Reconstruction of f (T ) gravity: Rip cosmology, finite-time future

singularities and thermodynamics

Kazuharu Bamba1, , Ratbay Myrzakulov2, ,

Shinichi Nojiri1,3, and Sergei D. Odintsov3,4,5,6,
1
Kobayashi-Maskawa Institute for the Origin of Particles and the Universe,
Nagoya University, Nagoya 464-8602, Japan
arXiv:1202.4057v3 [gr-qc] 12 Apr 2012

2
Eurasian International Center for Theoretical Physics,
Eurasian National University, Astana 010008, Kazakhstan
3
Department of Physics, Nagoya University, Nagoya 464-8602, Japan
4
Institucio Catalana de Recerca i Estudis Avancats (ICREA), Barcelona, Spain
5
Institut de Ciencies de lEspai (CSIC-IEEC),
Campus UAB, Facultat de Ciencies, Torre C5-Par-2a pl,
E-08193 Bellaterra (Barcelona), Spain
6
Tomsk State Pedagogical University, Tomsk, Russia

E-mail address: bamba@kmi.nagoya-u.ac.jp

E-mail addresses: rmyrzakulov@csufresno.edu; rmyrzakulov@gmail.com

E-mail address: nojiri@phys.nagoya-u.ac.jp

E-mail address: odintsov@ieec.uab.es

1
 Abstract
We demonstrate that there appear finite-time future singularities in f (T ) gravity with T being
the torsion scalar. We reconstruct a model of f (T ) gravity with realizing the finite-time future
singularities. In addition, it is explicitly shown that a power-low type correction term T ( > 1)
such as a T 2 term can remove the finite-time future singularities in f (T ) gravity. Moreover, we study
f (T ) models with realizing inflation in the early universe, the CDM model, Little Rip cosmology
and Pseudo-Rip cosmology. It is demonstrated that the disintegration of bound structures for
Little Rip and Pseudo-Rip cosmologies occurs in the same way as in gravity with corresponding
dark energy fluid. We also discuss that the time-dependent matter instability in the star collapse
can occur in f (T ) gravity. Furthermore, we explore thermodynamics in f (T ) gravity and illustrate
that the second law of thermodynamics can be satisfied around the finite-time future singularities
for the universe with the temperature inside the horizon being the same as that of the apparent
horizon.

PACS numbers: 04.50.Kd, 95.36.+x, 98.80.-k

2
I. INTRODUCTION

A number of cosmological observations, e.g., Type Ia Supernovae [1], cosmic microwave

background (CMB) radiation [2, 3], large scale structure (LSS) [4], baryon acoustic oscilla-
tions (BAO) [5], and weak lensing [6], have been implied the current accelerating expansion
of the universe. Approaches to account for the late time cosmic acceleration are classified
into two representative categories. The first is the introduction of unknown matters, i.e.,
so-called dark energy, in the framework of general relativity (for recent reviews, see [7, 8]).
The second is the modification of gravity such as f (R) gravity (for recent reviews, see [913]).
As a gravitational theory beyond general relativity, one could consider teleparallelism
with the Weitzenbock connection, which has torsion T and not the curvature R defined by
the Levi-Civita connection [14]. In modern cosmology, in order to explain both inflation [15]
and the late time accelerated expansion of the universe, the teleparallel Lagrangian density
represented by the torsion scalar T has been extended to a function of T as f (T ) [16, 17].
This idea is equivalent to the concept of f (R) gravity. In the recent literature, to check
whether f (T ) gravity can be an alternative gravitational theory to general relativity, its
various properties have been diversely explored [1823], e.g., the local Lorentz invariance [21],
non-trivial conformal frames and thermodynamics [22, 23].
Moreover, it is known that if (phantom/quintessence) dark energy dominates the universe,
in general there can appear finite-time future singularities, which have been classified into
four types [24]. The finite-time future singularities in f (R) gravity [25] have first been ob-
served and those in various modified gravity [26, 27] have also been investigated. Therefore,
it is important to examine models of f (T ) gravity in which finite-time future singularities
can exist.
In this paper, we concentrate on the two important theoretical features of f (T ) gravity:
the finite-time future singularities and thermodynamics in f (T ) gravity. First, we explicitly
reconstruct f (T ) gravity in which the finite-time future singularities appear by following the
procedure proposed in Refs. [2830]. We also study a correction term to the models of f (T )
gravity, so that such a term can remove the finite-time future singularities in analogy with
f (R) gravity. Moreover, we explore the reconstruction of f (T ) models which realize the
examples of inflation in the early universe, the CDM model, Little Rip cosmology [3037]
and Pseudo-Rip cosmology [38]. The Little Rip scenario is a kind of a mild phantom scenario

3
and considered in order to avoid a Big Rip singularity. On the other hand, the Pseudo-
Rip model is an intermediate case between the cosmological constant and the Little Rip
cosmology. In this model, the Hubble parameter asymptotically becomes constant as time
goes to infinity, although for the Big Rip singularity, the Hubble parameter diverges at finite
time, and in the Little Rip cosmology the Hubble parameter becomes infinity asymptotically
as time goes to infinity. Furthermore, we examine whether the time-dependent matter
instability in the star collapse [39] occurs in f (T ) gravity in analogy with f (R) gravity.
This instability has recently been found in the framework of f (R) gravity, in addition to
the well-known matter instability [40]. Next, we explore thermodynamics in f (T ) gravity,
especially, near to the finite-time future singularities. In particular, we demonstrate that the
second law of thermodynamics can be satisfied around the finite-time future singularities
if the temperature of the universe inside the horizon is the same as that of the apparent
horizon. We use units of kB = c = ~ = 1 and denote the gravitational constant 8G by
2 8/MPl 2 with the Planck mass of MPl = G1/2 = 1.2 1019 GeV.
The paper is organized as follows. In Sec. II, we explain the fundamental formulations and
basic equations in f (T ) gravity. In Sec. III, we investigate the finite-time future singularities,
using analogy with f (R) gravity. In Sec. IV, we reconstruct f (T ) gravity models with
realizing the finite-time future singularities. We also study a correction term which can
remove the finite-time future singularities. In addition, we reconstruct f (T ) models with
realizing inflation in the early universe, the CDM model, Little Rip cosmology and Pseudo-
Rip cosmology. The calculation of an inertial force which may lead to the dissolution of
bound structures is done in Little Rip and Pseudo-Rip cosmologies for the Earth-Sun (ES)
system. Furthermore, we discuss that the time-dependent matter instability in the star
collapse can occur in f (T ) gravity in analogy with f (R) gravity. In Sec. V, we explore
thermodynamics in f (T ) gravity. We demonstrate that the second law of thermodynamics
can be satisfied around the finite-time future singularities. Finally, conclusions are given in
Sec. VI.

4
II. f (T ) GRAVITY

A. Fundamental formulations

Orthonormal tetrad components eA (x ) are used in the teleparallelism. An index A runs

over 0, 1, 2, 3 for the tangent space at each point x of the manifold. Their relation to the
metric g is described as g = AB eA B
e . Here, and are coordinate indices on the

manifold, which also run over 0, 1, 2, 3, and eA forms the tangent vector of the manifold.
The torsion T and contorsion K tensors are defined as

T eA eA A


e , (2.1)
1
K T T T .


(2.2)
2
The teleparallel Lagrangian density is expressed by using the torsion scalar T , although in
general relativity the Lagrangian density is described by the Ricci scalar R. The torsion
scalar T is given by

T S T , (2.3)
1
S K + T T .


(2.4)
2
The modified teleparallel action describing f (T ) gravity [17] is as follows:
 
f (T )
Z
4
I = d x|e| + LM , (2.5)
22

where |e| = det eA


= g and LM is the Lagrangian of matter. The variation of the
action in Eq. (2.5) with respect to the vierbein vector field eA presents [16]
1 1 2
(eSA ) f eA T S f + SA (T ) f + eA f = eA T (M) , (2.6)
e 4 2

where T (M) is the energy-momentum tensor of all perfect fluids of ordinary matter, i.e.,
radiation and non-relativistic matter.

B. Basic equations

We take the four-dimensional flat Friedmann-Lematre-Robertson-Walker (FLRW) space-

time with the metric
ds2 = h dx dx + r 2 d2 . (2.7)

5
Here, r = a(t)r, x0 = t and x1 = r with the two-dimensional metric h = diag(1, a2 (t)),
a(t) is the scale factor, and d2 is the metric of two-dimensional sphere with unit radius.
In this background, we have g = diag(1, a2 , a2 , a2 ) and the tetrad components eA
=

(1, a, a, a). By using these relations, we find the exact value of torsion scalar T = 6H 2
with H = a/a being the Hubble parameter, where the dot denotes the time derivative, /t.
In the flat FLRW background, the gravitational field equations can be written in the
equivalent forms of those in general relativity:

2 2
H = (M + DE ) , (2.8)
3
2
H = (M + PM + DE + PDE ) , (2.9)
2

where F df /dT , F = dF/dT , and M and PM are the energy density and pressure of
all perfect fluids of generic matter, respectively. The perfect fluid satisfies the continuity
equation M + 3H (M + PM ) = 0. Moreover, the energy density and pressure of dark
components can be represented by

1
DE = J1 , (2.10)
22
1
PDE = 2 (4J2 + J1 ) , (2.11)
2

with

J1 T f + 2T F , (2.12)
J2 (1 F 2T F ) H . (2.13)

Here, DE in Eq. (2.10) and PDE in Eq. (2.11) satisfy the standard continuity equation

DE + 3H (DE + PDE ) = 0 . (2.14)

6
III. FINITE-TIME FUTURE SINGULARITIES IN f (T ) GRAVITY

A. Classification of the four types

In the FLRW background (2.1), the effective equation of state (EoS) for the universe is
given by [9]

Peff 2H
weff = 1 , (3.1)
eff 3H 2
3H 2
eff 2 , (3.2)

2H + 3H 2
Peff . (3.3)
2

Here, eff and Peff correspond to the total energy density and pressure of the universe,
respectively. When the energy density of dark energy becomes completely dominant over
that of matter, one can consider wDE weff . For H < 0 (> 0), weff > 1 (< 1),
representing the non-phantom, i.e., quintessence (phantom) phase, whereas weff = 1 for
H = 0, corresponding to the cosmological constant.
In Ref. [24], the finite-time future singularities has been classified into the following four
types. (i) Type I (Big Rip [41]): In the limit t ts , a , eff and |Peff | .
The case in which eff and Peff becomes finite values at t = ts [42] is also included. (ii) Type
II (sudden [43, 44]): In the limit t ts , a as , eff s and |Peff | . (iii) Type III:
In the limit t ts , a as , eff and |Peff | . (iv) Type IV: In the limit t ts ,
a as , eff 0, |Peff | 0, and higher derivatives of H diverge. The case in which eff
and/or |Peff | asymptotically approach finite values is also included. Here, ts , as (6= 0) and s
are constants.
It is important to mention that the Type I, i.e., Big Rip singularity, has recently been
extended by Little Rip [3037] and Pseudo Rip [38] scenarios. Furthermore, in addition to
the Type V (w) singularity, (v) Type V (w [4547]) singularity and earlier by parallel-
propagated (p.p.) curvature singularities [48] have now been proposed. For the Type V
(w) singularity, in the limit t ts , a as , eff 0, |Peff | 0, and the EoS for the
universe diverges. It should be cautioned that the Type V (w) singularity is similar to
the Type IV singularity, namely, the Type V singularity does not lead to the divergence of
physical quantities such as the scale factor, the energy density and pressure in the future,

7
but in the Type V higher derivatives of H do not diverge. This is the difference between
the Type IV and Type V singularities.
In what follows, since we examine the behavior of the universe around the finite-time
future singularities, in which dark energy dominates over matter as DE M and PDE PM
in Eqs. (2.8) and (2.9), we consider eff DE in Eq. (2.10), Peff PDE in Eq. (2.11), and
weff wDE = PDE /DE .

B. Finite-time future singularities

We consider the case in which the Hubble parameter H is represented by [the third
reference in Ref. [28]]

hs
H , for q > 0 , (3.4)
(ts t)q
hs
H Hs + , for q < 1 , 1 < q < 0 . (3.5)
(ts t)q

Here, hs (> 0) and Hs (> 0) are positive constants, q(6= 0, 1) is a non-zero constant. More-
over, ts is the time when the finite-time future singularity appears and only the period
0 < t < ts is considered due to the fact that H should be real number. When t ts , for
q > 0, both H hs (ts t)q and H qhs (ts t)(q+1) become infinity. For 1 < q < 0,
H is finite, but H becomes infinity. For q < 1, but q is not any integer, both H and H are
finite, but the higher derivatives of H can become infinity. It follows from Eq. (3.4) that
 
hs (q1)
a as exp (ts t) for 0 < q < 1 , 1 < q , (3.6)
q1
hs
a as for q = 1 , (3.7)
(ts t)hs

where as is a constant.
It can be seen from Eq. (3.6) that when t ts , for q 1, a , whereas for q < 0
and 0 < q < 1, a as . Moreover, it follows from Eqs. (3.2) and (3.4) that for q > 0,
H and therefore eff = 3H 2/2 , whereas for q < 0, H asymptotically becomes
finite and also eff asymptotically approaches a finite constant value s . On the other hand,
from H qhs (ts t)(q+1) and Eq. (3.3) we find that for q > 1, H and hence
 
Peff = 2H + 3H /2 . For q < 1, but q is not any integer, a, eff and Peff are
2

finite because both H and H are finite, whereas the higher derivatives of H diverges. As a

8
TABLE I: Conditions for the finite-time future singularities to exist on q in the expressions of H
in Eqs. (3.4) and (3.5), DE in Eq. (2.10) and PDE in Eq. (2.11), and the behaviors of H and H in
the limit of t ts .

q(6= 0, 1) H (t ts ) H (t ts ) DE PDE

q 1 [Type I (Big Rip) singularity] H H J1 6= 0 J1 6= 0

or J2 6= 0
0 < q < 1 [Type III singularity] H H J1 6= 0 J1 6= 0
1 < q < 0 [Type II (sudden) singularity] H Hs H J2 6= 0
q < 1, but q is not any integer H Hs H 0
[Type IV singularity] (Higher
derivatives
of H diverge.)

result, the properties of the finite-time future singularities described by the expressions of
H in Eqs. (3.4) and (3.5) are summarized as follows: For q 1, the Type I (Big Rip)
singularity, for 0 < q < 1, the Type III singularity, and for 1 < q < 0, the Type II
(sudden) singularity. In addition, for q < 1, but q is not any integer, the Type IV
singularity appears. We present the conditions for the finite-time future singularities to
exist on q in the expressions of H in Eqs. (3.4) and (3.5), DE in Eq. (2.10) and PDE in
Eq. (2.11), and the behaviors of H and H in the limit of t ts in Table I.

IV. RECONSTRUCTION OF f (T ) GRAVITY

In this section, first we reconstruct f (T ) gravity in which there appear the finite-time
future singularities discussed in Sec. III1 . Next, we examine a correction term removing the
finite-time future singularities.

1
In Appendix, we also describe a reconstruction method of f (T ) gravity by way of using a scalar field
through the extension of that of f (R) gravity [2830].

9
 A. Reconstruction

By using Eqs. (2.10) and (2.11), we find that the effective EoS for the universe at the
dark energy dominated stage is written as
h i

PDE 4 (1 F 2T F ) H + (T f + 2T F )
weff wDE = = . (4.1)
DE T f + 2T F

As another description, from Eqs. (2.10) and (2.11) we have

PDE = DE + I(H, H) , (4.2)

where
1 h
i
I 2 (1 F 2T F ) H . (4.3)
2
Since T = 6H 2 , the form of f (T ) is a function of H. It follows from Eqs. (3.2) and (3.3)
that Peff = eff 2H/2 . By comparing this equation with Eq. (4.2), we acquire the
differential equation
2
H + I(H, H) = 0 . (4.4)
2
The substitution of Eq. (4.3) into Eq. (4.4) yields

H (F + 2T F ) = 0 . (4.5)

For H in Eq. (3.4), Eq. (4.5) reads F + 2T F = 0 because H 6= 0.

1. Power-law model

As a form of f (T ), first we take a power-law model, given by

f (T ) = AT , (4.6)

where A(6= 0) and (6= 0) are non-zero constants. In this case, from Eq. (4.5) we have

F + 2T F = A (6)1 (2 1) H 2(1) = 0 . (4.7)

In the limit of t ts , Eq. (4.7) has to be satisfied. From Eqs. (3.4) and (3.5), we find that
for q > 0 (i.e., the Type I singularity [q 1] and the Type III singularity [0 < q < 1]),
< 1, so that Eq. (4.7) can be satisfied asymptotically, whereas for q < 0 (i.e., the Type II

10
singularity [1 < q < 0] and the Type IV singularity [q < 1]), = 1/2, in which Eq. (4.7)
is always satisfied.
Furthermore, we state the meaning of the condition F + 2T F = 0, which follows from
Eq. (4.5), and another condition that the Friedmann equation (2.8) can be satisfied, which
may be interpreted as a consistency condition. Equations (2.8) and (2.9) are expressed as

T + J1 = 0 , (4.8)

J2 H = 0 , (4.9)

which are rewritten to

f + 2T F = 0 , (4.10)
F 2T F = 0 . (4.11)

Here, we have used Eqs. (2.12) and (2.13) and H 6= 0 for H in Eqs. (3.4) and (3.5). It
is clearly seen that Eq. (4.10) corresponds to a consistency condition and that the relation
F + 2T F = 0 shown above is equivalent to Eq. (4.11). Thus, the first condition is to satisfy
the second gravitational equation (2.9). Moreover, from Eq. (4.10) with Eq. (4.6) we find
that the consistency condition becomes

f + 2T F = A (6) (2 1) H 2 = 0 . (4.12)

In the limit of t ts , Eq. (4.12) must be satisfied. For q > 0 (i.e., the Type I singularity
[q 1] and the Type III singularity [0 < q < 1]), < 0, so that Eq. (4.12) can be satisfied
asymptotically, whereas for q < 0 (i.e., the Type II singularity [1 < q < 0] and the Type
IV singularity [q < 1]), = 1/2, in which Eq. (4.12) is always satisfied.
In addition, by using Eqs. (2.12) and (2.13) with (4.6), we obtain

J1 = 6H 2 1 A (6)1 (2 1) H 2(1) ,
 
(4.13)

J2 = H 1 A (6)1 (2 1) H 2(1) .
 
(4.14)

For the expression of H in Eqs. (3.4) and (3.5), from Eqs. (4.13) and (4.14) we find that the
conditions on J1 and J2 for the finite-time future singularities to exist in Table I is always
satisfied. Thus, there can appear all the four types of the finite-time future singularities.
We note that for = 1/2, Eqs. (4.13) and (4.14) becomes J1 = T 6= 0 and J2 = H 6= 0,

11
respectively. Moreover, if A = 1 and = 1, this model in Eq. (2.12) is equivalent to general
relativity.
It is very important to note that the conditions: [for q > 0 (i.e., the Type I singularity
[q 1] and the Type III singularity [0 < q < 1]), < 0, whereas for q < 0 (i.e., the
Type II singularity [1 < q < 0] and the Type IV singularity [q < 1]), = 1/2] derived
in the above considerations are necessary conditions to produce the finite-time future
singularities and not sufficient conditions. Indeed, if < 0, the Type I singularity with
q 1 rather than the Type III singularity with 0 < q < 1 appears because in the limit of
t ts , both H and H with q 1 diverge more rapidly than those with 0 < q < 1. This
originates from the absolute value of the power q, namely, the absolute value of q for the
Type I singularity (q 1) is larger than that for the Type III singularity (0 < q < 1). As a
result, the Type I singularity is realized faster than the Type III singularity, and eventually
the Type I singularity appears. Similarly, if = 1/2, the Type IV singularity with q < 1
rather than the Type II singularity with 1 < q < 0 occurs because in the limit of t ts ,
H Hs and H 0 wiht q < 1 are realized more quickly than H Hs and H with
1 < q < 0. This also comes from the absolute value of the power q, namely, the absolute
value of q for the Type IV singularity (q < 1) is larger than that for the Type II singularity
(1 < q < 0). As a consequence, the Type IV singularity is produced faster than the Type
II singularity, and accordingly the Type IV singularity appears.
We also remark that in the Type V (w) singularity, a scale factor can be taken as [46]
 n1 !1
3 n1 1 2/ (3)
a(t) = as 1 +
2 n [2/ (3)] n 2/ (3)
 n1 1  2/(3)
!
2 n [2/ (3)] t
nas 1
3 n1 ts
 n1 !1  n
3 n1 1 2/ (3) t
+ as 1 1 , (4.15)
2 n [2/ (3)] n 2/ (3) ts

where and n are arbitrary constants. In the limit of t ts , H(t ts ) 0 and H(t
ts ) 0. On the other hand, the effective EoS for the universe weff = (1/3) (2qdec 1) .
Here, qdec aa/a2 is the deceleration parameter, which will again be redefined in Eq. (4.31)
in Sec. IV C 2. Thus, in the limit of t ts , since H(t ts ) = 0, from Eq. (4.5) we obtain
F + 2T F = 0. if we take a power-law model in Eq. (4.6) with A 6= 0 and > 1, Eq. (4.7)
can be satisfied asymptotically because H(t ts ) = 0. As a result, if we take a power-law

12
model in Eq. (4.6) with A 6= 0 and > 1, the Type V (w) singularity can appear.

2. Exponential model

Next, we examine an exponential model

f (T ) = C exp (T ) , (4.16)

where C(6= 0) and (6= 0) are non-zero constants. In this case, Eqs. (4.10) and (4.11)
becomes

f + 2T F = C (1 + 2T ) = 0 , (4.17)

F 2T F = C (2T + 1) exp (T ) = 0 . (4.18)

For the expression of H in Eqs. (3.4) and (3.5), in the limit of t ts , both Eqs. (4.17) and
(4.18) cannot be satisfied simultaneously. Thus, in an exponential model in Eq. (4.16) there
cannot apper the finite-time future singularities.

3. Logarithmic model

Next, we explore an logarithmic model

f (T ) = D ln (T ) , (4.19)

where D(6= 0) is a non-zero constant and (> 0) is a positive constant. In this case,
Eqs. (4.10) and (4.11) becomes

f + 2T F = D [ ln (T ) + 2) = 0 , (4.20)
D
F 2T F = = 0. (4.21)
T

For the expression of H in Eqs. (3.4) and (3.5), in the limit of t ts , both Eqs. (4.20) and
(4.21) cannot be satisfied simultaneously. Hence, in an logarithmic model in Eq. (4.19) the
finite-time future singularities cannot occur, similarly to the case of a exponential model
in Eq. (4.16) in Sec. IV A 2. Thus, in general the occurrence of the finite-time future
singularities in f (T ) gravity is realized in less cases than in f (R) gravity.

13
 B. Correction term removing the finite-time future singularities

It is known that in f (R) gravity, the addition of an R2 term can cure the finite-time
future singularities (see Ref. [9]). Recently, it has also been demonstrated in Ref. [27] that
the addition of an R2 term can remove the finite-time future singularities in non-local gravity.
In this subsection, we investigate a correction term for the form of f (T ) in Eq. (4.6) so that
the finite-time future singularities cannot appear. To execute this analysis, we explore an
additional term of as function of T to the form of f (T ) in Eq. (4.6) so that for H in Eqs. (3.4)
and (3.5), the gravitational field equations (2.8) and (2.9) with Eqs. (2.10) and (2.11) cannot
be satisfied.
As a straight forward procedure, we explore the case that the form of f (T ) represented
by Eq. (4.6) has a correction term fc (T ), and analyze whether Eqs. (4.10) and (4.11) can be
satisfied or not. As an example, we choose a correction term fc (T ) as

fc (T ) = BT , (4.22)

where B(6= 0) and (6= 0) are non-zero constants. For = 2, the correction term is similarly
to that in f (R) gravity, i.e., a T 2 term. By combining Eqs. (4.6) and (4.22), the total form
of f (T ) including the correction term is expressed as

f (T ) = AT + BT . (4.23)

By substituting Eq. (4.23) into Eqs. (4.10) and (4.11), we find

f + 2T F = A (2 1) T + B (2 1) T 6= 0 , (4.24)

F 2T F = A (2 1) T 1 B (2 1) T 1 6= 0 . (4.25)

From the considerations in Sec. IV A, we find that for q > 0 (i.e., the Type I singularity
[q 1] and the Type III singularity [0 < q < 1]), > 0, so that the second inequality in
Eq. (4.24) can be satisfied asymptotically, whereas for q < 0 (i.e., the Type II singularity
[1 < q < 0] and the Type IV singularity [q < 1]), 6= 1/2, in which the second inequality
in Eq. (4.25) is always satisfied. As a result, if > 1, for the Hubble parameter in Eqs. (3.4)
and (3.5), in the limit of t ts both of the gravitational field equations (2.8) and (2.9)
cannot be satisfied. This means that a power-low type correction term T with > 1 can
remove the finite-time future singularities in f (T ) gravity. In Table II, we describe necessary

14
TABLE II: Necessary conditions on the model parameters of a power-law model of f (T ) in Eq. (4.6)
with realizing the finite-time future singularities, the emergence of the finite-time future singular-
ities, and those of the correction term fc (T ) = BT in Eq. (4.22) with removing the finite-time
future singularities.

q(6= 0, 1) Emergence f (T ) = AT fc (T ) = BT
(A 6= 0, 6= 0) (B 6= 0, 6= 0)

q 1 [Type I (Big Rip) singularity] Yes <0 >1

0 < q < 1 [Type III singularity] <0 >1
1 < q < 0 [Type II (sudden) singularity] = 1/2 6= 1/2
q < 1, but q is not any integer Yes = 1/2 6= 1/2
[Type IV singularity]

conditions on the model parameters of a power-law model of f (T ) in Eq. (4.6) with realizing
the finite-time future singularities, the emergence of the finite-time future singularities, and
those of the correction term fc (T ) = BT in Eq. (4.22) with removing the finite-time future
singularities. It is interesting to emphasize that a T 2 term, i.e., = 2, which is the minimum
integer to satisfy the condition > 1, can remove all the four types of the finite-time future
singularities in f (T ) gravity. This consequence is the same as that in f (R) gravity.
It is interesting note that for the case of the Type V (w) singularity, since in the limit of
t ts H(t ts ) = 0, if we choose a power-law type correction term fc (T ) in Eq. (4.22) with
B 6= 0 and < 0 and substitute Eq. (4.23) into Eqs. (4.10) and (4.11), we find Eqs. (4.24)
and (4.25). In other words, both of the gravitational field equations (2.8) and (2.9) cannot
be satisfied asymptotically. This means that a power-law type correction term fc (T ) in
Eq. (4.22) with B 6= 0 and < 0 can remove the Type V (w) singularity.

C. Reconstructed models realizing cosmologies

In this subsection, we explicitly present the reconstruction of f (T ) models with realizing

(a) inflation in the early universe, (b) the CDM model, (c) Little Rip cosmology and (d)
Pseudo-Rip cosmology.

15
 1. Inflation in the early universe

For generality, we consider power-law inflation. We suppose that the Hubble parameter
is given by
hinf
H= , (4.26)
t
where hinf (> 1) is a constant larger than unity. It follows from Eq. (4.26) that the scale
factor is given by
a(t) = ainf thinf . (4.27)

From Eq. (4.27), we find that a = ainf hinf (hinf 1) thinf 2 > 0, and hence power-law inflation
occurs. In this case, since H = hinf /t2 6= 0, by using Eq. (4.5) we obtain the condition
F + 2T F = 0.
As explained in Sec. IV A 1, for a power-law model f (T ) = AT in Eq. (4.6), the
conditions to be satisfied, which originate from the gravitational equations (2.8) and (2.9),
are given by Eqs. (4.7) and (4.12). Therefore, if < 0, in the limit of t 0, H in Eq. (4.26)
diverges, so that Eqs. (4.7) and (4.12) can approximately be satisfied in the very early
universe. Moreover, if = 1/2, Eqs. (4.7) and (4.12) can always be met.

2. The CDM model

When we describe the CDM model by the action in Eq. (2.5), f (T ) = T 2, where >
0 is positive cosmological constant, as is in general relativity. In this case, by substituting
this form of f (T ) into Eqs. (2.8) and (2.9), we have

H2 = , (4.28)
3
H = 0 . (4.29)
p
Clearly, from Eqs. (4.28) and (4.29), we find H H = /3 = constant, where we have
defined the Hubble parameter at the cosmological constant dominated stage as H (> 0).
Furthermore, the scale factor is expressed as

a = a exp (H t) , (4.30)

where a (> 0) is a positive constant.

16
 In the CDM model, from Eq. (3.1) we find that the EoS is given by wDE = 1 due to
the fact that H is constant. Moreover, the deceleration parameter qdec , the jerk parameter
j and the snark parameter s are defined by [49]

1 d2 a
qdec , (4.31)
aH 2 dt2
1 d3 a
j , (4.32)
aH 3 dt3
j1
s . (4.33)
3 (qdec 1/2)

By using Eq. (4.30), we obtain qdec = 1, j = 1 and s = 0.

The limit on a constant EoS for dark energy in a flat universe has been estimated as
wDE = 1.10 0.14 (68% CL) [the second reference in Ref. [3]]. In addition, for a time-
dependent EoS for dark energy with a linear form wDE (a) = wDE 0 + wDE a (1 a) [50],
where wDE 0 and wDE a are the current value of wDE and its derivative, respectively, the
constraints have been analyzed as wDE 0 = 0.93 0.13 and wDE a = 0.41+0.72
0.71 (68% CL)

[the second reference in Ref. [3]]. Thus, the deviations of the values of (wDE , qdec , j, s) from
those for the CDM model (1, 1, 1, 0) show how the model is different form the CDM
model. In other words, we can use these four parameters as a observational test.
We remark that if we consider the early universe, the model f (T ) = T 2 can lead to
exponential inflation realizing de Sitter expansion of the universe, i.e., the Hubble parameter
is given by
H = Hinf = constant , (4.34)

where Hinf > 0. By using Eq. (4.34), we have

a(t) = ainf exp (Hinf t) , (4.35)

where ainf (> 0) is a positive constant.

3. Little Rip cosmology

Furthermore, we study Little Rip cosmology [30? 36], which corresponds to a mild
phantom scenario. The Little Rip scenario has been proposed to avoid the finite-time future
singularities, in particular a Big Rip singularity. In this scenario, the energy density of
dark energy increases in time with wDE being less than 1 and then wDE asymptotically

17
approaches wDE = 1. However, such a scenario eventually leads to the dissolution of bound
structures at some time in the future via the increase of an inertial force between objects.
This process is called the Little Rip.
As an example to realize Little Rip cosmology, we take the Hubble parameter as [33]

H = HLR exp (t) , (4.36)

where HLR (> 0) and (> 0) are positive constants. In this case, the scale factor a is
expressed as  
HLR
a = aLR exp exp (t) , (4.37)

where aLR (> 0) is a positive constant. Moreover, from Eq. (3.1) we obtain

2
wDE = 1 exp (t) . (4.38)
3HLR

Since H = HLR exp (t) > 0, wDE < 1, i.e., the universe is always in the phantom phase.
In the limit of t , we find wDE 1 and hence the Little Rip scenario can be realized.
In the expression of wDE in Eq. (4.38), if we take = H0 , at t = t0 H01 , we have
wDE = 1 2H0 / (3HLR e). Here, t0 is the present time, H0 is the current value of the
Hubble parameter given by H0 = 2.1h 1042 GeV [51] with h = 0.7 [the second reference
in Ref. [3], [52]], and e = 2.71828. By comparing this expression with the observational
constraint on wDE = 1.10 0.14 (68% CL) [the second reference in Ref. [3]], we find that
if HLR [2H0 / (3e)] (1/0.24) = 1.50 1042 GeV, the current value of wDE in this Little
Rip model is consistent with the observations. Here, we have used the fact that = H0 and
HLR are positive values.
By using Eqs. (4.31)(4.33), (4.36) and (4.37), we acquire

qdec = 1 , (4.39)
HLR exp (t)
 
1
j = 1+ +3 , (4.40)
HLR HLR exp (t) exp (t)
2 [ + 3HLR exp (t)]
s= . (4.41)
3HLR [2 + 3HLR exp (t)] exp (t)

For = H0 , at t = t0 H01, from Eqs. (4.38)(4.41), we describe the expression of wDE ,

18
qdec , j and s at the present time t0 as

2
wDE (t = t0 ) = 1 , (4.42)
3
qdec (t = t0 ) = 1 , (4.43)

j(t = t0 ) = 1 + ( + 3) , (4.44)
2 ( + 3)
s(t = t0 ) = , (4.45)
3 (2 + 3)

with
H0
0.36 , (4.46)
HLR e
where the second inequality in Eq. (4.46) follows from the observational constraint on wDE =
1.100.14 (68% CL) [the second reference in Ref. [3]] as (3/2) 0.24 = 0.36. As a result,
if we take 1 enough for the deviation of the values of the four parameters (wDE , qdec , j, s)
from those for the CDM model (1, 1, 1, 0) to be very small, this Little Rip model can
be compatible with the CDM model.
It follows from Eq. (4.36) that in the limit of t , H diverges. For a power-law model
f (T ) = AT in Eq. (4.6), if < 0, in the limit of t , Eqs. (4.7) and (4.12) can be
satisfied asymptotically. This is just an opposite case in the model with realizing inflation in
Sec. IV C 1 because the limit in terms of t is the opposite direction. In addition, if = 1/2,
Eqs. (4.7) and (4.12) can always be met, similarly to that in the case of inflation in Sec. IV
C 1.

4. Pseudo-Rip cosmology

We also investigate Pseudo-Rip cosmology [34, 38]. The above four cosmological models
can be classified by using the behavior of the Hubble parameter as follows [38]. (a) power-law
inflation:
H(t) , t 0. (4.47)

(b) the CDM model or exponential inflation:

H(t) = H(t0 ) . (4.48)

(c) Little Rip cosmology:

H(t) , t . (4.49)

19
(d) Pseudo-Rip cosmology, which is also phantom asymptotically de Sitter universe:

H(t) H < , t . (4.50)

Here, we consider t t0 . Moreover, H (> 0) is a positive constant. We also note that for a
Big Rip singularity, H(t) , t ts , as shown in Table I. As an example of a Pseudo-Rip
model, we take  
t
H(t) = HPR tanh , (4.51)
t0
where HPR (> 0) is a positive constant. In this case, the scale factor a is expressed as
 
t
a = aPR cosh , (4.52)
t0

where aPR (> 0) is a positive constant. From Eq. (4.51), we find that H(t) is monotonically
increasing function of t and H(t) HPR < , t . Thus, a behavior of H in the
Pseudo-Rip cosmology in (4.50) is realized. We also have H(t) = HPR / t0 cosh2 (t/t0 )
 

0, t . This means from Eq. (2.9) that P in the limit of t . For a power-law

model f (T ) = AT with = 1/2 in Eq. (4.6), i.e., f (T ) = A T , Eqs. (2.9) and (2.9) can
always be satisfied including in the limit of t .
For H in Eq. (4.51), from Eq. (3.1) we find that the EoS is given by

2 1
wDE = 1 2 . (4.53)
3t0 HPR sinh (t/t0 )

From Eq. (4.53), we see that wDE < 1, namely, the universe is always in the phantom
phase, because H(t) = HPR / t0 cosh2 (t/t0 ) > 0. In the limit of t , we find wDE 1,
 

similarly to that in Little Rip cosmology.

It follows from wDE in Eq. (4.53) that at t = t0 H01 , wDE = 1
h i
2
[2H0 / (3HPR )] 4/ (e e1 ) . In comparison with the observational constraint on wDE =
1.10 0.14 (68% CL) [the second reference in Ref. [3]], we find that if HPR
h i
2
(2H0 /3) 4/ (e e1 ) (1/0.24) = 2.961042 GeV, the current value of wDE in this Pseudo-
Rip model is compatible with the observations. Here, we have used the fact that t0 H01
and HPR are positive values.

20
 Moreover, by using Eqs. (4.31)(4.33), (4.51) and (4.52), we have

(t0 HPR )2 tanh2 (t/t0 ) 1

qdec = 1 + , (4.54)
(t0 HPR )2 tanh2 (t/t0 )
1 (t0 HPR )3 tanh2 (t/t0 )
j = 1+ , (4.55)
(t0 HPR )3 tanh2 (t/t0 )
2 (t0 HPR )3 tanh2 (t/t0 ) 1
s= . (4.56)
3t0 HPR (t0 HPR )2 tanh2 (t/t0 ) + 2

At t = t0 H01, from Eqs. (4.53)(4.56), we describe the expression of wDE , qdec , j and
s at the present time t0 as

2
wDE (t = t0 ) = 1 , (4.57)
3 sinh2 1
2 tanh2 1 1
qdec (t = t0 ) = 1 + , (4.58)
2 tanh2 1
1 3 tanh2 1
j(t = t0 ) = 1 + , (4.59)
3 tanh2 1
2 3 tanh2 1 1
s(t = t0 ) = , (4.60)
3 2 tanh2 1 + 2

with
H0
0.497196 , (4.61)
HPR
where the second inequality in Eq. (4.61) follows from the observational constraint on wDE =
1.10 0.14 (68% CL) [the second reference in Ref. [3]] as (3/2) 0.24 sinh2 1 = 0.497196.
Here, we use sinh2 1 = 1.3811 and tanh2 1 = 0.580026. As a consequence, we can take
an appropriate value of in order for the deviation of the values of the four parameters
(wDE , qdec , j, s) from those for the CDM model (1, 1, 1, 0) to be very small, so that this
Pseudo-Rip model can be consistent with the CDM model, similarly to that in the Little
Rip model discussed in Sec. IV C 3. In Table III, we display forms of H and f (T ) with
realizing (a) inflation in the early universe, (b) the CDM model, (c) Little Rip cosmology
and (d) Pseudo-Rip cosmology.
In the expanding universe, the relative acceleration between two points separated by a
distance l is given by la/a, where a is the scale factor. Suppose that there exists a particle
with mass m at each of the points, an observer at one of the masses would measure an

21
TABLE III: Forms of H and f (T ) with realizing (a) inflation in the early universe, (b) the CDM
model, (c) Little Rip cosmology and (d) Pseudo-Rip cosmology.

Cosmology H f (T )

(a) Power-law inflation H = hinf /t , f (T ) = AT ,

(In the limit of t 0) hinf (> 1) < 0 or = 1/2
p
(b) CDM model H = /3 = constant , f (T ) = T 2 ,
or exponential inflation >0 >0
(c) Little Rip cosmology H = HLR exp (t) , f (T ) = AT ,
(In the limit of t ) HLR > 0 and > 0 < 0 or = 1/2

(d) Pseudo-Rip cosmology H = HPR tanh (t/t0 ) , f (T ) = A T
HPR > 0

inertial force on the other mass. The inertial force Finert on a mass m is given by [31, 33]
a  
Finert = ml = ml H + H 2 (4.62)
a
2 2
 
dDE (a)
= ml (DE (a) + 3PDE (a)) = ml 2DE (a) + a , (4.63)
6 6 da
where in deriving the first equality in Eq. (4.63) we have used Eqs. (2.8) and (2.9). We take
the present value of a as a0 a(t = t0 ) = 1. We also provide that the two particles are
bound by a constant force Fb . When Finert (> 0) is a positive force and the amplitude is
larger than that of Fb , the two particles become unbound and hence the bound structure is
dissociated. In Pseudo-Rip cosmology, Finert is asymptotically finite.
For a Big Rip singularity realizing H in Eq. (3.4) with q 1, by using Eq. (4.62) we find
 
q hs
Finert = mlhs + , t ts . (4.64)
(ts t)q+1 (ts t)2q
Here, the reason of the divergence is that H and H diverge in the limit of t ts . Moreover.
for a Little Rip model in Eq. (4.36), we see that

Finert = mlHLR [ + HLR exp (t)] exp (t) , t . (4.65)

Similarly, the reason of the divergence is that H and H diverge in the limit of t .
Thus, a phenomenon of Rip is produced by the cosmic accelerated expansion at a Big Rip
singularity and in Little Rip cosmology.

22
 On the other hand, for a Pseudo-Rip model in Eq. (4.51), we obtain
  
1 2 t PR
Finert = mlHPR 2 + HPR tanh Finert , < , t , (4.66)
t0 cosh (t/t0 ) t0

where
PR 2
Finert , mlHPR . (4.67)

Therefore, Finert is asymptotically finite. Here, the reason why Finert becomes a finite value
PR
of Finert , in the limit of t is that H HPR and H 0.
PR ES 2
It is necessary for Finert , to be larger than the bound force Fb = GM M /r =
4.37 1016 GeV2 of the ES system in order that the ES system will be disintegrated and
hence the Pseudo-Rip scenario can be realized. Here, M = 3.357 1051 GeV [51] and
M = 1.116 1057 GeV [51] are masses of Earth and Sun, respectively, and r = 1AU =
7.5812 1026 GeV1 [51] is the distance between Earth and Sun, i.e., the Astronomical
unit. As an example, if we take m as the Earth mass m = M , l is the distance between
PR ES
Earth and Sun l = r , in order for Finert , > Fb , by using Eq. (4.67), we find HPR >
p
3
GM /r = 1.31 1031 GeV. If this condition is met, the disintegration of the ES
system can occur much before arriving at de Sitter universe, so that the Pseudo-Rip scenario
can be realized. In addition, if this constraint is satisfied, the current value of wDE in this
Pseudo-Rip model is also consistent with the observations because the constraint on HPR
from the current value of wDE is much weaker as HPR 2.96 1042 GeV.

D. Time-dependent matter instability and star singularity in f (T ) gravity

In modified gravity, the important property of the viability is the existence of the gravita-
tional bound objects (stars, planets). It is checked via matter instability [40]. For instance,
without such a property the relativistic star might not be formed because a corresponding
singularity appears (for example, it is known that a singularity may appear for stars [53]
in f (R) gravity). In a gravitating system with a time dependent mass density such as as-
tronomical massive objects, the instability in f (R) gravity has recently been studied [39].
Furthermore, the generation mechanism of the time-dependent matter instability in the star
collapse has also been investigated. It has been demonstrated that the time-dependent mat-
ter instability develops and consequently the curvature singularity could appear [54]. In
this subsection, by examining the process for the curvature singularity to be realized [54] in

23
analogy with f (R) gravity, we discuss whether the time-dependent matter instability in the
star collapse occurs in f (T ) gravity.
We examine a small region inside the star. We can regard this system as homogeneous
and isotropic and hence the space-time is locally described by the flat FLRW metric (2.7).
Here, the Hubble parameter H is negative because we are exploring the star collapse, so
that the space-time can be shrinking. Since the region is shrinking, the energy densities of
the matters automatically increase. In case of cosmology, as shown in Sec. III B, all the
four types of the finite-time future singularities in f (T ) gravity can appear. If H diverges
in the limit of t tst , where tst is the time when the time when the curvature singularity
in the star appears, T becomes infinity because T = 6H 2 . From Table I, we see that for
the Type III singularity, T diverges, although the scale factor a is finite. This phenomenon
can be applied to the star collapse. The energy density and the pressure from the matter
are finite and therefore these can be neglected near the singularity because a asymptotically
becomes a finite value. In this case, the Hubble parameter H is expressed as

hst
H , (4.68)
(tst t)q

where hst (> 0) is a positive constant. For the Type III singularity, we have 0 < q < 1. In
the limit of t tst , T becomes infinite because H diverges. This means that the naked
curvature singularity appears in the finite future. We note that in the above investigations,
we have supposed that the region is almost homogeneous and isotropic. When these settings
are adequate also near the singularity, the singularity simultaneously occurs in all of the
region. The naked singularities occur densely in the region, even though the homogeneity
and isotropy are broken. Moreover, the density of matter grows as in the region farther
from the surface of the star, namely, nearer to the center of it. Accordingly, first the naked
curvature singularity can appear near the center of the star. When the singularity produces
the attractive force, the shrinking of the star proceeds more and more, whereas when the
repulsive force is generated, eventually the explosion may happen. However, when the
explosion happens, the sign of H has to change from negative to positive. The realization
of this phenomenon seems to be difficult.
On the other hand, in the cosmological context, when the Hubble parameter H is given
by Eq. (3.4) with 0 < q < 1, the Type III singularity can appear. By using the initial
condition for H and H, we can determine the values of ts (and hs ) in Eq. (3.4) and tst (and

24
hst ) in Eq. (4.68). The absolute values of H and H inside the star would be larger than
those in the expanding universe because we are investigating the collapsing star. Hence,
we find ts > tst . In other words, the curvature singularity in the star appears before the
cosmological singularity. As a consequence, the time-dependent matter instability in the
star collapse can occur in f (T ) gravity, similarly to that in f (R) gravity.

V. THERMODYNAMICS AROUND THE FINITE-TIME FUTURE SINGULAR-

ITIES

In this section, we discuss thermodynamics in f (T ) gravity in order to examine whether

f (T ) gravity is a viable gravitational theory. In particular, by following the procedure in
Refs. [23, 55], we examine whether the second law of thermodynamics can be verified around
the finite-time future singularities . The fundamental connection between gravitation and
thermodynamics was implied by Black hole thermodynamics [56] (for recent reviews, see,
e.g., [57]). In general relativity, by using the proportionality of the entropy to the horizon
area, the Einstein equation was derived from the Clausius relation in thermodynamics [58].
This consequence has been app lied to more general extended gravitational theories [59, 60].

A. First law of thermodynamics

In Refs. [23, 55, 61], it has been shown that if the standard continuity equation (2.14) in
terms of dark component is satisfied, an equilibrium description of thermodynamics can be
obtained. In the flat FLRW space-time, the radius rA of the apparent horizon is written by
rA = 1/H. The dynamical apparent horizon is determined by the relation h r r = 0.
The time derivative of rA = 1/H leads to drA /rA3 = HHdt. By plugging Eq. (2.8) with
this equation, we find the relation [1/ (4G)] drA = rA3 H (t + Pt ) dt. Here, t DE + M
and Pt PDE + PM are the total energy density and pressure of the universe, respectively.
In general relativity, the Bekenstein-Hawking horizon entropy is expressed as S = A/ (4G)
with A = 4rA2 being the area of the apparent horizon [56]. By using the horizon entropy
and the above relation, we acquire

1
dS = 4rA3 H (t + Pt ) dt . (5.1)
2rA

25
The Hawking temperature TH = |sg |/ (2) corresponds to the associated temperature of
the apparent horizon. Here, the surface gravity sg is represented by [62]

1  
sg = hh r (5.2)
2 h
rA
 
1 rA  2  2G
= 1 = 2H + H = rA (t 3Pt ) (5.3)
rA 2H rA 2 3F
2G
= rA (1 3wt ) t , (5.4)
3F

where h is the determinant of the metric h and wt Pt /t is the EoS for the total of
energy and matter in the universe. It follows from Eq. (5.4) that for wt 1/3, we have
sg 0. Thus, by substituting Eq. (5.3) into TH = |sg |/ (2), we obtain

rA
 
1
TH = 1 . (5.5)
2rA 2H rA

By combining Eq. (5.1) with Eq. (5.5), we acquire

TH dS = 4rA3 H (t + Pt ) dt 2rA2 (t + Pt ) drA . (5.6)

The Misner-Sharp energy [63] is defined as E = rA / (2G) = V t with V = 4rA3 /3 being the
volume inside the apparent horizon. The last equality implies that E is equivalent to the
total intrinsic energy. By using this equation, we have

dE = 4rA3 H (t + Pt ) dt + 4rA2 t drA . (5.7)

(1/2) T (M) h + T (DE) h

We also define the work density [64]: W =
(1/2) (t Pt ). Here, T (M) and T (DE) are the energy-momentum tensor of matter and
that of dark components, respectively. By using Eq. (5.7) and the work density W , the first
law of equilibrium thermodynamics can be described.

TH dS = dE + W dV (5.8)

As a result, an equilibrium description of thermodynamics can be realized. We also remark

 
that from Eqs. (2.8), (2.9) and (5.1), we have S = 8 2 H rA4 (t + Pt ) = (6/G) T /T 2 =
h i
3
(2/G) H/ (3H ) > 0. This means that in the expanding universe, the horizon entropy
S always increases as long as the null energy condition t + Pt 0, i.e., H 0, is met.

26
 B. Second law of thermodynamics

Next, we investigate the second law of thermodynamics in the equilibrium description.

The Gibbs equation in terms of all matter and energy fluid is represented by TH dSt =
d (t V ) + Pt dV = V dt + (t + Pt ) dV . The second law of thermodynamics can be expressed
as dSsum /dt dS/dt + dSt /dt 0 with Ssum S + St , where St is the entropy of total
energy inside the horizon. By using V = 4rA3 /3, Eqs. (2.9), (5.5) and the relation S =
 
8 2 H rA4 (t + Pt ) = (6/G) T /T 2 , we acquire
!2
dSsum 6 T 1
= . (5.9)
dt G T 4HT + T

As a result, from the second law of thermodynamics with Eq. (5.9) we find the condition [23]
   
2
Y 4HT + T = 12H 2H + H 0. From the behaviors of H and H in the limit
of t ts described in Table I, it can be seen that in the expanding universe, where H > 0,
for all the four types of the finite-time future singularities, the relation 2H 2 + H 0 can
always be realized. It is also interesting to note that this relation is satisfied even in the
phantom phase (H > 0). Thus, the second law of thermodynamics around the finite-time
future singularities is always satisfied. However, it should be cautioned that at the exact
singularity t = ts such as a Big Rip singularity, in which the scale factor a(t) diverges as
a(t = ts ) = , a naive classical picture of thermodynamics might break down. We also
mention that this result can be verified only provided that the temperature of the universe
inside the apparent horizon is equal to that of the horizon [65].

VI. CONCLUSIONS

In the present paper, we have reconstructed f (T ) gravity in which finite-time future

singularities appear. Furthermore, it has been demonstrated that a T ( > 1) correction
term, e.g., = 2, to a model of f (T ) gravity where the finite-time future singularities occur,
can remove the finite-time future singularities, similarly to that in f (R) gravity. We have
also explored non-singular f (T ) models in which inflation in the early universe, the CDM
model, Little Rip cosmology and Pseudo-Rip cosmology are realized. It has been shown
that the dissolution of bound structures for Little Rip and Pseudo-Rip cosmologies happens
in the same manner as in gravity with corresponding dark energy fluid. Moreover, we have

27
considered that the time-dependent matter instability in the star collapse can occur in f (T )
gravity, similarly to that in f (R) gravity. In addition, we have investigated thermodynamics
in f (T ) gravity and shown that the second law of thermodynamics can be satisfied around
the finite-time future singularities, provided that the temperature of the universe inside the
horizon is equal to that of the apparent horizon.
It would be interesting to develop more complicated versions of f (T ) gravity and study
their cosmological applications. For instance, one can consider non-minimal coupling of
f1 (T ) with electrodynamics, where f1 (T ) is a function of T , in analogy with that in non-
minimal f1 (R) gravity [66]. This may lead to the emergence of a domain-wall solution and
variation of fine structure constant in non-minimal f1 (T ) gravity. From another side, it
seems to be very interesting to generalize an f (T ) theory in a consistent way so that one
can include the presence of curvature in the Lagrangian.
It should be emphasized that the illustration of the existence of the finite-time future
singularities in f (T ) gravity and the possibility of those removing due to an additional
power-low term are nontrivial and significant. These consequences are also found in other
alternative gravitational theories such as f (R) gravity. It is considered that the removal
possibility of the finite-time future singularities can be one of the tests of a successful al-
ternative gravitational theory to general relativity. Therefore, it is strongly expected that
through these analyses of meaningful theoretical properties of modified gravitational theo-
ries, we can obtain a successful alternative gravitational theory to general relativity which
explains the cosmic accelerated expansion of the universe in a geometrical way.

Acknowledgments

S.D.O. would like to appreciate the support and very kind hospitality at Eurasian National
University and also acknowledge the Japan Society for the Promotion of Science (JSPS)
Short Term Visitor Program S11135 and the very warm hospitality at Nagoya University
where the work was developed. The work is supported in part by Global COE Program of
Nagoya University (G07) provided by the Ministry of Education, Culture, Sports, Science
& Technology (S.N.); the JSPS Grant-in-Aid for Scientific Research (S) # 22224003 and
(C) # 23540296 (S.N.); and MEC (Spain) project FIS2010-15640 and AGAUR (Catalonia)
2009SGR-994 (S.D.O.).

28
 Appendix A: Reconstruction method

In this appendix, we explain the procedure of a reconstruction method of f (T ) gravity

by applying that of f (R) gravity [2830] to f (T ) gravity. The action of f (T ) gravity with
matter is given by Eq. (2.5). We can rewrite the action in Eq. (2.5) with proper functions
P () and Q() of a scalar field to

4 1
Z Z
S = d x g 2 (P ()T + Q()) + d4 xLM (g , M ) . (A1)
2
Since the scalar field does not have kinetic term, we may regard as an auxiliary scalar
field. From Eq. (A1), we derive the equation of motion of as 0 = (dP ()/d) T +dQ()/d.
The substitution of = (T ) into the action in Eq. (A1) leads to the expression of f (T ) as
f (T ) = P ((T ))T + Q((T )). It follows from Eq. (2.6) that the gravitational field equation
is described by
1 dP () d
(eSA ) P () eA T S P () + SA (T )
e d dT
1 2
+ eA (P ()T + Q()) = eA T (M) . (A2)
4 2
In the flat FLRW background space-time with the metric in Eq. (2.7), the components
of (, ) = (0, 0) and (, ) = (i, j) (i, j = 1, , 3) in Eq. (A2) yield two independent
differential equations in terms of P ((t)) and Q((t)). In principle, by eliminating Q()
from these two equations, we obtain an equation which P ((t)) obeys. We may take the
scalar field as = t by redefining it properly. We express a(t) as a(t) = a exp (g(t)),
where a is a constant and g(t) is a proper function of t. By using H = dg()/ (d), we
represent the equation in terms of P ((t)) derived above to be a form described by P ((t)),
the derivatives of P ((t)) in terms of , dg()/ (d), and the derivatives of dg()/ (d) in
terms of . Furthermore, from another equation we can also express Q((t)) with P ((t)),
the derivatives of P ((t)) in terms of , dg()/ (d), and the derivatives of dg()/ (d) in
terms of . We derive the solutions of P ((t)) and Q((t)) and substitute these solutions
to f (T ) = P ((T ))T + Q((T )), we acquire the expression of f (T ) as a function of only T .
Finally, we remark the following point. By redefining the auxiliary scalar field as
() with a proper function and defining P () P (()) and Q() Q(()), the

new action S = d4 x g f(T )/ (22 ) + d4 xLM (g , M ), where f(T ) P ()T + Q(),
R R

is equivalent to the action in Eq. (A1). This is because f(T ) = f (T ). Since is the inverse

29
function of , by using = (T ), can be solved with respect to T as = (T ) =
1 ((T )). Accordingly, there exist the choices in as a gauge symmetry and therefore
can be identified with time t as = t. We can interpret this fact as a gauge condition
which is equivalent to the reparameterization of = () [67]. As a result, we can find
the form of f (T ) by obtaining the relation t = t(T ). In addition, regarding the relation
between H and T in f (T ) gravity and that between H and R in f (R) gravity in the flat
FLRW background space-time, we note that for f (T ) gravity and f (R) gravity, we have
 
2 2
T = 6H and R = 6 2H + H , respectively, and therefore in comparison with f (R)
gravity, in f (T ) gravity the torsion scalar T depends on only H, although in f (R) gravity
the scalar curvature R depends on both H and H.

[1] S. Perlmutter et al. [SNCP Collaboration], Astrophys. J. 517, 565 (1999)

[arXiv:astro-ph/9812133];
A. G. Riess et al. [Supernova Search Team Collaboration], Astron. J. 116, 1009 (1998)
[arXiv:astro-ph/9805201].
[2] D. N. Spergel et al. [WMAP Collaboration], Astrophys. J. Suppl. 148, 175 (2003)
[arXiv:astro-ph/0302209]; Astrophys. J. Suppl. 170, 377 (2007) [arXiv:astro-ph/0603449].
[3] E. Komatsu et al. [WMAP Collaboration], Astrophys. J. Suppl. 180, 330 (2009)
[arXiv:0803.0547 [astro-ph]]; Astrophys. J. Suppl. 192, 18 (2011) [arXiv:1001.4538 [astro-
ph.CO]].
[4] M. Tegmark et al. [SDSS Collaboration], Phys. Rev. D 69, 103501 (2004)
[arXiv:astro-ph/0310723];
U. Seljak et al. [SDSS Collaboration], Phys. Rev. D 71, 103515 (2005)
[arXiv:astro-ph/0407372].
[5] D. J. Eisenstein et al. [SDSS Collaboration], Astrophys. J. 633, 560 (2005)
[arXiv:astro-ph/0501171].
[6] B. Jain and A. Taylor, Phys. Rev. Lett. 91, 141302 (2003) [arXiv:astro-ph/0306046].
[7] L. Amendola and S. Tsujikawa, Dark Energy (Cambridge University press, 2010).
[8] M. Li, X. D. Li, S. Wang and Y. Wang, Commun. Theor. Phys. 56, 525 (2011) [arXiv:1103.5870
[astro-ph.CO]].

30
 [9] S. Nojiri and S. D. Odintsov, Phys. Rept. 505, 59 (2011) [arXiv:1011.0544 [gr-qc]]; eConf
C0602061, 06 (2006) [Int. J. Geom. Meth. Mod. Phys. 4, 115 (2007)] [arXiv:hep-th/0601213].
[10] S. Capozziello and V. Faraoni, Beyond Einstein Gravity (Springer, 2010).
[11] S. Capozziello and M. De Laurentis, Phys. Rept. 509, 167 (2011) [arXiv:1108.6266 [gr-qc]].
[12] A. De Felice and S. Tsujikawa, Living Rev. Rel. 13, 3 (2010) [arXiv:1002.4928 [gr-qc]].
[13] T. Clifton, P. G. Ferreira, A. Padilla and C. Skordis, arXiv:1106.2476 [astro-ph.CO].
[14] F. W. Hehl, P. Von Der Heyde, G. D. Kerlick and J. M. Nester, Rev. Mod. Phys. 48, 393
(1976);
K. Hayashi and T. Shirafuji, Phys. Rev. D 19, 3524 (1979) [Addendum-ibid. D 24, 3312
(1982)];
E. E. Flanagan and E. Rosenthal, Phys. Rev. D 75, 124016 (2007); [arXiv:0704.1447 [gr-qc]];
J. Garecki, arXiv:1010.2654 [gr-qc].
[15] R. Ferraro and F. Fiorini, Phys. Rev. D 75, 084031 (2007) [arXiv:gr-qc/0610067]; Phys. Rev.
D 78, 124019 (2008) [arXiv:0812.1981 [gr-qc]].
[16] G. R. Bengochea and R. Ferraro, Phys. Rev. D 79, 124019 (2009) [arXiv:0812.1205 [astro-ph]].
[17] E. V. Linder, Phys. Rev. D 81, 127301 (2010) [Erratum-ibid. D 82, 109902 (2010)]
[arXiv:1005.3039 [astro-ph.CO]].
[18] P. Wu and H. W. Yu, Phys. Lett. B 693, 415 (2010) [arXiv:1006.0674 [gr-qc]];
R. Myrzakulov, Eur. Phys. J. C 71, 1752 (2011) [arXiv:1006.1120 [gr-qc]]; arXiv:1008.4486
[astro-ph.CO];
K. K. Yerzhanov, S. R. Myrzakul, I. I. Kulnazarov and R. Myrzakulov, arXiv:1006.3879 [gr-
qc];
P. Wu and H. Yu, Phys. Lett. B 692, 176 (2010) [arXiv:1007.2348 [astro-ph.CO]];
P. Y. Tsyba, I. I. Kulnazarov, K. K. Yerzhanov and R. Myrzakulov, Int. J. Theor. Phys. 50,
1876 (2011) [arXiv:1008.0779 [astro-ph.CO]];
S. H. Chen, J. B. Dent, S. Dutta and E. N. Saridakis, Phys. Rev. D 83, 023508 (2011)
[arXiv:1008.1250 [astro-ph.CO]];
G. R. Bengochea, Phys. Lett. B 695, 405 (2011) [arXiv:1008.3188 [astro-ph.CO]];
P. Wu and H. W. Yu, Eur. Phys. J. C 71, 1552 (2011) [arXiv:1008.3669 [gr-qc]];
R. J. Yang, Europhys. Lett. 93, 60001 (2011) [arXiv:1010.1376 [gr-qc]]; arXiv:1007.3571 [gr-
qc];

31
J. B. Dent, S. Dutta and E. N. Saridakis, JCAP 1101, 009 (2011) [arXiv:1010.2215 [astro-
ph.CO]];
T. Wang, Phys. Rev. D 84, 024042 (2011) [arXiv:1102.4410 [gr-qc]];
Y. Zhang, H. Li, Y. Gong and Z. H. Zhu, JCAP 1107, 015 (2011) [arXiv:1103.0719 [astro-
ph.CO]];
C. Deliduman and B. Yapiskan, arXiv:1103.2225 [gr-qc];
B. Li, T. P. Sotiriou and J. D. Barrow, Phys. Rev. D 83, 104017 (2011) [arXiv:1103.2786
[astro-ph.CO]];
Y. F. Cai, S. H. Chen, J. B. Dent, S. Dutta and E. N. Saridakis, arXiv:1104.4349 [astro-
ph.CO];
S. Chattopadhyay and U. Debnath, Int. J. Mod. Phys. D 20, 1135 (2011) [arXiv:1105.1091
[gr-qc]];
M. Sharif and S. Rani, Mod. Phys. Lett. A 26, 1657 (2011) [arXiv:1105.6228 [gr-qc]];
H. Wei, X. P. Ma and H. Y. Qi, Phys. Lett. B 703, 74 (2011) [arXiv:1106.0102 [gr-qc]];
arXiv:1108.0859 [gr-qc];
X. h. Meng and Y. b. Wang, arXiv:1107.0629 [astro-ph.CO];
C. G. Boehmer, A. Mussa and N. Tamanini, arXiv:1107.4455 [gr-qc];
S. Capozziello, V. F. Cardone, H. Farajollahi and A. Ravanpak, arXiv:1108.2789 [astro-
ph.CO];
M. H. Daouda, M. E. Rodrigues and M. J. S. Houndjo, arXiv:1108.2920 [astro-ph.CO];
arXiv:1109.0528 [gr-qc]; arXiv:1111.6575 [gr-qc]; arXiv:1202.1147 [gr-qc];
L. R. A. Belo, E. P. Spaniol, J. A. de Deus and V. C. de Andrade, arXiv:1108.3796 [gr-qc];
P. Wu and H. Yu, Phys. Lett. B 703, 223 (2011) [arXiv:1108.5908 [gr-qc]];
C. Q. Geng, C. C. Lee, E. N. Saridakis and Y. P. Wu, Phys. Lett. B 704, 384 (2011)
[arXiv:1109.1092 [hep-th]];
R. Ferraro and F. Fiorini, Phys. Rev. D 84, 083518 (2011) [arXiv:1109.4209 [gr-qc]]; Phys.
Lett. B 702, 75 (2011) [arXiv:1103.0824 [gr-qc]];
H. Wei, arXiv:1109.6107 [gr-qc];
C. Q. Geng, C. C. Lee and E. N. Saridakis, arXiv:1110.0913 [astro-ph.CO];
Y. P. Wu and C. Q. Geng, arXiv:1110.3099 [gr-qc];
P. A. Gonzalez, E. N. Saridakis and Y. Vasquez, arXiv:1110.4024 [gr-qc];

32
 C. G. Boehmer, T. Harko and F. S. N. Lobo, Phys. Rev. D 85, 044033 (2012) [arXiv:1110.5756
[gr-qc]];
H. Wei, X. J. Guo and L. F. Wang, arXiv:1112.2270 [gr-qc].
[19] K. Bamba, C. Q. Geng and C. C. Lee, arXiv:1008.4036 [astro-ph.CO];
K. Bamba, C. Q. Geng, C. C. Lee and L. W. Luo, JCAP 1101, 021 (2011) [arXiv:1011.0508
[astro-ph.CO]].
[20] R. Zheng and Q. G. Huang, JCAP 1103, 002 (2011) [arXiv:1010.3512 [gr-qc]].
[21] B. Li, T. P. Sotiriou and J. D. Barrow, Phys. Rev. D 83, 064035 (2011) [arXiv:1010.1041
[gr-qc]]; Phys. Rev. D 83, 104030 (2011) [arXiv:1012.4039 [gr-qc]];
M. Li, R. X. Miao and Y. G. Miao, JHEP 1107, 108 (2011) [arXiv:1105.5934 [hep-th]];
arXiv:1107.0515 [hep-th].
[22] K. Karami and A. Abdolmaleki, arXiv:1111.7269 [gr-qc]; arXiv:1201.2511 [gr-qc].
[23] K. Bamba and C. Q. Geng, JCAP 1111, 008 (2011) [arXiv:1109.1694 [gr-qc]].
[24] S. Nojiri, S. D. Odintsov and S. Tsujikawa, Phys. Rev. D 71 (2005) 063004
[arXiv:hep-th/0501025].
[25] M. C. B. Abdalla, S. Nojiri and S. D. Odintsov, Class. Quant. Grav. 22, L35 (2005)
[arXiv:hep-th/0409177];
F. Briscese, E. Elizalde, S. Nojiri and S. D. Odintsov, Phys. Lett. B 646, 105 (2007)
[arXiv:hep-th/0612220].
[26] K. Bamba, S. Nojiri and S. D. Odintsov, JCAP 0810 (2008) 045 [arXiv:0807.2575 [hep-th]];
K. Bamba, S. D. Odintsov, L. Sebastiani and S. Zerbini, Eur. Phys. J. C 67 (2010) 295
[arXiv:0911.4390 [hep-th]].
[27] K. Bamba, S. Nojiri, S. D. Odintsov and M. Sasaki, arXiv:1104.2692 [hep-th], to be published
in General Relativity and Gravitation.
[28] S. Nojiri and S. D. Odintsov, Phys. Rev. D 74 (2006) 086005 [arXiv:hep-th/0608008]; J.
Phys. Conf. Ser. 66 (2007) 012005 [arXiv:hep-th/0611071]; Phys. Rev. D 78, 046006 (2008)
[arXiv:0804.3519 [hep-th]].
[29] S. Nojiri, S. D. Odintsov and D. Saez-Gomez, Phys. Lett. B 681, 74 (2009) [arXiv:0908.1269
[hep-th]].
[30] S. Nojiri, S. D. Odintsov and D. Saez-Gomez, arXiv:1108.0767 [hep-th].
[31] P. H. Frampton, K. J. Ludwick and R. J. Scherrer, Phys. Rev. D 84 (2011) 063003

33
 [arXiv:1106.4996 [astro-ph.CO]].
[32] I. Brevik, E. Elizalde, S. Nojiri and S. D. Odintsov, Phys. Rev. D 84 (2011) 103508
[arXiv:1107.4642 [hep-th]].
[33] P. H. Frampton, K. J. Ludwick, S. Nojiri, S. D. Odintsov and R. J. Scherrer, Phys. Lett. B
708, 204 (2012) [arXiv:1108.0067 [hep-th]].
[34] A. V. Astashenok, S. Nojiri, S. D. Odintsov and A. V. Yurov, arXiv:1201.4056 [gr-qc], to
appear in Phys. Lett. B, 2012.
[35] L. N. Granda and E. Loaiza, arXiv:1111.2454 [hep-th];
M. Ivanov and A. Toporensky, arXiv:1112.4194 [gr-qc].
[36] Y. Ito, S. Nojiri and S. D. Odintsov, arXiv:1111.5389 [hep-th].
[37] M. H. Belkacemi, M. Bouhmadi-Lopez, A. Errahmani and T. Ouali, arXiv:1112.5836 [gr-qc],
accepted for publication in Phys. Rev. D; P. Xi, X. h. Zhai and X. z. Li, arXiv:1111.6355
[gr-qc];
A. N. Makarenko, V. V. Obukhov and I. V. Kirnos, arXiv:1201.4742 [gr-qc].
[38] P. H. Frampton, K. J. Ludwick and R. J. Scherrer, arXiv:1112.2964 [astro-ph.CO].
[39] E. V. Arbuzova and A. D. Dolgov, Phys. Lett. B 700, 289 (2011) [arXiv:1012.1963 [astro-
ph.CO]].
[40] A. D. Dolgov and M. Kawasaki, Phys. Lett. B 573, 1 (2003) [arXiv:astro-ph/0307285].
[41] R. R. Caldwell, M. Kamionkowski and N. N. Weinberg, Phys. Rev. Lett. 91, 071301 (2003)
[arXiv:astro-ph/0302506];
B. McInnes, JHEP 0208 (2002) 029 [arXiv:hep-th/0112066];
S. Nojiri and S. D. Odintsov, Phys. Lett. B 562, 147 (2003) [arXiv:hep-th/0303117]; Phys.
Lett. B 571, 1 (2003) [arXiv:hep-th/0306212];
V. Faraoni, Int. J. Mod. Phys. D 11, 471 (2002) [arXiv:astro-ph/0110067];
P. F. Gonzalez-Diaz, Phys. Lett. B 586, 1 (2004) [arXiv:astro-ph/0312579];
E. Elizalde, S. Nojiri and S. D. Odintsov, Phys. Rev. D 70, 043539 (2004)
[arXiv:hep-th/0405034];
P. Singh, M. Sami and N. Dadhich, Phys. Rev. D 68, 023522 (2003) [arXiv:hep-th/0305110];
C. Csaki, N. Kaloper and J. Terning, Annals Phys. 317, 410 (2005) [arXiv:astro-ph/0409596];
P. X. Wu and H. W. Yu, Nucl. Phys. B 727, 355 (2005) [arXiv:astro-ph/0407424];
S. Nesseris and L. Perivolaropoulos, Phys. Rev. D 70, 123529 (2004) [arXiv:astro-ph/0410309];

34
 M. Sami and A. Toporensky, Mod. Phys. Lett. A 19, 1509 (2004) [arXiv:gr-qc/0312009];
H. Stefancic, Phys. Lett. B 586, 5 (2004) [arXiv:astro-ph/0310904];
L. P. Chimento and R. Lazkoz, Phys. Rev. Lett. 91, 211301 (2003) [arXiv:gr-qc/0307111];
J. G. Hao and X. Z. Li, Phys. Lett. B 606, 7 (2005) [arXiv:astro-ph/0404154];
E. Elizalde, S. Nojiri, S. D. Odintsov and P. Wang, Phys. Rev. D 71, 103504 (2005)
[arXiv:hep-th/0502082];
M. P. Dabrowski and T. Stachowiak, Annals Phys. 321, 771 (2006) [arXiv:hep-th/0411199];
F. S. N. Lobo, Phys. Rev. D 71, 084011 (2005) [arXiv:gr-qc/0502099];
R. G. Cai, H. S. Zhang and A. Wang, Commun. Theor. Phys. 44, 948 (2005)
[arXiv:hep-th/0505186];
I. Y. Arefeva, A. S. Koshelev and S. Y. Vernov, Phys. Rev. D 72, 064017 (2005)
[arXiv:astro-ph/0507067];
H. Q. Lu, Z. G. Huang and W. Fang, arXiv:hep-th/0504038;
W. Godlowski and M. Szydlowski, Phys. Lett. B 623, 10 (2005) [arXiv:astro-ph/0507322];
J. Sola and H. Stefancic, Phys. Lett. B 624, 147 (2005) [arXiv:astro-ph/0505133];
B. Guberina, R. Horvat and H. Nikolic, Phys. Rev. D 72, 125011 (2005)
[arXiv:astro-ph/0507666].
[42] Y. Shtanov and V. Sahni, Class. Quant. Grav. 19, L101 (2002) [arXiv:gr-qc/0204040].
[43] J. D. Barrow, Class. Quant. Grav. 21, L79 (2004) [arXiv:gr-qc/0403084].
[44] S. Nojiri and S. D. Odintsov, Phys. Lett. B 595, 1 (2004) [arXiv:hep-th/0405078]; Phys.
Rev. D 70, 103522 (2004) [arXiv:hep-th/0408170]; Phys. Rev. D 72, 023003 (2005)
[arXiv:hep-th/0505215];
S. Cotsakis and I. Klaoudatou, J. Geom. Phys. 55, 306 (2005) [arXiv:gr-qc/0409022];
M. P. Dabrowski, Phys. Rev. D 71, 103505 (2005) [arXiv:gr-qc/0410033];
L. Fernandez-Jambrina and R. Lazkoz, Phys. Rev. D 70, 121503 (2004) [arXiv:gr-qc/0410124];
Phys. Lett. B 670, 254 (2009) [arXiv:0805.2284 [gr-qc]];
J. D. Barrow and C. G. Tsagas, Class. Quant. Grav. 22, 1563 (2005) [arXiv:gr-qc/0411045];
H. Stefancic, Phys. Rev. D 71, 084024 (2005) [arXiv:astro-ph/0411630];
C. Cattoen and M. Visser, Class. Quant. Grav. 22, 4913 (2005) [arXiv:gr-qc/0508045];
P. Tretyakov, A. Toporensky, Y. Shtanov and V. Sahni, Class. Quant. Grav. 23, 3259 (2006)
[arXiv:gr-qc/0510104];

35
 A. Balcerzak and M. P. Dabrowski, Phys. Rev. D 73, 101301 (2006) [arXiv:hep-th/0604034];
M. Sami, P. Singh and S. Tsujikawa, Phys. Rev. D 74, 043514 (2006) [arXiv:gr-qc/0605113];
M. Bouhmadi-Lopez, P. F. Gonzalez-Diaz and P. Martin-Moruno, Phys. Lett. B 659, 1 (2008)
[arXiv:gr-qc/0612135];
A. V. Yurov, A. V. Astashenok and P. F. Gonzalez-Diaz, Grav. Cosmol. 14, 205 (2008)
[arXiv:0705.4108 [astro-ph]];
T. Koivisto, Phys. Rev. D 77, 123513 (2008) [arXiv:0803.3399 [gr-qc]];
J. D. Barrow and S. Z. W. Lip, Phys. Rev. D 80, 043518 (2009) [arXiv:0901.1626 [gr-qc]];
M. Bouhmadi-Lopez, Y. Tavakoli and P. V. Moniz, JCAP 1004, 016 (2010) [arXiv:0911.1428
[gr-qc]].
[45] C. Kiefer, Annalen Phys. 19, 211 (2010).
[46] M. P. Dabrowski and T. Denkiewicz, Phys. Rev. D 79, 063521 (2009) [arXiv:0902.3107 [gr-qc]].
[47] M. P. Dabrowski and T. Denkiewicz, AIP Conf. Proc. 1241, 561 (2010) [arXiv:0910.0023
[gr-qc]].
[48] L. Fernandez-Jambrina and R. Lazkoz, Phys. Rev. D 74, 064030 (2006) [arXiv:gr-qc/0607073];
L. Fernandez-Jambrina, Phys. Rev. D 82, 124004 (2010) [arXiv:1011.3656 [gr-qc]].
[49] V. Sahni, T. D. Saini, A. A. Starobinsky and U. Alam, JETP Lett. 77, 201 (2003) [Pisma Zh.
Eksp. Teor. Fiz. 77, 249 (2003)] [arXiv:astro-ph/0201498].
[50] M. Chevallier and D. Polarski, Int. J. Mod. Phys. D 10, 213 (2001) [arXiv:gr-qc/0009008];
E. V. Linder, Phys. Rev. Lett. 90, 091301 (2003) [arXiv:astro-ph/0208512].
[51] E. W. Kolb and M. S. Turner, The Early Universe (Addison-Wesley, Redwood City, California,
1990).
[52] W. L. Freedman et al. [HST Collaboration], Astrophys. J. 553, 47 (2001)
[arXiv:astro-ph/0012376].
[53] T. Kobayashi and K. i. Maeda, Phys. Rev. D 78, 064019 (2008) [arXiv:0807.2503 [astro-ph]];
Phys. Rev. D 79, 024009 (2009) [arXiv:0810.5664 [astro-ph]];
A. Dev, D. Jain, S. Jhingan, S. Nojiri, M. Sami and I. Thongkool, Phys. Rev. D 78, 083515
(2008) [arXiv:0807.3445 [hep-th]].
[54] K. Bamba, S. Nojiri and S. D. Odintsov, Phys. Lett. B 698, 451 (2011) [arXiv:1101.2820
[gr-qc]].
[55] K. Bamba and C. Q. Geng, JCAP 1006, 014 (2010) [arXiv:1005.5234 [gr-qc]].

36
[56] J. M. Bardeen, B. Carter and S. W. Hawking, Commun. Math. Phys. 31, 161 (1973);
J. D. Bekenstein, Phys. Rev. D 7, 2333 (1973);
S. W. Hawking, Commun. Math. Phys. 43, 199 (1975) [Erratum-ibid. 46, 206 (1976)];
G. W. Gibbons and S. W. Hawking, Phys. Rev. D 15, 2738 (1977).
[57] T. Padmanabhan, arXiv:0910.0839 [gr-qc]; AIP Conf. Proc. 1241, 93 (2010) [arXiv:0911.1403
[gr-qc]]; Rept. Prog. Phys. 73, 046901 (2010) [arXiv:0911.5004 [gr-qc]].
[58] T. Jacobson, Phys. Rev. Lett. 75, 1260 (1995) [arXiv:gr-qc/9504004].
[59] C. Eling, R. Guedens and T. Jacobson, Phys. Rev. Lett. 96, 121301 (2006)
[arXiv:gr-qc/0602001];
E. Elizalde and P. J. Silva, Phys. Rev. D 78, 061501 (2008) [arXiv:0804.3721 [hep-th]];
K. Bamba, C. Q. Geng, S. Nojiri and S. D. Odintsov, Europhys. Lett. 89, 50003 (2010)
[arXiv:0909.4397 [hep-th]];
S. F. Wu, B. Wang, X. H. Ge and G. H. Yang, Phys. Rev. D 81, 044010 (2010) [arXiv:0909.1367
[gr-qc]];
Y. Yokokura, arXiv:1106.3149 [hep-th].
[60] R. Brustein and M. Hadad, Phys. Rev. Lett. 103, 101301 (2009) [arXiv:0903.0823 [hep-th]];
R. Brustein and A. J. M. Medved, arXiv:1201.5754 [hep-th].
[61] K. Bamba, C. Q. Geng and S. Tsujikawa, Phys. Lett. B 688, 101 (2010) [arXiv:0909.2159
[gr-qc]]; Int. J. Mod. Phys. D 20, 1363 (2011) [arXiv:1101.3628 [gr-qc]].
[62] R. G. Cai and S. P. Kim, JHEP 0502, 050 (2005) [arXiv:hep-th/0501055].
[63] C. W. Misner and D. H. Sharp, Phys. Rev. 136, B571 (1964);
D. Bak and S. J. Rey, Class. Quant. Grav. 17, L83 (2000) [arXiv:hep-th/9902173].
[64] S. A. Hayward, Class. Quant. Grav. 15, 3147 (1998) [arXiv:gr-qc/9710089];
S. A. Hayward, S. Mukohyama and M. C. Ashworth, Phys. Lett. A 256, 347 (1999)
[arXiv:gr-qc/9810006];
R. G. Cai and L. M. Cao, Phys. Rev. D 75, 064008 (2007) [arXiv:gr-qc/0611071].
[65] Y. Gong, B. Wang and A. Wang, JCAP 0701, 024 (2007) [arXiv:gr-qc/0610151];
M. Jamil, E. N. Saridakis and M. R. Setare, Phys. Rev. D 81, 023007 (2010) [arXiv:0910.0822
[hep-th]].
[66] K. Bamba, S. Nojiri and S. D. Odintsov, Phys. Rev. D 85, 044012 (2012) [arXiv:1107.2538
[hep-th]].

37
[67] K. Bamba, C. Q. Geng, S. Nojiri and S. D. Odintsov, Phys. Rev. D 79, 083014 (2009)
[arXiv:0810.4296 [hep-th]]; K. Bamba and C. Q. Geng, Prog. Theor. Phys. 122, 1267 (2009)
[arXiv:0909.1249 [astro-ph.CO]].

38