The acceleration from the Unimodular gravity view point: Background and linear perturbations

Miguel A. Garc´ıa-Aspeitia1,2,∗ A. Hern´andez-Almada3, Juan Maga˜na4, and V. Motta5

1Unidad Acad´emica de F´ısica, Universidad Aut´onomade Zacatecas, Calzada Solidaridad esquina con Paseo a la Bufa S/N C.P. 98060, Zacatecas, M´exico.

2Consejo Nacional de Ciencia y Tecnolog´ıa, Av. Insurgentes Sur 1582. Colonia Cr´editoConstructor, Del. Benito Ju´arez C.P. 03940, Ciudad de M´exico, M´exico.

3Facultad de Ingenier´ıa, Universidad Aut´onomade Quer´etaro, Centro Universitario Cerro de las Campanas, 76010, Santiago de Quer´etaro, M´exico.

4Instituto de Astrof´ısica & Centro de Astro-Ingenier´ıa,Pontificia Universidad Cat´olica de Chile, Av. Vicu˜naMackenna, 4860, Santiago, Chile and

5Instituto de F´ısica y Astronom´ıa, Facultad de Ciencias, Universidad de Valpara´ıso, Avda. Gran Breta˜na1111, Valpara´ıso, Chile. (Dated: May 28, 2021) With the goal of studying the cosmological constant (CC) problem, we present an exhaustive analysis of unimodular gravity as a possible candidate to resolve the CC origin and with this, the current Universe acceleration. In this theory, a correction constant (CC-like) in the field equations sources the late cosmic acceleration. This constant is related to a new parameter, zini, which is interpreted as the of CC-like emergence. By comparing with the CC value obtained from +0.01 and Supernovaes measurements, it is possible to estimate zini = 11.15−0.02 and zini = +0.03 11.43−0.06 respectively, which is close to the epoch. Moreover, we use the observational Hubble data (OHD), Type Ia Supernovae (SnIa), Baryon Acoustic Oscillations (BAO) and the Cosmic Microwave Background Radiation (CMB) distance data to constrain the UG cosmological +0.074 parameters. A Joint analysis (OHD+SnIa+BAO+CMB), results in zini = 11.47−0.073 within 1σ confidence level consistent with our estimation from Planck and Supernovae measurements. We also include linear perturbations, starting with scalar and tensor perturbations and complementing with the perturbed Boltzmann equation for photons. We show that the 00 term in the UG field equations and the Boltzmann equation for photons contains corrections, meanwhile the other equations are similar as those obtained in standard . Keywords: Unimodular gravity, cosmology, cosmological constant.

I. INTRODUCTION model representing the ∼ 69% of the total components in the Universe [4]. One feature of CC is that it is im- Universe acceleration is one of the most intriguing co- posed in the Einstein field equations, sustained only by nundrums in modern cosmology [1–3]. This feature has the Lovelock’s theorem [9], and its energy density value is been confirmed not only by Supernovae of the Type Ia adjusted by cosmological observations to reproduce the (SnIa), but also with observations related to Cosmic Mi- expected Universe dynamics. Nevertheless, the physics crowave Background (CMB) Radiation [4], among oth- of the CC is still unknown and the preferred interpreta- ers1. Nowadays there is a concentration of efforts to tion is that CC is associated with the quantum vacuum arXiv:1912.07500v2 [astro-ph.CO] 26 May 2021 understand the physics (at macro and micro scales) re- fluctuation of the space-time due to its auto-gravitating sponsible for such acceleration (see [7, 8] for a compi- characteristics [10, 11]. This hypothesis leads to one of lation of models) and its consequences in the Universe the biggest discrepancies between theory and observa- evolution. Despite that many hypotheses have been put tion in modern physics, because the theoretical predic- on the table, the best candidate is still the cosmologi- tion according to quantum field theory is ∼ 120 orders of cal constant (CC), an essential ingredient in the ΛCDM magnitude greater than the obtained from cosmological observations [10, 11]. On the other hand, unimodular gravity (UG) proposes ∗ aspeitia@fisica.uaz.edu.mx an interesting and natural origin of the CC [12–17], since 1 See other recent compilations of observations like Cosmic it appears as an integration constant, which can be cho- Chronometers [5] and Strong Gravitational Lensing (SGL) [6] sen without extra hypothesis and estimated with cos- to test theoretical models related with the Universe acceleration. mological observations. This particular permissiveness 2 comes from the consideration that the determinant of of this new constant term that later will act like an effec- the metric is a constant invariant volume form, reducing tive CC (hereafter CC-like). Hence, this constant term the degrees of freedom via diffeomorphism invariance in in should be the smoking gun of the Einstein-Hilbert action and allowing a non gravitat- the model that could be detected in future experiments ing CC. Using a constrained Hamiltonian dynamic, it is in order to refute or validate it. In this paper, we con- possible to show that UG has the same number of degrees strain the free parameters of the model through diverse of freedom than General Relativity (GR), even though it observation (OHD, SnIa, CMB and BAO data) conclud- has less symmetries, and thus avoiding the problems that ing that the epoch of CC-like emergence should be in the other theories encounter when working with third order reionization epoch as pointed out (based in CMB data) derivatives [18]. In addition, UG does not generate a by Ref. [25], contributing with observational evidence of natural energy-momentum conservation and its assump- these assertions. Even demonstrating that there is no tion should be imposed as an extra hypothesis, reducing tension between the different data samples strengthen- it automatically to the traditional Einstein field equa- ing the presented model. Moreover, we discuss diverse tions plus the integration constant discussed above [14]. phenomena that could lead to understand the emergence However, from our point of view, such extra hypothesis of the CC-like, such as those related to the continuous reduces important characteristics of UG, making blurry spontaneous localization. Finally, this particular scenario some aspects that could lead us to a most profound the- of emergent models (see for example [26]) could give us ory of the space-time itself while maintaining the CC in clues about the tension between the Hubble constant val- the same mystery as in the standard paradigm. ues estimated from the early and the late Universe data A radical point of view is to disregard the traditional [27] (see also [28–30]). hypothesis of energy-momentum conservation (as pre- On the other hand, linear perturbations in UG are re- vious works do [13, 14]), allowing a new mathematical quired to study those effects related to the growth of structure of conservation which not only contains the structure, the CMB power spectrum, among others, and energy-momentum tensor but also the Ricci and energy- to develop a comparison with the standard perturbations momentum scalar, and whose structure emerges natu- that rise from standard cosmology under GR context. rally from the UG field equations without extra hypoth- Recently, in [31], the authors studied linear perturba- esis. This new standpoint gives a mechanism to under- tions with only a scalar field concluding that UG and GR stand the CC from a different perspective, resolving the are similar at this level. However, the result is centered µ value of the energy density and tracing its presence as far in adding the hypothesis that ∇ [(R + 8πGT )gµν ] = 0, as the epoch of electro-weak transition [16, 17]. Indeed, which is an extra assumption and directly converge to in this approach the CC can be interpreted as a cumu- GR with subtle changes. Here, we develop linear per- lative small violation of the energy-momentum conserva- turbations in the Newtonian gauge line element to scalar tion, which in turn should be understood as granularity and tensor perturbations, showing that the 00 component of the space-time structure [17], being a possible evidence contains an extra term that involves the perturbations of a quantum space-time. for Ricci and energy-momentum scalars, meanwhile the In addition, UG could be the candidate to extend GR ij component and tensor perturbations are identically to because is compatible with observations of gravitational those predicted by GR. In contrast, the Boltzmann equa- waves [19, 20] or super massive black holes [21] and could tion for photons is exposed because it contains the energy be the key to explore the quantum regime [22, 23] or even momentum violations that characterize the UG. Notori- the quantum gravity realm [16, 17]. As final comments, ously, the extra term carries higher order derivatives in the UG theory has not only been tested in cosmology but the conformal time component for the scale factor and also in compact objects to constrain the violation of the for the scalar curvature. If we add the extra hypothesis energy-momentum conservation and its consequences in that the derivative of the extra term in UG is equal to astrophysics [24]. zero, we recover the standard linear equations for GR. A In this vein, in Ref. [25] the authors describe the CC in new jerk function must be constructed in order to follow the unimodular gravity context, obtaining a constant pa- the paths of the homogeneous and isotropic scenario pre- rameter related to the equation of state of radiation. This viously discussed and help us to discern among spurious new radical approach suggests that the causative of Uni- solutions caused by third order derivatives. verse acceleration is a constant related to the violation The present paper is organized as follows: In Sec. II of the energy-momentum conservation at a given compo- we present a review of UG based in previous studies. In nent, such as radiation, which emerges at a redshift zini, Sec. III we constrain UG free parameters through OHD, close the reionization epoch. Another characteristic is SnIa, CMB posterior distances and BAO samples; con- that, unlike GR, unimodular gravity can provide an Uni- taining at the end the respectively results of this section. verse acceleration without invoking an exotic fluid with a In Sec. IV we present the equations associated to lin- negative equation of state (EoS), which is an important ear perturbations, dividing our study in scalar, tensor advantage in comparison with GR. In addition, reference perturbations and the Boltzmann equations for photons, [25] suggests that the physics at early times could be dif- finalizing with a resume of the results at perturbative ferent from the physics at late times due to the emergence level. Finally, in Sec. V we give some conclusions and 3 outlooks. order derivatives, discarding spurious solutions with no physical meaning, because of its well-known behavior. The idea of the jerk parameter will be discussed later. II. REVIEW ON BACKGROUND COSMOLOGY On the other hand, the integral-transcendent- WITH UG Friedmann equation can be computed with the help of Eq. (2) and (4), obtaining the Friedmann equation for Unimodular gravity can be described by the following UG as field equation 2 8πG X 2 H = ρi + H . (5) 1  1  3 corr R − g R = 8πG T − g T , (1) i µν 4 µν µν 4 µν In addition, the acceleration equation is deduced from where all the tensors are the standards of GR and G is (2), obtaining the Newton’s gravitational constant. This equation can be deduced formally by Einstein-Hilbert action under the a¨  4πG X √ = − (ρ + 3p ) + H2 , (6) consideration −g = ξ, where ξ is a constant. a 3 i i corr In order to study the background cosmology, we con- i sider an isotropic, homogeneous and flat Friedmann- where the non-canonical extra term in Eqs. (5) and (6), Lemaitre-Robertson-Walker (FLRW) metric2, ds2 = i.e. the UG correction to the Friedmann and acceleration −dt2 + a(t)2d~x2, where a is the scale factor, the per- equations, is defined in the form fect fluid energy momentum tensor is written as Tµν = a(t) 0 pgµν + (ρ + p)uµuν , where p, ρ and uµ are the pressure, 8πG X 2 Z da H2 ≡ p + H(a0)2[j(a0) − 1] , (7) density and four-velocity of the fluid respectively. Hence, corr 3 i 3 a0 the integrability for FLRW metric in UG give us [13, 14] i aini where the sum runs over the different species in the Uni- a¨ X H˙ = − H2 = −4πG (ρ + p ), (2) verse and a is some constant initial value, which will a i i ini i be used as free parameter as discussed later.√ The con- stant term is possible because in UG we have −g = ξ, where the dots stands for time derivative. In addition, a being ξ a constant (see [13]), and thus allowing a to be general conservation for UG theory is now written in the ini a free parameter. The integral start from late epochs (to- form day) and explore any value in all epoch of the Universe µ evolution through a(t). ∇ [32πGTµν − (R + 8πGT )gµν ] = 0. (3) Although the Universe acceleration happens when R 2 −1 Without independently assuming the energy momentum H (j − 1)a da > 2πG(ρ + p), which allows common µ conservation (∇ Tµν = 0), the Eq. (3) introduces new fluids to accelerate the Universe, the presence of the scale Friedmann, acceleration and fluid equations coupled with factor in the expression restrict the acceleration epoch. third order derivatives in the scale factor due to second According to [25], it is plausible to consider an ansatz order derivatives in the Ricci scalar. Hence, in the case for the JP in terms of the redshift with the following of non traditional conservation of the energy-momentum characteristics tensor, Eq. (3) must be solved to obtain the characteris- 9(1 + w)w tic fluid equation. Hence, solving for (3) under a FLRW 3(w+1) j(z) = 2 Ω0i(z + 1) + 1, (8) metric and perfect fluid we have 2E(z) where w is the EoS for any fluid. This expression for X  d  H3 (ρ + p ) + 3H(ρ + p ) = (1 − j), (4) the jerk is not unique (other equations could be consid- dt i i i i 4πG i ered) and it is proposed to fulfill Eq. (4). Notice how the function reproduce both the acceleration (driven by where the sum is over all the species in the Universe ... a cosmological constant) and the matter stages. If we and j ≡ a /aH3 is the Jerk parameter (JP) [33, 34], additionally choose Ω0i → Ω0r and w → wr = 1/3, well known in cosmography and proposed by [25] for the this expression for the jerk will also reproduce the ra- study of UG. The jerk parameter is the key ingredient diation stage and, in combination with the other stages, to avoid problems with the initial-values related to third the ΛCDM jerk in all eras [25]. Physically, the jerk pa- rameter is a guide to reproduce the different stages be- cause it involves third order derivatives of the scale factor.

2 It also produce the expected expansion for the enthalpy An important clarification is that a change of variables (such as −4 a → b1/4 and dt → b−3/4dτ), applied by [32] to FLRW metric to (∝ a ) as suggested in [35]. In summary, this choice maintain the UG constraint, does not modify the mathematical for the jerk parameter not only recreates the benefits of consistency and it only impacts the physical interpretation (see ΛCDM model and alleviate the problems associated with details in [25] appendix). the initial conditions, but also unveils some aspects of the 4 nature of CC-like, i.e. introducing advantages over the agree with the expected value for Ω0Λ in GR, we re- CMB +0.01 standard cosmological model. quire that zini = 11.15−0.02. From here, we conclude As mentioned previously, a crucial characteristic is that the CC-like origin is approximately at the epoch of that this jerk parameter could come from third order reionization in the late Universe3, which is one of the derivatives in the equations of a more fundamental the- most important results of this theory. It is worth to ory of space-time (allowing the violation of the energy- notice that these results are under the assumption that +0.0056 +0.0042 momentum conservation) or by effects in the measure- Ω0Λ = 0.6889−0.0056 and h = 0.6766−0.0042 are taken ment problem in quantum mechanics (QM) as argued from Ref. [4]. Having said that, if we use the value by [16]. For example, regarding space-time, the jerk of Hubble constant from SnIa data measured by [41] +0.0182 could be originated by the causal sets (CS) approach to (h = 0.7422−0.0182), the result differs from the one from quantum gravity [15], where in the cosmological context Planck as expected. Hence, using this h estimation and 4 the diffusion in the phase-space leads to the violation of assuming that Ω0Λ coincide with the Planck value , we SnIa +0.03 the energy-momentum conservation. Alternatively, the obtain that zini moves to 11.43−0.06 (1σ CL), not too CMB jerk parameter could emerge from the continuous spon- far from the previous value of zini = 11.15. taneous localization (CSL), also related to the measure- ment problem in QM, which is connected to energy diver- gences in quantum collapse models, establishing a corre- III. BACKGROUND COSMOLOGICAL lation with UG as it is shown in [15, 36, 37]. In addition, CONSTRICTIONS Corral et al. [38] tackle the problem by arguing that the jerk parameter presented in Eq. (8) could be a diffusion At background level the parameters zini and h can parameter Q in the equation of state, which would mimic be constrained by performing a Bayesian Markov Chain the Universe acceleration. Monte Carlo (MCMC) analysis employing OHD, SnIa, From the previous equation and Eq. (5) it is possible CMB and BAO data. to deduce To perform the MCMC analysis, we use the emcee 2 3 4 Python module [43] choosing a flat prior over all the E(z) = Ω0m(z + 1) + Ω0r(z + 1) 2 parameters in the range h : [0.2, 1.0], Ωbh : [0, 0.04], 4 +Ω0exs(zini + 1) , (9) Ω0m : [0, 1]. To establish a bound over the parameter z , we use Eq. (10). We set a burn-in phase to achieve where E(z) ≡ H(z)/H ,Ω ≡ w Ω and z comes ini 0 0exs r 0r ini the convergence according to Gelman-Rubin criteria [44] from the constant term in the Eq. (7) integral. It is and 4000 MCMC steps with 250 walkers. worth mentioning that we have the following relation Then, we build a Gaussian log-likelihood as the merit- 2  1/4 of-function to minimize −2 log(Ldata) ∝ χdata, for each 1 − Ω0m − Ω0r zini = −1 + , (10) dataset mentioned previously. Additionally, a joint anal- wrΩ0r ysis can be constructed through the sum of them, i.e., 2 2 2 2 2 coming from the flatness condition. Notice that the χJoint = χSnIa + χOHD + χCMB + χBAO, (12) source of the Universe acceleration is the constant term in the previous equation (Eq. 9), which is a combina- where subscripts indicate the observational measure- ments under consideration. The rest of the section is tion of Ω0exs and zini, in where we can naturally relate 4 devoted to describe the different cosmological observa- Ω0Λ → Ω0exs(zini + 1) . Since our choice in Eq. (8) de- pends on the EoS and the energy density parameter of tions. the radiation, the constant inherits those terms. An up- per limit for z can be established through the equation ini A. Supernovae Type Ia −1/4 zini + 1 < Ω , (11) 0exs The largest compilation provided by Ref. [3], contains −6 −2 hence, for Ω0exs = 8.23 × 10 h (1 + 0.2271g∗), being the observations of the luminosity modulus from 1048 g∗ = 3.04 the standard number of relativistic species [39], SnIa spanned in the redshift region 0.01 < z < 2.3. The and h is the dimensionless Hubble constant. This result χ2-function is constructed as [45] is dictated under the assumption that Ω = Ω (z + 0Λ 0exs ini  e  b2 1)4 < 1, obtaining an upper limit for z to be z < χ2 = a + log − , (13) ini ini P anSnIa 12.54 at 95% confidence level (CL), using Planck mission 2π e +0.0042 value h = 0.6766−0.0042 [4] (notice that the result for this constriction is strongly dependent of what h value is 3 considered). In Ref. [16], the CC can be traced from electroweak epoch, due to some specific quantum violation related to the energy- ΛCDM model (which our theory mimics) considers momentum tensor. the reionization era in the region 6 < z < 20, pre- 4 In general, this assumption is not true, however Riess et al. [42] dicting a ionization fraction of ∼ 0.2 at z ∼ 12 (see does not report the expected value for Ω0Λ in the Supernovaes Ref. [40] for details). On the other hand, in order to observations. 5

T −1 T −1 where a = ∆µ˜ · CP · ∆µ˜, b = ∆µ˜ · CP · ∆1, e = TABLE I: Mean values for the model parameters (h, T −1 z ) and χ2 , derived from each data set and the joint ∆1 ·CP ·∆1, ∆µ˜ =µ ˜th −µ˜obs is the vector of residuals ini min between the model distance modulus and the observed analysis. −1 oneµ ˜obs, ∆1 is the unit vector, and CP is the inverse of the covariance matrix. The theoretical distance modulus 2 is estimated by Data set χmin h zini +0.016 +0.237 OHD 22.0 0.709−0.016 11.788−0.250 µ˜th(z) = M + 5 log [dL(z)/10 pc], (14) +0.270 +2.366 10 SnIa 1036.0 0.602−0.272 10.623−3.021 +0.006 +0.091 CMB 0.0001 0.678−0.006 11.259−0.092 where M is a nuisance parameter that has been marginal- +0.031 +0.979 BAO 12.9 0.701−0.033 10.847−1.383 ized in (13), and dL(z) is the dimensionless luminosity +0.005 +0.074 distance given by the following equation Joint 1097.6 0.692−0.005 11.473−0.073

Z z dz0 dL(z) = (1 + z)c 0 , (15) 0 H(z )  th  where c is the light velocity. R − 1.7502 th VCMB =  lA − 301.147  , (18) 2th Ωbh − 0.02236 B. Observational Hubble Data the superscripts th refer to the theoretical values, and Cov−1 represents the inverse of Another important sample, is the Observational Hub- CMB ble Data (OHD) which consist of cosmological model in-  2116.00 18938.20 −45.54  dependent measurements of the Hubble parameter H(z). −8 CovCMB = 10  18938.20 801025.00 −443.03  . We consider the OHD compilation provided by [5] com- −45.54 −443.03 2.25 prised by 51 points given by the differential age (DA) tool (19) (20 points) and BAO measurements (31 points) within which is the covariance matrix for V . the redshift region 0 < z < 2.36. The χ2-function for CMB OHD can be written as

51  2 D. Baryon Acoustic Oscillations 2 X Hth(zi) − Hobs(zi) χOHD = i , (16) σobs i Baryon Acoustic Oscillations (BAO) are considered as standard rulers, being primordial signatures of the inter- where Hth(z) and Hobs(zi) are the theoretical and obser- i actions between baryons and photons in a hot plasma vational Hubble parameter at the redshift zi, and σobs is the observational error. on the matter power spectrum in the pre-recombination epoch. In Ref. [47], the authors collected 15 transversal BAO scale measurements, obtained from luminous red galaxies located in the region 0.110 < z < 2.225. C. CMB from Planck 2018 measurements In order to be used as a method to constrain cosmo- logical models, it is useful to build the χ2-function as Recently, [31] performed a further perturbative analy- sis of the UG model. They found that, at first and sec- 2 15 i ! ond order of perturbative level, there are no differences 2 X θBAO − θth(zi) χBAO = , (20) σ i between ΛCDM and UG, hence the UG theory is not i=1 θBAO excluded by measuring either the growth factor, CMB lensing or the integrated Sachs-Wolfe effect. Therefore, i where θBAO is the BAO angular scale and its uncertainty as a first approach, we consider the CMB posterior dis- σ i measured at zi. The theoretical counterpart, θth, θBAO tances from ΛCDM to test this model. is estimated through the following equation We use the shift parameter, R = 1.7502 ± 0.0046, +0.089 2 the acoustic scale, lA = 301.471−0.090, and Ωb0h = rdrag θth(z) = . (21) 0.02236 ± 0.0015 obtained for a flat Λ-cold (1 + z)DA(z) model [46]. Thus, the figure-of-merit is built as In the latter, rdrag is defined by the sound horizon at 2 baryon drag epoch and DA = dL(z)/(1 + z) is the angu- 2 −1 T χCMB = VCMB · CovCMB · VCMB, (17) lar diameter distance at z, with dL(z) defined in equation (15). In addition, we use the rdrag = 147.21 ± 0.23 re- where VCMB is ported by [4]. 6

emergence of small violation of the energy-momentum SnIa conservation which can be accounted as the energy den- BAO sity of the CC. For example, the collapsed structures in OHD the epoch of reionization are a way to brake the sym- CMB metry of the space-time, which is extensively explored Joint by authors like in Ref. [53]. Furthermore, CSL or CS, processes could be other reasons to the apparition of the CC, as we discussed earlier. 12.5 i n i 11.5 IV. EQUATIONS ASSOCIATED TO LINEAR z PERTURBATIONS OF UG

10.5 We dedicate this section to elucidate the equations 0.65 0.70 0.75 10.5 11.5 12.5 associated to the linear perturbations for UG, starting z h ini with the field equations and later on with the Boltzmann equations, centering our attention in the photons case FIG. 1: 1D marginalized posterior distributions and the because, as discussed in the previous section, this is the 2D 68%, 95%, 99.7% of CL for the h, and zini equation coupled with the jerk parameter. parameters of the UG model. The star (square) marker In order to study the equations associated to the linear represents the best fit value of Joint (CMB) data perturbations, we rewrite Eq. (1) in the form

1 G + (R + 8πGT )g = 8πGT , (22) E. Results at Background Level µν 4 µν µν

1 Table I provides the mean values for h and zini ob- where Gµν ≡ Rµν − 2 gµν R, is the Einstein tensor. tained from each data set and a joint analysis. Further- We start using the conformal Newtonian gauge line more, Figure 1 presents the 1D marginalized posterior element given by the components distributions and 2D at 68%, 95%, 99.7% of CL for the UG parameters. The results for the five cases confirm g00(t, x) = −1 + h00(t, x), (23) that zini ∼ 12; hence, we conclude that the CC-like g0i(t, x) = a(t)h0i(t, x), (24) should emerge in the reionization epoch (see [40]) in con- 2 gij(t, x) = a (t)[δij + hij(t, x)], (25) trast to the standard cosmological model where the CC is always present. where hµν  1, are metric perturbations. More precisely, the joint analysis estimates zini = +0.07 +0.01 +0.03 11.47−0.07, while zini = 11.15−0.02 and 11.43−0.06 are those obtained with the h values from Planck, and SnIa A. Scalar Perturbations measurements respectively. Note that the 4.6σ between CMB and the joint analysis is due that we choose the value of h given by Riess et al. [42] instead the pre- For this particular case, we reduce the above equa- sented by Planck [4]. As mentioned before, these val- tions to h00(t, x) = −2Ψ(t, x), h0i = 0 and hij(t, x) = ues lie on the reionization region (6 < z < 20) [48] 2δijΦ(t, x), where Ψ(t, x) and Φ(t, x) are the Newtonian (see [40] for other regions with diverse conditions), and and curvature potential, respectively. are also in agreement with recent measurements [49] by The first equation that we present is associated with the Shaped Antenna of the background Radiation Spec- the 00 component, therefore we have the following parts trum (SARAS), which ruled out reionization in the re- gion 10 < z < 6. Several important events happened at 2k2 that time such as the apparition of high-redshift galaxies δG0 = −6HΦ˙ + 6H2Ψ − Φ, (26) [48] or the birth of the first stars in the Universe, the 0 a2 so-called Population III stars [50–52]. In that epoch, the  a¨  2k2 δR = −12Ψ H2 + + Ψ + 6Φ¨ − 6H(Ψ˙ − 4Φ)˙ combination of the above mentioned mechanisms reionize a a2 the intergalactic medium initiating the process of metal- 4k2 licity in the Universe (and bringing the end of the so- + 2 Φ, (27) called dark ages epoch). We remark that this epoch is a 0 X particularly important because non linear physics pro- δT0 = − ρiδi − ργ Θ0 − ρν N0, (28) cesses are involved. We speculate that the friction of i non linear structures (galaxies, clusters, etc) with the X δT = −3 ρiδi − 2(ργ Θ0 + ρν N0), (29) granularity of the space-time could have helped to the i 7 where k is the wave-number related to the perturbation being h+, h×  1 two components of the divergenceless- j scale, ∂/∂x → ikj, δi ≡ δρi/ρi are the fractional over- traceless symmetric tensor chosen in the x − y plane. densities for DM and baryons, while Θ0 and N0 are the The subscripts + and × indicates the wave polarizations. monopolar contributions for photons and neutrinos per- Therefore, the components to construct Eq. (22) are turbations. Notice also that δ indicates that we are work-  2  ing with first order terms (linear perturbations). i i ik 1 k 3 δGj = δRj = δ hkj,00 + hkj + Hhkj,0 ,(34) Combining previous equations in the 00 component of 2 2a 2 (22) at perturbative level we have δR = 0, (35) the ij components of the energy momentum tensor is as 1 h X −3HΦ0 + 3H2Ψ − k2Φ + Λ(η, k) = −4πGa2 ρ δ + always (avoiding anisotropic stress contributions) 2 i i i Z 3 i X d p p pj i T i(t, x) = g f (x, p, t), (36) ργ Θ0 + ρν N0 (30) j s (2π)3 E (p) s s s where where fs(x, p, t) is the distribution function, gs values are the degrees of freedom of the fluids, pi and E(p) are  a00  k2 3 3 the three momenta and energy of the particles involved, Λ(η, k) = 3Ψ H2 + − Ψ − Φ00 + H(Ψ0 − 4Φ0) a 2 2 2 respectively. Moreover, δT is the same as that shown in " # Eq. (29). Following the traditional guidelines we have 2 2 X that −k Φ + 2πGa 3 ρiδi + 2(ργ Θ0 + ρν N0) , (31) i 2 1 2 00 0 2 here the equations are transformed to the conformal time a [δG1 − δG2] = hi + 2Hhi + k hi, (37) through the relation dt = adη, primes denote derivatives 0 where i = +, ×, and with respect to η ( = ∂η) and H is the Hubble parameter in conformal time. Notice that when Λ(η, k) = C, where C  1 is a constant small perturbation, the equation that  1  8πG δT 1 − δT 2 − T (g1 − g2) = 0, (38) rise from GR plus a constant should be recovered. How- 1 2 4 1 2 ever, this is no trivial, because the previous imposition, generates another constraint over the functions that con- thus, tains (31). The hypothesis and mathematical arguments 00 0 2 used to infer Λ(η, k) = C needs the Boltzmann equations hi + 2Hhi + k hi = 0. (39) discussed later. On the other hand, for the ij equations we use the Therefore, the linear tensor perturbations does not show i ˆ ˆj any difference with the one from GR. longitudinal traceless (LT) part of Gj denoted by kik − j (1/3)δi . Considering that the term (R + 8πGT )/4 con- i tains a product with δj, this is neglected by the LT con- C. Boltzmann equations sideration, therefore the final form is equal to those ex- pected by GR As we discussed previously, Eq. (4) is the main 2 2 k (Φ + Ψ) = −32πGa [Θ2ργ + N2ρν ], (32) equation for the fluids with the sum indicating that all the components considered in the cosmology should be where Θ2 and N2 are the quadrupole contributions of added. For the matter component (with p = 0), the well photons and neutrinos, respectively. know equationρ ˙m+3Hρm = 0 is directly recovered, while for the radiation component the jerk is designed ad-hoc in such a way that the equationρ ˙r + 4Hρr = 0 for rel- B. Tensor Perturbations ativistic particles is recovered. Thus, it is expected that linear perturbations of Boltzmann equation for photons For tensor perturbations, we begin describing the met- should carry the information encrypted in the j term, ric components as h00 = −1, h0i = 0 and even in the free streaming epoch when Compton process could be safely neglected. In this epoch, the particle den- sity is low, implying that the optical depth is negligible δg (t, x) = a(t)2h (t, x), (33) R ij ij (τ  1) and defined as τ = neσT adη, being ne the elec- where tron number density and σT the Thompson cross section and hence, the interaction part described by the collision   h+ h× 0 function C[f(p)] will not play a role. µ hij =  h× −h+ 0  , From Eq. (3) it is possible to observe that ∇ Tµν = 0 0 0 ∂ν F(R,T ), being ∂ν F(R,T ) a function that involves 8 derivatives of the Ricci and energy-momentum scalars D. Results at Perturbative Level (the function absorbs the factor (32πG)−1 ), which im- plies higher order derivatives as it is shown in equation UG equations at perturbative level are presented, im- (4) encrypted in the jerk parameter. However, notice plementing the conformal Newtonian gauge line element, that the derivatives are only depending on time because showing the scalar and tensor perturbations as well as the ρ and p only depends on t, therefore we have ∂tF(R,T ) Boltzmann equation for photons. It is notorious that the and then we arrive to Eq. (4). scalar perturbations for 00 contains a term Λ(η, k), not Consequently, if photons carry the violations to the en- expected in standard GR, promoting another constric- ergy momentum tensor5, we expect that the linear per- tion into UG equations. Regarding the ij component, turbed equations in conformal time with τ  1, will take for scalar perturbations and due to the LT considera- the form tion, Eq. (32) maintains the same form as the one shown in GR. In the case of linear tensor perturbations, we con- 0 0 ˜ Θ + ikµΘ + Φ + ikµΨ = ∂ηF(δR, δT ), (40) clude that the equations do not change, having the same result as predicted by GR. where kµ = k · pˆ, and pˆ is the momentum direction, On the other hand, it is possible to deduce the Boltz- Θ ≡ δT/T is the photon perturbations, which cause dif- mann equation for the photons, assuming a negligible ferences in the temperature, and F˜ is the perturbed part optical depth, i.e. no Compton scattering. The left side of the function F, involving δR and δT given by Eqs. of the equation is the same as the one expected in stan- (27) and (29) for the scalar perturbations. Then, we have dard cosmology, meanwhile the right side will contain a for a scalar energy-momentum tensor composed only by function that depends on δR, δT and higher order con- photons formal temporal derivatives that now also involves the spatial curvature Φ. In this case, Eq. (8) is not useful as a guide to avoid problems with high order derivatives. h  a00  k2 ∂ F˜(δR, δT ) = (32πG)−1∂ 3Ψ H2 + − Ψ The classical way to solve this problem is to assume Eq η η a 2 (41) is equal to zero, therefore Λ(η, k) = C, which justify 3 3 i the previously mentioned assumption. − Φ00 + H(Ψ0 − 4Φ0) − k2Φ + 4πGa2ρ Θ 2 2 γ 0 Therefore, UG provides an extra equation which reads −1 = (32πG) ∂ηΛ(η, k). (41)  a00 k2  3 3 3 H2 + − Ψ − Φ00 + H(Ψ0 − 4Φ0) Notice that the derivatives involve the conformal time a 6 2 2 parameter, then, from Eq. (27) we will have higher or- 2 2 der derivatives for a and Φ. Thus, if we want to obtain −k Φ + 4πGa ργ Θ0 = C. (42) the standard perturbations for GR, the equation (8) is not useful anymore. In the same way, we need to pro- We will call it the constriction equation for UG. pose a new jerk parameter in order to face the problems Conversely, we can face Eq. (41) which implies dealing associated with third order derivatives encoded in the with higher order derivatives. In this case, the strategy Boltzmann equation for photons. should be the same as the one for the homogeneous Uni- As a final remark, we expect that the Boltzmann equa- verse, i.e. find a well known function as a guide (the jerk tions for the remaining species must be the same as these function). A direct manipulation of Eq. (40) generates obtained in the standard cosmology. Thus, the mathe- the following integro-differential equation, as it happens matical structure for the UG equations suggest that only for the background case, in the following way radiation (at background level) and photons (at pertur- bative level) are the only species coupled with the non Z η h  a00  conservative term. While we increase the energy, the ikµ (Θ + Ψ)dη = (32πG)−1 3Ψ H2 + a Compton scattering appears and therefore, Eq. (40) η0 must contain the term −τ 0[Θ − Θ + u µ], where u is k2 3 3 0 b b −32πG(Θ + Φ) − Ψ − Φ00 + H(Ψ0 − 4Φ0) the bulk velocity. Regarding the initial conditions associ- 2 2 2 i ated to the inflationary process, those must be studied in −k2Φ + 4πGa2ρ Θ , (43) a future work because, unlike the GR case, the analysis γ 0 requires extra contributions. which must be solved in combination with the fluids and Einstein equations.

5 At background level, we only claim that radiation (relativistic V. CONCLUSIONS AND OUTLOOKS particles) carry the energy-momentum violation, but we do not specify if those are the photons or the neutrinos. Here, we are adding another extra hypothesis assuming that photons specifi- Unimodular gravity, and its particular and natural cally carry energy-momentum violations. characteristic of violating the traditional form of the 9 energy-momentum conservation, not only suggests the ation of the Universe. As we argue previously, we suspect apparition of the CC-like term together with its radia- that non linear physics play an important role in the for- tion inherited characteristics (which is assumed coupled mation of complex structure (such as stars and galaxies), with the jerk parameter), but also the redshift at which generating friction with a possible non-continuous space- it emerges. In this scenario, the CC-like arises during the time and provoking the apparition of the CC-like. The epoch of reionization, when the newly form galaxies and latter assertion could be compatible with CS paradigm stars of the Population III reionizated the intergalactic or even CSL as we discussed previously. medium and, in general, when complex structures played Cosmological linear perturbations in UG is another a role in the cosmological dynamics and converting the subject that is studied in this paper. Under the Newto- Universe into a highly non-linear entity. nian conformal gauge we present scalar and tensor per- turbations, together with the Boltzmann equation for Towards of the order z & 11, the CC-like should generate a fingerprint that could be detected in photons. Under this scenario, we show how the 00 com- future observations. Hence, a particular form to refute ponent for scalar perturbations, presents an extra contri- or validate the model is the prediction that CC does bution call it Λ(η, k), meanwhile the ij and tensor per- not exist in all stages of the evolution of the Universe, turbations remains equal as those obtained by GR. As but it arises during the epoch of reionization, being the it is expected, the Boltzmann equation for photons, con- smoking-gun epoch to obtain evidence of CC and eluci- tains an additional term that contains third order deriva- date its future implications for the Universe evolution. tives, not only in the scale factor (as happens in the back- Models like UG where it is predicted a late CC, are ground) but also to the scalar curvature. The additional called emergent which the capability to alleviate the re- hypothesis of consider that Eq. (41) tends to zero, give us the constriction equation for UG. However, if we face cent problem related with observations of H0 at late and early cosmological times [27]. the problem of third order derivatives, it is necessary to solve Eq. (43) (coupled with the other fluids and field We constrained the UG parameters using OHD, SnIa, equations), which is an integro-differential equation for CMB and BAO data, together with a joint analysis. Our photons. Formally, it is the Boltzmann equations for main result estimates a value for z (Joint) of 11.473, in ini photons in UG scenario. As a final remark we mention good agreement with the theoretical values (11.5, 11.43) that it is not possible to generate a numerical analysis of using h measured by Planck 2018 and [42]; strengthening the linear perturbations presented previously and with our conclusions that important physics should emerge in this, the matter power spectrum, before study the initial the reionization epoch6. Furthermore, the estimated val- conditions in this scenario (inflationary Universe), which ues for parameter z are in agreement for the data used, ini will be studied elsewhere. which points towards a robust model at the background Finally, the confirmation of this model would not only level and motivates the study of other aspects (such as open the door to understand the current Universe accel- growth structure at perturbative level, etc). eration, but also could be a piece of evidence for the pos- Moreover, the Experiment to Detect the Global Epoch sible granularity or other primordial effects of the space- of reionization (EDGES, [54]) detects an interesting event time itself. at a redshift closer to the one we obtain (z ∼ 12): an excess of radiation centered at z ≈ 17 and spanning a re- gion 20 > z > 15. Thus, the EDGES event is consistent ACKNOWLEDGMENTS with our predictions and it could represent the detec- tion of the CC-like emergence. However, we caution that The authors acknowledge the enlightening conversa- EDGES result is still in the center of a controversy [49], tions with Luis Ure˜naand Daniel Sudarsky. M.A.G.-A. thus, we should wait for new experiments to confirm or acknowledges support from SNI-M´exico, CONACyT refute the observations. research fellow, COZCyT and Instituto Avanzado de Despite that UG sheds light into the nature of the CC Cosmolog´ıa(IAC) collaborations. A.H.A. thanks to the and establishes a redshift for its possible presence, it is PRODEP project, Mexico for resources and financial not clear why the CC-like emerges at the epoch of reion- support. J.M. acknowledges the support from CONI- ization. One possibility is that the non linearity in that CYT project Basal AFB-170002, V.M. acknowledges the epoch generates some kind of symmetry breakdown by support of Centro de Astrof´ısicade Valpara´ıso(CAV). a still unknown process and this, in turn, leads to the J.M., M.A.G.-A. and V.M. acknowledge CONICYT presence of the CC-like originated by the violation of the REDES (190147). energy-momentum conservation, resulting in the acceler-

6 Note that the 4.6σ between CMB and the joint analysis is due the presented by Planck [4]. that we choose the value of h given by Riess et al. [42] instead 10

[1] A. G. Riess, A. V. Filippenko, P. Challis, A. Clocchiatti, [28] A. Perez, D. Sudarsky, and E. Wilson-Ewing, (2020), A. Diercks, et al., The Astronomical Journal 116, 1009 arXiv:2001.07536 [astro-ph.CO]. (1998). [29] E. Di Valentino, O. Mena, S. Pan, L. Visinelli, W. Yang, [2] S. Perlmutter, G. Aldering, G. Goldhaber, R. A. Knop, A. Melchiorri, D. F. Mota, A. G. Riess, and J. Silk, P. Nugent, others, and T. S. C. Project, The Astrophys- (2021), arXiv:2103.01183 [astro-ph.CO]. ical Journal 517, 565 (1999). [30] G. Efstathiou, (2021), arXiv:2103.08723 [astro-ph.CO]. [3] D. M. Scolnic and et. al., The Astrophysical Journal 859, [31] A. Basak, O. Fabre, and S. Shankaranarayanan, Gen. 101 (2018). Rel. Grav. 48, 123 (2016), arXiv:1511.01805 [gr-qc]. [4] N. Aghanim et al. (Planck), (2018), arXiv:1807.06209 [32] E. Alvarez, S. Gonz´alez-Mart´ın, M. Herrero-Valea, and [astro-ph.CO]. C. P. Mart´ın, JHEP 08, 078 (2015), arXiv:1505.01995 [5] J. Maga˜na,M. H. Amante, M. A. Garc´ıa-Aspeitia, and [hep-th]. V. Motta, Mon. Not. Roy. Astron. Soc. 476, 1036 (2018), [33] M.-J. Zhang, H. Li, and J.-Q. Xia, Eur. Phys. J. C77, arXiv:1706.09848 [astro-ph.CO]. 434 (2017), arXiv:1601.01758 [astro-ph.CO]. [6] M. H. Amante, J. Maga˜na,V. Motta, M. A. Garc´ıa- [34] A. Al Mamon and K. Bamba, Eur. Phys. J. C78, 862 Aspeitia, and T. Verdugo, (2019), arXiv:1906.04107 (2018), arXiv:1805.02854 [gr-qc]. [astro-ph.CO]. [35] H. Velten and T. R. P. Caramˆes,Universe 7, 38 (2021), [7] E. J. Copeland, M. Sami, and S. Tsujikawa, Int. J. Mod. arXiv:2102.03457 [gr-qc]. Phys. D15, 1753 (2006), arXiv:hep-th/0603057 [hep-th]. [36] L. E. Ballentine, Phys. Rev. A 43, 9 (1991). [8] M. Li, X.-D. Li, S. Wang, and Y. Wang, Commun. [37] A. Bassi, E. Ippoliti, and B. Vacchini, Journal of Physics Theor. Phys. 56, 525 (2011), arXiv:1103.5870 [astro- A: Mathematical and General 38, 8017 (2005). ph.CO]. [38] C. Corral, N. Cruz, and E. Gonz´alez,Phys. Rev. D 102, [9] D. Lovelock, Journal of Mathematical Physics 12, 498 023508 (2020), arXiv:2005.06052 [gr-qc]. (1971), https://doi.org/10.1063/1.1665613. [39] E. Komatsu, K. M. Smith, J. Dunkley, C. L. Bennett, [10] S. Weinberg, Reviews of Modern Physics 61 (1989). B. Gold, G. Hinshaw, N. Jarosik, D. Larson, M. R. Nolta, [11] Y. B. Zeldovich, Soviet Physics Uspekhi 11 (1968). L. Page, D. N. Spergel, M. Halpern, R. S. Hill, A. Kogut, [12] J. L. Anderson and D. Finkelstein, Amer- M. Limon, S. S. Meyer, N. Odegard, G. S. Tucker, J. L. ican Journal of Physics 39, 901 (1971), Weiland, E. Wollack, and E. L. Wright, The Astrophys- https://doi.org/10.1119/1.1986321. ical Journal Supplement Series 192, 18 (2011). [13] G. F. R. Ellis, H. van Elst, J. Murugan, and J.-P. Uzan, [40] D. Paoletti, D. K. Hazra, F. Finelli, and G. F. Smoot, Classical and Quantum Gravity 28, 225007 (2011). JCAP 09, 005 (2020), arXiv:2005.12222 [astro-ph.CO]. [14] C. Gao, R. H. Brandenberger, Y. Cai, and P. Chen, [41] A. G. Riess et al., Astrophys. J. 826, 56 (2016), JCAP 1409, 021 (2014), arXiv:1405.1644 [gr-qc]. arXiv:1604.01424 [astro-ph.CO]. [15] T. Josset, A. Perez, and D. Sudarsky, Phys. Rev. Lett. [42] A. G. Riess, S. Casertano, W. Yuan, L. M. 118, 021102 (2017), arXiv:1604.04183 [gr-qc]. Macri, and D. Scolnic, Astrophys. J. 876, 85 (2019), [16] A. Perez, D. Sudarsky, and J. D. Bjorken, Int. J. Mod. arXiv:1903.07603 [astro-ph.CO]. Phys. D27, 1846002 (2018), arXiv:1804.07162 [gr-qc]. [43] D. Foreman-Mackey, D. W. Hogg, D. Lang, and J. Good- [17] A. Perez and D. Sudarsky, Phys. Rev. Lett. 122, 221302 man, PASP 125, 306 (2013), arXiv:1202.3665 [astro- (2019), arXiv:1711.05183 [gr-qc]. ph.IM]. [18] M. Henneaux and C. Teitelboim, Physics Letters B 222, [44] A. Gelman and D. Rubin, Statistical Science 67, 457 195 (1989). (1992). [19] B. P. Abbott et al. (LIGO Scientific, Virgo), Phys. Rev. [45] A. Conley, J. Guy, M. Sullivan, N. Regnault, P. Astier, Lett. 116, 061102 (2016), arXiv:1602.03837 [gr-qc]. C. Balland, S. Basa, R. G. Carlberg, D. Fouchez, [20] B. P. Abbott et al. (LIGO Scientific, Virgo), Phys. Rev. D. Hardin, I. M. Hook, D. A. Howell, R. Pain, Lett. 116, 241103 (2016), arXiv:1606.04855 [gr-qc]. N. Palanque-Delabrouille, K. M. Perrett, C. J. Pritchet, [21] T. E. C. et al., ApJL 875, 1 (2019). J. Rich, V. Ruhlmann-Kleider, D. Balam, S. Baumont, [22] R. Bufalo, M. Oksanen, and A. Tureanu, Eur. Phys. J. R. S. Ellis, S. Fabbro, H. K. Fakhouri, N. Fourmanoit, C75, 477 (2015), arXiv:1505.04978 [hep-th]. S. Gonz´alez-Gait´an, M. L. Graham, M. J. Hudson, [23] R. Percacci, Found. Phys. 48, 1364 (2018), E. Hsiao, T. Kronborg, C. Lidman, A. M. Mourao, J. D. arXiv:1712.09903 [gr-qc]. Neill, S. Perlmutter, P. Ripoche, N. Suzuki, and E. S. [24] J. A. Astorga-Moreno, J. Chagoya, J. C. Flores-Urbina, Walker, The Astrophysical Journal Supplement Series and M. A. Garc´ıa-Aspeitia, JCAP 2019, 005 (2019), 192, 1 (2010). arXiv:1905.11253 [gr-qc]. [46] L. Chen, Q.-G. Huang, and K. Wang, JCAP 02, 028 [25] M. A. Garc´ıa-Aspeitia, C. Mart´ınez-Robles, (2019), arXiv:1808.05724 [astro-ph.CO]. A. Hern´andez-Almada, J. Maga˜na, and V. Motta, [47] R. C. Nunes, S. K. Yadav, J. F. Jesus, and A. Bernui, Phys. Rev. D99, 123525 (2019), arXiv:1903.06344 (2020), arXiv:2002.09293 [astro-ph.CO]. [gr-qc]. [48] R. J. Bouwens, G. D. Illingworth, P. A. Oesch, J. Caru- [26] A. Hern´andez-Almada, G. Leon, J. Maga˜na, M. A. ana, B. Holwerda, R. Smit, and S. Wilkins, Astrophys. Garc´ıa-Aspeitia, and V. Motta, Mon. Not. Roy. Astron. J. 811, 140 (2015), arXiv:1503.08228 [astro-ph.CO]. Soc. 497, 1590 (2020), arXiv:2002.12881 [astro-ph.CO]. [49] A. Liu and J. R. Shaw, Publ. Astron. Soc. Pac. 132, [27] L. Verde, T. Treu, and A. G. Riess (2019) 062001 (2020), arXiv:1907.08211 [astro-ph.IM]. arXiv:1907.10625 [astro-ph.CO]. [50] M. Magg, R. S. Klessen, S. C. O. Glover, and H. Li, MN- 11

RAS 487, 486 (2019), arXiv:1903.08661 [astro-ph.GA]. N. Yoshida (2010) pp. 134–137, arXiv:1006.4434 [51] S. Glover, in The First Galaxies, Astrophysics and [astro-ph.CO]. Space Science Library, Vol. 396, edited by T. Wik- [53] L. Lombriser, (2018), 10.1088/1475-7516/2019/09/065, lind, B. Mobasher, and V. Bromm (2013) p. 103, [JCAP1909,no.09,065(2019)], arXiv:1805.05918 [astro- arXiv:1209.2509 [astro-ph.CO]. ph.CO]. [52] M. Trenti, in American Institute of Physics Conference Series,[54] J. D. Bowman, A. E. E. Rogers, R. A. Monsalve, T. J. American Institute of Physics Conference Series, Vol. Mozdzen, and N. Mahesh, Nature 555, 67 (2018), 1294, edited by D. J. Whalen, V. Bromm, and arXiv:1810.05912 [astro-ph.CO].