Arxiv:2107.12488V1 [Gr-Qc] 26 Jul 2021 ∗ ‡ † I.Toﬂiswt Osateutosof Equations Constant with ﬂuids Two III

arXiv:2107.12488v1 [gr-qc] 26 Jul 2021 ∗ ‡ † I.Tofliswt osateutosof equations constant with fluids Two III. V w ud ihcntn qain of equations constant with fluids Two IV. [email protected] [email protected] [email protected] I il qain,flis n clrfils3 fields scalar and fluids, equations, Field II. .Itouto 2 Introduction I. time conformal with solutions state—parametric tt—ooigtm 8 time state—comoving .Eoi ud 11 fluids Exotic D. .Rdcint udaue 7 quadratures to Reduction D. .Snl udadeetv uvtr / curvature effective and fluid Single A. .Frtcondition: First A. .Smay11 Summary C. .Mlifli omlg 6 cosmology Multi-fluid C. .Scn condition: Second B. .Saa ed5 field Scalar B. .Smere fteEinstein-Friedmann the of Symmetries E. .Si atrpu eod(elor (real second a plus matter Stiff 4. second a plus constant Cosmological 3. or (real second a plus Radiation 2. .Ds lsascn ra reffective) or (real second a plus Dust 1. .Si atrpu eod(elor (real second a plus matter Stiff 4. 3. or (real second a plus Radiation effective) 2. or (real second a plus Dust 1. Λ qain 7 equations -fluids ffcie ud10 9 10 fluid effective) fluid effective) or (real fluid effective) fluid ffcie ud11 10 10 fluid effective) Λ fluid effective) fluid oiae eiwwt oeneyes. modern vi with solut of review and a point conditions, motivates integrability betwee modern solutions, relations a Old and from fields. solutions solutions, Interesting the de of a literature. provide forms We parametric fields. and scalar state and of constant, equations cosmological constant the with fluids perfect effective) Friedmann-Lemaître-Robertson-Waor describe and functions erve nltclsltoso h iseneutoswh equations Einstein the of solutions analytical review We lsascn ra reetv)fli 11 fluid effective) or (real second a plus ut-udcsooyi isengaiy nltclsolu analytical gravity: Einstein in cosmology Multi-fluid CONTENTS p +1 r p aei Faraoni, Valerio +1 r 1 60CleeSre,Sebok,Qée,Cnd 1 1Z7 J1M Canada Québec, Sherbrooke, Street, College 2600 = eateto hsc srnm,Bso’ University, Bishop’s Astronomy, & Physics of Department + n q ∈ = Z n ∈ Z 1, ∗ oi Jose, Sonia 12 10 9 9 4 II w neatn ud 27 fluids interacting Two VIII. 1, I.Saa edsltos23 solutions field Scalar VII. I he ra reetv)flissltos21 solutions fluids effective) or (real Three VI. † .Epii ouin 14 solutions Explicit V. o ehd epbigrdsoee,which rediscovered, being keep methods ion n tv Dussault Steve and ffcieflisicuesailcurvature, spatial include fluids Effective . M. G. O. K. D. .Rdainpu pta curvature, spatial plus Radiation D. .Itrcigdr nryaddr atr27 matter dark and energy dark Interacting A. .Snl clrfil 23 field scalar Single A. .Rdainplus Radiation A. N. H. .Dust, A. A. .Mlil o-neatn clrfils27 fields scalar non-interacting Multiple C. C. .Radiation, C. .Itrcigrdainadnon-relativistic and radiation Interacting B. .Saa edpu ud25 fluid plus field Scalar B. B. .Ds lssailcurvature, spatial plus Dust B. .Any B. E. E. F. L. kruiesssucdb utpe(real multiple by sourced universes lker J. wicueitrcigflisadscalar and fluids interacting include ew I. cito ihuie oain explicit notation, unified with scription elementary of terms in expressed are ich ieetepesospeeti the in present expressions different n matter .Ohreatsltos24 23 solutions exact Other 2. potentials Exponential 1. udpu tfffli 23 fluid stiff plus fluid K fluid K K K K K K K w K 0 = Λ K K K K K 1. 2. 3. K 2 = 0 6= 0 = 0 = 0 = 0 6= 0 6= 0 = 0 = 0 = 0 = 0 = 0 = 0 = 0 = 0 = K K K K, 0 = +1 = = tfffli,any fluid, stiff , / K , , , , utpu aito 22 radiation plus dust , , tfffli plus fluid stiff , n elfli plus fluid real any , , , , aito plus radiation , utplus dust , , 3 0 = Λ 0 = Λ Λ 0 = Λ 0 = Λ 0 = Λ 0 = Λ 0 = Λ Λ 0 = Λ − , 0 = , 0 = Λ 0 6= lsasi udpu ut21 dust plus fluid stiff a plus 1 0 = Λ , , K 0 = Λ 0 = Λ , w , aito plus radiation , aito lssi ud22 fluid stiff plus radiation , w , aito lssi atr19 matter stiff plus radiation , utpu tffmte 18 matter stiff plus dust , utpu aito 15 radiation plus dust , any , 0 = Λ any , 0 = 1, Λ 1 w , 2 = ‡ 2 = , Λ w , w , lssailcurvature spatial plus 0 = Λ fli 13 -fluid w w / fli 13 -fluid fli 13 -fluid / ∈ 3 ∈ K 3 Λ Λ udpu aito 20 radiation plus fluid Q tions Q udpu stiff plus fluid Λ 0 = Λ P 2 2 = 0 = Λ ρ 2 / 3 29 21 20 19 17 17 18 16 16 15 14 14 14 13 12 2

C. Scalar field interacting with a perfect fluid 30 I. INTRODUCTION D. Einstein frame formulation of scalar-tensor gravity 31 E. Two interacting scalar fields 32 Analytical solutions of the Einstein-Friedmann equa- IX. Concluding remarks 33 tions describing cosmology in the context of Einstein’s Acknowledgments 34 theory of gravity have been known since its early days, beginning with those of de Sitter [152], Friedmann [183, A. Derivation of Eq. (2.31) 34 184], Lemaître [246–249], and Einstein himself, including the famous and ill-fated Einstein static universe [162] (see B. Equivalence of Eqs. (6.13) and (6.12) 35 the textbooks [94, 156, 165, 235, 238, 256, 296, 322, 398] and see Ref. [35] for a popular exposition). C. Single-fluid limit of the universe with dust plus radiation and K = 1 36 ± 1. Positive curvature 36 The qualitative analysis of the phase spaces of solutions 2. Negative curvature 36 of the relevant differential equations provide much information about the cosmological dynamics with multiple References 37 fluids (e.g., [127, 174, 178, 368, 372, 388, 389, 397]). How- ever, analytical solutions are often needed for the inves- tigation of physics beyond the cosmic dynamics. Various physical motivations to look for exact solutions include the need for toy models in theoretical research; testing numerical codes; describing the smooth transition between different eras (most notably, between radiation- and matter-dominated epochs, between inflation and radiation, or the onset of dark energy domination); obtaining solutions of the many inflationary scenarios of the early universe; modelling the acceleration of the cosmic expansion in the present era. The latter requires the simultaneous analysis of dark energy and dark/baryonic matter, and two-component dark energy models have also been proposed [194, 201, 392, 419] (in particular, the quintom model consisting of two interacting scalar fields, one regular and one phantom, has received much attention [14, 15, 77–82, 113, 158, 170, 179, 180, 202, 242, 243, 250, 251, 292, 294, 335, 346, 349, 350, 355– 361, 399, 400, 403, 404, 410–412, 420–424], see [75] for a review).

Most solutions of the Einstein-Friedmann equations found in the 1960s and 1970s and the conditions for their integrability are spread in a body of literature that spans over half a century and are being forgotten, as demon- strated by the fact that two- and three- (effective) fluid solutions found in the 1960s and 1970s have been re- investigated and rediscovered in recent years. Moreover, the theoretical motivation for studing multiple fluids or effective fluids has changed: scalar fields have assumed a primary role in cosmology with the advent of the theory of inflation and with the introduction of the dark energy concept to explain the 1998 discovery that the present expansion of the universe is accelerated [325, 326, 336]. This situation motivates reviewing analytical multi-fluid solutions of the Einstein-Friedmann equations from a modern point of view, including scalar fields and interacting fluids in the picture, the study of which was much less motivated in the 1970s. 3

II. FIELD EQUATIONS, FLUIDS, AND SCALAR Assuming the FLRW line element, the spatial symme- FIELDS tries reduce Eq. (2.1) to the Einstein-Friedmann equations We restrict our discussion to cosmology and analytical 8π Λ K solutions in the context of Einstein gravity, referring the H2 = ρ + , (2.6) 3 3 − a2 reader to [89, 148, 171, 186, 309, 370] for scalar-tensor (including f(R)) cosmology and to Refs. [123, 207, 212] a¨ 4π Λ = (ρ +3P )+ , (2.7) for more general theories of gravity. We follow the no- a − 3 3 tation of Ref. [398]: the signature of the metric tensor gab is +++ and units in which the speed of light c and ρ˙ +3H (P + ρ)=0 , (2.8) Newton’s− constant G are unity are used. The Einstein equations read where H a/a˙ is the Hubble function and an overdot denotes differentiation≡ with respect to the comoving time 1 t. Eq. (2.8) follows from the covariant conservation of Rab gabR +Λgab =8πTab , (2.1) b − 2 the matter energy-momentum tensor Tab = 0. Only two of the three Einstein-Friedmann∇ equations are in- where Rab and Tab are, respectively, the Ricci tensor and dependent. If two of them are given, the last one can the matter stress-energy tensor, while R gabR is the ≡ ab be deduced from them. Without loss of generality, we Ricci scalar and Λ is the cosmological constant. In or- can regard the Friedmann equation (2.6) and the energy der to discuss solution methods, one must necessarily re- conservation equation (2.8) as independent and the ac- strict the scope. We begin with the simplest situation, celeration equation (2.7) as derived. and the one most common in the research literature on If a fluid is not coupled explicitly to other matter or multi-fluid cosmology, consisting of two non-interacting1 fields and has equation of state (2.4), the conservation perfect fluids with stress-energy tensors equation (2.8) implies that its energy density scales as T (1) = (P + ρ ) u u + P g , (2.2) ab 1 1 a b 1 ab ρ(0) ρ(a)= (2.9) a3(w+1) (2) Tab = (P2 + ρ2) uaub + P2gab , (2.3) with ρ(0) a constant. Because fluids with different equa- c 2 where u is the common four-velocity, ρ1,2 are the energy tion of state parameters w scale differently with the scale densities, and P1,2 are the isotropic pressures. A constant factor a, in the presence of multiple fluids one of them linear barotropic equation of state will come to dominate the cosmic dynamics if the universe keeps expanding. For example, if there are only P = wρ, w = const. (2.4) radiation (w = 1/3) and dust (w = 0) in an expanding universe, the radiation fluid dominates early on for 4 is assumed for both fluids. small a but its energy density ρr a− decays faster ∼ 3 By imposing spatial homogeneity and isotropy to sat- than the energy density of dust ρm a− , which comes isfy the Copernican principle, the Einstein equations ad- to dominate later on if a(t) always increases.∼ Therefore, mit as a solution the Friedmann-Lemaître-Robertson- it is customary to approximate the history of the uni- Walker (FLRW) geometries [94, 156, 235, 238, 256, 296, verse with a radiation era followed by a matter era. This 398]. The FLRW line element in spherical comoving co- approximation is adequate for many problems, but there ordinates (t, r, ϑ, ϕ) is are situations in which one is interested in the detailed two-fluid solution. dr2 ds2 = dt2 + a2(t) + r2dΩ2 , (2.5) In cosmology, common terminology refers to a dark − 1 Kr2 (2) energy fluid if 1 w < 1/3 (which, according to − the acceleration− equation≤ (2.7)− implies that the universe where dΩ2 dϑ2 + sin2 ϑ dϕ2 is the line element on the accelerates, a>¨ 0), and to phantom matter if w< 1. (2) ≡ − unit 2-sphere, K is the curvature index (usually normal- If the relation (2.4) between the cosmic fluid pressure ized to 0, 1), and a(t) is the scale factor encoding the P and energy density ρ is not constant, one can still ± expansion history of the universe. We will also use the define an effective equation of state parameter weff ≡ conformal time η defined by dt adη. P/ρ. Non-linear barotropic equations of state P = P (ρ) ≡ have been studied in the literature, especially in relation with phantom fluids and sudden future singularities [20, 21, 34, 36, 48, 67, 68, 83, 87, 182, 195, 233, 304–306, 363, 1 Sec. VIII discusses interacting fluids. 365, 376] but here we restrict ourselves to linear equations 2 Again, the assumption that the two fluids are collinear is not of state. mandatory: mutually tilted fluids are the subject of a consider- The energy density of a matter component ρ can be able literature (e.g., [126, 128–131, 301, 302, 367, 393, 397]). expressed in terms of the dimensionless density parame- 4 ter tions, which necessarily commit to specific values of the ρ parameters and initial conditions, when analytical formu- Ω , (2.10) las are instead available. Our goal is to provide a bird’s ≡ ρ c eye view of this area of multi-fluid cosmology. 2 As said, the integration of the Einstein-Friedmann where ρ (t) 3H is the energy density of a spatially flat c 8π equations usually leads to solutions expressed by ellip- universe with≡ K =0 and Ω=1. tic integrals or hypergeometric functions which cannot The standard model of cosmology, the so-called Λ be reduced to simpler forms. These expressions are not Cold Dark Matter or ΛCDM model consists of a spa- useful in practice, making for a rather sterile catalogue tially flat FLRW universe containing approximately 70% of formal situations in which numerical solution of the dark energy (with a measured equation of state param- Einstein-Friedmann equations (2.16)-(2.18) is more con- eter consistent with the cosmological constant signature venient for practical purposes. Here we focus on situa- w = 1) and 30% matter, modelled as a dust, adding up tions in which the field equations are integrable and the to Ω − =1. tot solution can be expressed explicitly in the form of a finite Due to their non-linearity, there are significant math- number of elementary functions. ematical difficulties in solving analytically the Einstein- Friedmann equations of cosmology sourced by multiple fluids or effective fluids (the latter include the cosmological constant and spatial curvature, see below). It is A. Single fluid and effective curvature / Λ-fluids even more difficult to solve analytically the more general Einstein equations (without spatial homogeneity and If K = 0 and Λ=0 and there is a single fluid with isotropy) with two fluids as the matter source, for exam- constant equation of state (2.4), the well-known textbook ple in stellar models (e.g., [96, 97, 299]). In cosmology, solution for the scale factor is there are two approaches to the analytical solution: the 2 w first method attempts to integrate the Friedmann equa- a(t)= a0 (t t0) 3( +1) , (2.11) tion in comoving or in conformal time and deals with − an integral that, in general, can only be expressed in where a0 and t0 are constants and ρ(a) is given by terms or elliptic or hypergeometric functions. The sec- Eq. (2.9). Models with a single real fluid, possibly with ond method looks for solutions expressed in parametric cosmological constant and spatial curvature, were re- form employing the conformal time as the parameter. viewed and classified early on ([208, 218, 288, 289, 390], Also in this second approach, there is no guarantee that see also [156]). the integration can be performed in closed form in terms For a fluid mixture, one defines the total effective den- of elementary functions, which is instead the main goal sity by of the present paper. The two approaches are comple- 8π 8π Λ K 8π mentary: one approach may succeed in situations where ρ = ρ + (ρ + ρ + ρ ) , (2.12) the other fails, and vice-versa. 3 tot 3 3 − a2 ≡ 3 Λ K A large variety of situations, which is best described by qualitative methods and phase space analysis, can where present themselves for general forms of matter which may Λ 3K include perfect or imperfect fluids, tilted fluids [126, 128– ρΛ , ρK (2.13) ≡ 8π ≡−8πa2 131, 367, 393, 397], fluids with non-linear and/or non- constant equation of state [20, 21, 34, 36, 48, 67, 68, 83, can have any sign, as long as ρtot 0 (otherwise the 87, 182, 195, 233, 304–306, 363, 365, 376], scalar fields, Friedmann equation (2.6) cannot be satisfied≥ because the etc. One also has to distinguish between non-interacting left hand side is always non-negative). fluids and interacting (or explicitly coupled) fluids. There To be consistent in treating the curvature term K/a2 is now a large amount of literature devoted to the lat- in the Friedmann equation as an effective fluid, one− must ter, because of the hypothesis that dark energy and dark also attribute to it a pressure PK . This is done by im- matter may be coupled directly, which would in principle posing that the covariant conservation equation (2.8) be explain the coincidence between the orders of magnitude satisfied for this effective component of the cosmic fluid, of the density parameters ΩDE and ΩDM of these two which yields fluids. Even under the simplifying assumptions of spatial ρ K P = K = (2.14) homogeneity and isotropy, perfect fluids, and constant K − 3 8πa2 barotropic equations of state, the landscape of analytical FLRW solutions and their physics is rich and is cov- (this remark appears to have been made by Hughston & ered by a considerable amount of literature spanning five Shepley [218], McIntosh [288], and McIntosh & Foyster decades. The older literature is sometimes forgotten, [289] long ago and then apparently forgotten). Similarly, which results in analytical solutions and solution meth- it is customary to use the effective energy density ρΛ and ods being rediscovered, or in the use of numerical solu- pressure P , with P = ρ = Λ/(8π), for the cosmo- Λ Λ − Λ − 5 logical constant.3 The acceleration equation is automat- with a quintessence scalar field acting as dark energy, or ically satisfied since the combination ρK +3PK vanishes with modifications of gravity, among which f(R) gravity (as it should be, since the acceleration equation (2.7) does is very popular. In both cases, we have again a scalar not contain the curvature index K). By defining the total field dominating the dynamics of the universe. In the pressure late universe, this scalar field acts together with the dark matter fluid modelled as a zero pressure dust. Analytical Ptot = P + PΛ + PK , (2.15) solutions of the Einstein-Friedmann equations describing a scalar field and a fluid will be discussed in Sec. VII B: the Einstein-Friedmann equations can be rewritten as here we focus on a single scalar field. 8π H2 = ρ , (2.16) The cosmological literature offers also scalar field mod- 3 tot els of dark matter and unified dark energy-dark matter models [12, 23, 55, 58–60, 62, 133, 185, 205, 244, 245, 252, a¨ 4π = (ρtot +3Ptot) , (2.17) 275–278, 313, 339–341, 379, 380]. If scalar fields are truly a − 3 responsible for early inflation and late-time acceleration, then their physical nature amounts to two big mysteries ρ˙tot +3H (Ptot + ρtot)=0 . (2.18) of cosmology and fundamental physics. Certainly, scalar Formally, we have reduced spatially curved (K =0) uni- fields abound in high energy theories and gravitational verses, potentially with a cosmological constant6 Λ, to a scalars arise naturally in modifications of general relativ- fictitious spatially flat universe with vanishing Λ. The ity. The prototypical alternatives to Einstein theory are topology of the three-dimensional spatial sections can- Jordan-Brans-Dicke gravity [69, 222, 223] and its scalar- not change, of course, and this is only a formal trick tensor generalizations [54, 310, 396], in which the grav- used in the cosmology literature and textbooks to intro- itational coupling strength becomes a dynamical scalar field. While, with the exception of Sec. VIII D, we do not duce the total density parameter Ωtot ρtot/ρc. This formal reduction to a single effective fluid≡ is normally discuss these alternative theories of gravity here, there is little doubt that the scalar fields appearing in these the- not completed by including the pressure PK in modern literature and textbooks. Here, this extended definition ories, as well as in high energy physics, have contributed of effective curvature fluid is useful to study “simple” an- to the increasing use of scalars in cosmology. alytical two-fluid solutions of FLRW universes with any The energy-momentum tensor of a scalar field mini- spatial curvature, with or without Λ. mally coupled to the curvature is

(φ) 1 c Tab = aφ bφ gab φ cφ V (φ)gab , (2.19) B. Scalar field ∇ ∇ − 2 ∇ ∇ − where V (φ) is the potential energy density or self- Scalar fields are the simplest fundamental physical interaction potential. The stress-energy tensor (2.19) is fields: they abound in particle physics and are widely equivalent to that of a perfect fluid [25, 173, 176, 271, used in cosmology. Indeed, much of modern cosmology 272, 330, 354] with four-velocity is based on scalar fields as the form of matter propelling aφ the expansion of the universe. Inflation in the early uni- ua = , (2.20) ∇c verse [235, 257, 274, 296] is usually described as the ef- √ φ cφ fect of the dynamics of a single scalar field, dubbed in- −∇ ∇ c c flaton. Starobinsky inflation [374, 375], which histori- provided that the gradient φ is timelike (u is correctly normalized, ucu = 1). This∇ condition is always satis- cally was the first inflationary scenario (although Guth’s c − scenario [204] was much better known initially), is cur- fied in an unperturbed FLRW universe, where φ = φ(t) rently favoured by cosmological observations [5]. It re- in comoving coordinates in order to respect spatial ho- lies on quadratic corrections to the Einstein-Hilbert La- mogeneity. In general, if the scalar field has a potential grangian, but those can be reduced to an effective scalar V (φ), the effective equation of state of the equivalent field, as in all f(R) theories of gravity [148, 309, 370]. In fluid is dynamical. Choosing a potential V (φ) determines both standard and Starobinsky inflation, a single scalar a dynamical equation of state but there isn’t a one-to-one degree of freedom acts as the source of the Einstein- correspondence [47, 169, 271]. b (φ) Friedmann equations. What is more, the current accel- The covariant conservation equation Tab =0 gives eration of the cosmic expansion is commonly explained the Klein-Gordon equation obeyed by φ∇ dV φ =0 , (2.21) − dφ 3 Oddly, Harrison (who found many of the solutions reported in this review) distinguished between the cosmological constant and which reduces to a perfect fluid with w = −1, reporting solutions where only these two components source the Einstein-Friedmann equations simul- dV φ¨ +3Hφ˙ + =0 (2.22) taneously [208]. dφ 6 for a scalar φ(t) in FLRW space, or to C. Multi-fluid cosmology dV φ +2 φ + a2 =0 (2.23) Suppose that there are n non-interacting fluids with ,ηη H ,η dφ densities ρi and pressures Pi = wiρi, with wi = const. (i =1, 2, ..., n). Then the total density is in terms of conformal time, where a,η/a. In general, this equation is non-linear and relativelyH≡ few analytical n solutions are known. ρ = ρ (2.28) The effective energy density and pressure of a cosmo- tot i i=1 logical scalar field are X

2 and, according to Dalton’s law, the pressure is the sum (φ) φ˙ ρ = T uaub = + V (φ) , (2.24) of the partial pressures φ ab 2 n ˙2 ab (φ) Tii φ Ptot = wiρi . (2.29) Pφ = h Tab = = V (φ) , (2.25) gii 2 − i=1 X where ha is the projection operator on the 3-dimensional b The total effective equation of state parameter is defined spatial sections with Riemannian metric hab = gab+uaub. In general, the equation of state of the effective fluid as equivalent to φ is dynamical and time-dependent, how- P n P ever one can in principle impose that it remains constant w (a) tot = i=1 i tot ≡ ρ n ρ and determine the potential V (φ) that achieves this pe- tot Pj=1 j culiar situation (which may or may not not be physically n (0) wPρ a 3(wi+1) motivated), as done in [30, 373]. One obtains = i=1 i i − . (2.30) n (0) 3(wj +1) P j=1 ρj a− Pφ wφ = const. (2.26) P ≡ ρφ It follows that, in the presence of two or more perfect fluids, the effective equation of state of the mixture is time- if the kinetic and potential energy densities are propor- dependent even if the equation of state of each compo- ˙2 tional to each other, φ /2= αV (φ) with nent is constant. The time variation of the total effective equation of state parameter is given by [288] (w + 1) α = − φ (2.27) wφ 1 3H − w˙ = (w w )2 ρ ρ (2.31) tot −ρ2 i − j i j for wφ =1. The case wφ = 1 corresponds to V = 0 and tot i

D. Reduction to quadratures The symmetry transformations are given by [157]

Rewriting the Friedmann equation as (2.16), one ob- dt dt¯= f(ρ)dt , (2.37) → tains 1 f 2 ρ (1 f 2) ρ 8π ρ ρ¯ = − Λ+ = − ρ + , a˙ = a√ρ , (2.35) → 8πGf 2 f 2 f 2 Λ f 2 ± 3 tot r (2.38) where ρtot = ρtot(a). Formal integration yields [4πGf (8πGρ + Λ) f ′] P P¯ = ρ¯+ − (P + ρ) da 8π → − 4πGf 3 = (t t0) , (2.36) a ρtot(a) ±r 3 − Z [f 2 (ρ + ρ ) f ′] = ρ¯+ − Λ (P + ρ) , (2.39) p 3 where t0 is an integration constant. Usually, one cannot − f compute the integral in the left hand side explicitly in terms of elementary functions. Even when this is possi- with f(ρ) > 0 a dimensionless regular function (the dif- ble, one will have an analytical solution t = t(a) which, in ferential dt¯ = f(ρ(t))dt is exact because ρ = ρ(t)). The general, cannot be inverted to obtain a = a(t) explicitly. inverse is While physically there may be a big difference between dt¯ a source composed of two fluids (for example, a cosmic dt = , (2.40) radiation background that is still hot, inducing significant f microphysical effects in a universe already dominated by 2 2 dust) and a universe filled by a single fluid with exotic ρ = f ρ¯ + f 1 ρΛ , (2.41) − non-linear equation of state, the mathematics ruling the Einstein equations does not know the difference. From f 3 P + ρ = P¯ +ρ ¯ . (2.42) the purely formal point of view, one can define the total f 2 (ρ + ρ ) f − Λ ′ equation of state parameter for a universe sourced by two distinct fluids as in Eq. (2.30). Using

da da dt¯ da E. Symmetries of the Einstein-Friedmann a˙ = = f , (2.43) ¯ ¯ equations ≡ dt dt dt dt f da H = fH¯ , (2.44) In any physical theory, symmetries are important and ≡ a dt¯ the symmetries of the Einstein-Friedmann equations have been the subject of many works ([6, 7, 75, 77, 84, 88, and Eq. (2.41), the Friedmann equation (2.6) for the spa- 89, 98, 107–112, 114, 138, 140, 172, 333, 382, 399] and tially flat universe changes into referenes therein). A connection has been made between Einstein-Friedmann equations and methods of supersym- 8π Λ H¯ 2 = ρ¯ + , (2.45) metric quantum mechanics [311, 344, 345]. As in other 3 3 areas of physics, also in FLRW cosmology Lie, Noether, and other symmetries are often studied with the ultimate while the covariant conservation equation (2.8) becomes goal of generating first integrals of motion and exact so- [157] lutions [39, 66, 88, 89, 91, 92, 107, 120, 121, 154, 190, 314, 315, 317–319, 329, 387, 417], sometimes in relation dρ¯ +3H¯ P¯ +¯ρ with Penrose’s cyclic time cosmologies or conformal time dt¯ [51–53, 392]. The existence of such a symmetry does not automatically make the corresponding solution a physical f 2 (ρ + ρ ) f ′ = − Λ [ρ ˙ +3H (P + ρ)]=0 . one. f 4 A symmetry studied in [107, 314] (see also [6, 7, 98, (2.46) 108, 110–112, 114]) and generalized in [157] applies to K =0 universes sourced by perfect fluids. It consists of a rescaling of the comoving time t, the Hubble function H When Λ=0, one has and all first time derivatives, and of the fluid energy den- ρ sity and pressure which leaves the Einstein-Friedmann f(ρ)= , (2.47) ρ¯ equations invariant in form. The pressure is rescaled r differently than the energy density, which changes the equation of state. and the symmetry (2.37)-(2.39) reduces to the one of 8

Ref. [107] (see also [314]) given by s = 4/3, dust with w = 0 is mapped into radiation. If s =3/2 radiation becomes a stiff fluid. This symmetry is ρ ρ¯(ρ) , (2.48) → useful to generate fluid solutions with non-linear equation of state from seeds solution with linear one, but there is ρ¯ always the question of how physical the result of a formal H H¯ = H, (2.49) → ρ technique is. Generating new scalar field solutions for r new potentials starting from a given one, which would ρ dρ¯ be more interesting, proves difficult or impossible [157]. P + ρ P¯ +¯ρ = (P + ρ) , (2.50) → ρ¯ dρ r Another symmetry exists for spatially flat universes [177]: the transformation ρ f(ρ) = . (2.51) ρ¯ r a a˜ = aσ, (2.60) de Sitter universes with Λ > 0 are fixed points of the → transformation because, if P = ρ = 0, then P¯ = 3(w+1)(σ 1) 3(w+1)(σ 1) dt dt˜= σa 2 − dt = s a˜ 2σ − dt, ρ¯ = const., → − (2.61) 2 (1 f ) 3(w+1)(σ 1) ρ¯ = − ρ (2.52) ρ ρ˜ = a− − ρ, (2.62) f 2 Λ → with f = const. and, by changing units of time one leaves the form of the Einstein-Friedmann equations un- can set f to unity. These symmetries map perfect fluids changed and the perfect fluid retains its equation of FLRW universes with equation of state (2.4) into uni- state (2.4) with the same equation of state parameter w, verses with cosmological constant and fluids with non- in contrast with the previous symmetries. The symmetry linear equations of state [157] with σ =1, which leaves the comoving time unchanged, is the best known (e.g., [7, 107, 138, 172]). ρ f 2 (ρ + ρΛ) f ′ w¯ +1=(w + 1) 2 − . ρ + (1 f )ρΛ f − (2.53) One can impose that the new equation of state be con- III. TWO FLUIDS WITH CONSTANT stant, i.e., w¯ P/¯ ρ¯ = const. Then, denoting EQUATIONS OF STATE—COMOVING TIME ≡ w¯ +1 s = const. , (2.54) Let us consider the situation in which there are two ≡ w +1 non-interacting fluids with constant equations of state the non-linear ordinary differential equation P1 = w1ρ1, P2 = w2ρ2 with w1,2 = const. We use the subscripts m, r, and s to denote dust (non-relativistic (s 1)f 2 ρΛ matter), radiation, and a stiff fluid, respectively. It is f ′ + − + sf 1 f =0 (2.55) 2 (ρ + ρΛ) − 2ρ (ρ + ρΛ) assumed that the two fluids are collinear and not tilted with respect to each other, i.e., that they have the same must be satisfied by the transformation function f. One four-velocity uc in their stress-energy tensors. Since the possible solution is [157] spatial curvature and Λ are now treated as effective perfect fluids with constant equations of state, we include ρ + ρ f(ρ)= Λ , (2.56) spatially curved FLRW universes, with or without Λ, in γρs + ρ r Λ the discussion. 1 s In this section we restrict to the use of comoving time where γ is a constant with dimensions [γ]= ρ − . The energy density and pressure transform according t and we look for solutions of the Einstein-Friedmann to equations (2.16)-(2.18) of the form a = a(t) or t = t(a). The total energy density of the mixture is ρtot = ρ1 + ρ2, ρ¯ = γρs , (2.57) where the energy densities of the components scale as

s (0) (0) P¯ = γ [α(w + 1) 1] ρ , (2.58) ρ1 ρ2 − ρ1(a)= , ρ2(a)= , (3.1) a3(w1+1) a3(w2+1) and the equation of state parameter becomes (0) w¯ = s (w + 1) 1 , (2.59) and ρ are constants. Then − i For example, a de Sitter universe with w = 1 is mapped into another one for any α; if s =1, a perfect− fluid does 8π ρ(0) ρ(0) a˙ = 1 + 2 (3.2) not change the equation of state parameter (w¯ = w). If s 3w1+1 3w2+1 ±r 3 a a 9 and practice, we only need to assume that one of the parame- 1/2 ters wi Q; since we impose one of the conditions (3.12), 3w1+1 (0) (0) 2 3(w1 w2) − ∈ I da a ρ1 + ρ2 a − (3.19) below in our search for integrability, the second w- ≡ parameter is automatically rational if the first one is. Z h i 8π = (t t0) , (3.3) ± 3 − p+1 = r A. First condition: r n ∈ Z where t0 is an integration constant. This expression reduces to the single fluid equation (2.36) in the limit of a p+1 Imposing r = n Z in order to find integrability (0) ∈ single fluid ρ 0. cases in terms of elementary functions [103, 288], one 2 → For general values of the parameters w1,2, this integral obtains can be expressed in terms of the Gauss hypergeometric (2n 1) w 2nw =1 . (3.12) function 2F1 as − 1 − 2 q α p +1 p +1 β We can fix the first fluid (i.e., fix w1 Q) and look I = ap+1 F q; ; + 1; ar ∈ p +1 2 1 − r r −α for physically interesting values of the equation of state parameter w of the second fluid, as n Z runs. (3.4) 2 ∈ where α and β are constants and

p = (3w1 + 1)/2 , (3.5) 1. Dust plus a second (real or eﬀective) ﬂuid

q = 1/2 , (3.6) Suppose that the first fluid is a dust, w1 = 0, and we − look for a second fluid that gives “simple” integrability. Then, r = 3(w1 w2) . (3.7) − 1 1 1 w = (3.13) In general, this representation is not useful for practical − 2 ≤ 2 −2n ≤ 2 purposes in cosmology. However, as several authors (Jacobs when one of the fluids is a stiff fluid [220]; McIn- and, as n = , ... , 3, 2, 1, 1, 2, 3, ... , + , we ob- tosh [288] and MacIntosh & Foyster [289]; and, more tain the pairs−∞ − − − ∞ recently, Chen, Gibbons, Li & Yang [103]) have noticed, for special values of the parameters w1,2 the integral can (w1, w2) = (0, 0) (single dust fluid), be expressed in terms of elementary functions thanks ... to the truncation of the hypergeometric series or to (w1, w2) = (0, 1/6) , the Chebysev theorem of integration [101, 273]. This (w1, w2) = (0, 1/4) , theorem states that (w1, w2) = (0, 1/2) , (w , w ) = (0, 1/2) , 1 2 − The integral (w1, w2) = (0, 1/4) , − (w1, w2) = (0, 1/6) , ... − J dx xp (α + βxr)q r =0, p,q,r Q (3.8) ≡ 6 ∈ (w1, w2) = (0, 0) (again, a single dust fluid). Z (3.14) admits a representation in terms of elementary functions Both limits n produce a second dust, i.e., there is p+1 p+1 → ±∞ if and only if at least one of r , q, r + q is an integer. a single dust fluid in the FLRW universe. No value of w2 In our situation, we have particularly interesting from the physical point of view is obtained; although quintessence (but not phantom) fluids p +1 w1 +1 = , (3.9) are obtained this way, their equation of state parameter r 2(w1 w2) − w2 is not close to 1, as required by current observations − 1 [3]. q = , (3.10) −2

p +1 w2 +1 2. Radiation plus a second (real or effective) fluid + q = . (3.11) r 2(w w ) 1 − 2 Now suppose that the first fluid is radiation, w =1/3, The authors of [103, 220, 288, 289] have studied sys- 1 and we look for a second fluid giving integrability in terms tematically the simple integrability cases. To proceed we of elementary functions. Then, need w1 and w1 to be rational numbers, which always happens in cosmology, or else one can use the rational 1 n 2 w = − 1 (3.15) approximation to the equation of state parameter w. In − 3 ≤ 2 3n ≤ 10 and, letting n = , ... , 3, 2, 1, 1, 2, 3, ... , + , we arises, and the route followed by the authors of [103], who obtain the pairs −∞ − − − ∞ apparently were unaware of Refs. [288, 289]). This happens when [288, 289] (cf. Eqs. (22) and (24) of Ref. [288]) (w1, w2) = (1/3, 1/3) (a single radiation fluid), ... γ1 1 =1 , (3.20) (w1, w2) = (1/3, 5/9) , γ2 − m (w1, w2) = (1/3, 2/3) , where γ w +1 and m Z. This relation is equiv- (w1, w2) = (1/3, 1) (radiation plus stiff fluid), 1,2 ≡ 1,2 ∈ (w , w ) = (1/3, 1/3) (radiation plus spatial curvature),alent to 1 2 − (w1, w2) = (1/3, 0) (radiation plus dust), mw1 + (1 m)w2 = 1 . (3.21) (w1, w2) = (1/3, 1/9) , − − ... Since m Z, setting m =2n Z gives back Eq. (3.19). (w1, w2) = (1/3, 1/3) (again, a single radiation fluid). Let us∈ fix the first fluid and∈ look for a second fluid that (3.16) guarantees “simple integrability”.

3. Cosmological constant plus a second (real or effective) 1. Dust plus a second (real or effective) fluid fluid

If the first fluid is a dust, w1 =0, we have In this case w = 1, n disappears from Eq. (3.12), 1 − which gives w2 = 1, producing only a cosmological con- 1 − 1 w2 = 1 (3.22) stant with no other fluids. − ≤ −2n +1 ≤ and the pairs 4. Stiff matter plus a second (real or effective) fluid (w1, w2) = (0, 0) (a single dust fluid), ... In this case w1 =1, (w1, w2) = (0, 1/5) , (w , w ) = (0, 1/3) (dust plus radiation), (n 1) 1 2 0 w2 = − 2 , (3.17) (w1, w2) = (0, 1) (dust plus stiff fluid), ≤ n ≤ (w , w ) = (0, 1) (dust plus Λ), 1 2 − and we have the pairs (w1, w2) = (0, 1/3) (dust plus spatial curvature), (w , w ) = (0, −1/5) , 1 2 − (w1, w2) = (1, 1) (single stiff fluid), (w1, w2) = (0, 1/7) , ...... − (w1, w2) = (1, 4/3) , (w1, w2) = (0, 0) (again, a single dust fluid). (w1, w2) = (1, 3/2) , (3.23) (w1, w2) = (1, 2) , (w1, w2) = (1, 0) (stiff matter plus dust), (w1, w2) = (1, 1/2) , 2. Radiation plus a second (real or effective) fluid (w1, w2) = (1, 2/3) , ... If we start from radiation, w1 =1/3, we obtain (w1, w2) = (1, 1) (again, a single stiff fluid). (3.18) 2n 3 5 1 w = − (3.24) − ≤ 2 3(2n + 1) ≤ 3

p+1 + = B. Second condition: r q n ∈ Z and the pairs

(w1, w2) = (1/3, 1/3) (single radiation fluid ), Let us explore now the second possibility (3.11) of ex- ... pressing the integral I in terms of elementary functions. (w , w ) = (1/3, 3/5) , Imposing p+1 + q = n Z [103], one obtains 1 2 r ∈ (w1, w2) = (1/3, 7/9) , (w1, w2) = (1/3, 5/3) , 2nw1 1 w2 = − . (3.19) (w , w ) = (1/3, 1) (radiation plus stiff fluid), 2n +1 1 2 (w , w ) = (1/3, 1/9), 1 2 − Solving the Friedmann equation with an hypergeomet- (w1, w2) = (1/3, 1/15), ric function, McIntosh [288] and later McIntosh and Foys- (w1, w2) = (1/3, 1/7) , ter [289] noted that the latter truncates and simplifies to ... elementary functions when one of its arguments is an in- (w1, w2) = (1/3, 1/3) (again, a single radiation fluid). teger (which is the situation when the Chebysev theorem (3.25) 11

3. Λ plus a second (real or eﬀective) ﬂuid • radiation plus spatial curvature, Λ=0;

• K =0, Λ=0, dust plus radiation; Again, setting w = 1 makes n disappear from 1 − Eq. (3.19) and produces only w2 = 1: there is only a • K =0, dust plus Λ; cosmological constant in a spatially ﬂat− FLRW universe. • K =0, Λ=0, dust plus stiﬀ matter;

4. Stiff matter plus a second (real or effective) fluid • K =0, Λ=0, radiation plus stiff matter. Both single fluid solutions for K = 0, Λ=0 (studied Setting w1 = 1 (the equation of state parameter of a in [103] in D spatial dimensions) and6 K =0 FLRW uni- stiff fluid) yields verses with a single fluid and Λ = 0 can be obtained by regarding them as two-fluid solutions6 in a fictitious spa- 2n 1 1 w2 = − 3 (3.26) tially flat universe, provided that they fall within the list − ≤ 2n +1 ≤ of cases above. and the pairs

D. Exotic fluids (w1, w2) = (1, 1) (single stiff fluid), ... (w1, w2) = (1, 7/5) , In addition to the classic dust, radiation, non-diluting (w1, w2) = (1, 5/3) , Λ (with ρ˙Λ =0), and hypothetical phantom dark energy (w1, w2) = (1, 3) , which concentrates with the cosmic expansion (i.e., ρ˙ = (w , w ) = (1, 1) (stiff fluid plus Λ), 3(w + 1)ρ > 0 for w < 1), other exotic equations of 1 2 − − − (w1, w2) = (1, 1/3) (stiff fluid plus radiation), state are of physical interest. (w1, w2) = (1, 3/5) , A stiff fluid with w = 1 is believed to be appropriate (n, w2) = (1, 5/7) , to describe matter at high (nuclear) densities and it is ... used in astrophysics to model dense nuclear matter in (w1, w2) = (1, 1) (again, a single stiff fluid). the core of neutron stars. It is reasonable that, as the (3.27) universe cooled from high temperature and energy scales, it underwent a period at lower temperatures and nuclear densities described by the stiff equation of state. C. Summary A stiff fluid also corresponds to a free scalar field φ minimally coupled to the Ricci curvature. By setting In all cases, when n , the solutions degen- V (φ)=0 in Eqs. (2.24) and (2.25) it is Pφ = ρφ and erate into a single fluid→ solution. ±∞ If both conditions the scalar field behaves as a stiff perfect fluid. Then the (3.12) and (3.19) hold simultaneously, one necessarily has Klein-Gordon equation (2.22) admits the first integral w1 = w2 = 1, which is the case of a single cosmological − C0 constant with K =0 and no real fluids. φ˙ = , (3.28) a3 Based on the tables of values (w1, w2) obtained, in which we always find w2 1, one concludes that, when or one the fluids is phantom≥− and the other is dust, radi- C ation, or stiff matter, there are no analytical solutions 0 φ,η = 2 , (3.29) in the form t = t(a) or a = a(t) expressed in terms of a elementary functions (there can still be simple analyti- where C0 is an integration constant. Authors from the cal solutions in parametric form, or solutions given by a 1960s and 1970s, who were not particularly interested in hypergeometric function or elliptic integral, or solutions scalar fields,4 did not regard a stiff fluid as the realiza- with a time-dependent equation of state). The same con- tion of a scalar field and did not integrate Eq. (3.28) or clusion applies to quintessence fluids with realistic values Eq. (3.29). We will report the explicit form of φ in this of the equation of state parameter, other than Λ. In the situation, when it is given explicitly in terms of elemen- tables reported, “real” quintessence fluids (other than Λ) tary functions. A stiff fluid can also mimic anisotropy in have equation of state parameter w2 far from 1, while the universe [218, 220, 288, 293, 362]. current observations give w 1 [3]. − Looking at the previous tables,≃− the situations involving physically interesting fluids in which the two-fluid solution can be expressed in terms of elementary functions 4 Interest in scalar field cosmology arose with inflationary sce- are: narios in the 1980s [235, 257, 262, 274, 296] and, again, with the discovery of the present acceleration of the universe in 1998 • dust plus spatial curvature, Λ=0; [325, 326, 336]. 12

Finally, certain string-inspired effective gravitational A. K = 0, Λ 6= 0, any w ∈ Q field theories of the primordial universe [40–42, 280, 281] contain a massless Kalb-Ramond axion of gravitational For this combination of parameters, the Einstein- nature which is equivalent to an effective fluid with a stiff Friedmann equations (2.16)-(2.18) can always be inte- equation of state, and other stringy axions are possible. grated in terms of simple functions [103, 208, 288, 390]. These fields are linked to matter-antimatter asymmetry Let us assume that w = 1, 1/3, otherwise we fall into in the universe after inflation and to axionic dark matter. one of the cases previously6 − discussed.− The integral on the It is speculated that a stiff matter era precedes inflation left hand side of Eq. (4.2) has the form in these theories [281]. The value 1/3 of the equation of state parameter is 1/2 √3 da a 2 8πρ(0)a 3(w+1) +Λ − , (4.3) also motivated− in cosmology: a frustrated cosmic strings − − network produces an effective fluid with this equation of Z h i state [371, 394]. i.e., the form (3.8) with A fluid with equation of state parameter w = 2/3 − 1 describes domain walls and is interesting in cosmology p = 2 , r = 3(w + 1) =0 , q = , (4.4) [46, 71, 132] but, unfortunately, it does not appear in − − 6 −2 the lists of integrable cases found above. In general, a which are all rational if w is. The Chebysev theorem network of non-intercommuting n-dimensional topolog- applies if one of ical defects produces a fluid with effective equation of state parameter w = n/3 [371, 394]. p +1 1 p +1 3w +1 − = , + q = (4.5) The other values of the equation of state parameter r 3(w + 1) r −6(w + 1) w for which one has integrability in terms of elementary functions are not particularly interesting from the physi- is an integer. These conditions correspond, respectively, cal point of view, except perhaps as toy models. Among to these, solutions involving a w = 2/3 fluid were given by 1 Vajk [390]. w = 1+ , (4.6) − 3n 2 w = 1+ , (4.7) IV. TWO FLUIDS WITH CONSTANT − 3(2m + 1) EQUATIONS OF STATE—PARAMETRIC with n,m Z. These two conditions are mutually ex- SOLUTIONS WITH CONFORMAL TIME clusive and∈ they both allow to get arbitrarily close to the phantom divide w = 1 if n or m are sufficiently Several exact solutions, for both single- and two-fluid large, spanning both quintessence− | | ( 1| |w< 1/3) and solutions, can be obtained in parametric form using the phantom (w < 1) equations of state− ≤ near the− w = 1 conformal time η defined by dt adη as a parame- limit. − − ≡ ter. Two-fluid solutions in this form are perhaps not as Imposing the first condition (4.6) for integrability in widely known as single-fluid solutions (which are instead terms of elementary functions, as n spans the values reported in the pedagogical literature [94, 165, 168, 238, n = , ... , 3, 2, 1, 1, 2, 3, ... , + , one obtains the 264, 398]). corresponding−∞ − equation− − of state parameters∞ In terms of conformal time, the Friedmann equation 10 7 4 2 5 8 for a FLRW universe with a single matter fluid (satisfy- w = 1, ... , , , , , , , ... , 1 . (4.8) ing P = wρ, w = const.), cosmological constant Λ, and − − 9 −6 −3 −3 −6 −9 − spatial curvature reads Imposing instead the second condition (4.7), for m = , ... , 3, 2, 1, 0, 1, 2, 3, ... , + one obtains 8π ρ(0) Λa4 −∞ − − − ∞ (a )2 = + Ka2 , (4.1) ,η 3 a3w 1 3 17 11 5 1 7 13 19 − − w = 1, ... , , , , , , , , ... , 1 . − −15 − 9 −3 −3 −9 −15 −21 − which gives, along the lines of the discussion already (4.9) made using comoving time, The degenerate case w = 1 reproduces the empty, spatially flat, de Sitter universe,− while w = 1/3 gives a − da spatially curved universe filled with a string gas and with = (η η0) , (4.2) Λ already discussed. None of the other values of w corre- 8π ρ(0) Λa4 2 ± − Z 3w 1 + Ka 3 a − 3 − spond to dust, radiation, or stiff matter, but they cover q quintessence and phantom equations of state. There is where η0 is an integration constant. From now on, this no general formula expressing the solution that is valid section follows Refs. [103, 289] and we assume that the for all values of w corresponding to integrability in terms equation of state parameter w is rational. of elementary functions. 13

B. Any K, Λ = 0, any w ∈ Q 2. K = +1, Λ = 0,w-ﬂuid

This case gives rise to more physically interesting sit- In this case Eq. (4.10) yields uations. When the fluid is radiation or dust, this is 3w +1 a standard solution found in all cosmology textbooks f(a)= tan (η η ) . (4.12) ± 2 − 0 [94, 165, 238, 256, 398, 405]. For K = 0, the solution is elementary and given by Eq. (2.11). When K = 1 Using the expression of f(a) and squaring yields and there is only radiation, one can eliminate the param-± eter η to obtain a(t). 8π (0) (3w+1) 2 3w +1 ρ a− = cos− (η η ) (4.13) For K = 1 and dust, one can express both a and t 3 2 − 0 as simple functions± of η, but eliminating this parameter can only be done to obtain a relation t(a) that cannot be and finally [103, 208, 390] inverted explicitly to obtain a(t). 1 2 8π (0) 3w+1 3w+1 3w+1 For K = 1 and w = 0, 1/3, in general, one can a(η) = 3 ρ cos 2 (η η0) , compute a(η)±explicitly in6 terms of elementary functions, −  while t(η) is reduced to a quadrature and written as an  t(η) = a(η)dη . integral that may not have an explicit expression (see Sec. V). If the initial conditionR a(η =0)=0 is imposed, it must Explicitly, for any K, Λ=0, and w Q, it is possible be η0 = π/(3w + 1). The special cases of this solution to express a(η) in terms of elementary functions.∈ Various for dust and radiation were found by Friedmann in 1922 solutions have been proposed by many authors over the [183] and Tolman in 1931 [385], respectively. years (see the next sections). The general solution is given by [103] 3. K = −1, Λ = 0,w-fluid

1 tan− [f(a)] if K = +1 , − Two-(effective) fluid solutions for KΛ = 0 for dust and for radiation (separately) were reported6 in terms of  1 if K =0 ,  f(a) elliptic functions by Lemaître [249], Edwards [161] and 3w +1  Kharbedya [227]. The solution for K = 1, Λ=0 and (η η0)=  ± 2 −  1 f(a)+1 any w-fluid was reported by Tauber [384].− In this case,  tanh− f(a) 1  − the exponentiation of Eq. (4.10) gives if K = 1  f(a)+1 −  = 1 ln ,  2 f(a) 1 f(a)+1 (3w+1)(η η0)  − = e± − (4.14)  (4.10) f(a) 1  where  −

and it is straightforward to obtain

(3w+1)(η η0) 8π (0) (3w+1) 1+ e± − f(a)= ρ a− K. (4.11) f(a)= 3 − −1 e (3w+1)(η η0) r − ± − (3w + 1) = coth (η η ) . (4.15) − ± 2 − 0 = 0 Λ = 0 1. K , ,w-fluid Using the expression of f(a) and squaring yields [103, 390] For K =0, Λ=0, Eq. (4.10) yields [103, 208, 390] 1 2 a(η) = 8π ρ(0) 3w+1 sinh 3w+1 (η η ) 3w+1 , 3 2 − 0 1 2 a(η) = 2π ρ(0) 3w+1 [(3w +1)(η η )] 3w+1 ,  3 − 0  t(η) = a(η)dη .  3(w+1) (4.16) 1  (3w+1) 3w+1 3(w+1)  R  t(η) = 2π ρ(0) 3w+1 (η η ) 3w+1 If w > 1/3, the initial condition is a(η0) = 0. If  3(w+1) 3 0 −  − w = 1/3, we have an empty universe with hyperbolic 3- sections− and zero cosmological constant: this is the Milne  +t0 .  universe, which is nothing but Minkowski space sliced  e.g.  with a hyperbolic foliation ( , [165, 296, 398]). The initial condition at t = t0 (corresponding to η = η0) The most well known two-fluid solution in parametric is a(0) = 0; this universe expands forever and the fluid form is the one for dust plus radiation and K = 0 (see redshifts away if w> 1. Eq. (5.18)). − 14

V. EXPLICIT SOLUTIONS reads [13, 390]

In the following, we present the known explicit analyti- 8π (0) 2 8π (0) 1 3a t(a)= ρ a a + ρ sin− +t cal solutions of the Einstein-Friedmann equations (2.16)- 3 m 3 m (0) 0 −r − s8πρm ! (2.18) expressible in terms of elementary functions, (5.5) for two real or effective fluids with constant, linear, for K = +1 and barotropic equation of state P = wρ, w = const. Not all the situations, listed in the previous two sections, in 8π (0) 2 8π (0) 1 3a which integration in finite terms by means of elementary t(a)= ρm a + a ρ sinh− +t0 3 − 3 m (0) functions is possible are physically motivated. Therefore, r s8πρm ! we limit ourselves to the most physically significant val- (5.6) for K = 1 [13, 390]. Both solutions satisfy the Big ues of the equation of state parameters for the real fluids − (dust, radiation, or stiff fluid). We first provide the so- Bang initial condition a(t0)=0. lutions using comoving time t(a) or, whenever possible, The parametric forms of these solutions are [94, 256, their inverse a(t). Then, we provide the solution in the 398, 405] parametric form 4π (0) a(η) = ρm (1 cos η) , 3 − a = a(η) , (5.7)  4π (0) (5.1)  t(η) = ρm (η sin η)+ t0 ,  3 −  t = t(η) , for K = +1and using the conformal time η as the parameter, which may 4π (0) be useful in certain applications where conformal time is a(η) = 3 ρm (cosh η 1) , preferred for computational convenience (for example, to − (5.8)  solve the equations for scalar or tensor perturbations in 4π (0)  t(η) = ρm (sinh η η)+ t0 , slow-roll inflation [235, 257, 262, 274, 296]). We also re- 3 − port initial conditions, asymptotics, and single-fluid lim- for K = 1 where, in both cases, t = t0 corresponds to its for these solutions. η =0 and−a =0 and

(0) ρm 3K ρ (a)= , ρ (a)= − . (5.9) A. Dust, K = 0, Λ = 0 m a3 K 8πa2 The parametric forms make it clear that the solution for This well known single ﬂuid solution is 8π (0) K = +1 reaches a maximum size amax = 3 ρm at η = π 1/3 4π2 (0) (0) 2/3 (or t = 3 ρm ), while the K = 1 universe expands a(t)= 6πρm (t t0) (5.2) − − forever becoming curvature-dominated. ρ(0) 1 ρ(a)= m = , (5.3) a3 6π(t t )2 C. Radiation, K = 0, Λ = 0 − 0 or, in the less commonly encountered parametric form, This is another classic textbook solution

2π (0) 2 1/4 a(η) = ρm (η η ) , 32π 3 − 0 a(t)= ρ(0) √t t , (5.10) (5.4) 3 r − 0  2π (0) 3 t(η) = ρm (η η ) + t .  9 0 0 (0) − ρr  ρr(a)= , (5.11) The initial condition is a Big Bang a = 0 at t = t0, a4 corresponding to η = η . 0 or, in parametric form,

8π (0) a(η) = ρr (η η ) , B. Dust plus spatial curvature, Λ = 0 3 − 0  q (5.12)  2π (0) 2 In its parametric form using conformal time, the solu-  t(η) = ρr (η η ) + t , 3 − 0 0 tion is well known (Tauber [384] and Gilman [189] gave  q the solution for K =0, 1) and is found in standard cos- where tis an integration constant. The Big Bang initial ± 0 mology textbooks, e.g., [94, 256, 398]. It is rarer to ﬁnd condition a = 0 at t = t0 (or η = η0) has been imposed it expressed in terms of comoving time, in which case it and the universe expands forever. 15

D. Radiation plus spatial curvature, Λ = 0 to probe the thermal history of the universe, the physics must inserted in numerical calculations in detailed sce- This solution is well-known and found in standard cos- narios [10, 116, 117, 216, 321, 381, 418]. mology textbooks, e.g., [94, 256, 398] and as a special The solution with non-interacting dust and radiation case of more complicated solutions [13]. For K = +1, and Λ=0 was found by Tolman for K = +1 [386] and the scale factor is by Chernin [104] and McIntosh [286] for any K. The spatially flat, Λ=0, radiation plus dust solution 8π (0) 2 due to Jacobs [219] (see also [125, 390]) is probably the a(t)= ρr (t t ) (5.13) 3 − − 0 most well known solution with two real (as opposed to r effective) fluids [103]. In spite of its importance for the or, in parametric form, transition from the radiation to the dust era, it rarely appears [165] in modern textbooks. Setting w1 =0, w2 = 8π (0) 1/3, the relevant integral (3.3) becomes a(η) = 3 ρr sin η ,  q (5.14) da a 2  = ρ(0)a 2ρ(0) ρ(0)a + ρ(0)  8π (0) 2 m r m r t(η) = t0 3 ρr cos η . (0) (0) (0) − − Z ρm a + ρr 3 ρm q q  q  (0) This solution has a Big Bang at t = t0 8πρr /3 (or 8π − = (t t0) . (5.17) q (0) 3 − η = 0) and a Big Crunch at t = t0 + 8πρr /3 (or r η = π). q This K = 0 solution was given by Jacobs in 1967 [219] For K = 1, the scale factor is (see also [390]) as − (0) (0) (0) (0) 2 8π (0) t(a)= A ρm a 2ρr ρm a + ρr + t0 , (5.18) a(t)= (t t0) ρr (5.15) − − − 3 q r 1 (0) 2 − or, in parametric form, which coincides with (5.17) for A = √6π (ρm ) . In parametric form, the solution ish5 [13, 297, 390]i 8π (0) a(η) = 3 ρr sinh η , (0) 2π (0) 2 ρr (5.16) a(η) = 3 ρm η (0) ,  q − ρm   8π (0)  (5.19) t(η) = 3 ρr cosh η + t0 . (0)  2π (0) 3 ρr  t(η) = 9 ρm η (0) η ,  q − ρm There is a Big Bang at η = 0, corresponding to t =  (0) with the initial condition a(t = 0)= 0. This universe t + 8πρr /3. At late times t + , the solution is 0 → ∞ expands forever. The radiation fluid dominates at early approximatelyq linear, a(t) t. times, when a(t) √t t , while the dust dominates at ∼ ≃ − 0 late times, with a(t) (t t )2/3. ≃ − 0 = 0 Λ = 0 Let us consider the single-fluid limits. The limit to E. K , , dust plus radiation (0) dust is obtained trivially by setting ρr =0 in Eq. (5.17), obtaining the well known power-law a(t)= a (t t )2/3, Solutions describing dust plus radiation are, no doubt, 0 − 0 some of the most well-motivated because of the need i.e., Eq. (2.11) with w =0. to describe the simultaneous presence of non-relativistic The radiation-only limit is not as obvious: letting (0) matter and cosmic background radiation or, more in ρm 0 in Eq. (5.17) (or in the parametric form of this → general, neutrinos or ultra-relativistic particle species. solution) does not produce a meaningful result. Instead, (0) FLRW universes containing dust and radiation have been one has to take the limit ρm 0 in the integral (3.3), studied in the years immediately following the discov- which reduces to → ery of the cosmic microwave background [324], beginning with Alpher & Herman [16] and continuing with a 8π da = (t t0) , (5.20) the works of Chernin [104], Jacobs [219], Cohen [125], (0) 3 − Z ρr r Roeder [342], McIntosh [285], Vajk [390], Sapar [348], Sistero [366], May [279], Coquereaux & Grossman [124], q giving a(t) = a0√t t0 (or Eq. (2.11) with w = 1/3). and Dabrowski & Stelmach [139]. However this exact In general, the single-fluid− limit of a two- or three- (real solution for two non-interacting fluids cannot reproduce the complicated microphysics of ultrarelativistic particle species decaying, electrons combining with protons, etc. When an accurate description is needed, for example to 5 The authors of Ref. [13, 297] were apparently unaware of Vajk’s study imprints left in the cosmic microwave background paper [13, 297, 390]. 16 or effective) fluid solution in which radiation survives is more problematic than the analogous limit in which a non-radiative fluid remains, as will be seen in the follow- a0 3M √3Λ t ing. 2 a(t)= 2/3 1+ 1+ 3 e The parametric form of this spatially flat universe was 2 " s a0Λ ! rediscovered by Barrow & Saich [30] as a solution with 2/3 6 radiation plus a scalar field in the potential 3M √3Λ 2 t + 1 1+ 3 e− , (5.25) 2φ − s a0Λ ! # V e φ0 V (φ)= 0 (5.21) 2φ 2 φ e 0 + V1 where a0 = a(t0) at the present time t0, with V , φ constants, but imposing that the equation 8π 0,1 0 M =Ω H2a3 = ρ(0) , (5.26) of state of the effective fluid equivalent of the scalar field m0 0 0 3 m be constant, and φ˙2 = αV (φ) , (5.22) 2 ρm(t) 8πρm(t) Ωm(t) = 2 (5.27) ≡ ρc(t) 3H (t) which gives the effective equation of state parameter α 1 wφ = α−+1 for the scalar field fluid. As expected from is the dimensionless dust density parameter. At late the list (3.16), the case wφ = 0 equivalent to α = 1 is times the cosmological constant with constant energy integrable and gives the exact solution (5.19) for the po- density Λ/3 dominates over the dust that redshifts away 3 tential (5.21) [30]. The scalar field (not reported in [30]) as a− and the solution (5.25) converges to the de Sitter can be found by noting that the Klein-Gordon equation universe admits the first integral [30] 2/3 φ a0 3M ˙ 0 a(t + ) 1+ 1+ eH0t (5.28) φ = 3α (5.23) 2/3 3 a α+1 → ∞ ≃ 2 s a0Λ ! which, in our case, yields ρ = φ2a 3, dφ/dη = φ /√a φ 0 − 0 (where H = Λ/3 ), which is a phase space attractor. and 0 Indeed, this is the solution obtained by letting M 0 in Eq. (5.25). p → 3 2π (0) φ(η)= φ0 (0) arccosh (0) ρm η + φ1 , The limit to a spatially flat, dust-filled universe with- s2πρm s3ρr ! out cosmological constant is obtained straightforwardly (5.24) by a formal expansion of Eq. (5.25) as Λ 0, which where φ0,1 are integration constants. produces a(t) t2/3. → ∼ The K = 0 dust-plus-radiation universe was general- The solution for dust and Λ < 0 is obtained as the ized by Jacobs [220] to anisotropic Bianchi I models, but special case w2 =0 of Eq. (5.57): it is given in terms of elliptic integrals. The dust-plus- radiation solution describing K = 1 FLRW universes is ± ρ 1/3 3 Λ given by Eqs. (6.11) and (6.13) below. a(t)= m sin2/3 | | (t t ) . (5.29) ρ 2 − 0 | Λ| "p # F. K = 0, dust plus Λ There are a Big Bang at t = 0 and a Big Crunch at 2π t = t0 + . √3 Λ If Λ > 0, this is the standard ΛCDM model (with Λ as | | the simplest form of dark energy) and the analytical solution of the Einstein-Friedmann equations (2.16)-(2.18) is well-known (e.g., [13, 100, 103]): G. K = 0, radiation plus Λ

The solution for the spatially ﬂat FLRW universe containing radiation and with Λ > 0 is [208] 6 Barrow & Saich do not seem to have made the connection between their solution and one of the exact integrability cases (in 1/4 the list (3.16) of the Einstein-Friedmann equations—the results 8πρ(0) Λ of [220, 288, 289] discussed in our Sec. III are not mentioned in a(t)= r sinh 2 t . (5.30) v [30]. Λ ! u r 3 ! u t 17

It was rediscovered in the more complicated form [13] or, in parametric form,

1/4 1 2 3Γ 2 Λ t (0) a(t)= a + + a4 e √ 3 32π 0 0 a(η) = 3 ρs √η , √2" r Λ !  (5.39) 1/2  1/4  4 2π (0) 3/2 3Γ Λ  t(η) = ρs η + t0 . 2 4 2√ 3 t 3 3 + a0 + a0 e− (5.31) − r Λ ! #  The scalar field equivalent to the stiff fluid is (0) Λ ρr Λ = a2 cosh 2 t + + a4 sinh 2 t φ(t)= φ0 ln (t t0)+ φ1 , (5.40) v 0 0 − u 3 ! s ρΛ 3 ! u r r t (5.32) or where 3 4 2π 1/4 φ(η)= φ ln η + φ + φ ln ρ(0) . (5.41) 2 0 1 0 3 3 s a0 = a(0) , (5.33) " # Again, for K = 1 one cannot integrate t(η) in finite 2 4 8π (0) ± Γ=Ωr0H a = ρ , (5.34) form. For K = +1 we have [390] 0 0 3 r and a zero subscript denotes quantities evaluated at the 1/4 8π (0) present time. Again, this universe expands forever and a(η) = 3 ρs cos(2η) , (5.42) the cosmological constant dominates over the radiation  p at late times.  t(η) = dη a(η) , As for the single-fluid limits, a Taylor expansion of  Eq. (5.32) as Λ 0 reproduces the spatially flat radiative a universe with finiteR size that begins in a Big Bang and universe with zero→ cosmological constant and with scale ends in a Big Crunch. For K = 1, the solution is instead (0) − factor a(t) √t. Likewise, setting ρr = 0 reduces the ∼ 1/4 solution to the de Sitter space with pure Λ and scale 8π (0) a(η) = 3 ρs sinh(2η) , √ Λ t factor a(t)= a0 e 3 . (5.43)  p The solution for radiation and cosmological constant  t(η) = dη a(η) , can be presented also in the simple form [392] describing a universeR beginning in a Big Bang and eter- Λ nally expanding. a(t)= a sinh 2 t (5.35) eqv u r 3 ! u t where I. K = 0, stiff fluid plus Λ

(0) 1/4 8πρr One can alternatively regard this solution as being aeq = (5.36) Λ ! sourced by a stiff fluid plus Λ, or by a free scalar field plus Λ, or by a scalar field with constant potential is the value of the scale factor at the time of equivalence V (φ)=Λ/(8π). The solution is [100, 169] between the energy densities of radiation and of Λ (this is a special case of the more general solution (5.55)). The 1/3 a(t)= a0 sinh √3Λ t (5.44) expression of the solution in terms of conformal time requires elliptic integrals. where

(0) 1/6 1/6 Λ = 0 8πρ 4πC2 H. , stiff fluid, any K a = s = 0 (5.45) 0 Λ Λ ! These solutions are again given by Vajk [390]. For K = 0, we have the usual, forever expanding, solution (2.11) and C0 is the constant appearing in the first integral (3.28) of the Klein-Gordon equation. If a scalar field 1/6 a(t)= 24πρ(0) (t t )1/3 , (5.37) is the stiff fluid source, then [169] s − 0 (0) √3Λ t ρs φ(t)= φ ln tanh + φ (5.46) ρ(a)= , (5.38) 0 2 1 a6 " !# 18

1/2 with φ0 = (12π)− and φ1 is an integration constant. 208, 288]. For Λ > 0, the solution is This solution± has a Big Bang singularity at t =0, where 1 w a(t) t1/3 , (5.47) ρ(0) 3( 2+1) ≃ a(t)= 2 ρΛ !

√3Λ t 2 3(w +1) φ(t) φ0 ln . (5.48) (w2 + 1) 2 ≃ 2 ! sinh √3Λ (t t0) , × 2 − At late times t + , the solution asymptotes to the (5.55) phase space de Sitter→ ∞ attractor with constant scalar ﬁeld

Λ t where ρ =Λ/(8π) is the eﬀective energy density of the a(t) a e√ 3 , (5.49) Λ ≃ 0 cosmological constant. If w > 1, there is a Big Bang 2 − at t = t0. At late times the solution converges to the de φ(t) φ . (5.50) Sitter phase space attractor ≃ 1 The total energy density and pressure are 1 (0) 3(w2+1) 2 ρ φ˙ Λ 2 √Λ/3(t t0) ρ = , (5.51) a(t) e − . (5.56) tot 2 − 8π ≃ ρΛ !

φ˙2 Λ The limit Λ 0 reproduces, through a straightforward Ptot = + . (5.52) Taylor expansion→ of the hyperbolic sine, the single-fluid, 2 8π 2 w The effective equation of state of the mixture is time- no-Λ solution a(t)= a0 (t t0) 3( 2+1) . However, regard- − dependent, ing the other single-fluid limit, one cannot simply take ρ(0) 0 in Eq. (5.55) to obtain the empty de Sitter Ptot 2 w(t) =1 2 tanh2 √3Λ t , (5.53) space→ generated by Λ. ≡ ρtot − For Λ < 0 we have [208] interpolating between the stiff equation of state P ρ at ≃ 1 early times (when the cosmological constant has not yet (0) 3(w2+1) ρ2 had the time to influence the dynamics) and Ptot ρtot a(t)= ≃− ρ at late times when Λ dominates [169]. | Λ|! The limit Λ 0 in Eqs. (5.44) and (5.44) reproduces 2 → 3(w +1) the stiff fluid (or free scalar field) solution (w2 + 1) 2 sin 3 Λ (t t ) (5.57) × 2 | | − 0 1/3 a(t)= a0t (5.54) p (i.e., the scale factor (2.11) with w = 1). The scalar and there are a Big Bang at t0 and a Big Crunch at field (5.46), however, diverges in the limit Λ 0; the cor- → 2π rect free φ(t) can be obtained from the first integral (3.28) t = t0 + (5.58) ˙ 3 Λ (w2 + 1) which yields φ = C0/t and, finally, φ(t) = C0 ln (t/t0). | | (0) Similarly, setting ρs = 0 does not automatically re- if w > 1. p cover the de Sitter space corresponding to the surviving 2 − Again, the limit Λ 0 reproduces the single-fluid uni- Λ: one must notice that removing the stiff fluid is equiv- → 2 w alent to setting φ = const. while retaining its potential verse with scale factor a(t) = a0 (t t0) 3( 2+1) , but the (0) − Λ t simple limit ρ 0 in Eq. (5.55) fails to reproduce V (φ)=Λ/(8π), which generates a(t) e√ 3 . 2 ∝ anti-de Sitter space.→ For Λ=0 we have the usual, forever-expanding, solu- J. K = 0, any real fluid plus Λ tion (2.11)

2 3(w +1) Indeed, when K =0 and Λ =0, not only dust but any (0) 2 6 a(t)= (w2 + 1) 6πρ2 (t t0) , (5.59) single ﬂuid with constant equation of state parameter w2 − can be integrated explicitly using comoving time7 [103, q with a =0 at t = t if w > 1. 0 2 −

7 A similar solution with a speciﬁc value of the exponent of the hyperbolic sine for dust plus dark energy with a time-dependent K. K = 0, Λ = 0, dust plus stiﬀ matter equation of state is obtained in Ref. [332] after an ad hoc assumption on the functional form of the deceleration parameter q ≡ −aa/¨ a˙ 2 (which essentially amounts to a choice of the scale This possibility appears in both lists (3.18) and (3.23). factor). As a function of comoving time, the scale factor assumes 19 the simple form found by Vajk [390] and, integrating,

(0) 1/3 (0) (0) 2 ρs 3 ρ a(t)= 6πρ (t t0) , (5.60) 2 s m (0) t(a)= t0 a a + " − − ρm # (0)  v (0) ± s32πρr u ρr  u (0) (0) t ρm ρs ρ (a)= , ρ (a)= . (5.61) (0)  (0) m 3 s 6 ρ ρ a a s ln a2 + s + a (5.67) (0) v (0)  This universe expands forever, with the stiff fluid red- −ρr u ρr  u  shifting away considerably faster than the dust. This so- t  lution is a special case of the one found by Chavanis [100] (in this K =0 case, the scale factor a is dimensionless). and Dariescu et al. [145] for dust plus a stiff fluid plus At late times this universe is dominated by radiation and Λ, which is given by Eq. (6.9) with the constants α and a(t) √t. The parametric form of this solution is [390] β as in Eqs. (6.7) and (6.8). By taking the limit Λ 0 ∼ → and using the second order expansions sinh x = x + ..., (0) 8π (0) 2 ρs 2 a(η) = 3 ρr η (0) , cosh x =1+ x /2+ ... as x 0, Eq. (6.9) yields − ρr →  r 2 1/3  9βt  (0) a(t)= +3√αt (5.62)  η 8π (0) 2 ρs  t(η) = 2 3 ρr η (0) 4  − ρr  r  (5.68) 1/3  (0) 2 (0)  ρ(0) (0) = a0 6πρ t + 24πρ t + ... , (5.63)  s 3 8π 2 1 (0) (0) ln 3 ρr η − ρr 32πρr q  q q which reproduces the solution (5.60) for   (0)  8π (0) 2 ρs  + ρr η (0) + t0 ,  3 ρ (0)  − r ρs 1  r t = . (5.64)  0 (0)  s 6π ρm from which one sees that this universe begins in a Big Bang. Although the authors of both Refs. [145, 390] do not re- Equation (3.28) for the scalar field equivalent of the gard this as a free scalar field solution, it is straightfor- stiff fluid can be integrated giving φ(a). Using dφ/dt = ward to integrate Eq. (3.28), which yields a˙ dφ/da, one obtains φ 9β 3√α φ(t)= 0 ln + + φ , (5.65) dφ C0 3√α 4 t 1 = , (5.69) da (0) 2 (0) a ρr a + ρs where φ0,1 are integration constants. q (0) The single-fluid limit ρs 0 leaving only a dust, which integrates to applied to the scale factor (5.60)→ reproduces the correct 2/3 a(t) = a0 (t t0) and φ = const. disappears since − 2 (0) ρs = ρφ = Pφ = φ˙ /2 vanishes. However, one cannot 1 ρr φ(a)= φ tanh− 1+ a2 + φ , (5.70) (0) 0 v (0) 1 take the limit ρm 0 in Eq. (5.60) to obtain the pure u ρs  → u stiff fluid solution. t  (0) where φ = C / ρs and φ is an integration con- 0 − 0 1 L. K = 0, Λ = 0, radiation plus stiff matter stant. q (0) The limit ρs 0 in which the stiff fluid disappears, → This solution for K = 0, Λ=0, and radiation plus a leaving only radiation, transforms the relation (5.67) into 8 stiff fluid was given by Vajk [390]. We have (w1, w2)= the correct scale factor a(t)= a0√t t0. The other limit (0) − (1/3, 1), the Friedmann equation yields ρr 0 in Eq. (5.67) does not produce the corresponding stiff→ fluid solution, which has to be recovered directly 2 (0) da a 8π from the limit ρr 0 of the integral (5.66). I = = (t t0) (5.66) → (0) 2 (0) ± 3 − Z ρr a + ρs r q M. w = 2/3, Λ = 0

8 This solution has recently been rediscovered in its parametric This solution was given by Vajk [390] without apparent form, using conformal time as the parameter, in Ref. [17]. physical motivation and was reported by other authors 20

[288]. For K =0, it is the single-ﬂuid universe (2.11), The Big Bang occurs at

1/5 2 50π (0) 2/5 √ (0) (0) a(t)= ρ (t t0) , (5.71) 3 3 ρ2 ρ2 3 − t = t0 + (0) ln (0) (5.78) 4√π ρr ! 2ρr ! ρ(0) ρ(a)= 5 , (5.72) and the universe expands forever, becoming radiation- a dominated at late times. The parametric form of this or, in parametric form, solution in terms of conformal time is [390]

(0) 1/3 2/3 a(η) = 6πρ η , (0) (0) (0) ρ ρ 1 2ρ a2 + 2 a 2 cosh− r +1 (5.73) ρ(0) 2ρ(0) ρ(0)  r − r 2  3 (0) 1/3 5/3  r  t(η) = 6πρ η + t0 .  5  (5.79)  8π (0)  = ρr η , There is a Big Bang singularity a =0 at t = t0 (or η =0)  3 and the universe expands forever. q For K = 1, the comoving time cannot be expressed  t(η)= dη a(η) . ±  explicitly in terms of simple functions. When K = +1   (0) R we have [390] The limit ρ2 0 leaves only radiation; taking this limit in Eq. (5.76)→ reproduces the correct scale factor 1/3 a(η) = 8π ρ(0) cos2/3 3η , a(t) = a √t t for the single-fluid radiative universe. 3 2 0 − 0 (5.74) The limit ρ(0) 0, instead, fails to reproduce the corre-  r  t(η) = dη a(η) , sponding scale→ factor, which must be recovered directly from the usual integral. which has a Big BangR at η = π/3 and a Big Crunch at η = π/3. − For K = 1 the solution is [390] − O. K = 0, Λ = 0,w1 = 2/3 fluid plus stiff fluid 8π (0) 1/3 2/3 3η a(η) = 3 ρ sinh 2 , (5.75) This possibility appears as the second last entry of the  list (3.18). This solution was again given by Vajk [390]  t(η) = dη a(η) . using comoving time10 In this case there isR a Big Bang at η =0 and the universe expands forever. 1 3 ρ(0) t(a)= a + s 15 (0) v (0) s2πρ1 u ρ1 u N. K = 0, Λ = 0,w = 2/3 fluid plus radiation t (0) (0) 2 2 4ρs ρs 3a (0) a +8 (0) + t0 This case, with w1 =1/3, w2 =2/3, is the third entry ×  −  ρ1 ρ1 ! in the list (3.16). The comoving time form of this solution is again given by Vajk as9 [390]   (5.80) ρ(0) ρ(0) ρ (a)= 1 , ρ (a)= s , (5.81) 1 a5 s a6 3 3ρ(0) ρ(0) t(a)= a 2 a2 + 2 a (0) (0) v (0) or, in parametric form [390], s32πρr  − 2ρr ! u ρr u (0 (0  t ρs 2ρs (0) 2 a + a = 6πρ η , (0) (0) (0) ρ(0) ρ(0) 1 3 ρ ρ ρ 1 − 1 + 2 ln a2 + 2 a + a + 2  r q (5.82) 4 (0) v (0) (0)  ρr ! u ρr 2ρr   u t(η)= dη a(η) . t +t0 ,  (5.76)  The Big Bang occursR at

ρ(0) ρ(0) (0) 5/2 r 2 4 2 ρs ρr(a)= 4 , ρ2(a)= 5 . (5.77) t = t + , (5.83) a a 0 5 (0) (0) s3πρ1 ρ2 !

9 (0) (0) An error in the square root qa + ρs /ρr in the ﬁrst term on the right hand side of Eq. (5.76), present in the solution of 10 Vajk’s solution has a sign error in the square root, which was Ref. [390], was corrected in [288]. corrected by McIntosh [288]. 21 or at t =0 if the integration constant is chosen as A. Radiation plus Λ plus spatial curvature K 6= 0

(0) 5/2 4 2 ρs These solutions were found by Harrison [208] and sub- t = . (5.84) 0 (0) (0) sequently rediscovered [13]. For K = +1 and Λ > 0, we −5s3πρ ρ ! 1 1 have [208, 378] This universe expands eternally and is dominated by the 3 Λ w1 =2/3 fluid, which decays slower than the stiff fluid), a(t)= 1 cosh 2 (t t0) in its late history. r 2Λ ( − " r 3 − # We can find the scalar field φ associated with the stiff 1/2 fluid as a function of the scale factor a. By differentiating Λ Λ +2 sinh 2 (t t ) , (6.1) Eq. (5.80), one obtains Λ 3 0 r c " r − #) 2 dt 1 3 a where = (˙a)− = (5.85) da (0) (0) s8πρ1 a + ρs 9 (0) Λc = (6.2) ρ1 (0) r 32πρr which, substituted into dφ/dt =adφ/da ˙ , gives is the critical value of the cosmological constant required to have a static Einstein universe with the same amount dφ 3 dφ a2 = . (5.86) of radiation. da (0) dt (0) s8πρ1 ρs For K = 1 and Λ > 0, a + (0) − ρ1 r 3 Λ Equation (3.28) now yields a(t)= cosh 2 (t t0) 1 r 2Λ ( " r 3 − # −

dφ 3 C0 1/2 = (0) , (5.87) Λ Λ da (0) +2 sinh 2 (t t0) , (6.3) s8πρ1 ρs a a + (0) Λc 3 − ρ r " r #) r 1 Both solutions begin with a Big Bang at t = t and which integrates to 0 asymptote to the de Sitter space with scale factor a(t)= Λ (0) a0 exp 3 t at late times. 1 ρ1 φ(a)= φ tanh− 1+ a + φ , (5.88) 0 v (0) 1 For K =q +1 and Λ < 0, the solution is [208] u ρs  u t  3 Λ where φ are integration constants. a(t)= 1 cosh 2 | | (t t0) 0,1 2 Λ 3 (0) s ( − " r − # The single-fluid limit ρs 0 in Eq. (5.80) reproduces | | → 2/5 1/2 the correct scale factor a(t)= a0 (t t0) , but fails on − Λ Λ the expression (5.88) of the scalar field: this happens +2 | | sinh 2 | | (t t ) . (6.4) 0  because the limit corresponds to the value C0 =0 of the s Λc " r 3 − # integration constant in the first integral (3.28) leading to  Eq. (5.88).  B. K = 0, Λ plus a stiff fluid plus dust

VI. THREE (REAL OR EFFECTIVE) FLUIDS SOLUTIONS The FLRW universe with stiff matter plus dust plus Λ was found by Vajk [390] and, like several others, was rediscovered recently [100, 145]. Using the slightly different Known three-fluid solutions include: notation a 6 • (w1, w2, w3) = (1/3, 1, 1/3), radiation plus Λ ρ = ρ(0) 0 , (6.5) plus spatial curvature;− − 1 1 a 3 • (w1, w2, w3) = ( 1/3, 0, 1/3), spatial curvature (0) a0 − ρ2 = ρ , (6.6) K =0 plus dust plus radiation; 2 a 6 • (w1, w2, w3) = ( 1, 0, 1), or Λ plus a dust plus a 8π − α = ρ(0)a6 , (6.7) stiff fluid; 3 1 0

• (w1, w2, w3)=(1/3, 2/3, 1), or radiation plus P2 = 8π β = ρ(0)a3 , (6.8) 2ρ2/3 fluid plus a stiff fluid. 3 2 0 22 the solution for Λ > 0 is in the alternative form presented by Vajik [390]

8π 8π 3β t(a)= a2 + ρ(0)a + ρ(0) a(t)= cosh √3Λ t 1 3 m 3 r 2Λ − r h i 1/3 (0) 3α √ 4π (0) 1 a +4πρm /3 + sinh 3Λ t . (6.9) ρm sinh−   Λ − 3 (0) ) 4π (0) 4π (0) 2ρr r ρm − ρm +  3 3 ρ(0)   r m  This universe, which begins with a Big Bang at t = 0,  h i  asymptotes to the de Sitter phase space attractor a(t)= +t0 (6.13) √ Λ t a0 e 3 at late times. (see Appendix B for a proof of the equivalence), or in the parametric form11 [297, 390] C. K 6= 0, dust plus radiation

The solution for dust plus radiation in a spatially ﬂat 8π (0) 4π (0) universe is given by Eq. (5.18) using comoving time or a(η) = ρr sinh η + ρm (cosh η 1) , 3 3 − by Eqs. (5.19) in parametric form using conformal time.  q The corresponding solutions for spatially curved FLRW  8π (0) 4π (0)  t(η) = ρr (cosh η 1)+ ρm (sinh η η) , universes were found in [104, 125, 219, 286, 289, 390] 3 − 3 − (see also [165, 296]). The solution for K = +1 is given  q (6.14)  by (e.g., [125]) with the initial condition a(t =0)=0. (0) (0) When ρm or ρr are set to zero, the solutions reduce,

8π (0) 8π (0) 2 of course, to single-fluid spatially curved universes filled t(a)= t0 ρr + ρm a a only with radiation or matter, respectively. In the case − r 3 3 − of matter-only, however, it is not trivial to reproduce the single-fluid expressions (5.5) for K = +1 and (5.6) for 4π (0) 4π ρm + a (0) (0) 1 3 K = 1 as ρr 0. Appendix C discusses these limits. + ρm sin−  −  . 2 − → 3 (0) (0) The solutions of the Einstein-Friedmann equations for 4π ρ + 8π ρ  3 m 3 r  dust plus radiation with spatial curvature and Λ =0 were  r  6  (6.10) classified qualitatively by Payne [320].

This universe begins at a Big Bang, reaches a maximum size, and ends in a Big Crunch singularity. The parametric form of this solution is [13, 297, 390] D. K 6= 0, Λ = 0, radiation plus stiﬀ ﬂuid

4π (0) 8π (0) The spatially curved FLRW universes filled with radi- a(η) = ρm (1 cos η)+ ρr sin η , 3 − 3 ation plus a stiff fluid were given by Vajk in parametric  q form [390]. For K = +1,  4π (0) 8π (0)  t(η) = 3 ρm (η sin η)+ 3 ρr (cos η 1) , − − (6.11)  (0) For K = 1 the solution is [125, 289, 390] 4π (0) 4π (0) 4π (0) 2ρs − a(η) = 3 ρr + 3 ρr 3 ρr + (0) sin(2η) ,  s ρr  r 8π (0) 8π (0) 2  t(a)= t0 + ρr + ρm a + a  t(η) = dη a(η) . r 3 3  (6.15) 4π 8π 8π This cosmosR begins at a Big Bang, expands to a max- ρ(0) ln C ρ(0) + ρ(0)a + a2 + a m r m imum size, and collapses in a Big Crunch. The scalar − 3 ( "r 3 3 field equivalent of the stiff fluid can only be expressed in 4π terms of η, or of a, by means of elliptic integrals. + ρ(0) , (6.12) 3 m For K = 1, instead, we have the forever-expanding − where C is a constant (missing in [125]) needed to make the argument of the logarithm dimensionless. This universe begins at a Big Bang and expands forever, becom- 11 ing dust-dominated. The expression (6.12) can be cast Ref. [13] contains an error in this parametric solution. 23 universe VII. SCALAR FIELD SOLUTIONS

(0) A. Single scalar field 4π (0) 4π (0) 2ρs a(η) = ρr ρr + (0) sinh(2η) 3 3 ρr  r 1/2 We have already seen that a minimally coupled scalar  4π ρ(0) ,  3 r field is equivalent to a perfect fluid which, in general, has  −  a dynamical effective equation of state and that a free t(η) = dη a(η) scalar field is equivalent to a stiff fluid. Therefore, all   (6.16) stiff fluid solutions are automatically free scalar solutions,  R which, during its history, is first dominated by the stiff and we have provided the corresponding expression of fluid, then becomes radiation-dominated, and ends in a φ(t) when it is given explicitly in terms of elementary curvature-dominated era. functions. Originally, the study of analytical solutions of scalar field cosmology was motivated by early universe inflation. When the scalar field potential V (φ) is part of a high energy theory, exact solutions are usually not available E. K = 0, Λ = 0, radiation plus P = 2ρ /3 fluid plus 2 2 because of the non-linearity of the Klein-Gordon equa- stiff fluid tion (2.22) and phase space analyses are most informative about the dynamics (e.g., [127, 372, 397]). Several formal The solution for radiation plus a P2 =2ρ2/3 fluid plus solutions exist in the literature, although some of them a stiff fluid was again given by Vajk [390]. It reads are mostly of mathematical interest. Only solutions that are attractors in phase space, or that display particular physical properties (such as evading slow-roll constraints (0) (0) (0) part of the time, providing exact spectral indices or de- 3 3ρ ρ ρ t(a)= t + a 2 a2 + 2 a + s sired expansion histories a(t), etc.) are usually important 0 (0)  (0) v (0) (0) s32πρr − 2ρr ! u ρr ρr from the physical point of view. Many of the analytical  u t solutions found are only valid in restricted regions of ini- (0) 2  (0) (0) (0) tial conditions and parameters, usually due to the fact 3 ρ ρ ρ ρ + 2 s ln a2 + 2 a + s that variables changes are needed which are restricted to 4 (0) (0) v (0) (0)  ρr ! − ρr  u ρr ρr regions in which the scalar field φ(t) is monotonic and u   t invertible. Therefore, most of these solutions should be ρ(0) regarded as toy models (which may still be quite useful), +a + 2 (6.17) (0) with the notable exception of solutions that behave as 2ρr #) attractors in phase space. The following solutions12 have been found in the and describes a Big Bang universe which expands forever, context of inflationary scenarios of the early universe with the stiff fluid, then the w2 = 2/3 fluid, and then [235, 257, 274] (we do not report approximate solutions (0) radiation becoming dominant. In the limit ρs 0 in here). → which the matter content reduces to the w2 = 2/3 fluid plus radiation, Eq. (5.76) is reproduced. The parametric form corresponding to the uni- 1. Exponential potentials verse (6.17) using conformal time as the parameter is [390] The power-law inflationary scenario is described by the exact solution [1, 28, 72, 254, 268, 298]

(0) (0) (0) p 2 ρ2 ρs ρ2 a(t)= a0t , (7.1) a + (0) a + (0) (0) ρr ρr − 2ρr  r   φ(t)= φ0 + α ln ( t) , (7.2)   (0) (0)  1 2a+ρ2 /ρr 8π (0)  cosh−   = ρr η , 4π φ  (0) 2 3  × ρ (0) p mpl  2 4ρs V (φ)= V0e± , (7.3) v (0) (0) q q  u ρr ! − ρr    u    t    t(η)= dη a(η) .   (6.18) 12 These solutions have been used as seeds to generate correspond-  (0) R In the limit ρs 0 it reproduces Eq. (5.79) for a w2 = ing universes in scalar-tensor gravity by means of the conformal 2/3 fluid plus radiation.→ transformation from the Einstein to the Jordan frame (e.g., [2]). 24 where mpl is the Planck mass, V0, α, and p are constants, H0 and a0 are positive constants, and φ0 is an integration with p > 1 in order to have cosmic acceleration a¨ > 0. constant. The dimensionless density parameter The de Sitter solution is obtained if φ = const., which corresponds to a cosmological constant [28, 72, 298]. The 4πφ2 Ω(t)= 1+ 1 tanh2(H t) (7.13) Klein-Gordon equation with exponential potential has a 2 0 H0 sinh (H0t) long history in mathematics, beginning with early studies of the non-linear wave equation by Liouville in 1853 approaches unity at late times when the solution enters [267]. Power-law inflation has the advantage that the a slow-roll regime. spectral indices of scalar and tensor perturbations are Yet another exact solution by Ellis & Madsen [164] also determined exactly. for any curvature index K exhibits the unusual linear Spatially curved (K = 0) versions of power-law infla- expansion tion are possible, but these6 solutions cannot be put in explicit form even though the field equations can be in- a(t)= a0t , (7.14) tegrated explicitly [164]. The exponential potential is a trademark of higher- φ(t)= φ φ ln t , (7.15) dimensional compactified theories and of the subsequent 0 ± 1 transformation to the Einstein conformal frame and po- 2(φ φ ) tentials consisting of a sum of exponential terms appear 2 − 0 V (φ)= φ e φ1 , (7.16) in supergravity and superstring theories. A sum of ex- 1 ponentials is typical of perturbation expansions in super- with a0 a positive constant, string theories [312]. Two-exponential potentials 1 K V (φ)= V e2λφ + V e 2λφ 2V V (7.4) φ2 = 1+ (7.17) 0 1 − 0 1 1 4π a2 − 0 (with V positive constants) were studied in relation 0,1 and constant dimensionless density parameter Ω. This with Noether and Hojiman symmetries and other con- (non-slow-roll) solution can be regarded as a modern ver- served quantities [90, 106, 115, 149–151, 155]. sion of the Milne universe which is filled by a scalar field Kruger and Norbury [237] found the spatially flat and allows for any spatial curvature [164]. By contrast, FLRW solution the original Milne universe is a hyperbolic (K = 1) foliation of empty Minkowski spacetime [296]. − V (φ)= V0 [1 + cosh(λφ)] , (7.5) Easther [159] studied a potential of the form

1 eλ√V0 t +1 N φ(t)= ln , (7.6) λj γφ λ√V0 t V (φ)= V e (7.18) λ e 1! j − − j=1 1/3 X a(t)= exp 2λ V0 t 1 , (7.7) and gave solutions for K = 1 for a,t,φ in parametric − ± h p i form, recovering a previous solution of Ref. [312] and for 2 1 K =0. ρ (a)=2V 1+ + . (7.8) φ 0 a3 a6

Ellis and Madsen [164] found a solution, for any cur- 2. Other exact solutions vature index K, for the potential

2 The intermediate inflation scenario contains another 3H0 2 2 2H0 V (φ)= + φ sinh (φ φ0) (7.9) exact solution [29, 30, 298]: 8π 1 φ − 1 (determined a posteriori), which produces m2 β2 V (φ)= 1 , (7.19) a(t)= a sinh (H t) , (7.10) φβ − 6φ2 0 0 2 m β/2 H0t − φ1 e 1 ρ(a)= β/2 (ln a) , (7.20) φ(t)= φ0 ln H t − , (7.11) (2β) ± H0 e 0 +1 2 β+4 β+4 where φ = φ 2 A(t t ) , (7.21) 0 ± − 0 2 1 2 K φ1 = H0 + , (7.12) γ 4π a2 a(t)= a0 exp [α(t t0) ] , (7.22) 0 − 25 where m is a mass scale, β is a constant, and goes from unity to infinity and back to unity. Maartens, Taylor & Roussos used the number of e- β (β + 4) m A = , (7.23) foldings (or ln(a/ai), where ai is the scale factor at the 2√3 beginning of inflation), or ultimately the scale factor it- β +2 self, as the independent variable in order to find new γ =1 , (7.24) analytical solution of scalar field cosmology [270]. Their − β +4 solutions interpolate between exponential or power-law 2 A β+4 inflationary expansion and the radiation era, providing α = . (7.25) an effective description of the exit from inflation. Al- 2β though reheating and entropy production at the end of Various solutions were provided by Ellis & Madsen inflation cannot be taken into account by such a sim- [164] with the purpose of obtaining any cosmic history ple model, these analytical solutions offer a quick way to a(t) that could be reconstructed from cosmological ob- model a complicated transition. servations, or to evade slow-roll. Their method con- Methods to solve the coupled Einstein-Friedmann- sists of imposing the form of a(t) and then solving for Klein-Gordon equations produced other exact inflation- φ(t) and V (t), inverting to obtain t(φ), and deriving ary solutions [22, 31–33, 93, 99, 105, 163, 181, 206, 224, V (φ) = V (t(φ)) a posteriori. Their first solution is de 237, 255, 258–261, 316, 347, 353, 373, 416, 425]. In gen- Sitter expansion for K 0, given by eral, all these solution-generating methods have some ≥ restrictions and limitations, and they necessarily focus a(t)= a eH0t , (7.26) 0 on the mathematics, leading to new solutions which may not be very relevant for the physical aspects of φ 1 H0t the accelerated cosmic expansion during inflation or late φ(t)= φ0 e− , (7.27) ± H0 quintessence domination. We refer the reader to the re-

2 cent review by Martin, Ringeval & Vennin [274] for a 3H0 2 2 comprehensive survey of inﬂationary scenarios with exact V (φ)= + H0 (φ φ0) , (7.28) 8π − and approximate solutions of the relevant ﬁeld equations. where

K B. Scalar field plus fluid φ1 = 2 (7.29) s4πa0 Since a stiff fluid is equivalent to a free scalar field, and a0,H0 are positive constants, while φ0 is an integra- solutions describing universes sourced by a single perfect tion constant. The density parameter is [164] fluid decoupled from a free scalar field have already been 2 given as two-fluid solutions one of which obeys the stiff 4πφ1 2H0t Ω(t)=1+ 2 e− . (7.30) equation of state P = ρ. Whenever the scalar field can H0 be integrated explicitly, its expression has been provided. This solution enters a slow-roll regime only at late times, Let us turn now to the physical motivation for such so- as Ω 1 and reduces to the usual de Sitter solution with lutions. cosmological→ constant (contained in the potential (7.28)) Early universe inflation, which is believed to be driven and constant scalar field φ if K =0. by a scalar field in the early universe [235, 257, 262, 274], Another solution for K > 0 is [164] provided the first motivation for searching solutions that describe FLRW universes sourced by a single fluid plus a(t)= a0 cosh(H0t) , (7.31) a scalar field. However, until 1998 the interest was more mathematical than physical. Following the 1998 discov-

2φ1 H0t ery of the present acceleration of the universe made with φ(t)= φ0 arctan e , (7.32) ± H0 type Ia supernovae [325, 326, 336] and the introduction of dark energy in theoretical cosmology, many models of 2 3H0 2 2 2H0 scalar field (modelling dark energy) and a single fluid (a V (φ)= + φ1 sin (φ φ0) , (7.33) 8π φ1 − dust modelling dark matter) were introduced. The most popular ones (at least before the idea of modifying grav- where a0,H0 are positive constants, ity at large scales to get rid of the ad hoc dark energy al- 1 K together [148, 309, 370]) were, and remain, quintessence φ2 = H2 (7.34) models consisting of a scalar field (the “quintessence”) in- 1 4K a2 − 0 0 teracting with the dark matter fluid (a dust) only gravi- and the density parameter tationally. Indeed, the study of these models with inverse α power-law potentials V (φ) = V0/φ , α > 0, begun with 2 4πφ1 2 the work of Peebles & Ratra [323] predating the discov- Ω(t)= 1+ coth (H0t) (7.35) H2 cosh2(H t) ery of the cosmic acceleration. We refer the reader to 0 0 26 the excellent book by Amendola & Tsujikawa [18] for a [167, 331] of the form proper account of these quintessence models. d2b λ Pre-1998 analytical solutions include those of Barrow + Q(τ)b = , (7.37) & Saich [30] and Chimento & Jakubi [106]. Barrow & dτ 2 b3 Saich solved the Einstein-Friedmann equations for a spa- where λ is a constant and Q(t) is an arbitrary func- tially flat FLRW universe sourced by a scalar field in a tion. In fact, considering a spatially flat FLRW universe potential V (φ) and a perfect fluid. They imposed that sourced by a perfect fluid with a constant equation of the scalar field effective equation of state be constant and state P = wρ = wρ(0)/a3(w+1) and a minimally coupled determined the potential a posteriori. scalar field φ not interacting directly with it and with Attempts to solve the Einstein-Friedmann-Klein- potential V (φ), the Einstein-Friedmann equations (2.16)- Gordon equations with a potential usually amount to (2.18) can be combined to give imposing a certain scale factor (usually, but not always, power-law or exponential), substituting it into the field a¨ a˙ 2 ρ(0) = 4π φ˙2 + (w + 1) . (7.38) equations, and solving them while determining a scalar a − a2 − a3(w+1) field potential V (φ). The last step requires expressing V as a function of time t and inverting the functional rela- Setting n 3(w + 1) and a b2/n (with b > 0) and tion φ = φ(t) to find t(φ) and then V (φ) = V (t(φ)). In changing the≡ time coordinate≡ to τ defined by dτ = bdt, practice, this is not always possible to do analytically. In Eq. (7.38) becomes [209] this approach à la Synge [383], the potential V (φ) is not determined by physical considerations but is ad hoc. d2b dφ 2 2πn2ρ(0) +2πn b = , (7.39) For example, with loose motivation from inflation, dτ 2 dτ − 3b3 Ht Méndez [290] imposed a(t) = a0e while allowing for spatial curvature and a perfect fluid with constant equa- which does not contain V (φ) and is of the form (7.37) tion of state parameter in the range 0 w 1. He found with no solutions for K 0 since [290] ≤ ≤ ≤ dφ 2 Q(τ)=2πn , (7.40) dτ t 2Ht K e− ′ 3(w+1)Ht′ 2 (0) φ(t)= φ0 dt′ 2 (w + 1) ρ0 e− , 2πn ρ ± 0 s 4πa0 − λ = . (7.41) Z − 3 (7.36) and found one solution for K > 0, but the relation φ(t) Given two linearly independent solutions b1,2(τ) of the cannot be inverted to give V (φ) [290]. There are actu- associated linear homogeneous equation ally more analytical solutions φ(t) for K > 0 that can d2b be expressed in terms of elementary functions, as can be + Q(τ)b =0 , (7.42) seen by applying the Chebysev theorem to Eq. (7.36). dτ 2 Moreover, one can in principle find solutions for K 0 by allowing a phantom equation of state (w < ≤1) the general solution of the non-linear Ermakov-Pinney but invertibility of φ(t) is usually not possible. Mén-− equation (7.37) is [331] dez turned instead to a viscous imperfect fluid with 2 2 1/2 bulk viscosity proportional to a power of its energy den- bP (τ)= Ab1(τ)+ Bb2(τ)+2Cb1(τ)b2(τ) , (7.43) sity, finding analytical solutions with scalar field that is an inverse exponential of time and quadratic potential where A, B, and C are constants satisfying 2 V = V0 + V1 (φ φ0) . (This reproduces the solution − 2 λ (7.26)-(7.28) of Ellis & Madsen [164].) However, the AB C = 2 (7.44) chain of assumptions jeopardizes the physical significance − W of these solutions. and the Wronskian Hawkins & Lidsey discovered that the Einstein- db2 db1 Friedmann equations (2.16)-(2.18) describing the dynam- W = b1 b2 (7.45) ics of a spatially flat universe sourced by a perfect flu- dτ − dτ ids plus a scalar field can be reformulated in terms is constant [331]. of the non–linear Ermakov–Pinney equation [209]. Er- The Ermakov-Pinney equation is closely related to makov systems consisting of two coupled, second–order, the one-dimensional Schrödinger equation. Setting now non–linear ordinary differential equations arise in many n/2 ψ a− , changing the variable t to σ defined by fields of physics. In one dimension (the case of FLRW dσ ≡= ψ(t)dt, Eq. (7.38) assumes the Schrödinger form cosmology, in which the scale factor a(t) depends on only one variable), the Ermakov system reduces to a d2ψ + [E P (σ)] ψ(σ)=0 , (7.46) single (“Ermakov–Pinney” or “Milne-Pinney”) equation dσ2 − 27 where E 2πn2ρ(0)/3 and dict non-Gaussianity in the spectrum of density pertur- ≡− bations. The search for non-Gaussianity by measuring n- dφ 2 point correlation functions (with n 3) is an active line P (σ)=2πn (7.47) ≥ dσ of research in observational cosmology [8, 9, 26, 27, 37, 38, 44, 45, 56, 57, 73, 74, 118, 166, 217, 229, 230, 265, 266, is the potential. 269, 291, 300, 327, 337, 338, 351, 352, 395, 401, 414, 415]. In practice, the Ermakov-Pinney equation is solved by Since, in principle, one can mimic a perfect fluid with choosing the scale factor a(t), solving for φ(t), inverting constant equation of state with a scalar field φ by impos- this relation to obtain t(φ), and then determining the ing that its effective equation of state be constant and potential V (φ) = V (t(φ)) [209]. As a simple check, for determining its potential V (φ) accordingly, as done by φ = const. (which implies V (φ) = const., equivalent to Barrow & Saich [30], all two-fluids exact solutions pre- a cosmological constant Λ), one recovers Harrison’s so- sented in Secs. V and VI can in principle be reproduced lution (5.55) for a single fluid with Λ [209]. In general, in this way, although the two potentials determined in the potential V (φ) may turn out to be physically unmo- this way for the two scalar fields will, in general, not be tivated and care must be taken to restrict to situations physically motivated. A potential is necessary because a of physical interest. free scalar field with V (φ) 0 plus a stiff fluid is equiv- ≡ Several authors have used the Hawkins-Lidsey reduc- alent to a single stiff fluid and does not give rise to new tion to an Ermakov-Pinney equation [197–200, 328]. solutions beyond what already seen in Sec. V. D’Ambroise studied further the equivalence between the Einstein-Friedmann-Klein-Gordon system with a perfect fluid and non-linear Schrödinger equations [143]. Seven VIII. TWO INTERACTING FLUIDS exact solutions for a scalar field and a single barotropic fluid P = wρ with w = const. were found [143]. The modern interest in two interacting fluids in cos- The case of a scalar field and two perfect fluids with mology arises in relation with the dark energy problem. constant equation of state Pi = wiρi was studied in [236] Early interest in the 1960s and 1970s was motivated by using a non-linear Schrödinger equation representation, the need to model the conversion of radiation into non- and the D’Ambroise single-fluid solutions were general- relativistic matter in the early universe. We proceed with ized to include the second fluid. Again, the potentials some general considerations applied to interacting dark V (φ) found do not appear to be physical. energy scenarios before moving to the early literature. The reductions to the Ermakov-Pinney equation and to linear and non-linear Schrödinger equations have been applied also to spatially curved FLRW universes A. Interacting dark energy and dark matter [197, 409], anisotropic Bianchi models [142], braneworld models [209], and quantum cosmology [119, 144, 343]. The standard ΛCDM model of the universe contains, There is also a connection with analogue gravity: in the in addition to a small fraction of baryonic matter, dark K = +1 case, the Ermakov-Pinney equation establishes energy and dark matter, the two most abundant con- an analogy with two-dimensional Bose–Einstein conden- stituents the nature of which is completely unknown [4]. sates [263]. Dark energy and dark matter, the two biggest mysteries of ΛCDM cosmology, are normally treated as two sep- arate physical sectors. However, it has been speculated C. Multiple non-interacting scalar fields [64] that these two sectors could couple explicitly in the so-called “interacting dark energy” scenarios. The phe- Inflation with multiple non-interacting scalar fields nomenological implications of such an interaction would was studied before the 1998 discovery of the present be significant: the coincidence problem of why the cur- acceleration of the universe [325, 326, 336], when it rent densities ΩDE and ΩDM of dark energy and dark seemed that the total dimensionless density parameter matter are of the same order of magnitude today, al- Ωtot was approximately 0.3, i.e., before the inclusion though they evolve differently, would be alleviated. In in the standard ΛCDM model of dark energy bring- interacting dark energy models, the interaction is mod- ing Ωtot up to unity. In fact, Ωtot = 1 is a robust elled by modifying arbitrarily the equations of motion for prediction of single scalar field inflation and a natural dark energy and dark matter, described as interacting way to obtain inflation with Ωtot < 1 consists of us- perfect fluids. In this picture, these fluid stress-energy ing multiple inflaton fields. These models of multiple tensors are not covariantly conserved separately but the (tot) (DE) (DM) field inflation with Ωtot < 1 were practically abandoned total Tab = Tab + Tab is: soon after 1998. Modern interest in multiple field inflation originates from the fact that these two-field models b (DM) T = Qa , [8, 9, 26, 27, 37, 38, 56, 57, 73, 74, 118, 166, 265, 266, ∇ ab (8.1) 269, 291, 300, 327, 351, 352, 395, 401] and multiple field models [44, 45, 217, 229, 230, 337, 338, 414, 415] pre- bT (DE) = Q , ∇ ab − a 28 so that cannot have spatial components in comoving coordinates and points in the direction of comoving time, i.e., b (tot) Tab =0 . (8.2) ∇ qc = α(t)uc , (8.8) The four-vector Qa describes phenomenologically the in- where α is a function of time (with α 0 to keep qc teraction. Its detailed form is picked by hand in the lit- ≥ erature, reflecting our ignorance about the dark sectors, future-oriented). while a proper description would express the unknown The two fluids are imperfect fluids, but by all means physics of the interaction (see Refs. [24, 187, 426] for not in the usual sense of dissipative fluid used in the rela- e.g. attempts to derive the vector Qa from physical consider- tivity literature ( , [160, 377]). Usually the dissipative ations). What is more, it is not even clear how to provide term qaub + qbua in the imperfect fluid stress-energy tensor describes a purely spatial (therefore, non-causal) heat a Lagrangian description of the interaction, as should al- c c flux density q with q uc = 0 [160, 377]. Here, instead, ways be done in fundamental physics. With the exception c c of Einstein frame scalar-tensor gravity (see Sec. VIII D), q must point parallel to u , otherwise it violates spatial isotropy. The traces of the fluids stress-energy tensors we lack even an effective field theory Lagrangian for this (i) purpose, in spite of substantial theoretical efforts to de- Tab are scribe dark energy (and its contender, modified gravity) T (i) = ρ +3P 2α . (8.9) via effective field theories [65, 196, 212]. − i i ∓ The standard description in the large literature (e.g., (tot) (1) (2) While the total stress-energy tensor Tab = Tab + Tab [18, 43, 61, 70, 86, 102, 122, 134, 137, 188, 210, 211, 221, (i) 228, 253, 307, 308, 369, 391, 406, 413, 427], see [364, 402] is covariantly conserved, each component Tab of the mix- for reviews) assumes a spatially flat FLRW universe filled ture satisfies by two perfect fluids satisfying the equations bT (i) = u ub P + u u b (ρ 2α)+ P ∇ ab a ∇b i a b∇ i ± ∇a i ρ˙1 +3H (P1 + ρ1)= Q, (8.3) + (P + ρ 2α) ub u i i ± ∇b a

ρ˙2 +3H (P2 + ρ2)= Q. (8.4) − + (P + ρ 2α) u bu . (8.10) i i ± a∇ b where the quantity Q(t) models the interaction. The Here i =1, 2, the upper sign corresponds to fluid 1, and explicit forms of Q(t) adopted in the literature are arbi- the lower one to fluid 2. By projecting this equation on trary: usually this quantity is assumed to be a function the time direction one obtains of one or more of a(t), a˙(t),H(t),ρ1,2(t) or of powers of a b (i) them [43, 61, 70, 86, 102, 122, 134, 137, 188, 210, 211, u Tab = (ρ ˙i 2α ˙ )+3H (Pi + ρi 2α) , (8.11) 221, 228, 253, 307, 308, 369, 391, 406, 413, 427]. At- ∇ ± ± a tempts can be made to constrain the form of Q(t) using with the usual notation ρ˙i u aρi. If we require (i) ≡ ∇ cosmological observations [134, 210, 211, 413]. that ua bT = 0, the two fluids are covariantly con- ∇ ab The total fluid with energy density ρtot = ρ1 + ρ2 and served separately, but their perfect-fluid components pressure Ptot = P1 + P2 obeys the conservation equation (Pi + ρi) uaub + Pigab are not because ua b [(P + ρ ) u u + P g ]= 2 α˙ + α ub . ρ˙tot +3H (Ptot + ρtot)=0 . (8.5) i i a b i ab b ∇ ± ∇ (8.12) The extensive literature on interacting dark energy In a FLRW universe this equation assumes the form leaves the basic questions of a covariant and Lagrangian ρ˙ +3H (P + ρ )= 2 (α ˙ +3Hα) (8.13) formulation almost completely unanswered, presenting i i i ∓ models already set up in comoving coordinates in the and one naturally reads the right hand side as Q(t), FLRW universe. A speculative covariant description reproducing Eqs. (8.3) and (8.4). Then α(t) and± Q(t) of the possible dark energy-dark matter interaction has are related by been proposed in Ref. [175]. There, the two fluids are Q(t) described by the modified energy-momentum tensors α˙ +3Hα + =0 (8.14) 2 T (1) = (P + ρ ) u u + P g + q u + q u , (8.6) ab 1 1 a b 1 ab a b b a or, equivalently,

(2) 1 d 3 Q(t) T = (P2 + ρ2) uaub + P2gab qaub qbua , (8.7) αa + =0 , (8.15) ab − − a3 dt 2 a where u is a common 4-velocity pointing in the time yielding direction of the observers comoving with both fluids. The four-vector qc describes an energy flux density between 1 α(t)= dt a3(t)Q(t) . (8.16) the two fluids. In a spatially isotropic FLRW universe, qc −2a3(t) Z 29

A possible physical interpretation is proposed in [175]: interpolating between 1/3 at early times and 0 at late fluid 1, which is is not a perfect fluid, has effective energy times [286, 287], for example [286] density and pressure 1 w (t)= , (8.22) (1) a b tot α T u u = ρ +2α = ρ , 3(1+ t/t0) ab 1 6 1 where α and t0 are constants. May & McVittie al- 1 P = T (1)hab , (8.17) lowed for the possibility that Pm = 0 (all these early 1 3 ab works assumed Λ=0). Instead of prescribing6 a function while, for fluid 2, wtot(t), they introduced a non-negative function φ such that a˙ = g(φ), where [282] T (2)uaub = ρ 2α = ρ , (8.18) ab 2 − 6 2 cot φ if K = +1 ,

1 (2)  P = T hab , (8.19) g(φ)= 1/φ if K =0 , (8.23) 2 3 ab   coth φ if K = 1 . The two fluids are not perfect fluids because their stress- − (i)  enery tensors Tab contain the terms (qaub + qbua) de- φ satisfies the differential equation [282] scribing an energy transfer occurring± simultaneously at dg all points of space without three-dimensional flux. The + g2 = K (8.24) energy lost by fluid 1 per unit time and per unit vol- dφ − ume is transferred to fluid 2, according to the splitting and the scale factor is given by (8.3)-(8.4). An alternative interpretation [175] is the following: 3w (t)+1 a(t)= tot . (8.25) when α > 0, the term 2 (α ˙ +3Hα) appearing in fluid 1 2φ˙ corresponds to a dust with zero pressure and energy density 2α which transfers energy instantaneously to fluid 1 The covariant conservation equation for the fluid mixture taking it from fluid 2. According to fluid 2, a dust with ρ˙tot +3H (Ptot + ρtot)=0 is equivalent to negative energy density 2α removes energy to transfer d 3wtot(t)+1 it to fluid 1. This second− dust obviously violates the weak =2g(φ) , (8.26) dt ˙ energy condition, but this problem is not more significant φ than the lack of causality of usual imperfect fluids with which determines φ(t) and a(t) once wtot(t) is assigned. spacelike heat currents [160, 377]. The (non-)conservation equations for radiation and mat- The relation (8.15) between the quantities Q(t) and ter are α(t) becomes, in this picture, ρ˙r +3H (Pr + ρr)= Qr , (8.27) 2 αa3 ˙ Q(t)= 3 . (8.20) − a ρ˙m +3H (Pm + ρm)= Qm , (8.28) For a three-dimensional region with unit comoving vol- where Qr is the radiation energy per unit volume con- ume and physical volume a3, 2αa3 is the energy trans- verted into matter energy per unit time and ferred between the two fluids− in this volume, (2αa3)˙ is the energy transfer rate, while Q(t) is the rate− of energy Qr + Qm =0 . (8.29) transferred per unit volume. A mixture of non-interacting radiation and dust (Pm =0) is obtained as the trivial case Qr = Qm =0, with B. Interacting radiation and non-relativistic matter 1 wtot(t)= (0) . (8.30) ρm 1+ (0) a Early work on interacting fluids was motivated by the ρr need to model the conversion of the cosmic fluid from Eq. (8.24) can be rewritten as [282] radiation (with energy density and pressure ρ , P ) to r r d dt 2g(φ) non-relativistic matter (with ρm and Pm = 0) and is ln (3w + 1) = (8.31) dφ tot dφ 3w +1 mainly due to Davidson & Narlikar [141] and McIntosh tot [284, 286, 287] for spatially flat universes, May & McVit- and integrated twice to tie [282] for K = 1, 0 and again to May & McVittie − 3w +1 g [283] for K = +1. McIntosh assumed an ad hoc form of tot = B exp 2 dφ , (8.32) the total equation of state parameter φ˙ 3wtot +1 Z

Ptot Pr dφ dφ′ g wtot(t) = (8.21) t t0 = B exp 2 (8.33) ≡ ρ ρ + ρ − 3wtot +1 3wtot +1 tot r m Z Z 30 with B and t0 integration constants [282]. Then Historically, interest in a scalar field directly coupled Eq. (8.25) gives to matter can be found in the Hoyle-Narlikar steady state theory [214, 215] in which matter is created by the so- B g a(t)= exp 2 dφ . (8.34) called C-field (for “creation”), a scalar field tranferring its 2 3wtot +1 energy into matter. Although is usually referred to as a Z scalar-tensor theory, the energy transfer from the scalar Using C-field requires only a direct coupling in the context of g general relativity [214, 287]. As the steady-state theory F exp 2 dφ , (8.35) was abandoned by the cosmological community, the in- ≡ − 3w +1 Z tot terest in a direct coupling between a scalar and ordinary one has matter waned. In the 1980s literature devoted to reheating the uni- 3F 2 2 verse after it is cooled by the inflationary expansion, an ρtot(t)= 2 g + K , (8.36) 2πB interaction between the inflaton scalar field and radia- 3F 2 tion was introduced in a phenomenological way in order P (t)= w g2 + K , (8.37) to model the decay of the inflaton into ultrarelativistic tot 2πB2 tot particles. This process should happen due to the infla- and ton’s coupling to other particles and should reheat the universe to allow for primordial nucleosynthesis and the ρ = ρ +3ρ(0)F 2 g2 + K , (8.38) m − r tot formation of what is now the cosmic microwave back- where ground [234, 235]. (Later research produced more so- phisticated scenarios for ending inflation, most notably 1 ρ(0) . (8.39) parametric amplification during the so-called preheating, tot ≡ 2πB2 see Refs. [11, 19] for reviews.) The phenomenological in- By further introducing [282] teraction has the form Q =Γφ˙2 , (8.44) P h(φ) m , (8.40) ≡ ρ(0) where Γ is a positive constant. The equation of motion tot (8.43) for the scalar is modified to the transfer rate is computed as dV φ˙ φ¨ +3Hφ˙ +Γφ˙ + =0 . (8.45) (0) ˙ 4 d h dφ Qm =3ρ φ F tot dφ F 4 Since13 φ˙ =0 this equation becomes a Klein-Gordon one 6 6wtot(1 3wtot) dwtot 2 2 with a potential augmented by a friction term of strength − g +3 F g + K . Γ proportional to the inflaton speed φ˙. Taking this phe- − 3wtot +1 dφ nomenological description as paradigmatic for a scalar (8.41) field-fluid interaction leads to The relation between wtot(t) and g(φ) in [282] is not phys- dV φ¨ +3Hφ˙ +Γφ˙ + =0 (8.46) ically transparent and is dictated more by the mathemat- dφ ics than the physics. Other models of (possibly tilted) interacting radiation and dust fluids were studied over for the scalar and to e.g. the years, see [128, 129, 301, 302]. ρ˙1 +3H (P1 + ρ1)=Γφ˙ (8.47) for the fluid 1 which has now a source Γφ˙ on the right C. Scalar field interacting with a perfect fluid hand side. This description can be related to the covariant description of two interacting fluids [175] described A scalar field can couple to a fluid or to another field. in Sec. VIII A. Then, Consider a perfect fluid with energy density ρ1 and pres- Γ α(t)= dt a3φ˙2 (8.48) sure P1 coupling explicitly to a scalar field φ. Their in- −2a3 teraction is again described by the equations Z depends only on the kinetic energy density φ˙2/2 of φ. ρ˙1 +3H (P1 + ρ1)= Q, (8.42) The decay of the field φ into fluid stops when φ becomes static. ρ˙ +3H (P + ρ )= Q ; (8.43) φ φ φ − 13 adding them leads to the conservation equation for the It is φ˙ =6 0 because, if φ = φ0 = const., the scalar field effective total fluid with energy density ρtot = ρ1 +ρφ and pressure fluid reduces to a pure cosmological constant Λ = V (φ0) and decouples from the perfect fluid. Ptot = P1 + Pφ. 31

D. Einstein frame formulation of scalar-tensor form [153] gravity ˜ 4 R 1 ab S = d x g˜ g˜ ˜ aφ˜˜ bφ˜ U φ˜ Although we confine ourselves to Einstein gravity in − 16πG − 2 ∇ ∇ − Z ( " # the rest of this review, here we make an exception to p 8 πG φ˜ mention one Lagrangian and covariant fomulation of in- +e− 2ω+3 (m) [˜g] , (8.52) q L teracting dark energy. In the context of alternative theories of gravity, when scalar-tensor [54, 69, 310, 396] or ˜ f(R) [85, 95, 148, 309, 370] gravity is reformulated in the where a is the covariant derivative operator of the rescaled∇ metric g˜ and Einstein conformal frame, the scalar field degree of free- ab dom looks like an ordinary scalar field in general relativ- πG ity, except for the fact that it couples explicitly to matter. U φ˜ = V φ φ˜ exp 8 φ˜ . (8.53) If the latter is the dark matter fluid in a FLRW cosmol- − r2ω +3 ! ogy, one then obtains a direct interaction of this dark h i matter with the extra scalar degree of freedom (which In the Einstein conformal frame the scalar field φ˜ couples could be used to model dark energy) contained in the directly to matter through the exponential factor multi- theory in addition to the two massless spin two modes of plying the Lagrangian density (m) in the action (8.52). Einstein theory. As a consequence, in this frameL the matter energy- Let us illustrate how this works using Jordan-Brans- momentum tensor is not covariantly conserved. Mathe- Dicke gravity [69, 222, 223] and cosmology [171, 186]. b (m) matically, this happens because the equation Tab =0 Brans-Dicke’s is the simplest scalar-tensor framework is not conformally invariant [398]. The conformally∇ trans- and the prototypical alternative to general relativity. The (m) formed stress-energy tensor T˜ obeys the corrected Jordan frame Brans-Dicke action is [69] ab equation

(BD) 1 4 ω ab (m) d S = d x √ g φR g aφ bφ V (φ) ˜ b ˜ ˜(m) ˜ 16π − − φ ∇ ∇ − Tab = [ln Ω(φ)] T aφ . (8.54) Z ∇ − dφ ∇ +S(m) , (8.49) Only conformally invariant matter with vanishing trace T (m) =0 remains covariantly conserved after the confor- where S(m) = d4x √ g (m) describes the matter sec- mal transformation. tor and the Brans-Dicke− L scalar field φ > 0 is, roughly Consider the Einstein frame stress-energy tensor de- speaking, the inverseR of the effective gravitational cou- rived from the usual expression [398] pling strength Geff, which becomes a field in this theory. The dimensionless parameter ω is the “Brans-Dicke cou- 2 δ √ g˜ (m) ˜(m) − L pling” [69] and the Jordan frame is the pair of dynamical Tab = − ab ; (8.55) √ g˜ δg˜ variables (gab, φ). − There is a second representation of the theory, the so- it is easy to see that called Einstein conformal frame, i.e., the pair of variables (m) 2 (m) T˜ =Ω− T , (8.56) gãb, φ˜ where the new metric gãb is obtained from gab ab ab by means of the conformal transformation ^b(m) 4 b(m) Ta =Ω− Ta , (8.57) g g˜ =Ω2 g , Ω= φ , (8.50) ab → ab ab p ãb 6 ab(m) while the new scalar field comes from the non-linear re- T =Ω− T , (8.58) definition14 and

2ω +3 φ (m) 4 (m) φ˜(φ)= ln (8.51) T˜ =Ω− T . (8.59) 16πG φ r 0 In particular, a perfect fluid stress-energy tensor maps to (where ω > 3/2). Under this transformation, the − Brans-Dicke action (8.49) assumes its Einstein frame ˜(m) ˜(m) (m) ˜(m) Tab = P +ρ ˜ uãu˜b + P gãb (8.60) a b under the conformal transformation, and gãb u˜ u˜ = 1 yields − 14 This non-linear redefinition turns the kinetic energy density of φ˜ into its canonical form. a 1 a u˜ =Ω− u , uã =Ω ua . (8.61) 32

As a consequence, which shows explicitly how the scalar field interacts with ordinary matter through its gradient. To conclude, this ˜(m) (m) ˜(m) P +ρ ˜ uãu˜b + P gãb picture gives a completely covariant and Lagrangian formulation of a scalar field interacting with a matter fluid. 2 (m) (m) (m) In the Jordan frame this scalar has gravitational na- =Ω− P + ρ uaub + P gab , (8.62) ture but this information does not matter in the Ein- as follows fromh Eqs. (8.58), (8.60), and (8.61). Therefore,i stein frame, in which φ˜ could be regarded as standard it must be scalar field, except for its anomalous coupling to (non- conformal) matter. (m) 4 (m) ˜(m) 4 (m) ρ˜ =Ω− ρ , P =Ω− P . (8.63) In principle, when performing the conformal transfor- Due to Eq. (8.63), a Jordan frame fluid satisfying the mation one should also rescale all fundamental and de- barotropic equation of state (2.4) is mapped into an Ein- rived units, according to Dicke’s prescription [153], with stein frame fluid with the same equation of state. the result that the two conformal frames would then be Specializing to FLRW universes, the conformal trans- physically equivalent [153]. This physical equivalence has formation maps the Jordan frame fluid conservation been the subject of a debate filling a fairly large litera- equation ture. This equivalence issue is immaterial here since there is no doubt that the transformations (8.50) and (8.51) dρ(m) constitute a well-defined map between the two worlds +3H P (m) + ρ(m) =0 (8.64) dt and here we are interested in mathematical solutions for a into scalar field coupled to a fluid (or for two coupled effective fluids). Therefore, Jordan frame cosmological solutions dρ˜(m) d (ln Ω) 1 +3H˜ P˜(m) +˜ρ(m) = φ˙ 3P˜(m) ρ˜(m) . in the Jordan frame (where Geff φ− varies) trans- dt dφ − late into mathematical solutions in≃ the Einstein frame, (8.65) where formally the theory can be interpreted as general relativity with the anomalous coupling of an (otherwise Stepping back to general spacetimes, the stress-energy canonical) scalar field φ˜ to the cosmological fluid(s). tensor T (m) scales under the conformal transformation ab One could think of using a tensor-multi-scalar theory according to [398] to obtain multi-fluids in the Einstein conformal frame. Tãb =Ωs T ab , T˜(m) =Ωs+4 T (m) , (8.66) However, in general, one cannot recast the transformed (m) (m) ab ab fields so that more than one has canonical kinetic energy where s is an appropriate conformal weight. In four [225]. spacetime dimensions, the Jordan frame covariant conservation equation b T (m) =0 corresponds to [171, 398] ∇ ab E. Two interacting scalar fields s ab s ab s 1 ab ˜ Ω T =Ω T + (s +6)Ω − T Ω ∇a (m) ∇a (m) (m)∇a There are cosmological scenarios in which both inter- s 1 ab (m) Ω − g T aΩ . (8.67) acting fluids are actually effective fluids coming from − ∇ scalar fields φ and ψ. In the previous picture, if the first The choice of conformal weight s = 6 is consistent with fluid is the ψ-fluid with self-interaction potential U(ψ), Eq. (8.56) and gives − then ˜(m) ab ˜(m) 4 (m) T g˜ Tab =Ω− T , (8.68) ψ˙ 2 ≡ ρ = + U(ψ) , (8.72) while Eq. (8.67) has the conformal cousin 1 2

˜ ãb ˜(m) ab ˜ 2 aT(m) = T g˜ a (ln Ω) (8.69) ψ˙ ∇ − ∇ P1 = U(ψ) , (8.73) in the tilded world. For Brans-Dicke theory, where Ω = 2 − √φ, one has [69, 171, 186, 396] and the equation of motion for ψ becomes 1 4πG ˜ Tãb = T˜(m) ˜ bφ = T˜(m) ˜ bφ˜ . φ˙2 dU a (m) ψ¨ +3Hψ˙ Γ + =0 (8.74) ∇ −2φ ∇ −r2ω +3 ∇ ˙ (8.70) − ψ dψ Consider now a dust fluid describing dark and baryonic (m) (assuming ψ˙ =0). When ψ˙ is large and increasing and matter in a FLRW universe. Setting P =0 yields 6 | | Γ > 0, the extra term Γ φ˙2/ψ˙ enhances the motion of u˜ u˜ ˜ bρ˜(m) +˜ρ(m) u˜ ˜ bu˜ +ρ ˜(m) u˜ ˜ c u˜ − a b ∇ a ∇ b c ∇ a ψ and plays the role of antifriction for ψ. If ψ decreases, this term describes an effective friction opposing the motion of ψ. The quantity α given by (8.48) can still be 4πG (m) ˜ ˜ ρ˜ aφ =0 , (8.71) introduced. −r2ω +3 ∇ 33

More generally, one can couple the two scalars by intro- the seeds for structure formation in the form of quan- ducing a common potential V (φ, ψ) that modifies their tum fluctuations of the inflaton. In most inflationary equations of motion to scenarios, the inflaton driving the rapid expansion of the universe is a scalar field. This is true also for R2 inflation ∂V φ¨ +3Hφ˙ + =0 , (8.75) [374, 375] which, although correcting Einstein theory, can ∂φ be seen as the result of the dynamics of the effective ex- ∂V tra scalar degree of freedom. It is now well known that a ψ¨ +3Hψ˙ + =0 . (8.76) scalar field φ with timelike gradient φ is equivalent to a ∂ψ a perfect fluid which, in general, has a∇ dynamical equation This is done in preheating scenarios designed to end in- of state. It was realized rather early that a free scalar is flation and raise the temperature of the universe [11, 19]. equivalent to a stiff fluid, which appears in several inte- One of the two fields is the inflaton and the other denotes grability cases of the Einstein-Friedmann equations, but any field to which the inflaton couples (bosonic fields are interesting physical scenarios for the scalar field were still generated in the early stages and fermionic fields near the missing. end of preheating). The interaction potential commonly Having learned the lesson from inflation, the scalar used in preheating has the form field was naturally the main ingredient for modelling dark energy after the 1998 discovery of the cosmic ac- m2 m2 g V (φ, ψ)= φ φ2 + ψ ψ2 + φ2ψ2 . (8.77) celeration with type Ia supernovae [325, 326, 336]. In 2 2 2 all these situations dark and baryonic matter form one Quintom models of dark energy (see Ref. [75] for a fluid and dark energy, or its scalar field version, another review) are two-component dark energy models in which fluid. It is possible, and it would be even convenient a phantom scalar field interacts with a second scalar field. for modellers if these dark fluids were interacting di- Sometimes, for simplicity, the two scalars are modelled rectly, which has been the subject of a considerable liter- with two fluids [419]. ature. In the meantime, scalar-tensor gravity originally proposed by Brans & Dicke [69] and later generalized [54, 310, 396], has been the subject of much interest and IX. CONCLUDING REMARKS many other theories of gravity have been studied (see the reviews [89, 123, 207, 212, 407, 408]). After 2004, The scientist’s view of multiple fluid FLRW cosmol- f(R) gravity (a subclass of scalar-tensor gravity) was ogy has changed significantly since the discovery of the studied intensely to explain the present acceleration of first mathematical solutions describing this physics in the the universe [148, 309, 370]. Nowadays, testing grav- 1960s and 1970s. Early interest was in radiation and ity constitutes added value for astrophysical and cosmo- dust filling a (possibly spatially curved) FLRW universe logical observations and satellite missions at vastly dif- and was motivated by the recent discovery of the cos- ferent scales, as well as theoretical research programs. mic microwave background radiation by Penzias and Wil- In the last decade or so, much interest has been de- son [324], and expanded to consider any two fluids that voted to “second generation” scalar-tensor theories of could give exact solutions in terms of elementary func- gravity (see [232, 241] for reviews) known as Horndeski tions. The search for FLRW solutions quickly expanded [146, 147, 213, 231, 303] and beyond-Horndeski and De- to include effective fluids describing spatial curvature and generate Higher Order Scalar-Tensor (DHOST) theories the cosmological constant. [49, 50, 135, 191, 192, 239, 240, 295]. They explore the The modern view is quite different: the standard vast lanscape of the most general scalar-field based theo- ΛCDM model of the universe includes dark energy and ries of gravity that have second order equations and are dark matter fluids in a spatially flat FLRW universe. At- compatible with theoretical and observational constraints tempts to explain the current acceleration of the universe [63]. The motivation for studying these theories is pri- without the ad hoc dark energy by modifing gravity at marily cosmological and, eventually, multi-fluid FLRW large scales instead [148, 309, 370] end up being not so universes enter the description of these cosmologies. different from dark energy scenarios because in a FLRW While trying to connect the old with the new litera- universe the deviations of gravity from general relativity ture, we have reported analytical multi-(effective) fluid in the field equations can be grouped in the right hand solutions expressed by an integral in finite form contain- side, producing an effective perfect fluid in a wide spec- ing only elementary functions. Either the scale factor is trum of theories of gravity [203]. a function a(t) of the comoving time t, or the latter is Another ingredient that was hardly present in cosmol- a function t(a) (this relation can sometimes be inverted ogy in the 1960s is the scalar field: since 1980 we have to obtain a(t) explicitly). Alternatively, the solution is witnessed an enormous interest in inflationary scenarios expressed in parametric form with the conformal time η of the early universe. Inflation solves the three prob- as the parameter. lems of standard Big Bang cosmology, i.e., the horizon, A key element of the discussion is the treatment of the flatness, and monopole problems and provides a mech- spatial curvature of FLRW universes and of the cosmo- anism to generate density fluctuations which constitute logical constant Λ as effective fluids, reducing formally 34 the Einstein-Friedmann equations to those for an effec- ground and through large-scale structure surveys. Per- tive K =0, Λ=0, multi-fluid situation. While this pro- turbations of multi-fluid FLRW universes constitute an cedure is followed routinely for the cosmological constant important part of cosmology. and for the effective energy density of spatial curvature, Even restricting the scope of this work, however, left it is not normally implemented for the effective pressure many interesting multi-(effective) fluids to contemplate. of the latter. This richness is due to the variety of real or effective matter sources, including real fluids with barotropic equa- Following the key realization of Refs. [103, 220, 288, tions of state, scalar fields with a large variety of physi- 289] that the integral expressing the scale factor has an cally interesting potentials, the gravitational scalar field hypergeometric series representation which truncates for of scalar-tensor gravity in the Einstein frame, the cosmo- special values of its arguments (or, alternatively, using logical constant Λ, and spatial curvature. The knowledge the Chebysev theorem of integration), the two-fluid situ- of analytical solutions must be supplemented by phase ations in which t(a) or a(t) can be obtained analytically space analysis to determine which ones are attractors in in terms of elementary functions can be determined. One phase space and, therefore, physically important. This can then compile a catalogue of two-fluid solutions ex- type of study is fundamental in theories of inflation and pressed in terms of elementary functions and select those dark energy. Nevertheless, the proliferation of inflation- of physical interest. ary and dark energy models, and the associated phase To keep this review manageable, we restricted its space studies has sometimes obscured the full covariance scope by excluding from it certain extensions, and re- of a theory or its Lagrangian formulation, as we have lated topics that are of interest in modern cosmol- pointed out in the context of interacting dark energy. ogy. First, we did not discuss tilted fluids [126, 128– Finally, it is rather unfortunate that in recent years 131, 301, 302, 367, 393, 397] and imperfect fluids, which old solutions, integrability conditions, and methods have are interesting possibilities to consider. FLRW cosmolo- been forgotten and needed to be rediscovered, which can gies do not necessarily represent perfect fluid solutions be partially explained by the very different points of view but can also describe a viscous fluid in a non-comoving on this subject in the last half century. We set out with frame. Imperfect fluids are characterized by spacelike en- the goal of providing a starting point for the less ex- ergy fluxes and anisotropic stresses [160, 377, 398], which perienced reader and a reference for the expert one, al- are excluded by spatial isotropy in a FLRW universe, though a review cannot replace the thorough study of but bulk viscosity is allowed and has been studied ex- the relevant literature. We hope that highlighting multi- tensively in the literature. Bulk viscosity could be due, fluid cosmology will bring i) a better understanding of for example, to particle production in the early universe. the relations between the various possible components of Likewise, we did not discuss multi-fluid or scalar field the universe; ii) more efficient research in the context of anisotropic Bianchi universes and cosmic no-hair theo- Einstein theory; iii) will facilitate the comparison with rems [193]. Finally, multi-fluid FLRW solutions in al- analytical solutions describing similar universes in alter- ternative theories of gravity have been omitted, except native theories of gravity. for the Einstein frame description of scalar-tensor gravity, which is formally equivalent to a cosmological scalar field always present in the theory that couples explicitly ACKNOWLEDGMENTS to the matter sector and could explain interacting dark energy in a fully covariant and Lagrangian way. The lit- This work is supported by the Natural Sciences & En- erature on modified gravity and cosmology is, however, gineering Research Council of Canada (grant no. 2016- huge and a comprehensive review of multi-fluid or multi- 03803 to V.F.) and by a Bishop’s University Graduate effective fluid cosmology in all of these theories would Entrance Scholarship to S.J. be a book-size undertaking which goes well beyond the purposes of this work. We did not mention Chaplygin gas cosmology, in which Appendix A: Derivation of Eq. (2.31) a single gas interpolates between different fluids during the evolution of the universe [226]. This is an interesting alternative and the relations between its FLRW solutions Following Ref. [288], we have and those of multi-fluids are not completely explored. At n (0) 3(wi+1) the same time, we restricted to exact FLRW universes, i=1 wiρi a− wtot(a)= n (A.1) ignoring their perturbations. However, the latter are the ρ a 3(wj +1) P j=1 j − main source of information in modern cosmology through the imprints that they left in the cosmic microwave back- and P 35

3 (0) 3wi 4 (0) 3(wj +1) (0) 3(wi+1) (0) 3wj 4 w˙ (a,H)= w ρ (w + 1)a− − a˙ ρ a− w ρ a− ρ (w + 1)a− − a˙ tot −ρ2  i i i  i  − i i j j  i j i j X X X X      3 (0) (0) 3(wi+wj ) 6 a˙ = w (w w ) ρ ρ a− − . (A.2) −ρ2  i i − j i j  a i,j X   

(0) 3(wi+1) Remembering that ρi = ρi a− , we have

wi (wi wj ) ρiρj = (wi wj + wj ) (wi wj ) ρiρj 4π (0) − − − 1 a + 3 ρm i,j i,j sinh− x = ln  X X (0) 4π (0) 4π (0) 2ρm 2 ρm ρm + = (wi wj ) ρiρj + wj (wi wj ) ρiρj , (A.3)  3 3 (0) − −  − ρr i,j i,j r X X  2 4π (0) therefore, a + 3 ρm +v (0) +1  u 4π (0) 4π (0) 2ρm wi (wi wj ) ρiρj wj (wi wj ) ρiρj u 3 ρm 3 ρm + (0) − − − u − ρr  i,j i,j u  X X t  2 = (wi wj ) ρiρj , (A.4) 2 8π (0) 8π (0) 4π (0) − a + 3 ρm a + 3 ρr + a + 3 ρm i,j = ln   X q (0) 4π ρ(0) 4π ρ(0) + 2ρm or  3 m 3 m (0)   − ρr   r  1 2 wi (wi wj ) ρiρj = (wi wj ) ρiρj . (A.5) 8π 8π − 2 − 2 (0) (0) i,j i,j = ln C a + ρm a + ρr + a X X ( "r 3 3 Equation (A.2) then becomes 4π + ρ(0) (B.3) 3 m 3H 2 (0) (0) 3(wi+wj ) 6 w˙ = (w w ) ρ ρ a− − tot −2ρ2 i − j i j tot i,j where X 3H 2 1 = 2 (wi wj ) ρiρj , (A.6) C . (B.4) −ρ − (0) tot i

n n Therefore, we have 1 = . (A.7) 2 8π (0) 8π (0) i,j=1 i

Appendix B: Equivalence of Eqs. (6.13) and (6.12) 4π (0) 4π (0) 1 a + 3 ρm ρm sinh−   3 (0) (0) (0) − 4πρm 4πρm 2ρm Set 3 3 + (0)  − ρr  r  4π (0)   a + 3 ρm 8π 8π x (B.1) = a2 + ρ(0)a + ρ(0) ≡ (0) m r 4π (0) 4π (0) 2ρm 3 3 3 ρm 3 ρm + (0) r − ρr r 4π (0) 2 8π (0) 8π (0) and use the inverse hyperbolic function identity ρm ln C a + ρm a + ρr + a − 3 ( "r 3 3 4π 1 2 + ρ(0) (B.5) sinh− x = ln x + x +1 (B.2) 3 m p to write or, Eq. (6.13) coincides with Eq. (6.12). 36

Appendix C: Single-ﬂuid limit of the universe with dust plus radiation and K = ±1 8π (0) 1 3a 4π (0) 1 3a − ρm sin (0) = ρm cos− 1 (0) Here we discuss the single-ﬂuid limits of the solu- 3 s8πρr ! 3 − 4πρr ! tions (6.10) and (6.12) of the Einstein-Friedmann equations (2.16)-(2.18) describing spatially curved universes 4π (0) π 1 3a = ρm sin− 1 (0) sourced by dust plus radiation. As expected, these lim- 3 " 2 − − 4πρr !# its generate the well known solutions for radiation with curvature, however the limits to dust plus curvature are 2 2π (0) 4π (0) 1 3a − not trivial in both cases K = 1. = ρm + ρm sin (0) 1 , (C.7) ± 3 3 4πρr − ! which yields 1. Positive curvature

8π (0) 2 4π (0) 1 3a t(a)= ρ a a + ρ sin− 1 In the positively curved case K = +1, the relation m m (0) − 3 − 3 4πρr − ! between comoving time and scale factor is given by r Eq. (6.10), which we reproduce here for convenience: +t0′ (C.8)

2π2 (0) 8π (0) 8π (0) 2 (where t0′ = t0 + 3 ρm ): this coincides with the well t(a)= t ρr + ρm a a 0 − 3 3 − known equation (5.5). r (0) Limit to radiation and K = +1: By setting ρm = 0, 4π (0) Eq. (C.1) gives 4π (0) 1 3 ρm + a + ρm sin−  −  . 3 2 8π 4π (0) 8π (0) t = t ρ(0) a2 , (C.9)  3 ρm + 3 ρr  0 r   − r 3 −  r  (C.1) which is inverted to

(0) Limit to dust and K = +1: Setting ρr =0, one obtains 8π (0) 2 a(t)= ρr (t t0) ; (C.10) the limit r 3 − − this equation coincides coincides with Eq. (5.13). 8π (0) 2 4π (0) 1 3a t(a)= ρm a a + ρ sin− 1 − 3 − 3 m (0) − r 4πρm ! 2. Negative curvature +t0 , (C.2) In the negatively curved universe K = 1, the relation that looks diﬀerent from Eq. (5.5), which is instead between comoving time and scale factor− is (Eq. (6.12))

8π (0) 8π (0) 2 8π (0) 2 8π (0) 1 3a t(a)= t0 + ρr + ρm a + a t(a)= ρm a a + ρ sin− 3 3 3 3 m (0) r −r − s8πρm !

4π (0) 8π (0) 8π (0) 2 +t0 , (C.3) ρm ln C ρr + ρm a + a + a − 3 ( "r 3 3 To proceed, use the trigonometric identities 4π + ρ(0) . (C.11) 3 m (0) 1 1 1 2 Limit to dust and K = 1: Setting ρr = 0 in sin− x = cos− 1 2x , (C.4) − 2 − Eq. (C.11), one obtains 1 π 1 cos− x = sin− x , (C.5) 8π (0) 2 2 − t(a)= t0 + ρm a + a r 3 1 1 sin− ( x)= sin− x , (C.6) − − 4π (0) 8π (0) 2 4π (0) ρm ln C ρm a + a + a + ρm , − 3 ( "r 3 3 #) 3a for x (0) to obtain ≡ 8πρr (C.12) q 37 which does not coincide with Eq. (5.6); the latter gives However, using the inverse hyperbolic function identities instead

1 1 1 2 sinh− x = cosh− 2x +1 , (C.14) 8π (0) 2 8π (0) 1 3a 1 2 t = t + ρm a + a ρ sinh− . cosh− x = ln x + x 1 , (C.15) 0 3 − 3 m (0) r s8πρm ! − p (C.13) one has

8π (0) 1 3a 4π (0) 1 3a − − ρm sinh (0) = ρm cosh (0) +1 3 s8πρm ! 3 4πρm !

2 4π 3a 3a = ρ(0) ln +1+ +1 1 m  (0) v (0)  3 4πρm u 4πρm ! − u  t    4π 3 4π 8π = ρ(0) ln a + ρ(0) + a2 + ρ(0)a 3 m (0) 3 m 3 m " 4πρm r !#

4π (0) 4π (0) 2 8π (0) 4π (0) 3 = ρ ln a + ρ + a + ρm a + ρ ln . (C.16) 3 m 3 m 3 3 m (0) r ! " 4πρm !#

The last and constant term is incorporated in the con- obtains immediately the radiation-only limit stant t0 and Eq. (C.16) then coincides with Eq. (5.6). 8π (0) 2 t(a)= t0 + ρr + a (C.17) r 3 that is inverted to give

2 8π (0) a(t)= (t t0) ρr , (C.18) r − − 3 (0) Limit to radiation and K = 1: Setting ρm = 0, one which coincides with Eq. (5.15). −

[1] L.F. Abbott and M.B. Wise, Nucl. Phys. B 244, 541-548 [arXiv:astro-ph/0403157 [astro-ph]]. (1984) doi:10.1016/0550-3213(84)90329-8 [7] J.M. Aguirregabiria, L.P. Chimento, A.S. Jakubi [2] J.P. Abreu, P. Crawford, and J.P. Mimoso, and R. Lazkoz, Phys. Rev. D 67, 083518 (2003) Class. Quant. Grav. 11, 1919-1940 (1994). doi:10.1103/PhysRevD.67.083518 [arXiv:gr-qc/0303010 http://dx.doi.org/10.1088/0264-9381/11/8/002, [gr-qc]]. arXiv:gr-qc/9401024. [8] L. Alabidi and D. Lyth, JCAP 08, 006 [3] P.A.R. Ade et al. (Planck collaboration), Astron. Astro- (2006) doi:10.1088/1475-7516/2006/08/006 phys. 571, A1 (2014). [arXiv:astro-ph/0604569 [astro-ph]]. [4] P.A.R. Ade et al. [Planck], Astron. Astrophys. [9] L. Alabidi, JCAP 10, 015 (2006) doi:10.1088/1475- 571, A16 (2014) doi:10.1051/0004-6361/201321591 7516/2006/10/015 [arXiv:astro-ph/0604611 [astro-ph]]. [arXiv:1303.5076 [astro-ph.CO]]. [10] Y. Ali-Haimoud and C.M. Hirata, Phys. Rev. D [5] P.A.R. Ade et al. [Planck Collaboration], As- 83, 043513 (2011) doi:10.1103/PhysRevD.83.043513 tron. Astrophys. 594, A20 (2016) doi:10.1051/0004- [arXiv:1011.3758 [astro-ph.CO]]. 6361/201525898 [arXiv:1502.02114 [astro-ph.CO]]. [11] R. Allahverdi, R. Brandenberger, F.Y. Cyr-Racine [6] J.M. Aguirregabiria, L.P. Chimento and and A. Mazumdar, Ann. Rev. Nucl. Part. Sci. 60, R. Lazkoz, Phys. Rev. D 70, 023509 27-51 (2010) doi:10.1146/annurev.nucl.012809.104511 (2004) doi:10.1103/PhysRevD.70.023509 [arXiv:1001.2600 [hep-th]]. 38

[12] M. Alcubierre, F.S. Guzman, T. Matos, D. Nunez, [35] J.D. Barrow, The Book of Universes (W.W. Norton & L.A. Urena-Lopez and P. Wiederhold, Class. C., New York, 2011). Quant. Grav. 19, 5017 (2002) doi:10.1088/0264- [36] J.D. Barrow, G. Galloway, and F.J.T. Tipler, Mon. Not. 9381/19/19/314 [arXiv:gr-qc/0110102 [gr-qc]]. Roy. Astron. Soc. 223, 835 (1986). [13] R. Aldrovandi, R.R. Cuzinatto and L.G. Medeiros, [37] N. Bartolo, S. Matarrese and A. Riotto, Phys. Rev. Found. Phys. 36, 1736-1752 (2006) doi:10.1007/s10701- D 65, 103505 (2002) doi:10.1103/PhysRevD.65.103505 006-9076-6 [arXiv:gr-qc/0508073 [gr-qc]]. [arXiv:hep-ph/0112261 [hep-ph]]. [14] M. Alimohammadi and H.M. Sadjadi, Phys. Lett. B [38] N. Bartolo, S. Matarrese and A. Riotto, Phys. Rev. 648, 113-118 (2007) doi:10.1016/j.physletb.2007.03.014 D 69, 043503 (2004) doi:10.1103/PhysRevD.69.043503 [arXiv:gr-qc/0608016 [gr-qc]]. [arXiv:hep-ph/0309033 [hep-ph]]. [15] M. Alimohammadi, Gen. Rel. Grav. 40, 107-115 [39] S. Basilakos, M. Tsamparlis and A. Paliathana- (2008) doi:10.1007/s10714-007-0514-3 [arXiv:0706.1360 sis, Phys. Rev. D 83, 103512 (2011) [gr-qc]]. doi:10.1103/PhysRevD.83.103512 [arXiv:1104.2980 [16] R.A. Alpher and R.C. Herman, Phys. Rev. 75, no.7, [astro-ph.CO]]. 1089-1095 (1949) doi:10.1103/physrev.75.1089 [40] S. Basilakos, N.E. Mavromatos and J. Solà Per- [17] F.G. Alvarenga, R. Fracalossi, R.C. Freitas and acaula, Phys. Rev. D 101, no.4, 045001 (2020) S.V.B. Gonçalves, Gen. Rel. Grav. 49, no.11, 136 (2017) doi:10.1103/PhysRevD.101.045001 [arXiv:1907.04890 doi:10.1007/s10714-017-2301-0 [arXiv:1607.03478 [gr- [hep-ph]]. qc]]. [41] S. Basilakos, N.E. Mavromatos and J. Solà Peracaula, [18] L. Amendola and S. Tsujikawa, Dark Energy: Theory Int. J. Mod. Phys. D 28, no.14, 1944002 (2019) and Observations, (Cambridge University Press, Cam- doi:10.1142/S0218271819440024 [arXiv:1905.04685 bridge, UK, 2010). [hep-th]]. [19] M.A. Amin, M.P. Hertzberg, D.I. Kaiser and [42] S. Basilakos, N. E. Mavromatos and J. Solà Per- J. Karouby, Int. J. Mod. Phys. D 24, 1530003 (2014) acaula, and Axion Dark Matter,” Phys. Lett. B doi:10.1142/S0218271815300037 [arXiv:1410.3808 [hep- 803, 135342 (2020) doi:10.1016/j.physletb.2020.135342 ph]]. [arXiv:2001.03465 [gr-qc]]. [20] K.N. Ananda and M. Bruni, Phys. Rev. D 74, [43] M. Bastero-Gil, A. Berera, R.O. Ramos and 023524 (2006) doi:10.1103/PhysRevD.74.023524 J.G. Rosa, JCAP 01, 016 (2013) doi:10.1088/1475- [arXiv:gr-qc/0603131 [gr-qc]]. 7516/2013/01/016 [arXiv:1207.0445 [hep-ph]]. [21] K.N. Ananda and M. Bruni, Phys. Rev. D 74, [44] T. Battefeld and R. Easther, JCAP 03, 023523 (2006) doi:10.1103/PhysRevD.74.023523 020 (2007) doi:10.1088/1475-7516/2007/03/020 [arXiv:astro-ph/0512224 [astro-ph]]. [arXiv:astro-ph/0610296 [astro-ph]]. [22] A.A. Andrianov, F. Cannata and A.Y. Kamen- [45] D. Battefeld and T. Battefeld, JCAP 05, shchik, JCAP 10, 004 (2011) doi:10.1088/1475- 012 (2007) doi:10.1088/1475-7516/2007/05/012 7516/2011/10/004 [arXiv:1105.4515 [gr-qc]]. [arXiv:hep-th/0703012 [hep-th]]. [23] A. Arbey, Phys. Rev. D 74, 043516 [46] R.A. Battye, M. Bucher and D. Spergel, (2006) doi:10.1103/PhysRevD.74.043516 [arXiv:astro-ph/9908047 [astro-ph]]. [arXiv:astro-ph/0601274 [astro-ph]]. [47] S.S. Bayin, F.I. Cooperstock and V. Faraoni, As- [24] G. Ballesteros, B. Bellazzini and L. Mercolli, JCAP trophys. J. 428, 439-446 (1994) doi:10.1086/174256 05, 007 (2014) doi:10.1088/1475-7516/2014/05/007 [arXiv:astro-ph/9402033 [astro-ph]]. [arXiv:1312.2957 [hep-th]]. [48] J. Beltrán Jiménez, D. Rubiera-Garcia, D. Sáez-Gómez, [25] G. Ballesteros, D. Comelli and L. Pilo, Phys. Rev. D 94, and V. Salzano, Phys. Rev. D 94, 123520 (2016). no.2, 025034 (2016) doi:10.1103/PhysRevD.94.025034 [49] J. Ben Achour, D. Langlois and K. Noui, [arXiv:1605.05304 [hep-th]]. Phys. Rev. D 93, no.12, 124005 (2016) [26] N. Barnaby and J.M. Cline, Phys. Rev. D 73, doi:10.1103/PhysRevD.93.124005 [arXiv:1602.08398 106012 (2006) doi:10.1103/PhysRevD.73.106012 [gr-qc]]. [arXiv:astro-ph/0601481 [astro-ph]]. [50] J. Ben Achour, M. Crisostomi, K. Koyama, D. Lan- [27] N. Barnaby and J. M. Cline, Phys. Rev. D glois, K. Noui and G. Tasinato, JHEP 12, 100 (2016) 75, 086004 (2007) doi:10.1103/PhysRevD.75.086004 doi:10.1007/JHEP12(2016)100 [arXiv:1608.08135 [hep- [arXiv:astro-ph/0611750 [astro-ph]]. th]]. [28] J.D. Barrow, Phys. Lett. B 187, 12-16 (1987) [51] J. Ben Achour and E.R. Livine, JHEP 12, 031 (2019) doi:10.1016/0370-2693(87)90063-3 doi:10.1007/JHEP12(2019)031 [arXiv:1909.13390 [gr- [29] J.D. Barrow, Phys. Lett. B 235, 40-43 (1990) qc]]. doi:10.1016/0370-2693(90)90093-L [52] J. Ben Achour and E.R. Livine, Class. Quant. Grav. [30] J.D. Barrow and P. Saich, Class. Quant. Grav. 10, 279- 37, no.21, 215001 (2020) doi:10.1088/1361-6382/abb577 283 (1993) doi:10.1088/0264-9381/10/2/009 [arXiv:2004.05841 [gr-qc]]. [31] J.D. Barrow, Phys. Rev. D 48, 1585-1590 (1993) [53] J. Ben Achour and E.R. Livine, JHEP 03, 067 (2020) doi:10.1103/PhysRevD.48.1585 doi:10.1007/JHEP03(2020)067 [arXiv:2001.11807 [gr- [32] J.D. Barrow and A.R. Liddle, Phys. Rev. D 47, qc]]. no.12, R5219 (1993) doi:10.1103/PhysRevD.47.R5219 [54] P.G. Bergmann, Int. J. Theor. Phys. 1, 25-36 (1968) [arXiv:astro-ph/9303011 [astro-ph]]. doi:10.1007/BF00668828 [33] J.D. Barrow, Phys. Rev. D 49, 3055-3058 (1994) [55] A. Bernal and F. Siddhartha Guzman, Phys. Rev. doi:10.1103/PhysRevD.49.3055 D 74, 103002 (2006) doi:10.1103/PhysRevD.74.103002 [34] J.D. Barrow, Class. Quant. Grav. 21, L79 (2004). [arXiv:astro-ph/0610682 [astro-ph]]. 39

[56] F. Bernardeau and J.-P. Uzan, Phys. Rev. D [76] Y.F. Cai, M. Li and X. Zhang, Phys. Lett. B 66, 103506 (2002) doi:10.1103/PhysRevD.66.103506 718, 248-254 (2012) doi:10.1016/j.physletb.2012.10.065 [arXiv:hep-ph/0207295 [hep-ph]]. [arXiv:1209.3437 [hep-th]]. [57] F. Bernardeau and J.-P. Uzan, Phys. Rev. D [77] Y.F. Cai, H. Li, Y.S. Piao and 67, 121301 (2003) doi:10.1103/PhysRevD.67.121301 X.m. Zhang, Phys. Lett. B 646, 141- [arXiv:astro-ph/0209330 [astro-ph]]. 144 (2007) doi:10.1016/j.physletb.2007.01.027 [58] D. Bertacca, S. Matarrese and M. Pietroni, [arXiv:gr-qc/0609039 [gr-qc]]. Mod. Phys. Lett. A 22, 2893-2907 [78] Y.F. Cai, T. Qiu, R. Brandenberger, Y.S. Piao and (2007) doi:10.1142/S0217732307025893 X. Zhang, JCAP 03, 013 (2008) doi:10.1088/1475- [arXiv:astro-ph/0703259 [astro-ph]]. 7516/2008/03/013 [arXiv:0711.2187 [hep-th]]. [59] D. Bertacca, N. Bartolo and S. Matarrese, JCAP [79] Y.F. Cai, M.Z. Li, J.X. Lu, Y.S. Piao, 05, 005 (2008) doi:10.1088/1475-7516/2008/05/005 T.t. Qiu and X.m. Zhang, Phys. Lett. B 651, [arXiv:0712.0486 [astro-ph]]. 1-7 (2007) doi:10.1016/j.physletb.2007.05.056 [60] D. Bertacca, N. Bartolo, A. Diaferio and S. Matar- [arXiv:hep-th/0701016 [hep-th]]. rese, JCAP 10, 023 (2008) doi:10.1088/1475- [80] Y.F. Cai, T. Qiu, Y.S. Piao, M. Li and X. Zhang, JHEP 7516/2008/10/023 [arXiv:0807.1020 [astro-ph]]. 10, 071 (2007) doi:10.1088/1126-6708/2007/10/071 [61] O. Bertolami, C.G. Boehmer, T. Harko and [arXiv:0704.1090 [gr-qc]]. F.S.N. Lobo, Phys. Rev. D 75, 104016 (2007) [81] Y.F. Cai and X. Zhang, JCAP 06, 003 (2009) doi:10.1103/PhysRevD.75.104016 [arXiv:0704.1733 doi:10.1088/1475-7516/2009/06/003 [arXiv:0808.2551 [gr-qc]]. [astro-ph]]. [62] O. Bertolami, C. Cosme and J.G. Rosa, Phys. Lett. [82] Y.F. Cai and J. Wang, Class. Quant. Grav. 25, B 759, 1-8 (2016) doi:10.1016/j.physletb.2016.05.047 165014 (2008) doi:10.1088/0264-9381/25/16/165014 [arXiv:1603.06242 [hep-ph]]. [arXiv:0806.3890 [hep-th]]. [63] D. Bettoni, J. M. Ezquiaga, K. Hinterbichler and M. Zu- [83] G. Calcagni, Phys. Rev. D 69, 103508 malacárregui, Phys. Rev. D 95, no.8, 084029 (2017) (2004) doi:10.1103/PhysRevD.69.103508 doi:10.1103/PhysRevD.95.084029 [arXiv:1608.01982 [arXiv:hep-ph/0402126 [hep-ph]]. [gr-qc]]. [84] G. Calcagni, Phys. Rev. D 71, 023511 (2005) [64] A.P. Billyard and A.A. Coley, Phys. Rev. D 61, 083503 doi:10.1103/PhysRevD.71.023511 [arXiv:gr-qc/0410027 (2000) [astro-ph/9908224]. [gr-qc]]. [65] J.K. Bloomﬁeld, É.É. Flanagan, M. Park, and S. Wat- [85] S. Capozziello, S. Carloni and A. Troisi, Re- son, JCAP 1308, 010 (2013) [arXiv:1211.7054[astro- cent Res. Dev. Astron. Astrophys. 1, 625 (2003) ph.CO]]. [arXiv:astro-ph/0303041 [astro-ph]]. [66] A. Borowiec, S. Capozziello, M. De Laurentis, [86] S. Capozziello, V.F. Cardone, E. Elizalde, S. No- F.S.N. Lobo, A. Paliathanasis, M. Paolella and jiri and S.D. Odintsov, Phys. Rev. D 73, A. Wojnar, Phys. Rev. D 91, no.2, 023517 (2015) 043512 (2006) doi:10.1103/PhysRevD.73.043512 doi:10.1103/PhysRevD.91.023517 [arXiv:1407.4313 [gr- [arXiv:astro-ph/0508350 [astro-ph]]. qc]]. [87] S. Capozziello, V.F. Cardone, E. Elizalde, S. Nojiri, and [67] M. Szydlowski, A. Stachowski, A. Borowiec, and A. Wo- S.D. Odintsov, Phys. Rev. D 73, 043512 (2006). jnar, Eur. Phys. J. C 76, 567 (2016). [88] S. Capozziello, E. Piedipalumbo, C. Rubano and [68] M. Bouhmadi-López, A. Errahmani, P. Martin-Moruno, P. Scudellaro, Phys. Rev. D 80, 104030 (2009) T. Ouali, and Y. Tavakoli, Int. J. Mod. Phys. D 24, doi:10.1103/PhysRevD.80.104030 [arXiv:0908.2362 1550078 (2015). [astro-ph.CO]]. [69] C. Brans and R.H. Dicke, Phys. Rev. 124, 925-935 [89] S. Capozziello, V. Faraoni, Beyond Einstein Gravity: (1961) doi:10.1103/PhysRev.124.925 A Survey of Gravitational Theories For Cosmology and [70] I. Brevik, V.V. Obukhov and A.V. Timoshkin, Astrophysics (Springer, New York, 2010). Mod. Phys. Lett. A 29, no.15, 1450078 (2014) [90] S. Capozziello and M. Roshan, Phys. Lett. B doi:10.1142/S0217732314500783 [arXiv:1404.1887 [gr- 726, 471-480 (2013) doi:10.1016/j.physletb.2013.08.047 qc]]. [arXiv:1308.3910 [gr-qc]]. [71] M. Bucher and D.N. Spergel, Phys. Rev. D [91] S. Capozziello, M. De Laurentis and R. Myrzakulov, Int. 60, 043505 (1999) doi:10.1103/PhysRevD.60.043505 J. Geom. Meth. Mod. Phys. 12, no.09, 1550095 (2015) [arXiv:astro-ph/9812022 [astro-ph]]. doi:10.1142/S0219887815500954 [arXiv:1412.1471 [gr- [72] A.B. Burd and J.D. Barrow, Nucl. Phys. B 308, 929-945 qc]]. (1988) [erratum: Nucl. Phys. B 324, 276-276 (1989)] [92] S. Capozziello, M. De Laurentis and K.F. Dialek- doi:10.1016/0550-3213(88)90135-6 topoulos, Eur. Phys. J. C 76, no.11, 629 (2016) [73] C.T. Byrnes, K.Y. Choi and L.M.H. Hall, JCAP doi:10.1140/epjc/s10052-016-4491-0 [arXiv:1609.09289 10, 008 (2008) doi:10.1088/1475-7516/2008/10/008 [gr-qc]]. [arXiv:0807.1101 [astro-ph]]. [93] B.J. Carr and J.E. Lidsey, Phys. Rev. D 48, 543-553 [74] C.T. Byrnes, K.Y. Choi and L.M.H. Hall, JCAP (1993) doi:10.1103/PhysRevD.48.543 02, 017 (2009) doi:10.1088/1475-7516/2009/02/017 [94] S.M. Carroll, Spacetime and Geometry: An Introduction [arXiv:0812.0807 [astro-ph]]. to General Relativity (Addison Wesley, San Francisco, [75] Y.F. Cai, E.N. Saridakis, M.R. Setare and 2004). J.Q. Xia, Phys. Rept. 493, 1-60 (2010) [95] S.M. Carroll, V. Duvvuri, M. Trodden and doi:10.1016/j.physrep.2010.04.001 [arXiv:0909.2776 M.S. Turner, Phys. Rev. D 70, 043528 [hep-th]]. (2004) doi:10.1103/PhysRevD.70.043528 40

[arXiv:astro-ph/0306438 [astro-ph]]. Roy. Astron. Soc. 456, no.4, 3494-3508 (2016) [96] B. Carter and D. Langlois, Nucl. Phys. B 454, doi:10.1093/mnras/stv2691 [arXiv:1510.03877 [astro- 402-424 (1995) doi:10.1016/0550-3213(95)00425-R ph.CO]]. [arXiv:hep-th/9611082 [hep-th]]. [118] K. Y. Choi, L. M. H. Hall and C. van de Bruck, JCAP [97] B. Carter and D. Langlois, Nucl. Phys. B 531, 02, 029 (2007) doi:10.1088/1475-7516/2007/02/029 478-504 (1998) doi:10.1016/S0550-3213(98)00430-1 [arXiv:astro-ph/0701247 [astro-ph]]. [arXiv:gr-qc/9806024 [gr-qc]]. [119] T. Christodoulakis, C. Helias, P.G. Kevrekidis, [98] M. Cataldo and L.P. Chimento, Int. J. Mod. Phys. D I.G. Kevrekidis and G.O. Papadopoulos, 17, 1981-1989 (2008) doi:10.1142/S0218271808013790 [arXiv:gr-qc/0302120 [gr-qc]]. [arXiv:gr-qc/0506090 [gr-qc]]. [120] T. Christodoulakis, N. Dimakis and P.A. Terzis, [99] T. Charters and J.P. Mimoso, JCAP 08, 022 (2010) J. Phys. A 47, 095202 (2014) doi:10.1088/1751- doi:10.1088/1475-7516/2010/08/022 [arXiv:0909.2282 8113/47/9/095202 [arXiv:1304.4359 [gr-qc]]. [hep-ph]]. [121] T. Christodoulakis, A. Karagiorgos and [100] P.H. Chavanis, Phys. Rev. D 92, 103004 (2015). A. Zampeli, Symmetry 10, no.3, 70 (2018) [101] P.L. Chebyshev, “Sur l’integration des diﬀérentielles ir- doi:10.3390/sym10030070 rationnelles”, Journal de Mathematiques (series 1) 18, [122] T. Clemson, K. Koyama, G.B. Zhao, R. Maartens 87–111 (1853). and J. Valiviita, Phys. Rev. D 85, 043007 (2012) [102] X.m. Chen, Y.g. Gong and E.N. Saridakis, JCAP doi:10.1103/PhysRevD.85.043007 [arXiv:1109.6234 04, 001 (2009) doi:10.1088/1475-7516/2009/04/001 [astro-ph.CO]]. [arXiv:0812.1117 [gr-qc]]. [123] T. Clifton, P.G. Ferreira, A. Padilla and [103] S. Chen, G. W. Gibbons, Y. Li and Y. Yang, JCAP C. Skordis, Phys. Rept. 513, 1-189 (2012) 12, 035 (2014) doi:10.1088/1475-7516/2014/12/035 doi:10.1016/j.physrep.2012.01.001 [arXiv:1106.2476 [arXiv:1409.3352 [astro-ph.CO]]. [astro-ph.CO]]. [104] A.D. Chernin, Sov. Astron. 9, 871 (1966). [124] R. Coquereaux and A. Grossmann, Ann. Phys. (USA) [105] S.V. Chervon, I.V. Fomin and A. Beesham, Eur. Phys. 143, 296 (1982). J. C 78, no.4, 301 (2018) doi:10.1140/epjc/s10052-018- [125] J.M. Cohen, Nature 216, 249 (1967). 5795-z [arXiv:1704.08712 [gr-qc]]. [126] A.A. Coley, Astrophys. Space Sci. 155, 193-201 (1989). [106] L.P. Chimento and A.S. Jakubi, Int. J. Mod. Phys. [127] A.A. Coley, Dynamical Systems and Cosmology D 5, 71-84 (1996) doi:10.1142/S0218271896000084 (Kluwer, Dordrecht, 2003). [arXiv:gr-qc/9506015 [gr-qc]]. [128] A.A. Coley and B.O.J. Tupper, J. Math. Phys. 27, 406 [107] L.P. Chimento, Phys. Rev. D 65, 063517 (2002) (1986). doi:10.1103/PhysRevD.65.063517 [129] A.A. Coley and B.O.J. Tupper, Can. J. Phys. 64, 204 [108] L.P. Chimento and R. Lazkoz, Phys. Rev. Lett. (1986). 91, 211301 (2003) doi:10.1103/PhysRevLett.91.211301 [130] A.A. Coley and J. Wainwright, Class. Quant. Grav. 9, [arXiv:gr-qc/0307111 [gr-qc]]. 651–665 (1992). [109] L.P. Chimento and R. Lazkoz, Class. Quant. Grav. [131] A.A. Coley and R.J. van den Hoogen, Class. 23, 3195-3204 (2006) doi:10.1088/0264-9381/23/9/027 Quant. Grav. 12, 1977-1994 (1995) doi:10.1088/0264- [arXiv:astro-ph/0505254 [astro-ph]]. 9381/12/8/015 [arXiv:gr-qc/9605061 [gr-qc]]. [110] L.P. Chimento and D. Pavon, Phys. Rev. D [132] L. Conversi, A. Melchiorri, L. Mersini-Houghton 73, 063511 (2006) doi:10.1103/PhysRevD.73.063511 and J. Silk, Astropart. Phys. 21, 443-449 [arXiv:gr-qc/0505096 [gr-qc]]. (2004) doi:10.1016/j.astropartphys.2004.02.006 [111] L.P. Chimento and W. Zimdahl, Int. J. Mod. Phys. D [arXiv:astro-ph/0402529 [astro-ph]]. 17, 2229-2254 (2008) doi:10.1142/S0218271808013820 [133] C. Cosme, J.G. Rosa and O. Bertolami, JHEP [arXiv:gr-qc/0609104 [gr-qc]]. 05, 129 (2018) doi:10.1007/JHEP05(2018)129 [112] L.P. Chimento, F.P. Devecchi, M.I. Forte and [arXiv:1802.09434 [hep-ph]]. G.M. Kremer, Class. Quant. Grav. 25, 085007 (2008) [134] A.A. Costa, X.D. Xu, B. Wang, E.G.M. Ferreira and doi:10.1088/0264-9381/25/8/085007 [arXiv:0707.4455 E. Abdalla, Phys. Rev. D 89, no.10, 103531 (2014) [gr-qc]]. doi:10.1103/PhysRevD.89.103531 [arXiv:1311.7380 [113] L.P. Chimento, M.I. Forte, R. Lazkoz and [astro-ph.CO]]. M.G. Richarte, Phys. Rev. D 79, 043502 (2009) [135] M. Crisostomi, K. Koyama and G. Tasinato, JCAP doi:10.1103/PhysRevD.79.043502 [arXiv:0811.3643 04, 044 (2016) doi:10.1088/1475-7516/2016/04/044 [astro-ph]]. [arXiv:1602.03119 [hep-th]]. [114] L.P. Chimento, R. Lazkoz and M.G. Richarte, [136] M. Crisostomi, R. Klein and D. Roest, JHEP Phys. Rev. D 83, 063505 (2011) 06, 124 (2017) doi:10.1007/JHEP06(2017)124 doi:10.1103/PhysRevD.83.063505 [arXiv:1011.2345 [arXiv:1703.01623 [hep-th]]. [astro-ph.CO]]. [137] N. Cruz, S. Lepe and F. Pena, Phys. Lett. B [115] L.P. Chimento, M.I. Forte and M.G. Richarte, 646, 177-182 (2007) doi:10.1016/j.physletb.2006.12.070 Mod. Phys. Lett. A 28, 1250235 (2013) [arXiv:gr-qc/0609013 [gr-qc]]. doi:10.1142/S0217732312502355 [arXiv:1106.0781 [138] M.P. Dabrowski, T. Stachowiak and [astro-ph.CO]]. M. Szydlowski, Phys. Rev. D 68, 103519 [116] J. Chluba and R.M. Thomas, Mon. Not. Roy. (2003) doi:10.1103/PhysRevD.68.103519 Astron. Soc. 412, 748 (2011) doi:10.1111/j.1365- [arXiv:hep-th/0307128 [hep-th]]. 2966.2010.17940.x [arXiv:1010.3631 [astro-ph.CO]]. [139] M. Dabrowski, and J. Stelmach, Ann. Phys. (USA) 166, [117] J. Chluba and Y. Ali-Haimoud, Mon. Not. 422 (1986). 41

[140] M.P. Dabrowski, C. Kiefer and B. Sandhofer, Phys. Rev. [165] G.F.R. Ellis, R. Maartens, and M.A.H. MacCallum, D 74, 044022 (2006) doi:10.1103/PhysRevD.74.044022 Relativistic Cosmology (Cambridge University Press, [arXiv:hep-th/0605229 [hep-th]]. Cambridge, UK, 2012). [141] W. Davidson and J.V. Narlikar Progr. Phys. 29, 539 [166] K. Enqvist and A. Vaihkonen, JCAP 09, (1966). 006 (2004) doi:10.1088/1475-7516/2004/09/006 [142] J. D’Ambroise, [arXiv:0711.3916 [hep-th]]. [arXiv:hep-ph/0405103 [hep-ph]]. [143] J. D’Ambroise, [arXiv:1005.1410 [gr-qc]]. [167] V.P. Ermakov, Univ. Izv. Kiev 20, 1 (1880). [144] J. D’Ambroise and F.L. Williams, Int. J. Pure Appl. [168] V. Faraoni, Am. J. Phys. 67, 732 (1999) Maths. 34, 117 (2007) doi:10.1119/1.19361 [arXiv:physics/9901006 [physics]]. [145] M.-A. Dariescu, D.-A. Mihu, and C. Dariescu, Roma- [169] V. Faraoni, Am. J. Phys. 69, 372-376 (2001) nian J. Phys. 62, 101 (2017). doi:10.1119/1.1290250 [arXiv:physics/0006030 [146] C. Deﬀayet, G. Esposito-Farese and [physics]]. A. Vikman, Phys. Rev. D 79, 084003 (2009) [170] V. Faraoni, Phys. Rev. D 69, 123520 (2004) doi:10.1103/PhysRevD.79.084003 [arXiv:0901.1314 doi:10.1103/PhysRevD.69.123520 [arXiv:gr-qc/0404078 [hep-th]]. [gr-qc]]. [147] C. Deﬀayet, X. Gao, D.A. Steer and G. Za- [171] V. Faraoni, Cosmology in Scalar-Tensor Gravity hariade, Phys. Rev. D 84, 064039 (2011) (Kluwer Academic, Dordrecht, 2004), doi:10.1007/978- doi:10.1103/PhysRevD.84.064039 [arXiv:1103.3260 1-4020-1989-0 [hep-th]]. [172] V. Faraoni, Phys. Lett. B 703, 228-231 (2011) [148] A. De Felice and S. Tsujikawa, Living Rev. Rel. 13 doi:10.1016/j.physletb.2011.08.018 [arXiv:1108.2102 (2010) 3 doi:10.12942/lrr-2010-3 [arXiv:1002.4928 [gr- [gr-qc]]. qc]]. [173] V. Faraoni, Phys. Rev. D 85, 024040 (2012) [149] R. de Ritis, G. Marmo, G. Platania, C. Rubano, doi:10.1103/PhysRevD.85.024040 [arXiv:1201.1448 [gr- P. Scudellaro and C. Stornaiolo, Phys. Rev. D 42, 1091- qc]]. 1097 (1990) doi:10.1103/PhysRevD.42.1091 [174] V. Faraoni and C.S. Protheroe, Gen. Rel. Grav. [150] R. de Ritis, G. Marmo, G. Platania, C. Rubano, P. 45, 103-123 (2013) doi:10.1007/s10714-012-1462-0 Scudellaro, and C. Stornaiolo, Phys. Lett. 149A 79-83 [arXiv:1209.3726 [gr-qc]]. (1990). [175] V. Faraoni, J.B. Dent and E.N. Saridakis, [151] R. de Ritis. G. Platania, C. Rubano, and R. Sabatino, Phys. Rev. D 90, no.6, 063510 (2014) Phys. Lett. 16lA 230 (1991). doi:10.1103/PhysRevD.90.063510 [arXiv:1405.7288 [152] W. de Sitter, Mon. Not. Roy. Astron. Soc. 78, 3-28 [gr-qc]]. (1917) [176] V. Faraoni and J. Coté, Phys. Rev. D 98, [153] R.H. Dicke, Phys. Rev. 125, 2163-2167 (1962) no.8, 084019 (2018) doi:10.1103/PhysRevD.98.084019 doi:10.1103/PhysRev.125.2163 [arXiv:1808.02427 [gr-qc]]. [154] N. Dimakis, T. Christodoulakis and [177] V. Faraoni, Symmetry 12, no.1, 147 (2020) P.A. Terzis, J. Geom. Phys. 77, 97-112 (2014) doi:10.3390/sym12010147 [arXiv:2001.02126 [gr-qc]]. doi:10.1016/j.geomphys.2013.12.001 [arXiv:1311.4358 [178] J. E. Felten and R. Isaacman, Rev. Mod. Phys. 58, 689- [gr-qc]]. 698 (1986) doi:10.1103/RevModPhys.58.689 [155] N. Dimakis, A. Karagiorgos, A. Zampeli, [179] B. Feng, M. Li, Y.S. Piao and X. Zhang, Phys. Lett. B A. Paliathanasis, T. Christodoulakis and 634, 101-105 (2006) doi:10.1016/j.physletb.2006.01.066 P. A. Terzis, Phys. Rev. D 93, no.12, 123518 (2016) [arXiv:astro-ph/0407432 [astro-ph]]. doi:10.1103/PhysRevD.93.123518 [arXiv:1604.05168 [180] B. Feng, [arXiv:astro-ph/0602156 [astro-ph]]. [gr-qc]]. [181] I.V. Fomin, Russ Phys J 61, 843–851 (2018). [156] R. d’Inverno, Introducing Einstein’s Relativity (Claren- https://doi.org/10.1007/s11182-018-1468-5 don Press, Oxford, 1992). [182] P.H. Frampton, K.J. Ludwick and R.J. Scher- [157] S. Dussault and V. Faraoni, Eur. Phys. J. C 80, rer, Phys. Rev. D 84, 063003 (2011) no.11, 1002 (2020) doi:10.1140/epjc/s10052-020-08590- doi:10.1103/PhysRevD.84.063003 [arXiv:1106.4996 8 [arXiv:2009.03235 [gr-qc]]. [astro-ph.CO]]. [158] S. Dutta, M. Lakshmanan and S. Chakraborty, [183] A. Friedmann, Zeit. Physik A 10, 377-386 (1922) Int. J. Mod. Phys. D 25, no.14, 1650110 (2016) doi:10.1007/BF01332580 doi:10.1142/S0218271816501108 [arXiv:1607.03396 [gr- [184] A. Friedmann, Zeit. Physik A 21, 326-322 (1924, qc]]. doi:10.1007/BF01328280 Translated in Gen. Rel. Grav. [159] R. Easther, Class. Quant. Grav. 10, 2203- 31, 31 (1999). 2216 (1993) doi:10.1088/0264-9381/10/11/005 [185] B. Fuchs and E.W. Mielke, Mon. Not. Roy. Astron. Soc. [arXiv:gr-qc/9308010 [gr-qc]]. 350, 707 (2004) doi:10.1111/j.1365-2966.2004.07679.x [160] C. Eckart, Phys. Rev. 58, 919-924 (1940) [arXiv:astro-ph/0401575 [astro-ph]]. doi:10.1103/PhysRev.58.919 [186] Y. Fujii and K. Maeda, The Scalar–Tensor Theory of [161] D. Edwards, Mon. Not. R. Astron. Soc. 159, 51 (1972). Gravity (Cambridge University Press, Cambridge, UK, [162] A. Einstein, Relativitätstheorie". Sitzungsb. König. 2003). Preuss. Akad. 142-152 (1917). [187] A. Füzfa and J.M. Alimi, Phys. Rev. D 75, [163] G.F.R. Ellis, Class. Quant. Grav. 5, 891-901 (1988) 123007 (2007) doi:10.1103/PhysRevD.75.123007 doi:10.1088/0264-9381/5/6/010 [arXiv:astro-ph/0702478 [astro-ph]]. [164] G.F.R. Ellis and M. Madsen, Class. Quant. Grav. 8, 667 [188] L.A. Gergely and S. Tsujikawa, Phys. Rev. D 89, 064059 (1991). (2014) [arXiv:1402.0553 [hep-th]]. 42

[189] R.C. Gilman, Phys. Rev. D 2, 1400-1410 (1970) doi:10.1016/j.physrep.2018.11.006 [arXiv:1807.01725 doi:10.1103/PhysRevD.2.1400 [gr-qc]]. [190] G. Gionti, S.J. and A. Paliathanasis, Mod. [213] G.W. Horndeski, Int. J. Theor. Phys. 10, 363-384 (1974) Phys. Lett. A 33, no.16, 1850093 (2018) doi:10.1007/BF01807638 doi:10.1142/S0217732318500931 [arXiv:1711.11106 [214] F. Hoyle and J.V. Narlikar, Proc. Roy. Soc. (London) [gr-qc]]. A273, 1 (1963). [191] J. Gleyzes, D. Langlois, F. Piazza and F. Vernizzi, [215] F. Hoyle and J.V. Narlikar, Proc. Roy. Soc. (London) Phys. Rev. Lett. 114, no.21, 211101 (2015) A294, 138 (1966). doi:10.1103/PhysRevLett.114.211101 [arXiv:1404.6495 [216] W. Hu and J. Silk, Phys. Rev. D 48, 485-502 (1993) [hep-th]]. doi:10.1103/PhysRevD.48.485 [192] J. Gleyzes, D. Langlois, F. Piazza and F. Vernizzi, [217] Q.G. Huang, JCAP 05, 005 (2009) doi:10.1088/1475- JCAP 02, 018 (2015) doi:10.1088/1475- 7516/2009/05/005 [arXiv:0903.1542 [hep-th]]. 7516/2015/02/018 [arXiv:1408.1952 [astro-ph.CO]]. [218] L.P. Hughston and L.C. Shepley, Astrophys. J. 160, 333 [193] D.S. Goldwirth and T. Piran, Phys. Rept. 214, 223-291 (1970). (1992) doi:10.1016/0370-1573(92)90073-9 [219] K.C. Jacobs, Nature 215, 1156 (1967). [194] Y. Gong and X. Chen, Phys. Rev. D 76, 123007 (2007) [220] K.C. Jacobs, Astrophys. J. 153, 661 (1968). doi:10.1103/PhysRevD.76.123007 [arXiv:0708.2977 [221] M. Jamil, E.N. Saridakis and M.R. Setare, Phys. Rev. [astro-ph]]. D 81, 023007 (2010) doi:10.1103/PhysRevD.81.023007 [195] V. Gorini, A. Kamenshchik, U. Moschella, and V. [arXiv:0910.0822 [hep-th]]. Pasquier, Phys. Rev. D 69, 123512 (2004). [222] P. Jordan, Naturwiss. 26, 417 (1938). [196] G. Gubitosi, F. Piazza, and F. Vernizzi, JCAP 1302, [223] P. Jordan, Z. Phys. 157, 112-121 (1959) 032 (2013) [arXiv:1210.0201[hep-th]]. doi:10.1007/BF01375155 [197] B. Gumjudpai, Gen. Rel. Grav. 41, 249-265 (2009) [224] A. Joseph and R. Saha, [arXiv:1912.06782 [gr-qc]]. doi:10.1007/s10714-008-0665-x [arXiv:0710.3598 [gr- [225] D. I. Kaiser, Phys. Rev. D 81, 084044 (2010) qc]]. doi:10.1103/PhysRevD.81.084044 [arXiv:1003.1159 [gr- [198] B. Gumjudpai, Astropart. Phys. 30, 186-191 qc]]. (2008) doi:10.1016/j.astropartphys.2008.09.006 [226] A.Y. Kamenshchik, U. Moschella and V. Pasquier, [arXiv:0708.3674 [gr-qc]]. Phys. Lett. B 511, 265 (2001) [gr-qc/0103004]. [199] B. Gumjudpai, JCAP 09, 028 (2008) doi:10.1088/1475- [227] L.I. Kharbediya, Astron. Zh. (USSR) 53, 1145 (1976) 7516/2008/09/028 [arXiv:0805.3796 [gr-qc]]. (in Russian). [200] B. Gumjudpai, [arXiv:0904.2746 [gr-qc]]. [228] M. Khurshudyan, E. Chubaryan and B. Pourhassan, Int. [201] Z.K. Guo, Y.S. Piao, X.M. Zhang and J. Theor. Phys. 53, 2370 (2014) doi:10.1007/s10773-014- Y.Z. Zhang, Phys. Lett. B 608, 177-182 2036-6 [arXiv:1402.2385 [gr-qc]]. (2005) doi:10.1016/j.physletb.2005.01.017 [229] S.A. Kim and A.R. Liddle, Phys. Rev. D 74, [arXiv:astro-ph/0410654 [astro-ph]]. 063522 (2006) doi:10.1103/PhysRevD.74.063522 [202] Z.K. Guo, Y.S. Piao, X. Zhang and [arXiv:astro-ph/0608186 [astro-ph]]. Y.Z. Zhang, Phys. Rev. D 74, 127304 [230] S.A. Kim, A.R. Liddle and D. Seery, (2006) doi:10.1103/PhysRevD.74.127304 Phys. Rev. Lett. 105, 181302 (2010) [arXiv:astro-ph/0608165 [astro-ph]]. doi:10.1103/PhysRevLett.105.181302 [arXiv:1005.4410 [203] M. Gurses and Y. Heydarzade, Eur. Phys. J. C 80, [astro-ph.CO]]. no.11, 1061 (2020) doi:10.1140/epjc/s10052-020-08641- [231] T. Kobayashi, M. Yamaguchi and J. Yokoyama, 0 [arXiv:2009.12825 [gr-qc]]. Prog. Theor. Phys. 126, 511-529 (2011) [204] A.H. Guth, Phys. Rev. D 23, 347-356 (1981) doi:10.1143/PTP.126.511 [arXiv:1105.5723 [hep-th]]. doi:10.1103/PhysRevD.23.347 [232] T. Kobayashi, Rept. Prog. Phys. 82, no.8, 086901 (2019) [205] F.S. Guzman and L.A. Urena-Lopez, Phys. Rev. D doi:10.1088/1361-6633/ab2429 [arXiv:1901.07183 [gr- 68, 024023 (2003) doi:10.1103/PhysRevD.68.024023 qc]]. [arXiv:astro-ph/0303440 [astro-ph]]. [233] G. Koﬁnas, R. Maartens, and E. Papantonopoulos, J. [206] T. Harko, F.S.N. Lobo and M.K. Mak, Eur. Phys. J. High Energy Phys. 2003, 066 (2003). C 74, 2784 (2014) doi:10.1140/epjc/s10052-014-2784-8 [234] L.A. Kofman, A.D. Linde and A.A. Starobinsky, [arXiv:1310.7167 [gr-qc]]. Phys. Lett. B 157, 361-367 (1985) doi:10.1016/0370- [207] T. Harko and F.S.N. Lobo, Extensions of f(R) Gravity 2693(85)90381-8 (Cambridge University Press, Cambridge, 2018) [235] E.W. Kolb and M.S. Turner, The Early Universe [208] E.R. Harrison, Mon. Not. Roy. Astron. Soc. 137, 69 (Addison-Wesley, Redwood City, CA, 1990). (1967). [236] C. Kritpetch, J. Sanongkhun, P. Vanichchapong- [209] R.M. Hawkins and J.E. Lidsey, Phys. Rev. D jaroen and B. Gumjudpai, Mod. Phys. Lett. A 35, 66, 023523 (2002) doi:10.1103/PhysRevD.66.023523 no.19, 2050157 (2020) doi:10.1142/S0217732320501576 [arXiv:astro-ph/0112139 [astro-ph]]. [arXiv:1908.11265 [gr-qc]]. [210] J.H. He, B. Wang and Y.P. Jing, JCAP 07, 030 (2009) [237] A.T. Kruger and J.W. Norbury, Phys. Rev. D doi:10.1088/1475-7516/2009/07/030 [arXiv:0902.0660 61, 087303 (2000) doi:10.1103/PhysRevD.61.087303 [gr-qc]]. [arXiv:gr-qc/0004039 [gr-qc]]. [211] J.H. He, B. Wang and E. Abdalla, Phys. Rev. D [238] L.D. Landau and E.M. Lifschitz, The Classical Theory 83, 063515 (2011) doi:10.1103/PhysRevD.83.063515 of Fields (Pergamon Press, Oxford, 1989), pp. 363–367. [arXiv:1012.3904 [astro-ph.CO]]. [239] D. Langlois and K. Noui, JCAP 02, 034 (2016) [212] L. Heisenberg, Phys. Rept. 796, 1-113 (2019) doi:10.1088/1475-7516/2016/02/034 [arXiv:1510.06930 43

[gr-qc]]. doi:10.1119/1.1405506 [arXiv:astro-ph/0109215 [astro- [240] D. Langlois and K. Noui, JCAP 07, 016 (2016) ph]]. doi:10.1088/1475-7516/2016/07/016 [arXiv:1512.06820 [265] A.D. Linde and V.F. Mukhanov, Phys. Rev. D [gr-qc]]. 56, R535-R539 (1997) doi:10.1103/PhysRevD.56.R535 [241] D. Langlois, Int. J. Mod. Phys. D 28, no.05, [arXiv:astro-ph/9610219 [astro-ph]]. 1942006 (2019) doi:10.1142/S0218271819420069 [266] A.D. Linde and V. Mukhanov, JCAP 04, [arXiv:1811.06271 [gr-qc]]. 009 (2006) doi:10.1088/1475-7516/2006/04/009 [242] R. Lazkoz and G. Leon, Phys. Lett. B 638, [arXiv:astro-ph/0511736 [astro-ph]]. 303-309 (2006) doi:10.1016/j.physletb.2006.05.075 [267] J.J. Liouville, Math. Pures Appl. Paris 18 (1), 71 (1853). [arXiv:astro-ph/0602590 [astro-ph]]. [268] F. Lucchin and S. Matarrese, Phys. Rev. D 32, 1316 [243] R. Lazkoz, G. Leon and I. Quiros, Phys. Lett. B (1985) doi:10.1103/PhysRevD.32.1316 649, 103-110 (2007) doi:10.1016/j.physletb.2007.03.060 [269] D.H. Lyth, C. Ungarelli and D. Wands, Phys. Rev. [arXiv:astro-ph/0701353 [astro-ph]]. D 67, 023503 (2003) doi:10.1103/PhysRevD.67.023503 [244] J.W. Lee and S. Lim, JCAP 01, 007 (2010) [arXiv:astro-ph/0208055 [astro-ph]]. doi:10.1088/1475-7516/2010/01/007 [arXiv:0812.1342 [270] R. Maartens, D.R. Taylor and N. Roussos, Phys. Rev. [astro-ph]]. D 52, 3358-3364 (1995) doi:10.1103/PhysRevD.52.3358 [245] J.W. Lee, J. Korean Phys. Soc. 54, 2622 (2009) [271] M.S. Madsen, Astrophys. Space Sci. 113, 205 (1985). doi:10.3938/jkps.54.2622 [arXiv:0801.1442 [astro-ph]]. [272] M.S. Madsen, Class. Quant. Grav. 5, 627-639 (1988) [246] G. Lemaître, Ann. Soc. Sci. Bruxelles A 47, 49 (1927). doi:10.1088/0264-9381/5/4/010 [247] G. Lemaître, Bull. Astron. Inst. Neth. 5, No. 200 (1930). [273] E.A. Marchisotto and G.-A. Zakeri Coll. Math. J. 25, [248] G. Lemaître, Mon. Not. Roy. Astron. Soc. 95, 483 295-308 (1994). (1931). [274] J. Martin, C. Ringeval and V. Vennin, Phys. Dark [249] G. Lemaître, Ann. Soc. Sci. Bruxelles IA 53, 51 (1933). Univ. 5-6, 75-235 (2014) doi:10.1016/j.dark.2014.01.003 [250] G. Leon, Y. Leyva and J. Socorro, Phys. Lett. B [arXiv:1303.3787 [astro-ph.CO]]. 732, 285-297 (2014) doi:10.1016/j.physletb.2014.03.053 [275] T. Matos and F.S. Guzman, Class. Quant. Grav. [arXiv:1208.0061 [gr-qc]]. 17, L9-L16 (2000) doi:10.1088/0264-9381/17/1/102 [251] G. Leon, A. Paliathanasis and J.L. Morales- [arXiv:gr-qc/9810028 [gr-qc]]. Martínez, Eur. Phys. J. C 78, no.9, 753 (2018) [276] T. Matos, F.S. Guzman and L.A. Urena-Lopez, Class. doi:10.1140/epjc/s10052-018-6225-y [arXiv:1808.05634 Quant. Grav. 17, 1707-1712 (2000) doi:10.1088/0264- [gr-qc]]. 9381/17/7/309 [arXiv:astro-ph/9908152 [astro-ph]]. [252] B. Li, T. Rindler-Daller and P.R. Shapiro, [277] T. Matos, F.S. Guzman, L.A. Urena-Lopez Phys. Rev. D 89, no.8, 083536 (2014) and D. Nunez, doi:10.1007/0-306-47115-9_16 doi:10.1103/PhysRevD.89.083536 [arXiv:1310.6061 [arXiv:astro-ph/0102419 [astro-ph]]. [astro-ph.CO]]. [278] T. Matos, A. Vazquez-Gonzalez and J. Magana, [253] Y.H. Li, J.F. Zhang and X. Zhang, Phys. Rev. D 90, Mon. Not. Roy. Astron. Soc. 393, 1359-1369 (2009) no.6, 063005 (2014) doi:10.1103/PhysRevD.90.063005 doi:10.1111/j.1365-2966.2008.13957.x [arXiv:0806.0683 [arXiv:1404.5220 [astro-ph.CO]]. [astro-ph]]. [254] A.R. Liddle, Phys. Lett. B 220, 502-508 (1989) [279] T.L. May, Astrophys. J. 199, 322 (1975). doi:10.1016/0370-2693(89)90776-4 [280] N.E. Mavromatos, J. Solà Peracaula and [255] A.R. Liddle and R.J. Scherrer, Phys. Rev. D S. Basilakos, Universe 6, no.11, 218 (2020) 59, 023509 (1999) doi:10.1103/PhysRevD.59.023509 doi:10.3390/universe6110218 [arXiv:2008.00523 [gr- [arXiv:astro-ph/9809272 [astro-ph]]. qc]]. [256] A. Liddle, An Introduction to Modern Cosmology (Wi- [281] N.E. Mavromatos and J. Solà Peracaula, ley, New York, 2015). doi:10.1140/epjs/s11734-021-00197-8 [arXiv:2012.07971 [257] A.R. Liddle and D.H. Lyth, Cosmological Inﬂation [hep-ph]]. and Large-Scale Structure (Cambridge University Press, [282] T.K. May and G.C. McVittie, Mon. Not. Roy. Astron. Cambridge, UK, 2000). Soc. 148, 407 (1070). [258] J.E. Lidsey, Phys. Lett. B 273, 42-46 (1991) [283] T.L. May and G.C. McVittie, Mon. Not Roy. Astron. doi:10.1016/0370-2693(91)90550-A Soc. 153, 491-500 (1971). [259] J.E. Lidsey, Class. Quant. Grav. 8, 923-933 (1991) [284] C.B.G. Mclntosh, Nature 216, 1297–1298 (1967). doi:10.1088/0264-9381/8/5/016 [285] C.G.B. McIntosh, Mon. Not. Roy. Astron. Soc. 138, 423 [260] J.E. Lidsey, Gen. Rel. Grav. 25, 399-407 (1993) (1968). doi:10.1007/BF00757120 [286] C.G.B. McIntosh, Mon. Not. Roy. Astron. Soc. 140, 461 [261] J.E. Lidsey and I. Waga, Phys. Rev. D 51, (1968). 444-449 (1995) doi:10.1103/PhysRevD.51.444 [287] C.B.G. McIntosh J. Math. Phys. 11, 250-252 (1970). [arXiv:astro-ph/9408016 [astro-ph]]. [288] C.B.G. McIntosh, Austral. J. Phys. 25, 75-82 (1972). [262] J.E. Lidsey, A.R. Liddle, E.W. Kolb, E.J. Copeland, [289] C.B.G. McIntosh and J.M. Foyster, Austral. J. Phys. T. Barreiro and M. Abney, Rev. Mod. Phys. 25, 83-89 (1972). 69, 373-410 (1997) doi:10.1103/RevModPhys.69.373 [290] V. Méndez, Class. Quant. Grav. 13, 3229-3239 (1996) [arXiv:astro-ph/9508078 [astro-ph]]. doi:10.1088/0264-9381/13/12/013 [263] J.E. Lidsey, Class. Quant. Grav. 21, 777-786 (2004) [291] J. Meyers and N. Sivanandam, Phys. Rev. D doi:10.1088/0264-9381/21/4/002 [arXiv:gr-qc/0307037 83, 103517 (2011) doi:10.1103/PhysRevD.83.103517 [gr-qc]]. [arXiv:1011.4934 [astro-ph.CO]]. [264] J.A.S. Lima, Am. J. Phys. 69, 1245-1247 (2001) [292] S. Mishra and S. Chakraborty, Eur. Phys. J. C 78, 44

no.11, 917 (2018) doi:10.1140/epjc/s10052-018-6405-9 doi:10.1103/PhysRevD.90.103524 [arXiv:1410.4930 [gr- [arXiv:1811.08279 [gr-qc]]. qc]]. [293] C.W. Misner, Astrophys. J. 151, 431-457 (1968) [317] A. Paliathanasis and M. Tsamparlis, Phys. Rev. D 90, doi:10.1086/149448 no.4, 043529 (2014) doi:10.1103/PhysRevD.90.043529 [294] H. Mohseni Sadjadi and M. Alimohammadi, Phys. Rev. [arXiv:1408.1798 [gr-qc]]. D 74, 043506 (2006) doi:10.1103/PhysRevD.74.043506 [318] A. Paliathanasis, M. Tsamparlis, S. Basilakos and [arXiv:gr-qc/0605143 [gr-qc]]. J.D. Barrow, Phys. Rev. D 93, no.4, 043528 (2016) [295] H. Motohashi, K. Noui, T. Suyama, M. Yamaguchi and doi:10.1103/PhysRevD.93.043528 [arXiv:1511.00439 D. Langlois, JCAP 07, 033 (2016) doi:10.1088/1475- [gr-qc]]. 7516/2016/07/033 [arXiv:1603.09355 [hep-th]]. [319] A. Paliathanasis and S. Capozziello, Mod. [296] V. Mukhanov, Physical Foundations of Cosmology Phys. Lett. A 31, no.32, 1650183 (2016) (Cambridge University Press, Cambridge, UK, 2005). doi:10.1142/S0217732316501832 [arXiv:1602.08914 [297] V. F. Mukhanov, H. A. Feldman and R. H. Bran- [gr-qc]]. denberger, Phys. Rept. 215, 203-333 (1992) [320] A.D. Payne, Austral. J. Phys. 23, 177-185 (1970). doi:10.1016/0370-1573(92)90044-Z [321] P.J.E. Peebles, Astrophys. J. 153, 1 (1968) [298] A.G. Muslimov, Class. Quant. Grav. 7, 231-237 (1990) doi:10.1086/149628 doi:10.1088/0264-9381/7/2/015 [322] P.J.E. Peebles, Principles of Physical Cosmology [299] N.F. Naidu, S. Carloni and P.K.S. Dunsby, (Princeton University Press, Princeton, 1993). [arXiv:2102.05693 [gr-qc]]. [323] P.J.E. Peebles and B. Ratra, Astrophys. J. Lett. 325, [300] A. Naruko and M. Sasaki, Prog. Theor. Phys. 121, 193- L17 (1988) doi:10.1086/185100 210 (2009) doi:10.1143/PTP.121.193 [arXiv:0807.0180 [324] A.A. Penzias and R.W. Wilson, Astrophys. J. 142, 419- [astro-ph]]. 421 (1965) doi:10.1086/148307 [301] A.V. Nesteruk, Class. Quant. Grav. 11, L15 (1994). [325] S. Perlmutter et al. [Supernova Cosmology Project], [302] A.V. Nesteruk and A.C. Ottewill, Class. Quant. Grav. Nature 391, 51-54 (1998) doi:10.1038/34124 12, 51-57 (1995) doi:10.1088/0264-9381/12/1/005 [arXiv:astro-ph/9712212 [astro-ph]]. [303] A. Nicolis, R. Rattazzi and E. Trincherini, Phys. Rev. [326] S. Perlmutter et al. [Supernova Cosmology Project], D 79, 064036 (2009) doi:10.1103/PhysRevD.79.064036 Astrophys. J. 517, 565-586 (1999) doi:10.1086/307221 [arXiv:0811.2197 [hep-th]]. [arXiv:astro-ph/9812133 [astro-ph]]. [304] S. Nojiri and S.D. Odintsov, Phys. Rev. D 70, [327] C.M. Peterson and M. Tegmark, Phys. Rev. D 103522 (2004) doi:10.1103/PhysRevD.70.103522 84, 023520 (2011) doi:10.1103/PhysRevD.84.023520 [arXiv:hep-th/0408170 [hep-th]]. [arXiv:1011.6675 [astro-ph.CO]]. [305] S. Nojiri, S.D. Odintsov, and S. Tsujikawa, Phys. Rev. [328] T. Phetnora, R. Sooksan and B. Gumjudpai, Gen. Rel. D 71, 063004 (2005). Grav. 42, 225-240 (2010) doi:10.1007/s10714-009-0831- [306] S. Nojiri and S.D. Odintsov, Phys. Rev. D 72, 023003 9 [arXiv:0805.3794 [gr-qc]]. (2005). [329] E. Piedipalumbo, M. De Laurentis and [307] S. Nojiri and S.D. Odintsov, Phys. Lett. B 639, S. Capozziello, Phys. Dark Univ. 27, 100444 (2020) 144-150 (2006) doi:10.1016/j.physletb.2006.06.065 doi:10.1016/j.dark.2019.100444 [arXiv:1912.08089 [arXiv:hep-th/0606025 [hep-th]]. [gr-qc]]. [308] S. Nojiri and S.D. Odintsov, Phys. Lett. B 649, [330] L.O. Pimentel, Class. Quant. Grav. 6, L263 (1989). 440-444 (2007) doi:10.1016/j.physletb.2007.02.042 [331] E. Pinney, Proc. Amer. Math. Soc. 1, 681 (1950). [arXiv:hep-th/0702031 [hep-th]]. [332] A. Pradhan, Indian J. Phys. 88, 215-223 (2014), [309] S. Nojiri and S. D. Odintsov, Phys. Rept. 505 (2011) 59 arXiv:1211.1882 [physics.gen-ph]. doi:10.1016/j.physrep.2011.04.001 [arXiv:1011.0544 [gr- [333] M.L. Pucheu and M. Bellini, Nuovo Cim. B qc]]. 125, 851-859 (2010) doi:10.1393/ncb/i2010-10888-0 [310] K. Nordtvedt, Jr., Astrophys. J. 161, 1059-1067 (1970) [arXiv:0906.1824 [gr-qc]]. doi:10.1086/150607 [334] O. Pujolas, I. Sawicki and A. Vikman, JHEP 11, 156 [311] M. Nowakowski and H. Rosu, Phys. Rev. E (2011) doi:10.1007/JHEP11(2011)156 [arXiv:1103.5360 65, 047602 (2002) doi:10.1103/PhysRevE.65.047602 [hep-th]]. [arXiv:physics/0110066 [physics]]. [335] T. Qiu, Mod. Phys. Lett. A 25, 909-921 (2010) [312] M. Özer and M.O. Taha, Phys. Rev. D 45, R997-R999 doi:10.1142/S021773231000006X [arXiv:1002.3971 [hep- (1992) doi:10.1103/PhysRevD.45.R997 th]]. [313] T. Padmanabhan and T.R. Choudhury, Phys. Rev. [336] A.G. Riess et al. [Supernova Search Team], As- D 66, 081301 (2002) doi:10.1103/PhysRevD.66.081301 tron. J. 116, 1009-1038 (1998) doi:10.1086/300499 [arXiv:hep-th/0205055 [hep-th]]. [arXiv:astro-ph/9805201 [astro-ph]]. [314] T. Pailas, N. Dimakis, A. Paliathanasis, P.A. Terzis [337] G.I. Rigopoulos, E.P.S. Shellard and and T. Christodoulakis, Phys. Rev. D 102, no.6, B.J.W. van Tent, Phys. Rev. D 73, 083522 063524 (2020) doi:10.1103/PhysRevD.102.063524 (2006) doi:10.1103/PhysRevD.73.083522 [arXiv:2005.11726 [gr-qc]]. [arXiv:astro-ph/0506704 [astro-ph]]. [315] A. Paliathanasis, M. Tsamparlis and S. Basi- [338] G.I. Rigopoulos, E.P.S. Shellard and lakos, Phys. Rev. D 84, 123514 (2011) B.J.W. van Tent, Phys. Rev. D 76, 083512 doi:10.1103/PhysRevD.84.123514 [arXiv:1111.4547 (2007) doi:10.1103/PhysRevD.76.083512 [astro-ph.CO]]. [arXiv:astro-ph/0511041 [astro-ph]]. [316] A. Paliathanasis, M. Tsamparlis and S. Basi- [339] T. Rindler-Daller and P.R. Shapiro, Mod. lakos, Phys. Rev. D 90, no.10, 103524 (2014) Phys. Lett. A 29, no.2, 1430002 (2014) 45

doi:10.1142/S021773231430002X [arXiv:1312.1734 [363] Y. Shtanov and V. Sahni, Class. Quant. Grav. 19, L101- [astro-ph.CO]]. L107 (2002). [340] V.H. Robles and T. Matos, Mon. Not. Roy. As- [364] A.S. Silbergleit and A.D. Chernin, Interacting Dark En- tron. Soc. 422, 282-289 (2012) doi:10.1111/j.1365- ergy and the Expansion of the Universe, Springer Briefs 2966.2012.20603.x [arXiv:1201.3032 [astro-ph.CO]]. in Physics (Springer, New York, 2017). [341] V.H. Robles, J.S. Bullock and M. Boylan-Kolchin, [365] S. Silva e Costa, Mod. Phys. Lett. A 24, 531-540 (2009) Mon. Not. Roy. Astron. Soc. 483, no.1, 289-298 (2019) doi:10.1142/S021773230902845X. doi:10.1093/mnras/sty3190 [arXiv:1807.06018 [astro- [366] R.F. Sistero, Astrophys. Space Sci. 17, 150 (1972). ph.CO]]. [367] G.F. Smoot, M.V. Gorenstein, and R.A. Muller, Phys. [342] R.C. Roeder, Nature 216, 774 (1967). Rev. Lett. 39, 898-901 (1977). [343] H.C. Rosu, P. Espinoza and M. Reyes, Nuovo Cim. B [368] S. Sonego and V. Talamini, Am. J. Phys. 80, 114, 1439-1444 (1999) [arXiv:gr-qc/9910070 [gr-qc]]. 670-679 (2012) doi:10.1119/1.4731258 [arXiv:1112.4319 [344] H.C. Rosu and P. Ojeda-May, Int. J. Theor. Phys. [physics.ed-ph]]. 45, 1191-1196 (2006) doi:10.1007/s10773-006-9123-2 [369] T.P. Sotiriou and V. Faraoni, Class. Quant. Grav. 25, [arXiv:gr-qc/0510004 [gr-qc]]. 205002 (2008) doi:10.1088/0264-9381/25/20/205002 [345] H.C. Rosu and K. V. Khmelnytskaya, SIGMA 7, 013 [arXiv:0805.1249 [gr-qc]]. (2011) doi:10.3842/SIGMA.2011.013 [arXiv:1012.1920 [370] T.P. Sotiriou and V. Faraoni, Rev. Mod. Phys. [gr-qc]]. 82, 451-497 (2010) doi:10.1103/RevModPhys.82.451 [346] J. Sadeghi, M.R. Setare, A. Banijamali and [arXiv:0805.1726 [gr-qc]]. F. Milani, Phys. Lett. B 662, 92-96 (2008) [371] D. Spergel and U.L. Pen, Astrophys. J. Lett. 491, L67- doi:10.1016/j.physletb.2008.02.062 [arXiv:0804.0553 L71 (1997) doi:10.1086/311074 [arXiv:astro-ph/9611198 [hep-th]]. [astro-ph]]. [347] D.S. Salopek and J.R. Bond, Phys. Rev. D 42, 3936- [372] R. Stabell and S. Refsdal, Mon. Not. Roy. Astron. Soc. 3962 (1990) doi:10.1103/PhysRevD.42.3936 132, 379 (1966). [348] A. Sapar, Astron. Zh. (USSR) 47, 503 (1970) (in Rus- [373] S.P. Starkovich and F.I. Cooperstock, Astrophys. J. 398, sian). 1 (1992). [349] E.N. Saridakis and J.M. Weller, Phys. Rev. D [374] A.A. Starobinsky, Phys. Lett. B 91, 99-102 (1980) 81, 123523 (2010) doi:10.1103/PhysRevD.81.123523 doi:10.1016/0370-2693(80)90670-X [arXiv:0912.5304 [hep-th]]. [375] A.A. Starobinsky, JETP Lett. 30 (1979) 682 [Pisma Zh. [350] E.N. Saridakis, Nucl. Phys. B 830, 374-389 (2010) Eksp. Teor. Fiz. 30 (1979) 719]. doi:10.1016/j.nuclphysb.2010.01.005 [arXiv:0903.3840 [376] H. Stefan˘ci˘c,˘ Phys. Rev. D 71, 084024 [astro-ph.CO]]. (2005) doi:10.1103/PhysRevD.71.084024 [351] M. Sasaki, J. Valiviita and D. Wands, Phys. Rev. [arXiv:astro-ph/0411630 [astro-ph]]. D 74, 103003 (2006) doi:10.1103/PhysRevD.74.103003 [377] H. Stephani, General Relativity: An Introduction to the [arXiv:astro-ph/0607627 [astro-ph]]. Theory of the Gravitational Field (Cambridge Univer- [352] M. Sasaki, Prog. Theor. Phys. 120, 159-174 (2008) sity Press, Cambridge, UK, 1982). doi:10.1143/PTP.120.159 [arXiv:0805.0974 [astro-ph]]. [378] H. Stephani, D. Kramer, M. MacCallum, C. Hoense- [353] F.E. Schunck and E.W. Mielke, Phys. Rev. D laers, E. Herlt, Exact Solutions of Einstein’s Field Equa- 50, 4794-4806 (1994) doi:10.1103/PhysRevD.50.4794 tions, 2nd edition (Cambridge University Press, Cam- [arXiv:gr-qc/9407041 [gr-qc]]. bridge, UK, 2003). [354] I. Semiz, Phys. Rev. D 85, 068501 (2012) [379] A. Suarez and T. Matos, Mon. Not. Roy. Astron. Soc. doi:10.1103/PhysRevD.85.068501 416, 87 (2011) doi:10.1111/j.1365-2966.2011.19012.x [355] M.R. Setare, Phys. Lett. B 641, 130- [arXiv:1101.4039 [gr-qc]]. 133 (2006) doi:10.1016/j.physletb.2006.08.039 [380] A. Suárez, V.H. Robles and T. Matos, Astrophys. Space [arXiv:hep-th/0611165 [hep-th]]. Sci. Proc. 38, 107-142 (2014) doi:10.1007/978-3-319- [356] M.R. Setare and E.N. Saridakis, Phys. Lett. B 02063-1_9 [arXiv:1302.0903 [astro-ph.CO]]. 668, 177-181 (2008) doi:10.1016/j.physletb.2008.08.033 [381] R.A. Sunyaev and Y.B. Zeldovich, Astrophys. Space Sci. [arXiv:0802.2595 [hep-th]]. 7, 20-30 (1970). [357] M.R. Setare and E.N. Saridakis, JCAP 09, 026 (2008) [382] M. Szydlowski, W. Godlowski and R. Wojtak, Gen. doi:10.1088/1475-7516/2008/09/026 [arXiv:0809.0114 Rel. Grav. 38, 795 (2006) doi:10.1007/s10714-006-0265- [hep-th]]. 6 [arXiv:astro-ph/0505202 [astro-ph]]. [358] M.R. Setare and E.N. Saridakis, Phys. Rev. D [383] J.L. Synge, Relativity: The General Theory (North Hol- 79, 043005 (2009) doi:10.1103/PhysRevD.79.043005 land, Amsterdam, 1955). [arXiv:0810.4775 [astro-ph]]. [384] G.E. Tauber, J. Math. Phys. 8, 118 (1967). [359] M.R. Setare and E.N. Saridakis, Int. J. Mod. Phys. [385] R.C. Tolman, Phys. Rev. 38, 1758 (1931). D 18, 549-557 (2009) doi:10.1142/S0218271809014625 [386] R.C. Tolman, Relativity, Thermodynamics and Cosmol- [arXiv:0807.3807 [hep-th]]. ogy (Clarendon Press, Oxford, 1934), pp. 407-411. [360] M.R. Setare and J. Sadeghi, Int. J. Theor. Phys. [387] M. Tsamparlis and A. Paliathanasis, Symmetry 47, 3219-3225 (2008) doi:10.1007/s10773-008-9757-3 10, no.7, 233 (2018) doi:10.3390/sym10070233 [arXiv:0805.1117 [gr-qc]]. [arXiv:1806.05888 [gr-qc]]. [361] S.G. Shi, Y. S. Piao and C. F. Qiao, JCAP [388] J.-P. Uzan and R. Lehoucq, Eur. J. Phys. 22, 371 04, 027 (2009) doi:10.1088/1475-7516/2009/04/027 (2001). [arXiv:0812.4022 [astro-ph]]. [389] L. A. Urena-López, [arXiv:physics/0609181 [physics]]. [362] I.S. Shikin, Sov. Phys. (Doklady) 13, 320 (1968). [390] J.P. Vajk, J. Math. Phys. 10, 1145 (1969). 46

[391] J. Valiviita, E. Majerotto and R. Maartens, JCAP 14, 1873-1882 (2005) doi:10.1142/S0218271805007486 07, 020 (2008) doi:10.1088/1475-7516/2008/07/020 [arXiv:gr-qc/0509036 [gr-qc]]. [arXiv:0804.0232 [astro-ph]]. [411] J.Q. Xia, B. Feng and X.M. Zhang, Mod. Phys. Lett. A [392] J.A. Vazquez, S. Hee, M.P. Hobson, A.N. Lasenby, 20, 2409-2416 (2005) doi:10.1142/S0217732305017445 M. Ibison and M. Bridges, JCAP 07, 062 (2018) [arXiv:astro-ph/0411501 [astro-ph]]. doi:10.1088/1475-7516/2018/07/062 [arXiv:1208.2542 [412] H.H. Xiong, Y.F. Cai, T. Qiu, Y.S. Piao and [astro-ph.CO]]. X. Zhang, Phys. Lett. B 666, 212-217 (2008) [393] A.K. Verma, Astrophys. Space Sci. 321, 73-77 (2009). doi:10.1016/j.physletb.2008.07.053 [arXiv:0805.0413 [394] A. Vilenkin, Phys. Rev. Lett. 53, 1016-1018 (1984) [astro-ph]]. doi:10.1103/PhysRevLett.53.1016 [413] X.D. Xu, B. Wang, P. Zhang and F. Atrio- [395] A.C. Vincent and J.M. Cline, JHEP 10, 093 (2008) Barandela, JCAP 12, 001 (2013) doi:10.1088/1475- doi:10.1088/1126-6708/2008/10/093 [arXiv:0809.2982 7516/2013/12/001 [arXiv:1308.1475 [astro-ph.CO]]. [astro-ph]]. [414] S. Yokoyama, T. Suyama and T. Tanaka, JCAP [396] R.V. Wagoner, Phys. Rev. D 1, 3209-3216 (1970) 07, 013 (2007) doi:10.1088/1475-7516/2007/07/013 doi:10.1103/PhysRevD.1.3209 [arXiv:0705.3178 [astro-ph]]. [397] J. Wainwright and G.F.R. Ellis (eds.) Dynamical Sys- [415] S. Yokoyama, T. Suyama and T. Tanaka, Phys. Rev. tems in Cosmology (Cambridge University Press, Cam- D 77, 083511 (2008) doi:10.1103/PhysRevD.77.083511 bridge, UK, 1997). [arXiv:0711.2920 [astro-ph]]. [398] R.M. Wald, General Relativity (Chicago University [416] A.V. Yurov and V.A. Yurov, J. Math. Phys. 51, Press, Chicago, 1984). 082503 (2010) doi:10.1063/1.3460856 [arXiv:0809.1216 [399] J. Wang and S.p. Yang, J. Theor. Phys. 1, 62-75 (2012) [hep-th]]. [arXiv:0901.1441 [gr-qc]]. [417] A. Zampeli, T. Pailas, P.A. Terzis and [400] J. Wang, S.w. Cui and S.p. Yang, Phys. Lett. B T. Christodoulakis, JCAP 05, 066 (2016) 683, 101-107 (2010) doi:10.1016/j.physletb.2009.11.064 doi:10.1088/1475-7516/2016/05/066 [arXiv:1511.08382 [arXiv:0901.1439 [gr-qc]]. [gr-qc]]. [401] T. Wang, Phys. Rev. D 82, 123515 (2010) [418] Y.B. Zeldovich and R.A. Sunyaev, Astrophys. Space Sci. doi:10.1103/PhysRevD.82.123515 [arXiv:1008.3198 4, 301-316 (1969) doi:10.1007/BF00661821 [astro-ph.CO]]. [419] X. Zhang, Commun. Theor. Phys. 44, 762-768 (2005) [402] B. Wang, E. Abdalla, F. Atrio-Barandela and doi:10.1088/6102/44/4/762 D. Pavon, Rept. Prog. Phys. 79, no.9, 096901 (2016) [420] X.F. Zhang and T. Qiu, Phys. Lett. B 642, doi:10.1088/0034-4885/79/9/096901 [arXiv:1603.08299 187-191 (2006) doi:10.1016/j.physletb.2006.09.038 [astro-ph.CO]]. [arXiv:astro-ph/0603824 [astro-ph]]. [403] H. Wei, R.G. Cai and D.F. Zeng, Class. Quant. Grav. [421] S. Zhang and B. Chen, Phys. Lett. B 669, 4-8 (2008) 22, 3189-3202 (2005) doi:10.1088/0264-9381/22/16/005 doi:10.1016/j.physletb.2008.09.025 [arXiv:0806.4435 [arXiv:hep-th/0501160 [hep-th]]. [hep-ph]]. [404] H. Wei and S.N. Zhang, Phys. Rev. D 76, 063005 (2007) [422] X. Zhang, Phys. Rev. D 79, 103509 (2009) doi:10.1103/PhysRevD.76.063005 [arXiv:0705.4002 [gr- doi:10.1103/PhysRevD.79.103509 [arXiv:0901.2262 qc]]. [astro-ph.CO]]. [405] S. Weinberg, Gravitation and Cosmology (J. Wiley & [423] G.B. Zhao, J.Q. Xia, M. Li, B. Feng Sons, New York, 1972). and X. Zhang, Phys. Rev. D 72, 123515 [406] C. Wetterich, Lect. Notes Phys. 892, 57 (2015) (2005) doi:10.1103/PhysRevD.72.123515 doi:10.1007/978-3-319-10070-8_3 [arXiv:1402.5031 [arXiv:astro-ph/0507482 [astro-ph]]. [astro-ph.CO]]. [424] W. Zhao, Phys. Rev. D 73, 123509 [407] C.M. Will, Theory and Experiment in Gravitational (2006) doi:10.1103/PhysRevD.73.123509 Physics (Cambridge University Press, Cambridge, [arXiv:astro-ph/0604460 [astro-ph]]. 1993). [425] V.M. Zhuravlev, S.V. Chervon and V.K. Shchigolev, [408] C.M. Will, Living Rev. Rel. 17, 4 (2014) J. Exp. Theor. Phys. 87, 223-228 (1998) doi:10.12942/lrr-2014-4 [arXiv:1403.7377 [gr-qc]]. doi:10.1134/1.558649 [409] F.L. Williams, P.G. Kevrekidis, T. Christodoulakis, C. [426] H. Ziaeepour, Phys. Rev. D 86, 043503 (2012) Helias, G.O. Papadopoulos and T. Grammenos, Trends doi:10.1103/PhysRevD.86.043503 [arXiv:1112.6025 in General Relativity and Quantum Cosmology, edited [astro-ph.CO]]. by Ch.V. Benton (Nova Science, New York, 2006), p. 37. [427] W. Zimdahl, Int. J. Geom. Meth. Mod. Phys. [410] P.x. Wu and H.w. Yu, Int. J. Mod. Phys. D 11, 1460014 (2014) doi:10.1142/S0219887814600147 [arXiv:1404.7334 [astro-ph.CO]].