Entropy of Black Holes in \mathcal{N}=2 Supergravity

Entropy of Black Holes in Supergravity


Using the techniques of the isolated horizon formalism, we construct space of solutions of asymptotically flat extremal black holes in pure supergravity in dimensions. We prove the laws of black hole mechanics. Further, restricting to constant area phase space, we show that the spherical horizons admit a Chern- Simons theory. Standard way of quantizing this topological theory and counting states confirms that entropy is indeed proportional to the area of horizon.

Supergravity, Black Holes, Entropy

1 Introduction

Black holes are “the simplest macroscopic objects” made out of spacetime [1]. This stimulates the hope that black holes will turn out to provide crucial clues for a quantum theory of gravity just as the hydrogen atom helped us unravel the secrets of the atomic system. The most important development in the last few decades was to establish that black holes behave as macroscopic states in thermal equilibrium and their dynamical laws (laws of black hole mechanics) being similar to laws of thermodynamics [2]. This observation prompted Bekenstein and Hawking to argue that black holes indeed have temperature related to the surface gravity and their entropy is related to the area of the black hole horizon [3, 4]. However, one needs statistical interpretation for such thermodynamic arguments. It is expected that any quantum theory of gravity should be able to specify the microstates of the black hole spacetime and the leading term in Boltzmann definition of entropy would be proportional to the area of the horizon. Supergravity and string theories are leading candidates of quantum theory. Black holes occurring in these theories are subjects of intense study. Moreover, these solutions can be interpreted as self-gravitating solitons interpolating between different vacua of the theory [5, 6, 7]. The extremal Reissner-Nordstrom solution arising in pure supergravity in -dimensions is the simplest example [8, 9, 10, 11, 12]. The solitons in string theory (and low energy effective actions) play a central role in understanding string dualities. Alternatively, string theory has been used to investigate quantum properties of extremal black holes [13, 14, 15, 16, 17]. Extremal black holes are BPS solitons as they have some residual supersymmetry of the extended supersymmetric theories (see [18] for other interpretations of the term BPS). Using this BPS property, one can do computations like entropy for example, in perturbative regime and extrapolate to non-perturbative region. These BPS solutions have high degree of supersymmetry as isometries and this shields the counting of states over large variation of modular parameters. More solutions of the BPS type exist in higher dimensions with various degree of supersymmetry. They have proved to be of great interest for establishing various subtleties related to the entropy calculation (see [19, 20, 21, 22, 23]).

A popular way of finding the ‘macroscopic’ entropy of black holes (in different theories of gravity) is to use the Killing horizon formalism and the Wald formula [24, 25]. However, it is well known that this approach is difficult to implement convincingly for extremal black holes (see [26, 27]). Indeed, the laws of mechanics depend on the assumption that the Killing horizon admit bifurcation spheres (see [24, 28]). These are special spheres which lie in the intersection of the past and the future horizon. Wald’s proposal for interpreting the Noether charge (corresponding to the diffeomorphism invariance associated with the Killing vector field generating the horizon) as the “entropy” crucially depends on the existence of the bifurcation sphere [24, 25, 26]. But, extremal black holes are not past complete and hence do not have any bifurcation sphere. This implies that the proof of the laws of mechanics and the determination of the entropy of these black holes via the Wald formula remains a suspect. One might consider taking extremal limit of the entropy of corresponding non-extremal black holes. Arguments presented above also question such procedure. Consider the space of solution (of any theory of gravity) containing Killing horizons. If the laws of black hole mechanics hold in this phase space, then it does not contain extremal black holes. The extremal black holes do not exist as a limit point in this phase space. In other words, no sequence of solution can be constructed in this space of solution which will have their surface gravity limiting to zero. So in this phase space, taking extremal limit of the Noether charge for non-extremal black holes remains ill defined. There are at least two ways to solve this problem: first, to find a way to deal with extremal black holes in the Wald formulation or second, to consider the isolated horizon formalism which supports extremal horizons in its phase space. Our aim in this paper is to show that the second possibility can be naturally used to calculate the entropy of spherical black holes in pure supergravity.

Isolated horizon (IH) [29, 30, 31, 32, 33, 34, 35] is a local definition of black hole horizon. Unlike event horizons or the Killing horizon, this definition does not require the knowledge of spacetime external to the horizon. The knowledge of the entire spacetime is redundant . IH boundary conditions are the minimal set of conditions that any generic black hole horizon (extremal/nonextremal) is expected to satisfy. The most important characterization of IH is that they are expansion-free. This condition separates an arbitrary null surface from a black hole horizon. For example, the Minkowski light cone expands in the future (and also in the past) and hence is not an IH. This is expected because the Minkowski null-cone behaves as a horizon only for the Rindler observers. The condition of expansion-freeness implies that no matter field falls inside the horizon. Moreover, the boundary conditions imply that there exists Killing vectors fields on the black hole horizon only. Thus there might be radiation arbitrarily close to the horizon but are not allowed to cross it. This allows a large class of solutions to satisfy the IH boundary conditions. Indeed, the space of solutions of any theory of gravity admitting an IH as an inner boundary is larger than that of the Killing or the event horizon (these definitions require Killing vector fields outside the horizon too). The original formulation of IH however, had extremal and non- extremal horizons in distinct phase spaces. The Weak Isolated Horizon (WIH) uses a set of milder conditions than IH and puts these two classes of solutions in equal footing [35, 27]. The IH (or the WIH) formulation does not require bifurcation spheres to establish the laws of black holes mechanics or to determine the entropy. In other words, the IH (or the WIH) formlation provides the ideal set-up to study mechanics of extremal black hole horizons arising in string theory or supergravity. The IH formulation is also useful to compute the entropy of horizons in loop quantum gravity (LQG) approach. The point of view of this approach is that the horizon supports the effective degrees of freedom that arise out of a well defined interaction between the bulk and the boundary configurations. These microstates residing on the boundary capture all the essential features of the spacetime. In other words, the microstates relevant for entropy counting are localized on the horizon only. It is also not very difficult to guess the nature of the theory on horizon. Since the surface is null, it does not support a metric theory. It is only natural that the effective theory on this null surface be a topological theory. The theory turns out to be a Chern-Simons theory [36]. Entropy of the horizon can then be obtained by directly geometrically quantizing this Chern-Simons theory [37, 38]. Alternatively, the authors of [39, 40] used techniques of conformal field theory to obtain the entropy and corrections to all orders in Planck length.

We shall use the IH formalism approach to deal with the black hole solutions in supergravity. In [35], it was proposed that the IH formalism introduces an ideal set-up to study black hole solutions in string theory and supergravity. In [41, 42, 43], the authors formulated the precise notions required to make IH amiable to supergravity solutions. However, the detail compatibility study between the IH boundary conditions and the conditions arising from the supergravity theory is still to be carried out. Moreover, the derivation of the laws of black hole mechanics and the calculation of entropy (as is done for IH in GR) are needed for detailed understanding of the non-perturbative aspects of spacetime. The aim of the present paper is to fill this gap.

We will be interested in classical spacetimes which are purely bosonic solutions of supergravity theories. The solutions which retain some supersymmetry of the theory are lebelled BPS saturated. The BPS condition bounds the mass (measured at infinity) from below by a function of the asymptotic charges of the fields (for , by charges of the Maxwell field) (however, see [18]). When the bound is attained, the classical spacetime admits Killing spinor fields. The supersymmetry transformations generated by these Killing spinors are such that the bosonic fields are left invariant while the supersymmetry transformations of the fermionic fields vanish. These conditions on the fields can be turned into a set of first order differential equation called the Killing spinor equations (KSE)2. Alternatively, given the KSEs for a supergravity theory, their solution leads to classical configurations with unbroken supersymmetry. However, only a few of these configurations actually solve the corresponding supergravity equations of motion. Solutions of KSEs for different supergravity theories are of great interest.

Our main interest in this paper is to study the black holes in supergravity using the IH formalism. As a concrete example, we shall consider the extremal Reissner-Nordstrom black hole solution in supergravity. This is also a well solution in the Einstein-Maxwell theory which in fact, is a consistent truncation of the supergravity. For these global solutions, the mass (and charge) measured by an asymptotic observer equals the mass (and charge) defined on the horizon. Thus, this particular solution is supersymmetric or BPS for any observer, at asymptopia or at the horizon. These kind of solutions will be called globally supersymmetric. In the context of the IH formalism, the most general cases are those where one has access only to mass and charge defined locally on the horizon and does not have any knowledge of the nature of the exterior spacetime. In that case, it is natural to consider solutions (or configurations) which saturate the BPS bound solely on the horizon, i.e. local horizon mass equal the charge of the field equated on the horizon. Indeed, IH formalism can incorporate solutions which are supersymmetric on the horizon only while the bulk spacetime may have no residual supersymmetry of the theory. We repeat that, while it is enough for the IH formalism to require KSEs to hold just on the horizon, the configurations which solve the KSEs globally (like the extremal RN solution) will also naturally be part of the IH phase space.

As mentioned before, we shall investigate the applicability of the IH formalism for black hole solutions arising in supergravity. We will use the global and supersymmetric Reissner-Nordstrom solution arising in pure supergravity for consistency study. First, we need to check that in the region outside the horizon (when the Killing vector is timelike), the solution of KSEs (equations arising because of BPS condition) imply a Reissner-Nordstrom like configuration i.e. static with asymptotically flat geometry, invariant under half of the the supersymmetries. Second, when the Killing vector is null, for e.g. on the horizon, the KSEs give rise to configurations whose geometric structures are consistent with the ones derived from IH boundary conditions.3 We shall call a horizon supersymmetric weak isolated horizon (SWIH) if the conditions for existence of Killing spinors on the horizon are compatible with the IH boundary conditions. In other words, on a SWIH, the KSEs arising because of the BPS nature of the horizon will be consistent with the IH boundary conditions. Once the SWIH is defined, the next task is to construct the phase-space of the theory of gravity in hand with appropriate boundary conditions. Our phase space will consist of all solutions of supergravity which satisfy the SWIH boundary condition at the horizon and are asymptotically flat at infinity.4 In this paper, we shall use the generalization of the Holst action [44] as the action for pure supergravity [45]. Using well known techniques of covariant phase space, one can construct the symplectic structure on this SWIH phase space [46, 47]. The first law for the SWIH will then follow from this symplectic structure. It will also follow from this symplectic structure that the topological theory on fixed area phase space of SWIH is a Chern-Simons theory.

The plan of this paper is as follows: First, we spell out the isolated boundary conditions to be imposed on a generic null surface. Second, we study the constraints arising from Killing spinor equations and show that they are consistent with IH boundary conditions. Thirdly, we construct the space of solution of pure supergravity which satisfy the SWIH boundary conditions at horizon and are asymptotically flat at infinity. We construct the symplectic structure and prove the laws of black hole mechanics. Next, we shall go to fixed area phase space and and identify the Chern-Simons theory as the boundary theory.

2 Isolated horizon boundary conditions

We consider a - dimensional spacetime manifold equipped with a metric having Lorentzian signature and a null hypersurface . Let be a future directed null normal on . However, if is a future directed null normal, then so is , where is any positive function on . Two null normals and on will be called equivalent if . Thus, naturally admits an equivalence class of null normals. We shall denote this equivalence class by . Let us denote by the degenerate intrinsic metric on which is induced by the spacetime metric (indices that are not intrinsic on will be pulled back, denoted by an arrow under them, and signifies that the equality holds only on ). Thus has a signature . A tensor will be called an inverse of if it obeys the condition . The inverse metric , however, is not unique as one can redefine it as , where is any vector field tangential with . The expansion of the null normal is then defined by , where is the spacetime covariant derivative compatible with . Note that the expansion is insensitive to the ambiguity in the inverse metric but it varies under the scaling of the null normal in the equivalence class by .

In what follows, we shall work with the Newmann-Penrose (NP) null tetrad basis , being normal to the foliation of by and are tangential to - spheres. The basis vector obey the orthonormality conditions , others being zero. This is specially suited for the present study because one of the null-normals in the equivalence class coencide with the basis vector . Moreover, in this basis, many components of the connection vanish making the calculations much simpler than that in the coordinate basis. For future calculations with supergravity, it will be convenient to use the spinor basis alongside [48, 30, 36]. We can find a spin dyad with normalization condition . The null vectors are related to these dyad basis by the following relations:


where is called the soldering form.

The isolated horizon boundary conditions for spherically symmetric cases can be stated in terms of the spin dyads as follows [30, 36]. The surface will be called a non-expanding horizon (NEH) if

  1. is topologically .

  2. The spin dyads are constrained to satisfy 5


    where, is a real, nowhere-vanishing, spherically symmetric function, and denotes the unique torsion-free connection compatible with .

  3. All equations of motion hold on and the forms of the fields are such that is causal and is spherically symmetric.

We will study the consequences of these conditions after we have pointed out the restrictions from pure Supergravity.

3 Conditions from pure supergravity

The purpose of this section is to establish the compatibility of the isolated horizon boundary conditions to the conditions obtained from the KSEs of supergravity. In other words, we intend to show that on the horizon, the KSEs can be put in a form which are precisely the same as the IH boundary conditions6. This approach was also addressed in [41, 42, 43]. However, for our purpose, which includes construction of the symplectic structure, deriving the first law of black hole mechanics and to understand the origin of entropy of the black holes arising in supergravity, further details about spacetime connection and its curvature will be required. This needs a study of all the constraints available from the KSEs.

Before diving into formal calculations, let us try to comprehend the method. As mentioned previously, we are interested in BPS configurations. In other words, we look for classical configurations which have some residual supersymmetry and hence satisfy the KSEs for supergravity. All such configurations (not necessarily solutions of equation of motion) for have already been determined and classified in [8, 10, 11] by explicitly solving the KSEs. It then remains to identify, in the above classification, weather there exists BPS configuration(s) (i.e. member(s) of solution of KSEs) which also satisfy the isolated horizon boundary conditions. In an elaborate way, we can pick each configuration and check weather it actually satisfies the IH boundary conditions. Alternatively and equivalently, we might show that (in case when the Killing vector is null, see the paragraph just above eqn. (10) below), the conditions arising out of the KSEs can be put in a form which will be exactly identical to the isolated horizon boundary conditions. This will easily establish that under these circumstances, there exists classical configurations of pure supergravity which will satisfy the isolated horizon boundary conditions. All such configurations might not be solutions of equation of motion, only a few will be. We shall argue that there also exists solutions of equations of motion, for the theory under consideration, in this space of configurations. As an example, shall explicitly show that the Reissner-Nordstrom black hole, which solution of pure supergravity equations of motion, solves the KSEs and its horizon satisfies the isolated horizon boundary conditions. This will be done in two steps: first, we shall show that when there is a timelike static Killing vector, the configuration obtained by solving the KSEs is identical to the external region (of the horizon) of the Reissner-Nordstrom spacetime. Secondly, when the Killing vector is null, the KSEs can be put in a form identical to the IH boundary conditions which are also the ones satisfied by the horizon of Reissner- Nordstrom spacetime. Thus, the Reissner- Nordstrom spacetime will become consistent with IH fromalism. This will also demonstrate that other black hole solutions in supergravity theories can be understood using the isolated horizon formulation.

Let us now look into the KSEs of the pure supergravity. This theory has graviton and Maxwell field as the bosonic fields and two gravitini as their fermionic counterparts [49]. We are interested in the bosonic sector, with the gravitini fields set to zero since this sector will give us the classical spacetime solutions. Supersymmetry transformations are generated by gauge-spinor fields where is the internal index and is the spinor index. We start with the standard Killing spinor equations [10, 11, 41, 42].


where is the anti self-dual part of , i.e.,


For further calculations, it is useful to define the function . Also define the vector fields 7:


We first concentrate on the case . It follows from the equations in (3) that and are local gradients. A combination of the two vector fields and can be used to define the plane [10, 11]. From (3), we also get that the vector field is timelike and Killing:


All the static configurations admitting solution to the eqn. (3) are in the Majumdar-Papapetrou class [8, 10, 11]. If defines the horizon, the space of solutions of eqn. (3) include the part of the extremal Reissner-Nordstrom spacetime. This is the only static solution (in the Majumdar-Papapetrou class) with a single horizon. It follows that when the Killing vector is timelike, a solution of the Killing spinor equation is indeed the spacetime exterior to the Reissner-Nordstrom horizon.

The degenerate sector, is somewhat subtle and needs care. Before going into that, let us study the case in greater detail. From equation (6), we get:


Since the vector field is orthogonal to i.e. , we get from (7) that:


where is some function8. For , the vector fields and can be used to define two orthogonal directions. Indeed the these can describe the plane. A combination of the other two vector fields and define the plane.

On the horizon, become null and hence, is automatically geodetic:


where is some function on the horizon. Eqn.(9) and (8) together imply that on the horizon, . Thus the entire plane degenerates (to a line) on the horizon. We will then identify the horizon to be the surface where, (modulo rescaling by functions). We shall take this as our criterion for defining the horizon . We will see below that in this precise sense, defines a horizon.

When is vanishing, , where the function on is such that (see 3)


From equation (5), we also obtain that implies (modulo rescaling by functions). The configurations which solve the Killing spinor equations (3) for actually describe null surfaces. The standard coordinate system used in the exterior collapses on such surfaces. For example, in the Reissner-Nordstrom solution, the plane degenerates on the horizon. The trick to lift such degeneracy of coordinate systems is to introduce by hand an auxiliary vector field 9. We shall use this option. We introduce two normalised spinors and as the such that . The spinor is such that , where, is a real function on . These spinors can be used to construct a null tetrad basis (compare with eqn. (2)).

For further calculations, we need the form of the Maxwell field (). Observe that the equation (10) can be rewritten as:


Since, is a function on , the right hand side of (11) can depend on . Then, can be of the form (see eqn. (2))


However, we shall see that and there cannot be any term proportional to . This is because, the surface is null and the energy-momentum tensor () must be such that the vector field is causal or null. For the Maxwell fields, the energy-momentum tensor is given by:


Using the eqns. (4), (13) and form of in terms of spin-dyads (see eqn. (2)), the abovementioned restriction on implies that . For future convenience, we shall call . The other component of the Maxwell field will be kept unrestricted.

Let us now determine the constraints on the null-normals on . Using eqn. (3) and (12), we get


From the normalization condition and eqn. (14), we can obtain the action of the gradiant operator on . We restrict the form to be


where is a function on . With the equations (14) and (15) in hand, we can proceed to study their consequences. Note that these equations are precisely of the same form as the IH boundary conditions. Moreover, the constraints on the Maxwell field derived from eqn. (12) are such that they satisfy the conditions matter field must satisfy on an IH (see section 2).

4 Consequences of boundary conditions

In this section, we shall study the kinematical consequences of the boundary conditions. In what follows, we shall always restrict to horizons which are spherical, i.e. the null surface will be foliated by spheres. Using eqn. (2), we find that the equations (14) and (15) imply


where, and are one-forms on . The superscripts and indicate that and depend on the transformation of these vector fields. Also, note that is purely imaginary.

Several consequences follow from these equations [30, 27]. Firstly, since is null (and generates ), it is automatically geodetic and shear-free. Moreover from eqn. (16), the expansion vanishes on . All these restrictions, the Raychaudhuri equation for and the the energy condition further imply that is also shear-free. From eqn. (16), we get that the null-vector field is a Killing vector on ,


i.e. the IH boundary conditions imply that it is enough to have a Killing vector only on . Further, the volume form of the -spheres foliating given by , is also lie dragged: . To see this, use the Cartan formula , and the equation (18). The surface gravity of is denoted by :


The equations (16) and (20) together imply that .

The properties of the vector field follow similarly. It is twist-free, shear-free, has spherically symmetric expansion and vanishing on . This now shows that the function is actually related to the expansion of the vector field .

Before proceeding further, let us discuss the issues related to the available gauge freedom for the spin-dyads on . The most general transformation that preserves the normalization of the dyad is [30]


where and are real functions on . Under (2), transformations, the null vectors are


where, and are functions on . The one-forms and transform like gauge fields under rescaling of and respectively:


Consequently the surface gravity also depends on the rescaling of , . In this paper, we are interested in asymptotically flat global solutions, i.e where the observer one has access to the infinity. In these special cases, the vector field can always be normalized with respect to infinity. In other words, we can set so that there is no scaling ambiguity in the evaluation of the surface gravity . The function however remains unrestricted.

For further calculations, we shall need the curvature of the one-form fields and . They are given by [36, 27]:


where, and are components of the Weyl and the Ricci tensor respectively (see [1, 48]). The spherical symmetry of the horizon implies that or in other words, is a pure gauge and can be made to vanish by a choice of gauge. We are interested in extremal black holes. These have vanishing surface gravity. For future calculations, we shall always set . Eqn. (14) shows that this can be done for some special choice of the function .

5 Laws of black hole mechanics and entropy

In this section, we shall derive the zeroth and the first law of black hole mechanics. We will construct the symplectic structure for supergravity with appropriate boundary conditions and derive the first law. Thereafter, we will restrict to fixed area part of the phase space and derive the effective theory residing on the horizon. This will give us clues to calculate the entropy.

5.1 The Zeroth Law

We call a Weakly Isolated Horizon (WyIH) if . The requirement can be justified as follows. The quantity is analogue of the extrinsic curvature on the null hypersurface [50]. Since is a Killing vector on , the above condition implies that the entire data on the phase space is lie dragged by . Using the Cartan equation for lie derivative, it follows that the surface gravity is constant


throughout the horizon. For extremal and spherical black holes, . The restriction of WyIH implies that the surface gravity is zero and remains constant on .

5.2 The First Law

For the first law, we need to study the dynamics. For this, we need an action which will specify the dynamics and interactions of the geometric and matter degrees of freedom. We will use the Holst action modified for supergravity theory [45].

Holst Type Modification of Supergravity Action

The supergravity has an Abelian gauge field , the tetrad fields and two superpartner gravitinos. These gravitinos have chiral projections. One of them, denoted by , has positive chirality and the other, denoted by , has negative chirality. For writing thr first order action, we also introduce Lie algebra-valued connection one form . The modified action for supergravity is given by [45]:


where, the term is the standard supergravity action for :


where, , and is curvature of , i.e . The supercovariant field strength is given by


and the and signs denote the self dual and anti self dual fields, and . The other part of the action is given by:


where is the Barbero-Immirzi parameter. Equations of motion and other ramifications can be found in [45].

We are interested in black hole solutions. Moreover, we are restricting to Reissner-Nordstrom type of configurations. These are global supersymmetric solutions. From the point of view of classical solutions, the degrees of freedom of this theory are equivalent to the Holst action of the Einstein-Maxwell system, which is a consistent truncation of the modified Holst action for supergravity given in eqn. (28). In other words, for studying global classical solutions, we can consistently put the fermion fields to be zero. The KSEs for (28) are identical to that of the Einstein-Maxwell system. In what follows, we shall use the following action:


where, is a -form and is the completely antisymmetric tensor in internal space. Variation of the action with respect to the connection leads to:


It then follows from (33) that is the spin connection. Then the variation of the action (32) with respect to the tetrads give Einstein-Maxwell equations [51]. The boundary terms that arise from the variation of the action get contributions from the inner and outer boundaries. However, IH (or SWIH) boundary conditions and asymptotic flatness ensure that these boundary terms vanish, making the action principle well-defined [27].

We will need to specify the form of the Lorentz Lie-algebra valued connection one-form on . Introduce a fixed set of internal null vectors on such that while the other inner products vanish. These internal vectors are such that . Given these internal null vectors and the tetrad , we can construct the null vectors through . We can use these information to find the connection. To do this, we first note that since the internal null vectors are fixed, (), for the internal vector we get


Similar expressions can be obtained for the other internal vectors. Using the equations (16), the full connection turns out to be:


We define the following Lie algebra valued connection for ease of further computation [51, 27]


We shall also need the expression of the product of tetrads on the horizon . It is easily determined to be:


Symplectic structure

Given the lagrangian -form, there exists specific prescription for constructing the symplectic structure on the space of solutions [24, 25, 46, 47]. One obtains on-shell, the symplectic one-form (a spacetime -form) from the variation of the Lagrangian, where is an arbitrary vector field in the phase space. For the case in hand, we get


From , one then constructs the symplectic current . This is closed on-shell and integrating over the entire spacetime, we get (see figure 1):


To construct the symplectic structure, we must be careful that no data flows out of the phase space because of our choice of foliation. To ensure this, we will check that the symplectic structure is independent of the choice of foliation. We introduce potentials and .


A straightforward calculation shows that (see [35, 27]):


So only a special combination of the bulk and boundary symplectic current is independent of the choice of foliation. The symplectic structure is that of a Einstein- Maxwell system (see [27] for detailed derivation):

Figure 1: are two partial Cauchy surfaces enclosing a region of space-time and intersecting in the -spheres respectively and extend to spatial infinity . Another Cauchy slice M is drawn which intersects in

The first law can now be derived using this symplectic structure (42). Let us understand the conceptual basis of this proof. IH is a local definition of a horizon and the first law is expected to relate variations of local quantities that are defined only at the horizon without any reference to the rest of the spacetime 10. For example, the surface gravity is defined locally at the horizon. For the first law, the IH formalism enables us to define local energy (for horizons carrying other charges, such as angular momentum, electric potential etc., we must also provide local definitions for them). In spacetime, energy is associated with a timelike Killing vector field. Given any vector field in spacetime, it naturally induces a vector field in the phase space. The phase space vector field is the generator of time translation in the phase space. If time translation is a canonical transformation in the phase space then defines a Hamiltonian function . So to find out the Hamiltonian function associated with energy, we must look for phase space transformations that keep the symplectic structure invariant (canonical transformations). The vector fields tangent to these canonical flows are the Hamiltonian vector fields. To check wheather a vector field in the phase space is Hamiltonian, one constructs a one-form where , where is the lie flow generated by the spacetime vector field when tensor fields are varied. The necessary and sufficient condition for the vector field to be a globally Hamiltonian vector field is that the one-form is to be exact, where is the exterior derivative in phase space and is the corresponding Hamiltonian function. In other words, the vector field is globally Hamiltonian if and only if for any vector field in the phase space. The vector fields are also restricted by the condition that it should be tangential on . Now being a null surface, the WIH has only three tangential directions, one null and the two other spacelike. The closest analog of ‘time’ translation on WIH is therefore translation along the null direction. It is generated by the vector field . For global solutions this null normal vector field becomes timelike outside the horizon and is expected to match with the asymptotic time-translation for asymptotically flat spacetimes.

Using the above considerations, the first law of supersymmetric horizons turns out to be:


where is the ADM energy obtained when matches with the time translation at infinity and is the charge of the electromagnetic field on the horizon.. The right hand side of (45) is an exact variation if and only if is a function of alone. The phase space is characterized by charge and so is a function of . Define a quantity where


such that where is the associated Hamiltonian function . It is natural to interprete as the locally defined energy of the WIH and (45) as the first law of the WIH. The quantity receives contributions both from the bulk as well as the boundary symplectic structures and stands for the energy of the region between the WIH and the spatial infinity. The ADM energy is the sum total of these two energies. For the global solutions we are interested in, it is well known that the first law is equivalent to:


i.e., . To see this, observe that for all global solutions, . This is because when there is a global Killing vector field, induces infinitesimal gauge transform and is thus a gauge direction,


for all on the phase space. So, for any connected component of the phase space consisting of the spacetimes with global Killing vector field, is a constant. This constant can only be some spacetime quantity and can be the cosmological constant. For the present case, the cosmological constant vanishes and hence vanishes too and hence the energy measured at is same as that measured by any ADM observer. This means that this first law is exactly equivalent to that for the event horizons, with ADM replaced by in (45). This is a consistency check for the IH formulation.

5.3 Chern-Simons theory and entropy

In the introduction, we have said that the effective theory residing on the horizon can only be a topological theory. In this section, we shall outline the derivation of the Chern-Simons theory on . Detail calculations are similar to the ones in [27].

Let us now restrict to fixed area and fixed charge phase space. Define the connection component . In this case of spherical symmetry, it can be shown that the Gauss-Bonnet theorem implies that the equation (26) reduce to [36, 27]


This condition is also called the quantum horizon condition. The subscript indicates that we are in spherically symmetric phase space. Putting equation (47) in (39) and integrating by parts, we see that:


where . Note that the Maxwell field does not give any contribution to the entropy (see [27]). The boundary symplectic structure turns out to be that of Chern-Simons theory. The level of the theory takes integer values on quantization.

The entropy of the horizon can be obtained by quantization of Chern-Simons theory and thence counting states. The details of the quantization technique and various ramifications have been calculated in details [37]. The counting of states and entropy computation was first done in [37]. Better state counting methods have since been proposed [52, 53, 54] and the one put forward in [55] has carefully reconsidered some intricacies in the counting. The essential idea is the following: Consider a horizon of area . To compute the entropy, those states are relevant which satisfy the quantum horizon condition and have the fixed area of value . Entropy is obtained by taking logarithm of this value. The detailed counting of the microscopic quantum states of black hole is based on loop quantum gravity. It is proposed that the states are characterized by means of spin network basis [51]. If an edge with lebel ends at the horizon , it creates a puncture with label . The area of the horizon will be given by the value , being the Planck length. The punctures are also lebelled by the half-integers where . The quantum horizon condition relates this eigenstates to that of Chern-Simons theory. The requirement that the horizon is a sphere imposes the constraint . Thus the quantum state associated with cross-section of horizon are characterized by punctures and spin quantum numbers associated to each punctures lebel the states. Counting of states establishes that the entropy is indeed proportional to the area of the horizon.

6 Discussions

The objective of this paper was to introduce a new way of calculating the entropy of extremal black holes in supergravity theories. We observe that the standard Wald formulation [24, 25] fails to address the issue of entropy for extremal black holes. This is because the formulation depends on the existence of bifurcation spheres which are absent for these special black holes (and also for black holes formed out of collapse). It is then becomes necessary to formulate new ways to address this problem. Instead of modifying the Wald’s Killing horizon (KH) formulation, we reconsidered the isolated horizon (IH) formulation of black hole horizon. We matched the boundary conditions precisely and showed that it is possible to include the black holes arising in pure supergravity in the space of solutions this theory with IH as an inner boundary. Moreover, we proved the laws of black hole mechanics for these black holes and then went on to show that tye entropy of these black holes can be easily determined by quantizing the effective Chern-Simons theory that resides on the inner boundary of these black holes.

The advantage of this framework is that it doesnot require the entire spacetime to be supersymmetric. It is very much a possibility that the spacetime just outside the horizon is non-supersymmetric (in the sense that there are no Killing spinors that generate supersymmetry as isometries) because of presence of time dependent fields like electromagnetic and gravitational while only the horizon itself is supersymmetric (i.e. the horizon supports some Killing spinors) because the horizon is in equilibrium. So, this formulation admits a larger class of spacetime in its phase-space than the KH or the event horizon formulation which require some or the entire spacetime to be supersymmetric respectively. The laws of black hole mechanics thus proved on the larger phase space can encompass solutions which were otherwise difficult to address. Secondly, the method of determining the entropy is direct. It does not depend on the near-horizon/asymptotic structures (as is done for example in Kerr-CFT approach [56]) but is based on the quantization of the horizon topological theory induced as a result of the bulk-boundary gravitational interaction.

The present method however can be extended in various directions. Firstly, the present calculation is restricted to black holes in pure supergravity and can be repeated for black holes in extended supergravity. Secondly, black holes in higher dimensions are becoming more and more important. It will be an interesting problem to address this method for higher dimensions.


The author thanks A. Ghosh for discussions and encouragement at various stages of the work. He also thanks P. Majumdar for encouragement. The initial stage of the work began at IFT, UAM/CSIC Madrid. The author thanks T. Ortin and P. Meessen at IFT for stimulating discussions on reference [10], black holes and various other issues. He also thanks I. Booth for comments.


  1. preprint: Aaaa/Mm/Yy
  2. The KSEs are different for different theories but since we are interested in classical configuration, it remains true for any other set of fermion field
  3. This kind of construction can be also done for general relativity. For example, we might want to construct configurations which solve the Killing equation where is a timelike vector field. The static configurations includes the extremal Reissner-Nordstrom like spacetime. However, not all of these configurations are solutions of the Einstein equation. The solutions are those for which the constants and in the configuration can be identified with mass and charge respectively (for , we get the extremal Reissner-Nordstrom solution). In case becomes null (for example on the horizon), the solutions of the Killing equation will comprise of configurations which satisfy the IH boundary conditions.
  4. It is important to note that the phase space will only consist of the solutions of the equations of motion of supergravity. These are a subset of all solutions of the KSEs of the given supergravity theory.
  5. Quantities which are not intrinsic to are pulled back and denotes equality holds only on .
  6. Since the IH boundary conditions deal only with the horizon, regardless of structure of the exterior spacetime, it is enough to check the KSEs on the horizon. However, in this paper, we intend study the global and completely supersymmetric configurations like the Reissner-Nordstrom spacetime and hence we shall also look for solutions of KSEs in the exterior.
  7. From now on, we shall omit the soldering form . Double spinor indices of the same type will indicate one spacetime index, for e.g. .
  8. More generally, since and are also orthogonal to , etc. the equation 8 should be . However, we are working with static solutions having timelike Killing vector field . This implies that will not include the components. In other words, we can concentrate only on the deformations of the plane, keeping aside the ‘sphere’ () part
  9. If generates the null surface, one can introduce the auxiliary null vector field such that .
  10. For Reissner- Nordstrom like global configurations, the first law will be valid for the entire spacetime.


  1. S. Chandrasekhar, The mathematical theory of black holes, . Oxford, UK: Clarendon (1992) 646 p.
  2. J. M. Bardeen, B. Carter, and S. W. Hawking, The Four laws of black hole mechanics, Commun. Math. Phys. 31 (1973) 161–170.
  3. J. D. Bekenstein, Black holes and entropy, Phys. Rev. D7 (1973) 2333–2346.
  4. S. W. Hawking, Particle Creation by Black Holes, Commun. Math. Phys. 43 (1975) 199–220.
  5. G. W. Gibbons and P. K. Townsend, Vacuum interpolation in supergravity via super p-branes, Phys. Rev. Lett. 71 (1993) 3754–3757, [hep-th/9307049].
  6. K. S. Stelle, Lectures on supergravity p-branes, hep-th/9701088.
  7. M. J. Duff, R. R. Khuri, and J. X. Lu, String solitons, Phys. Rept. 259 (1995) 213–326, [hep-th/9412184].
  8. G. W. Gibbons and C. M. Hull, A Bogomolny Bound for General Relativity and Solitons in N=2 Supergravity, Phys. Lett. B109 (1982) 190.
  9. P. Breitenlohner, D. Maison, and G. W. Gibbons, Four-Dimensional Black Holes from Kaluza-Klein Theories, Commun. Math. Phys. 120 (1988) 295.
  10. K. P. Tod, All Metrics Admitting Supercovariantly Constant Spinors, Phys. Lett. B121 (1983) 241–244.
  11. K. P. Tod, More on supercovariantly constant spinors, Class. Quant. Grav. 12 (1995) 1801–1820.