Journal of Cosmology and

Astroparticle Physics

Related content
- Gravitational radiation from compact
Scalar-tensor extension of the ΛCDM model binaries in scalar-tensor gravity
R N Lang

To cite this article: W.C. Algoner et al JCAP11(2016)034 - A h3 Gpc3 study of cosmic homogeneity
using BOSS DR12 quasar sample
Pierre Laurent, Jean-Marc Le Goff,
Etienne Burtin et al.

- Constraining the dynamical dark energy

View the article online for updates and enhancements. parameters: Planck-2013 vs WMAP9
B. Novosyadlyj, O. Sergijenko, R. Durrer et
al.

Recent citations
- Non-minimally coupled scalar field
cosmology with torsion
Antonella Cid et al

- Diagnostic of Horndeski theories

Louis Perenon et al

This content was downloaded from IP address 200.238.166.198 on 12/06/2018 at 19:09

J ournal of Cosmology and Astroparticle Physics
An IOP and SISSA journal

Scalar-tensor extension of the ΛCDM

model

JCAP11(2016)034
W.C. Algoner, H.E.S. Velten and W. Zimdahl
Departamento de Fı́sica, Universidade Federal do Espı́rito Santo,
Av. Fernando Ferrari, 514, Campus de Goiabeiras, CEP 29075-910,
Vitória, Espı́rito Santo, Brazil
E-mail: w.algoner@cosmo-ufes.org, velten@pq.cnpq.br,
winfried.zimdahl@pq.cnpq.br

Received July 20, 2016

Revised November 7, 2016
Accepted November 8, 2016
Published November 18, 2016

Abstract. We construct a cosmological scalar-tensor-theory model in which the Brans-Dicke

type scalar Φ enters the effective (Jordan-frame) Hubble rate as a simple modification of
the Hubble rate of the ΛCDM model. This allows us to quantify differences between the
background dynamics of scalar-tensor theories and general relativity (GR) in a transparent
and observationally testable manner in terms of one single parameter. Problems of the
mapping of the scalar-field degrees of freedom on an effective fluid description in a GR
context are discused. Data from supernovae, the differential age of old galaxies and baryon
acoustic oscillations are shown to strongly limit potential deviations from the standard model.

Keywords: dark energy theory, modified gravity

ArXiv ePrint: 1607.03952

c 2016 IOP Publishing Ltd and Sissa Medialab srl doi:10.1088/1475-7516/2016/11/034
Contents

1 Introduction 1

2 Basics of scalar-tensor theories 2

3 Einstein-frame description 4
3.1 General relations 4
3.2 Modeling the interaction 5

JCAP11(2016)034
3.3 Effective EoS and Hubble expansion rate for a linear dependence b(ϕ) = Kϕ 5
3.4 Consistency check 8

4 Jordan-frame description 8

5 Statistical (observational) analysis 18

6 Discussion 20

1 Introduction
In scalar-tensor theories the gravitational interaction is mediated both by a metric tensor and
a scalar field. The interest in this type of theories of gravity is connected with the expectation
that the observed late-time accelerated expansion of the Universe may be understood without
a dark-energy (DE) component [1–4]. Instead, it is the modified (compared with Einstein’s
theory) geometrical sector which is supposed to provide the desired dynamics [5–8]. This
may be seen as a geometrization of DE. Different aspects of scalar-tensor theories in general
or subclasses of them have been investigated in [9–22].
Scalar-tensor theories are formulated either in the Einstein frame or in the Jordan frame.
Both frames are related by a conformal transformation. While matter and scalar field energies
are separately conserved in the Jordan frame, the dynamics of both components is coupled in
the Einstein frame for any equation of state (EoS) different from that of radiation. Because
of the complex structure of scalar-tensor theories, simple solutions are difficult to obtain,
even if the symmetries of the cosmological principle are imposed. Hence, in practice, the
background expansion rate is usually obtained via numerical integration of the equations of
motion. In general, the scalar-tensor-theory based cosmological dynamics may substantially
differ from standard cosmology. Our focus here is on the simplest possible extension of the
standard ΛCDM model that scalar-tensor theory can provide. In this minimalist approach
we remain in the vicinity of the standard model at the present epoch and we aim to quantify
the differences between scalar-tensor theory and general relativity (GR) by establishing a
structure in which the scalar field Φ explicitly enters an analytic solution of the dynamics
such that for Φ = 1 the standard ΛCDM limit is recovered. To this purpose we construct a
simple model which is analytically solved in the Einstein frame. With the help of a conformal
transformation we then demonstrate how the field Φ, which is given as a certain power of
the scale factor, enters the (Jordan-frame) Hubble rate. Here we rely on an effective GR
description of the Jordan-frame dynamics to determine the geometric equivalent of DE.
In more detail, our starting point is a simple, analytically tractable expression for the
coupling between nonrelativistic matter and the (Einstein frame) scalar field which modifies

–1–
the standard decay of the matter energy density with the third power of the cosmic scale
factor. This interaction-triggered deviation of the standard decay in the Einstein frame is
modeled by a power-law behavior in terms of the Einstein-frame scale factor. Under this
condition and if additionally an effective energy density and an effective pressure, linked by
a constant “bare” EoS parameter, are assumed, an explicit solution of the Einstein-frame
dynamics is obtained with the mentioned power as an additional parameter. A straightfor-
ward conformal transformation then allows us to obtain the Hubble rate and the deceleration
parameter in terms of this parameter in the Jordan frame as well. For the value zero of such
new parameter, corresponding to Φ = 1, both frames become indistinguishable and repro-
duce the dynamics of the standard ΛCDM model. Otherwise one has a variable Φ and the

JCAP11(2016)034
dynamics in both frames becomes different, deviating from that of the standard model. The
analytic expression for the Hubble rate which explicitly clarifies the impact of the scalar field
on the cosmological dynamics is the main achievement of this paper. We shall confront the
deviations from the standard model with data from supernovae of type Ia (SNIa), the dif-
ferential age of old galaxies that have evolved passively (using H(z), where H is the Hubble
rate and z is the redshift parameter) and baryon acoustic oscillations (BAO).
The structure of the paper is as follows. In section 2 we recall basic general relations for
scalar-tensor theories and specify them to the homogeneous and isotropic case. In section 3 we
set up an effective two-component description in the Einstein frame, introduce our interaction
model and find the Einstein-frame Hubble rate. The transformation to the Jordan frame is
performed in section 4 where we also discuss the implications of a mapping of the scalar-field
degrees of freedom on the effective fluid dynamics in a GR context. Section 5 is devoted to
a Bayesian statistical analysis on the basis of observational data of SNIa, H(z) and BAO.
Finally, in section 6 we summarize our results.

2 Basics of scalar-tensor theories

Scalar-tensor theories are based on the (Jordan-frame) action (see, e.g., [18, 20, 21])

4 √
Z  
1 ω(Φ) 2
S(gµν , Φ) = 2 d x −g ΦR − (∇Φ) − U (Φ) + Sm (gµν ) (2.1)
2κ Φ
with a minimally coupled matter part
√
Z
Sm = d4 x −gLm (gµν ) , (2.2)

where κ2 = 8πG and Lm denotes the matter Lagrangian. The dynamical field equations are
   
1 2 ω(Φ) 1 2 1
Φ Rµν − gµν R = κ Tµν + ∂µ Φ∂ν Φ − gµν (∇Φ) + ∇µ ∇ν Φ − gµν Φ − gµν U,
2 Φ 2 2
(2.3)
where the energy-momentum tensor of the matter is obtained as usual via
2 δSm
Tµν = − √ . (2.4)
−g δg µν
The dynamics of the field Φ is dictated by
 
1 2 dω(Φ) 2 dU
Φ = κ T− (∇Φ) + Φ − 2U , (2.5)
2ω(Φ) + 3 dΦ dΦ
which implies a coupling to the trace T ≡ T r(Tµν ) of the matter energy-momentum tensor.

–2–
 Adopting the conformal transformation

1
gµν = g̃µν = e2 b(ϕ) g̃µν (2.6)
Φ
of the metric tensor gµν as well as a redefinition of the potential term

U (Φ)
V (ϕ) = (2.7)
2κ2 Φ2
and 2 2
κ2
 
1 dΦ db

JCAP11(2016)034
= = , (2.8)
4Φ2 dϕ dϕ 4ω(Φ) + 6
one obtains the Einstein-frame action
Z   Z
4
p 1 1  ˜ 2 4
p 
2b(ϕ)

S (g̃µν , ϕ) = d x −g̃ R̃ − ∇ϕ − Ṽ (ϕ) + d x −g̃ L̃m e g̃ µν , (2.9)
2κ2 2

in which the matter part is non-minimally coupled to the gravitational sector. Throughout
the paper, quantities without a tilde refer to the Jordan frame, quantities with a tilde have
their meaning in the Einstein frame.
Restricting ourselves to spatially flat homogeneous and isotropic cosmological mod-
els with a Robertson-Walker metric and assuming a perfect-fluid structure for the energy-
momentum tensor of the matter with energy density ρm , pressure pm and four-velocity uµ ,

Tµν = ρm uµ uν + pm hµν , hµν = gµν + uµ uν , (2.10)

the relevant Jordan-frame equations are

" #
κ2 ρm 1 1 ω(Φ) ∂Φ 2
 
2 ∂Φ 1
H = + − 3H + U , (2.11)
3 Φ 3Φ 2 Φ ∂t ∂t 2

1 da
where H = a dt is the Hubble rate of the Jordan frame and a is the Jordan-frame scale factor,
" #
κ2 ρm + pm 1 ω(Φ) ∂Φ 2 ∂Φ d2 Φ
 
dH
=− − −H + 2 , (2.12)
dt 2 Φ 2Φ Φ ∂t ∂t dt
!
2
d2 Φ
 
dΦ 1 dω dΦ dU
+ 3H = κ2 (ρm − 3pm ) − −Φ + 2U , (2.13)
dt2 dt 2ω + 3 dΦ dt dΦ

as well as the matter conservation

dρm
+ 3H (ρm + pm ) = 0. (2.14)
dt
The corresponding relations for the Einstein frame are
" #
κ2 1 dϕ 2
 
2
H̃ = ρ̃m + + Ṽ , (2.15)
3 2 dt̃

–3–
 1 dã
where H̃ = ã dt̃ is the Einstein-frame Hubble rate and ã is the Einstein-frame scale factor,
as well as
"  2 #
dH̃ κ2 dϕ
=− ρ̃m + p̃m + , (2.16)
dt̃ 2 dt̃
dρ̃m dϕ db
+ 3H̃ (ρ̃m + p̃m ) = − (−ρ̃m + 3p̃m ) (2.17)
dt̃ dt̃ dϕ
and
d2 ϕ dϕ db
+ 3H̃ + Ṽ,ϕ = (−ρ̃m + 3p̃m ) , (2.18)
dt̃2 dt̃ dϕ

JCAP11(2016)034
where we have introduced the notation Ṽ,ϕ = ∂∂ϕ
Ṽ
.
Different from the Jordan frame, the matter component is not separately conserved
here. The time coordinates and the scale factors of the FLRW metrics of both frames are
related by
dt = eb dt̃ and a = eb ã , (2.19)
respectively. The matter pressure and the matter energy density transform as
pm = e−4b p̃m and ρm = e−4b ρ̃m , (2.20)
pm p̃m
respectively. This means ρm = ρ̃m , i.e., the EoS parameter of the matter remains invariant.

3 Einstein-frame description
3.1 General relations
The Einstein-frame equations (2.15)–(2.18) are of an effective two-component structure in
which matter interacts with a scalar field. We associate an effective energy density ρ̃ϕ and
an effective pressure p̃ϕ to the scalar field by
1 dϕ 2 1 dϕ 2
   
ρ̃ϕ = + Ṽ and p̃ϕ = − Ṽ , (3.1)
2 dt̃ 2 dt̃
respectively. Equations (2.17) and (2.18) can then be written as
dρ̃m dϕ db
+ 3H̃ (1 + w̃m ) ρ̃m = Q ≡ (1 − 3w̃m ) ρ̃m (3.2)
dt̃ dt̃ dϕ
and
dρ̃ϕ dϕ db
+ 3H̃ (1 + w̃) ρ̃ϕ = −Q = − (1 − 3w̃m ) ρ̃m , (3.3)
dt̃ dt̃ dϕ
p̃
respectively, where w̃m = ρ̃p̃mm
is the matter EoS parameter and w̃ = ρ̃ϕϕ is the Einstein-
frame EoS parameter for the scalar field. Notice the difference in notation between the EoS
parameter w and the coupling ω in the action of the scalar-tensor theory. The interaction
vanishes for the special case w̃m = 1/3. The total energy density ρ̃ and the total pressure p̃ are
ρ̃ = ρ̃m + ρ̃ϕ , p̃ = p̃m + p̃ϕ , (3.4)
for which the conservation equation
dρ̃
+ 3H̃ (ρ̃ + p̃) = 0 (3.5)
dt̃
holds.

–4–
3.2 Modeling the interaction
In the special case p̃m = 0 the solution of (3.2) can be written as
dρ̃m ρ̃m df
ρ̃m = ρ̃m0 ã−3 f (ã) ⇒ + 3H̃ ρ̃m = , (3.6)
dt̃ f dt̃
where the function f encodes the effects of an interaction between matter and scalar field.
Comparing (3.2) and (3.6), it follows that
1 df dϕ db
= ⇒ f = eb(ϕ) . (3.7)
f dt̃ dt̃ dϕ

JCAP11(2016)034
The absence of an interaction means a constant f , equivalent to ϕ = ϕ0 = constant. Writing
the time derivative of f (ã) as
df df dã df
= = ãH̃ , (3.8)
dt̃ dã dt̃ dã
Eq. (3.3) becomes
dρ̃ϕ df
+ 3 (1 + w̃) ρ̃ϕ = −ρ̃m0 ã−3 . (3.9)
d ln ã d ln ã
For a constant EoS parameter w̃ the solution of this equation is
Z ã
−3(1+w̃) −3(1+w̃) df
ρ̃ϕ = ρ̃ϕ0 ã − ρ̃m0 ã dã ã3w̃ . (3.10)
ã0 dã

For a given interaction f (ã), the Hubble rate (2.15) is then determined by the sum of ρ̃m
from (3.6) with (3.8) and ρ̃ϕ from (3.10).
To construct an explicitly solvable model we shall assume a power-law behavior of the
interaction function f (ã),
f (ã) = ã3m . (3.11)
Clearly, for m = 0 the function f becomes constant and the interaction is absent.
With (3.11) the explicit solution of (3.10) then is
 
m m
ρ̃ϕ = ρ̃ϕ0 + ρ̃m0 ã−3(1+w̃) − ρ̃m ã−3(1−m) . (3.12)
w̃ + m w̃ + m 0
Via
dϕ db dϕ db
= 3mH̃, = 3m, (3.13)
dt̃ dϕ d ln ã dϕ
the interaction term Q becomes
Q = 3mH̃ ρ̃m . (3.14)

3.3 Effective EoS and Hubble expansion rate for a linear dependence b(ϕ) = Kϕ
In order to obtain analytical expressions for the dynamics we consider the simple case of a
linear dependence r
κ2
b = b(ϕ) = Kϕ, K= , (3.15)
4ω + 6
where the expression for K follows from (2.8). This implies also the relations
3m db
ϕ= ln ã, b = 3m ln ã, eb = ã3m , = 3mH̃, (3.16)
K dt̃

–5–
as well as
Φ = e−2Kϕ = e−2b = ã−6m . (3.17)
The power m is a direct measure of the interaction strength. The interaction-free case m = 0
corresponds to ϕ = ϕ0 = constant, i.e., Ṽ = ρ̃ϕ and, consequently, to w̃ = −1, equivalent to
the ΛCDM model. For m > 0 one has Q > 0 while for m < 0 the opposite, Q < 0, is valid.
The first equation in (3.17) relates the scalars Φ and ϕ without specifying a potential V (ϕ).
For the special case of an exponential potential,

V (ϕ) = V0 e−λϕ , (3.18)

JCAP11(2016)034
one has, from (2.8),
1 λ
ϕ=− ln Φ ⇒ V = V0 Φ 2K (3.19)
2K
and relation (2.7) provides us with the power-law potential
λ
U (Φ) = U0 Φ2+ 2K , U0 = 2κ2 V0 . (3.20)

With (3.13) the balance equation (3.3) for ρ̃ϕ can be written

dρ̃ϕ  
+ 3H̃ 1 + w̃eff ρ̃ϕ = 0, (3.21)
dt̃
with an effective EoS parameter
ρ̃m
w̃eff = w̃ + m , (3.22)
ρ̃ϕ
corresponding to an effective pressure p̃eff eff
ϕ = w̃ ρ̃ϕ . The interaction modifies the “bare”
EoS parameter w̃. The modification is linear in the interaction parameter m. Likewise, the
matter balance (3.2) takes the form of a conservation equation

dρ̃m
+ 3H̃ (1 − m) ρ̃m = 0 ⇒ ρ̃m = ρ̃m0 ã−3(1−m) . (3.23)
dt̃
K
Since from (3.17) one has ã = e 3m ϕ , it follows that the matter energy density can be written
in terms of the scalar field as
1−m
ρ̃m = ρ̃m0 ã−3(1−m) = ρ̃m0 e− m
Kϕ
. (3.24)

This exponential structure allows us to write

m
ρ̃m = − ρ̃m,ϕ , (3.25)
K (1 − m)

which is of interest in defining an effective potential. Then, eq. (3.3) (for p̃m = 0) is
equivalent to
d2 ϕ dϕ
2
+ 3H̃ + Ṽ,ϕeff = 0, (3.26)
dt̃ dt̃
where Ṽ eff now includes the interaction,
m
Ṽ eff = Ṽ + Ṽint ≡ Ṽ − ρ̃m . (3.27)
(1 − m)

–6–
Because of the representation in (3.24), the interaction potential is of the exponential type.
One may introduce effective quantities

1 dϕ 2 1 dϕ 2
   
eff eff eff
ρ̃ϕ = + Ṽ and p̃ϕ = − Ṽ eff , (3.28)
2 dt̃ 2 dt̃
for which, from (3.26),
dρ̃eff
ϕ
 
+ 3H̃ 1 + W̃ eff ρ̃eff
ϕ =0 (3.29)
dt̃
is valid with
m m
p̃eff
ϕ p̃ϕ + 1−m ρ̃m w̃ + 1−m r
W̃ eff =

JCAP11(2016)034
eff
= m = m . (3.30)
ρ̃ϕ ρ̃ϕ − 1−m ρ̃m 1 − 1−m r
For w̃ = −1 we have W̃ eff = −1 as well.
Together with (3.24) the total energy becomes
 
1 −3(1−m) m
ρ̃ = ρ̃m + ρ̃ϕ = ρ̃m0 ã + ρ̃ϕ0 + ρ̃m ã−3(1+w̃) . (3.31)
1 + m/w̃ w̃ + m 0
For the actual value we have ρ̃0 = ρ̃m0 + ρ̃ϕ0 . Introducing the fractional quantities
8πGρ̃m0 8πGρ̃ϕ0
Ω̃m0 = , Ω̃ϕ0 = = 1 − Ω̃m0 , (3.32)
3H̃02 3H̃02
the square of the Hubble rate is

H̃ 2 1 n
−3(1−m)
h m i
−3(1+w̃)
o
= Ω̃ m0 ã + 1 + − Ω̃ m0 ã . (3.33)
H̃02 1 + m/w̃ w̃
The non-interacting case m = 0 corresponds to a wCDM model. As already mentioned, since
then ϕ = constant, the only possibility here is w̃ = −1, i.e., the ΛCDM model.
In the following we shall focus on the case w̃ = −1 but admitting m to be non vanishing.
Under this condition H̃ 2 consists of a constant part like the ΛCDM model but modified by
the presence of m and a matter part in which the parameter m modifies the conventional
ã−3 behavior:
H̃ 2 −3
Kϕ Ω̃m0 ã Ω̃m0
= e +1− . (3.34)
H̃02 1−m 1−m
These small modifications make the Einstein-frame dynamics different from that of the
ΛCDM model. (Alternatively, this solution may be interpreted in a purely GR context
with the ϕ component belonging to the matter part of the field equations, interacting with
nonrelativistic matter.)
For the deceleration parameter

ã d H̃
q̃ = −1 − (3.35)
H̃ d ã
we find h i
(1 − 3m) Ω̃ ã−3(1−m) − 2 1 − m − Ω̃
1 m0 m0
q̃ = h i . (3.36)
2 Ω̃m0 ã−3(1−m) + 1 − m − Ω̃m0

For m = 0 its present value consistently reduces to q̃0 = −1 + 32 Ω̃m0 .

–7–
3.4 Consistency check
Now, a consistency check can be performed as follows. With (3.16) we have
 2
dϕ 9m2 2
= H̃ . (3.37)
dt̃ K2
κ2
Taking into account H̃ 2 = (ρ̃m + ρ̃ϕ ) results in
3
 2
1 dϕ
2 dt̃ −V 3m2 (2ω + 3) (ρ̃m + ρ̃ϕ ) − V
w̃ =  2 = , (3.38)
1 dϕ 3m2 (2ω + 3) (ρ̃m + ρ̃ϕ ) + V
+V

JCAP11(2016)034
2 dt̃

equivalent to
(ρ̃m +ρ̃ϕ )
6m2 (2ω + 3)
w̃ = −1 + V
(ρ̃ +ρ̃ )
= −1 + O(m2 ). (3.39)
1 + 3m2 (2ω + 3) mV ϕ
In general, our initial assumption of a constant w̃ does not seem to be compatible with the
dynamics of ϕ in (3.37). However, the corrections to the constant value w̃ = −1 are quadratic
in the interaction parameter m, there does not appear a term linear in m in (3.39). Our
approach will therefore be consistent, if we restrict ourselves to w̃ = −1 and to modifications
of the effective equation of state (3.22) which are linear in m. Since m quantifies deviations
from the ΛCDM model, one expects m to be small.

4 Jordan-frame description
With H = a1 da
dt the transformations (2.19) allow us to establish the relation between the
Hubble rates of both frames:
H = e−b (1 + 3m) H̃. (4.1)
With
a
f = eb = (4.2)
ã
and (3.11) we have
1 ã 3m
a = ã3m+1 ⇒ ã = a 3m+1 ⇒ = a− 3m+1 , (4.3)
a
i.e., the power m quantifies the difference of the scale factors in both frames. For m = 0
one has ã = a and the dynamics in both frames reduces to that of the ΛCDM model.
Combining (3.16) and (4.3) yields b in terms of the Jordan-frame scale factor a,
3m db 3m
b= ln a, = H. (4.4)
1 + 3m dt 1 + 3m
For the Hubble rate it follows that
ã
H= (1 + 3m) H̃. (4.5)
a
The matter energy densities in both frames are related by
 4
ã 12m
ρm = e−4b ρ̃m = ρ̃m0 ã−3(1−m) = a− 1+3m ρ̃m0 ã−3(1−m) = ρm0 a−3 , (4.6)
a

–8–
where we have used ρ̃m0 = ρm0 . As expected, we recover the characteristic a−3 behavior for
separately conserved non-relativistic matter in the Jordan frame.
Using (4.3) and (3.33) in (4.5) we obtain the Jordan-frame Hubble rate square
( " #)
Ω̃m0 3m+3 Ω̃m0 6m+3(1+w̃)
H 2 = (1 + 3m)2 H̃02 a− 3m+1 + 1 − a− 3m+1 . (4.7)
1 + m/w̃ 1 + m/w̃

Obviously, H02 = (1 + 3m)2 H̃02 . This implies Ω̃m0 = (1 + 3m)2 Ωm0 . Then, for w̃ = −1,

H2 6m 6m (1 + 3m)2
2 = AΩm0 a−3+ 1+3m + [1 − AΩm0 ] a− 1+3m , A≡ , (4.8)
H0 1−m

JCAP11(2016)034
or, since
6m
a 1+3m = e2b = Φ−1 , (4.9)
the square of the Hubble rate can be written as
H2 AΩm0 a−3
= + [1 − AΩm0 ] Φ. (4.10)
H02 Φ
This expression is our main result. It represents the (Jordan-frame) Hubble rate of our scalar-
tensor-theory model. From the structure of (4.10) it is obvious, how the scalar Φ modifies the
cosmological dynamics compared with the GR based standard model. For m = 0, equivalent
to Φ = 1, we recover the ΛCDM model and the Jordan-frame dynamics coincides with the
dynamics in the Einstein frame. For any m 6= 0, equivalent to Φ 6= 1, the expression (4.10)
(or (4.8)) represents a testable, alternative model with presumably small deviations from the
ΛCDM model. Notice that the modifications of the ΛCDM model are different from those
in the Einstein frame. In particular, there is no constant part even for w̃ = −1.
6m
It should be emphasized that the solution Φ = a− 1+3m is a consequence of the solution
of the macroscopic fluid dynamical equations for the energy densities under the assump-
tion (3.11). It is not a solution of equation (2.13), which would require an explicit expression
for the potential U . The unknown exact solution of the scalar field equation (2.13) is replaced
6m
here by the effective solution Φ = a− 1+3m which was obtained using the approximations (3.11)
and (3.15) in the macroscopic fluid dynamics. Our procedure allows us to obtain an explicit
expression for Φ and for the Hubble rate without solving the scalar field equation (2.13).
This can be seen as a major advantage of the approach. By direct calculation one verifies
that Φ obeys the simplified effective equation of motion
d2 Φ dΦ m
= −9H02 (1 + m) AΩm0 a−3 + 2 (1 − AΩm0 ) Φ2
 
2
+ 3H 2 (4.11)
dt dt (1 + 3m)
instead of (2.13). Since
1+3m
a−3 = Φ 2m , (4.12)
equation (4.11) is of the form
d2 Φ dΦ eff
+ 3H + U,Φ = 0, (4.13)
dt2 dt
where
eff m h 1+3m
i
U,Φ = 9H02 (1 + m) AΩ m0 Φ 2m + 2 (1 − AΩ
m0 ) Φ2
. (4.14)
(1 + 3m)2

–9–
This means, the dynamics of our Φ is that of a scalar field with potential U eff . For m = 0
eff vanishes, corresponding to a constant effective potential and compatible
the derivative U,Φ
with Φ = 1, thus recovering the ΛCDM model. It should be kept in mind, however, that the
dynamics of Φ describes deviations from the ΛCDM model, not this model itself.
The total energy density corresponding to (4.10) may be written as
ρm
ρ=A + ρ0 (1 − AΩm0 ) Φ, (4.15)
Φ
or, separating the matter part as in (2.11)
ρm ρm

JCAP11(2016)034
ρ= + (A − 1) + ρ0 [1 − AΩm0 ] Φ. (4.16)
Φ Φ
The appearance of the scalar Φ in these expressions changes the relative contributions of
matter and the dark-energy equivalent. Our model encodes the deviations of the scalar-
tensor description from the ΛCDM model entirely in the constant parameter m which is
supposed to be small. To be more definite we shall assume |m| < 31 . Under this condition Φ
decays with a for m > 0 while it increases for m < 0. It is only exactly at the present epoch
a = 1 that Φ(a = 1) = 1. For m > 0 we have Φ > 1 in the past (a < 1), for m < 0, on the
other hand, Φ increases from Φ < 1 at a < 1 to the present Φ(a = 1) = 1. For a  1 the
energy density approaches
6m
ρ ≈ Aρm0 a−3 a 1+3m (a  1). (4.17)

In the far-future limit a  1 we have

6m
ρ ≈ ρ0 (1 − AΩm0 ) a− 1+3m (a  1). (4.18)

Depending on the sign of m it may either decay (m > 0) or increase (m < 0).
The matter fraction becomes
ρm Ωm0 a−3
Ωm = = . (4.19)
ρ AΦ−1 Ωm0 a−3 + [1 − AΩm0 ] Φ

We may introduce an effective GR description by defining a component ρx = ρ − ρm ,

ρx 6m
h 6m
i
= [1 − AΩm0 ] a− 1+3m + Ωm0 a−3 A a 1+3m − 1 , (4.20)
ρ0

equivalent to  
A
ρx = − 1 ρm + ρ0 [1 − AΩm0 ] Φ (4.21)
Φ
with the fractional contribution
AΦ−1 − 1 Ωm0 a−3 + [1 − AΩm0 ] Φ


ρx
Ωx = = . (4.22)
ρ AΦ−1 Ωm0 a−3 + [1 − AΩm0 ] Φ

Then, at high redshift,

ρm 1 6m ρx 1 6m
Ωm = ≈ a− 1+3m , Ωx = ≈ 1 − a− 1+3m (a  1). (4.23)
ρ A ρ A

– 10 –
For m > 0 the effective energy density ρx becomes negative for small values of a. While the
combination 1 − AΩm0 remains always positive for small values of m and Ωm0 of the order
6m
of 0.3, the combination A a 1+3m − 1 will change its sign with the increase of the scale factor.
This sign change will occur at a value ac , given by
  1+3m
6m
1+3m 1−m 6m
A ac −1=0 ⇒ ac = . (4.24)
(1 + 3m)2

It is straightforward to check that values O |m| & 0.1 drastically change the dynamics com-
6m
pared with that of the standard model. Namely, for m > 0 the combination A a 1+3m − 1 is

JCAP11(2016)034
negative for small values of a. With increasing a one obtains a change to positive values at
ac ≈ 0.25 for m = 0.1, corresponding to a redshift zc ≈ 3 and at ac ≈ 0.31 (redshift zc ≈ 2.2)
6m
for m = 0.01. For m < 0 the term A a 1+3m − 1 is positive for a  1 but will change the
sign as well. For m = −0.1 this happens at ac ≈ 0.39 (redshift zc ≈ 1.6), for m = −0.01
the corresponding values are ac ≈ 0.08 (zc ≈ 11.6). The behavior of the fractional energy
densities is shown in figure 1. The right panels use logarithmic units showing the asymptotic
values of Ωm and Ωx at early times. It should be emphasized again that the “energy density”
ρx does not represent a material substratum, it is of geometric origin. The fact that the
quantity ρx may become negative does not jeopardize the model. What matters here is the
Hubble rate which is always well behaved.
To complete the analogy, the component ρx is supposed to obey the conservation
equation
dρx
+ 3H (1 + wx ) ρx = 0 (4.25)
dt
dρx dρx da
with an effective EoS parameter wx . With dt = da dt ≡ ρ0x aH we obtain

1 ρ0x a
ρ0x a + 3 (1 + weff ) ρx = 0 ⇒ 1 + wx = − . (4.26)
3 ρx
A direct calculation yields
h i
2m
1+3m [1 − AΩm0 ] Φ + Ωm0 a−3 1+m
1+3m AΦ
−1 −1
wx = −1 + . (4.27)
[1 − AΩm0 ] Φ + Ωm0 a−3 [AΦ−1 − 1]

Equation (4.27) establishes a relation between the constant, “bare” EoS parameter w̃ = −1 in
the Einstein frame and the generally time-dependent effective Jordan-frame EoS parameter
wx . For m = 0 we recover wx = w̃ = −1. At high redshift
h i
1+m
AΦ −1 − 1
1+3m
wx ≈ −1 + (a  1). (4.28)
[AΦ−1 − 1]
The value of wx may be close to zero at high redshift, i.e., mimicking dust, but the effective
energy density will be negative for m > 0. Already a rather small value of |m| 6= 0 will
substantially change the behavior of this dark-energy equivalent compared with the standard-
model dark energy. The present value of the effective EoS parameter is
2m
1+3m + 3mΩm0
wx = −1 + (a = 1). (4.29)
1 − Ωm0

– 11 –
 1.0 1

m = -0.001 Ωm
0.50
Ωm m = -0.01
0.8

0.6
Ωx
0.10
Ω Ω
0.4 0.05

0.2
Ωx
m = -0.01

JCAP11(2016)034
0.01
m = -0.001
0.0
0.0 0.2 0.4 0.6 0.8 1.0 10-4 0.001 0.010 0.100 1

a a
1.2
1

1.0 Ωm
m = 0.001 0.50
Ωm m = 0.01
0.8

0.6
Ωx
Ω Ω 0.10
0.4
0.05

0.2 Ωx

0.0
m = 0.01
0.01
m = 0.001
-0.2
0.0 0.2 0.4 0.6 0.8 1.0 10-4 0.001 0.010 0.100 1

a a

Figure 1. Matter fraction Ωm of the total energy and geometric energy fraction Ωx for negative (top
panels) and positive (bottom panels) values of m. Logarithmic units are used in the right panels.
This allows for a better visualization for the asymptotic values at early times.

Given that |m| is small, this remains in the vicinity of wx = −1. In the far future wx
approaches
2m
wx ≈ −1 + (a  1). (4.30)
1 + 3m
The evolution of the effective EoS parameter wx is visualized in figure 2 for different values
of m. For small values of a one has wx > 0 but there will be a change to wx < 0 well before
the present epoch. The effective energy density changes from the phantom regime to a later
phase with ρx > 0. Due to the sign change in ρx the transition point ρx = 0 is accompanied
by a singularity in wx at this point for m positive.
0
The result for the deceleration parameter q = −1 − a HH is
1−3m −3 −1 − 2
1 1+3m AΩm0 a Φ 1+3m [1 − AΩm0 ] Φ
q= . (4.31)
2 AΩm0 a−3 Φ−1 + [1 − AΩm0 ] Φ
As shown in figure 3, it changes from a high-redshift value
1 1 − 3m
q≈ (a  1), (4.32)
2 1 + 3m

– 12 –
 0.5

0.0
0.0

m = -0.01 m = 0.01
m = -0.001 m = 0.001
wx -0.5 wx -0.5

-1.0
-1.0

JCAP11(2016)034
-1.5
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0

a a

Figure 2. Effective EoS parameter as function of the scale factor for negative (left panel) and positive
(right panel) values of m.

0.6 0.6
EdS EdS

0.4 0.4

0.2 0.2

q 0.0 q 0.0

-0.2 -0.2

m = -0.01 m = 0.01
-0.4 -0.4
m = -0.001 m = 0.001

-0.6 -0.6

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0

a a

Figure 3. Deceleration parameter for various negative (left panel) and positive (right panel)
values of m.

close to 12 , to the value close to −1,

1
q≈− (a  1), (4.33)
1 + 3m

at late times.
The explicit expression for the pressure is
 
2m 1+m
px = weff ρx = −ρx + ρ0 [1 − AΩm0 ] Φ + ρ0 Ωm0 a−3 AΦ−1 − 1 . (4.34)
1 + 3m 1 + 3m

– 13 –
 1.5

1.0
m = -0.01
1.0
m = -0.001 m = 0.01
m = 0.001
0.5
0.5
 
px px
 0.0  0.0
ρx ρx
-0.5
-0.5

-1.0

JCAP11(2016)034
-1.0

-1.5
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0

a a

Figure 4. Effective adiabatic sound speed for various negative (left panel) and positive (right panel)
values of m.

6m
Differentiating and using Φ̇ = − 1+3m HΦ yields
( 2 " 2 #)
2m −3 1+m −1
ṗx = −ρ̇x − 3Hρ0 [1 − AΩm0 ] Φ + Ωm0 a AΦ −1 . (4.35)
1 + 3m 1 + 3m

Since
  
2m 1+m
ρ˙x = −3Hρ0 [1 − AΩm0 ] Φ + Ωm0 a−3 AΦ−1 − 1 , (4.36)
1 + 3m 1 + 3m

we find  
 2 2
2m
1+3m [1 − AΩm0 ] Φ + Ωm0 a−3 1+m
1+3m AΦ−1
−1
ṗx
= −1 + h i . (4.37)
ρ˙x 2m
[1 − AΩm0 ] Φ + Ωm0 a−3 1+m
AΦ −1 − 1
1+3m 1+3m

At high redshift this quantity (which corresponds to the adiabatic sound speed) is consider-
ably smaller than 1, it may even be close to zero. The far-future limit is

ṗx 2m
≈ −1 + (a  1). (4.38)
ρ˙x 1 + 3m

While this does not seem to differ substantially from the standard model, the intermediate
behavior does. As visualized in figure 4, this quantity changes its sign at the points with
ρ̇x = 0 which implies a singularity in ρṗ˙xx . Recall that the energy density ρx is an effective
quantity which simulates DE but does not belong to the matter part of the field equations.
It is of entirely geometric origin.
We may define a total EoS parameter w ≡ ρp = pρx which results in

p px 1 (1 + m) [1 − AΩm0 ] Φ + 2mAΦ−1 Ωm0 a−3

w≡ = =− . (4.39)
ρ ρ 1 + 3m [1 − AΩm0 ] Φ + AΦ−1 Ωm0 a−3

– 14 –
 0.2 0.2

0.0 0.0

-0.2 -0.2

m = -0.01 m = 0.01
w -0.4 w -0.4
m = -0.001 m = 0.001

-0.6 -0.6

JCAP11(2016)034
-0.8 -0.8

-1.0 -1.0
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0

a a

Figure 5. Total EoS parameter for various negative (left panel) and positive (right panel) values of m.

The limits are

2m
w≈− (a  1) (4.40)
1 + 3m
and
1+m
w≈− (a  1), (4.41)
1 + 3m
i.e., at high redshift w is close to zero, in the far-future limit it is close to −1. The total EoS
parameter is visualized in figure 5. It is well behaved for all values of m.
With
ρ+p 1+m 2m
= AΦ−1 Ωm0 a−3 + [1 − AΩm0 ] Φ, (4.42)
ρ0 1 + 3m 1 + 3m
 
ρ̇ 1+m −1 −3 2m ρ+p
= −3H AΦ Ωm0 a + [1 − AΩm0 ] Φ = −3H (4.43)
ρ0 1 + 3m 1 + 3m ρ0

and
ṗ ṗx 2m (1 + m)  −3 −1

= = 3H (1 − AΩ m0 ) Φ + Ω m0 a AΦ (4.44)
ρ0 ρ0 (1 + 3m)2
we find
ṗ 2m (1 + m) (1 − AΩm0 ) Φ + Ωm0 a−3 AΦ−1
=− . (4.45)
ρ̇ 1 + 3m 2m [1 − AΩm0 ] Φ + (1 + m) AΦ−1 Ωm0 a−3
This effective adiabatic sound speed of the cosmic medium as a whole has the high-redshift
limit
ṗ 2m
≈− (a  1), (4.46)
ρ̇ 1 + 3m
i.e., a very small value, and
ṗ 1+m
≈− (a  1), (4.47)
ρ̇ 1 + 3m
in the far future, a value close to −1.

– 15 –
 0.10 0.10

m = 0.001
0.05 0.05
m = 0.01
 
p p
 0.00  0.00
ρ ρ
m = -0.01
-0.05 -0.05
m = -0.001

-0.10 -0.10
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0

a a

JCAP11(2016)034
Figure 6. Scale-factor dependence of the total effective adiabatic sound speed for various negative
(left panel) and positive (right panel) values of m.

While for m > 0 the derivative ṗ is always positive and ρ̇ is always negative, the quantity
ṗ
ρ̇remains negative in the entire range. For m < 0, on the other hand, ṗ is always negative
but ρ̇ may change its sign for sufficiently large values of a, which gives rise to a divergency
in ρ̇ṗ at the critical value
  1 1−3|m|
(1 − |m|) AΩm0 3 1+|m|
as = (m < 0). (4.48)
2|m| (1 − AΩm0 )
For a < as the ratio ρ̇ṗ is positive, for a > as it is negative since ρ starts to grow at a = as .
This is consistent with the far-future limit in (4.18). Except for the rather large deviation
from the standard model with m = −0.1, these singularities occur in the distant future.
Figure 6 shows the dependence of ρ̇ṗ on the scale factor for various values of m.
We finish this section with a simplified stability check of the solution for Φ, leaving a true
perturbation analysis to be the subject of a subsequent paper. We introduce a decomposition
Φ = Φb + Φ1 (4.49)
with our “background” solution Φb and a (homogeneous) perturbation Φ1 according to
6m
Φb = a− 1+3m , Φ1  Φb . (4.50)
Linearizing yields
1+3m 1+3m 1+3m 1 + 3m 1+m
Φ 2m = (Φb + Φ1 ) 2m ≈ Φb 2m + Φb 2m Φ1 (4.51)
2m
and
Φ2 = (Φb + Φ1 )2 ≈ Φ2b + 2Φb Φ1 . (4.52)
Introducing these decompositions into (4.11) provides us, at first order, with an equation
for Φ1 ,
d2 Φ1
 
dΦ1 2 1 1+m
1+m 4m
+ 3H + 9H0 AΩm0 Φb 2m
+ (1 − AΩm0 ) Φb Φ1 = 0. (4.53)
dt2 dt 2 1 + 3m (1 + 3m)2
This is an equation for a damped harmonic oscillator with a time-dependent frequency ω,
given by  
2 1 1+m 4m
1+m
2
ω = 9H0 AΩm0 Φb 2m
+ (1 − AΩm0 ) Φb . (4.54)
2 1 + 3m (1 + 3m)2

– 16 –
 Observationally
Allowed ( 2σ )
Φb <1
0.3

Stable Φb =1
0.2 (ω 2 >0)

ω2 0.1

JCAP11(2016)034
0.0

Unstable (ω 2 <0)

-0.1

-0.04 -0.02 0.00 0.02 0.04

Figure 7. Dependence of ω 2 on m for the present value Φb = 1 (solid line) and for a constant value
Φb < 1 (broken line) which, for m < 0, corresponds to a scale factor in the past, i.e. a < 1. The
instability range (m-values left to the intersection of the Φb curve with the axis ω 2 = 0) shrinks
towards the past, since the intersection point moves to the left for Φb < 1. The grey stripe marks
the instability region. The observationally allowed values lie entirely in the stable part of the ω 2 -m
plane. For m > stability is always guaranteed.

The perturbation Φ1 is bounded, i.e. there is no instability for any ω 2 > 0. For positive
values of m this is always guaranteed. For m < 0 the second term in the square brackets had
to dominate the first one to violate the condition ω 2 > 0. But since we imposed |m|  1 this
may happen only under the extreme condition a > acr where acr is a critical value, given by

1+|m|
3 1 − 3|m| AΩm0
acr1−3|m| = (1 − |m|) . (4.55)
8|m| 1 − AΩm0

It follows that an instability may occur at the most in the far-future limit a > acr where
acr  1 due to the inverse dependence on |m|. This value of acr is of the same order as the
previously derived critical scale factor (4.48). In figure 7 we depict the dependence (4.54) of
ω 2 on m for the present value Φb = 1 (solid line) and for a constant value Φb < 1 (broken
line) which corresponds to a scale factor in the past, i.e. a < 1, for m < 0. The gray strip
represents the unstable part ω 2 < 0. The observationally allowed region for m (see the
statistical analysis of the following section) lies entirely in the stable range, even for the
admitted negative m-values. Towards the past, the range of admissible negative m-values
increases since the intersection of the broken line with ω 2 = 0 is shifted to the left.
With the help of the decomposition (4.49) the Hubble rate (4.10) may be split into
a background part Hb and linear, homogeneous perturbations H1 about this background
1+3m
as well. Since with a−3 = Φ 2m the time evolution of the Hubble rate (4.10) is entirely
determined by the dynamics of Φ, we find H1 ∝ Φ1 , i.e., the perturbed Hubble rate inherits
the stability properties of the scalar field perturbations.

– 17 –
5 Statistical (observational) analysis
The purpose of this section is to test the viability of the background expansion predicted
by (4.8), equivalent to (4.10), based on the available observations. A particular interest here
is to obtain constraints on the parameter m which is responsable to dictating deviations of
our model from ΛCDM.
In order to compare (4.8) (or (4.10)) with the data we constrain the model parameters
with the following observational data sets.
Supernovae. First, we use Supernovae data from the JLA compilation [23]. We will use
the binned data set provided by reference [23] with the corresponding covariance matrix C.

JCAP11(2016)034
This test is based on the observed distance modulus µobs (z) of each binned SN Ia data at a
certain redshift z,
dL (z)
µth (z) = 25 + 5 log10 , (5.1)
Mpc
where the luminosity distance, in a spatially flat FLRW metric, is given by the formula
dz 0
Z z
dL (z) = c(1 + z) 0
. (5.2)
0 H(z )
Knowledge of the Hubble expansion rate allows us to compute the predicted theoretical value
µth (zi ) for a given redshift zi . The binned JLA data contains 31 data-points. For the JLA
Supernova sample the χ2 function is constructed according to
χ2SN = (µth (z) − µobs (z))† C −1 (µth (z) − µobs (z)) . (5.3)
Differential age. A second observational source comes from the evaluation of differential
age data of old galaxies that have evolved passively [24–26]. Indeed, the expansion rate is
defined as
1 dz
H(z) = − . (5.4)
1 + z dt
Since spectroscopic redshifts of galaxies are known with very high accuracy, one just needs
a differential measurement of time dt at a given redshift interval in order to obtain values
for H(z). The data used in this work consist of 28 data points listed in [27], but previously
compiled in [28].
Baryon acoustic oscillations. The baryon acoustic oscillation (BAO) scale is calcu-
lated via
cz 1/3
 
2 2
DV (z) = (1 + z) DA (z) , (5.5)
H(z)
where DA (z) is the angular-diameter distance. The values for DV have been reported in the
literature by several galaxy surveys. In our analysis we use data at z = 0.2 and z = 0.35
from the SDSS [29, 30], data at z = 0.44, 0.6 and 0.73 from the WiggleZ [31] and one data
point at z = 0.106 from the 6dFGRS [32] surveys.
For our statistical analysis we construct the chi-square function of each sample
N
2
2
X f th (zi ) − f obs (zi )
χ = , (5.6)
i=1
σi2
where f = (H, DV ) for the H(z) and BAO datasets, respectively. The number of data points
zi , f obs (zi ) in each set is, respectively, NH and NBAO whereas σi is the observational


– 18 –
 0.20 0.20
35

0.15 H0 = 67.8 km/s/Mpc 0.15 H0 = 73.2 km/s/Mpc

30

0.10 0.10
25

0.05 
0.05

PDF
20

m
m

ΛCDM ΛCDM
0.00 0.00
15


-0.05 -0.05
10

-0.10 -0.10
5

-0.15 -0.15
0
0.1 0.2 0.3 0.4 0.5 0.6 0.1 0.2 0.3 0.4 0.5 0.6 -0.04 -0.02 0.00 0.02 0.04 0.06

Ωm0 Ωm0 m

JCAP11(2016)034
Figure 8. Observational constraints on the model parameters. In the left panel we fix the hyper-
surface with the Planck prior H0 = 67.8 km/s/Mpc and display 1σ and 2σ confidence level contours
in the plane Ωm0 × m. The central panel displays the same contours with the recent HST prior
H0 = 73.2 km/s/Mpc. H(z) data only is represented by red regions, BAO data in green and Super-
novae data in blue. The contours obtained from the total χ2 are denoted by the black solid lines.
In the right panel we display the one-dimensional probability distribution function (PDF) for the
parameter m after marginalization over H0 and Ωm0 .

error associated to each observation f obs and f th is the theoretical value predicted by the
scalar-tensor model.
Adding up information from all data samples we can construct the total chi-square
function as
χ2Total = χ2SN + χ2H + χ2BAO . (5.7)
We consider the expansion rate (4.8) as a model with three free parameters, H0 , Ωm0
and m. Our main interest is in constraints on the latter.
We will fix two hypersurfaces for the parameter H0 : the Planck prior H0 =
68.7 Km/s/Mpc [33] and the recent determination from the Hubble Space Telescope (HST)
H0 = 73.2 Km/s/Mpc [34]. These priors on H0 allows us to span the Ωm0 × m plane. This
can be seen in the left and central panels in figure 8 where the 1σ and 2σ confidence regions
are shown in red for the H(z) data, blue for the Supernovae data and green for the BAO data,
respectively, for both H0 priors. The combined contours obtained from the total chi-square
function (5.7) are given by the solid black lines. From these plots it is clear that the preferred
m-values depend on the H0 prior. With the Planck prior positive values of m are preferred,
while the HST prior results in a preference for negative m.
It is worth noting that in both two-dimensional plots the ΛCDM model, expressed
by the horizontal line at m = 0, lies outside the 2σ confidence level region of the total
(χ2Total ) function.
In order to obtain specific constraints on m we apply Bayesian statistical analysis.
With this procedure we obtain a one-dimensional probability distribution function (PDF)
after marginalizing the likelihood function
2
L = Ae−χTotal (H0 ,Ωm0 ,m)/2 (5.8)
over the parameters H0 and Ωm0 .
The resulting PDF for the parameter m is shown in the right panel of figure 8. Although
positive values for the parameter m are slightly preferred, the peak of the distribution has
+0.011(1σ) +0.017(2σ)
been obtained at m = 0.004−0.011(1σ) −0.017(2σ) which indicates the agreement of the model
with the ΛCDM model.

– 19 –
6 Discussion

We have established a class of scalar-tensor-theory-based cosmological models which are

simple extensions of the ΛCDM model. The background dynamics of this class has been
discussed in detail. Our main result is an analytic expression (4.10) for the Hubble rate which
explicitly quantifies the difference to the standard model through a constant parameter m
which determines the dynamics of the scalar field Φ. The solution for Φ is a consequence of the
solution of the macroscopic fluid dynamics. It corresponds to a scalar-field dynamics (4.13)
with an effective potential given by (4.14). We identified a geometric equivalent of the DE
component of the standard model. The corresponding effective energy density may be positive
or negative, including a transition between both regimes. The effective EoS parameter of

JCAP11(2016)034
this component diverges at the transition point but the overall dynamics is well behaved. A
similar comment holds for the effective adiabatic sound speed square of the geometrical DE.
Such behavior is unavoidable in any model in which the energy density of an DE equivalent
and its time derivative change their signs during the cosmic evolution. This seems to limit
the usefulness of an effective fluid description of parts of the geometrical sector and may
cause computational problems. But since the overall dynamics and the dynamics of the
matter component are smooth, the mentioned apparently unwanted features do not really
jeopardize the model. They just demonstrate the fact that the fluid picture for parts of the
geometry is a formal description which may well differ from that of a real fluid.
Our tests of the model parameter m with data from SNIa, H(z) and BAO constrain
m to values very close to the ΛCDM value m = 0 with a slight preference for positive
values. We expect the existence of the analytic background solution (4.10) to be useful for
future investigation of the perturbation dynamics of this scalar-tensor extension of the ΛCDM
model. Even a very small non-vanishing value of |m| will certainly modify the standard
scenario of structure formation.

Acknowledgments

Partial financial support by CNPq, CAPES and FAPES is gratefully acknowledged.

References
[1] S.M. Carroll, V. Duvvuri, M. Trodden and M.S. Turner, Is cosmic speed-up due to new
gravitational physics?, Phys. Rev. D 70 (2004) 043528 [astro-ph/0306438] [INSPIRE].
[2] S. Nojiri and S.D. Odintsov, Where new gravitational physics comes from: M-theory?, Phys.
Lett. B 576 (2003) 5 [hep-th/0307071] [INSPIRE].
[3] S. Nojiri and S.D. Odintsov, Introduction to modified gravity and gravitational alternative for
dark energy, eConf C 0602061 (2006) 06 [Int. J. Geom. Meth. Mod. Phys. 4 (2007) 115]
[hep-th/0601213] [INSPIRE] [INSPIRE].
[4] R. Gannouji, D. Polarski, A. Ranquet and A.A. Starobinsky, Scalar-Tensor Models of Normal
and Phantom Dark Energy, JCAP 09 (2006) 016 [astro-ph/0606287] [INSPIRE].
[5] E.J. Copeland, M. Sami and S. Tsujikawa, Dynamics of dark energy, Int. J. Mod. Phys. D 15
(2006) 1753 [hep-th/0603057] [INSPIRE].
[6] D.F. Torres, Quintessence, superquintessence and observable quantities in Brans-Dicke and
nonminimally coupled theories, Phys. Rev. D 66 (2002) 043522 [astro-ph/0204504] [INSPIRE].
[7] F.S.N. Lobo, The dark side of gravity: modified theories of gravity, arXiv:0807.1640 [INSPIRE].

– 20 –
 [8] R.R. Caldwell and M. Kamionkowski, The Physics of Cosmic Acceleration, Ann. Rev. Nucl.
Part. Sci. 59 (2009) 397 [arXiv:0903.0866] [INSPIRE].
[9] R. Catena, M. Pietroni and L. Scarabello, Einstein and Jordan reconciled: a frame-invariant
approach to scalar-tensor cosmology, Phys. Rev. D 76 (2007) 084039 [astro-ph/0604492]
[INSPIRE].
[10] V. Faraoni and S. Nadeau, The (pseudo)issue of the conformal frame revisited, Phys. Rev. D
75 (2007) 023501 [gr-qc/0612075] [INSPIRE].
[11] S. Capozziello, S. Nojiri, S.D. Odintsov and A. Troisi, Cosmological viability of f(R)-gravity as
an ideal fluid and its compatibility with a matter dominated phase, Phys. Lett. B 639 (2006)
135 [astro-ph/0604431] [INSPIRE].

JCAP11(2016)034
[12] A.D. Dolgov and M. Kawasaki, Can modified gravity explain accelerated cosmic expansion?,
Phys. Lett. B 573 (2003) 1 [astro-ph/0307285] [INSPIRE].
[13] T. Chiba, 1/R gravity and scalar-tensor gravity, Phys. Lett. B 575 (2003) 1
[astro-ph/0307338] [INSPIRE].
[14] R. Durrer and R. Maartens, Dark Energy and Modified Gravity, in Dark Energy: Observational
& Theoretical Approaches, P. Ruiz-Lapuente ed., Cambridge University Press (2010), pg. 48,
arXiv:0811.4132 [INSPIRE].
[15] T.P. Sotiriou and V. Faraoni, f(R) Theories Of Gravity, Rev. Mod. Phys. 82 (2010) 451
[arXiv:0805.1726] [INSPIRE].
[16] T.P. Sotiriou, 6+1 lessons from f(R) gravity, J. Phys. Conf. Ser. 189 (2009) 012039
[arXiv:0810.5594] [INSPIRE].
[17] A.W. Brookfield, C. van de Bruck and L.M.H. Hall, Viability of f(R) Theories with Additional
Powers of Curvature, Phys. Rev. D 74 (2006) 064028 [hep-th/0608015] [INSPIRE].
[18] N. Agarwal and R. Bean, The Dynamical viability of scalar-tensor gravity theories, Class.
Quant. Grav. 25 (2008) 165001 [arXiv:0708.3967] [INSPIRE].
[19] C.E.M. Batista and W. Zimdahl, Power-law solutions and accelerated expansion in
scalar-tensor theories, Phys. Rev. D 82 (2010) 023527 [arXiv:0912.0998] [INSPIRE].
[20] T. Clifton, P.G. Ferreira, A. Padilla and C. Skordis, Modified Gravity and Cosmology, Phys.
Rept. 513 (2012) 1 [arXiv:1106.2476] [INSPIRE].
[21] T. Chiba and M. Yamaguchi, Conformal-Frame (In)dependence of Cosmological Observations
in Scalar-Tensor Theory, JCAP 10 (2013) 040 [arXiv:1308.1142] [INSPIRE].
[22] A. Joyce, L. Lombriser and F. Schmidt, Dark Energy vs. Modified Gravity, Ann. Rev. Nucl.
Part. Sci. 66 (2016) 95 [arXiv:1601.06133] [INSPIRE].
[23] SDSS collaboration, M. Betoule et al., Improved cosmological constraints from a joint analysis
of the SDSS-II and SNLS supernova samples, Astron. Astrophys. 568 (2014) A22
[arXiv:1401.4064] [INSPIRE].
[24] R. Jimenez and A. Loeb, Constraining cosmological parameters based on relative galaxy ages,
Astrophys. J. 573 (2002) 37 [astro-ph/0106145] [INSPIRE].
[25] R. Jimenez, L. Verde, T. Treu and D. Stern, Constraints on the equation of state of dark
energy and the Hubble constant from stellar ages and the CMB, Astrophys. J. 593 (2003) 622
[astro-ph/0302560] [INSPIRE].
[26] D. Stern, R. Jimenez, L. Verde, M. Kamionkowski and S.A. Stanford, Cosmic Chronometers:
Constraining the Equation of State of Dark Energy. I: H(z) Measurements, JCAP 02 (2010)
008 [arXiv:0907.3149] [INSPIRE].
[27] O. Farooq, D. Mania and B. Ratra, Hubble parameter measurement constraints on dark energy,
Astrophys. J. 764 (2013) 138 [arXiv:1211.4253] [INSPIRE].

– 21 –
[28] M. Moresco, L. Verde, L. Pozzetti, R. Jimenez and A. Cimatti, New constraints on
cosmological parameters and neutrino properties using the expansion rate of the Universe to
z ∼ 1.75, JCAP 07 (2012) 053 [arXiv:1201.6658] [INSPIRE].
[29] SDSS collaboration, W.J. Percival et al., Baryon Acoustic Oscillations in the Sloan Digital Sky
Survey Data Release 7 Galaxy Sample, Mon. Not. Roy. Astron. Soc. 401 (2010) 2148
[arXiv:0907.1660] [INSPIRE].
[30] N. Padmanabhan et al., A 2% distance to z = 0.35 by reconstructing baryon acoustic
oscillations — I. Methods and application to the Sloan Digital Sky Survey, Mon. Not. Roy.
Astron. Soc. 427 (2012) 2132 [arXiv:1202.0090] [INSPIRE].
[31] C. Blake et al., The WiggleZ Dark Energy Survey: mapping the distance-redshift relation with

JCAP11(2016)034
baryon acoustic oscillations, Mon. Not. Roy. Astron. Soc. 418 (2011) 1707 [arXiv:1108.2635]
[INSPIRE].
[32] F. Beutler et al., The 6dF Galaxy Survey: Baryon Acoustic Oscillations and the Local Hubble
Constant, Mon. Not. Roy. Astron. Soc. 416 (2011) 3017 [arXiv:1106.3366] [INSPIRE].
[33] Planck collaboration, P.A.R. Ade et al., Planck 2015 results. XIII. Cosmological parameters,
Astron. Astrophys. 594 (2016) A13 [arXiv:1502.01589] [INSPIRE].
[34] A.G. Riess et al., A 2.4% Determination of the Local Value of the Hubble Constant, Astrophys.
J. 826 (2016) 56 [arXiv:1604.01424] [INSPIRE].

– 22 –