more blue-shifted than perturbations of χ-bilinears; Dirac equation (25) in terms of the spinorial functions

2. cross-correlation between perturbations of the ψ ˜ 1 f±h = √ [˜uL,h(k, η) +u ˜R,h(k, η)] , field and perturbations of the χ field are negligi- 2 ble, since they are due to an interaction involving 1 a graviton loop, the latter being suppressed by the g˜±h = √ [˜vL,h(k, η) +v ˜R,h(k, η)]. (26) 2 forth power of the Planck mass Mp. These are recovered by rescaling densitized spinors up to The perturbations to the χ field are then computed u˜ = a3/2u andv ˜ = a3/2v. Densitized spinors are in turn resorting to the same kind of assumptions outlined in expressed in terms of their chiral and helical components Refs. [77, 78], but following the procedure outlined in [41]. The analysis we report below is showing explicitly X X u˜L,h(k, η) that at perturbative level fluctuations that are recovered u˜(t, k) = u˜h(t, k) = ξh , (27) u˜R,h(k, η) are free of gradient and ghost instability, which often h h exists in other bouncing cosmologies. This refines our X X v˜R,h(k, η) v˜(t, k) = v˜h(t, k) = ξh , (28) previous heuristic argument exposed in Sec. III. v˜L,h(k, η) First, we notice that away from the bounce the scale h h 2 2 factor reads a(η) ' η /η0, with in which we have introduced the helicity 2-eigenspinor, written in terms of the unit vector kˆ, which is −1/2 η0 = [κ(mψnψ + mχnχ)] , (21) ˆ ˆ 1 h(kx − ıky) ˆ ξh = q ˆ ˆ , k·σ ξh = h ξh, (29) ˆ ıkx − h ky which becomes, because of the requirement specified in 2(1 − h kz) equation (19), σ standing for the Pauli matrices. −1/2 ˜ η0 ' (κmψnψ) . (22) We can recast equation (25) in terms of the fh functions: The dynamics of the perturbations of the χ-field bilinear, ˜00 2 ˜ which differently than in [78] is now provided with a four f±h + ω (k, η)f±h = 0 , (30)

6 in which we have introduced an effective frequency, de- c(k) being a k dependent coefficient that must be fined by determined by matching at the moment of Hubble crossing the above two asymptotic solutions (32) 2 ω (k, η) = (31) and (34). m a0 = k2 +m2 a2 +ım a0 + 2ξκ(n + n ) χ −ı . χ χ χ ψ a a3 The asymptotic solution at super-Hubble scales is therefore determined to be

In this framework χ is a curvaton, and does not con- √ r 1+γ− 1+4γ mχ √ tribute to drive the dynamics of spacetime background. ˜ 4 3− 1+4γ f±h ' (kη) , (35) So the condition mχ mψ holds naturally, and the sec- 2k ond term of the effective frequency can be then neglected. While for what concerns the third and the last term, having now normalized the expectation value of a which are imaginary, we may proceed to smooth them fermionic bilinears in the asymptotic future. out, by taking the time-averaged evolution at super- Hubble scales. Thanks to this consideration, the effective After the phase of contraction, the universe would frequency will turn out to depend mainly on the gradient eventually enter the nonsingular bouncing phase by 2 avoiding the spacetime singularity due to the help of the term k and the effective mass term 2ξκ(nχ + nψ)mχ/a, in which the effect of the new interaction between the background fermionic condensate. During this phase, two fermionic species is evident. the evolution of the Hubble parameter H can be ap- Notice also that the imaginary part of the effective proximated as a linear function of the cosmic time t [85, 87]. Accordingly, the scale factor behaves roughly as frequency can be suppressed in the contracting phase 2 far away from the bounce, which ensures initial states a ∼ exp(t ). In this case, one can read that the evolution for fermionic perturbations to be close to the vacuum of the nonsingular bounce can occur when the scale fac- fluctuations in Minkowski space. Eventual deviations tor reaches the minimal value. By inserting back into the from scale invariance of the perturbations (before exiting perturbation equation (30), the latter can be solved both the Hubble radius), which may arise along with the numerically and semi-analytically. The procedure is sim- universe’s contraction because of the presence of the ilar to the analyses performed in [66] (see also [46] for the imaginary part in (31) ca ne be switched off by a proper observational constraint on the growth of primordial fluc- fine tuning, requiring the imaginary part to be at most tuations during the bounce). We can learn that, for a fast of the same order of the real one. bounce with the energy scale much lower than the Planck scale, the perturbations passing through the nonsingular Finally, we can find solutions that interpolate among bouncing phase would not be much affected by the back- two different limits: ground evolution. In the present study, for simplicity we assume that the perturbations were almost conserved 1. when the gradient term is dominant, namely at sub- during the nonsingular bounce and can be inherited from Hubble scales with |kη| 1, we impose the initial the contracting phase to the expanding phase, smoothly. condition for the Fourier modes of the fermionic bilinear, which are the observable quantity, resorting to a Wentzel-Kramers-Brillouin (WKB) approxi- C. The power spectrum of scalar gravitational mation. This yields perturbations

r mχ We dispose now of all the necessary ingredients to calcu- f˜ ' 4 e−ikη . (32) ±h 2k late the power spectrum of the primordial scalar gravitational perturbations. Indeed, if we substitute (35) into This initial condition exactly coincides with the (20), we find that the power spectrum of the primordial vacuum fluctuations. Second, we study the asymp- curvature perturbations is expressed by the relation totic solution to the perturbation equation in the √ 3 2 2 4(1+γ− 1+4γ) mχ|δαχ| k √ limit of |kη| 1, i.e. at super-Hubble scales. To 3− 1+4γ 2 2 PS = 2 2 2 (kη) , (36) apply the relation a(η) ' η /η0 and Eq. (22), one mψnψ 4π can write down the effective mass term as in which we have used relations (9). Exact scale invari- γ 2ξ(nχ + nψ)mχ ance of the power spectrum corresponds to the value − 2 with γ = − . (33) η nψmψ γ = 2 (i.e. ξ = −nψmψ/[(nχ + nψ)mχ]). In this case, during the matter contracting phase, the amplitude scales as η−2 and the power spectrum generated in the 2. when the (other) asymptotic solution of the equa- fermion curvaton mechanism casts tion of motion has a leading term of the form 3 √ mχδnχ 1 2(1+γ− 1+4γ) √ PS = (37) ˜ 3− 1+4γ 2 2 2 2 2 f±h ' c(k) η , (34) mψnψ 4π a η

7 2 in which we have assumed δnχ = |δαχ| to retain only a of the Hubble rate to be of the same order of magnitude m √ψ mild, and phenomenologically negligible, dependence on of ξ . So at the fermionic matter-bounce phase the am- k. plitude of the Hubble parameter can not be bigger than m Evaluating (37) at the end of the matter contracting √ψ ξ , which allows us to approximate it with its maxi- phase t , when the scale factor equals the value a , en- mψ E E mal value |H | ' √ , and find the corresponding power able us to recast it as E ξ spectrum 3 2 mχδnχ HE 2 PS = , (38) 1 m m2 n2 16π2 ψ ψ ψ PT ' 2 2 . (41) ϑ |ξ|Mp in which we have used η = 2/H = 2/(a H ). In E E E E Finally the tensor to scalar ratio, which is by definition (38) we used the notation H and a respectively for the E E r ≡ P /P , si easily recovered to be values of the comoving Hubble parameter and of the scale T S factor at the end of matter contracting phase, just right 2 2 2 16π mψ nψ before the time t at which the phase transition takes r = , (42) E ϑ2 m3 δn M 2 place and perturbations, before reentering the Hubble χ χ p horizon, become constant. in which the condition of scale invariance γ ' 2 has been Slight deviations of γ from 2 in (36) entails to derive assumed. This result is the same that was already derived phenomenologically allowed relations for the spectral in- in [78]. dex, i.e. Cosmological observations constraint the power spec- −9 d ln P 2 trum of scalar perturbations to be PS ' 2.2 × 10 [90], n − 1 ≡ S ' − (γ − 2) , (39) and set an upper bound to the detection of primordial S d ln k 3 gravitational waves, allowing for the tensor-to-scalar ra- in which the power spectrum appears to be red-tilted. tio values within the range r < 0.12 (95%) [45]. If we neglect, as working assumption, the null hypothesis for r and estimate it with its higher bound r ∼ 0.12, we can V. CMBR PHENOMENOLOGY AND DARK constraint up to two parameters of the theory. Therefore, MATTER deploying the experimental value and bound respectively for PS and r, we use equations (38) and (42) to derive In this section, we show how hot dark matter constrains constraints of the masses of the fermion fields involved. can be satisfied in a heuristic but successful fashion The first constraint we can derive is on the mass of the within the two-fermion-species toy-model we have dis- heavy species, which is the same as in [78], i.e. cussed so far. But before tackling hot dark matter con- 2 −11 2 strains we first focus on the phenomenological conse- mψ . 10 |ξ| Mp . (43) quences of the toy-model for CMBR observables. Differently than [78], if we now assume the total mass hierarchy (19) we find that large values of |ξ| = A. Constraints for the masse from CMBR (nψmψ)/[(nχ + nψ)mχ] are favored, once we look at values of γ that allows for a nearly scale-invariant power spectrum (γ ' 2). Constraint (43) is then linking now We start by commenting on the production of primordial the mass of the heavy species to the GUT scale, if we gravitational waves, and hence on the testable implica- make a proper choice of ξ ' 104. tions for the scalar to tensor ratio. We notice that the The second constraint we can derive on the masses of dynamics of primordial gravitational waves is uniquely the two fermionic species arise from combining equations determined by the spacetime background dynamics, and (38) and (42) into remind that their evolution decouple from other perturbation modes at linear order. Thus the derivation m2 n2 of the tensor perturbations power spectrum goes along ψ ψ 2 3 2 ∼ O(10 ) . (44) Ref. [87, 94], and allows to find mχ δnχMp 1 H2 This value is actually slightly different than the one re- P = E , where ϑ = 8π(2q − 3)(1 − 3q).(40) T ϑ2 a2 M 2 ported in [78], and can be easily achieved within this E p scenario, while linking the mass of the light species to The coefficient q is typically required to be less than the mass of the heavy one. unity, and is determined by the detailed procedure of the phase transition, since it represents a background parameter associated with the contracting phase. B. Including dark matter If we assume that the universe is evolving through the bounce, and neglect for the moment its fermionic con- So far we have developed a fermion curvaton mechanism traction phase, we can estimate the maximal amplitude consistent with the latest cosmological data. The gist of

8 this framework is in the realization of a see-saw mecha- A. Minimal coupling with SM leptons nism that has phenomenological consequences in cosmology. Assuming that the fluctuation on the abundance of The ψ-species can be then identified with the right the light χ fermions is related to the abundance of the handed (RH) neutrino, considering a see-saw type I heavy ψ fermions by mechanism for the left-handed (LH) neutrino mass. A Fukugita-Yanagida leptogenesis mechanism [105, 106] n2 ψ −7 can be then realized, once all the Sakharov’s conditions δnχ ' 3 10 , (45) mχ [107] are satisfied. A minimal Lagrangian for this instantiation of the see- we can concretely realize in this framework a see-saw saw mechanism reads mechanism. The two fermionic species accounted for α T −1 T −1 should indeed correspond to: L = yψlαφ + mψψ C ψ + mχχ C χ 1. a regular neutrino, which would correspond in this framework to the light χ-fermion species, the mass ξ µ −3 + 2 (ψγ5γµψ)(χγ γ5χ) + h.c. , of which would be then mχ < 10 eV , fulfilling in MP l this way the constraints from Big-Bang nucleosyn- 2 −1 thesis and being consistent with all the experimen- in which we have reinserted MP l = κ . The decay chan- tal data [101]; nels of the heavy fermion to the SM particles are

2. a sterile neutrino, which would be driving the pri- ψ → lφ, ψ → ¯lφ¯ , mordial spacetime background dynamics, namely the ψ species. Its mass would be at the GUT in which l are lepton fields and φ denotes the Higgs, y is scale for a choice of the ratio that appears in (45). the Yukawa matrix of ψ and l generations. In particular, Nonetheless, even smaller values than the GUT we will assume that the number of ψ-generations will be scale would be consistent in this model for the mass more than one. Such a Lagrangian mediates L-violating of the background species ψ. channels with several different L-number assignment for ψ. For instance, if L : ψ = 0, the first Yukawa term We end this section with a comment on the perturba- violates the L-number of ∆L = 1. However, such an tion theory in fermion fields cosmologies. For this type assignment for the lepton number of ψ is problematic, of cosmologies, the stress energy tensor behaves as a per- since ψ could also couple to quarks, destabilizing the fect fluid only at the background level. But if we con- proton. A strong fine-tuning for the coupling of ψ to sider perturbations, anisotropic components may appear the quarks should be then invoked. The most elegant in the stress-energy tensor. These latter would not affect choice is to consider a L-preserving Yukawa term, while our conclusions on the curvature perturbations generated the L-number is violated by the Majorana mass term. in the primordial era, but may affect the propagation of This choice coincides with L : ψ = 1, as for RH neu- primordial gravitational waves, and thus the estimate of trino. On the other hand, the CP violation is encoded in the scalar to tensor ratio. This is a very interesting topic the complex phases of the Yukawa coupling y, which is a which deserves a further investigation, especially in light matrix in the space of leptonic generations of ψ. of the considerations in [41]. Below we will then consider the case of a negligible four fermion interaction among the extra massive fermion and the light one. This imposes a bound on the suppression −2 VI. BARYOGENESIS IN THE CURVATON scale ξMP l . FERMI-BOUNCE SCHEME The decay width of the heavy neutrino ψ at tree level is In this section, we discuss a scenario in which baryo- ¯¯ 1 † genesis is realized from the decay of the ultra-massive ΓD = Γ(ψ → φl) + Γ(ψ → φl) = y y mψ , (46) 8π fermions, described by the ψ-species, into the SM particles, corresponding to the χ-species. Thus ψ is now while the annihilation cross section σ(ψψ → χχ) is as- defined as a singlet of the SM gauge group. Conse- sumed to be subdominant, for the moment. Once the quently, this ultra massive field is not protected by the temperature of the universe drops below the critical value electroweak symmetry, and for the Georgi’s missing sin- T¯ = mψ, the heavy neutrinos cannot follow the rapid glet mechanism, it should have a mass much higher than variation of the equilibrium distribution. The deviation the SM vacuum expectation value (VEV) scale [104]. We from thermal equilibrium is related to a too large number will not only show that the correct Baryon asymmetry is density of heavy fermions, compared to the equilibrium reproduced, but we also set sever limits on the bounce density. However, the heavy fermion decay, and a lepton scale. In Fig.1, we obtain a number density asymmetry asymmetry, could be generated owing to the presence |nB−L| which is compatible with the baryogenesis consis- of CP-violating transitions. CP-violating transitions in- −10 tency condition nB/s ∼ 10 . volves the quantum interference between the tree-level

9 amplitude and the one-loop diagrams (vertex and self- where H is the Hubble parameter, dΦ1,..,n denotes the energy contributions). The resulting CP violation pa- phase space integration over particles in the initial and rameters can be conveniently deﬁned as ﬁnal states,

3 3 Γ(ψ → lφ) − Γ(ψ → ¯lφ¯) d p1 d pn 4 4 Da ≡ . (47) dΦ1,..,n = 3 ... 3 (2π) δ (p1 + ... − pn) Γ(ψ → lφ) + Γ(ψ → ¯lφ¯) (2π) 2E1 (2π) 2En (56) At tree-level, this parameter is equal to zero. However, and the weights the CP asymmetry appears at the leading order of pertur- Z d3p bation theory y2 (1-loop), entailing the two contributions f (p) = exp(−βE (p)), n (p) = g f (p) (57) i i i i (2π)3 i 1 Im(y†y)2 M ik represent the Boltzmann distribution -for simplicity we a = − , (3level) + (self energy) , (48) 8π (y†y) use Boltzmann distribution rather than Bose-Einstein and Fermi-Dirac distributions, neglecting the distribu- where i, k are ψ-generation indices, tion functions in the final state, which is a good approxi- 1 m Im(y†y)2 mation for small number densities) and the number den- V = − f ψi ik , (49) sity of particle i = N, l, φ at temperature T = 1 , respec- a 8π m (y†y) β ψi tively, M and M0 denote the scattering matrix elements and of the indicated processes at T = 0 (the prime indicated that in 2 → 2 scatterings the contribution of the internal √ 1 + x f(x) = x 1 − (1 + x)ln (50) resonance state has been subtracted). x The ratio of number density and entropy density (namely YX = nX /s) remains constant for an expanding These results hold for mψ >> |Γ| -for small mass dif- universe in thermal equilibrium. The heavy fermions are ferences, we may expect an enhancement of the mixing weakly coupled to the thermal bath. So that they freeze- contribution. out of the thermal equilibrium at T¯ = mψ. This implies that the decay rate is too slow to follow the rapidly de- creasing equilibrium distribution fN ∼ exp(−βmψ). As B. The Boltzmann equations a consequence, the system will evolve toward an excess eq of the number density nN > nN . The generation of a baryon/lepton asymmetry can be As is known, the Boltzmann equations are classical treated with Boltzmann equations. While accounting dynamical equations. However, they are endowed with S- for these latter, we have to consider the main processes, matrix elements, in turn containing quantum mechanical which are the decays and inverse decays of the heavy interferences of different amplitudes. The S-matrices are fermions, and the lepton number conserving ∆L = 0 and contained in collisions terms in the Boltzmann equations. violating ∆L = 2 processes. These processes read, re- The scattering matrix elements are evaluated at zero spectively, temperature. However, the quantum mechanical interferences must be affected by interactions with the thermal dn ψ + 3Hn =−γ(ψ → lφ) + γ(lφ → ψ) (51) bath. For 2 → 2 scatterings, one finds dt ψ ¯¯ 2 0 ¯¯ 2 ¯¯ 2 −γ(ψ → ¯lφ¯) − γ(¯lφ¯ → ψ) , |M(lφ → lφ)| = |M (lφ → lφ)| + |Mres(lφ → lφ)| , (58) where the resonance contribution reads dnl + 3Hnl =γ(ψ → lφ) − γ(lφ → ψ) (52) dt M(lφ → ¯lφ¯) ∼ M(lφ → ψ)M(ψ → ¯lφ¯)∗ = |M(lφ → ψ)|2 . +γ(¯lφ¯ → lφ) − γ(lφ → ¯lφ¯) , (59) The particles involved in the process may be treated as massless ones: their distribution functions will coincide dn¯l + 3Hn¯ =γ(ψ → ¯lφ¯) − γ(¯lφ¯ → ψ) (53) with the equilibrium distribution. On the other hand, dt l resonances may be treated as on-shell particles, falling +γ(lφ → ¯lφ¯) − γ(¯lφ¯ → lφ) , out of thermal equilibrium. Because of CPT invariance, with the reaction densities expressed as it follows for thermal equilibrium distributions we must have

Z d(nl − n¯l) eq 2 + 3H(n − n¯) = ∆γ = 0 . (60) γ(ψ → lφ) = dΦ123fN (p1)|M(ψ → lφ)| , (54) dt l l and The resonance contributions (the decay and inverse decay) Z ¯¯ 0 2 γ(lφ → lφ) = dΦ1234fl(p1)fφ(p2)|M (ψ → lφ)| , (55) eq eq ∆γres = −2γ (ψ → lφ) ,

10 have a compensating term from 2 → 2 contribution processes Z eq eq ∆γ2→2 = 2 dΦ1234fl (p1)fφ (p2) ×(|M0(lφ → ¯lφ¯)|2 − |M0(¯lφ¯ → lφ)|2) , (61) Such a lφ ↔ ¯lφ¯ compensation is a consequence of CPT symmetry/unitarity. From unitarity, one can find X (|M(lφ → X)|2 − |M(X → lφ)|2) = 0 , X where X denotes all possible generic channels. In the case of weak coupling y, the channel is restricted to 2-particles states to the leading order of perturbation expansion. Therefore, at the leading order y2, one obtains FIG. 1. The evolution of the number densities as functions Z eq eq eq of z = mψ/T are displayed (log10 n, log10 z) scale. In dashed ∆2→2 =2 dΦ1234fl (p1)fφ (p2) (62) black lines, we show the evolution of the number density of massive fermions ψ starting from a thermal initial density 2 ¯¯ 2 × − |M(lφ → ψ)| |M(ψ → lφ)| nψ(z << 1) ' 3/4. In thick black lines, we plot the evolution of the number density of massive fermions ψ starting from a 2 2 π 2 +|M(¯lφ¯ → ψ)| |M(ψ → lφ)| δ(s − m ) thermal initial density nψ(z <<1) ' 0. In the red dashed and m Γ ψ ψ thick red lines the B − L number density |nB−L| is pictured eq eq = 2γ (ψ → lφ) = −∆γ . from initial thermal and zero nψ respectively. The charac- res 10 teristic parameters chosen in this plot are mψ ' 10 GeV, The incorporation of off-shell effects requires a formalism −6 −3 −6 −8 ' 10 , mχ ≡ mν1 ' 10 eV, ξ ≤ 10 ÷ 10 . that goes beyond the Boltzmann equations (Kadanoff- Baym equations [109]). However, the corrections are expected to be negligible from Boltzmann’s model. However, the plasma of the early Universe is very dif- The numerical integration of the Boltzmann equations ferent from a weakly coupled relativistic gas. This be- for reasonable parameters is displayed in Fig.1. In par- cause of unscreened non-abelian gauge interactions which ticular, the number density of ψ-particle decreases in cos- have very important nonperturbative effects to take into mological time, generating a small Lepton number asym- account. metry. Leptons, Quarks and Higgs will interact via perturbative operators Yukawa and gauge couplings. However, they will also interact via the nonperturbative sphaleron C. Baryon number asymmetry from sphalerons processes. These processes lead to a set of constraints between the SM chemical potentials. The effective inter- Let us assign a chemical potential µ to each SM particles: actions induced by sphalerons (SU(2) electro-weak in- quarks, Higgs, and leptons. In the SM -with Nf genera- stantons) must imply 1 tions and one Higgs doublet - we must assign 5Nf + 1 chemical potentials. The asymmetry in the particle and X (3µqi + µui − µdi) = 0 . (64) antiparticle number densities, for βµi << 1, reads i gT 3 n − n¯ = βµ + O[(βµ )3] (for fermions) , (63) i i 6 i i On the other hand, SU(3) QCD instanton-mediated processes must generate effective interactions between LH and and RH quarks, which leads to constrains gT 3 n − n¯ = 2βµ + O[(βµ )3] (for bosons) . X i i 6 i i (2µqi − µui − µdi) = 0 . (65) i having considered the thermal bath as a non-interacting gas of massless particles. A third condition (valid in all the range of temperatures) arises from the requirement that the total hypercharge of the plasma must vanish. So that, from (63) and the SM hypercharges we obtain 1 Let us note that, in addition to the Higgs doublet, the two left- handed doublets q and l and the three right-handed singlets u , i i i 2 di and ei of each generation, will be treated with independent X (µqi + 2µui − µdi − µli − µei + µφ) = 0 . (66) chemical potential (each one). N i f

11 3 2 Finally, from the Yukawa couplings, one finds, in turn estimated at s < ξMP l. This process becomes out-of- 2 2 equilibrium for s ' T ≤ 4mψ, which means that χχ µqi −µφ −µdj = 0, µqi +µφ −µuj = 0, µli −µφ −µej = 0 . collisions can not reproduce two heavy ψ particles. (67) In the scenario we deepened here, we imposed the hi- −1/2 These relations are valid if and only if the system is in erarchy mψ < χ MP l, i.e. the bounce scale must be thermal equilibrium2, at temperatures 100 GeV < T < higher than the heavy fermion mass. On the other hand, 1012 GeV. a successful leptogenesis requests a fermion mass scale 9 Let us define the Baryon number density nB = of mψ ∼ 10 GeV or similar. This imposes an indirect 2 2 gBT /6 and the lepton number densities nli = LigT /6. bound on the bounce scale. In particular, from numeri- So that, we express Baryon and Lepton numbers B and cal calculations, we checked that the safety bound for a −1/2 Li in terms of the chemical potentials, i.e. good leptogenesis is ξ MP l ≥ (10 ÷ 15)mψ, assuming −1 X natural initial conditions nψ(z < 10 ) ' 0, 3/4) respec- B = (2µqi + µui + µdi) , (68) tively. From these values, the same plots displayed in i Fig.1 are obtained. On the other hand, the case in which mψ ≥ (10 ÷ X 15)ξ−1/2M cannot correspond to a successful lepto- L = 2µ + µ ,L = L . P l i li ei i genesis, without an unnatural initial superabundance i of heavy fermions. In particular, the case of mψ ' −1/2 Solving this simple system of algebric equations, with ξ MP l is undesired, first of all because the unitarity respect to µl, one obtains and the calculability of z < 1 in Fig.1 can not be con- trolled, and second because the annihilation process will 2 4Nf 14Nf + 9Nf dominate over all other channels around z ' 1, provoking B = − µl,L = µl , (69) 3 6Nf + 3 a too fast number density decay of ψ. They yield to the important connections between the B, B-L and L asymmetries: VII. DISCUSSION

B = a(B − L); L = (a − 1)(B − L) , (70) In this paper we have shown that the matter bounce scenario, as an alternative framework to inflation, allows to where a = (8N + 4)/(22N + 13). f f encode dark matter in a pretty natural way when fermion matter fields are taken into account. We have further D. Four fermion interaction between the two shown that the model is able to generate leptogenesis, species thus to explain baryogenesis. Specifically, we have fo- cused on a toy-model for the curvaton mechanism, which is an instantiation of a Fermi-bounce cosmology that sin- We wish now to comment on the possible relevance of the gles out as a dark matter candidate a sterile neutrino-like torsion mediated four fermion interactions, namely field, hence deriving phenomenological consequences of ξ these assumptions. It is quite remarkable that this sce- O = ψγ γ ψ χγµγ χ . (71) Γ M 2 µ 5 5 nario comes out to be falsifiable, since it predicts a non P l vanishing value of r. If χ is weakly interacting with the SM particles, ψψ → χχ A peculiarity of the model we have shown here is that represent entropy leaking collisions that could affect the the results we have derived are not confined to an ef- −1 leptogenesis scenario if mψ ∼ ξ MP l. In particular, if fective analysis in which the dynamics of dark matter is the fermion ψ has a lepton number assignment L = 0, the only addressed from a hydrodynamical perspective. Con- OΓ operator violates the lepton number as ∆L = 2. For versely, the Fermi matter bounce scenario described here example, the initial number density of nψ can be affected encodes microphysical description in terms of fermionic by the annihilation process fields, which provides a natural way to overcome short- comings that usually arise in cosmology because of the ξs use of auxiliary scalar fields, with consequent issues of σ(ψψ → χχ) ∼ 2 , 8πMP l arbitrariness as it happens in inflation. It might be indeed look surprising at a first sight that matter fields belonging to the SM and to its sim-

2 Let us note that Yukawa interactions are in equilibrium only within a more restricted temperature window depending on the 3 2 strength/magnitude of the Yukawa couplings. In the following We notice that for s > ξMP l the ultra-violet (UV) completion discussions, we will ignore these technical complications. In fact, of such a term is not yet understood. However, our analysis only a small eﬀect on our discussion of Leptogenesis is expected concerns energies which are supposed to be lower than UV cutoﬀ from taking them into account. scale, where unitarity issues are not expected.

12 plest extension reproduce desired background and per- ACKNOWLEDGMENTS turbation features. But it should be not surprising that in this framework retrieved from particle physics it is possible to tackle questions that concern the nature of The work of AA was supported in part by the MIUR dark matter and the origin of baryogenesis. Indeed the research grant Theoretical Astroparticle Physics PRIN model we have deepened here turns out to be compatible 2012CPPYP7 and by SdC Progetto speciale Multiasse La with hot dark matter constrains. Several issues continue Societ`adella Conoscenza in Abruzzo PO FSE Abruzzo to be unexplored, and we can not deny that this line 2007-2013. The work of YFC is supported in part by of research will require in the future more detailed in- the Chinese National Youth Thousand Talents Program vestigations. Nonetheless we wish to mention that the (Grant No. KJ2030220006), by the USTC start-up fund- Fermi bounce scenario might entail as distinctive phe- ing (Grant No. KY2030000049), by the NSFC (Grant nomenological predictions, able to falsify this paradigm No. 11421303 and No. 11653002), and by the Fund for among the others in the literature, the appearance of Fostering Talents in Basic Science of the NSFC (Grant non-vanishing cross-correlation functions between polar- No. J1310021). AM wishes to acknowledge support ization modes, sourced by the parity-violating elements by the Shanghai Municipality, through the grant No. of the Cliﬀord algebra bilinears [41]. But more work is KBH1512299, and by Fudan University, through the still also required even in this direction. grant No. JJH1512105.

[1] S. Alexander, C. Bambi, A. Marcianò and [13] Y. F. Cai, F. Duplessis, D. A. Easson and D. G. Wang, L. Modesto, Phys. Rev. D 90, no. 12, 123510 (2014) “Searching for a matter bounce cosmology with low red- doi:10.1103/PhysRevD.90.123510 [arXiv:1402.5880 shift observations,” Phys. Rev. D 93, no. 4, 043546 [gr-qc]]. (2016) [arXiv:1512.08979 [astro-ph.CO]]. [2] P. Donà, A. Marcianò, Y. Zhang and C. An- [14] J. de Haro and Y. F. Cai, “An Extended Matter Bounce tolini, Phys. Rev. D 93, no. 4, 043012 (2016) Scenario: current status and challenges,” Gen. Rel. doi:10.1103/PhysRevD.93.043012 [arXiv:1509.05824 Grav. 47, no. 8, 95 (2015) [arXiv:1502.03230 [gr-qc]]. [gr-qc]]. [15] S. D. Odintsov and V. K. Oikonomou, “ΛCDM [3] A. Addazi, P. Donàand A. Marcianò,arXiv:1602.01772 Bounce Cosmology without ΛCDM: the case of mod- [gr-qc]. ified gravity,” Phys. Rev. D 91, no. 6, 064036 (2015) [4] S. Alexander, A. Marcianò and Z. Yang, [arXiv:1502.06125 [gr-qc]]. arXiv:1602.06557 [hep-th]. [16] J. L. Lehners and E. Wilson-Ewing, “Running of the [5] A. Addazi, A. Marcianò and S. Alexander, scalar spectral index in bouncing cosmologies,” JCAP arXiv:1603.01853 [gr-qc]. 1510, no. 10, 038 (2015) [arXiv:1507.08112 [astro- [6] S. Dubovsky, T. Gregoire, A. Nicolis and R. Rattazzi, ph.CO]]. JHEP 0603, 025 (2006) [hep-th/0512260]. [17] P. C. Ferreira and D. Pavon, “Thermodynamics of non- [7] V. A. Rubakov, Phys. Rev. D 88 (2013) 044015 singular bouncing universes,” Eur. Phys. J. C 76, no. 1, doi:10.1103/PhysRevD.88.044015 [arXiv:1305.2614 37 (2016) [arXiv:1509.03725 [gr-qc]]. [hep-th]]. [18] R. H. Brandenberger, Y. F. Cai, S. R. Das, E. G. M. Fer- [8] B. Elder, A. Joyce and J. Khoury, Phys. Rev. D 89 reira, I. A. Morrison and Y. Wang, “Fluctuations in (2014) no.4, 044027 doi:10.1103/PhysRevD.89.044027 a Cosmology with a Space-Like Singularity and their [arXiv:1311.5889 [hep-th]]. Gauge Theory Dual Description,” arXiv:1601.00231 [9] V. A. Rubakov, Phys. Usp. 57 (2014) 128 [hep-th]. [Usp. Fiz. Nauk 184 (2014) no.2, 137] [19] S. Nojiri, S. D. Odintsov and V. K. Oikonomou, doi:10.3367/UFNe.0184.201402b.0137 [arXiv:1401.4024 “Bounce universe history from unimodular F (R) [hep-th]]. gravity,” Phys. Rev. D 93, no. 8, 084050 (2016) [10] A. Addazi and A. Marcianò, in preparation. [arXiv:1601.04112 [gr-qc]]. [11] F. W. Hehl, Phys. Lett. A 36, 225 (1971); F. W. [20] S. D. Odintsov and V. K. Oikonomou, “Deformed Mat- Hehl, Gen. Relativ. Gravit. 4, 333 (1973); F. W. Hehl, ter Bounce with Dark Energy Epoch,” Phys. Rev. D 94, Gen. Relativ. Gravit. 5, 491 (1974); P. von der Heyde, no. 6, 064022 (2016) [arXiv:1606.03689 [gr-qc]]. G. D. Kerlick and J. M. Nester, Rev. Mod. Phys. 48, [21] Y. F. Cai, A. Marciano, D. G. Wang and E. Wilson- 393 (1976); E. A. Lord, Tensors, Relativity and Cosmol- Ewing, “Bouncing cosmologies with dark matter and ogy (McGraw-Hill, New Delhi, 1976); V. de Sabbata and dark energy,” arXiv:1610.00938 [astro-ph.CO]. M. Gasperini, Introduction to Gravitation (World Scien- [22] A. Trautman, Bull. Acad. Pol. Sci. Ser. Math. Astron. tific, Singapore, 1985); V. de Sabbata and C. Sivaram, Phys. “On the Einstein-Cartan equations”, 20 (1972) Spin and Torsion in Gravitation (World Scientific, Sin- 185, 503, 895. gapore, 1994); I. L. Shapiro, Phys. Rep. 357, 113 (2002); [23] P. van Nieuweinhuizen, Physics Reports 68 (1981) 189. R. T. Hammond, Rep. Prog. Phys. 65, 599 (2002). [24] S. W. Hawking and G. F. R. Ellis, “The Large scale [12] Y. F. Cai, R. Brandenberger and X. Zhang, JCAP structure of space-time,” Cambridge University Press, 1103, 003 (2011) doi:10.1088/1475-7516/2011/03/003 Cambridge, 1973. [arXiv:1101.0822 [hep-th]].

13 [25] J. Khoury, B. A. Ovrut, P. J. Steinhardt and N. Turok, [44] Planck, P. Ade et al., “Planck 2015 results. XIII. Cos- “The ekpyrotic universe: Colliding branes and the origin mological parameters,” arXiv:1502.01589. of the hot big bang,” Phys. Rev. D 64, 123522 (2001), [45] BICEP2, Planck, P. Ade et al., “Joint Analysis of arXiv:hep-th/0103239. BICEP2/KeckArray and P lanck Data,” Phys. Rev. [26] R. H. Brandenberger and C. Vafa, ”Superstrings in Lett. 114 (2015) 101301, arXiv:1502.00612. the Early Universe,” Nucl. Phys. B 316, 391 (1989); [46] J. Quintin, Z. Sherkatghanad, Y.-F. Cai, and R. H. S. Alexander, R. H. Brandenberger and D. A. Eas- Brandenberger, “Evolution of cosmological perturba- son, “Brane gases in the early universe,” Phys. Rev. tions and the production of non-Gaussianities through D 62, 103509 (2000); arXiv:hep-th/0005212; A. Nay- a nonsingular bounce: Indications for a no-go theorem eri, R. H. Brandenberger and C. Vafa, “Producing a in single field matter bounce cosmologies,” Phys. Rev. scale-invariant spectrum of perturbations in a Hagedorn D92 (2015) 063532, arXiv:1508.04141. phase of string cosmology,” Phys. Rev. Lett. 97, 021302 [47] E. Wilson-Ewing, “The Matter Bounce Scenario in (2006), arXiv:hep-th/0511140. Loop Quantum Cosmology,” JCAP 1303 (2013) 026, [27] F. Finelli and R. Brandenberger, “On the generation arXiv:1211.6269. of a scale invariant spectrum of adiabatic fluctuations [48] Y.-F. Cai and E. Wilson-Ewing, “A ΛCDM bounce sce- in cosmological models with a contracting phase,” Phys. nario,” JCAP 1503 (2015) 006, arXiv:1412.2914. Rev. D 65, 103522 (2002), arXiv:hep-th/0112249. [49] E. Wilson-Ewing, “Separate universes in loop quantum [28] D. Wands, “Duality invariance of cosmological perturba- cosmology: framework and applications,” to appear in tion spectra,” Phys. Rev. D 60, 023507 (1999), arXiv:gr- Int. J. Mod. Phys. D, arXiv:1512.05743. qc/9809062. [50] V. F. Mukhanov, H. Feldman, and R. H. Brandenberger, [29] C. Bambi, D. Malafarina, A. Marcianòand L. Modesto, “Theory of cosmological perturbations,” Phys. Rept. “Singularity avoidance in classical gravity from four- 215 (1992) 203–333. fermion interaction,” Phys. Lett. B 734, 27 (2014), [51] V. Mukhanov, Physical Foundations of Cosmology. arXiv:1402.5719 [gr-qc]. Cambridge University Press, Cambridge, 2005. [30] A. Perez and C. Rovelli, “Physical effects of the Immirzi [52] G. Calcagni, “Cosmology of the Lifshitz universe,” parameter,” Phys. Rev. D 73, 044013 (2006), arXiv:gr- JHEP 0909, 112 (2009) [arXiv:0904.0829 [hep-th]]. qc/0505081. [53] E. Kiritsis and G. Kofinas, “Horava-Lifshitz Cosmol- [31] S. Alexander, T. Biswas and G. Calcagni, “Cosmological ogy,” Nucl. Phys. B 821, 467 (2009) [arXiv:0904.1334 Bardeen-Cooper-Schrieffer condensate as dark energy,” [hep-th]]. Phys. Rev. D 81, 043511 (2010) [Erratum-ibid. D 81, [54] R. Brandenberger, “Matter Bounce in Horava-Lifshitz 069902 (2010)], arXiv:0906.5161 [astro-ph.CO]. Cosmology,” Phys. Rev. D 80, 043516 (2009) [32] L. Freidel, D. Minic and T. Takeuchi, “Quantum gravity, [arXiv:0904.2835 [hep-th]]. torsion, parity violation and all that,” Phys. Rev. D 72, [55] Y.-F. Cai and E. N. Saridakis, “Non-singular cosmology 104002 (2005), arXiv:hep-th/0507253. in a model of non-relativistic gravity,” JCAP 0910, 020 [33] C. Armendariz-Picon and P. B. Greene, “Spinors, (2009) [arXiv:0906.1789 [hep-th]]. inflation, and nonsingular cyclic cosmologies,” Gen. [56] Y.-F. Cai, S.-H. Chen, J. B. Dent, S. Dutta and Rel. Grav. 35, 1637 (2003), hep-th/0301129, hep- E. N. Saridakis, “Matter Bounce Cosmology with the th/0301129. f(T) Gravity,” Class. Quant. Grav. 28, 215011 (2011) [34] N. J. Poplawski, “Nonsingular, big-bounce cosmology [arXiv:1104.4349 [astro-ph.CO]]. from spinor-torsion coupling,” Phys. Rev. D 85, 107502 [57] N. J. Poplawski, “Nonsingular, big-bounce cosmology (2012), arXiv:1111.4595 [gr-qc]. from spinor-torsion coupling,” Phys. Rev. D 85, 107502 [35] M. Bojowald and R. Das, “Canonical gravity (2012) [arXiv:1111.4595 [gr-qc]]. with fermions,” Phys. Rev. D 78, 064009 (2008), [58] Y.-F. Cai and E. N. Saridakis, “Cyclic cosmology from arXiv:0710.5722 [gr-qc]. Lagrange-multiplier modified gravity,” Class. Quant. [36] M. Bojowald and R. Das, “Fermions in Loop Quantum Grav. 28, 035010 (2011) [arXiv:1007.3204 [astro- Cosmology and the Role of Parity,” Class. Quant. Grav. ph.CO]]. 25, 195006 (2008), arXiv:0806.2821 [gr-qc]. [59] Y.-F. Cai and E. N. Saridakis, “Non-singular Cyclic [37] S. Mercuri, “Fermions in Ashtekar-Barbero connec- Cosmology without Phantom Menace,” J. Cosmol. 17, tions formalism for arbitrary values of the Immirzi pa- 7238 (2011) [arXiv:1108.6052 [gr-qc]]. rameter,” Phys. Rev. D 73, 084016 (2006), arXiv:gr- [60] Y.-F. Cai, C. Gao and E. N. Saridakis, “Bounce and qc/0601013. cyclic cosmology in extended nonlinear massive grav- [38] S. Alexander and T. Biswas, “The Cosmological BCS ity,” JCAP 1210, 048 (2012) [arXiv:1207.3786 [astro- mechanism and the Big Bang Singularity,” Phys. Rev. ph.CO]]. D 80, 023501 (2009), arXiv:0807.4468 [hep-th]. [61] R. Durrer, M. Gasperini, M. Sakellariadou and [39] J. Magueijo, T. G. Zlosnik and T. W. B. Kibble, “Cos- G. Veneziano, Phys. Rev. D 59 (1999) 043511 mology with a spin,” Phys. Rev. D 87, 6, 063504 (2013), doi:10.1103/PhysRevD.59.043511 [gr-qc/9804076]. arXiv:1212.0585 [astro-ph.CO]. [62] T. Biswas, A. Mazumdar and W. Siegel, “Bouncing [40] P. Donà,A. Marcianòand X. Gan, in preparation. universes in string-inspired gravity,” JCAP 0603, 009 [41] P. Donàand A. Marcianò,arXiv:1605.09337 [gr-qc]. (2006) [hep-th/0508194]. [42] R. H. Brandenberger, “The Matter Bounce Alternative [63] T. Biswas, R. Brandenberger, A. Mazumdar and to Inflationary Cosmology,” arXiv:1206.4196. W. Siegel, “Non-perturbative Gravity, Hagedorn [43] R. Brandenberger and P. Peter, arXiv:1603.05834 [hep- Bounce & CMB,” JCAP 0712, 011 (2007) [hep- th]. th/0610274].

14 [64] Y.-F. Cai, T. Qiu, Y. -S. Piao, M. Li and X. Zhang, [84] J. K. Erickson, D. H. Wesley, P. J. Steinhardt and “Bouncing universe with quintom matter,” JHEP 0710, N. Turok, “Kasner and mixmaster behavior in universes 071 (2007) [arXiv:0704.1090 [gr-qc]]. with equation of state w ≥ 1,” Phys. Rev. D 69, 063514 [65] Y.-F. Cai, T. Qiu, R. Brandenberger, Y.-S. Piao and (2004) [hep-th/0312009]. X. Zhang, “On Perturbations of Quintom Bounce,” [85] Y.-F. Cai, D. A. Easson and R. Brandenberger, “To- JCAP 0803, 013 (2008) [arXiv:0711.2187 [hep-th]]. wards a Nonsingular Bouncing Cosmology,” JCAP [66] Y.-F. Cai, T. -t. Qiu, R. Brandenberger and X.- 1208, 020 (2012) [arXiv:1206.2382 [hep-th]]. m. Zhang, “A Nonsingular Cosmology with a Scale- [86] Y.-F. Cai, R. Brandenberger and P. Peter, “Anisotropy Invariant Spectrum of Cosmological Perturbations from in a Nonsingular Bounce,” Class. Quant. Grav. 30, Lee-Wick Theory,” Phys. Rev. D 80, 023511 (2009) 075019 (2013) [arXiv:1301.4703 [gr-qc]]. [arXiv:0810.4677 [hep-th]]. [87] Y.-F. Cai, E. McDonough, F. Duplessis and R. H. Bran- [67] K. Bhattacharya, Y.-F. Cai and S. Das, “Lee-Wick ra- denberger, “Two Field Matter Bounce Cosmology,” diation induced bouncing universe models,” Phys. Rev. JCAP 1310, 024 (2013) [arXiv:1305.5259 [hep-th]]. D 87, no. 8, 083511 (2013) [arXiv:1301.0661 [hep-th]]. [88] M. Koehn, J.-L. Lehners and B. A. Ovrut, “A Cosmo- [68] E. I. Buchbinder, J. Khoury and B. A. Ovrut, “New logical Super-Bounce,” arXiv:1310.7577 [hep-th]. Ekpyrotic cosmology,” Phys. Rev. D 76, 123503 (2007) [89] Y.-F. Cai and E. Wilson-Ewing, “Non-singular bounce [hep-th/0702154]. scenarios in loop quantum cosmology and the ef- [69] P. Creminelli and L. Senatore, “A Smooth bouncing cos- fective field description,” JCAP 1403, 026 (2014) mology with scale invariant spectrum,” JCAP 0711, [arXiv:1402.3009 [gr-qc]]. 010 (2007) [hep-th/0702165]. [90] P. A. R. Ade et al. [Planck Collaboration], “Planck [70] C. Lin, R. H. Brandenberger and L. Perreault Levasseur, 2013 results. XVI. Cosmological parameters,” “A Matter Bounce By Means of Ghost Condensation,” arXiv:1303.5076 [astro-ph.CO]. JCAP 1104, 019 (2011) [arXiv:1007.2654 [hep-th]]. [91] P. A. R. Ade et al. [BICEP2 Collaboration], “BICEP2 [71] R. H. Brandenberger, C. Kounnas, H. Partouche, I: Detection Of B-mode Polarization at Degree Angular S. P. Patil and N. Toumbas, “Cosmological Pertur- Scales,” arXiv:1403.3985 [astro-ph.CO]. bations Across an S-brane,” JCAP 1403, 015 (2014) [92] M. J. Mortonson and U. Seljak, arXiv:1405.5857 [astro- [arXiv:1312.2524 [hep-th]]. ph.CO]. [72] J. Martin and P. Peter, “Parametric amplification of [93] R. Flauger, J. C. Hill and D. N. Spergel, metric fluctuations through a bouncing phase,” Phys. arXiv:1405.7351 [astro-ph.CO]. Rev. D 68, 103517 (2003) [hep-th/0307077]. [94] Y. -F. Cai, J. Quintin, E. N. Saridakis and E. Wilson- [73] D. M. Solomons, P. Dunsby and G. Ellis, “Bounce be- Ewing, “Nonsingular bouncing cosmologies in light of haviour in Kantowski-Sachs and Bianchi cosmologies,” BICEP2,” arXiv:1404.4364 [astro-ph.CO]. Class. Quant. Grav. 23, 6585 (2006) [gr-qc/0103087]. [95] C. Armendariz-Picon and P. B. Greene, “Spinors, [74] M. Novello and S. E. P. Bergliaffa, “Bouncing Cosmolo- inflation, and nonsingular cyclic cosmologies,” Gen. gies,” Phys. Rept. 463, 127 (2008) [arXiv:0802.1634 Rel. Grav. 35, 1637 (2003) [hep-th/0301129, hep- [astro-ph]]. th/0301129]. [75] J.-L. Lehners, “Ekpyrotic and Cyclic Cosmology,” Phys. [96] J. Magueijo, T. G. Zlosnik and T. W. B. Kibble, “Cos- Rept. 465, 223 (2008) [arXiv:0806.1245 [astro-ph]]. mology with a spin,” Phys. Rev. D 87, 6, 063504 (2013), [76] Y.-F. Cai, “Exploring Bouncing Cosmologies with Cos- arXiv:1212.0585 [astro-ph.CO]. mological Surveys,” arXiv:1405.1369 [hep-th]. [97] Y.-F. Cai and J. Wang, “Dark Energy Model with [77] S. Alexander, C. Bambi, A. Marciano and Spinor Matter and Its Quintom Scenario,” Class. Quant. L. Modesto, Phys. Rev. D 90, no. 12, 123510 (2014) Grav. 25, 165014 (2008) [arXiv:0806.3890 [hep-th]]. doi:10.1103/PhysRevD.90.123510 [arXiv:1402.5880 [98] S. Alexander, T. Biswas and G. Calcagni, “Cosmological [gr-qc]]. Bardeen-Cooper-Schrieffer condensate as dark energy,” [78] S. Alexander, Y. F. Cai and A. Marciano, Phys. Lett. Phys. Rev. D 81, 043511 (2010) [Erratum-ibid. D 81, B 745, 97 (2015) doi:10.1016/j.physletb.2015.04.026 069902 (2010)] [arXiv:0906.5161 [astro-ph.CO]]. [arXiv:1406.1456 [gr-qc]]. [99] Y.-F. Cai, M. Li and X. Zhang, “Emergent Universe [79] S. Alexander and T. Biswas, “The Cosmological BCS Scenario via Quintom Matter,” Phys. Lett. B 718, 248 mechanism and the Big Bang Singularity,” Phys. Rev. (2012) [arXiv:1209.3437 [hep-th]]. D 80, 023501 (2009) [arXiv:0807.4468 [hep-th]]. [100] Y.-F. Cai, Y. Wan and X. Zhang, “Cosmology of [80] D. Wands, “Duality invariance of cosmological pertur- the Spinor Emergent Universe and Scale-invariant bation spectra,” Phys. Rev. D 60, 023507 (1999) [gr- Perturbations,” Phys. Lett. B 731, 217 (2014) qc/9809062]. [arXiv:1312.0740 [hep-th]]. [81] F. Finelli and R. Brandenberger, “On the generation of [101] R. A. Battye and A. Moss, Phys. Rev. Lett. 112, 051303 a scale invariant spectrum of adiabatic fluctuations in (2014) [arXiv:1308.5870 [astro-ph.CO]]. cosmological models with a contracting phase,” Phys. M. Fukugita, T. Yanagida, Phys.Lett. B 174 45 (1986). Rev. D 65, 103522 (2002) [hep-th/0112249]. A. D. Sakharov, JETP Lett. 5, 24 (1967). [82] V. A. Belinsky, I. M. Khalatnikov and E. M. Lifshitz, V. A. Kuzmin, V. A. Rubakov, N. E. Shaposhnikov, “Oscillatory approach to a singular point in the rela- Phys. Lett.155B 36 (1985). tivistic cosmology,” Adv. Phys. 19, 525 (1970). L. P. Kadanoff, G. Baym, Quantum Statistical Mechan- [83] J. Khoury, B. A. Ovrut, P. J. Steinhardt and N. Turok, ics (Benjamin, New York,1962). “The Ekpyrotic universe: Colliding branes and the ori- [102] S. D. Odintsov and V. K. Oikonomou, “Matter Bounce gin of the hot big bang,” Phys. Rev. D 64, 123522 (2001) Loop Quantum Cosmology from F (R) Gravity,” Phys. [hep-th/0103239]. Rev. D 90, no. 12, 124083 (2014) [arXiv:1410.8183 [gr-

15 qc]]. [106] W. Buchmuller, R. D. Peccei and T. Yanagida, [103] K. Bamba, A. N. Makarenko, A. N. Myagky and Ann. Rev. Nucl. Part. Sci. 55 (2005) 311 S. D. Odintsov, “Bouncing cosmology in modiﬁed doi:10.1146/annurev.nucl.55.090704.151558 [hep- Gauss-Bonnet gravity,” Phys. Lett. B 732, 349 (2014) ph/0502169]. [arXiv:1403.3242 [hep-th]]. [107] A. D. Sakharov, JETP Lett. 5, 24 (1967). [104] H. Georgi, Phys. Lett. 108B, 283 (1982) [108] V. A. Kuzmin, V. A. Rubakov, N. E. Shaposhnikov, [105] M. Fukugita, T. Yanagida, Phys.Lett. B 174 45 (1986). Phys. Lett. 155B 36 (1985). [109] L. P. Kadanoﬀ, G. Baym, Quantum Statistical Mechan- ics (Benjamin, New York,1962).