Entropy in Born-Infeld Gravity

1 1 1
Gokcen Deniz Ozen , Sahin Kurekci , and Bayram Tekin
1
Department of Physics, Middle East Technical University, 06800, Ankara, Turkey
arXiv:1710.01110v1 [hep-th] 3 Oct 2017

Abstract
There is a class of higher derivative gravity theories that are in some sense natural extensions
of cosmological Einsteins gravity with a unique maximally symmetric classical vacuum and only
a massless spin2 excitation about the vacuum and no other perturbative modes. These theories
are of the Born-Infeld determinantal form. We show that the macroscopic entropy as defined by
Wald for bifurcate Killing horizons in these theories are equivalent to the Bekenstein-Hawking
entropy but with an effective gravitational constant which encodes all the information about
the background spacetime and the underlying theory. We carry out the computations in generic
ndimensions including the particularly interesting limits of three, four and infinite number of
dimensions.

Contents
1 Introduction 2

2 Wald Entropy in Generic Born-Infeld Gravity 6

3 Two Special BI gravities 9

4 Born-Infeld extension of New Massive Gravity 10

5 Conclusions 11

6 Acknowledgment 11

dozen@metu.edu.tr

kurekci@metu.edu.tr

btekin@metu.edu.tr

1
1 Introduction

It is perhaps not possible to write a classical theory of gravity that can supersede Einsteins gravity
in elegance and simplicity, which, in its most succinct form, reads in the absence of matter as

Z
g dn x gR = 0. (1)
M

In words, pure general relativity (GR) is a statement about the topological Lorentzian manifold
M: finding the metrics that are critical points of the total scalar curvature on the manifold with
prescribed boundary and/or asymptotic conditions to model various physical cases. This statement
is valid in generic ndimensions but trivial for n = 1 and n = 2. For all other dimensions, pure GR
boils down to searching for solutions of the non-linear partial differential equation R = 0, that is
finding Ricci-flat metrics. In n = 3, the Riemann tensor and the Ricci tensor have equal number of
components (six) and so Ricci-flat metrics are trivially locally Riemann-flat and hence there is no
local gravity besides the location of the sources. But beyond n > 3, the innocent looking equation
(1) and its equivalent partial differential equation form R = 0 are highly complicated to solve
and since their first appearance 100 years ago, many fascinating predictions of these equations,
perhaps the most remarkable ones being the global dynamics of the Universe and the existence of
black holes have been observed. What is rather fascinating about the classical equation R = 0 is
that it leads to thermodynamics type equations such as 1

M = T S + J (2)

for black hole solutions where T = 2 is the horizon temperature and is the surface gravity, is
the angular velocity of the horizon and J is the angular momentum of the black hole while M is
the mass of it. S is the entropy given by the Bekenstein-Hawking formula [1, 2]
AH
S= , (3)
4G
where AH is the horizon area. The temperature T and entropy S are defined up to a multiplicative
constant which can be fixed by the help of a semi-classical analysis such as the Hawking radiation
[3, 4, 5]. But the crucial point is that, classically R = 0 equations for a black hole encode the
equation (2) which is considered as the first law of black hole thermodynamics. In fact, all four
laws of black hole mechanics (or thermodynamics) follow from Einsteins equations [2]. It was also
a rather remarkable suggestion and demonstration that the arguments can be turned around and
thermodynamics of Rindler horizons (with some assumptions) can lead to the Einsteins equations
as the equation of state [6].
One of the current problems in gravity (or quantum gravity) research is to find a possible micro-
scopic explanation of the entropy S in this formula. Namely, from pure statistical mechanical point
of view, S must be given as S = ln N where N refers to the microstates that have the same M, T,
and J values in (2). Namely, a black hole has much more internal states than meets the eye. But
classical considerations suggest that black holes have no classical hair or microscopic degrees of
freedom to account for the entropy. Therefore even though Einsteins theory yields a macroscopic
thermodynamical equation such as (2), it does not seem to explain it. Of course if the classical
picture were to be correct, the four laws of black hole mechanics are not true thermodynamical
1
If Maxwell field is coupled to gravity one has a more general thermodynamical equation including the charge
variation dQ on the right-hand side where is the electric potential on the horizon.

2
equations but just a mere analogy. There is mounting evidence from various considerations (such
as the microscopic degree of freedom counting for extremal black holes in string theory in terms of
D-brane charges [7, 8], in 2+1 dimensional gravity in terms of asymptotic symmetries etc.) that
this is not the case and black holes do have temperature and entropy.
In trying to understand the actual content of equations (2) & (3), it is extremely important to figure
out how universal they are: namely, if the underlying theory is modified from Einsteins gravity
to something else, what remains of these thermodynamical equations? Besides, we also know that
both at large distance scales and short distance scales (low and high energies) GR, in its simplest
form, is not sufficient to explain the observable Universe. These deficiencies are well-known: for
example the accelerated expansion of the Universe and the observed age (or homogeneity and
isotropy) cannot be explained by pure gravity alone, hence there are various attempts to modify
it. Of course, ideally one would like to have a quantum version of gravity in order to explain the
beginning of the Universe or the space and time and the singularity issues of black holes. But at
the moment we do not know exactly what the symmetries, principles etc., of such a theory are. In
any case, the idea is that Einsteins theory captures a great deal about the gravitating Universe or
subsystems (such as the solar system) but it is not yet the final word on it.
In trying to go beyond Einsteins theory there are many routes to follow. One such route that we
have advocated recently in several works [9, 10] is the following: let us write down all the easily
detectable low energy properties of Einsteins theory (1) and try to implement them in the new
modified gravity theory that we are searching for. It is quite possible that at very small distances
gravity (or spacetime) will be nothing like the one proposed by Einsteins theory, so any of its low
virtues may not survive at high energies. For example superstring theory is such an example where
the spacetime or the metric of the gravity is an emergent quantity. But this does not deter us from
our discussion because we shall be interested in the intermediate energy scales in string theory
where Einsteins gravity appears with modified curvature corrections and the pseudo-Riemannian
description of spacetime is accurate.
To accomplish our goal of writing a new modified theory we may list the virtues of Einsteins theory
as follows:
(1) uniqueness of the vacuum: source-free GR, R = 0, has a maximally symmetric solution
which is the Minkowski spacetime. Classically it is the unique vacuum and other consid-
erations, such as the positive energy theorem, suggest that this is the lowest energy state
with zero energy (under the proper decay assumptions and several other physically motivated
assumptions such as the cosmic censorship, dominant energy condition etc.)
(2) masslessness of the spin2 graviton: about its unique vacuum if one expands the metric as
g = + h, one observes that h obeys a massless Klein-Gordon type equation.
(3) diffeomorphism invariance of the theory: the spacetime is a topological Lorentzian manifold
without a prior geometry before gravity is switched on and there are no preferred coordinates
even with gravity switched on.
Certainly one could add further, harder-to-see properties of the theory to this list. For example,
the theory has the local causality property which can be shown in various ways such as proving
that there is always a Shapiro time-delay as opposed to time advance. It has a 3 + 1 dimensional
splitting property and a well-defined initial value (Cauchy) formulation where the full equation
R = 0 can be split into constraints on a spacelike Cauchy surface and evolutions off the surface in
the time direction. But these require further assumptions on the topology of M and we shall not

3
be interested in them here. Our basic working principle will be to write down a modified version of
Einsteins gravity that has the above three properties. The third item on the list is easy to satisfy
but the first two are quite hard.
One of the earliest modifications satisfying the above requirements was introduced by Einstein
when he added a cosmological constant to his theory:

Z
g dn x g(R 2) = 0. (4)
M

What is interesting about this theory is that, in contrast to (1), this one has a dimensionful
parameter which will set the sizes of the gravitating system.2 Now the field equations are
summarized as
2
R = g , (5)
n2
where the number of dimensions n also makes an explicit appearance. All three listed properties
above are also valid for this theory: for > 0 de Sitter (dS) is the unique vacuum and for < 0
anti-de Sitter (AdS) is the unique vacuum, about which there is a single massless spin2 excitation.
Many solutions (such as black holes) of the R = 0 equations also solve (in modified forms) the
cosmological Einsteins theory. Of course currently due to the accelerated expansion of the Universe,
> 0 is preferred. For dS, similar to the black holes, one can define a temperature and entropy
for the cosmological horizon.
Other earlier modifications of Einsteins theory are of the following form

Z
g dn x g (R 20 + R2 + R 2 2
+ R + . . . ), (6)
M

which generically do not have the first two properties in our list: they have many different vacua
which cannot be compared with each other and there are many massive modes besides the massless
spin-2 graviton. Typically some of these modes are ghosts and so the theory is hard to make sense
both at the classical and the quantum levels.
In our search for a higher order gravity theory that has the listed three properties we used the ideas
laid out by Deser and Gibbons [11] who were inspired by Eddingstons idea of using generalized
volumes in the action as

Z q
g, dn x det (g + R ()) g , (7)
M

where one assumes that the metric g and the connection are independent. This type of theories
gained some interest recently (under the title Eddington inspired Born-Infeld gravity even though
strictly speaking Born-Infelds work in electrodynamics came much later than Eddingtons work in
gravity). For this path of the developments, we refer the reader to [12] and a recent review [13].
Deser-Gibbonss suggestion was to consider actions of the type
Z q
g dn x det(g + R (g) + ), (8)
M

where the connection is the unique Levi-Civita connection, not to be varied independently. Of
course the resulting theories are quite different: the action (8) generically is not unitary nor does
2
Note that, Newtons constant is also a dimensionful parameter that arises in the coupling of matter to gravity.
While at this stage we consider pure and classical gravity for the sake of keeping the discussion simple, we shall
restore it below.

4
it have a unique vacuum. Only recently, the most general theory satisfying these requirements was
found in [14, 15] for 3dimensions and in [9, 10] for four and higher dimensions. The 3dimensional
theory is particularly elegant as it is formed by taking the determinant of the Einstein tensor plus
the metric
4m2
r

Z   
3
 
I= d x det g + 2 G 1 g , (9)
G3 M m 2
where G = R 12 g R. Here 2 = 1 and the theory has a unique maximally symmetric vacuum
with an effective cosmlogical constant given as
 
= m2 1 (10)
2
and a massive spin-2 excitation about this vacuum with a mass Mg2 = m2 + . The Wald entropy,
the c function and the c charge of this theory was studied in [10], but to set the stage for the
generic Born-Infeld (BI) gravity in higher dimensions, we reproduce the results and give some more
details on the computations and also compute the second Born-Infeld type extension of new massive
gravity in 3dimensions.
For higher dimensions, the simple form (9) does not lead to a unitary theory, therefore one needs
to add at least quadratic terms in curvature inside the determinant. After a rather tedious com-
putation laid out in detail in [9, 10] the following theory was found:

n
q 
1
Z

I= d x g det( + A ) (0 + 1) , (11)
8G M
where the Atensor reads
 
2
A = R + S + a1 C C + a3 R R + g b1 C 2


+ b2 R . (12)
n
Here, S = R n1 g R is the traceless-Ricci tensor and C is the Weyl tensor given as

2
2R
C = R + g[ R] g[ R] + g g . (13)
n2 (n 1)(n 2) [ ]

Observe that there are three other possible quadratic corrections C R , S S and S
2 which

do not survive since they lead to massive particles and/or destroy the uniqueness of the vacuum.
The theory has five dimensionless parameters , ai , bi and a dimensionful one which is the Born-
Infeld parameter with dimensions of L2 . Unitarity of the excitations of the theory and the
requirement that the theory has a unique viable vacuum with only one massless spin2 graviton
about this vacuum leads to an elimination of three of these parameters. Namely, the following
constraints must be satisfied
(n 1)2 +1
a1 + b1 = , a3 = , b2 = . (14)
4(n 2)(n 3) 4 4
Therefore after all the constraints are applied, there are only two free dimensionless parameters one
of which can be chosen as and a1 or b1 . And for any values of these parameters, the theory has
a massless graviton and no other particle around the unique viable vacuum given by the following
equation
 1n

1 + (0 + 1) +1 = 0, (15)
n2 n2

5
where = 0 is the dimensionless effective cosmological constant. Although looking quite cum-
bersome, remarkably for any n, vacuum equation gives a unique viable solution consistent with
the unitarity requirements. This subtle fact requires a long analysis of the root structure of this
equation and we shall not repeat it here as it was done in the Appendix-C of [10].
As expected, the theory reproduces cosmological Einsteins gravity at the first order expansion
in curvature; it reproduces Einstein-Gauss-Bonnet theory at the second order expansion and a
particular cubic and quartic theory in the next two order expansions. The important point here
is that, all these perturbative theories are unitary on their own with the same particle content as
Einsteins theory, namely they just have one massless spin2 excitation about their unique vacuum
solution. Moreover, the theory keeps all of its properties at any truncated order in an expansion in
powers of the curvature. The form of the vacuum equation and the effective gravitational constant
are modified up to and including the O(R4 ) expansion but beyond that, namely at O(R4+k ) with
k 1, the structure of the vacuum equation or the propagator of the theory is not affected. This
is due to the fact that the action is defined with a determinant operation and only up to second
order terms are kept inside the determinant.
For the Wald entropy calculation of this model, the following exact relation will be employed:
4 2n 4
C C = R( R) + R( ) R 2
R ( R) + RR
n2 (n 2) (n 2)2
 
2 2 1
+ 2
R R2 g , (16)
(n 2) n1
whose contraction yields
2 2 4 2 2
C = R R + R2 . (17)
n2 (n 1)(n 2)

The layout of this paper is as follows: In section 2, we define the Wald entropy for a generic theory
and carry out the computation of the ndimensional Born-Infeld gravity (11). In section 3, we
study a minimal version of BI gravity and consider the n limit of the theory and show that
the resulting exponential gravity has the same area-law for the entropy. In section 4, we study the
3dimensional Born-Infeld gravity which has a massive graviton in its perturbative spectrum.

2 Wald Entropy in Generic Born-Infeld Gravity

In this paper, we shall consider the locally (A)dS metrics, but for the sake of generality and possible
further work, in this section we consider a general spherically symmetric and static metric and give
its curvature properties which will be relevant to black holes as well as (A)dS spacetimes. Therefore
we consider the following metric
n2
1 X
ds2 = a(r)2 dt2 + dr 2
+ r 2 hij (xk )dxi dxj , (18)
a(r)2
i,j=2

where hij (xk ) is the induced metric on (n 2)dimensional sphere. The Christoffel symbols for
this metric are

0 a 1 3 1 a 1 2 i
ji
10 = , 00 =a a, 11 = , ij = ra hij , 1j = , (19)
a a r

6
where a is the shorthand notation for a(r) and prime denotes derivative with respect to the radial
coordinate r. From the definition of the Riemann tensor

R = + , (20)

and its contractions, the corresponding curvature terms follow

ra
R0101 = (a )2 + aa , R0i0j = ra3 a hij ,
R1i1j = hij ,
 a 
2 2 2 2 n2
Rijkl = r (a 1)(hik hjl hil hjk ), R00 = a (a ) + aa + aa ,
r
 
1 2 n2

Rij = (n 3)(a2 1) + 2raa ) hij ,
 
R11 = 2 (a ) + aa + aa ,
a r
 
2 4(n 2) (n 2)(n 3) 2
R = 2(a ) + 2aa + aa + (a 1) . (21)
r r2

Note that since de Sitter spacetime is maximally symmetric its curvature terms can be written
collectively as
2 2 2n
R = (g g g g ) , R = g , R= , (22)
(n 1)(n 2) n2 n2

which will be very useful through the calculations.

Now we turn to the computation of Wald entropy for the BI theory (11). Recall that [16, 17, 18, 19]
Wald entropy arises as a Noether charge of diffeomorphism symmetry of the theory and for a generic
action of the form
1
Z
I= dn x gL (gab , Rabcd ) , (23)
16G M
it reads as I  
L
SW = 2 ab cd dVn2 , (24)
Rabcd
where r = rH denotes the bifurcate Killing horizon and the integral is to be evaluated on-shell. The
volume element dVn2 is the volume element on the bifurcation surface and ab are the binormal
vectors to . For the metric (18) the binormal vectors can quite easily be found as

01 = 1, 10 = 1, (25)

by using the timelike Killing vector (1, 0, 0, . . .) in the formula

1
ab = a b , (26)
k
q
where k is defined as k = 21 ( )( ).

To find the Wald entropy of the de Sitter spacetime which is given by the metric
1
2r 2 2r 2
  
2 2
ds = 1 dt + 1 dr 2 + r 2 d2n2 , (27)
(n 1)(n 2) (n 1)(n 2)

7
we needpto compute the derivative of the
p Lagrangian which
 1 can be done as follows. By defining
f (A) = det(g + A ) and using det M = exp 2 Tr ln M we obtain

L   A

= f (A) (g + A)1 . (28)
Rabcd Rabcd
Here, when calculating the derivative one needs to use
2 2
R
R R
= g c[a gb]d , = g a[c Rd]b g b[c Rd]a , = 2Rabcd , (29)
Rabcd Rabcd Rabcd

and it is a lot easier if A is written in Riem-Ric-R (RRR) basis as it is done in [9]. By denoting a
bar on top of the background elements we can evaluate all the terms in de Sitter background with
the help of (22). The derivative of the Atensor reads
   
A 1+ b d ac b c ad a d bc a c bd

= + a3 ( ) g ( ) g ( ) g + ( ) g
Rabcd R 4 n2
 
2 
+ + b2 g ac g bd gad gbc g , (30)
2n n(n 2)

where = > 0. The value of the other tensors can be computed as

g + A = g (1 + ) , (31)

2 42
where = n2 + (n2)2 (a3 + b2 ) and finally one has

f (A)R = (1 + )n/2 . (32)

Using these with the implementation of the binormal vectors we obtain the Wald entropy for the
ndimensional de Sitter spacetime as
 n2
42
 
AH 4 2 2
SW = 1+ (a3 + b2 ) 1+ + (a 3 + b 2 ) . (33)
4G n2 n 2 (n 2)2

After inserting the constraints (14) on a3 and b2 for the unitarity of the theory, Wald entropy takes
the form  n1
AH
SW = 1+ , (34)
4G n2
over which we can easily take two important limits of small and large
AH
lim SW = = SBH , (35)
0 4G
 n1
AH
lim SW = n1 . (36)
4G n 2

As expected when the Born-Infeld parameter is taken to be very small the theory gives the results
of Einsteins theory and the Bekenstein-Hawking and Wald results match. In the large limit the
crucial result to note is the increase in entropy when increases, i.e. when we add more curvature
to the spacetime, entropy increases dramatically with the cubic order in four dimensions.

8
The effective gravitational constant of the theory again might be calculated from the derivative of
the Lagrangian with respect to Riemann tensor: 3
 
1 1 R L
= , (37)
eff G R R R
which in our case gives
 n1
1 1
= 1+ . (38)
eff G n2
This result has some important implications since it allows us to rewrite the Wald entropy in a
nicer form
AH
SW = , (39)
4eff
which is consistent with the results of [21]. Namely, as far as the dynamical entropy is concerned,
BI gravity that has the same particle content of Einsteins theory has the Bekenstein-Hawking form.
The only change is that instead of G, eff must be used. And since eff < G for dS, as can be seen
from (38), the entropy is increased due to the high curvature terms.

3 Two Special BI gravities

Observe that the Wald entropy (34) does not depend on the parameters , a3 and b2 . Namely,
these parameters are not constrained further by the positivity of the Wald entropy. Obeying the
constraints they must satisfy, one can write down a minimal theory without any parameters which
reads as [9]
" n/2 #
2 2 (n 1)2 2 (n 2) 2
L= 1+ R+ GB R 0 1 , (40)
n 4n(n 2)(n 3) 4n2

where GB = R 2 2 + R2 is the Gauss-Bonnet term and the Wald entropy for the dS
4R
spacetime is still given by (34). This theory has further studied in [22] where it was shown that
the Schwarzschild-Tangherlini black hole is only an approximate solution.
General relativity in the large number of dimensions has some interesting properties (see [23]). Here
we can also consider the n limit of the BI gravity given by (40) to get [9]
2  2
   
2 2
L= exp R + R 4R 0 1 , (41)
2 8
where the Wald entropy reads as
    
AH 2(2n 3) n (2n 3)
SW = 1 exp 1 . (42)
4G (n 1)(n 2) n2 (n 1)(n 2)

Here, another important limit, the largen limit, comes into the play and also the results of other
two limits (small and large limits) change in a slighter way:
AH
lim SW = = SBH , (43)
0 4G
AH
lim SW = e . (44)
n 4G
3
For more details on this see [20].

9
 AH 1 e
Note that one can easily check the relation SW = 4eff where eff can be found as eff = G by
taking the n limit in (38).

4 Born-Infeld extension of New Massive Gravity

As we noted in the Introduction, the Born-Infeld gravity in 2 + 1 dimensions has a particularly

simple form which was obtained as an infinite order expansion of the quadratic new massive gravity
[24, 25, 26]. BI theory has the nice property that it coincides at any order in curvature expansion
with the theory obtained from the requirements of AdS/CFT [27, 28] and moreover it appears as
a counter term in AdS4 [29]. Both BI and new massive gravity describe a massive spin2 particle
with 2 degrees of freedom in 2 + 1 dimensions and have very rich solution structures, including the
BTZ black hole and various other black holes.
For the theory given in (9), we can define the Wald entropy of the horizon of the BTZ black hole
in terms of the derivative of the Lagrangian with respect to the Ricci tensor as
I  
L
SW = 2 g d3 x. (45)
R
To carry out the computation, let us define
r  
f (R ) = det g + 2 G , (46)
m
which yields
1
 
f
+ g g ,


= f g + 2G (47)
R 4m2 m
R 1
+ .


where the derivative of the Ricci tensor is taken by respecting its symmetry, R = 2
Locally the BTZ solution is AdS with the following curvature values
R = (g g g g ) , R = 2g , R = 6. (48)
From here one can calculate the following relevant quantities
 
1
g + 2 G = 1 2 g, (49)
m m

f (R )R = g(1 2 )3/2 , (50)
m
which leads to   r
L
= 1 2 g . (51)
R R 16G3 m
Since the binormal vectors are the same as before, one arrives at
r
AH
SW = 1 2, (52)
4 G3 m
which is positive for = +1. To put this result into the Bekenstein-Hawking form, let us compute
the effective gravitational constant which reads
  r
1 1 R L
= = 1 2, (53)
eff G3 R R R G3 m

10
 AH
leading to the relation SW = 4eff . Since < 0, higher curvature terms increases the entropy.
As it is noted in [14], there is another theory that can be considered as an extension of the new
massive gravity whose action reads
s !
4m2
  
1
Z

I = d3 x det g + 2 R g R g . (54)
G3 m 6

By carrying out the analogous computations for this action one arrives at the Wald entropy of the
BTZ solution as r
AH
SW = 1 + 2, (55)
4 G3 m
AH
where this time entropy is positive for = 1 and the relation SW = 4eff is again satisfied with
q

effective gravitational constant calculated as 1eff = G3 1 + m2.

5 Conclusions

We have studied the Wald entropy for Born-Infeld type gravity theories which were constructed to
carry the vacuum and excitation properties of Einsteins gravity. Namely, these BI theories have
a unique maximally symmetric solution (despite the many powers of curvature) and a massless
spin2 graviton about the vacuum for n 3 + 1 dimensions (and a massive graviton for n =
2 + 1 dimensions). Wald entropy (dynamical entropy coming from diffeomorphism invariance)
usually differs from the Bekenstein-Hawking entropy for generic gravity theories. But here we have
shown that these two notions of entropy coincide for BI gravity theories in generic dimensions with
the slight modification that instead of the Newtons constant of Einsteins theory, the effective
gravitational coupling appears. Basically the following expression is valid:
AH
SW = SBH = , (56)
4eff
where eff encodes all the information about the underlying theory and the background spacetime.
This is a rather interesting result and one might conjecture that it might be a property of gravity
theories that have only massless gravity in their spectrum. (The 3dimensional case is somewhat
different as there is no massless bulk graviton and one necessarily considers massive gravity). To
be able to understand this potentially deep connection between various notions of entropy and the
particle content of the underlying gravity theory, further solutions of BI gravity theories must be
constructed. This work is underway.

6 Acknowledgment

S.K. is supported by the TUBITAK 2211-A Scholarship.

References

[1] J. D. Bekenstein, Black Holes and Entropy, Phys. Rev. D 7, 2333 (1973).

11
[2] J. M. Bardeen, B. Carter and S. W. Hawking, The four laws of black hole mechanics, Com-
mun. Math. Phys. 31, 161 (1973).
[3] Ya. B. Zeldovich, Generation of Waves by a Rotating Body, JETP Lett. 14, 180 (1971).
[4] S. W. Hawking, Black hole explosions?, Nature 248, 30 (1974).
[5] S. W. Hawking, Particle creation by black holes, Commun. Math. Phys. 43, 199 (1975).
[6] T. Jacobson, Thermodynamics of spacetime: the Einstein equation of state, Phys. Rev. Lett.
75, 1260 (1995).
[7] A. Strominger and C. Vafa, Microscopic origin of the Bekenstein-Hawking entropy, Phys.
Lett. B 379, 99 (1996).
[8] A. Sen, Microscopic and macroscopic entropy of extremal black holes in string theory, Gen.
Rel. Grav. 46, 1711 (2014).
[9] I. Gullu, T. C. Sisman and B. Tekin, Born-Infeld Gravity with a Unique Vacuum and a Massless
Graviton, Phys. Rev. D 92, 104014 (2015).
[10] I. Gullu, T. C. Sisman and B. Tekin, Born-Infeld Gravity with a Massless Graviton in Four
Dimensions, Phys. Rev. D 91, 044007 (2015).
[11] S. Deser and G. W. Gibbons, Born-Infeld-Einstein actions?, Class. Quant. Grav. 15, L35
(1998).
[12] T. C. Sisman, Born-Infeld Gravity Theories in D-dimensions, PhD Thesis, Middle East
Technical University (2012).
[13] J. B. Jimenez, L. Heisenberg, G. J. Olmo and D. Rubiera-Garcia, Born-infeld inspired modi-
fications of gravity, [arXiv:1704.03351].
[14] I. Gullu, T. C. Sisman and B. Tekin, Born-Infeld Extension of New Massive Gravity, Class.
Quantum Grav. 27, 162001 (2010).
[15] I. Gullu, T. C. Sisman and B. Tekin, c-functions in the Born-Infeld extended new massive
gravity, Phys. Rev. D 82, 024032 (2010).
[16] R. M. Wald, Black Hole Entropy is Noether Charge, Phys. Rev. D 48, 3427-3431 (1993).
[17] V. Iyer and R. M. Wald, A Comparison of Noether Charge and Euclidean Methods for Com-
puting the Entropy of Stationary Black Holes, Phys. Rev. D 52, 4430-4439 (1995).
[18] T. Jacobson, G. Kang and R. C. Myers, On Black Hole Entropy, Phys. Rev. D 49, 6587-6598
(1994).
[19] T. Jacobson and R. C. Myers, Entropy of Lovelock Black Holes, Phys. Rev. Lett. 70, 3684-
3687 (1993).
[20] H. Adami, M. R. Setare, T. C. Sisman and B. Tekin, Conserved charges in extended theories
of gravity (To appear)
[21] R. Brustein, D. Gorbonos and M. Hadad, Walds entropy is equal to a quarter of the horizon
area in units of the effective gravitational coupling, Phys. Rev. D 79, 044025 (2009).
[22] A. Karasu, E. Kenar and B. Tekin, Minimal extension of Einsteins theory: The quartic grav-
ity, Phys. Rev. D 93, 084040 (2016).

12
[23] R. Emparan, R. Suzuki and K. Tanabe, The large D limit of General Relativity, JHEP 06,
009 (2013).
[24] E. A. Bergshoeff, O. Hohm and P. K. Townsend, Massive Gravity in Three Dimensions,
Phys. Rev. Lett. 102, 201301 (2009).
[25] E. A. Bergshoeff, O. Hohm and P. K. Townsend, More on massive 3D gravity, Phys. Rev.
D 79, 124042 (2009).
[26] I. Gullu, T. C. Sisman and B. Tekin, Canonical Structure of Higher Derivative Gravity in
3D, Phys. Rev. D 81, 104017 (2010).
[27] A. Sinha, On the new massive gravity and AdS/CFT, JHEP 1006, 061 (2010).
[28] M. F. Paulos, New massive gravity extended with an arbitrary number of curvature correc-
tions, Phys. Rev. D 82, 084042 (2010).
[29] D. P. Jatkar and A. Sinha, New Massive Gravity and AdS4 counterterms, Phys. Rev. Lett.
106, 171601 (2011).

13