Going Beyond Bekenstein and Hawking Exact and Asymptotic Degeneracies of Small Black Holes Tata Institute of Fundamental Research Kolymbari June 2005 Atish Dabholkar

28/06/2005Kolymbari2 A. Dabholkar PRL, 2005 A. D., Kallosh, & Maloney JHEP 0412:059,2004 A. D., Denef, Moore, & Pioline JHEP, 2005 A. D., Denef, Moore, & Pioline

28/06/2005Kolymbari3 S(Q) = k log  (Q) ? Macroscopic (from effective action) Microscopic (from brane configurations)

28/06/2005Kolymbari4 Several developments over the past few years have made it possible to carry out this comparison rather explicitly to all orders in an asymptotic expansion in inverse area. For a class of special BPS black holes in 4d, N=4 string theories, one can compute both {a i } and {b i } exactly to find that

28/06/2005Kolymbari5 Ingredients Action: Topological string, N=2 sugra Solution: Generalized attractor mechanism Entropy: Bekenstein-Hawking-Wald Ensemble: Mixed OSV ensemble Counting: Heterotic perturbative BPS states

28/06/2005Kolymbari6 Results Microscopic partition function is a generically is modular form of negative weight –w. The asymptotics are governed by a (hyperbolic) Bessel function

28/06/2005Kolymbari7 Macroscopic counting is governed by the number of abelian vector fields n v in the low energy N=2 supergravity and is also given by a Bessel function The overall normalization cannot yet be fixed.

28/06/2005Kolymbari8 Let’s recall a few properties of Bessel functions. It has the integral representation This satisfies the (hyperbolic) Bessel function. Moreover

28/06/2005Kolymbari9 The asymptotic expansion is given by with  = 4 2. For us, z » A(Q).

28/06/2005Kolymbari10 Microscopic BPS States Consider Heterotic String on T 4 £ T 2. Let the T 2 to be a product of two circles S 1 £ S 1. Take a winding string along a circle with winding number w and momentum n. It is BPS if it in the right-moving ground state. Dabholkar & Harvey

28/06/2005Kolymbari11 It can carry arbitrary left-moving oscillations subject to Virasoro contraint where N L is the number operator of 24 left-moving bosons. The spectrum is summarized by the partition function where d N is the degeneracy at level N.

28/06/2005Kolymbari12 It can be readily evaluated in terms of the Jacobi discriminant function where  is the Dedekind function. The degeneracy is given by the inverse Laplace transform.

28/06/2005Kolymbari13 To get the leading behavior at large N, we evaluate it at large temperature,  ! 0. It is convenient to use the modular property

28/06/2005Kolymbari14 Then Use asymptotics

28/06/2005Kolymbari15 From this integral representation we see where I 13 (z) is a Bessel function. The leading exponential can be found by saddle point approximation. The saddle occurs at

28/06/2005Kolymbari16 The degeneracy has the characteristic exponential growth This infinite tower of states has played a crucial role in furthering our understanding of stringy duality and black hole physics.

28/06/2005Kolymbari17 Thus the entropy is given by

28/06/2005Kolymbari18 Heterotic on T 4 $ IIA on K 3 Initial evidence came from matching low lying spectrum and effective action but a far more stringent test is obtained by matching the entire tower of BPS states. The configuration (n,w) is dual to a collection of w D4-branes with n D0-branes sprinkled on them. Their spectrum must exhibit this exponential growth. Vafa

28/06/2005Kolymbari19 The discriminant function made its appearance also in topological YM on K 3 which counts instantons in the 4-brane worldvolume gauge theory. This provided early hints of string-string duality. Vafa & Witten

28/06/2005Kolymbari20 Action In supergravity there are many moduli fields Vector multiplet A , X Hyper multiplet  1,  2 The vector-multiplet moduli couple to the electromagnetic field and vary in the of charged black hole background.

28/06/2005Kolymbari21 The vector-multiplet moduli space with n v vector multiplets is parameterized by n v +1 complex projective coordinates X I, I = 0, 1,…, n v. There are an infinite number of higher derivative corrections to the Einstein-Hilbert action that are expected to be relevant for the computation of the entropy. These F-type corrections to the effective action are summarized by the string-loop corrected holomorphic prepotential.

28/06/2005Kolymbari22 where F h are computed by the topological string amplitudes at h-loop and give terms in the effective action that go as Here W 2 is the graviphoton field. Witten; Bershadsky, Cecotti, Ooguri, Vafa, Antoniadis, Gava, Narain, Taylor

28/06/2005Kolymbari23 Solution The vector multiplet moduli can take arbitrary value at asymptotic infinity. In a black hole background, they also begin to vary as a function of r. The black hole entropy would then seem to depend on these continuous parameters. How can it possibly equal microscopic counting?

28/06/2005Kolymbari24 Attractor Formalism Noticed first for black holes with finite area in N=2 supergravity. Ferrara, Kallosh, Strominger The near horizon geometry of a black hole is Bertotti-Robinson (AdS 2 £ S 2 ) that preserves the full N=2 supersymmetry.Gibbons

28/06/2005Kolymbari25 The attractor equations are given by and C 2 W 2 =256. The scaling field C is introduced so that (CX I, CF I ) is non-projective and transforms like (p I, q I ) as a vector under the Sp(2n v +2; Z) symplectic duality group. The attractor equations are then determined essentially by symplectic invariance.

28/06/2005Kolymbari26 The metric near the horizon is given in terms of the central charge of the susy algebra. It has AdS 2 £ S 2 geometry. The radius of the sphere is equal to the value of the central charge at the horizon.

28/06/2005Kolymbari27 Entropy Consider an action with higher derivative corrections like R 2 etc in addition to the Einstein Hilbert term. Then the Bekenstein-Hawking entropy formula is modified. If there is a Killing horizon, then one can associate an entropy, Wald

28/06/2005Kolymbari28 such that the first law of thermodynamics is satisified. Thus the higher corrections will modify not only the solution but the area formula itself. The temperature is defined geometrically in terms of surface gravity

28/06/2005Kolymbari29 This beautifully general formula assumes its full significance only in a complete quantum theory of gravity such as string theory. But even with string theory, to really compute the corrections we would have to know all higher order terms, which might be possible in principle but not in practice. For computational control we need the action and should be able to solve it. With supersymmetric attractor and topological string, we have all the necessary ingredients.

28/06/2005Kolymbari30 The quantum-corrected entropy is given by Cardoso, de Wit, and Mohaupt (Kappeli)

28/06/2005Kolymbari31 The first set of attractor equations can be solved In terms of `potentials’  I. Then the entropy can be written in a suggestive form.

28/06/2005Kolymbari32 In terms of a `free energy’ function The potentials are determined by the second set of attractor equations Ooguri, Strominger, Vafa

28/06/2005Kolymbari33 Our System We have a special case when the Calabi-Yau is K 3 £ T 2. There are 23 2-cycles of which we take w 1 to be the 2-torus itself and w a, a =2, …, 23 to be the 22 2-cycles of K 3.

28/06/2005Kolymbari34 The perturbative state (n, w) on the heterotic side is dual to w 4-branes wrapping K 3 with n 0- branes sprinkled on it. The 4-cycle is dual to the 2-form w 1 and hence we have a nonzero magnetic charge and all other magnetic charges are zero. The 0-brane couples electrically to the graviphoton field and hence q 0 = n and all other electric charges are zero.

28/06/2005Kolymbari35 In the large volume limit we can approximate the second term by and solve the attractor equations for the charges q A =(q 0,0, 0, …, 0) and p A =(0, p 1, 0, …, 0). The solution is with  1 undetermined.

28/06/2005Kolymbari36 Leading order entropy The leading order entropy now comes not from the Einstein-Hilbert term but from the corrections and is given by In precise agreement with the logarithm of degeneracy at large N

28/06/2005Kolymbari37 Given the free energy it is natural to define a partition function  q I, p I ) are the black hole degeneracies. Can we test this?? Black Hole Partition Function

28/06/2005Kolymbari38 Black hole degeneracy The exact black hole degeneracy is proportional to Use the free energy. Do the 22 Gaussian {  a } integrals. The  0 integral is trivial at large N, the integrand is independent of  0. We are left with a single integral over  1.

28/06/2005Kolymbari39 We get the same integral representation that we encountered in our microscopic counting for d N. Hence black hole degeneracy  (p 1, q 0 ) then has is given by

28/06/2005Kolymbari40 We thus see that the Boltzmann entropy of the OSV ensemble matches with the microscopic entropy to all orders

28/06/2005Kolymbari41 Black Hole Entropy Microscopic Counting Sugra Attractor Topological String Wald’s Formalism

28/06/2005Kolymbari42 Stringy Cloak for a Null Singularity Classically, the spacetime geometry of these black holes is singular. There is a null singularity that is not separated from the asymptotic observer by a regular horizon. Once we include the corrections, the singularity is `cloaked’ and we obtain a spacetime with a regular horizon.

28/06/2005Kolymbari43 The near horizon geometry is AdS 2 £ S 2. The radius of the 2-sphere is large in Einstein frame and goes as. The string coupling is determined by the attractor equations and vanishes as on the horizon. Thus in the string frame, the horizon is still string scale consistent with the heuristic idea of a `stretched horizon’. Is there an exact conformal field theory?

28/06/2005Kolymbari44 The entropy is no longer given by the Bekenstein-Hawking relation but rather by Here A is the quantum-corrected area. Wald’s correction is of the same order of magnitude as Bekenstein-Hawking and together we have A/4 + A/4 = A/2.

28/06/2005Kolymbari45 This works for a large class of black holes in CY3 compactifications that have classically vanishing area. On the heterotic side for N=2 supersymmetry, this can be seen by a `universality’ argument of Sen. The microscopic entropy depends only on the central charge of the left-moving CFT, which is always 24. On the macroscopic side, the  ’ corrections of the N=2 sigma model are inherited from the N=4 theory.

28/06/2005Kolymbari46 Conclusions Black Holes that have zero classical area have precisely computable entropy using Wald’s formalism and sugra attractor which is in agreement with the microscopic counting. The black hole partition function agrees with the microscopic partition function in an asymptotic expansion to all orders in inverse charge but there are nonperturbative corrections.