Preprint typeset in JHEP style - HYPER VERSION arXiv:0804.

3811

The Entropy Function for the extremal

Kerr-(anti-)de Sitter Black Holes
arXiv:0804.3811v2 [hep-th] 12 May 2008

Jin-Ho Cho1,2 , Yumi Ko1 and Soonkeon Nam1

1 Department of Physics & Research Institute for Basic Sciences,
Kyung Hee University, Seoul 130-701, Korea
2 Center for Quantum Space Time, Sogang University, Seoul 121-742, Korea

E-mail: [Link]@[Link], koyumi@[Link], nam@[Link]

Abstract: Based on the entropy function formalism, we consider the extremal Kerr-
(anti-)de Sitter black holes in 4-dimensions. Solving differential equations exactly,
which are obtained by extremizing the entropy function, we find agreement of the re-
sult with Bekenstein-Hawking entropy. Concerning the higher derivative corrections,
we extend the computation to the case with Gauss-Bonnet term.

Keywords: Kerr-(anti-)de Sitter black hole, black hole entropy, entropy function,
higher derivative corrections.
Contents
1. Introduction 1

2. Kerr-(anti-)de Sitter Black Holes 3

2.1 Preliminary : Four Dimensional Kerr Black Holes 3
2.2 Horizon Structure of Kerr-(anti-)de Sitter Black Holes 4
2.3 Extremal Kerr-(anti-)de Sitter Case 5
2.4 Hawking Temperatures 9
2.5 Bekenstein-Hawking Entropy 10
2.6 Isometry of the Near Horizon Geometry 10

3. Entropy Function of Kerr-(anti-)de Sitter Black Holes 11

3.1 An Ansatz for the Metric 12
3.2 Euler-Lagrange Equations 13
3.3 Exact Form of the Near Horizon Metric 15
3.4 Extremum Value of the Entropy Function 16
3.5 Effect of the Gauss-Bonnet Term on the Entropy 18

4. Conclusions 19

A. Signs and Ranges of Parameters 20

A.1 Determining the Choice of Sign in β 2 20
A.2 Ranges of Parameters r02 and b 21

1. Introduction
The Bekenstein-Hawking entropy of black holes is a gateway to the quantum theory of
gravity [1, 2]. Although the entropy is given by a simple expression, SBH = A/4GN ,
where A is the area of the black hole event horizon, we do not have a complete
understanding of its microscopic origin.
String theory gives a good microscopic description but its low energy effective
theory gives rise to many additional fields as well as higher order terms for the gravity
part [3]. The entropy of a general black hole in the context of the higher order gravity
is hard to compute even macroscopically when we do not have a concrete form of the
solution.

–1–
 Recently, this problem has been remedied by the entropy function formalism
of Ashoke Sen for the extremal black holes [4, 5]. This allows us to compute the
Bekenstein-Hawking entropy without full knowledge about the metric. The only
necessary information is the near horizon isometry of a black hole. The extremal
Reissner-Nordström black hole in 4-dimensions, which has the isometry SO(1, 2) ×
SO(3), is a simple example where the entropy function formalism is applicable. There
have been several works on non-supersymmetric extremal black holes. The examples
include charged black holes in de Sitter background [6], rotating black holes [7, 8, 9],
and black holes of Nariai class [10]. In these cases, the entropy function formalism
gives the correct values of the Bekenstein-Hawking entropy.
In this paper, we would like to study yet another example, that of rotating
black holes in asymptotically (anti-)de Sitter space-time in 4-dimensions. We are
generalizing the case of rotating black holes in the asymptotically flat space-time [7].
We will see that the entropy function formalism is also applicable in the Kerr-(anti-
)de Sitter black hole case. We will also see the effect of Gauss-Bonnet term in the
Lagrangian on the black hole entropy.
Entropy function formalism is not easy to apply for the rotating black holes.
It is because the near horizon geometry of the rotating black hole is no longer two
dimensional anti-de Sitter space-time ( AdS2 ) times a round two-sphere ( S2 ). Rather,
the near horizon geometry has SO(1, 2)×U (1) isometry, and both the AdS2 part and
the circle ( S1 ) part are warped by a function depending on the latitudinal angle θ.
A rotating black hole has a horizon whose shape is a smooth deformation of a round
sphere. Unlike static cases, metric coefficients are now functions of θ. Despite this
difficulty, the entropy function formalism is still valid for the rotating black holes in
the flat background, as long as the near horizon geometry contains AdS2 part [7].
It is meaningful to check whether the entropy function formalism works for the
rotating black holes in (anti-)de Sitter background. These black holes are more re-
alistic in the sense that we are now in an accelerating universe and most of the
existing black holes are rotating. Exploring Kerr-de Sitter case is also important
on the theoretical side. It provides another example of non-supersymmetric attrac-
tor1 . Regarding the non-supersymmetric attractor, there have been many studies on
non-BPS but extremal black holes in the supersymmetric background [11, 12, 13].
Reissner-Nordström-de Sitter black hole was studied in Ref.[6], as the first example
showing the feature of non-supersymmetric attractor in non-supersymmetric vacuum.
We organize our paper as follows: Sec. 2 discusses general properties of the
extremal Kerr-(anti-)de Sitter black hole. A relation between Bekenstein-Hawking
entropy S, and the angular momentum J, is obtained (See Eq.(2.32)). The near
horizon geometry of the extremal black hole in (anti-)de Sitter spacetime contains
AdS2 symmetry. In Sec. 3, making an ansatz for the near horizon geometry of
1
It is an attractor in the sense that the physical quantities like the entropy are independent of
the asymptotic boundary structure of the geometry.

–2–
the extremal Kerr-(anti-)de Sitter black hole, we write down the entropy function.
We solve the ‘Euler-Lagrange equations’ minimizing the entropy function and obtain
the black hole entropy as an extremum value of the entropy function. We verify
the entropy function formalism for the case at hand, by recovering the relation of
Eq.(2.32) between the entropy and the angular momentum. (See Eq.(3.34)). We
then apply this method to obtain higher order corrections to the entropy. The case
with Gauss-Bonnet term is considered.

2. Kerr-(anti-)de Sitter Black Holes

In this section, we consider the general properties of extremal Kerr-(anti-)de Sitter
black holes such as the horizon structure, Hawking temperature, and the entropy.
In particular, we obtain a relation between the Bekenstein-Hawking entropy and the
angular momentum. The near horizon geometry of these black holes has SO(1, 2) ×
U (1) isometry. Using this isometry, we will write down an ansatz for the near horizon
geometry of rotating black holes in asymptotically (anti-)de Sitter spacetime in the
next section.

2.1 Preliminary : Four Dimensional Kerr Black Holes

Prior to our main issues on the Kerr-(anti-)de Sitter black holes, we will warm up
our discussion by considering the basic properties of the extremal Kerr black hole.
We will observe that the rotation gives the warping of AdS2 and S 1 by functions
depending on θ.
Four Dimensional Kerr metric in Boyer-Lindquist form is given as [14]

∆ ρ2 2 sin2 θ  2 2
ds2 = − (dt − a sin 2
θdφ) 2
+ dr + ρ2
dθ 2
+ (r + a2
)dφ − adt , (2.1)
ρ2 ∆ ρ2

where

ρ2 = r2 + a2 cos2 θ, and ∆ = r2 − 2mr + a2 . (2.2)

Here, the angular momentum parameter a = J/m, where m denotes ADM mass and
J is angular momentum. We use the geometrical unit that sets c = 1 and GN = 1.
With non vanishing angular momentum (a 6= 0) the metric function is not spherically
symmetric.
The near horizon geometry of the extremal case contains the universal factor
AdS2 . When r02 = a2 = m2 , we obtain the extremal Kerr black hole with the event
horizon at r = r0 . Now let us introduce the near horizon coordinates (t̄, r̄, φ̄) as

t̄ t̄
r = r0 + r̄, t= , φ = φ̄ + , (2.3)
 2r0 

–3–
where  is a dimensionless parameter. By shifting φ to φ̄, ∂/∂ t̄ becomes null at the
horizon r̄ = 0. It implies that the coordinate co-rotates with the horizon [15]. By
taking a limit  → 0, we can obtain the near horizon geometry of the extremal Kerr
black holes as
 1 + cos2 θ  r̄2 2r2  4r2 sin2 θ  r̄ 2
ds2 = − 2 dt̄2 + 20 dr̄2 + 2r02 dθ2 + 0 d φ̄ + dt̄ . (2.4)
2 2r0 r̄ 1 + cos2 θ 2r02
By rescaling r̄ as
r̄
r̃ = , (2.5)
2r02
we can rewrite Eq.(2.4) in a form that shows manifestly AdS2 × S1 structures but
warped by functions depending on θ;
 dr̃2  4r2 sin2 θ
2
ds2 = r02 (1 + cos2 θ) −r̃2 dt̄2 + 2 + dθ2 + 0 d φ̄ + r̃d t̄ . (2.6)
r̃ 1 + cos2 θ
2.2 Horizon Structure of Kerr-(anti-)de Sitter Black Holes
In d-dimensional gravity theories with cosmological constant, we have Kerr-(anti-)
de Sitter black holes, that is, a generic form of the stationary black hole solutions.
There are two kinds of parameters characterizing the solutions; the parameter ai
(i = 1, 2, · · ·) concerning various rotations, and the length scale l parametrizing the
size of the (anti-)de Sitter spacetime as
η 2
2
= Λd , (d ≥ 3). (2.7)
l (d − 1)(d − 2)
Here, we have η = +1/ − 1 for de Sitter/anti-de Sitter respectively2 .
Kerr-(anti-)de Sitter solutions in various dimensions are summarized as follows
[16, 17]:
∆r  a 2 ρ 2 ρ2 2
ds2 = − 2 dt − sin2 θdφ + dr2 + dθ
ρ Ξ ∆r ∆θ
∆θ sin2 θ  (r2 + a2 ) 2
+ a dt − dφ + r2 cos2 θ dΩd−4 , (2.8)
ρ2 Ξ
where
 η 
∆r ≡ ∆r (r) = (r2 + a2 ) 1 − 2 r2 − 2mr5−d ,
l
η 2
∆θ = 1 + 2 a cos2 θ,
l
η
Ξ = 1 + 2 a2 ,
l
ρ = r + a2 cos2 θ.
2 2
(2.9)
2
Introducing η, we can have unified expressions for both de Sitter and anti-de Sitter cases. There
is a big difference, however, when we consider the horizon structure. For example, only for de Sitter
case, we have a cosmological horizon.

–4–
We note that the solution approaches Kerr solution as l → ∞ or Schwarzschild-
(anti-)de Sitter solution when a → 0. Here, for simplicity, we just consider the case
with a single parameter a for the rotation. It is assumed to satisfy a2 < l2 so that
∆θ > 0 for any value of θ. When the cosmological constant is non-vanishing, one
cannot give physical meaning to the parameters a and m directly. The meaning will
be given below.
In the case of Kerr-(anti-) de Sitter black holes in 4-dimensions, the conserved
charges (in the spirit of Abbott and Deser [18]) of a rotating black hole are the ‘mass’,
M , and the ‘angular momentum’, J. These quantities, associated with the black hole
event horizon, are related with the parameters m, a and l (of Eqs.(2.8) and (2.9)) as
[19, 20];
m ma
M= , J= 2. (2.10)
Ξ Ξ
One can check that the ‘mass’ M coincides with the ADM mass in the asymptotically
flat spacetime as by limiting l → ∞.
In the rest of this paper, we will focus on the 4-dimensional case only, for sim-
plicity. The horizons are located at the zeros of ∆r (r). We will draw the form of the
function ∆r (r). A simple analysis on the derivative of the function ∆r (r) shows that
it has at most three extremal values in de Sitter case (η = 1) as in Fig. 1, while it
allows only one in anti-de Sitter case (η = −1) as in Fig. 2.
In de Sitter case, the generic form of the function ∆r (r) can be written as follows:
1
∆r (r) = − 2 (r − r− )(r − r+ )(r − rc )(r + r−− ), 0 < r− ≤ r+ ≤ rc . (2.11)
l
The three positive roots at r = r− , r+ and rc correspond to the inner, the outer
and the cosmological horizon respectively, while the other negative root at r =
−r−− (r−− > 0) is considered unphysical.
In anti-de Sitter case, the function ∆r (r) takes generically the following form;
1
∆r (r) = (r − r− )(r − r+ )(r2 + pr + q), 0 < r− ≤ r+ , (2.12)
l2
with p2 − 4q < 0. Here, we have at most two horizons.

2.3 Extremal Kerr-(anti-)de Sitter Case

Let us specify the black holes to be discussed hereon. First, we are interested in the
black hole with the degenerate horizons because its geometry is nicely factorized (thus
making the entropy function analysis manageable) near those degenerate horizons.
In principle, there could be many different cases with the degenerate horizons. For
example in Kerr-de Sitter black holes, there are two different cases; the extremal
black hole with r− coincident with r+ (Fig. 1(b)), or a Nariai type black hole3 if
3
Nariai type black hole is a black hole whose near horizon geometry contains two dimensional
de Sitter space-time as defined in Ref.[10].

–5–
Figure 1: The metric function ∆r (r) describing various horizons of 4-dimensional Kerr-de
Sitter black holes. (a) Non-degenerate horizons (b) The horizons of an extremal Kerr-de
Sitter black hole (c) The horizons of a Nariai type black hole.

Figure 2: Various horizons of 4-dimensional Kerr-anti-de Sitter black holes. (a) Non-
degenerate horizons (b) The horizons of an extremal Kerr-anti-de Sitter black hole.

r+ touches rc (Fig. 1(c)). We will focus on the former case, that is the extremal
black holes. The other case of Nariai type black holes was discussed in Ref.[10]. We
also consider the extremal Kerr-anti-de Sitter black hole, which allows up to one
degenerate horizon, since there is no cosmological horizon. (See Fig. 2(b).)
In each of these cases, the position of horizons can be expressed in terms of the
parameters m, a and l. When ∆r (r) has a double zero at r = r0 , we can factorize
the function as follows:

–6–
  η 
∆r (r) = (r2 + a2 ) 1 − 2 r2 − 2mr
 l
 − 12 (r − r0 )2 (r − rc )(r + (2r0 + rc )) if η = 1,
= l (2.13)
 12 (r − r0 )2 (r2 + 2r0 r + 3r02 + a2 + l2 ) if η = −1.
l
Here, the value η = 1 is for Kerr-de Sitter black holes while η = −1 is for Kerr-
anti-de Sitter black holes. To be more general, let us assume rc as a positive simple
zero of ∆r (r) not necessarily to be the cosmological horizon for the moment. In
this factorization, we can find the relations between the horizons and the parameters
(m, a and l). We will consider de Sitter case and anti-de Sitter case one by one.

(i) de Sitter case (η = 1)

From Eq.(2.13), we find the relations
r04  a2  2
− 1 − r + 2mr0 − a2 = 0, (2.14)
l2 l2 0
4 3  a2 
r − 2 1 − r0 + 2m = 0, (2.15)
l2 0 l2
1 2 2

 a2 
3r0 + 2r r
0 c + r c − 1 − = 0. (2.16)
l2 l2
The extremality heavily restricts the range of r0 . Solving Eq.(2.16) for rc in terms
of r0 , we have
q
rc = −r0 ± −2r02 + (l2 − a2 ) . (2.17)
We take the positive value of rc (the upper sign). For the following range of r0 ,
1
0 ≤ r02 < (l2 − a2 ), (2.18)
6
we have rc > r0 , which corresponds to an extremal black hole. The other case of
rc < r0 corresponds to a Nariai type black hole.
One can go further and write the range of r0 in terms of l2 only. Eqs.(2.14) and
(2.15) enable us to eliminate m, and write r0 in terms of l and a;
r
2 n 2
l a a2 2 12a2 o
r02 = 1− 2 ± 1− 2 − 2 . (2.19)
6 l l l
The lower sign corresponds to the extremal case of our interest. On the other hand,
the upper sign describes a Nariai type black hole.
We observe that the upper bound in Eq.(2.18) is nothing but the point of r02
where a2 takes the extremum value. Indeed it is the very point where the square
root part of Eq.(2.19) vanishes. Here we have the relation,
 √ 
a2 = l 2 7 ± 4 3 . (2.20)

–7–
Figure 3: Parameters in Kerr-de Sitter black holes (a) Mass parameter m as a function
of the position r0 of the degenerate horizon (b) Angular momentum parameter a2 as a
function of the position r02 of the degenerate horizon

For the lower sign in this equation, we can rewrite Eq.(2.19) as follows:
 
2
0≤ r02 <l 2
√ −1 . (2.21)
3

The upper sign of Eq.(2.20) gives negative r02 , so neglected.

Figure 4: Parameters in Kerr-anti-de Sitter black holes (a) Mass parameter m as a func-
tion of the position r0 of the degenerate horizon (b) Angular momentum parameter a2 as
a function of the position r02 of the degenerate horizon

–8–
(i) anti-de Sitter case (η = −1)
From Eq.(2.13), we find the relations

r04  a2  2
2
+ 1 + 2
r0 − 2mr0 + a2 = 0, (2.22)
l l
4 3  a2 
r + 2 1 + r0 − 2m = 0. (2.23)
l2 0 l2
In this case, there is no bound coming from the ‘cosmological’ horizon. However,
there still is a restriction on the range of r0 . First of all, r0 has to be positive. Fig.
4 shows that we must have 0 ≤ r0 < l in order to have m and a2 positive. Also, to
2
satisfy a2 < l2 assumed in Sec. 2.2, we must have 0 ≤ r02 < l3 . Therefore, we can
conclude that range of r0 for the extremal Kerr-anti-de Sitter black hole is

l2
0 ≤ r02 < . (2.24)
3

From Eqs.(2.22) and (2.23) the horizon size of the extremal Kerr-anti-de Sitter
black hole is proportional to
r
2 n 2
2 l a a2 2 12a2 o
r0 = − 1+ 2 ± 1+ 2 + 2 . (2.25)
6 l l l

Only the lower sign gives physical r02 .

2.4 Hawking Temperatures

Let us consider the Hawking temperatures of the extremal Kerr-(anti-)de Sitter black
holes. One simple way of computing Hawking temperature is to use the Euclidean
method [21]. It is given by
|∆0r (rH )|
T = 2
, (2.26)
4π(rH + a2 )
where rH stands for zeros of ∆r (r).
So the extremal black hole has zero temperature. For the extremal Kerr-de Sitter
case, we have two horizons. Each horizon has different temperature. The event
horizon (at r0 ) has zero temperature. However, there is non-vanishing temperature
of the cosmological horizon,

(rc − r0 )2 (rc + r0 )
Tr0 = 0, Trc = . (2.27)
2πl2 (rc2 + a2 )

We call this ‘cold’ black hole. Of course, in the special case of r0 = rc , Trc is also
zero, and this case is called an ‘ultra cold black hole’ [6].

–9–
2.5 Bekenstein-Hawking Entropy
At the event horizon of the 4-dimensional extremal Kerr-(anti-)de Sitter black hole,
the Bekenstein-Hawking Entropy is given as
A πl2 (r02 + a2 )
S= = . (2.28)
4 l2 + ηa2
Eliminating a by using a relation Eq.(2.19) or Eq.(2.25), the above result can be
rewritten as
2πl2 r02
S= 2 . (2.29)
l + 3ηr02
The entropy of an extremal rotating black hole is determined by angular momentum
of the black hole thanks to the relation between mass and angular momentum. Using
Eqs.(2.10) and (2.19) (or (2.25)), one can write the angular momentum as
p
a (r02 + a2 )  η 2  η 2 −2 r02 l2 (l2 − 3ηr02 )(l2 + ηr02 )
J = 1 − 2 r0 1 + 2 a = . (2.30)
2r0 l l (l2 + 3ηr02 )2
We will obtain an algebraic relation between the entropy and the angular momentum.
We can invert Eq.(2.29) and get

l2 S
r02 = . (2.31)
2πl2 − 3ηS
Inserting this into Eq.(2.30) results in

3S 4 − 4ηλ2 S 3 + λ4 S 2 − 4π 2 λ4 J 2 = 0, (2.32)

where λ2 = πl2 . Given an angular momentum, one can determine the entropy. In
principle, one can solve this equation for the entropy. However, we will not give the
messy expression of the entropy here because the above relation is sufficient for our
purpose. At the final stage, we will compare the above relation with the one obtained
by the entropy function formalism.

2.6 Isometry of the Near Horizon Geometry

Let us consider a near horizon geometry of the extremal Kerr-(anti-)de Sitter black
holes. In the near horizon coordinates (t̄, r̄, φ̄) defined as
t̄ aΞt̄
r = r0 + r̄, t= , φ = φ̄ + , (2.33)
 (r02 + a2 )
the near horizon geometry is obtained by taking a limit  → 0. Since ∆r (r) has a
double zero for the extremal case, its expansion starts from the second order in ;
∆00r (r0 )
∆r (r) = (r̄)2 + O(3 ). (2.34)
2

– 10 –
Then the metric of Eq.(2.8) takes the following form;

2 ρ20  ∆00 (r0 ) 2 2 2(r02 + a2 ) dr̄2 

ds = 2 − 2 r̄ dt̄ +
r0 + a2 2(r0 + a2 ) ∆00 (r0 ) r̄2
ρ2 ∆θ (a2 + r02 )2 sin2 θ  2ar0 Ξ 2
+ 0 dθ2 + d φ̄ + r̄d t̄ , (2.35)
∆θ Ξ2 ρ20 (a2 + r02 )2
where
 η

∆00r (r0 ) 2 2
= 2 1 − 2 6r0 + a , ρ20 = r02 + a2 cos2 θ. (2.36)
l
Though in the case of the extremal black hole of which the horizon size r0 and the
angular momentum parameter a are related to each other by Eq.(2.19), we keep both
parameters, for the moment, to make expressions simpler.
Introducing a rescaled radial coordinate r̃ as
∆00 (r0 )
r̃ = r̄, (2.37)
2(a2 + r02 )

we can simplify the metric Eq.(2.35) as

2ρ20  2 2 dr̃2  ρ20 2 ∆θ (r02 + a2 )2 sin2 θ  ar0 Ξ r̃dt̄ 2

ds2 = −r̃ d t̄ + + dθ + d φ̄ + .
∆00 (r0 ) r̃2 ∆θ Ξ2 ρ20 (a2 + r02 )∆00 (r0 )
(2.38)

It clearly shows that the near horizon geometry around the double zero r0 is AdS2 ×
S1 with the isometry SO(1, 2) × U (1).

3. Entropy Function of Kerr-(anti-)de Sitter Black Holes

In this section, we will demonstrate that Sen’s entropy function method works for
Kerr-(anti-)de Sitter black holes. We recover the result of Eq.(2.32) without the
knowledge of the exact form of the solution of the Einstein equation. In Sen’s entropy
function method, all we need for the computation of the entropy of an extremal black
hole is its near horizon geometry. In the static extremal cases, the near horizon
geometry, respecting the assumed isometry of SO(1, 2) × SO(3), is characterized by
some parameters depending on the mass, the charge, and the angular momentum
of the black holes. The near horizon geometry of the stationary black holes is more
complicated than the static cases. The above parameters are generalized to some
unknown functions like f (θ), g(θ), and h(θ), which depend not only on the mass, the
charge, and the angular momentum, but also on the latitudinal angle θ.
In order to carry out Sen’s entropy function method for the extremal Kerr-(anti-
)de Sitter black holes, we first make an ansatz for their near horizon geometries.
The unknown functions f (θ), g(θ) and h(θ) appearing in the ansatz are governed by

– 11 –
the ‘Euler-Lagrange equations’ extremizing the Lagrangian. We then solve the Euler-
Lagrange equations with proper boundary conditions, and determine these functions.
Inserting the results into the entropy function, we obtain the black hole entropy,
which coincides with the Bekenstein-Hawking entropy. Finally we also consider the
correction of the black hole entropy coming from the Gauss-Bonnet term in the
Lagrangian.

3.1 An Ansatz for the Metric

Wald’s entropy formula [22, 23, 24], applied to the near horizon geometry of an
extremal black hole, results in the entropy function. It takes the form of some
Legendre transformation of Lagrangian density (over AdS2 spacetime) with respect
to some parameters characterizing the near horizon geometry [4, 5]. This method
can be generalized to the stationary ones with SO(1, 2) × U (1) near horizon isometry
[7] or to the static cases in de Sitter background [6]. We will verify that the method
is also successful for the case of the stationary black holes in the background with
cosmological constant.
Let us start with the action for the gravity theory in (anti-)de Sitter background,
√
Z
1
S= d4 x −g (R − 2Λ4 ). (3.1)
16π
Based on the near horizon isometry SO(1, 2) × U (1) of extremal stationary black
holes, we make an ansatz for the near horizon geometry as

dr2
ds2 = f 2 (θ) (−r2 dt2 + 2
) + g 2 (θ) dθ2 + h2 (θ) sin2 θ (dφ + α r dt)2 . (3.2)
r
Here α is a dimensionless constant and f, g, h (with the length dimension) are func-
tions of θ. The latitudinal angle θ takes values 0 ≤ θ ≤ π while the azimuthal angle
φ is a periodic coordinate with period 2π.
We give some physical conditions on the metric coefficients, f 2 (θ), g 2 (θ), and
h2 (θ). Since the geometry is symmetric across the equator, we first require that

f 2 (θ) = f 2 (π − θ), g 2 (θ) = g 2 (π − θ), h2 (θ) = h2 (π − θ). (3.3)

Second, in order to make the metric, in Λ4 → 0 limit, approach the near-horizon

metric of the extremal Kerr black hole, the functions, f 2 (θ), g 2 (θ), and h2 (θ) should
be regular and also be non-vanishing. Lastly, we give the following condition

g(θ)h(θ) = k, (3.4)

where k is a nonzero constant carrying the dimension of length squared. This con-
dition means that we keep the volume of the sphere part constant as it is deformed
by the rotation. It will greatly simplify the equations of motion which we will solve

– 12 –
later. We will see this in the next subsection. Furthermore, this accords with the
property of Kerr metric if we let Λ4 → 0.
Let us comment on the boundary conditions. Since the rotation does not distort
the geometry near the north-pole and the south-pole, we have g 2 (θ) = h2 (θ) at
θ = 0, π. Combined with the condition of Eq.(3.4), it leads to;
g 2 (θ) = h2 (θ) = k, at θ = 0, π. (3.5)
Now we have three unknowns f (θ), h(θ) and α to be determined.

3.2 Euler-Lagrange Equations

Given the cosmological constant Λ4 , and angular momentum J which is conjugate
to the parameter α, we will determine the forms of α, f (θ), and h(θ) so that they
extremize the entropy function.
In terms of the angular momentum J, the entropy function, being defined as the
Routhian, takes the form
F = 2π(Jα − L). (3.6)
Here, the function L is the Lagrangian density over two-dimensional anti-de Sitter
spacetime. We write down its explicit expressions and the equations of motion below.
√
Z
1
L= dθdφ −g L
16π
f 2 cot θ 3ηf 2 gh o
Z
1 n
1
= dθ sin θ f 2 h + f 02 h + 2f f 0 h0 + (g 0 h − gh0 ) − gh −
4 g g2 l2
1h
1 iθ=π
− 2f f 0 h sin θ + f 2 h0 sin θ (3.7)
4 g θ=0
The boundary terms arise from integration by parts. In the above, for convenience,
we used simplified notations for functions of θ such as f and h instead of f (θ) and
h(θ), and the notation 0 for the derivative with respect to the coordinate θ. The
black hole entropy is given by the extremum value of the entropy function F . As the
first step to obtain this, we follow the standard variational procedure which results
in the Euler-Lagrange equations as
δF 0
=0 : (f 0 gh) − 2f 0 g 0 h + (2f gh0 − f g 0 h + f 0 gh) cot θ
δf (θ)
α2 g 3 h3 sin3 θ 3ηf g 3 h
+ (g 0 h0 − gh − gh00 ) f + + = 0, (3.8)
4f 3 l2
δF 3ηf 2 g 2 2f f 0 g 0 3α2 g 2 h2 sin2 θ
=0 : g 2 + f 02 + + 2f f 00
− − = 0, (3.9)
δg(θ) l2 g 4f 2
δF 2 02 3ηf 2 g 2 0 2f f 0 h0 α2 g 2 h2 sin2 θ
=0 : g +f + + 2f f cot θ + − = 0, (3.10)
δh(θ) l2 h 4f 2
gh3 sin3 θ
Z
∂F α
=J : dθ = J. (3.11)
∂α 8 f2

– 13 –
Notice that when we combine Eq.(3.9) and Eq.(3.10), we have rather a simple dif-
ferential equation,

00
 0 0  α2 (gh)2 sin2 θ
f − f (ln(gh)) + cot θ − = 0. (3.12)
4f 2
This equation completely decouples from g(θ) and h(θ), if we use the condition
g(θ)h(θ) = k. This condition also simplifies Eq.(3.8). One can justify the validity of
the condition by checking whether the results of the unknowns thus obtained indeed
solve the equations of motion. We have checked that this is the case.
Let us now give the simpler version of the expressions incorporating the condition
g(θ)h(θ) = k as follows:
√
Z
1
L= dθdφ −g L
16π
α2 k 2 sin2 θ  3ηk 2 f 2 o
Z
1 sin θ n 2
= dθ f + f 02 + p + (f f 0
− f 2
cot θ)p 0
− k 2
−
4 k 4f 2 l2
1h iθ=π
− 4f f 0 p sin θ + f 2 p0 sin θ (3.13)
8k θ=0 ,

where p(θ) ≡ h2 (θ). Now we have three unknowns, f (θ), p(θ) and α, satisfying the
following equations of motion:
 2f 0   α2 k 2 sin2 θ 2(f 00 + f 0 cot θ)  6ηk 2
00 0
p + + 3 cot θ p + + − 2 p + = 0,
f 2f 4 f l2
(3.14)
2 2 2
α k sin θ
f 00 − f 0 cot θ − = 0, (3.15)
4f 3
p sin3 θ
Z
αk
dθ = J. (3.16)
8 f2
The variations for the boundary terms are
πh iθ=π
2 0 2 2 0 0
δF = 2( f cos θ + f f sin θ ) δp + f sin θ δp + 4f p sin θ δf . (3.17)
4k θ=0

It vanishes due to the conditions in Eq.(3.3).

Eq.(3.15), written purely in terms of f (θ), is easy to solve despite its nonlinearity.
In terms of x ≡ cos θ and β 2 ≡ α2 k 2 /4, the equation gets simplified as

d2 f (x)
− β 2 f −3 (x) = 0. (3.18)
dx2
This allows us to determine the form of f 2 (θ), up to two integration constants b and
c, as
β2
f 2 (θ) = + b (cos θ − c)2 . (3.19)
b

– 14 –
Here, the constant b should be positive and carries the dimension of the length
squared. The other constant c vanishes because of Eq.(3.3).
Solving Eq.(3.14) is a bit tricky. For this, it is very important to note that the
equation is nothing but the θθ-component of the Einstein equation Rθθ − Λ4 gθθ = 0
applied for the metric ansatz (3.2). (In order to show this, one needs to use Eq.
(3.15).) This is clear in the explicit form of the component Rθθ for the geometry
(3.2):
1 n 4f 00   2f 0 
0 00
o
Rθθ = 2− p− + 3 cot θ p − p . (3.20)
2p f f
As it is, the expression is a second order differential equation for p(θ). One can
rewrite this into a first order differential equation, which is much easier to solve.
The trick is based on the fact that the equation, Rθθ − Λ4 gθθ = 0, is equivalent to
Gθθ + Λ4 gθθ = 0, for the Einstein tensor Gµν = Rµν − 21 Rgµν . Since

f 0 n 0  f 0 β 2 sin2 θ  k2 o
Gθθ = p + − + 2 cot θ p + , (3.21)
fp f f 3f 0 ff0
we have
f0 β 2 sin2 θ  k2 3ηk 2 f
Gθθ + Λ4 gθθ = p0 + − + 2 cot θ p + + = 0, (3.22)
f f 3f 0 ff0 l2 f 0
which is a first order differential equation. Since, f (θ) was already calculated in
Eq.(3.19), the explicit form for p(θ) can be obtained, as we will see in the next
subsection.

3.3 Exact Form of the Near Horizon Metric

The exact solution of Eq.(3.22) is
k2 n β2
2 n cos θ 3η  β 4 2 2 b2 cos4 θ o
p(θ) = 2 −b cos θ − + 2 − 2β cos θ − ,
(β + b2 cos2 θ) sin2 θ b k2 l b2 3
(3.23)
where n is another integration constant which should vanish owing to the condition
of Eq.(3.3).
One can remove other constants β and k further so that only the constant b
remain in the final results of the unknowns. For this, we use first the condition that
the function be regular at θ = 0, π. To cancel the possible singularity at θ = 0, π due
to sin2 θ factor, the numerator must vanish at θ = 0, π. This results in the relation
among the constants;
3β 4 + b(ηl2 − 6b)β 2 − b3 (b + ηl2 ) = 0. (3.24)
This allows us to write β in terms of b;
b  √ 
β2 = 6b − η l2 ± l4 + 48b2 (3.25)
6

– 15 –
and the function p(θ) gets simplified as
k2  β 2 (b l2 + 3ηβ 2 ) η b2
2

p(θ) = + cos θ . (3.26)
β 2 + b2 cos2 θ b2 l 2 l2
The choice of multi-sign in Eq.(3.25) depends on the sign of the cosmological
constant; the upper sign for de Sitter and the lower sign for anti-de Sitter case. The
choice ensures the positivity of β 2 and p(θ)(≡ h2 (θ)). We defer the details to the
appendix A.1. There, we also find that anti-de Sitter case restricts the range of b as
0 < b < l2 /4.
Having these conditions in mind, we can rewrite Eq.(3.25) for both de Sitter and
anti-de Sitter cases at once using an unified notation inserting η as
b  √ 
β2 = 6b − η l2 + η l4 + 48b2 . (3.27)
6
So far, we have obtained form of f (θ) and p(θ) up to unknown constants k and b.
From the boundary condition (3.5), k can be written in terms of b as;
 β 2 (bl2 + 3ηβ 2 ) ηb2 −1 l2 12b + ηl2 − η √l4 + 48b2

k = (β 2 + b2 ) + 2 = . (3.28)
l 2 b2 l 6 (l2 + 4ηb)
Therefore, the unknowns f (θ), p(θ) and α are written only in terms of b as follows:
1 √ 
f 2 (θ) = 6b − ηl2 + η l4 + 48b2 + b cos2 θ,
6
√
2 √
l2 12b + ηl2 − η l4 + 48b2  7b + η l4 + 48b2 + b cos2 θ 
p(θ) = √ ,
6 (4ηb + l2 )2 6ηb − l2 + l4 + 48b2 + 6ηb cos2 θ
√
√
1/2
2β 8ηb + ηl2 l4 + 48b2 6b − ηl2 + η l4 + 48b2
α= = √ . (3.29)
k 24b l2
The interpretation of the constant b is unclear in the case at hand, but we will see
later that it corresponds to the angular momentum J in the limit of zero cosmological
constant.

3.4 Extremum Value of the Entropy Function

Let us remind ourselves that we set p(θ) = h2 (θ) and β 2 = α2 k 2 /4 above and only one
undetermined constant b is remained to be fixed. Now we need to consider Eq.(3.16),
the last Euler-Lagrange equation to be solved. Inserting Eq.(3.29) into Eq.(3.16) we
get a relation between the angular momentum J and the constant b as
p(θ) sin3 θ
Z
β
J = dθ
8 f 2 (θ)
√ √
√
1/2
l2 b l2 − 12ηb + 2 l4 + 48b2 6b − ηl2 + η l4 + 48b2
= √ . (3.30)
3 6 (4ηb + l2 )2

– 16 –
From this relation, we can obtain b written in terms of J although the form is very
complicated. We will not write down here the exact form of b, which is one-page
long, but keep the above relation instead. Eq.(3.29) also can be written in terms of
angular momentum J eventually. It means that all unknowns we introduced in our
ansatz are determined with a given angular momentum J.
Now, we note that the Eq.(3.29) coincides with the metric coefficients of the near
horizon geometry in Eq.(2.38) if we replace b with

l2 r02 (l2 − 3 η r02 )

b= 4 . (3.31)
l − 6 η l2 r02 − 3 r04

Deriving this result, we used the extremal conditions of Eq.(2.19) and Eq.(2.25).
Furthermore, we find that the ranges of b, (for de Sitter case, b > 0 and for anti-de
Sitter case, 0 < b < l2 /4), match exactly those of r02 . See appendix A.2 for details.
Now, we will obtain a relation between the angular momentum J and the ex-
tremal value of the entropy function F of Eq.(3.13);
√

πl2 12 η b + l2 − l4 + 48b2
F = . (3.32)
6 ( 4b + η l2 )

By inverting the equation, we can write b in terms of F as

λ2 (λ2 − 3ηF )F
b= . (3.33)
2π(6F 2 − 6ηλ2 F 2 + λ4 )

Inserting this result into Eq.(3.30), we obtain an algebraic relation between F and
J;
3F 4 − 4ηλ2 F 3 + λ4 F 2 − 4π 2 λ4 J 2 = 0, (3.34)

where λ2 = πl2 . This relation has exactly the same form with Eq.(2.32). Since it is
an algebraic equation, we can conclude that F = S, that is to say, F coincides with
the Bekenstein-Hawking entropy for the extremal Kerr-(anti-)de Sitter black holes.
The results coincide with those of the extremal Kerr black holes if we let l → ∞.
In the limit, we get

4b
f 2 (θ) → b(1 + cos2 θ), h2 (θ) → ,
(1 + cos2 θ)
k → 2b, α = 1, J → b, S → 2πb, (3.35)

which are nothing but those of the extremal Kerr solution. Replacing b with r02 , the
above results are identified with the ones in the near horizon geometry of the Kerr
black holes of Eq.(2.6). Eq.(3.35) agrees with the results of Eq.(4.30) in Ref.[7] if we
set b = J/16π. (Newton’s constant GN = 1/16π was used in Ref.[7].)

– 17 –
3.5 Effect of the Gauss-Bonnet Term on the Entropy
In this section, we will consider the higher order corrections to the entropy of Kerr-
(anti-)de Sitter black holes. We first note that the near horizon ansatz of Eq.(3.2)
works for the case involving higher derivative terms in the Lagrangian[25]. However,
as for the generic higher order terms, the variation of the entropy function leads us to
higher order differential equation, which is difficult to solve. We will consider Gauss-
Bonnet combination of the terms[26], which, being topological in 4-dimensions, does
not affect the equations, so that one can solve it.
The contribution of the Gauss-Bonnet term to the action is
√
Z
1
∆S = d4 x −gLGB
16π
√
Z
ξ
= d4 x −g (RM N P Q RM N P Q − 4RM N RM N + R2 ). (3.36)
16π
Here, ξ is a constant whose dimension is length squared. Using the metric ansatz of
Eq.(3.2), LGB is obtained as
4ξ h α2 k 2 p2 sin2 θ  3   3α2 k 2 sin3 θ 
0 0 02
LGB = 4 2 f sin θ − f cos θ + pp f sin θ −
k f sin θ f3 2 4f 2
i0
+ k 2 p0 sin θ + 2f 02 p2 cos θ + 2 k 2 p cos θ . (3.37)

Inserting this result into Eq.(3.36), we get the Lagrangian density over AdS2 ;
ξ h α2 k 2 p2 sin2 θ  0 3 
0

02 3α2 k 2 sin3 θ 
∆L = f sin θ − f cos θ + pp f sin θ −
2k 3 f3 2 4f 2
iθ=π
+ k 2 p0 sin θ + 2f 02 p2 cos θ + 2 k 2 p cos θ . (3.38)
θ=0

Since it contributes only to the boundary terms, it does not change the equations
of motion of Eqs.(3.14)-(3.16) and therefore the solutions. However, it has a non-
vanishing boundary value which makes the correction to the extremum value of the
entropy function as
2πξ h iθ=π
∆F = − p(θ) cos θ = 4πξ. (3.39)
k θ=0

The total entropy of the extremal Kerr-(anti-)de Sitter black holes with Gauss-
Bonnet correction terms becomes

Stot = S + ∆F
√

πl2 12b + l2 − l4 + 48b2
= + 4πξ . (3.40)
6(4b + ηl2 )
This results in a constant addition to the entropy, with its value independent of any
charges characterizing the black holes. As was discussed in Ref.[6] one can understand

– 18 –
the entropy as a relative quantity so that the constant addition is nothing but a shift
of the reference point of the black hole entropy.
It is very interesting to see that the value of the constant addition to the entropy
is the same as the ones for Reissner-Nordström-(anti-)de Sitter black holes[6] and for
the black holes of Nariai class[10]. We will see later that this universal contribution
of Gauss-Bonnet term to the black hole entropy comes as the result of the topological
nature of Gauss-Bonnet term in 4-dimensions.
Though the parameter ξ introduced in this paper is an arbitrary free parameter,
there is an argument determining its value[27, 28]. If we require the whole action
including the Gauss-Bonnet term be extremal for the class of the asymptotically
(anti-)de Sitter geometries, it should be related to the cosmological constant as

l2 3
ξ=− =− , (3.41)
4η 4Λ4

where η = 1 for de Sitter and −1 for anti-de Sitter. This guarantees the vanishing
of the boundary terms of the variation if the geometry is asymptotically (anti-)de
Sitter4 .

4. Conclusions
In this paper, we extended the validity of the entropy function formalism to the
case of the extremal Kerr-(anti-)de Sitter black holes. The formalism generates the
correct value of the Bekenstein-Hawking entropy. It generalizes the result in flat
case. In l → ∞ limit, we recover the result of Kerr black holes given in Ref.[7] with
b = J/16π (in 16πGN = 1 unit). The result suggests the possibility of extending the
rotating attractor to the case with the cosmological constant by including nontrivial
scalar fields like the dilaton. Despite the non-supersymmetric background and the
warped factorization of the near-horizon geometry, the entropy does not concern the
asymptotic data of the geometry.
The way we showed the validity of the entropy formalism in the extremal Kerr-
(anti-)de Sitter black holes was via Eq.(3.34). The result is exactly the same as
the one in Eq.(2.32) that was obtained from the black hole solution. Although the
algebraic equation, being of fourth order in F , allows four roots in principle, only
one of them corresponds to the value of F in Eq.(3.32). For example, one can see
the relation gets truncated to second order in F in the limit of λ2 = πl2 → ∞, thus
allows two different values F = ±2πJ. Only the upper choice of the sign gives the
value of F in Eq.(3.32), that is, 2πb in l → ∞ limit.
We also showed that the Gauss-Bonnet terms of the Lagrangian adds a constant
correction to the entropy. Since the value is independent of any hair property of the
4
We thank Rodrigo Olea for pointing this out.

– 19 –
black hole, we interpret its meaning just as the shift of the base point of the entropy.
This shift is not physically sensible because what concerns us in some process is the
change of the entropy rather than its value itself.
This feature of the Gauss-Bonnet correction is peculiar in 4-dimensions. Being
topological in nature, the Gauss-Bonnet term of the Lagrangian does not affect the
equation of motion either.
The universal value of the shift in the entropy can be understood with the same
reasoning. The value ∆F = 2πξ in Eq.(3.39) is independent of the angular mo-
mentum. This is because the topological quantity ∆F concerns the topology of the
near horizon geometry, AdS2 ×S2 , which is the same for all extremal black holes in
4-dimensions. (We also have the same topology in the stationary case where the
factorization AdS2 ×S1 is warped by a function of θ.) The value is the same even
for the black holes of Nariai class, where the near-horizon geometry is factorized as
a two-dimensional de Sitter spacetime and a two-sphere [10]. The topology of this
latter near-horizon geometry is the same as that of the extremal black holes. This
is clear in their isometries; SO(1, 2) for two-dimensional anti-de Sitter, and SO(2, 1)
for two-dimensional de Sitter.
In this formalism of computing the black hole entropy, the extremality of the
metric function is still important. The next mission would be to make this ‘extreme’
condition looser. For example, we have to check the attractor behavior of the near-
extremal black holes and finally of the Schwarzschild black hole. This issue will be
treated elsewhere.

A. Signs and Ranges of Parameters

A.1 Determining the Choice of Sign in β 2
In Eq.(3.25), we have a multi-sign in its expression. The choice of the sign is deter-
mined by the conditions that β 2 > 0 and p(θ) = h2 (θ) > 0. Eqs.(3.25) and (3.26)
specify the conditions as
√
6b − η l2 ± l4 + 48b2 > 0, (A.1)
β 2 (b l2 + 3ηβ 2 ) + η b4 cos2 θ > 0. (A.2)
Let us analyze these conditions for de Sitter case and for anti-de Sitter case sep-
arately.

(i) de Sitter case (η = 1)

Since l4 + 48b2 > (6b − l2 )2 , we must choose the upper sign of Eq.(A.1). This
case satisfies Eq.(A.2) automatically so long as b > 0. Therefore for de Sitter case,
we have
b  √ 
β2 = 6b − l2 + l4 + 48b2 , 0 < b. (A.3)
6

– 20 –
(ii) anti-de Sitter case (η = −1)
Eq.(A.2) gives a range for β 2 ;
b  2 √4  b  2 √4 
l − l − 12b2 cos2 θ < β 2 < l + l − 12b2 cos2 θ . (A.4)
6 6
To make sense of the square root, the parameter b should be confined to the region
√
0 < b < l2 / 12. The range guarantees Eq.(A.1) for both signs. However, inserting
Eq.(3.26) into Eq.(A.4), we can refine the range further. We discard the upper sign
of Eq.(3.25), otherwise Eq.(A.4) cannot be satisfied. As for the lower sign, Eq.(A.4)
can be satisfied as long as b < l2 /4. Summing up the result for the anti-de Sitter
case, we have
b √  l2
β2 = 6b + l2 − l4 + 48b2 , 0 < b < . (A.5)
6 4
A.2 Ranges of Parameters r02 and b
We have introduced two different parameters r0 and b. The parameter r0 appeared in
Eq.(2.38) describing the horizon of an extremal Kerr-(anti)de Sitter black hole while
b appeared as an integration constant in the entropy function formalism. Comparing
the near-horizon geometry of Eq.(3.2) (with the unknown functions determined in
Eq.(3.29) via the entropy function formalism,) with the near-horizon geometry (2.38),
one can find the following relation:
l2 r02 (l2 − 3ηr02 )
b= . (A.6)
l4 − 6 η l2 r02 − 3 r04
We showed in Eqs.(2.21) and (2.24) that r0 was restricted to be in a certain range:
 
2
2
0 ≤ r0 < √ − 1 l2 , when η = 1, (A.7)
3
l2
0 ≤ r02 < , when η = −1. (A.8)
3
We want to check whether these conditions on r0 are consistent with the conditions
on b derived in appendix A.1.

(i) de Sitter case (η = 1)

From the range b > 0, we have
l2 r02 (l2 − 3r02 )
> 0. (A.9)
l4 − 6 l2 r02 − 3 r04
Eq.(A.6) gives either

l2 − 3r02 > 0, and l4 − 6l2 r02 − 3r04 > 0, (A.10)

2
or, l − 3r02 < 0, and 4
l − 6l2 r02 − 3r04 < 0. (A.11)

– 21 –
  
Eq.(A.10) gives 0 ≤ r02 < √2 − 1 l2 , which is exactly the same as Eq.(A.7). On
3
the other hand, Eq.(A.11) gives r02 > l2 /3, which corresponds to unphysical solution
as can be seen clearly in Fig. 3(b).

(ii) anti-de Sitter case (η = −1)

The allowed range of b was 0 < b < l2 /4. This leads to the range on r02 ,
2
0 ≤ r02 < l3 , which is exactly the range we have as in Eq.(A.8). Therefore we
conclude that the relation between r0 and b as given in (A.6) is correct.

Acknowledgments
This work is supported by the SRC program of KOSEF through CQUeST with
grant number R11-2005-021 and the Korea Research Foundation Grant funded by
the Korean Government(MOEHRD) (KRF-2007-314-C00056).

References
[1] J. D. Bekenstein, Black holes and entropy, Phys. Rev. D7 (1973) 2333–2346.

[2] S. W. Hawking, Particle Creation by Black Holes, Commun. Math. Phys. 43 (1975)
199–220.

[3] J. Callan, Curtis G., E. J. Martinec, M. J. Perry, and D. Friedan, Strings in

Background Fields, Nucl. Phys. B262 (1985) 593.

[4] A. Sen, Black hole entropy function and the attractor mechanism in higher derivative
gravity, JHEP 09 (2005) 038, [hep-th/0506177].

[5] A. Sen, Entropy function for heterotic black holes, JHEP 03 (2006) 008,
[hep-th/0508042].

[6] J.-H. Cho and S. Nam, Non-supersymmetric Attractor with the Cosmological
Constant, JHEP 07 (2007) 011, [arXiv:0705.2892].

[7] D. Astefanesei, K. Goldstein, R. P. Jena, A. Sen, and S. P. Trivedi, Rotating

attractors, JHEP 10 (2006) 058, [hep-th/0606244].

[8] D. Astefanesei and H. Yavartanoo, Stationary black holes and attractor mechanism,
Nucl. Phys. B794 (2008) 13–27, [arXiv:0706.1847].

[9] G. L. Cardoso, J. M. Oberreuter, and J. Perz, Entropy function for rotating extremal
black holes in very special geometry, JHEP 05 (2007) 025, [hep-th/0701176].

[10] J.-H. Cho and S. Nam, The Entropy Function for the Black Holes of Nariai Class,
JHEP 03 (2008) 027, [arXiv:0711.2514].

– 22 –
[11] K. Goldstein, N. Iizuka, R. P. Jena, and S. P. Trivedi, Non-supersymmetric
attractors, Phys. Rev. D72 (2005) 124021, [hep-th/0507096].

[12] P. K. Tripathy and S. P. Trivedi, Non-supersymmetric attractors in string theory,

JHEP 03 (2006) 022, [hep-th/0511117].

[13] S. Nampuri, P. K. Tripathy, and S. P. Trivedi, On The Stability of

Non-Supersymmetric Attractors in String Theory, JHEP 08 (2007) 054,
[arXiv:0705.4554].

[14] R. H. Boyer and R. W. Lindquist, Maximal analytic extension of the Kerr metric, J.
Math. Phys. 8 (1967) 265.

[15] J. M. Bardeen and G. T. Horowitz, The extreme Kerr throat geometry: A vacuum
analog of AdS2 × S 2 , Phys. Rev. D60 (1999) 104030, [hep-th/9905099].

[16] B. Carter, Hamilton-Jacobi and Schrodinger separable solutions of Einstein’s

equations, Commun. Math. Phys. 10 (1968) 280.

[17] M. H. Dehghani, Kerr-de Sitter spacetimes in various dimension and dS/CFT

correspondence, Phys. Rev. D65 (2002) 104003, [hep-th/0112002].

[18] L. F. Abbott and S. Deser, Stability of Gravity with a Cosmological Constant, Nucl.
Phys. B195 (1982) 76.

[19] R.-G. Cai, Cardy-Verlinde formula and thermodynamics of black holes in de Sitter
spaces, Nucl. Phys. B628 (2002) 375–386, [hep-th/0112253].

[20] S. W. Hawking, C. J. Hunter, and M. Taylor, Rotation and the AdS/CFT

correspondence, Phys. Rev. D59 (1999) 064005, [hep-th/9811056].

[21] G. W. Gibbons and S. W. Hawking, Action Integrals and Partition Functions in

Quantum Gravity, Phys. Rev. D15 (1977) 2752–2756.

[22] R. M. Wald, Black hole entropy is the Noether charge, Phys. Rev. D48 (1993)
3427–3431, [gr-qc/9307038].

[23] T. Jacobson, G. Kang, and R. C. Myers, On black hole entropy, Phys. Rev. D49
(1994) 6587–6598, [gr-qc/9312023].

[24] V. Iyer and R. M. Wald, Some properties of Noether charge and a proposal for
dynamical black hole entropy, Phys. Rev. D50 (1994) 846–864, [gr-qc/9403028].

[25] H. K. Kunduri, J. Lucietti, and H. S. Reall, Near-horizon symmetries of extremal

black holes, Class. Quant. Grav. 24 (2007) 4169–4190, [arXiv:0705.4214].

[26] B. Zwiebach, Curvature Squared Terms and String Theories, Phys. Lett. B156
(1985) 315.

– 23 –
[27] R. Aros, M. Contreras, R. Olea, R. Troncoso and J. Zanelli, Phys. Rev. Lett. 84,
1647 (2000) [arXiv:gr-qc/9909015].

[28] R. Olea, JHEP 0506, 023 (2005) [arXiv:hep-th/0504233].

– 24 –