Cosmological Models Based On A Complex Scalar Field With A Power-Law Potential With Polytropic Equation - Chavanis2022

PHYSICAL REVIEW D 106, 043502 (2022)
Cosmological models based on a complex scalar field with a power-law

potential associated with a polytropic equation of state
Pierre-Henri Chavanis
Laboratoire de Physique Théorique, Université de Toulouse, CNRS, UPS, Toulouse, France
(Received 9 November 2021; accepted 28 June 2022; published 1 August 2022)
We construct cosmological models based on a complex scalar field with a power-law potential
K m 2γ
V ¼ γ−1 ð ℏ Þ jφj2γ associated with a polytropic equation of state P ¼ Kργ (the potential associated with an
isothermal equation of state P ¼ ρkB T=m is V ¼ mkℏB2 T jφj2 ½lnðm2 jφj2 =ρ ℏ2 Þ − 1 and the potential
associated with a logotropic equation of state P ¼ A lnðρ=ρP Þ is V ¼ −A½lnðm2 jφj2 =ℏ2 ρP Þ þ 1). We
consider a fast oscillation regime of “spintessence” where the equations of the problem can be simplified.
We study all possible cases with arbitrary (positive and negative) values of the polytropic constant and
polytropic index. The ΛCDM model, the Chaplygin gas model, and the Bose-Einstein condensate model
are recovered as particular cases of our study corresponding to a constant potential (γ ¼ 0), an inverse
square-law potential (γ ¼ −1), and a quartic potential (γ ¼ 2). We also derive the two-fluid representation
of the Chaplygin gas model.
DOI: 10.1103/PhysRevD.106.043502
I. INTRODUCTION Different forms of SFs were introduced such as quintes-

Scalar fields (SFs) have been invoked in different sence [11–13], k-essence [14,15], phantom (ghost) fields
domains of particle physics, quantum field theory, astro- [16,17], tachyons [18–23], quintom [24–26], etc. They
physics, and cosmology. In particle physics and string usually involve a mass of the order of the current Hubble
theory, they arise in a natural manner as bosonic spin-0 scale (m ∼ H 0 ℏ=c2 ∼ 10−33 eV=c2 ). It has also been pro-
particles described by the Klein-Gordon (KG) equation posed that dark matter (DM) and superfluid stars may be
[1,2]. Examples include the Higgs particle, the inflaton, the described in terms of a SF which represents the wave
dilaton field of superstring theory, pseudo-Nambu- function of a self-gravitating Bose-Einstein condensate
Goldstone bosons, tachyons, and so on. SFs also arise (BEC). This leads to the concept of boson stars [27–29]
in the Kaluza-Klein and Brans-Dicke theories [3] as well and BEC stars [30] that could describe DM stars or
as in quantum gravity, supergravity, and superstring the superfluid core of neutron stars. DM halos may also
models (e.g., hidden sector fields and moduli). In cosmol- be interpreted as giant self-gravitating BECs. In particular,
ogy, SFs were first introduced to explain the phase of the fuzzy dark matter (FDM) model is based on the
inflation in the primordial universe assumed to be driven assumption that DM is made of extremely light scalar
by vacuum energy [4]. The inflaton, which has its origin particles (ultralight axions) with mass m ∼ 10−22 eV=c2
in the quantum fluctuations of the vacuum, is usually [31,32]. At large scales, FDM behaves as CDM but at
associated with a nonequilibrium phase transition. In small scales (≲1 kpc), the wave (quantum) properties of
canonical SF models based on a Lagrangian of the form the bosons manifest themselves and may solve the prob-
L ¼ X − VðφÞ, where X ¼ 12 ∂μ φ∂μ φ is the kinetic term lems of CDM such as the core-cusp problem [33], the
and VðφÞ is the potential term, the physical information on missing satellite problem [34–36], and the “too big to fail”
the system is contained in the potential VðφÞ of the SF. problem [37].
In k-inflation models based on a Lagrangian of the form In cosmology, most models attempting to describe the
L ¼ LðXÞ [5,6], the physical information on the system is evolution of the universe by a SF consider a real SF.
encapsulated in the nonstandard kinetic term.1 After the However, complex SFs should be considered as well
discovery of the accelerating expansion of the universe because they can be given a physical justification in relation
[7–10], SFs were used to describe dark energy (DE). to the Higgs mechanism in particle physics or to the
phenomenon of Bose-Einstein condensation in ultracold
gases. In addition, complex SFs are potentially more
1
General Lagrangians of the form LðX; φÞ have also been relevant than real SFs because they can form stable DM
considered in Refs. [5,6]. halos while DM halos made of real SFs are either
2470-0010=2022=106(4)=043502(39) 043502-1 © 2022 American Physical Society

PIERRE-HENRI CHAVANIS PHYS. REV. D 106, 043502 (2022)
dynamically unstable [38] or oscillating [39].2 This is radiationlike era, and a DM era.5 It is also possible to
basically due to the fact that the charge of a complex SF account for the present acceleration of the universe (DE)
is conserved while real SFs have no conserved charge. by introducing a cosmological constant Λ or by shifting
Therefore, a complex SF seems more promising than a real the origin of the SF potential to a small positive value
SF to account both for the cosmic evolution of the universe equal to the cosmological energy density ϵΛ ¼ ρΛ c2 so
2 2 λ
(cosmological background) and the formation of large-scale that V tot ¼ ρΛ c2 þ m2ℏc2 jφj2 þ 4ℏc jφj4 (see Appendix E of
structures (galaxies and DM halos). [52]). The radiationlike era only exists for sufficiently
The cosmological evolution of a spatially homogeneous large values of the self-interaction parameter [40,41]. Self-
complex SF with a quartic self-interaction potential (pos- interacting SFs may be more relevant than noninteracting
sibly representing the wave function of a relativistic BEC) SFs. Indeed, it is found in [40,41] that the FDM model
was recently studied by Li et al. [40] and Suárez and which assumes that the bosons are noninteracting is not
Chavanis [41].3 These authors considered a slow oscil- consistent with cosmological constraints from cosmic
lation regime where the SF rolls down the potential without microwave background (CMB) and big bang nucleosyn-
oscillating and a fast oscillation regime of “spintessence” thesis (BBN).
λ
[42] where the direction (phase) of the SF rotates rapidly in A complex SF with an attractive 4ℏc jφj4 (λ < 0) self-
the complex plane while its modulus changes slowly interaction can display, in the fast oscillation regime, two
(adiabatically). In the slow oscillation regime, there is types of behaviors which are associated with two branches
an interesting situation of “kination” where the kinetic of solutions. These solutions start at a nonzero scale factor
term dominates the potential term and the SF behaves with a finite energy density. In this sense, the SF emerges
as stiff matter [43]. In the fast oscillation regime, the “suddenly” in the universe. Initially, there is a very short
equations of the problem can be averaged over the cosmic stringlike era (P ∼ −ϵ=3). On the normal (non-
oscillations of the SF and simplified. Using a hydro- relativistic) branch, the energy density decreases to zero as
dynamic representation of the KG equation, Suárez and a−3 and the SF behaves as pressureless DM (P ≃ 0). On the
Chavanis [41] showed that the fast oscillation regime of peculiar (ultrarelativistic) branch, the energy density slowly
spintessence is equivalent to the Thomas-Fermi (TF) decreases and tends to a constant at late times giving rise to
approximation where the quantum potential can be
neglected (this amounts to taking ℏ ¼ 0 in the hydro- 5
It is well known that a noninteracting real SF undergoes a stiff
dynamic equations). They investigated the domain of matter era followed by an inflation era and, finally (in the fast
validity of this regime in detail. oscillation regime), by a matter era [44–47]. This evolution (stiff
A complex SF with a repulsive 4ℏc λ
jφj4 (λ > 0) self- matter → inflation → matter) was also found by Scialom and
Jetzer [48,49] for a self-interacting complex SF. These authors
interaction undergoes, in the fast oscillation regime, a showed that the inflation era occurring in the slow-roll regime is
radiationlike4 era (with an equation of state P ∼ ϵ=3) sufficiently long provided that the bosonic charge Q is close to
followed by a matterlike era (P ≃ 0). However, the fast zero (in that case, the phase of the SF remains approximately
oscillation regime is not valid at very early times. It is constant and thus inflation is essentially driven by one component
of the SF as if it were real). On the other hand, for a massive
preceded, in the primordial universe, by a stiff matter era complex SF, inflation is less effective when the self-interaction λ
(P ¼ ϵ) which is valid in the slow oscillation regime. increases. This is because the potential becomes less flat and, as a
Therefore, a complex SF with a repulsive jφj4 self- consequence, the slow-roll approximation needed for inflation to
interaction undergoes successively a stiff matter era, a occur is no longer well satisfied. Therefore, inflation takes place
only if the charge and the self-interaction of the SF are
sufficiently small. Otherwise, the background passes directly
from the stiff matter era to the oscillatory phase [40,41] (there
2 may, however, be an inflation era prior to the stiff matter era [50]).
These dynamical instabilities occur in the relativistic regime
of boson stars. Although DM halos can usually be described by Arbey et al. [51] also considered (independently from [48,49]) a
nonrelativistic equations, there are certain situations in which self-interacting complex SF. For λ ¼ 0 they found a stiff matter
relativity may be important. In addition, bosons described by a era, followed by an epoch where the SF is constant (like the
complex SF with a global Uð1Þ symmetry associated with a inflation era reported previously but occurring during the stan-
conserved charge (Noether theorem) can form BECs even in the dard radiation era), and a DM era. For λ ≠ 0 they found, in the
early universe while this is more difficult for a system of bosons fast oscillation regime, a dark radiation era followed by a DM era.
described by a real SF like the QCD axion. For a complex SF in the fast oscillation regime, only the phase θ
3 of the SF changes (spintessence). The modulus jφj of the SF
Li et al. [40] focused on a repulsive quartic self-interaction
evolves slowly without oscillating. By contrast, for a real SF in
while Suárez and Chavanis [41] considered both repulsive and the fast oscillation regime, φðtÞ oscillates rapidly by taking
attractive quartic self-interactions. They also developed a general positive and negative values. Arbey et al. [51] considered a fast
formalism that is valid for an arbitrary potential of interaction oscillation regime different from spintessence where the complex
Vðjφj2 Þ. SF behaves as an effective oscillating real SF (axion). In the
4
The radiationlike era is due to the jφj4 self-interaction of the present paper, when considering the fast oscillation regime of a
SF. It is different from the standard radiation era due to photons or complex SF, we shall systematically assume that it corresponds to
ultrarelativistic particles like neutrinos. the spintessence regime.
043502-2
COSMOLOGICAL MODELS BASED ON A COMPLEX SCALAR … PHYS. REV. D 106, 043502 (2022)
a de Sitter evolution. In that case, the SF behaves as DE In the present paper, we do not necessarily try to construct
(P ∼ −ϵ).6 On the normal branch, the fast oscillation regime a realistic model of universe. We consider an arbitrary
is not valid at early times so that quantum mechanics power-law potential V ∼ Kjφj2γ and describe the different
(ℏ ≠ 0) must be taken into account. In the very early types of evolutions that it produces depending on the value
universe, a complex SF with an attractive self-interaction of γ and on the sign of K. Our work therefore complements
may undergo an inflation era. If the self-interaction is the study of the jφj4 potential performed in [40,41].
sufficiently small, the inflation era is followed by a stiff The paper is organized as follows. In Sec. II we consider
matter era. On the peculiar branch, if the self-interaction is a spatially homogeneous SF in an expanding universe.
sufficiently large, the fast oscillation regime is valid Following our previous paper [41] we establish the general
initially but it ceases to be valid at late times so that equations of the problem in the fast oscillation (or TF)
quantum mechanics (ℏ ≠ 0) must be taken into account regime for an arbitrary potential of interaction Vðjφj2 Þ and
ultimately. As a result, the de Sitter regime may stop and the show that the SF behaves as a barotropic fluid described by
SF may eventually enter in a stiff matter era implying that an equation of state PðρÞ determined by the potential. In
the late universe passes from a phase of acceleration to a Sec. III we show that the energy density ϵ can be written as
phase of deceleration. the sum of a rest-mass energy density ρm c2 and an internal
In the present paper, we study how the evolution of the energy density u which play respectively the roles of DM
universe depends on the form of the SF potential. Many and DE. We relate the internal energy density u to the SF
types of potentials have been studied in the past in the case potential V. We also introduce a two-fluid model associated
of a real SF. The novelty of our approach is to consider the with the single dark fluid (or SF) model. In Sec. IV we
cosmic evolution of a complex SF. Furthermore, we focus consider power-law potentials associated with polytropic
on the fast oscillation regime where the SF experiences the and isothermal equations of state. In Sec. V we describe all
process of spintessence. The general equations of the possible cosmological evolutions corresponding to a pos-
problem, valid for an arbitrary self-interaction potential, itive polytropic constant K > 0. In Sec. VI we describe all
are given in [41]. In the present paper, we consider possible cosmological evolutions corresponding to a neg-
K m 2γ
algebraic potentials of the form V ¼ γ−1 ð ℏ Þ jφj2γ where ative polytropic constant K < 0. In Sec. VII we investigate
K and γ can be positive or negative. These power-law the validity of the fast oscillation regime in these models.
potentials are associated with the polytropic equation of In Sec. VIII we briefly consider the problem of structure
2
state P ¼ Kργ where ρ ¼ mℏ2 jφj2 is the pseudorest-mass formation. A summary of our main results is provided in
density. The ΛCDM model is recovered for a constant the conclusion. Complements about the general formalism
potential V ¼ ρΛ c2 (γ ¼ 0 and K ¼ −ρΛ c2 ) corresponding are given in the appendixes.
to a constant equation of state P ¼ −ρΛ c2 . The Chaplygin
gas is recovered for an inverse square law potential V ¼ II. THEORY OF A COMPLEX SF
A ℏ 2 1
2 ðmÞ jφj2 (γ ¼ −1 and K ¼ −A) corresponding to the
In this section, we recall the basic equations governing
equation of state P ¼ −A=ρ. The standard BEC model is the cosmological evolution of a spatially homogeneous
2
recovered for a quartic potential V ¼ 2πams3ℏ ðmℏ Þ4 jφj4 (γ ¼ 2 complex SF with an arbitrary self-interaction potential in a
and K ¼ 2πas ℏ2 =m3 ) corresponding to the equation of Friedmann-Lemaître-Robertson-Walker (FLRW) universe.
2 We also recall how these equations can be simplified in the
state P ¼ 2πams3ℏ ρ2 . We also consider a potential of the form
fast oscillation regime (equivalent to the classical or TF
V ¼ mkℏB2 T jφj2 ½lnðm2 jφj2 =ρ ℏ2 Þ − 1 associated with the approximation) that is considered in the following sections.
isothermal equation of state P ¼ ρkB T=m. This is the limit We refer to our previous papers [41,52,55] and references
of the polytropic equation of state when γ → 1. In a therein for a more detailed discussion.
companion paper [52], we specifically study a potential
of the form V ¼ −A½lnðm2 jφj2 =ℏ2 ρP Þ þ 1 associated with
A. Spatially homogeneous SF
the logotropic equation of state P ¼ A lnðρ=ρP Þ. This is the
limit of the polytropic equation of state when γ → 0 and Let us consider a complex SF φ with a self-interaction
K → þ∞ with A ¼ Kγ constant (see Sec. 3 of [54] and potential Vðjφj2 Þ described by the KG equation. For a
Appendix A of [52] for a precise statement). The logotropic spatially homogeneous SF φðtÞ evolving in an expanding
model gives a good agreement with the cosmological background, the KG equation takes the form7
observations of our universe and is able to account for
the universal value of the surface density Σobs 1 d2 φ 3H dφ m2 c2 dV
0 ¼ þ 2 þ 2 φþ2 φ ¼ 0; ð1Þ
141þ83 M ⊙ =pc 2
of DM halos without free parameter. c2 dt2 c dt ℏ djφj2
−52
6 7
This intriguing behavior was first reported by Suárez and See Appendix A for the general expression of the KG
Chavanis [41] and further discussed by Carvente et al. [53]. equation valid for a spatially inhomogeneous SF.
043502-3
where H ¼ a=a _ is the Hubble parameter and aðtÞ is the B. Charge of the SF
scale factor. The second term in Eq. (1) is the Hubble drag. Writing the complex SF as
The rest-mass term (third term) can be written as φ=λ2C
where λC ¼ ℏ=mc is the Compton wavelength (m is the φ ¼ jφjeiθ ; ð7Þ
mass of the SF). The total potential including the rest-mass
term and the self-interaction term reads where jφj is its modulus and θ its phase (angle), inserting
this decomposition into the KG equation (1), and separating
m2 c2 2
V tot ðjφj2 Þ ¼ jφj þ Vðjφj2 Þ: ð2Þ the real and imaginary parts, we obtain the following pair of
2ℏ2 equations:
The energy density ϵðtÞ and the pressure PðtÞ of the SF are
1 djφj dθ d2 θ 3H dθ
given by (see, e.g., [55] for details) 2
2 þ jφj 2 þ 2 jφj ¼ 0; ð8Þ
c dt dt dt c dt

1 dφ2 m2 c2 2 2
ϵ¼ 2 þ jφj þ Vðjφj2 Þ; ð3Þ 1 d2 jφj dθ 3H djφj
2c dt 2ℏ2 − jφj þ 2
c2 dt2 dt c dt

1 dφ2 m2 c2 2 m2 c2 dV
P¼ − jφj − Vðjφj2 Þ: ð4Þ þ jφj þ 2 jφj ¼ 0: ð9Þ
2c2 dt 2ℏ2 ℏ 2
djφj2
The equation of state parameter is defined by w ¼ P=ϵ. Equation (8) can be rewritten as a conservation equation
The Friedmann equations determining the evolution of
the homogeneous background are d 3 2 dθ
a jφj ¼ 0: ð10Þ
dt dt
dϵ
þ 3Hðϵ þ PÞ ¼ 0 ð5Þ
dt _ we get
Introducing the pulsation ω ¼ −θ,
and
Qℏc2
ω¼ ; ð11Þ
8πG a3 jφj2
H2 ¼ 2 ϵ: ð6Þ
3c
where Q is a constant of integration which represents the
Equation (6) is valid in a flat universe (k ¼ 0) without charge of the SF [40,41,51,56,57]. This conservation law
cosmological constant (Λ ¼ 0). The Friedmann equations results from the identity Dμ Jμ ¼ 0 where Jμ is the KG
can be derived from the Einstein field equations by using current. The conservation of the charge is equivalent to the
the FLRW metric. The energy conservation equation (5) conservation of the boson number provided that antibosons
results from the identity Dν T μν ¼ 0, where T μν is the are counted negatively [58]. Equation (9) can be rewritten as
energy-momentum tensor. It can also be obtained from
the KG equation (1) by using Eqs. (3) and (4) (see d2 jφj 2 4
2 jφj þ 3H djφj þ m c jφj þ 2c2 dV jφj ¼ 0:
− ω
Appendix G of [55]). Inversely, the KG equation can be dt2 dt ℏ2 djφj2
derived from Eqs. (3)–(5). Once the SF potential Vðjφj2 Þ is ð12Þ
given, the Klein-Gordon-Friedmann (KGF) equations pro-
vide a complete set of equations that can in principle be Substituting Eq. (11) into Eq. (12) we obtain the differential
solved to obtain the evolution of the universe assuming that equation
the energy density is entirely due to the SF (for simplicity
we do not consider here the effect of other species like
d2 jφj djφj m2 c4 2 dV Q2 ℏ2 c4
standard radiation and baryonic matter). þ 3H þ jφj þ 2c jφj − ¼ 0:
Remark: From Eq. (5) we see that the energy density dt2 dt ℏ2 djφj2 a6 jφj3
decreases with the scale factor when w > −1 and increases ð13Þ
with the scale factor when w < −1. In the second case, the
universe has a phantom behavior (we will see that this This equation is exact. It determines the evolution of the
regime is not allowed for a complex SF in the fast modulus of the complex SF. It differs from the KG equation
oscillation regime). On the other hand, when k ¼ Λ ¼ 0, of a real SF by the presence of the last term and the fact that
the deceleration parameter q ¼ −̈aa=a_ 2 is given by φ is replaced by jφj. On the other hand, substituting Eq. (7)
q ¼ ð1 þ 3wÞ=2. Therefore, the universe is decelerating into Eqs. (3) and (4) we find that the energy density and the
when w > −1=3 and accelerating when w < −1=3. pressure are given by
043502-4
2
1 djφj 2 ω m2 c2 dV
ϵ¼ 2 þ þ jφj2 þ Vðjφj2 Þ; ð14Þ jφj2 ≃ V 0 ðhjφj2 iÞhjφj2 i; ð19Þ
2c dt 2c2 2ℏ2 djφj2
2
1 djφj 2 ω m2 c2 we get
P¼ þ − jφj2 − Vðjφj2 Þ: ð15Þ
2c2 dt 2c2 2ℏ2
m2 c2
hϵi ¼ hjφj2 i þ V 0 ðhjφj2 iÞhjφj2 i þ Vðhjφj2 iÞ; ð20Þ
ℏ2
C. Fast oscillation regime and spintessence
hPi ¼ V 0 ðhjφj2 iÞhjφj2 i − Vðhjφj2 iÞ: ð21Þ
_
In the fast oscillation regime ω ¼ jdθ=dtj ≫ H ¼ a=a,
where the pulsation is high with respect to the Hubble
expansion rate, Eq. (12) reduces to8 The equation of state parameter is then given by
m2 c4 dV P V 0 ðhjφj2 iÞhjφj2 i − Vðhjφj2 iÞ

ω ¼ 2 þ 2c2
2
: ð16Þ w¼ ¼ m2 c2 : ð22Þ
ℏ djφj2 ϵ 2 0 2 2
2 hjφj i þ V ðhjφj iÞhjφj i þ Vðhjφj iÞ
2
ℏ
This equation can be interpreted as a condition of equi- We note that the averages are not strictly necessary in
librium between the centrifugal force ω2 jφj and the force Eqs. (20)–(22) since the modulus of the SF changes slowly
c2 dV tot =djφj produced by the total SF potential (see with time. Equations (20) and (21) can also be obtained from
Sec. V. A. of [41]). When this condition is satisfied, the Eqs. (14) and (15) by using Eq. (16) and neglecting the
direction (phase) of the SF rotates rapidly in the complex term ðdjφj=dtÞ2 .
plane while its modulus changes slowly (adiabatically).
This is what Boyle et al. [42] call “spintessence.” There is
E. Hydrodynamic variables and TF approximation
no relation such as Eq. (16) for a real SF. Combining
Eqs. (11) and (16), we obtain Instead of working with the SF φðtÞ, we can use hydro-
dynamic variables.9 We write the SF in the de Broglie form
Q2 ℏ2 c4 m2 c4 dV (for a homogeneous SF)
¼ 2 þ 2c2 : ð17Þ
a6 jφj4 ℏ djφj2
ℏ pffiffiffiffiffiffiffiffi iStot ðtÞ=ℏ
φðtÞ ¼ ρðtÞe ; ð23Þ
This equation relates the modulus jφj of the SF to the scale m
factor a in the fast oscillation regime. The pulsation ω of
the SF is then given by Eq. (11) or (16). where
m2 2
D. Equation of state in the fast oscillation regime ρ¼ jφj ð24Þ
ℏ2
To establish the equation of state of the SF in the fast
oscillation regime, we can proceed as follows [40–42,59–65]. is the pseudorest-mass density10 and Stot ¼ ð1=2Þiℏ
Multiplying the KG equation (1) by φ and averaging over a lnðφ =φÞ is the total action of the SF. In terms of the
time interval that is much longer than the field oscillation pseudorest-mass density the total potential (2) can be
period ω−1 , but much shorter than the Hubble time H −1 , we written as
obtain
1
1 dφ2 m2 c2 V tot ðρÞ ¼ ρc2 þ VðρÞ: ð25Þ
2
¼ 2 hjφj i þ 2
dV 2
jφj : ð18Þ 2
c2 dt ℏ djφj2
On the other hand, the total energy of the SF (including its
This relation constitutes a sort of virial theorem. On the other rest-mass energy mc2 ) is
hand, for a spatially homogeneous SF, the energy density and
the pressure are given by Eqs. (3) and (4). Taking the average 9
See Appendixes A and B, and our previous works
value of the energy density and pressure, using Eq. (18), and [41,55,57,66,67], for a general hydrodynamical description of
making the approximation the SF valid for possibly inhomogeneous systems.
10
We stress that it is only in the nonrelativistic limit c → þ∞
that ρ has the interpretation of a rest-mass density (in this limit,
8
For a free field (V ¼ 0), the pulsation ω is proportional to the we also have ϵ ∼ ρc2 ). In the relativistic regime, ρ does not have a
mass of the SF (ω ¼ mc2 =ℏ) and the fast oscillation condition clear physical interpretation but it can always be introduced as a
becomes mc2 =ℏ ≫ H. convenient notation.
043502-5
dStot VðρÞ. Inversely, for a given equation of state, the potential is

Etot ðtÞ ¼ − : ð26Þ
dt given by13
Substituting Eq. (23) into the KG equation (1) and taking Z

PðρÞ
the imaginary part, we obtain the conservation equation VðρÞ ¼ ρ dρ: ð35Þ
ρ2
[41,57]
d The various correspondences between the results of this

ðρEtot a3 Þ ¼ 0: ð27Þ section and the results of the previous sections show that
dt the fast oscillation regime (ω ≫ H) is equivalent to the TF
It expresses the conservation of the charge of the SF.11 It or semiclassical approximation (ℏ → 0). We note that we
can be integrated into cannot directly take ℏ ¼ 0 in the KG equation (this is why
we have to average over the oscillations) while we can take
Etot Qm ℏ ¼ 0 in the hydrodynamic equations (see Refs. [41,55,57]
ρ ¼ 3 ; ð28Þ
mc2 a for details). This is an interest of the hydrodynamic
representation of the SF. It can be shown (see, e.g., [55]
where Q is the charge of the SF. These equations are and Appendix A) that Eqs. (33) and (34) remain valid for a
equivalent to Eqs. (10) and (11).12 Next, substituting spatially inhomogeneous SF in the TF approximation. They
Eq. (23) into the KG equation (1), taking the real part, determine the equation of state P ¼ PðϵÞ in parametric
and making the TF approximation ℏ → 0, we obtain the form. The equation of state parameter can be written as
Hamilton-Jacobi (or Bernoulli) equation [41,57]
P ρV 0 ðρÞ − VðρÞ
E2tot ¼ m2 c4 þ 2m2 c2 V 0 ðρÞ: ð29Þ w¼ ¼ 2 ; ð36Þ
ϵ ρc þ VðρÞ þ ρV 0 ðρÞ
This equation is equivalent to Eq. (16). It can be rewritten as which is equivalent to Eq. (22). We note that the condition
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi from Eq. (31) implies w > −1 so that a universe described
2
Etot ¼ mc 1 þ 2 V 0 ðρÞ:
2
ð30Þ by a complex SF in the fast oscillation regime has never a
c phantom behavior. On the other hand, the universe is
decelerating (w > −1=3) when 4ρV 0 ðρÞ − 2VðρÞ > −ρc2
Note that Eq. (29) requires that
[or equivalently 2ρV 0tot ðρÞ > V tot ðρÞ] and accelerating in
2 0 the opposite case. Finally, the pseudosquared speed of
V 0tot ðρÞ ¼ 1 þ V ðρÞ > 0: ð31Þ sound is
c2
Combining Eqs. (28) and (30), we obtain c2s ¼ P0 ðρÞ ¼ ρV 00 ðρÞ; ð37Þ
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 Qm while the true squared speed of sound is
ρ 1 þ 2 V 0 ðρÞ ¼ 3 ; ð32Þ
c a
ρV 00 ðρÞc2
c2s ¼ P0 ðϵÞc2 ¼ : ð38Þ
which corresponds to Eq. (17). Finally, writing Eqs. (3) and c þ 2V 0 ðρÞ þ ρV 00 ðρÞ
2
(4) in terms of hydrodynamic variables, making the TF
approximation ℏ → 0, and using the Bernoulli equa- There is a tachyonic instability when c2s < 0, i.e.,
tion (29), we get [41,57] V 00 ðρÞ < 0, possibly leading to Q-ball [68] formation
(a nontopological soliton), in agreement with the instability
ϵ ¼ ρc2 þ VðρÞ þ ρV 0 ðρÞ; ð33Þ criterion V 00 ðjφjÞ − V 0 ðjφjÞ=jφj < 0 of [42,63]. This is a
dynamical instability different from the gravitational Jeans
P ¼ ρV 0 ðρÞ − VðρÞ; ð34Þ instability.
Remark: More general equations valid beyond
in agreement with Eqs. (20) and (21). Equation (34) the TF limit (i.e. taking into account quantum terms
determines the equation of state PðρÞ for a given potential
13
11
We note that the potential is defined from the pressure up to a
For a spatially homogeneous SF, the density of charge term of the form Aρ, where A is a constant. If we add a term Aρ in
(or rest-mass density ρm ) is proportional to ρEtot (see Sec. III the potential V, we do not change the pressure P but we introduce
and Appendix B).
12
a term 2Aρ in the energy density. On the other hand, if we add a
To make the link between the SF variables
pffiffiffi and the hydro- constant term C in the potential V (cosmological constant), this
dynamical variables, we use jφj ¼ ðℏ=mÞ ρ, θ ¼ Stot =ℏ and adds a term −C in the pressure and a term þC in the energy
ω ¼ −θ_ ¼ −S_ tot =ℏ ¼ Etot =ℏ. density.
043502-6
depending on ℏ) are given in [41,55,57,66,67] and in SF which can be represented, through the de Broglie
Appendixes A and B. transformation, as a single dark fluid. It therefore provides
a unified dark matter and dark energy (UDM) model. In this
F. Cosmological evolution of a spatially homogeneous section, we interpret this dark fluid as a barotropic gas at
complex SF in the fast oscillation regime T ¼ 0 and we determine its rest-mass density ρm, internal
energy density u, and equation of state Pðρm Þ in terms of
The differential equation determining the evolution of the SF potential. This equivalence is valid only in the TF
the scale factor aðtÞ of the Universe induced by a spatially approximation. Following [50,54,69], we argue that the
homogeneous complex SF in the regime where its oscil- rest-mass density ρm plays the role of DM and the internal
lations are faster than the Hubble expansion is given by [see energy density u plays the role of DE. This provides a
Eqs. (6), (32) and (33)] simple physical interpretation of these two mysterious
2 components.
3 a_ 1
¼ ρ þ 2 ½VðρÞ þ ρV 0 ðρÞ ð39Þ
8πG a c
A. First principle of thermodynamics
with
The first principle of thermodynamics for a relativistic
2 Qm gas can be written as
ρ 1 þ 2 V 0 ðρÞ ¼ 3 : ð40Þ
c a
ϵ 1 s
d ¼ −Pd þ Td ; ð45Þ
Actually, it is not convenient to solve the differential ρm ρm ρm
equation (39) for the scale factor a because we have to
inverse Eq. (40) in order to express ρ as a function of a in where s is the entropy density, P is the pressure, T is the
the right-hand side of Eq. (39). It is more convenient to temperature, and
view a as a function of ρ, given by Eq. (40), and transform
Eq. (39) into a differential equation for ρ. Taking the ϵ ¼ ρm c2 þ uðρm Þ ð46Þ
logarithmic derivative of Eq. (40), we get
is the energy density including the rest-mass energy
a_ 1 ρ_ ρV 00 ðρÞ density ρm c2 (where ρm ¼ nm is the rest-mass density)
¼− 1þ 2 : ð41Þ
a 3ρ c þ 2V 0 ðρÞ and the internal energy density uðρm Þ. We assume that
Tdðs=ρm Þ ¼ 0. This corresponds to cold (T ¼ 0) or isen-
Substituting this expression into Eq. (39), we obtain the tropic (s=ρm ¼ cst) gases. In that case, Eq. (45) reduces to
differential equation

2 ϵ 1 P
c 2
ρ_ ρc2 þ VðρÞ þ ρV 0 ðρÞ d ¼ −Pd ¼ 2 dρm : ð47Þ
¼ 00 : ð42Þ ρm ρm ρm
24πG ρ 2
½1 þ 2ρV ðρÞ
0 c þ2V ðρÞ
This equation can be rewritten as
For a given SF potential VðρÞ, this equation can be solved
easily as it is just a first-order differential equation for ρ. dϵ Pþϵ
¼ ; ð48Þ
Therefore, tðρÞ can be expressed in the form of an integral. dρm ρm
The temporal evolution of the scale factor aðtÞ is then
obtained in parametric form a ¼ aðρÞ and t ¼ tðρÞ from where the term on the right-hand side is the enthalpy h (the
Eqs. (40) and (42). This yields chemical potential is μ ¼ mh). We have
ðQmÞ1=3 Pþϵ dϵ dP
a¼ ð43Þ h¼ ; h¼ ; dh ¼ : ð49Þ
; ρm dρm ρm
ρ1=3 ½1 þ c22 V 0 ðρÞ1=6
2 1=2 Z 00 ðρÞ Equation (47) can be integrated into

c 1 þ c2ρV þ2V 0 ðρÞ
t¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi dρ: ð44Þ Z
24πG Pðρm Þ
ρ ρc2 þ VðρÞ þ ρV 0 ðρÞ ϵ ¼ ρm c2 þ ρm dρm ; ð50Þ
ρ2m
III. EFFECTIVE DARK MATTER establishing that
AND DARK ENERGY
Z
Pðρm Þ
In our model, there is no DM and no DE considered as uðρm Þ ¼ ρm dρm : ð51Þ
two different and independent species. There is just a single ρ2m
043502-7
This equation determines the internal energy density as a potential V of the corresponding SF in the TF approxima-
function of the equation of state Pðρm Þ. Inversely, the tion. According to Eq. (48), we have
equation of state is determined by the internal energy Z
density uðρm Þ through the relation ϵ0 ðρÞ
ln ρm ¼ dρ: ð57Þ
ϵþP
dðϵ=ρm Þ 2 uðρm Þ 0
Pðρm Þ ¼ − ¼ ρm ¼ ρm u0 ðρm Þ − uðρm Þ:
dð1=ρm Þ ρm Using Eqs. (33) and (34), which are valid in the TF
approximation, we obtain
ð52Þ
Z
c2 þ ρV 00 ðρÞ þ 2V 0 ðρÞ
We note that ln ρm ¼ dρ: ð58Þ
ρc2 þ 2ρV 0 ðρÞ
P0 ðρm Þ ¼ ρm u00 ðρm Þ: ð53Þ
This equation determines the function ρm ðρÞ. It can be
explicitly integrated into
B. Application to cosmology rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2
The previous results are general. We now consider a ρm ¼ ρ 1 þ 2 V 0 ðρÞ; ð59Þ
c
spatially homogeneous gas in an expanding universe.
Combining the energy conservation equation (5) with where the constant of integration has been determined in
the first principle of thermodynamics [Eq. (48)], we obtain order to obtain ρm ¼ ρ in the nonrelativistic limit. This
relation can also be established in the manner described in
dρm Appendix A [see Eq. (A37)]. We emphasize that Eq. (59) is
þ 3Hρm ¼ 0: ð54Þ
dt always valid in the TF approximation, even for an inho-
mogeneous SF. Eliminating ρ between PðρÞ [see Eq. (34)]
This equation expresses the conservation of the particle
and ρm ðρÞ [see Eq. (59)] we obtain the equation of state of
number (or rest mass). It can be integrated into ρm ∝ a−3 . the gas under the form Pðρm Þ. On the other hand, according
Inserting this relation into Eq. (46), we see that ρm plays the to Eq. (46), its internal energy density can be obtained from
role of DM while u plays the role of DE. This decom- the relation
position provides therefore a simple interpretation of DM
and DE in terms of a single DF [50,54,69]. Owing to this u ¼ ϵ − ρm c2 : ð60Þ
interpretation, we can write
From Eqs. (33) and (59), we get
Ωm;0 ϵ0
ρm c2 ¼ ð55Þ rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
a3
2
u ¼ ρc þ ρV ðρÞ þ VðρÞ − ρc 1 þ 2 V 0 ðρÞ;
2 0 2
ð61Þ
and c
where ρ ¼ ρðρm Þ can be obtained from Eq. (59). Therefore,
Ωm;0 ϵ0 Ωm;0 ϵ0
ϵ¼ þu ; ð56Þ for a given SF potential V, the rest-mass density (DM) is
a3 c2 a3
determined by Eq. (59) and the internal energy density
where ϵ0 is the present energy density of the universe and (DE) is determined by Eq. (61). Eliminating ρ between
Ωm;0 is the present proportion of DM. For given Pðρm Þ or Eqs. (59) and (61) we obtain uðρm Þ. We can also obtain
uðρm Þ we can get ϵðaÞ from Eq. (56). We can then solve the uðρm Þ from Pðρm Þ, or the converse, by using Eqs. (51) and
Friedmann equation (6) to obtain the temporal evolution of (52). As mentioned above, ρm mimics DM and u mimics
the scale factor aðtÞ [50,54,69]. DE. However, in our model, we have just one fluid (or just
Remark: Eqs. (46) and (52) determine the equation of one complex SF).
state P ¼ PðϵÞ. As a result, we can obtain Eq. (55) directly For a spatially homogeneous SF in a cosmological
from Eqs. (46), (52) and the energy conservation equa- context and in the TF approximation, combining the
tion (5). Indeed, combining these equations we obtain Hamilton-Jacobi (or Bernoulli) equation (30) with
Eq. (54) which integrates to give Eq. (55). Eq. (59), we find that the relation between the rest-mass
density ρm and the pseudorest-mass density ρ is
C. Determination of the rest-mass density and internal
Etot
energy density ρm ¼ ρ : ð62Þ
mc2
We now relate the rest-mass density ρm (proportional to
the charge density) and the internal energy density u of the Using Eqs. (59) and (62), Eqs. (28) and (32) can be
barotropic gas to the pseudorest-mass density ρ and rewritten as
043502-8
Qm equivalent for the evolution of the homogeneous back-

ρm ¼ : ð63Þ
a3 ground, they may differ for what concerns the formation of
the large-scale structures of the Universe.
The rest-mass density (or the charge density) decreases as
a−3 . This expresses the conservation of the charge of the SF IV. POWER-LAW POTENTIALS
or, equivalently, the conservation of the boson minus
antiboson number. Comparing Eqs. (55) and (63), we The equations of Secs. II and III are general. We now
establish that consider specific potentials Vðjφj2 Þ associated with poly-
tropic and isothermal equations of state.
Qmc2 ¼ Ωm;0 ϵ0 : ð64Þ
A. Polytropic equation of state
2
This shows that the constant Qmc (which is proportional We first consider a power-law SF potential that we write
to the charge of the SF or to the boson number) is equal to under the form (see also Appendix I of [41])
the present energy density of DM ϵm;0 ¼ Ωm;0 ϵ0 . 2γ
Remark: In the homogeneous case, we can directly 2 K m
Vðjφj Þ ¼ jφj2γ ðγ ≠ 1Þ: ð68Þ
deduce the relation γ−1 ℏ
Etot 2 For the sake of generality, we consider arbitrary values of K
ρm ¼ ρ 2 ¼ ρ 1 þ 2 V 0 ðρÞ ð65Þ (positive and negative) and arbitrary values of γ. The total
mc c
potential including the rest-mass term is
from Eqs. (28) and (32) if we assume Eq. (64), but this 2γ
derivation is less general than the calculations detailed 2 1 m2 c2 2 K m
V tot ðjφj Þ ¼ 2
jφj þ jφj2γ : ð69Þ
above which remain valid for inhomogeneous systems. 2 ℏ γ−1 ℏ
D. Two-fluid model Introducing the pseudorest-mass density defined by

Eq. (24), we get
As mentioned above, in our model, we have a single dark
fluid with an equation of state P ¼ Pðρm Þ. Still, the energy K γ
density (46) is the sum of two terms, a rest-mass density VðρÞ ¼ ρ ð70Þ
γ−1
term ρm which mimics DM and an internal energy density
term uðρm Þ which mimics DE. It is interesting to consider a and
two-fluid model which leads to the same results as our
single dark fluid model, at least for what concerns the 1 K γ
evolution of the homogeneous background. In this two- V tot ðρÞ ¼ ρc2 þ ρ: ð71Þ
2 γ−1
fluid model, one fluid corresponds to pressureless DM with
an equation of state Pm ¼ 0 and a density ρm c2 ¼ Using Eqs. (34), (37) and (70), we find that the derivative of
Ωm;0 ϵ0 =a3 determined by the energy conservation equation the potential, the pressure and the pseudosquared speed of
for DM, and the other fluid corresponds to DE with an sound are given by
equation of state Pde ðϵde Þ and an energy density ϵde ðaÞ
determined by the energy conservation equation for DE. Kγ γ−1
V 0 ðρÞ ¼ ρ ; ð72Þ
We can obtain the equation of state of DE yielding the same γ−1
results as the one-fluid model by taking
P ¼ Kργ ; ð73Þ
Pde ¼ Pðρm Þ; ϵde ¼ uðρm Þ ð66Þ
c2s ¼ Kγργ−1 : ð74Þ
or, equivalently,
We see that the equation of state (73) associated with the
Pde ¼ PðρÞ; ϵde ¼ uðρÞ: ð67Þ power-law potential (68) is that of a polytrope with
polytropic constant K and polytropic index γ ¼ 1 þ 1=n.
In other words, the equation of state Pde ðϵde Þ of DE in the The pressure is positive when K > 0 and negative when
two-fluid model corresponds to the relation PðuÞ in the K < 0. Negative pressures play an important role in
single fluid (or SF) model. An explicit example of cosmology. They are necessary to account for the early
the equivalence between the one- and two-fluid models inflation and the present accelerating expansion of the
for the homogeneous background is given in Secs. V F universe (DE). We note that the potential V given by
and VI F in connection to the (anti-) Chaplygin gas. We Eq. (70) is similar to the Tsallis free energy density
note that although the one- and two-fluid models are V ¼ −Ksγ , where the polytropic constant K plays the role
043502-9
1
of a generalized temperature and sγ ¼ − γ−1 ðργ − ρÞ is the in parametric form with parameter ρ. As we have recalled
Tsallis entropy density. in Sec. III, the rest mass density ρm of the SF mimics
For the power-law potential from Eq. (70) the equations DM and the internal energy density u of the SF mimics
of the problem [Eqs. (30), (32), (33), (36), (38) and (39)] DE [50,54,69].
become The foregoing equations determine the cosmological
evolution of a spatially homogeneous complex SF
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi described by the potential (68) in the fast oscillation regime
2 Kγ γ−1 Qm
ρ 1þ 2 ρ ¼ 3 ; ð75Þ for any values of K and γ. Some values of γ are of particular
c γ−1 a interest.
(i) For γ ¼ −1 (n ¼ −1=2), we obtain
γþ1 γ
ϵ ¼ ρc2 þ Kρ ; ð76Þ
γ−1 V¼−
K
; ð85Þ
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2ρ
Etot 2 Kγ γ−1
¼ 1þ 2 ρ ; ð77Þ K
mc 2
c γ−1 P¼ ; ð86Þ
ρ
3H2 γþ1K γ ϵ ¼ ρc2 ; ð87Þ
¼ρþ ρ; ð78Þ
8πG γ − 1 c2
leading to
K γ−1
c2
ρ
w¼ γþ1 K γ−1
; ð79Þ Kc2
1 þ γ−1 c2 ρ P¼ : ð88Þ
ϵ
Kγργ−1 c2 This is the equation of state of the Chaplygin
c2s ¼ : ð80Þ
c2 þ γþ1
γ−1 Kγρ
γ−1
(K < 0) or anti-Chaplygin (K > 0) gas [70–73].
The total SF potential is
From Eqs. (73) and (76), we obtain the equation of state
PðϵÞ of the SF under the inverse form ϵðPÞ as 1 K
V tot ¼ ρc2 − ð89Þ
1=γ 2 2ρ
P γþ1
ϵ¼ c2 þ P: ð81Þ or, equivalently,
K γ−1

1 m2 c2 2 K ℏ 2 1
On the other hand, the differential equation governing the V tot ðjφj2 Þ ¼ jφj − : ð90Þ
temporal evolution of the pseudorest-mass density [see 2 ℏ2 2 m jφj2
Eq. (42)] is
It corresponds to an inverse square law self-interaction
2 potential Vðjφj2 Þ ∼ jφj−2 . The rest-mass density and
c2 ρ_ ρc2 þ Kðγþ1Þ
γ−1 ρ
γ
¼ 2
: ð82Þ the internal energy density are given by
24πG ρ ½1 þ c2Kγρ
γ−1

þ2Kγργ−1 γ−1 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ρm c2 ¼ ðρc2 Þ2 þ Kc2 ; ð91Þ
The rest-mass density (DM) and the internal energy
density (DE) of the SF are determined by Eqs. (59) and (61) qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
with Eq. (70). We get ρc2 ¼ ðρm c2 Þ2 − Kc2 ; ð92Þ
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Kc2
2γ K γ−1 P ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ; ð93Þ
ρm ¼ ρ 1 þ ρ ; ð83Þ ðρm c2 Þ2 − Kc2
γ − 1 c2
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
γþ1 γ 2γ K γ−1 u ¼ ρc2 − ðρc2 Þ2 þ Kc2 ; ð94Þ
2 2
u ¼ ρc þ Kρ − ρc 1 þ ρ : ð84Þ
γ−1 γ − 1 c2 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u¼ ðρm c2 Þ2 − Kc2 − ρm c2 : ð95Þ
We note that the rest-mass density (83) can be read off
directly from Eq. (75) if we assume Eq. (64). In the present context, the Chaplygin gas model is
Equations (73), (83), and (84) define Pðρm Þ and uðρm Þ justified from a complex SF theory. Furthermore,
043502-10
the pseudorest-mass energy density coincides with rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

4K
the energy density (ϵ ¼ ρc2 ). Inversely, the con- 2 2
u ¼ ρc þ 3Kρ − ρc 1 þ 2 ρ: 2
ð105Þ
dition ϵ ¼ ρc2 in Eq. (33) implies Eq. (85). c
(ii) For γ ¼ 2 (n ¼ 1), we obtain With Eq. (97), they define Pðρm Þ and uðρm Þ in
parametric form. Actually, Eq. (104) is a cubic
V ¼ Kρ2 ; ð96Þ equation for ρ which can be solved by standard
means to get ρðρm Þ. We can then obtain Pðρm Þ and
P ¼ Kρ2 ; ð97Þ uðρm Þ explicitly.
(iii) For γ ¼ 0 (n ¼ −1), we obtain
ϵ ¼ ρc2 þ 3Kρ2 ; ð98Þ
V ¼ −K; ð106Þ
leading to
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi P ¼ K; ð107Þ
−c2 c4 þ 12Kϵ
ρ¼ ð99Þ
6K ϵ ¼ ρc2 − K: ð108Þ
and The pressure is constant. This is the equation of state

of the ΛCDM (K < 0) or anti-ΛCDM (K > 0)
1 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2
P¼ ð−c2 c4 þ 12KϵÞ : ð100Þ model interpreted as a UDM model [73,75,76]. In
36K that case K ¼∓ ρΛ c2 where ρΛ is the cosmological
density. The pressure can be rewritten as P ¼∓ ρΛ c2
We must select the sign þ when K > 0 to have ρ ≥ 0 and the energy density as ϵ ¼ ρc2 ρΛ c2 . In the
while the two signs are allowed when K < 0. present case, the pseudorest-mass density ρ plays the
Alternatively, we can write the equation of state
role of DM and ρΛ c2 the role of DE. The total SF
under the inverse form
potential is
rffiffiffiffi
ϵ¼
P 2
c þ 3P: ð101Þ 1
V tot ¼ ρc2 − K ð109Þ
K 2
The total SF potential is or, equivalently,
1 1 m2 c2 2
V tot ¼ ρc2 þ Kρ2 ð102Þ V tot ðjφj2 Þ ¼ jφj − K: ð110Þ
2 2 ℏ2
or, equivalently, It corresponds to a constant self-interaction potential
4 Vðjφj2 Þ ¼ −K ¼ ρΛ c2 equal to the cosmological
2 1 m2 c2 2 m energy density. The rest-mass density and the
V tot ðjφj Þ ¼ 2
jφj þ K jφj4 : ð103Þ
2 ℏ ℏ internal energy density are given by
It corresponds to a quartic self-interaction potential P ¼ K; ρm ¼ ρ; u ¼ −K ¼ V: ð111Þ

Vðjφj2 Þ ∼ jφj4 . This is the standard potential
of a relativistic BEC. It takes into account two- In the present context, the ΛCDM model is justified
body interactions in a weakly interacting micros- from a complex SF theory. Furthermore, the pseu-
copic theory of superfluidity. In that case, K ¼ dorest-mass density coincides with the rest-mass
2πas ℏ2 =m3 . The self-interaction is repulsive when density (ρm ¼ ρ).
K > 0 and attractive when K < 0. The relativistic (iv) For γ ¼ 3 (n ¼ 1=2), we obtain
BEC model has been studied in detail in the context
1
of boson stars [29,30,74] and in cosmology [40,41]. V ¼ Kρ3 ; ð112Þ
For ϵ → þ∞ we have P ∼ ϵ=3 (dark radiation) and 2
for ϵ → 0 we have P ∼ Kðϵ=c2 Þ2 (matter). The rest-
P ¼ Kρ3 ; ð113Þ
mass density and the internal energy density are
given by
ϵ ¼ ρc2 þ 2Kρ3 : ð114Þ
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
4K Equation (114) is a cubic equation for ρ which can
ρm ¼ ρ 1 þ 2 ρ; ð104Þ
c be solved by standard means to get ρðϵÞ. Using
043502-11
Eq. (113), we can then obtain the equation of state and

PðϵÞ explicitly. Alternatively, we can write the
equation of state under the inverse form 3K 2 K pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
P¼ 9K 2 þ 4c2 ϵ: ð126Þ
1=3 2c2 2c2
P
ϵ¼ c2 þ 2P: ð115Þ Alternatively, we can write the equation of state PðϵÞ
K
under the inverse form
The total SF potential is 2
P
ϵ¼ c2 − 3P: ð127Þ
1 1 K
V tot ¼ ρc2 þ Kρ3 ; ð116Þ
2 2
The total SF potential is
or, equivalently,
1 pffiffiffi
2 2
6 V tot ¼ ρc2 − 2K ρ; ð128Þ
2 1m c 2 1 m 2
V tot ðjφj Þ ¼ jφj þ K jφj6 : ð117Þ
2 ℏ2 2 ℏ
or, equivalently,
It corresponds to a sextic self-interaction potential
1 m2 c2 2 m
Vðjφj2 Þ ∼ jφj6 (see, e.g., [77]). It takes into account V tot ðjφj2 Þ ¼ jφj − 2K jφj: ð129Þ
three-body interactions in a weakly interacting 2 ℏ2 ℏ
microscopic theory of superfluidity. It can also It corresponds to a linear self-interaction potential
describe an exotic DM superfluid with a different
Vðjφj2 Þ ∼ jφj. The rest-mass density and the internal
interpretation [78]. The rest-mass density and the
energy density are given by
internal energy density are explicitly given by
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2K
3K ρm ¼ ρ2 − 2 ρ3=2 ; ð130Þ
ρm ¼ ρ 1 þ 2 ρ2 ; ð118Þ c
c
p ffiffiffi
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi1=2 u ¼ ρc2 − 3K ρ − ρm c2 : ð131Þ
c2 c2 12K
ρ¼ − 1 þ 2 ρ2m ; ð119Þ
6K 6K c With Eq. (123) they define Pðρm Þ and uðρm Þ in
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi3=2 parametric form but the results cannot be made more

c2 c2 12K explicit.
P¼K − 1 þ 2 ρ2m ; ð120Þ Remark: We emphasize that the previous equations are
6K 6K c
valid for a possibly inhomogeneous SF. We also note that
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi1=2 the polytropic indices γ ¼ 0 and γ ¼ −1 are the only ones
c2 c2 12K
u¼ − 1 þ 2 ρ2m for which the polytropic equation of state P ¼ Kργ yields a
6K 6K c polytropic equation of state P ¼ Kðϵ=c2 Þγ .
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 1 12K
× 1 þ 2 ρ2m c2 − ρm c2 : ð121Þ
3 3 c B. Isothermal equation of state
We now consider a potential of the form
(v) For γ ¼ 1=2 (n ¼ −2),14 we obtain
2 2
pffiffiffi kB Tm 2 m jφj
V ¼ −2K ρ; ð122Þ Vðjφj2 Þ ¼ jφj ln − 1 : ð132Þ
ℏ2 ρ ℏ2
pffiffiffi
P ¼ K ρ; ð123Þ For the sake of generality, we consider arbitrary values of T
pffiffiffi (positive and negative). The total potential including the
ϵ ¼ ρc2 − 3K ρ: ð124Þ rest-mass term is
2 2
Equation (124) can be reversed to give 2 1 m2 c2 2 kB Tm 2 m jφj
V tot ðjφj Þ ¼ jφj þ 2 jφj ln −1 :
2 2 ℏ2 ℏ ρ ℏ2
3K 1 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 2
ρ¼ 9K þ 4c ϵ ð125Þ ð133Þ
2c2 2c2
Introducing the pseudorest-mass density defined by
14
An interpretation of this index is given in Sec. VI C. Eq. (24), we get
043502-12

kB T ρ kB T 2
VðρÞ ¼ ρ ln −1 ð134Þ m c
m ρ c2s ¼ 2kB T ρ
: ð144Þ
c2 þ m lnðρ Þ þ km
BT
and
From Eqs. (137) and (140), we obtain the equation of state
PðϵÞ of the SF under the inverse form ϵðPÞ as
1 k T ρ
V tot ðρÞ ¼ ρc2 þ B ρ ln −1 : ð135Þ
2 m ρ
mc2 mP
ϵ¼ P þ 2P ln − P: ð145Þ
Using Eqs. (34), (37), and (134), we find that the derivative kB T ρ kB T
of the potential, the pressure and the pseudosquared speed
of sound are given by Finally, the differential equation governing the temporal
evolution of the pseudorest-mass density [see Eq. (42)] is

0 kB T ρ
V ðρÞ ¼ ln ; ð136Þ 2
m ρ c2 ρ_ ρc2 þ 2kmB T ρ lnðρρ Þ − km
BT
ρ
¼ 2
: ð146Þ
24πG ρ kB T
kB T ½1 þ 2kBmcT2 ρ
1þ lnðρ Þ
PðρÞ ¼ ρ ; ð137Þ mc2
m
The rest-mass density (DM) and the internal energy
k T density (DE) of the SF are determined by Eqs. (59) and (61)
c2s ¼ B : ð138Þ
m with Eq. (134). We get
We see that the equation of state (137) associated with the sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi
potential (132) is the isothermal equation of state with an 2kB T ρ
ρm ¼ ρ 1 þ 2
ln ; ð147Þ
effective temperature T. It corresponds to a polytrope mc ρ
of index γ ¼ 1 (n → þ∞). The pressure is positive when
T > 0 and negative when T < 0. Negative pressures play
2kB T
2 ρ k T
an important role in cosmology in relation to the early u ¼ ρc þ ρ ln − B ρ
m ρ m
inflation and the present accelerating expansion of the sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi
universe (DE). On the other hand, a positive temperature 2k T ρ
(T > 0) can account for (effective) thermal effects in DM − ρc2 1 þ B 2 ln : ð148Þ
mc ρ
[79–83]. We note that the potential V given by Eq. (134) is
similar to the Boltzmann free energy density V ¼ −TsB,
where T is the temperature and sB ¼ −kB ðρ=mÞ Equations (137), (147), and (148) determine Pðρm Þ and
½lnðρ=ρ Þ − 1 is the Boltzmann entropy density. uðρm Þ in parametric form with parameter ρ. As we have
For the potential (132) the equations of the problem recalled in Sec. III, the rest mass density ρm of the SF
[Eqs. (30), (32), (33), (36), (38) and (39)] become mimics DM and the internal energy density u of the SF
mimics DE [50,54,69].
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi The foregoing equations determine the cosmological
2kB T ρ Qm evolution of a spatially homogeneous complex SF
ρ 1þ 2
ln ¼ 3 ; ð139Þ
mc ρ a described by the potential (132) in the fast oscillation
regime for any value of T.

2kB T ρ k T Remark: We emphasize that the previous equations
ϵ ¼ ρc2 þ ρ ln − B ρ; ð140Þ (except those involving a) are valid for a possibly inho-
m ρ m
mogeneous SF. The results of this section (isothermal
ffi systems) can be recovered from the results of Sec. IVA
Etot 2kB T ρ (polytropes) in the limit γ → 1. For example,
¼ 1þ ln ; ð141Þ
mc2 mc 2 ρ
K γ Kρ ðγ−1Þ ln ρ
V¼ ρ ¼ e
γ−1 γ−1
3H2 2kB T ρ k T
¼ρþ 2
ρ ln − B 2 ρ; ð142Þ Kρ
8πG mc ρ mc ≃ ½1 þ ðγ − 1Þ ln ρ þ ¼ Kρ ln ρ þ cst: ð149Þ
γ−1
1
w ¼ mc2 ; ð143Þ This corresponds to the passage from the Tsallis to the
kB T þ 2 lnðρρ Þ − 1 Boltzmann free energy when γ → 1.
043502-13
V. THE CASE K > 0 8

10 K>0
α era
In this section, we consider the case of a positive γ>1
polytropic constant (K > 0), corresponding to a positive 10
4
pressure. Since P > 0, the universe is always decelerat-

ing. In the figures, we take c ¼ Qm ¼ 4πG ¼ 1 and 0 ρt
ρ
10
K ¼ 1.15
-4
10
Matter era
A. The case γ > 1
-8
For γ > 1 the equations determining the pseudorest-mass 10
density and the energy density as a function of the scale 1e-06 0.0001 0.01 1 100 10000
a
factor can be written as
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi FIG. 1. Evolution of the pseudorest-mass density as a function
2 Kγ γ−1 Qm of the scale factor for γ > 1 (specifically γ ¼ 2). The pseudorest-
ρ 1þ 2 ρ ¼ 3 ; ð150Þ mass density decreases more rapidly in the matterlike era than in
c γ−1 a the α era.
γþ1 γ
ϵ ¼ ρc2 þ Kρ : ð151Þ This corresponds to the nonrelativistic regime P=ϵ ≪ 1
γ−1
(matterlike era) valid for a → þ∞. The equation of state is
When ρ → þ∞, Eqs. (150) and (151) reduce to P ∼ Kðϵ=c2 Þγ . The pseudorest-mass density and the energy
density decrease to zero as a increases to þ∞.
2 2 2 When K > 0 and γ > 1, the universe evolves from an α
Q m c γ − 1 1=ð1þγÞ 1
ρ∼ ; ð152Þ era (P ∼ αϵ) in the early universe to a matterlike era (P ≃ 0)
2 Kγ a6=ð1þγÞ in the late universe. The case γ ¼ 2 and K > 0, corre-
sponding to a relativistic BEC with a repulsive jφj4 self-
γþ1 γ interaction, is treated in detail in Sec. III of [41]. In that
ϵ∼ Kρ ; ð153Þ
γ−1 case, the universe evolves from a dark radiation era to a
2 2 2 matterlike era. The curves ρðaÞ and ϵðaÞ are plotted in
γþ1 Q m c γ − 1 γ=ð1þγÞ 1 Figs. 1 and 2.
ϵ∼ K : ð154Þ
γ−1 2 Kγ a6γ=ð1þγÞ The temporal evolution aðtÞ of the scale factor is
represented in Fig. 3. It is obtained by integrating
This corresponds to the ultrarelativistic regime valid for Eq. (82) numerically. Starting from a singularity at t ¼ 0
a → 0. Starting from þ∞ when a → 0, the pseudorest- where a ¼ 0 and ϵ → þ∞ (big bang) the scale factor first
mass density and the energy density decrease as a grows as a ∝ tðγþ1Þ=3γ (corresponding to a ∝ t2=½3ð1þαÞ ) in
increases. Using Eqs. (73) and (153), we obtain the the α era then as a ∝ t2=3 in the matterlike era [Einstein-de
equation of state Sitter (EdS) solution].
γ−1
P∼ ϵ: ð155Þ
γþ1 18
10
The pressure is a linear function P ∼ αϵ of the energy K>0

α era
density with coefficient α ¼ ðγ − 1Þ=ðγ þ 1Þ. For γ ¼ 2, 10
12
γ>1
we recover the equation of state of the dark radiation
P ¼ ϵ=3 due to a complex SF with a repulsive jφj4 self- 10
6
interaction [40,41].
ε
When ρ → 0, Eqs. (150) and (151) reduce to

10
0
εt
Qm Matter era
ρ∼ ; ð156Þ
a3 -6
10
ϵ ∼ ρc2 : ð157Þ 1e-06 0.0001 0.01

a
1 100 10000
FIG. 2. Evolution of the energy density as a function of the

The case K ¼ 0 corresponds to pressureless DM with
15
scale factor for γ > 1 (specifically γ ¼ 2). The energy density
P ¼ V ¼ u ¼ 0 and ϵ ¼ ρc2 ¼ ρm c2 ¼ Ωm;0 ϵ0 =a3 . decreases more rapidly in the α era than in the matterlike era.
043502-14
1e+04
density and the energy density as a function of the scale
K>0
factor are
100 γ>1 Matter era
ffi
2kB T ρ Qm
1 at ρ 1þ 2
ln ¼ 3 ; ð160Þ
mc ρ a
a
0.01
α era
22kB T ρ k T
ϵ ¼ ρc þ ρ ln − B ρ: ð161Þ
0.0001 m ρ m
When ρ → þ∞, Eqs. (160) and (161) reduce to

1e-06
1e-08 0.0001 1
t rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2kB T Qm
ρ ln ρ ∼ 3 ; ð162Þ
FIG. 3. Evolution of the scale factor as a function of time for mc2 a
γ > 1 (specifically γ ¼ 2). The scale factor increases more
rapidly in the matterlike era than in the α era. The dashed line 2kB T
corresponds to the energy density. ϵ∼ ρ ln ρ: ð163Þ
m
The transition between the two regimes typically This regime is valid for a → 0. Starting from þ∞ when
occurs at a → 0, the pseudorest-mass density and the energy density
decrease as a increases. This corresponds to a matterlike era
1 (ϵ ≃ a−3 ) modified by logarithmic corrections. Using
1þγ 2 Kγ 3ðγ−1Þ
at ¼ ðQmÞ3ðγ−1Þ 2 2 2γ−1
; ð158Þ Eqs. (137) and (163), we obtain the equation of state
Qmc
2 1 ϵ
c γ − 1 γ−1 2 P≃ : ð164Þ
2
ϵt ¼ ρt c ¼ c: ð159Þ 2 ln ϵ
2 Kγ
The pressure is an approximately linear function of the
2
The total SF potential V tot ðjφj Þ is represented in Fig. 4. energy density P ≃ αϵ with a logarithmic correction yield-
It behaves as jφj2 for small jφj and as jφj2γ for large jφj. The ing a small effective coefficient α ∼ 1=ð2 ln ϵÞ ≪ 1. For
SF slowly descends the potential towards jφj ¼ 0 (at which a → 0, we have P=ϵ ≪ 1.
ϵ ¼ V tot ¼ 0) by rapidly spinning about the vertical axis. In Considering now small values of ρ, we see that Eq. (160)
this manner, the universe evolves from the α era to the imposes the condition ρ ≥ ρMin with
matterlike era.
mc2
ρMin ¼ ρ e−2kB T : ð165Þ
B. The case γ = 1
The pseudorest-mass density decreases as a increases and
The case γ ¼ 1 must be treated specifically. It corre- tends to ρMin when a → þ∞. However, we must be careful
sponds to the SF potential from Eq. (132) associated with that the energy density vanishes before ρ reaches its absolute
the isothermal equation of state (137). In the present case minimum value. Indeed, ϵMin ¼ −ρMin kB T=m < 0 when
T > 0. The equations determining the pseudorest-mass ρ ¼ ρMin . According to Eq. (161), the energy density
vanishes (ϵ ¼ 0) at
16
10 K>0 1 mc2
γ>1 ρmin ¼ ρ e2−2kB T : ð166Þ
2γ
~|φ|
16 This happens when the scale factor reaches the value
10
sffiffiffiffiffiffiffiffi
Vtot
1=3
Qm mc2 −12þ2kmc2T
amax ¼ e B : ð167Þ
10
15
ρ kB T
2
~|φ|
The physical solution of Eqs. (160) and (161) is defined only
0 for a ≤ amax. The pseudorest-mass density and the energy
-10000 -5000 0 5000 10000
density decrease as a increases. When a ¼ amax , the energy
density vanishes while the pseudorest-mass density reaches
FIG. 4. Total SF potential for γ > 1 (specifically γ ¼ 2). its minimum accessible value ρmin . This evolution is
043502-15
100
illustrated in the following section which presents similar
features. Matter era K>0
80 0<γ<1
C. The case 0 < γ < 1

60
For 0 < γ < 1 the equations determining the pseudorest-
ρ
mass density and the energy density as a function of the 40
scale factor can be written as
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 20
ρmin
2 Kγ 1 Qm ρMin
ρ 1− 2 1−γ
¼ 3 ; ð168Þ
c 1−γρ a 0
0 0.2 0.4 0.6 0.8 1
a
γþ1 γ
ϵ ¼ ρc2 − Kρ : ð169Þ FIG. 5. Evolution of the pseudorest-mass density as a function
1−γ of the scale factor for 0 < γ < 1 (specifically γ ¼ 0.5). The
region ρ < ρmin is unphysical.
Qm
ρ∼ ; ð170Þ The temporal evolution aðtÞ of the scale factor is
a3
ϵ ∼ ρc2 : ð171Þ Eq. (82) numerically. Starting from a singularity at t ¼ 0
where a ¼ 0 and ϵ → þ∞ (big bang), the scale factor first
This corresponds to the nonrelativistic regime P=ϵ ≪ 1 grows as a ∝ t2=3 in the matterlike era (EdS solution) until
(matterlike era) valid for a → 0. The equation of state is it reaches a maximum value amax at which the energy
P ∼ Kðϵ=c2 Þγ . Starting from þ∞ when a → 0, the pseu- density vanishes (ϵ ¼ 0). After that moment, the universe
dorest-mass density and the energy density decrease as a collapses and forms a singularity at tbc where a ¼ 0 and
increases. ϵ → þ∞ (big crunch). This process repeats itself periodi-
Considering now small values of ρ (ultrarelativistic cally in time. This solution describes a cyclic universe
regime), we see that Eq. (168) imposes the condition presenting phases of expansion and contraction separated
ρ > ρMin with by critical points where the energy density is either infinite
1=ð1−γÞ (when a ¼ 0) or zero (when a ¼ amax ). At the latter point
2 Kγ the universe “disappears.”
ρMin ¼ : ð172Þ The total SF potential V tot ðjφj2 Þ is represented in Fig. 8.
c2 1 − γ
It behaves as −jφj2γ for small jφj and as jφj2 for large jφj.
The pseudorest-mass density decreases as a increases and It has a minimum at ρ ¼ ρMin where V tot ¼ −ð1 − γÞ
tends to ρMin when a → þ∞. However, as in the previous ρMin c2 =2γ < 0. The region where V 0tot ðjφj2 Þ < 0 is forbid-
section, the energy density vanishes before ρ reaches its den [see Eq. (31)]. The SF slowly descends the potential
absolute minimum value. Indeed, ϵMin ¼ − 1−γ 2
2γ ρMin c < 0
when ρ ¼ ρMin . According to Eq. (169), the energy
density vanishes (ϵ ¼ 0) at 100

γ þ 1 K 1=ð1−γÞ K>0
ρmin ¼ : ð173Þ 80
Matter era 0<γ<1
1 − γ c2
60
This happens when the scale factor reaches the value
ε
40
1=3 1 − γ ðγþ1Þ=½6ð1−γÞ c2 1=½3ð1−γÞ
amax ¼ ðQmÞ : ð174Þ
γþ1 K 20
amax
The physical solution of Eqs. (168) and (169) is defined
only for a ≤ amax. The pseudorest-mass density and the 0
0 0.2 0.4 0.6 0.8 1
a
energy density decrease as a increases. When a ¼ amax ,
the energy density vanishes while the pseudorest-mass FIG. 6. Evolution of the energy density as a function of the
density reaches its minimum accessible value ρmin . The scale factor for 0 < γ < 1 (specifically γ ¼ 0.5). The energy
curves ρðaÞ and ϵðaÞ are plotted in Figs. 5 and 6. density vanishes at amax .
043502-16
fluid with an equation of state Pde ¼ −ϵde yielding a

amax K>0
0.6
0<γ<1 constant energy density ϵde ¼ −ρΛ c2 < 0. This agreement
0.5 is expected since, when γ ¼ 0, the pressure is constant
(P ¼ K) and we know that a constant positive pressure
0.4
P ¼ ρΛ c2 returns the anti-ΛCDM model [73]. In the
a
0.3 present context, the anti-ΛCDM model is obtained from

a complex SF theory with a constant negative potential
0.2
Vðjφj2 Þ ¼ −ρΛ c2 (anti-vacuum energy). For this particular
0.1 Matter
model, we see that the pseudorest-mass density decreases
era as ρ ∝ a−3 and behaves as DM (see the Remark below).
0
0 0.5 1 1.5 2 2.5 3 3.5 The anti-ΛCDM model has been studied in Sec. 6
t
of [73]. To make the connection with this study, we set
FIG. 7. Evolution of the scale factor as a function of time for K ¼ ρΛ c2 and a2 ¼ ðQm=ρΛ Þ1=3 . Equation (177) can then
0 < γ < 1 (specifically γ ¼ 0.5). The dashed lines correspond to be rewritten as
the energy density.
3
a2
ϵ ¼ ρΛ c2 −1 : ð178Þ
towards ρ ¼ ρmin (at which ϵ ¼ 0 and V tot ¼ −½ð1 − γÞ= a
ð1 þ γÞρmin c2 =2 < 0) by rapidly spinning about the verti-
cal axis. Starting from þ∞ when a → 0, the energy density
decreases as a increases and vanishes at
D. The case γ = 0 (anti-ΛCDM model)
Qmc2 1=3
For γ ¼ 0, the equations determining the pseudorest- amax ¼ ¼ a2 : ð179Þ
K
mass density and the energy density as a function of the
scale factor reduce to This is the maximum scale factor. At that point, the
pseudorest-mass density reaches its minimum accessible
Qm value
ρ¼ ; ð175Þ
a3
K
ϵ ¼ ρc2 − K: ð176Þ ρmin ¼ ¼ ρΛ : ð180Þ
c2
They can be combined to give In the nonrelativistic regime a ≪ a2 , corresponding to the
matterlike era where P=ϵ ≪ 1, we have
Qmc2
ϵ¼ − K: ð177Þ 3
a3 Qmc2 2 a2
ϵ∼ ∼ ρΛ c : ð181Þ
a3 a
This equation is equivalent to the one obtained in the anti-
ΛCDM model which assumes that the universe is filled The temporal evolution of the scale factor is obtained by
with pressureless DM (P ¼ 0) and that the cosmological solving the Friedmann equation (6) with Eq. (178). In that
constant is negative (Λ < 0) or that it is represented by a case, the solution can be obtained analytically yielding [73]

20
a pffiffiffiffiffiffi t
K>0 ¼ sin2=3 6π ; ð182Þ
0<γ<1
a2 tΛ
15
ρΛ c2
10 ϵ¼ 2
pffiffiffiffiffiffi ; ð183Þ
tan ð 6π ttΛ Þ
Vtot
5
ρΛ c2
ρ¼ pffiffiffiffiffiffi ; ð184Þ
0 sin2 ð 6π ttΛ Þ
ρmin
ρMin pffiffiffiffiffiffiffiffiffi
-5 where tΛ ¼ 1= GρΛ is the cosmological time. The evo-
-10 -5 0 5 10
|φ| lution of the universe is similar to the one described in
Sec. V C with the particularity that ρMin ¼ 0. The anti-
FIG. 8. Total SF potential for 0 < γ < 1 (specifically γ ¼ 0.5). ΛCDM model is studied in more detail in Sec. 6 of [73].
043502-17
ðγþ1Þ=½6ð1−γÞ 2 1=½3ð1−γÞ
Remark: In Eq. (177) the first term is the rest-mass energy 1−γ c
density ρm c2 [see Eqs. (55) and (63)] and the second term is amax ¼ ðQmÞ1=3 :
γþ1 K
the internal energy density u [see Eq. (56)]. As discussed in
Sec. III, the rest-mass density ρm can be interpreted as DM ð191Þ
and the internal energy density u can be interpreted as DE
[50,54,69]. In the present case, the pseudorest-mass density In that case, the evolution of the universe is similar to
coincides with the rest-mass density (ρ ¼ ρm ) and the the one described in Sec. V C with the particularity
that ρMin ¼ 0.
internal energy density is constant (u ¼ −K ¼ −ρΛ c2 ).
(ii) The case γ ¼ −1 (anti-Chaplygin gas) is specifically
It coincides with the SF potential (u ¼ V). In the two-fluid
treated in the next subsection.
model associated with the anti-ΛCDM model (see
(iii) When γ < −1, the asymptotic behavior from
Sec. III D), DM has an equation of state Pm ðϵm Þ ¼ 0 and
Eq. (189) is valid for small values of a. The
DE has an equation of state Pde ðϵde Þ ¼ −ϵde .
pseudorest-mass density decreases as a decreases
and tends to zero as ρ ∝ a6=j1þγj when a → 0. In
E. The case γ < 0 parallel, the energy density increases as a decreases
For γ < 0 the equations determining the pseudorest-mass and tends to þ∞ as
factor can be written as jγ þ 1j 1 1
ϵ∼ K jγj ∝ 6jγj=j1þγj ð192Þ
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1−γ ρ a
2 Kjγj 1 Qm
ρ 1þ 2 1−γ
¼ 3 ; ð185Þ when a → 0. Using Eqs. (73) and (192), we obtain
c 1−γρ a the equation of state
γþ1 1 1−γ
ϵ ¼ ρc2 − K : ð186Þ P¼ ϵ: ð193Þ
1 − γ ρjγj jγ þ 1j
When ρ → þ∞, Eqs. (185) and (186) reduce to The pressure is a linear function P ∼ αϵ of the
energy density with coefficient α ¼ ð1 − γÞ=
Qm jγ þ 1j. This determines an α-era. Strikingly, we
ρ∼ ; ð187Þ
a3 have two branches of solutions for sufficiently small
values of the scale factor: the nonrelativistic branch
ϵ ∼ ρc2 : ð188Þ from Eqs. (187) and (188) and the ultrarelativistic
branch described above. The two branches merge at
This corresponds to the nonrelativistic regime P=ϵ ≪ 1
a maximum scale factor
(matterlike era) valid for a → 0. The equation of state is
P ∼ Kðϵ=c2 Þγ . Starting from þ∞ when a → 0, the pseu-
jγ þ 1j jγþ1j=½6ð1−γÞ
dorest-mass density and the energy density decrease as a amax ¼ ðQmÞ1=3
1−γ
increases. 2 1=½3ð1−γÞ
When ρ → 0 (ultrarelativistic regime), Eqs. (185) and 1 c
× 1=½3ð1−γÞ ; ð194Þ
(186) reduce to jγj K
2 2 2
Q m c 1 − γ 1=ð1þγÞ 1 corresponding to
ρ∼ : ð189Þ
2 Kjγj a6=ð1þγÞ 1=ðγ−1Þ
21−γ 1
ρc ¼ c ð195Þ
We have to distinguish three subcases: Kjγj j1 þ γj
(i) When γ > −1, the asymptotic behavior from
Eq. (189) is valid for large values of a. The and
pseudorest-mass density decreases as a increases
and tends to zero as ρ ∝ a−6=ð1þγÞ when a → þ∞. 1−γ
ϵc ¼ ρ c2 : ð196Þ
However, the energy vanishes before. According to jγj c
Eq. (186), it vanishes at
The curves ρðaÞ and ϵðaÞ are plotted in Figs. 9 and 10.
γ þ 1 K 1=ð1−γÞ The temporal evolution aðtÞ of the scale factor is
ρmin ¼ : ð190Þ
1 − γ c2 represented in Fig. 11. It is obtained by integrating
Eq. (82) numerically, assuming that the SF follows the
This happens when the scale factor reaches the value curve of Fig. 9 from ρ ¼ þ∞ to ρ ¼ 0. Starting from a
043502-18
10
equals ϵc . This phase of expansion corresponds to the
K>0 nonrelativistic branch of Fig. 9. After that moment, the
8 Matter era γ < −1
universe collapses (α era) and forms a singularity at tbc
where a ¼ 0 and ϵ → þ∞ (big crunch). This phase of
6
contraction corresponds to the ultrarelativistic branch of
Fig. 9. The energy density increases as a−6jγj=j1þγj while the
ρ
scale factor decreases as a ∝ ðtbc − tÞjγþ1j=3jγj [correspond-

4
ing to a ∝ ðtbc − tÞ2=½3ð1þαÞ ] in the α era. This process

2
repeats itself periodically in time. This solution describes
α-era amax
an asymmetric cyclic universe presenting phases of expan-
0
0 0.2 0.4 0.6 0.8 1 sion (explosion) and contraction (implosion) separated by
a
critical points where the energy density is either infinite
FIG. 9. Evolution of the pseudo-rest mass density as a function (when a ¼ 0) or equal to ϵc (when a ¼ amax ).
of the scale factor for γ < −1 (specifically γ ¼ −2). The total SF potential V tot ðjφj2 Þ is represented in Fig. 12.
It behaves as −jφj−2jγj for small jφj and as jφj2 for large jφj.
50
On the nonrelativistic branch, the SF slowly descends the
potential towards ρ ¼ ρc (at which ϵ ¼ ϵc ) by rapidly
40
K>0 spinning about the vertical axis. On the ultrarelativistic
γ < -1 branch, the SF slowly descends the potential from ρ ¼ ρc to
jφj ¼ 0 (at which ϵ → þ∞ and V tot → −∞) by rapidly
30
spinning about the vertical axis. We find
α-era Matter era
ε
20
ðγ − 1Þðγ þ 2Þ
V tot ðρc Þ ¼ ρc c2 : ð197Þ
10 2γð1 þ γÞ
amax
0
0 0.2 0.4 0.6 0.8 1
a Therefore, V tot ðρc Þ ≤ 0 when −2 ≤ γ < −1 and V tot ðρc Þ ≥ 0
when γ ≤ −2.
FIG. 10. Evolution of the energy density as a function of the Remark: Other possible evolutions can be contemplated.
scale factor for γ < −1 (specifically γ ¼ −2).
The SF could follow the curve of Fig. 9 in the reverse sense,
from ρ ¼ 0 to ρ ¼ þ∞. This would lead to an asymmetric
singularity at t ¼ 0 where a ¼ 0 and ϵ → þ∞ (big bang), cyclic universe where the phase of expansion corresponds
the energy density decreases as a−3 and the scale factor to the ultrarelativistic branch (α era) and the phase of
grows as a ∝ t2=3 in the matter era (EdS solution) until it contraction corresponds to the nonrelativistic branch (mat-
reaches a maximum value amax at which the energy density terlike era). We could also consider the case of a symmetric
cyclic universe (like in Sec. V C) where the SF follows
1 the same branch (nonrelativistic or ultrarelativistic) during
amax K>0 the phases of expansion and contraction. In that case, there
γ < -1
0.8 would be two possible evolutions.
50
α-era
0.6
a
0.4 25
Matter era
Vtot
0.2 0 ρc
0
0 0.5 1 1.5 2
t -25 K>0
γ < −1
FIG. 11. Evolution of the scale factor as a function of time for
γ < −1 (specifically γ ¼ −2). The dashed lines correspond to the -50
-10 -5 0 5 10
energy density divided by 10. We note that a_ ≠ 0 at a ¼ amax |φ|
(contrary top ffiffiffi case of Fig. 7) because ϵc ≠ 0 at that point [recall
the
_ ∝ ϵ according to the Friedmann equation (6)].
that a=a FIG. 12. Total SF potential for γ < −1 (specifically γ ¼ −2).
043502-19
F. The case γ = − 1 (anti-Chaplygin gas) The integral can be calculated explicitly and is given by
For γ ¼ −1, the equations determining the pseudorest- Eq. (30) of [73]. The evolution of the universe is similar to
mass density and the energy density as a function of the the one described in Sec. V C with the particularity that ρ
scale factor reduce to vanishes at amax (i.e. ρmin ¼ 0). The anti-Chaplygin gas
model is studied in more detail in Sec. 4.3 of [73].
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Remark: We can rewrite Eq. (200) as
Q2 m2 K
ρ¼ − 2; ð198Þ "sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi #
a6 c
Qmc2 Q2 m2 c4 Qmc 2
ϵ¼ þ − Kc2 − ; ð205Þ
ϵ ¼ ρc2 : ð199Þ a3 a6 a3
They can be combined to give where the first term is the rest-mass energy density ρm c2
[see Eqs. (55) and (63)] and the second term is the internal
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
energy density u [see Eqs. (56) and (61)]. As discussed in
Q2 m2 c4
ϵ¼ − Kc2 : ð200Þ Sec. III, the rest-mass density ρm can be interpreted as DM
a6 and the internal energy density u can be interpreted as DE
[50,54,69]. In the two-fluid model associated with the anti-
This equation is equivalent to the one obtained in the anti- Chaplygin gas (see Sec. III D), DM has an equation of state
Chaplygin gas model. This agreement is expected because, Pm ðϵm Þ ¼ 0 and DE has an equation of state
when γ ¼ −1, we have ϵ ¼ ρc2 . Therefore, the equation of
state P ¼ K=ρ from Eq. (73) can be written as P ¼ Kc2 =ϵ 2Kc2 ϵde
which is the equation of state of the anti-Chaplygin gas Pde ðϵde Þ ¼ ; ð206Þ
ϵ2de − Kc2
when K > 0 [72,73]. In the present context, the anti-
Chaplygin gas model is obtained from a complex SF which is obtained by eliminating ρ between Eqs. (86) and
theory with a potential Vðjφj2 Þ ¼ − 12 Kðmℏ Þ2 jφj1 2 . For this (94), and by identifying PðuÞ with Pde ðϵde Þ. Solving the
particular model, the pseudorest-mass density coincides energy conservation equation (5) with the equation of state
with the energy density (ρ ¼ ϵ=c2 ). (206), we recover the expression of DE (second term)
The anti-Chaplygin gas has been studied in Sec. 4.3 of from Eq. (205).
[73]. Toffi make the connection with this study, we set ρ ¼
pffiffiffiffiffiffiffiffiffiffi
K=c2 and a ¼ ðQ2 m2 c2 =KÞ1=6 . Equation (200) can VI. THE CASE K < 0
then be rewritten as
In this section, we consider the case of a negative
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
6 ffi polytropic constant (K < 0), corresponding to a negative
a
ϵ ¼ ρ c2 − 1: ð201Þ pressure. This allows the universe to experience a phase of
a accelerating expansion. In the figures, we take c ¼ Qm ¼
4πG ¼ 1 and K ¼ −1.
Starting from þ∞ when a → 0, the energy density
decreases as a increases and vanishes at
A. The case γ > 1
2 2 2 1=6
Qmc For γ > 1 the equations determining the pseudorest-mass
amax ¼ ¼ a : ð202Þ density and the energy density as a function of the scale
K
This is the maximum scale factor. In the nonrelativistic sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
regime a ≪ a , corresponding to the matterlike era where 2 jKjγ γ−1 Qm
P=ϵ ≪ 1, we have ρ 1− 2 ρ ¼ 3 ; ð207Þ
c γ−1 a
3
Qmc2
2 2 a γþ1
ϵ ¼ ρc ∼ ¼ ρ c : ð203Þ ϵ ¼ ρc2 − jKjργ : ð208Þ
a3 a γ−1
The temporal evolution of the scale factor is obtained by When ρ → 0, Eqs. (207) and (208) reduce to
solving the Friedmann equation (6) with Eq. (201). This
yields [73] Qm
ρ∼ ; ð209Þ
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Z þ∞ a3
dx
96πGρ t ¼ : ð204Þ
a 6 xðx − 1Þ1=4 ϵ ∼ ρc2 : ð210Þ
ðaÞ
043502-20
This corresponds to the nonrelativistic regime P=ϵ ≪ 1 Dark energy era ρmax
(matterlike era) valid for a → þ∞. The equation of state is
P ∼ Kðϵ=c2 Þγ . The pseudorest-mass density and the energy
density decrease to zero as a increases. 0.2
Considering now large values of ρ (ultrarelativistic amin
ρ
regime), we see that Eq. (207) imposes the condition
ρ ≤ ρmax with 0.1
2
c γ − 1 1=ðγ−1Þ K<0
ρmax ¼ : ð211Þ γ>1 Matter era
2 jKjγ
0
0 1 2 3 4 5
The pseudorest-mass density increases as a increases and a
tends to ρmax when a → þ∞. In parallel, the energy density
FIG. 13. Evolution of the pseudorest-mass density as a function
decreases as a increases and tends to a constant of the scale factor for γ > 1 (specifically γ ¼ 2).
γ−1
ϵmin ¼ ρ c2 ð212Þ
2γ max universe expands as it enters into the matterlike era. This
phase of expansion corresponds to the nonrelativistic branch
when a → þ∞. This leads to a DE era corresponding to a of Fig. 13. The energy density decreases as ϵ ∝ a−3 while
de Sitter evolution. the scale factor increases as a ∝ t2=3 (EdS). This solution
Therefore, we have two branches of solutions for describes an asymmetric bouncing universe presenting a
sufficiently large values of the scale factor: the nonrelativ-
istic branch from Eqs. (209) and (210) and the ultra-
0.1
relativistic branch described above. The two branches
K<0
merge at a minimum scale factor γ>1
amin
0.08

1=3 γ þ 1 ðγþ1Þ=½6ðγ−1Þ Dark energy era
amin ¼ ðQmÞ
γ−1 0.06 εmin
1=½3ðγ−1Þ
ε
1=½3ðγ−1Þ jKj
×γ ; ð213Þ 0.04
c2
Matter era
0.02
corresponding to
1=ðγ−1Þ
2γ−1 1
0
0 1 2 3 4 5
ρi ¼ c ð214Þ a
jKjγ 1 þ γ
and scale factor for γ > 1 (specifically γ ¼ 2).
γ−1 2
ϵi ¼ ρi c : ð215Þ
γ
K<0
The case γ ¼ 2 and K < 0, corresponding to a relativistic 8
γ>1
BEC with an attractive jφj4 self-interaction, is treated in
detail in Sec. IV of [41]. The curves ρðaÞ and ϵðaÞ are 6 Matter era
plotted in Figs. 13 and 14. Dark energy era
a
The temporal evolution aðtÞ of the scale factor is

4
Eq. (82) numerically, assuming that the SF follows the
curve of Fig. 13 from ρmax to ρ ¼ 0. Starting from a DE era 2
amin
at t ¼ −∞ where a ¼ þ∞ and ϵ ¼ ϵmin , the energy density
slowly increases while the scale factor decreases exponen- 0
-5 -2.5 0 2.5 5 7.5 10
2 1=2 t
tially rapidly as a ∝ e−ð8πGϵmin =3c Þ t (de Sitter) until it
reaches a minimum value amin at which the energy density FIG. 15. Evolution of the scale factor as a function of time for
equals ϵi . This phase of contraction corresponds to the γ > 1 (specifically γ ¼ 2). The dashed line corresponds to the
ultrarelativistic branch of Fig. 13. After that moment, the energy density multiplied by 100.
043502-21
0.1 sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi
K<0 2kB jTj ρ Qm
γ>1
ρ 1− 2
ln ¼ 3 ; ð217Þ
ρmax
mc ρ a
0.1 ρi
2kB jTj
2 ρ k jTj
ϵ ¼ ρc − ρ ln þ B ρ: ð218Þ
ρ
Vtot
m m
0.0 When ρ → 0, Eqs. (217) and (218) reduce to
2kB jTj Qm
ρ 2
j ln ρj ∼ 3 ; ð219Þ
-0.1
-1.0 -0.5 0.0 0.5 1.0
mc a
|φ|
2kB jTj
ϵ∼ ρj ln ρj: ð220Þ
FIG. 16. Total SF potential for γ > 1 (specifically γ ¼ 2). m
This regime is valid for a → þ∞. The pseudorest-mass
phase of contraction and a phase of expansion. There is no density and the energy density decrease to zero as a
big bang singularity in this model. increases. This corresponds to a matterlike era (ϵ ≃ a−3 )
The total SF potential V tot ðjφj2 Þ is represented in Fig. 16. modified by logarithmic corrections. Using Eqs. (137) and
It behaves as jφj2 for small jφj and as −jφj2γ for large jφj. (220), we obtain the equation of state
It has a maximum at ρ ¼ ρmax where V tot ¼ ðγ − 1Þρmax c2 =
2γ > 0. The region where V 0tot ðjφj2 Þ < 0 is forbidden [see ϵ
P≃− : ð221Þ
Eq. (31)]. On the nonrelativistic branch, the SF slowly 2j ln ϵj
descends the potential from ρ ¼ ρi (where ϵ ¼ ϵi ) towards
jφj ¼ 0 (at which ϵ ¼ V tot ¼ 0) by rapidly spinning about The pressure is an approximately linear function of the
the vertical axis. On the ultrarelativistic branch, the SF slowly energy density P ≃ αϵ with a logarithmic correction yield-
descends the potential from ρ ¼ ρmax (at which ϵ ¼ ϵmin ) ing a small effective coefficient α ∼ −1=ð2j ln ϵjÞ ≪ 1. For
towards ρ ¼ ρi (where ϵ ¼ ϵi ) by rapidly spinning about the a → þ∞, we have jPj=ϵ ≪ 1.
vertical axis. We find Considering now large values of ρ, we see that Eq. (217)
imposes the condition ρ ≤ ρmax with
ðγ − 1Þðγ þ 2Þ 2
V tot ðρi Þ ¼ ρi c > 0: ð216Þ mc2
2γð1 þ γÞ ρmax ¼ ρ e2kB jTj : ð222Þ
The pseudorest-mass density increases as a increases and

Remark: Other possible evolutions can be contemplated. tends to ρmax when a → þ∞. In parallel, the energy density
The SF could follow the curve of Fig. 13 in the reverse decreases as a increases and tends to a constant
sense, from ρ ¼ 0 to ρmax . This would lead to an asym-
metric bouncing universe where the phase of contraction kB jTj
corresponds to the nonrelativistic branch (matterlike era) ϵmin ¼ ρmax ð223Þ
m
and the phase of expansion corresponds to the ultrarela-
tivistic branch (DE era). We could also consider the case of when a → þ∞.
a purely expanding universe where the SF follows the same Therefore, we have two branches of solutions for
branch (nonrelativistic or ultrarelativistic). In that case, the sufficiently large values of the scale factor: the nonrela-
SF suddenly appears at a finite scale factor ai and behaves tivistic branch from Eqs. (219) and (220) and the ultra-
either as DM (normal branch) or as DE (peculiar branch). relativistic branch described above. The two branches
This corresponds to the two possible evolutions (DM and merge at a minimum scale factor
DE) described in [41].
sffiffiffiffiffiffiffiffiffiffiffi 1=3
Qm mc2 12−2kmcjTj 2
amin ¼ e B ; ð224Þ
B. The case γ = 1 ρ kB jTj
The case γ ¼ 1 must be treated specifically. It corre-
sponds to the SF potential from Eq. (132) associated with corresponding to
the isothermal equation of state (137). In the present case 2
−12þ2kmcjTj
T < 0. The equations determining the pseudorest-mass ρi ¼ ρ e B ð225Þ
factor are and
043502-22
2kB jTj
ϵi ¼ ρi : ð226Þ
8
10 Matter era K<0
m 0<γ<1
4
This situation is similar to the case described previously. 10
ρt
0
C. The case 0 < γ < 1
ρ
10
For 0 < γ < 1 the equations determining the pseudorest- -4

10
mass density and the energy density as a function of the α era
scale factor can be written as -8
10
2 jKjγ 1 Qm 0.0001 0.01
a
1 100
ρ 1þ 2 1−γ
¼ 3 ; ð227Þ
c 1−γρ a
FIG. 17. Evolution of the pseudorest-mass density as a function
γþ1 of the scale factor for 0 < γ < 1 (specifically γ ¼ 0.5). The
ϵ ¼ ρc2 þ jKjργ : ð228Þ pseudorest-mass density decreases more rapidly in the α era than
1−γ in the matterlike era.
Qm
ρ∼ ; ð229Þ 9
a3 10
K<0
Matter era 0<γ<1
2
ϵ ∼ ρc : ð230Þ 10
6
This corresponds to the nonrelativistic regime P=ϵ ≪ 1 ε

10
3

P ∼ Kðϵ=c2 Þγ . Starting from þ∞ when a → 0, the pseu- 0
εt
10
dorest-mass density and the energy density decrease as a α era
increases.
When ρ → 0, Eqs. (227) and (228) reduce to 10
-3
1=ð1þγÞ 0.0001 0.01

a
1 100
Q2 m2 c2 1 − γ 1
ρ∼ ; ð231Þ
2 jKjγ a6=ð1þγÞ FIG. 18. Evolution of the energy density as a function of the
scale factor for 0 < γ < 1 (specifically γ ¼ 0.5). The energy
γþ1 density decreases more rapidly in the matterlike era than in
ϵ∼jKjργ ; ð232Þ the α era.
1−γ
2 2 2
γþ1 Q m c 1 − γ γ=ð1þγÞ 1
ϵ∼ jKj : ð233Þ
1−γ 2 jKjγ a6γ=ð1þγÞ
This corresponds to the ultrarelativistic regime valid for K<0

100 α era
a → þ∞. The pseudorest-mass density and the energy 0<γ<1
density decrease with a and tend to zero when a → þ∞.

Using Eqs. (73) and (232) we obtain the equation of state 1
at
a
1−γ
P¼− ϵ: ð234Þ
γþ1 0.01
Matter era
The pressure is a linear function P ∼ αϵ of the energy

density with coefficient α ¼ −ð1 − γÞ=ðγ þ 1Þ.
When K < 0 and 0 < γ < 1, the universe evolves from a 0.0001
0.0001 0.01 1 100
t
matterlike era (P ≃ 0) in the early universe to an α era
(P ∼ αϵ) in the late universe. The curves ρðaÞ and ϵðaÞ are FIG. 19. Evolution of the scale factor as a function of time for
plotted in Figs. 17 and 18. 0 < γ < 1 (specifically γ ¼ 0.5). The scale factor increases more
The temporal evolution aðtÞ of the scale factor is rapidly in the α era than in the matterlike era. The dashed line
represented in Fig. 19. It is obtained by integrating corresponds to the energy density.
043502-23
Eq. (82) numerically. Starting from a singularity at t ¼ 0 This equation is equivalent to the one obtained in the
where a ¼ 0 and ϵ → þ∞ (big bang) the scale factor first ΛCDM model which assumes that the universe is filled
grows as a ∝ t2=3 in the matterlike era (EdS solution) then with pressureless DM (P ¼ 0) and that the cosmological
as a ∝ tðγþ1Þ=3γ (corresponding to a ∝ t2=½3ð1þαÞ ) in the constant is positive (Λ > 0) or that it is represented by a
α era. fluid with an equation of state Pde ¼ −ϵde yielding a
The transition between the two regimes typically occurs at constant energy density ϵde ¼ ρΛ c2 > 0. This agreement
is expected since, when γ ¼ 0, the pressure is constant
2 2 2 1
1 Q m c 1 − γ 3ð1−γÞ (P ¼ K) and we know that a constant negative pressure
at ¼ ; ð235Þ P ¼ −ρΛ c2 returns the ΛCDM model [73,75,76]. In the
1þγ
ðQmÞ3ð1−γÞ 2 jKjγ
present context, the ΛCDM model is obtained from a
1 complex SF theory with a constant positive potential
2 jKjγ 1−γ
ϵt ¼ ρt c2 ¼ c2 : ð236Þ Vðjφj2 Þ ¼ ρΛ c2 (vacuum energy). For this particular
c2 1 − γ model, we see that the pseudorest-mass density decreases
as ρ ∝ a−3 and behaves as DM (see the Remark below).
The expansion of the universe is decelerating (α > −1=3)
The ΛCDM model has been studied in Sec. 5 of [73]. To
in the α era when 1=2 < γ < 1 and accelerating (α < −1=3)
make the connection with this study, we set K ¼ −ρΛ c2
when 0 < γ < 1=2. This provides a physical interpretation
of the particular index γ ¼ 1=2 considered in Sec. IVA. and a2 ¼ ðQm=ρΛ Þ1=3 . Equation (239) can then be rewrit-
The total SF potential V tot ðjφj2 Þ is represented in Fig. 20. ten as
It behaves as jφj2γ for small jφj and as jφj2 for large jφj. The 3
SF slowly descends the potential towards jφj ¼ 0 (at which a2
ϵ ¼ ρΛ c2 þ1 : ð240Þ
ϵ ¼ V tot ¼ 0) by rapidly spinning about the vertical axis. a
In this manner, the universe evolves from the matterlike era
to the α era. Starting from þ∞ when a → 0, the energy density
decreases as a increases and tends to a constant value
K ¼ ρΛ c2 when a → þ∞. The temporal evolution of
D. The case γ = 0 (ΛCDM model)
the scale factor is obtained by solving the Friedmann
For γ ¼ 0, the equations determining the pseudorest- equation (6) with Eq. (240). In that case, the solution
mass density and the energy density as a function of the can be obtained analytically yielding [73]
scale factor reduce to

a pffiffiffiffiffiffi t
Qm ¼ sinh2=3 6π ; ð241Þ
ρ¼ 3 ; ð237Þ a2 tΛ
a
ϵ ¼ ρc2 þ jKj: ð238Þ ρΛ c2

ϵ¼ pffiffiffiffiffiffi ; ð242Þ
tanh2 ð 6π ttΛ Þ
They can be combined to give
ρΛ
Qmc2 ρ¼ pffiffiffiffiffiffi ;
2
ð243Þ
ϵ¼ þ jKj: ð239Þ sinh ð 6π ttΛ Þ
a3
pffiffiffiffiffiffiffiffiffi
8 where tΛ ¼ 1= GρΛ is the cosmological time. The evo-
10
K<0 lution of the universe is similar to the one described in
0<γ<1 Sec. VI E below with the particularity that ρ → 0 for
2
~|φ| a → þ∞. The ΛCDM model is studied in more detail in
10
8
Sec. 5 of [73].
Remark: In Eq. (239) the first term is the rest-mass energy
Vtot
density ρm c2 [see Eqs. (55) and (63)] and the second term is
7
the internal energy density u [see Eq. (56)]. As discussed in
10
Sec. III, the rest-mass density ρm can be interpreted as DM
2γ
~|φ| and the internal energy density u can be interpreted as DE
[50,54,69]. In the present case, the pseudorest-mass density
0 4 4 0 4 4
coincides with the rest-mass density (ρ ¼ ρm ) and the
internal energy density is constant (u ¼ jKj ¼ ρΛ c2 ). It
-10 -10 10 10
|φ|
coincides with the SF potential (u ¼ V). In the two-fluid
FIG. 20. Total SF potential for 0 < γ < 1 (specifically γ ¼ 0.5). model associated with the ΛCDM model (see Sec. III D),
043502-24
10
DM has an equation of state Pm ðϵm Þ ¼ 0 and DE has an
equation of state Pde ðϵde Þ ¼ −ϵde . K<0
8 γ<0
Matter era
E. The case γ < 0 6

For γ < 0 the equations determining the pseudorest-mass
ρ
density and the energy density as a function of the scale 4
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 Dark energy era
2 jKjjγj 1 Qm ρmin
ρ 1− 2 ¼ 3 ; ð244Þ
c 1 − γ ρ1−γ a 0
0 2 4 6 8 10
a
γþ1 1 FIG. 21. Evolution of the pseudorest-mass density as a function

ϵ ¼ ρc2 þ jKj jγj : ð245Þ of the scale factor for γ < 0 (specifically γ ¼ −2).
1−γ ρ

10
Qm
ρ∼ ; ð246Þ Matter era K<0
a3 8
γ<0
ϵ ∼ ρc2 : ð247Þ 6
ε
This corresponds to the nonrelativistic regime jPj=ϵ ≪ 1 4
P ∼ Kðϵ=c2 Þγ . Starting from þ∞ when a → 0, the pseu- 2
Dark energy era
dorest-mass density and the energy density decrease as a
εmin
increases. 0
0 2 4 6 8 10
Considering now small values of ρ (ultrarelativistic a
regime) we see that Eq. (244) imposes the condition
ρ ≥ ρmin with
scale factor for γ < 0 (specifically γ ¼ −2).
1=ð1−γÞ
2 jKjjγj 1
ρmin ¼ : ð248Þ
c2 1 − γ 2jγjQm 1=3 c2 1 − γ 3ð1−γÞ
at ¼ : ð250Þ
1−γ 2 jKjjγj
The pseudorest-mass density decreases as a increases and
tends to ρmin when a → þ∞. Similarly, the energy density
decreases as a increases and tends to a constant The total SF potential V tot ðjφj2 Þ is represented in Fig. 24.
It behaves as jφj−2jγj for small jφj and as jφj2 for large jφj.
1−γ
ϵmin ¼ ρ c2 ð249Þ
2jγj min 3
K<0
Dark energy era
γ<0
when a → þ∞. We see that the evolution of the universe in
this regime is similar to the one induced by a cosmological 2
constant or by a constant energy density. This corresponds
to the DE era. The curves ρðaÞ and ϵðaÞ are plotted in
a
Figs. 21 and 22.

1 Matter era
The temporal evolution aðtÞ of the scale factor is
Eq. (82) numerically. Starting from a singularity at t ¼ 0
where a ¼ 0 and ϵ → þ∞ (big bang) the scale factor first 0
0 0.5 1 1.5 2
grows as a ∝ t2=3 in the matterlike era (EdS) then as a ∝ t
2 1=2
eð8πGϵmin =3c Þ t in the DE era (de Sitter). FIG. 23. Evolution of the scale factor as a function of time for
The transition between the two regimes typically occurs γ < 0 (specifically γ ¼ −2). The dashed line corresponds to the
when Qmc2 =a3 ∼ ϵmin yielding energy density.
043502-25
50 sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
6
K<0 a
40 γ<0 ϵ ¼ ρ c2 þ 1: ð254Þ
a
30
In the nonrelativistic regime where ρ; ϵ → þ∞, we have
Vtot
20
3
Qmc2 2 a
ϵ ¼ ρc2 ∼
10
¼ ρ c : ð255Þ
a3 a
0 ρmin
-10 This corresponds to the matterlike era (a ≪ a ) where

-10 0 10
-5
|φ|
5
jPj=ϵ ≪ 1. When a → þ∞, the energy density tends to a
constant
FIG. 24. Total SF potential for γ < 0 (specifically γ ¼ −2).
qffiffiffiffiffiffiffiffiffiffiffi
ϵmin ¼ jKjc2 ¼ ρ c2 : ð256Þ
It has a minimum at ρ ¼ ρmin where V tot ¼ ðγ − 1Þρmin c2 =
2γ > 0. The region where V 0tot ðjφj2 Þ < 0 is forbidden [see
Eq. (31)]. The SF slowly descends the potential towards This corresponds to the DE era (a ≫ a ). The Chaplygin
ρ ¼ ρmin [at which ϵ ¼ ϵmin and V tot ¼ ðγ − 1Þρmin c2 = equation of state can mimic the effect of a cosmological
2γ > 0] by rapidly spinning about the vertical axis. constant at late times. The temporal evolution of the scale
factor is obtained by solving the Friedmann equation (6)
with Eq. (254). This yields [73]
F. The case γ = − 1 (Chaplygin gas)
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Z
For γ ¼ −1, the equations determining the pseudorest- þ∞ dx
mass density and the energy density as a function of the 96πGρ t ¼ : ð257Þ
ðaa Þ6 xðx þ 1Þ1=4
scale factor reduce to
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi The integral can be calculated explicitly and is given by
Q2 m2 jKj Eq. (38) of [73]. The evolution of the universe has been
ρ¼ þ 2; ð251Þ
a6 c described in the previous section. The Chaplygin gas model
is studied in more detail in Sec. 4.4 of [73].
Remark: We can rewrite Eq. (253) as
ϵ ¼ ρc2 : ð252Þ
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2

Qmc Q2 m2 c4 2 Qmc2
They can be combined to give ϵ¼ þ þ jKjc − ; ð258Þ
a3 a6 a3
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Q2 m2 c4 where the first term is the rest-mass energy density ρm c2
ϵ¼ þ jKjc2 : ð253Þ
a6 [see Eqs. (55) and (63)] and the second term is the internal
energy density u [see Eqs. (56) and (61)]. As discussed in
This equation is equivalent to the one obtained in the Sec. III, the rest-mass density ρm can be interpreted as DM
Chaplygin gas model. This agreement is expected because, and the internal energy density u can be interpreted as DE
when γ ¼ −1, we have ϵ ¼ ρc2 . Therefore, the equation of [50,54,69]. In the two-fluid model associated with the
state P ¼ K=ρ from Eq. (73) can be written as P ¼ Kc2 =ϵ Chaplygin gas (see Sec. III D), DM has an equation of state
which is the equation of state of the Chaplygin gas when Pm ðϵm Þ ¼ 0 and DE has an equation of state
K < 0 [70–73]. In the present context, the Chaplygin gas
model is obtained from a complex SF theory with a −2jKjc2 ϵde
potential Vðjφj2 Þ ¼ − 12 Kðmℏ Þ2 jφj1 2 . For this particular model, Pde ðϵde Þ ¼ ; ð259Þ
ϵ2de þ jKjc2
the pseudorest-mass density coincides with the energy
density (ρ ¼ ϵ=c2 ). which is obtained by eliminating ρ between Eqs. (86) and
The Chaplygin gas has been studied in Sec. 4.4 of [73]. (94), and by identifying PðuÞ with Pde ðϵde Þ. Solving the
To make the connection with this study, we set ρ ¼
pffiffiffiffiffiffiffiffiffiffiffiffiffi energy conservation equation (5) with the equation of state
jKj=c2 and a ¼ ðQ2 m2 c2 =jKjÞ1=6 . Equation (253) can (259), we recover the expression of DE (second term)
then be rewritten as in Eq. (258).
043502-26
VII. VALIDITY OF THE FAST A. Standard BEC model

OSCILLATION REGIME For a standard BEC (γ ¼ 2 and K ¼ 2πas ℏ2 =m3 ), the
The previous results are valid in the fast oscillation fast oscillation regime is valid provided that
regime (ω ≫ H) of the complex SF (spintessence). This is
equivalent to making the TF approximation (ℏ → 0). In this 1 3ρ̃ 3jas jc2
ρ̃ ≪σ≡ : ð265Þ
section we determine the domain of validity of this 1 4ρ̃ 4Gm
approximation and apply it to specific models of physical
interest (see Refs. [64,65] for a detailed study of the validity This condition has been studied in detail in Secs. III.E and
of the fast oscillation regime for models based on a real IV.E of [41] (see also Sec. VII D below).
self-interacting SF).
The fast oscillation regime is valid provided that ω ≫ H, B. ΛFDM model
where ω ¼ −θ_ ¼ −S_ tot =ℏ ¼ Etot =ℏ is the pulsation of the For the (anti-) ΛFDM model (γ ¼ 0 and K ¼ ρΛ c2 ),16
SF and H ¼ a=a _ is the Hubble constant which is related to the fast oscillation regime is valid provided that
the energy density ϵ by the Friedmann equation
2
H2 ¼ ð8πG=3c2 Þϵ. Therefore, the fast oscillation regime 3m2 c4 m
is valid provided that ρ̃ ∓ 1 ≪ σ ≡ ¼ ; ð267Þ
8πGℏ2 ρΛ mΛ
H2 8πGℏ2 ϵ where Λ ¼ 8πGρΛ is the cosmological constant and

¼ ≪ 1: ð260Þ
ω2 3c2 E2tot rffiffiffiffi
ℏ 8πGρΛ 1=2 ℏ Λ
mΛ ¼ 2 ¼ 2 ¼ 1.20 × 10−33 eV=c2
Using Eqs. (30) and (33), this condition can be rewritten as c 3 c 3
ð268Þ
8πGℏ2 ρc2 þ VðρÞ þ ρV 0 ðρÞ
≪ 1: ð261Þ
3m2 c6 1 þ c22 V 0 ðρÞ is the cosmon mass [52]. Using the results of Secs. V D and
VI D, we find that ρ̃ ¼ ρ=ρΛ ¼ ða2 =aÞ3 . As a result, the
For the power-law potential (68) we obtain condition from Eq. (267) can be rewritten as
3
2 γþ1 γ a2
8πGℏ2 ρc þ γ−1 Kρ ∓ 1 ≪ σ: ð269Þ
Kγ γ−1 ≪ 1: ð262Þ a
3m2 c6 1 þ c22 γ−1 ρ
For the anti-ΛFDM model (K > 0) the fast oscillation
At this stage, it is convenient to introduce the dimensionless regime is valid provided that σ ≫ 1 (i.e. m ≫ mΛ )
density ρ̃ ¼ ðjKj=c2 Þ1=ðγ−1Þ ρ in terms of which the fore- and a2 =σ 1=3 ≪ a ≤ a2 .
going equation becomes For the ΛFDM model (K < 0) the fast oscillation regime
is valid provided that σ ≫ 1 (i.e. m ≫ mΛ ) and
1 a ≫ a2 =σ 1=3 , where a2 ¼ ½Ωm;0 =ð1 − Ωm;0 Þ1=3 ¼ 0.765
1 γþ1
γ−1 ρ̃
γ−1
3m2 c4 jKj γ−1
ρ̃ 2γ γ−1
≪σ≡ ; ð263Þ is the value of the scale factor corresponding to the
1 γ−1 ρ̃ 8πGℏ2 c2 transition between the DM era and the DE era (see
Appendix E. 2 of [52]). The fast oscillation regime is
where the upper sign corresponds to K > 0 and the lower not valid at very early times. In that case, the ΛCDM model
sign corresponds to K < 0. We note that a necessary (which corresponds to the TF approximation of the ΛFDM
condition for the fast oscillation regime to be valid on a model) is not valid and we have to take into account the
relatively large period of time is that presence of the quantum potential in the hydrodynamic
equations. This may lead to a stiff matter era or even to an
1
3m2 c4 jKj γ−1 inflation era in the primordial universe [41,52].
σ≡ ≫ 1: ð264Þ
8πGℏ2 c2
We denote by (anti-) ΛFDM model a complex SF with a
16
If the fast oscillation condition (260) is not fulfilled we have potential of the form
to take into account quantum effects and come back to the
1 m2 c2 2
original Eqs. (1)–(6) [or Eq. (13)] for the complex SF or V tot ðjφj2 Þ ¼ jφj ∓ ρΛ c2 : ð266Þ
take into account the presence of the quantum potential in 2 ℏ2
the hydrodynamic equations (27) and (29) (i.e. we have to In the TF approximation, this model is equivalent to the (anti-)
go beyond the TF approximation) [41,57]. ΛCDM model (see Appendix E of [52]).
043502-27
C. Chaplygin gas model When γ ¼ 0, the fast oscillation regime is valid for
For the (anti-) Chaplygin gas model (γ ¼ −1), the fast sufficiently small ρ, hence for sufficiently large a, provided
oscillation regime is valid provided that that σ is large enough.
When γ < 0, the fast oscillation regime is valid for
sffiffiffiffiffiffiffi intermediate values of ρ (neither too small nor too large),
ρ̃3 3m2 c4 c2
≪ σ ≡ : ð270Þ hence for intermediate values of a (neither too small nor too
ρ̃2 1 8πGℏ2 jKj large), provided that σ is large enough.
Using the results ffiof Secs. V F and VI F, we find that

pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi E. Nonrelativistic limit
ρ̃ ¼ ða =aÞ6 ∓ 1. As a result, the condition from
Eq. (270) can be rewritten as The nonrelativistic limit is valid when ϵ ∼ ρc2 . According
to Eqs. (75) and (76) this condition corresponds to
6 6 3=2
a a
∓1 ≪ σ: ð271Þ 2
a a c
ρ γ−1
≪ : ð272Þ
K
For the anti-Chaplygin gas model (K > 0), the fast
oscillation regime is valid provided that σ ≫ 1 and
a =σ 1=3 ≪ a ≤ a . When γ > 1, the nonrelativistic limit is valid if ρ is
For the Chaplygin gas model (K < 0), the fast oscillation sufficiently small and when γ < 1 it is valid if ρ is
regime is valid provided that σ ≫ 1 and a =σ 1=3 ≪ sufficiently large. This corresponds to the matter era.
a ≪ a σ 1=6 . The fast oscillation regime is not valid at very In order to determine if the fast oscillation (or TF)
early and very late times. In these cases, we have to take approximation is valid in the nonrelativistic regime, we
into account the presence of the quantum potential in the have to compare the criteria from Eqs. (262) and (272).
hydrodynamic equations. Since the precise range of validity depends on the value of σ
we cannot give general results (a specific study would be
D. Other models required case by case). However, we can give the following
qualitative results:
We now briefly consider other models (different values When γ > 1, the fast oscillation regime and the non-
of K and γ) without going into too much detail. A more relativistic regime are always compatible at sufficiently
accurate study should be made specifically for each model large a.
along the lines sketched above. When γ < 1, the fast oscillation regime and the non-
relativistic regime may or may not be compatible.
1. K > 0
When γ > 1, the fast oscillation regime is valid for
VIII. STRUCTURE FORMATION
sufficiently small ρ, hence for sufficiently large a.
When −1 ≤ γ ≤ 1, the fast oscillation regime is valid for The study of structure formation in the cosmological
sufficiently small ρ (down to ρmin ), hence for sufficiently models discussed previously is outside the scope of this
large a (up to amax ). paper (see Ref. [64] for a detailed study of cosmological
When γ < −1, the fast oscillation regime is valid for perturbations in the case of a real self-interacting SF). The
sufficiently small ρ. In the α era, it is always valid if σ is general equations describing the development of perturba-
large enough and it is valid only for sufficiently small a tions for a complex SF in an expanding background are
otherwise. In the DM era, the fast oscillation regime is given in [57] using an hydrodynamic approach (see also
never valid if σ is too small and it is valid for sufficiently [84–88]). They have been studied in detail in the non-
large a (up to amax ) otherwise. relativistic limit both for fully quantum systems and for
semiclassical systems in the TF approximation [57,85].
2. K < 0 Here, we restrict ourselves to the nonrelativistic and
When γ ≥ 1, the fast oscillation regime is valid for semiclassical limit.17 In that case, the gravitational Jeans
sufficiently small ρ. In the DM era (normal branch), it is length is given by the usual expression [85]
always valid if σ is large enough and it is valid only for
17
sufficiently large a otherwise. In the DE era (peculiar The domain of validity of the TF (or fast oscillation) regime
branch), the fast oscillation regime is never valid if σ is too has been determined in Sec. VII. Actually, these results are valid
small and it is valid for sufficiently small a (starting from for the homogeneous background. It is not clear if they are also
valid for the inhomogeneous SF during the process of structure
amin ) otherwise. formation. Indeed, quantum effects may be important at small
When 0 < γ < 1, the fast oscillation regime is valid for scales even when they can be neglected at large scales (back-
sufficiently small ρ, hence for sufficiently large a. ground). A specific study is required to settle that point.
043502-28
1=2
c2s matter spectrum that are not observed [75] (see the
λJ ¼ 2π ; ð273Þ detailed discussion in Sec. XVI. F of [52]).
4πGρ
Remark: The above results are valid for a massive
where c2s ¼ P0 ðρÞ is the squared speed of sound. For a complex SF with a power-law self-interaction potential
polytropic equation of state P ¼ Kργ , we have in the fast oscillation regime (ω ≫ H) and in the non-
relativistic limit (ϵ ∼ ρc2 ). In that case, the equation of state
c2s ¼ Kγργ−1 : ð274Þ is polytropic: P ¼ Kργ ¼ Kðϵ=c2 Þγ . In the ultrarelativistic
limit, we obtain instead the linear equation of state (see
Secs. V and VI)
We need to impose c2s ≥ 0 in order to avoid tachyonic
(dynamical) instabilities.18 For K > 0 this demands γ ≥ 0 γ−1
and for K < 0 this demands γ ≤ 0. On the other hand, in P ¼ αϵ ¼ ϵ: ð277Þ
γþ1
order to have favorable conditions for structure formation,
the comoving Jeans length λcJ ¼ λJ =a should be sufficiently This equation of state is valid in the α era when (i) K > 0
small. Combining Eqs. (273) and (274), we get [85] and γ > 1 (early universe); (ii) K > 0 and γ < −1;
(iii) K < 0 and 0 < γ < 1 (late universe). It coincides
Kγργ−2 1=2 1 with the equation of state obtained in [59,63,64] for a
λcJ ¼ 2π : ð275Þ
4πG a massless real SF with a power-law self-interaction poten-
tial in the fast oscillation regime. The squared speed of
Recalling that ρ ∝ a−3 in the matter era, we find that sound is c2s ¼ P0 ðϵÞc2 ¼ ðγ − 1Þ=ðγ þ 1Þc2 . When −1 <
λcJ ∝ a−ð3γ−4Þ=2 . We see that the comoving Jeans length γ < 1 there is a tachyonic instability in agreement with
decreases with a when γ > 4=3, is constant when γ ¼ 4=3, [63,64] for a real SF. This instability invalidates the models
and increases with a when γ < 4=3 except when γ ¼ 0 with K < 0 and 0 < γ < 1 (late universe) for explaining
(ΛCDM model) in which case the Jeans length vanishes.19 cosmic acceleration (when 0 < γ < 1=2).
Note that the comoving quantum Jeans length
1=4 IX. CONCLUSION
ℏ2 1
λcQ ¼ 2π ð276Þ In this paper, we have considered a complex SF with a
16πGρm2 a
self-interaction potential Vðjφj2 Þ described by the KGE
decreases as a−1=4 (the quantum Jeans length increases as equations (A1) and (A3). In the nonrelativistic regime,
these equations reduce to the Gross-Pitaevskii-Poisson
a3=4 ) [52,85,88]. It is dominant (λQ > λJ ) at sufficiently
(GPP) equations (A48) and (A49). We have determined
late times when γ > 3=2 and at sufficiently early times
the equation of state PðρÞ associated with the self-
when γ < 3=2. In conclusion:
(i) When K > 0, we need γ ≥ 0 to avoid tachyonic interaction potential Vðjφj2 Þ in the TF approximation.
instabilities and we need γ ≥ 4=3 and K small to Some examples of SF potentials, and the corresponding
favor the formation of structures during the matter equations of state, are given in Appendix C. In this paper,
era (the comoving Jeans length should decrease with we have specifically considered the cosmological evolu-
the scale factor). The standard BEC model (as > 0 tion of a spatially homogeneous complex SF with a
and γ ¼ 2) fulfills these conditions [40,41]. polytropic or an isothermal equation of state in the fast
(ii) When K < 0, we need γ ≤ 0 to avoid tachyonic oscillation regime (equivalent to the TF approximation).
instabilities and we need γ ≃ 0 and K small to favor In the case K > 0, corresponding to a positive pressure,
the formation of structures. Only UDM models close we have found the following results. The models with γ > 1
to the ΛCDM model (γ ¼ 0) are consistent with the describe the transition between an α era and a pressureless
observations. In the Chaplygin gas model (γ ¼ −1) DM era. For a jφj4 potential (γ ¼ 2), corresponding to the
the speed of sound and the Jeans length increase standard BEC with a repulsive self-interaction, there is a
with time preventing the formation of large scale transition between a dark radiation era (due to the SF) and
structures and leading to spurious oscillations in the a matter era. This model [40,41] is consistent with the
observations (see Sec. VIII). The models with γ ≤ 1
18
Actually, tachyonic instabilities (e.g. for a BEC with as < 0 describe a cyclic universe presenting periods of expansion
and γ ¼ 2) may be relevant in the early universe to trigger the from a big bang and periods of contraction to a big crunch.
formation of structures through a nongravitational instability The evolution is symmetric for −1 ≤ γ ≤ 1 (with a
[57,85,88,89].
19
The Jeans length behaves as λJ ∝ a−3ðγ−2Þ=2 . It decreases
vanishing energy density at the maximum scale factor)
with a when γ > 2, is constant when γ ¼ 2, and increases with a and asymmetric for γ < −1 (with a nonvanishing energy
when γ < 2. density at the maximum scale factor). Symmetric cyclic
043502-29
universes include the anti-ΛCDM model (γ ¼ 0) and the APPENDIX A: INHOMOGENEOUS

anti-Chaplygin gas model (γ ¼ −1). These models are not RELATIVISTIC COMPLEX SF
consistent with the observations. IN A CURVED SPACETIME
In the case K < 0, corresponding to a negative pressure, 1. Klein-Gordon-Einstein equations
we have found the following results. The models with γ ≥ 1
describe a bouncing universe dominated by DE when t < 0 The evolution of a possibly spatially inhomogeneous
and by pressureless DM when t > 0, or the converse. At relativistic complex SF φðxμ Þ ¼ φðx; y; z; tÞ, which may
t ¼ 0 the universe achieves its minimum radius amin and its represent the wave function of a relativistic BEC, is
governed by the KG equation
maximum energy density ϵi. There is no singularity. This
model can also describe a peculiar cosmic evolution with
dV tot
two branches [41]. The SF suddenly emerges at some finite □φ þ 2 φ ¼ 0; ðA1Þ
scale factor amin and energy density ϵi and follows either djφj2
the nonrelativistic branch (DM) or the ultrarelativistic pffiffiffiffiffiffi μν
where □ ¼ Dμ ∂μ ¼ p1ffiffiffiffi
−g ∂μ ð −gg ∂ν Þ is the d’Alembertian
branch (DE). Therefore, we get either an asymmetric
bouncing universe or a universe with two branches of operator in a curved spacetime. The potential V tot ðjφj2 Þ can
solutions. These models are not consistent with the be decomposed into a rest-mass energy term and a self-
observations. The models with γ < 1 describe the tran- interaction energy term as
sition between a pressureless DM era and a DE era. They
provide UDM models. For 0 < γ < 1, the energy density m2 c2 2
V tot ðjφj2 Þ ¼ jφj þ Vðjφj2 Þ: ðA2Þ
decreases to zero in the DE era (α era) like in quintessence 2ℏ2
models; the universe is decelerating when 1=2 < γ < 1
and accelerating when 0 < γ < 1=2. For γ ≤ 0, the energy The KG equation is coupled to the Einstein field equations
density tends to a constant in the DE era like in the
1 8πG
presence of a cosmological constant. This gives rise to a de Rμν − gμν R ¼ 4 T μν ; ðA3Þ
Sitter era at late times. These UDM models include the 2 c
ΛCDM model (γ ¼ 0) and the Chaplygin gas model where
(γ ¼ −1). The Chaplygin gas model does not give a good
agreement with the observations. Only polytropic models 1
with γ sufficiently close to 0, i.e., sufficiently close to the T μν ¼ ð∂μ φ ∂ν φ þ ∂ν φ ∂μ φÞ
2
ΛCDM model are consistent with the observations (see
1 ρσ 2
Sec. VIII). − gμν g ∂ρ φ ∂σ φ − V tot ðjφj Þ ðA4Þ
2
A limitation of our study is that we have not studied in
detail (case by case) the validity of the fast oscillation
is the energy-momentum tensor of the SF (the term in
regime (this is done in [41] for the quartic potential and in
brackets corresponds to the Lagrangian density of the SF).
Appendix E of [52] for the ΛFDM model). This is partly This leads to the Klein-Gordon-Einstein (KGE) equations.
due to reasons of conciseness and partly due to the fact that The energy-momentum tensor satisfies the equation
most models are not consistent with the observations so it Dμ T μν ¼ 0, expressing the local conservation of energy
may not be necessary to perform a more detailed study
and momentum. On the other hand, the quadricurrent
than the one given here. However, in a companion paper
[52], we consider the logotropic model which is consistent m
with the observations. In that case we study the validity of Jμ ¼ − ðφ ∂μ φ − φ∂μ φ Þ ðA5Þ
2iℏ
the fast oscillation regime in detail.
In Ref. [55] we have shown that the equation of state of a satisfies the equation Dμ Jμ ¼ 0, expressing the local con-
relativistic barotropic fluid can be specified in different servation of charge (see, e.g., [55] for details). The charge
manners depending on whether the pressure is expressed in Z
terms of the energy density ϵ (model I), the rest-mass e pffiffiffiffiffiffi
Q¼ J0 −gd3 x ðA6Þ
density ρm (model II), or the pseudorest-mass density ρ mc
(model III). In the present paper, we have considered the
cosmological evolution of fluids described by a polytropic of the SF is proportional to the number N of bosons
equation of state of type III. The cosmological evolution of (Q ¼ Ne) provided that antibosons are counted negatively
fluids described by a polytropic equation of state of type I [58]. Therefore, the equation Dμ Jμ ¼ 0 also expresses the
has been studied in [73,90,91] and the cosmological local conservation of the boson number. For a real SF the
evolution of fluids described by a polytropic equation of quadricurrent vanishes implying that the particle number is
state of type II has been studied in [50]. not conserved.
043502-30
2. The de Broglie transformation Introducing the pseudoquadrivelocity22

We can write the KG equation (A1) under the form of
∂μ Stot
hydrodynamic equations by making the de Broglie [92–94] vμ ¼ − ¼ −∂μ θ; ðA15Þ
transformation. To that purpose, we write the SF as m
ℏ pffiffiffi iStot =ℏ we can rewrite Eqs. (A12) and (A13) as

φ¼ ρe ; ðA7Þ
m
Dμ ðρvμ Þ ¼ 0; ðA16Þ
where ρ is the pseudorest-mass density 20
and Stot is the
action. They are given by 1
mv vμ − QdB − mV 0tot ðρÞ ¼ 0: ðA17Þ
2 μ

m2 ℏ φ
ρ ¼ 2 jφj2 and Stot ¼ ln : ðA8Þ Taking the gradient of the quantum Hamilton-Jacobi
ℏ 2i φ equation (A17) we obtain [67]
We also have dvν 1

≡ vμ Dμ vν ¼ ∂ν QdB þ ∂ν V 0 ðρÞ; ðA18Þ
dt m
1
V tot ðρÞ ¼ ρc2 þ VðρÞ: ðA9Þ which can be interpreted as a relativistic quantum Euler
2
equation (with the limitation mentioned in footnotes 20
For convenience, we define θ ¼ Stot =m. In that case, and 22). The first term on the right-hand side can be
Eq. (A7) becomes interpreted as a quantum force and the second term as a
pressure force ð1=ρÞ∂ν P. The pressure PðρÞ satisfies the
ℏ pffiffiffi imθ=ℏ relation ð1=ρÞP0 ðρÞ ¼ h0 ðρÞ ¼ V 00 ðρÞ, where hðρÞ ¼ V 0 ðρÞ
φ¼ ρe : ðA10Þ is the pseudoenthalpy. Integrating this relation, we get
m

The angle (phase) and the pulsation of the SF are given by21 0 VðρÞ 0 2
PðρÞ ¼ ρhðρÞ−VðρÞ ¼ ρV ðρÞ−VðρÞ ¼ ρ : ðA19Þ
ρ
_ _
Θ¼
Stot mθ
¼ ; _ ¼ − Stot ¼ − mθ ¼ Etot ;
ω ¼ −Θ ðA11Þ This equation determines the pseudoequation of state PðρÞ
ℏ ℏ ℏ ℏ ℏ
as a function of the potential VðρÞ. Inversely, the potential
is determined by the pseudoequation of state according to
where Etot ¼ −S_ tot is the energy.
Substituting the de Broglie transformation from Z
PðρÞ
Eq. (A10) into the KG equation (A1), and separating the VðρÞ ¼ ρ dρ: ðA20Þ
real and the imaginary parts, we get ρ2
The pseudosquared speed of sound is c2s ¼ P0 ðρÞ ¼

Dμ ðρ∂μ θÞ ¼ 0; ðA12Þ ρV 00 ðρÞ.
Using the de Broglie transformation (A7) the energy-
pffiffiffi momentum tensor (A4) is given, in the hydrodynamic
1 ℏ2 □ ρ
∂ θ∂ θ − 2 pffiffiffi − V 0tot ðρÞ ¼ 0:
μ
ðA13Þ representation, by
2 μ 2m ρ

ℏ2 1
Equation (A12) can be interpreted as a continuity equation T μν ¼ ρ∂μ θ∂ν θ þ 2 ∂μ ρ∂ν ρ − gμν gμν ρ∂μ θ∂ν θ
4m ρ 2
and Eq. (A13) can be interpreted as a quantum relativistic
Hamilton-Jacobi (or Bernoulli) equation with a relativistic ℏ2
þ 2 gμν ∂μ ρ∂ν ρ − V tot ðρÞ : ðA21Þ
covariant quantum potential 8m ρ
pffiffiffi
ℏ2 □ ρ Similarly, the quadricurrent (A5) can be written as
QdB ¼ pffiffiffi : ðA14Þ
2m ρ
Jμ ¼ −ρ∂μ θ ¼ ρvμ : ðA22Þ
20
We stress that ρ is not the rest-mass density ρm ¼ nm (see
below). It is only in the nonrelativistic regime c → þ∞ that ρ The pseudoquadrivelocity vμ does not satisfy vμ vμ ¼ c2 so it
22
coincides with the rest-mass density ρm. is not guaranteed to be always timelike. Nevertheless, vμ can be
21
The angle was noted θ instead of Θ in Sec. II and in [41]. introduced as a convenient notation.
043502-31
Therefore, the continuity equation (A12) or (A16) is Dμ ðρ∂μ θÞ ¼ 0; ðA32Þ

equivalent to Dμ Jμ ¼ 0. It expresses the conservation of
the charge Q of the complex SF (or the conservation of the 1
∂ θ∂μ θ − V 0tot ðρÞ ¼ 0: ðA33Þ
boson number N) 2 μ
Z
e pffiffiffiffiffiffi Equation (A32) can be interpreted as a continuity equation
Q ¼ Ne ¼ − ρ∂0 θ −gd3 x: ðA23Þ
mc and Eq. (A33) can be interpreted as a classical relativistic
Hamilton-Jacobi (or Bernoulli) equation.
In the following we take e ¼ 1 so that Q ¼ N. Assuming In order to determine the rest mass density, we can repeat
∂μ θ∂μ θ > 0, we can introduce the fluid quadrivelocity the same procedure as before. Assuming V 0tot > 0, and
using Eq. (A33), we introduce the fluid quadrivelocity
∂μ θ
uμ ¼ − pffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi c; ðA24Þ
∂μ θ∂μ θ ∂μ θ
uμ ¼ − pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi c; ðA34Þ
2V 0tot ðρÞ
which satisfies the identity
which satisfies the identity (A25). Using Eqs. (A22) and
uμ uμ ¼ c2 : ðA25Þ (A34), we get
Using Eqs. (A22) and (A24), we get ρ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Jμ ¼ 2V 0tot ðρÞuμ ; ðA35Þ
ρ qffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi c
Jμ ¼ ∂μ θ∂μ θuμ ; ðA26Þ
c and we can write the continuity equation (A32) as
and we can write the continuity equation (A12) as pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Dμ ½ρ 2V 0tot ðρÞuμ ¼ 0: ðA36Þ
qffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Dμ ½ρ ∂μ θ∂μ θuμ ¼ 0: ðA27Þ Comparing the expressions of Jμ given by Eqs. (A28) and
(A35), we find that the rest-mass density ρm ¼ nm is given
The rest-mass density ρm ¼ nm (which is proportional to in the TF approximation, by
the charge density ρe ) is defined by
ρ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
μ μ ρm ¼ 2V 0tot ðρÞ: ðA37Þ
J ¼ ρm u : ðA28Þ c
The continuity equation Dμ J μ ¼ 0 can be written as This is the limit form of Eq. (A31) with QdB ¼ 0. We note
that ρm ≠ J0 =c in general. When V ¼ ρΛ c2 is constant
Dμ ðρm uμ Þ ¼ 0: ðA29Þ (corresponding to the ΛCDM model), the rest-mass density
coincides with the pseudorest-mass density (ρm ¼ ρ). We
Comparing the expressions of Jμ given by Eqs. (A26) and also have ρm ¼ ρ in the nonrelativistic limit c → þ∞.
(A28), we find that the rest-mass density ρm ¼ nm of the In the TF approximation, the energy-momentum tensor
SF is given by from Eq. (A21) reduces to
ρ qffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi
ρm ¼ ∂μ θ∂μ θ: ðA30Þ 1 μν
T μν ¼ ρ∂μ θ∂ν θ − gμν g ρ∂μ θ∂ν θ − V tot ðρÞ : ðA38Þ
c 2
We note that ρm ≠ J0 =c in general. Using the Bernoulli
Using Eq. (A34), we get
equation (A13), we obtain
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
pffiffiffi ffi uμ uν
T μν ¼ 2ρV 0tot ðρÞ − gμν ½ρV 0tot ðρÞ − V tot ðρÞ: ðA39Þ
ρ ℏ2 □ ρ c2
ρm ¼ pffiffiffi þ 2V 0tot ðρÞ: ðA31Þ
c m2 ρ
The energy-momentum tensor can be written under the
perfect fluid form
uμ uν
3. TF approximation T μν ¼ ðϵ þ PÞ − Pgμν ; ðA40Þ
c2
In the classical limit (ℏ → 0) or in the TF approximation
where the quantum potential can be neglected, the hydro- where ϵ is the energy density and P is the pressure,
dynamic equations (A12) and (A13) reduce to provided that we make the identifications
043502-32
ϵ ¼ ρV 0tot ðρÞ þ V tot ðρÞ ¼ ρc2 þ ρV 0 ðρÞ þ VðρÞ; ðA41Þ 2

ΔΦ 2 − 3H ;
¼ jψj ðA49Þ
P ¼ ρV 0tot ðρÞ − V tot ðρÞ ¼ ρV 0 ðρÞ − VðρÞ; ðA42Þ 4πGa2 8πG
where we have used Eq. (A9) to get the second equalities. where 3H2 =8πG ¼ ρb is the background density
Eliminating ρ between these equations, we obtain the [see Eq. (6)].
equation of state PðϵÞ. The true squared speed of sound is Making the Madelung [95] transformation
pffiffiffiffiffiffiffiffiffiffiffiffi
ρV 00 ðρÞc2 ψðr; tÞ ¼ ρðr; tÞeiSðr;tÞ=ℏ ; ðA50Þ
c2s ¼ P0 ðϵÞc2 ¼ 2 : ðA43Þ
c þ ρV 00 ðρÞ þ 2V 0 ðρÞ
∇S
ρ ¼ jψj2 ; vðr; tÞ ¼ ; ðA51Þ
Remark: We note that the expression of the pressure is the ma
same as in Eq. (A19). It is also the same as the one obtained
in the nonrelativistic limit c → þ∞ where the KGE where ρ is the mass density, S is the action and v is the
equations reduce to the GPP equations (see below). By velocity field, we obtain the system of hydrodynamic
contrast, the enthalpy differs from the pseudoenthalpy equations [66,84,85]:
except in the nonrelativistic limit. Using Eqs. (49),
(A33), (A37), (A41), and (A42) we find that the enthalpy ∂ρ 1
þ 3Hρ þ ∇ · ðρvÞ ¼ 0; ðA52Þ
is given by ∂t a
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffi

h¼ 2V 0tot ðρÞc ¼ c ∂μ θ∂μ θ: ðA44Þ ∂S ð∇SÞ2 ℏ2 Δ ρ
þ ¼ pffiffiffi − mΦ − mhðρÞ; ðA53Þ
∂t 2ma2 2ma2 ρ
Substituting Eq. (A9) into Eq. (A44), subtracting c2 , and pffiffiffi
taking the nonrelativistic limit c → þ∞, we obtain ∂v 1 ℏ2 Δ ρ 1 1
þ Hv þ ðv · ∇Þv ¼ 2 3 ∇ pffiffiffi − ∇Φ − ∇P;
h ¼ V 0 ðρÞ. ∂t a 2m a ρ a ρa
ðA54Þ
4. Nonrelativistic limit
To take the nonrelativistic limit, it is convenient to work ΔΦ 3H 2
with the conformal Newtonian gauge which is a perturbed 2
¼ρ− ; ðA55Þ
4πGa 8πG
form of the FLRW line element
where hðρÞ ¼ V 0 ðρÞ is the enthalpy and PðρÞ is the pressure
2 2 Φ 2 2 Φ defined by the relation hR0 ðρÞ ¼ P0 ðρÞ=ρ. It is explicitly
ds ¼ c 1 þ 2 2 dt − aðtÞ 1 − 2 2 δij dxi dxj ; ðA45Þ
c c given by PðρÞ ¼ ρhðρÞ − hðρÞdρ, i.e.,
where Φðr; tÞ=c2 ≪ 1 represents the gravitational potential PðρÞ ¼ ρV 0 ðρÞ − VðρÞ: ðA56Þ
of classical Newtonian gravity. Making the Klein trans-
formation The squared speed of sound is c2s ¼ P0 ðρÞ ¼ ρV 00 ðρÞ. The
hydrodynamic equations (A52)–(A55) have a clear physi-
ℏ −imc2 t=ℏ cal interpretation. Equation (A52), corresponding to the
φðr; tÞ ¼ e ψðr; tÞ ðA46Þ
m imaginary part of the GP equation, is the continuity
equation. Equation (A53), corresponding to the real part
in the KGE equations (A1) and (A3), using the relation of the GP equation, is the Bernoulli or Hamilton-Jacobi
equation. Equation (A54), obtained by taking the gradient
m2 2 of Eq. (A53), is the Euler (momentum) equation.
ρ ¼ jψj2 ¼ jφj ; ðA47Þ
ℏ2 Equation (A55) is the Poisson equation. It can be written
as ΔΦ ¼ 4πGa2 ðρ − ρb Þ where ρb is the background
and taking the nonrelativistic limit c → þ∞, we get the density of the expanding universe.
generalized GPP equations in an expanding universe (see, Remark: For a complex SF, we note that the potential
e.g., [57,66,67] for details) Vðjψj2 Þ that appears in the GP equation (A48) is the same as
the potential Vðjφj2 Þ that appears in the KG equation (A1)
∂ψ 3 ℏ2 dV
iℏ þ iℏHψ ¼ − 2
Δψ þ mΦψ þ m ψ; ðA48Þ with Eq. (A2) (see, e.g., [57,66,67]). For a real SF, the two
∂t 2 2ma djψj2 potentials generally differ (see, e.g., [77,96]).
043502-33
APPENDIX B: HOMOGENEOUS COMPLEX Etot ¼ −S_ tot ¼ −mθ_ ¼ ℏω ðB9Þ

SF IN AN EXPANDING UNIVERSE
In this appendix, we apply the results of Appendix A to a is the energy of the SF. Equation (B8) expresses the
cosmological context, namely for a homogeneous complex conservation of the charge of the complex SF (or equiv-
SF in an expanding background, and we recover the results alently the conservation of the boson number). It can be
of Sec. II. written as
Qm2 c2
1. General results ρEtot ¼ ; ðB10Þ
a3
The cosmological evolution of a spatially homogeneous
complex SF in an expanding universe is governed by the where Q ¼ Ne is a constant of integration representing
KGF equations the charge of the SF (proportional to the boson number N)
[40–42,51,56,57,97]. This equation can also be directly
1 d2 φ 3H dφ dV obtained from Eq. (A23).
2 2
þ 2 þ 2 tot2 φ ¼ 0; ðB1Þ
c dt c dt djφj The quantum Hamilton-Jacobi (or Bernoulli) equation
(A13) takes the form
8πG
H2 ¼ ϵ; ðB2Þ pffiffiffi pffiffiffi
3c2 1 d2 ρ 2 1
d ρ
E2tot ¼ ℏ2 pffiffiffi þ 3Hℏ pffiffiffi þ 2m2 c2 V 0tot ðρÞ: ðB11Þ
ρ dt2 ρ dt
which can be deduced from the KGE equations (A1) and
(A3). The energy density ϵðtÞ and the pressure PðtÞ of the
Finally, we have established in the general case that
SF are given by
the rest-mass density of the SF is given by Eq. (A30). For
a spatially homogeneous SF in an expanding background,
1 dφ2
ϵ ¼ 2 þ V tot ðjφj2 Þ; ðB3Þ we get
2c dt
ρ 1 ℏω S_
1 dφ2 E
ρm ¼ − ∂0 θ ¼ − 2 ρθ_ ¼ ρ 2 ¼ ρ tot2 ¼ −ρ tot2 : ðB12Þ
P ¼ 2 − V tot ðjφj2 Þ: ðB4Þ c c mc mc mc
2c dt
In the following, we use the hydrodynamic representation We note that, in this special case, ρm ¼ J0 =c, where J0 ¼
of the SF (see Appendix A 2). Making the de Broglie −ρ∂0 Stot =m is the time component of the quadricurrent of
transformation (A10), the energy density and the pressure charge, but this relation is not true for an inhomogeneous SF.
of the SF can be written as Combining Eq. (B12) with the Bernoulli equation (B11), we
obtain the relation
1 _2 ℏ2 sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ϵ¼ ρθ þ 2 2 ρ_ 2 þ V tot ðρÞ; ðB5Þ pffiffiffi pffiffiffi
2c 2
8m ρc ρ ℏ2 1 d2 ρ 3Hℏ2 1 d ρ
ρm ¼ pffiffiffi þ pffiffiffi þ 2V 0tot ðρÞ;
c m2 c2 ρ dt2 m2 c2 ρ dt
1 _2 ℏ2
P¼ ρθ þ ρ_ 2 − V tot ðρÞ: ðB6Þ ðB13Þ
2c2 8m2 ρc2
The equation Dν T μν ¼ 0 leads to the energy conserva- which corresponds to Eq. (A31). On the other hand, using
tion equation Eq. (B12), Eqs. (B8) and (B10) can be rewritten as
dϵ dρm
þ 3Hðϵ þ PÞ ¼ 0: ðB7Þ þ 3Hρm ¼ 0 ðB14Þ
dt dt
This equation can also be obtained from the KG and

equation (B1) with Eqs. (B3) and (B4).
The equation Dμ Jμ ¼ 0, which is equivalent to the ρm ¼
Qm
: ðB15Þ
continuity equation (A12), can be written as a3
d Equation (B14) can also be obtained by combining the first

ðE ρa3 Þ ¼ 0; ðB8Þ law of thermodynamics for a cold fluid with the energy
dt tot
conservation equation (see Sec. III). These equations express
where the conservation of the particle number.
043502-34
Z
2. TF approximation 2V 0tot ðρÞ þ ρV 00tot ðρÞ
¼ −6 ln a: ðB24Þ
In the TF approximation (ℏ → 0), the energy density ϵ ρV 0tot ðρÞ
and the pressure P of the SF [see Eqs. (B5) and (B6)]
reduce to Equation (B24) integrates to give Eq. (B19).
1 _2 ρE2tot 3. Nonrelativistic limit

ϵ¼ 2
ρθ þ V tot ðρÞ ¼ þ V tot ðρÞ; ðB16Þ
2c 2m2 c2 Combining Eqs. (A46) and (A50) and comparing
the resulting expression with Eq. (A7), we see that
1 _2 ρE2tot
P¼ ρ θ − V ðρÞ ¼ − V tot ðρÞ; ðB17Þ Stot ¼ −mc2 tþS, giving Etot ¼ mc2 þE with E ¼ −dS=dt.
2c2 2m2 c2
tot
We can then rewrite the previous equations in terms of S
where we have used Eq. (B9) to get the second equalities. and E instead of Stot ¼ mθ and Etot . Then, taking the
On the other hand, the quantum Hamilton-Jacobi (or nonrelativistic limit c → þ∞, we obtain
Bernoulli) equation (B11) reduces to
dρ 3H2
þ 3Hρ ¼ 0; E ¼ mV 0 ðρÞ; ¼ ρ: ðB25Þ
E2tot ¼ 2m2 c2 V 0tot ðρÞ: ðB18Þ dt 8πG
Combining Eqs. (B10) and (B18), we obtain These equations can also be obtained from Eqs. (A52)–
(A55) by considering a homogeneous SF. They are easily
ρ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Qm solved to give ρ ∝ a−3 , a ∝ t2=3 and ρ ¼ 1=ð6πGt2 Þ. This is
2V 0tot ðρÞ ¼ 3 : ðB19Þ the EdS solution. Therefore, in the nonrelativistic limit, the
c a
homogeneous SF/BEC behaves as CDM. For the power-law
This equation determines the relation between the pseudo potential (68), we find that E ¼ Kmγ=½ðγ − 1Þð6πGt2 Þγ−1 .
rest-mass density ρ and the scale factor a. In particular, for the Chaplygin gas (γ ¼ −1), we get
According to Eqs. (B12) and (B18), the rest-mass E ¼ 18π 2 KmG2 t4 ; for the standard BEC (γ ¼ 2), we get
density is given by E ¼ 2as ℏ2 =ð3Gm2 t2 Þ; for the ΛCDM model (γ ¼ 0), we
get E ¼ 0; for the superfluid (γ ¼ 3), we get E ¼ Km=
ρ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ρm ¼ 2V 0tot ðρÞ: ðB20Þ ð24π 2 G2 t4 Þ. On the other hand, for the potential (132), we
c find that E ¼ −kB T lnð6πGρ t2 Þ.
This shows that (B19) is equivalent to the conservation of
the rest-mass [see Eq. (B15)]. Note that Eq. (B20) is always
true in the TF approximation even for inhomogeneous APPENDIX C: EXAMPLES OF POTENTIALS
systems (see Appendix A 3). OF SELF-INTERACTION
Finally, inserting the Bernoulli equation (B18) into Let us consider a complex SF φ governed by the
Eqs. (B16) and (B17), we find that the energy density nonlinear KG equation
and the pressure of the SF in the TF approximation are
given by m2 c2 dV
□φ þ 2
φþ2 φ¼0 ðC1Þ
ℏ djφj2
ϵ¼ ρV 0tot ðρÞ þ V tot ðρÞ; ðB21Þ
with a self-interaction potential Vðjφj2 Þ. In the TF
P ¼ ρV 0tot ðρÞ − V tot ðρÞ: ðB22Þ approximation, this potential is associated with a baro-
tropic equation of state PðρÞ ¼ ρV 0 ðρÞ − VðρÞ where ρ ¼
These relations are always true in the TF approximation even jψj2 ¼ ðm2 =ℏ2 Þjφj2 is the pseudorest-mass density (see
for inhomogeneous systems (Appendix A 3). Appendix A). In the nonrelativistic limit, the nonlinear KG
Remark: Eqs. (B21) and (B22) determine the equation of equation reduces to the generalized GP equation
state P ¼ PðϵÞ. As a result, we can obtain Eq. (B19)
directly from Eqs. (B21), (B22) and the energy conserva-
∂ψ 3 ℏ2 dV
tion equation (B7). Indeed, combining these equations we iℏ þ iℏHψ ¼ − 2
Δψ þ mΦψ þ m ψ; ðC2Þ
obtain ∂t 2 2ma djψj2
dρ which involves the same self-interaction potential Vðjψj2 Þ.

½2V 0tot ðρÞ þ ρV 00tot ðρÞ ¼ −6HρV 0tot ðρÞ; ðB23Þ We give below some examples of self-interaction potentials
dt
and their corresponding equations of state PðρÞ. A more
leading to detailed discussion of these results is given in [55].
043502-35
The power-law potential (see Sec. IVA) (iii) The index γ ¼ 0 corresponds to the ΛFDM model
interpreted as a UDM model
2γ
2 K m
Vðjφj Þ ¼ jφj2γ ðγ ≠ 1Þ ðC3Þ Vðjφj2 Þ ¼ −K; P ¼ K; ðC15Þ
γ−1 ℏ
is associated with the polytropic equation of state P ¼ Kργ . m2 c2

□φ þ φ ¼ 0; ðC16Þ
The KG equation becomes ℏ2

m2 c2 2Kγ m 2γ 2ðγ−1Þ Vðjψj2 Þ ¼ −K; ðC17Þ
□φ þ 2 φ þ jφj φ ¼ 0: ðC4Þ
ℏ γ−1 ℏ
∂ψ 3 ℏ2
In the nonrelativistic limit, using iℏ þ iℏHψ ¼ − Δψ þ mΦψ: ðC18Þ
∂t 2 2ma2
K
Vðjψj2 Þ ¼ jψj2γ ðγ ≠ 1Þ; ðC5Þ (iv) The index γ ¼ 3 corresponds to a superfluid taking
γ−1 into account three-body interactions
we obtain the generalized GP equation
K m 6 6
Vðjφj2 Þ ¼ jφj ; P ¼ Kρ3 ; ðC19Þ
2 2 ℏ
∂ψ 3 ℏ Kmγ 2ðγ−1Þ
iℏ þ iℏHψ ¼ − Δψ þ mΦψ þ jψj ψ:
∂t 2 2ma 2 γ−1 6
m2 c2 m
ðC6Þ □φ þ 2 φ þ 3K jφj4 φ ¼ 0; ðC20Þ
ℏ ℏ
K 6
(i) The index γ ¼ −1 corresponds to the Chaplygin gas Vðjψj2 Þ ¼ jψj ; ðC21Þ
2

K ℏ 2 1 K
Vðjφj2 Þ ¼ − ; P¼ ; ðC7Þ ∂ψ 3 ℏ2 3Km 4
2 m jφj2 ρ iℏ þ iℏHψ ¼ − 2
Δψ þ mΦψ þ jψj ψ:
∂t 2 2ma 2
2 ðC22Þ
m2 c2 ℏ 1
□φ þ 2 φ þ K φ ¼ 0; ðC8Þ
ℏ m jφj4
(v) For γ ¼ 1=2 we get
K 1
Vðjψj2 Þ ¼ − ; ðC9Þ Vðjφj2 Þ ¼ −2K
m
jφj; P ¼ Kρ1=2 ; ðC23Þ
2 jψj2 ℏ
∂ψ 3 ℏ2 Km 1 m2 c2 m 1
iℏ þ iℏHψ ¼ − Δψ þ mΦψ þ ψ: □φ þ φ − 2K φ ¼ 0; ðC24Þ
∂t 2 2ma2 2 jψj4 ℏ 2 ℏ jφj
ðC10Þ
Vðjψj2 Þ ¼ −2Kjψj; ðC25Þ
(ii) The index γ ¼ 2 corresponds to the standard BEC
taking into account two-body interactions ∂ψ 3 ℏ2 1
4 iℏ þ iℏHψ ¼ − 2
Δψ þ mΦψ − Km ψ:
∂t 2 2ma jψj
2 m
Vðjφj Þ ¼ K jφj4 ; P ¼ Kρ2 ; ðC11Þ ðC26Þ
ℏ
4
m2 c2 m
□φ þ 2 φ þ 4K jφj2 φ ¼ 0; ðC12Þ The potential (see Sec. IV B)
ℏ ℏ
2 2
Vðjψj2 Þ ¼ Kjψj4 ; ðC13Þ kB Tm 2
2 m jφj
Vðjφj Þ ¼ jφj ln −1 ðC27Þ
ℏ2 ρ ℏ2
∂ψ 3 ℏ2
iℏ þ iℏHψ ¼ − Δψ þ mΦψ is associated with the isothermal equation of state
∂t 2 2ma2
P ¼ ρkB T=m. It can take into account finite temperature
þ 2Kmjψj2 ψ: ðC14Þ effects in DM [79–83]. The KG equation becomes
043502-36
2 2
m2 c2 2kB Tm m jφj on the KG and GP equations but it adds a constant term ϵΛ
□φ þ 2 φ þ 2
ln φ ¼ 0: ðC28Þ
ℏ ℏ ρ ℏ2 in the energy density, which is interpreted as DE.
Of course, we can consider a multitude of more general
In the nonrelativistic limit, using SF models by summing the potentials [like, e.g., those in
Eqs. (C3), (C27), (C31), and (C35)] or, equivalently, by
2
kB T 2 jψj summing the corresponding pressures (polytropic, iso-
Vðjψj2 Þ ¼ jψj ln −1 ; ðC29Þ thermal, logotropic, and constant). This is a form of
m ρ
Dalton’s law. The mixed equation of state generically reads
γ kB T 2 ρ
2 P ¼ Kρ þ ρ − ρΛ c þ A ln ; ðC36Þ
∂ψ 3 ℏ2 jψj m ρP
Δψ þ mΦψ þ kB T ln ψ:
∂t 2 2ma ρ
where we can add several polytropic terms with different
ðC30Þ index γ. This mixed equation of state is associated with a
complex SF potential of the form
The logarithmic potential [52]
2γ
2 2 2 K m
m jφj Vðjφj Þ ¼ jφj2γ
Vðjφj2 Þ ¼ −A ln þ 1 ðC31Þ γ−1 ℏ
ℏ2 ρP 2 2
k Tm m jφj
þ B 2 jφj2 ln − 1
is associated with the logotropic equation of state ℏ ρ ℏ2
2 2
P ¼ A ln ðρ=ρP Þ. The KG equation becomes 2 m jφj
þ ρΛ c − A ln þ1 ; ðC37Þ
ℏ2 ρ P
m2 c2 2A
□φ þ 2
φ − 2 φ ¼ 0: ðC32Þ
ℏ jφj which can be inserted in the KG equation (C1) to give

In the nonrelativistic limit, using m2 c2 2Kγ m 2γ 2ðγ−1Þ
□φ þ φ þ jφj φ
2 ℏ2 γ−1 ℏ
2 jψj 2 2
Vðjψj Þ ¼ −A ln þ1 ; ðC33Þ 2kB Tm m jφj 2A
ρP þ 2
ln 2
φ − 2 φ ¼ 0: ðC38Þ
ℏ ρ ℏ jφj
In the nonrelativistic, we obtain the potential
∂ψ 3 ℏ2 mA 2
Δψ þ mΦψ − 2 ψ: ðC34Þ 2 K 2γ kB T 2 jψj
∂t 2 2ma jψj Vðjψj Þ ¼ jψj þ jψj ln −1
γ−1 m ρ
2
This model is studied in detail in [52]. 2 jψj
Instead of the logarithmic potential we can consider the þ ρΛ c − A ln þ1 ; ðC39Þ
ρP
constant potential
which can be inserted in the GP equation (C2) to give
Vðjφj2 Þ ¼ V 0 ¼ ϵΛ ; ðC35Þ
∂ψ 3 ℏ2 Kmγ 2ðγ−1Þ
mimicking a cosmological constant like in Appendix E of iℏ þ iℏHψ ¼ − 2
Δψ þ mΦψ þ jψj ψ
∂t 2 2ma γ−1
[52]. It is associated with the constant equation of state 2
P ¼ −ϵΛ . It leads to the usual KG and GP equations (C16) jψj mA
þ kB T ln ψ − 2 ψ: ðC40Þ
and (C18). We note that a constant potential has no effect ρ jψj
043502-37
[1] E. W. Kolb and M. S. Turner, The Early Universe (Addison- [38] P. Jetzer and D. Scialom, Phys. Lett. A 169, 12 (1992).
Wesley Publishing Company, Reading, MA, 1989). [39] E. Seidel and W. M. Suen, Phys. Rev. Lett. 66, 1659 (1991).
[2] A. Zee, Quantum Field Theory in a Nutshell (Princeton [40] B. Li, T. Rindler-Daller, and P. R. Shapiro, Phys. Rev. D 89,
University Press, Princeton, NJ, 2003). 083536 (2014).
[3] P. S. Wesson, Space-Time-Matter, Modern Kaluza-Klein [41] A. Suárez and P. H. Chavanis, Phys. Rev. D 95, 063515
Theory (World Scientific, Singapore, 1999). (2017).
[4] A. Linde, Particle Physics and Inflationary Cosmology [42] L. A. Boyle, R. R. Caldwell, and M. Kamionkowski, Phys.
(Harwood, Chur, Switzerland, 1990). Lett. B 545, 17 (2002).
[5] C. Armendariz-Picon, T. Damour, and V. Mukhanov, Phys. [43] M. Joyce, Phys. Rev. D 55, 1875 (1997).
Lett. B 458, 209 (1999). [44] V. A. Belinsky, L. P. Grishchuk, I. M. Khalatnikov, and
[6] J. Garriga and V. Mukhanov, Phys. Lett. B 458, 219 (1999). Ya. B. Zeldovich, Phys. Lett. 155B, 232 (1985).
[7] A. G. Riess et al., Astron. J. 116, 1009 (1998). [45] V. A. Belinsky, L. P. Grishchuk, Ya. B. Zeldovich, and I. M.
[8] S. Perlmutter et al., Astrophys. J. 517, 565 (1999). Khalatnikov, Sov. Phys. JETP 62, 195 (1985).
[9] P. de Bernardis et al., Nature (London) 404, 955 (2000). [46] V. A. Belinsky and I. M. Khalatnikov, Sov. Phys. JETP 66,
[10] S. Hanany et al., Astrophys. J. 545, L5 (2000). 441 (1987).
[11] B. Ratra and P. J. E. Peebles, Phys. Rev. D 37, 3406 (1988). [47] T. Piran and R. M. Williams, Phys. Lett. 163B, 331 (1985).
[12] J. Frieman, C. Hill, A. Stebbins, and I. Waga, Phys. Rev. [48] D. Scialom and P. Jetzer, Phys. Rev. D 51, 5698 (1995).
Lett. 75, 2077 (1995). [49] P. Jetzer and D. Scialom, Phys. Rev. D 55, 7440 (1997).
[13] R. R. Caldwell, R. Dave, and P. J. Steinhardt, Phys. Rev. [50] P. H. Chavanis, Phys. Rev. D 92, 103004 (2015).
Lett. 80, 1582 (1998). [51] A. Arbey, J. Lesgourgues, and P. Salati, Phys. Rev. D 65,
[14] T. Chiba, T. Okabe, and M. Yamaguchi, Phys. Rev. D 62, 083514 (2002).
023511 (2000). [52] P. H. Chavanis, arXiv:2201.05908.
[15] C. Armendariz-Picon, V. Mukhanov, and P. J. Steinhardt, [53] B. Carvente, V. Jaramillo, C. Escamilla-Rivera, and D.
Phys. Rev. Lett. 85, 4438 (2000). Núñez, Mon. Not. R. Astron. Soc. 503, 4008 (2021).
[16] R. R. Caldwell, Phys. Lett. B 545, 23 (2002). [54] P. H. Chavanis, Eur. Phys. J. Plus 130, 130 (2015).
[17] R. R. Caldwell, M. Kamionkowski, and N. Weiberg, Phys. [55] P. H. Chavanis, arXiv:2109.05963.
Rev. Lett. 91, 071301 (2003). [56] J.-A. Gu and W.-Y. P. Hwang, Phys. Lett. B 517, 1
[18] A. Sen, Mod. Phys. Lett. A 17, 1797 (2002). (2001).
[19] G. Gibbons, Phys. Lett. B 537, 1 (2002). [57] A. Suárez and P. H. Chavanis, Phys. Rev. D 92, 023510
[20] A. Frolov, L. Kofman, and A. Starobinsky, Phys. Lett. B (2015).
545, 8 (2002). [58] L. D. Landau and E. M. Lifshitz, Fluid Mechanics
[21] T. Padmanabhan, Phys. Rev. D 66, 021301 (2002). (Pergamon Press, London, 1959).
[22] A. Feinstein, Phys. Rev. D 66, 063511 (2002). [59] M. S. Turner, Phys. Rev. D 28, 1243 (1983).
[23] T. Padmanabhan and T. R. Choudhury, Phys. Rev. D 66, [60] L. H. Ford, Phys. Rev. D 35, 2955 (1987).
081301 (2002). [61] P. J. E. Peebles and A. Vilenkin, Phys. Rev. D 60, 103506
[24] Z. K. Guo, Y. S. Piao, X. Zhang, and Y. Z. Zhang, Phys. (1999).
Lett. B 608, 177 (2005). [62] T. Matos and L. A. Ureña-López, Phys. Rev. D 63, 063506
[25] M. R. Setare, J. Sadeghi, and A. R. Amani, Phys. Lett. B (2001).
660, 299 (2008). [63] M. C. Johnson and M. Kamionkowski, Phys. Rev. D 78,
[26] A. R. Amani, Int. J. Theor. Phys. 50, 3078 (2011). 063010 (2008).
[27] D. J. Kaup, Phys. Rev. 172, 1331 (1968). [64] J. A. R. Cembranos, A. L. Maroto, and S. J. Núñez Jareño,
[28] R. Ruffini and S. Bonazzola, Phys. Rev. 187, 1767 (1969). J. High Energy Phys. 03 (2016) 013.
[29] M. Colpi, S. L. Shapiro, and I. Wasserman, Phys. Rev. Lett. [65] J. A. R. Cembranos, A. L. Maroto, S. J. Núñez Jareño, and
57, 2485 (1986). H. Villarrubia-Rojo, J. High Energy Phys. 08 (2018) 073.
[30] P. H. Chavanis and T. Harko, Phys. Rev. D 86, 064011 [66] A. Suárez and P. H. Chavanis, J. Phys. Condens. Matter 654,
(2012). 012008 (2015).
[31] W. Hu, R. Barkana, and A. Gruzinov, Phys. Rev. Lett. 85, [67] P. H. Chavanis and T. Matos, Eur. Phys. J. Plus 132, 30
1158 (2000). (2017).
[32] L. Hui, J. Ostriker, S. Tremaine, and E. Witten, Phys. Rev. D [68] S. Coleman, Nucl. Phys. B262, 263 (1985).
95, 043541 (2017). [69] P. H. Chavanis, Phys. Lett. B 758, 59 (2016).
[33] B. Moore, T. Quinn, F. Governato, J. Stadel, and G. Lake, [70] A. Kamenshchik, U. Moschella, and V. Pasquier, Phys. Lett.
Mon. Not. R. Astron. Soc. 310, 1147 (1999). B 511, 265 (2001).
[34] G. Kauffmann, S. D. M. White, and B. Guiderdoni, Mon. [71] N. Bilic, G. B. Tupper, and R. D. Viollier, Phys. Lett. B 535,
Not. R. Astron. Soc. 264, 201 (1993). 17 (2002).
[35] A. Klypin, A. V. Kravtsov, and O. Valenzuela, Astrophys. J. [72] V. Gorini, A. Kamenshchik, U. Moschella, and V. Pasquier,
522, 82 (1999). Phys. Rev. D 69, 123512 (2004).
[36] M. Kamionkowski and A. R. Liddle, Phys. Rev. Lett. 84, [73] P. H. Chavanis, Eur. Phys. J. Plus 129, 222 (2014).
4525 (2000). [74] P. H. Chavanis, Eur. Phys. J. Plus 130, 181 (2015).
[37] M. Boylan-Kolchin, J. S. Bullock, and M. Kaplinghat, Mon. [75] H. B. Sandvik, M. Tegmark, M. Zaldarriaga, and I. Waga,
Not. R. Astron. Soc. 415, L40 (2011). Phys. Rev. D 69, 123524 (2004).
043502-38
[76] P. P. Avelino, L. M. G. Beca, J. P. M. de Carvalho, and [86] A. Suárez and P. H. Chavanis, Phys. Rev. D 98, 083529
C. J. A. P. Martins, J. Cosmol. Astropart. Phys. 09 (2003) (2018).
002. [87] P. H. Chavanis, Universe 6, 226 (2020).
[77] P. H. Chavanis, Phys. Rev. D 98, 023009 (2018). [88] P. H. Chavanis, Phys. Rev. D 103, 123551 (2021).
[78] E. G. M. Ferreira, Astron. Astrophys. Rev. 29, 7 (2021). [89] P. Brax, P. Valageas, and J. A. R. Cembranos, Phys. Rev. D
[79] P. H. Chavanis, M. Lemou, and F. Méhats, Phys. Rev. D 91, 102, 083012 (2020).
063531 (2015). [90] P. H. Chavanis, Eur. Phys. J. Plus 129, 38 (2014).
[80] P. H. Chavanis, M. Lemou, and F. Méhats, Phys. Rev. D 92, [91] P. H. Chavanis, arXiv:1208.1185.
123527 (2015). [92] L. de Broglie, J. Physique 8, 225 (1927).
[81] P. H. Chavanis, Phys. Rev. D 100, 083022 (2019). [93] L. de Broglie, C.R. Hebd. Seances Acad. Sci. 185, 380 (1927).
[82] P. H. Chavanis, arXiv:2112.07726. [94] L. de Broglie, C.R. Hebd. Seances Acad. Sci. 185, 1118 (1927).
[83] P. H. Chavanis, Eur. Phys. J. B 95, 48 (2022). [95] E. Madelung, Zeit. F. Phys. 40, 322 (1927).
[84] P. H. Chavanis, Phys. Rev. D 84, 043531 (2011). [96] P. H. Chavanis, Phys. Rev. D 102, 083531 (2020).
[85] P. H. Chavanis, Astron. Astrophys. 537, A127 (2012). [97] S. Kasuya, Phys. Lett. B 515, 121 (2001).
043502-39

Cosmological Models Based On A Complex Scalar Field With A Power-Law Potential With Polytropic Equation - Chavanis2022

Uploaded by

Copyright:

Available Formats

Cosmological Models Based On A Complex Scalar Field With A Power-Law Potential With Polytropic Equation - Chavanis2022

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Cosmological Models Based On A Complex Scalar Field With A Power-Law Potential With Polytropic Equation - Chavanis2022

Uploaded by

Copyright:

Available Formats

PHYSICAL REVIEW D 106, 043502 (2022)

Cosmological models based on a complex scalar field with a power-law

(Received 9 November 2021; accepted 28 June 2022; published 1 August 2022)

I. INTRODUCTION Different forms of SFs were introduced such as quintes-

2470-0010=2022=106(4)=043502(39) 043502-1 © 2022 American Physical Society

m2 c4 dV P V 0 ðhjφj2 iÞhjφj2 i − Vðhjφj2 iÞ

dStot VðρÞ. Inversely, for a given equation of state, the potential is

Substituting Eq. (23) into the KG equation (1) and taking Z

d The various correspondences between the results of this

2 1=2 Z 00 ðρÞ Equation (47) can be integrated into

Qm equivalent for the evolution of the homogeneous back-

D. Two-fluid model Introducing the pseudorest-mass density defined by

the pseudorest-mass energy density coincides with rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

and The pressure is constant. This is the equation of state

It corresponds to a quartic self-interaction potential P ¼ K; ρm ¼ ρ; u ¼ −K ¼ V: ð111Þ

Eq. (113), we can then obtain the equation of state and

V. THE CASE K > 0 8

pressure. Since P > 0, the universe is always decelerat-

The pressure is a linear function P ∼ αϵ of the energy K>0

When ρ → 0, Eqs. (150) and (151) reduce to

ϵ ∼ ρc2 : ð157Þ 1e-06 0.0001 0.01

FIG. 2. Evolution of the energy density as a function of the

When ρ → þ∞, Eqs. (160) and (161) reduce to

C. The case 0 < γ < 1

fluid with an equation of state Pde ¼ −ϵde yielding a

0.3 present context, the anti-ΛCDM model is obtained from

scale factor decreases as a ∝ ðtbc − tÞjγþ1j=3jγj [correspond-

ing to a ∝ ðtbc − tÞ2=½3ð1þαÞ ] in the α era. This process

Considering now large values of ρ (ultrarelativistic amin

The temporal evolution aðtÞ of the scale factor is

The pseudorest-mass density increases as a increases and

For 0 < γ < 1 the equations determining the pseudorest- -4

This corresponds to the nonrelativistic regime P=ϵ ≪ 1 ε

(matterlike era) valid for a → 0. The equation of state is

1=ð1þγÞ 0.0001 0.01

This corresponds to the ultrarelativistic regime valid for K<0

density decrease with a and tend to zero when a → þ∞.

The pressure is a linear function P ∼ αϵ of the energy

ϵ ¼ ρc2 þ jKj: ð238Þ ρΛ c2

E. The case γ < 0 6

γþ1 1 FIG. 21. Evolution of the pseudorest-mass density as a function

When ρ → þ∞, Eqs. (244) and (245) reduce to

Figs. 21 and 22.

-10 This corresponds to the matterlike era (a ≪ a ) where

VII. VALIDITY OF THE FAST A. Standard BEC model

H2 8πGℏ2 ϵ where Λ ¼ 8πGρΛ is the cosmological constant and

Using the results ﬃof Secs. V F and VI F, we find that

universes include the anti-ΛCDM model (γ ¼ 0) and the APPENDIX A: INHOMOGENEOUS

2. The de Broglie transformation Introducing the pseudoquadrivelocity22

ℏ pffiffiffi iStot =ℏ we can rewrite Eqs. (A12) and (A13) as

We also have dvν 1

The pseudosquared speed of sound is c2s ¼ P0 ðρÞ ¼

Therefore, the continuity equation (A12) or (A16) is Dμ ðρ∂μ θÞ ¼ 0; ðA32Þ

ϵ ¼ ρV 0tot ðρÞ þ V tot ðρÞ ¼ ρc2 þ ρV 0 ðρÞ þ VðρÞ; ðA41Þ 2

pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffi

APPENDIX B: HOMOGENEOUS COMPLEX Etot ¼ −S_ tot ¼ −mθ_ ¼ ℏω ðB9Þ

This equation can also be obtained from the KG and

d Equation (B14) can also be obtained by combining the first