Ramanujan and Quantum Black Holes

Atish Dabholkar,1,2,3
1 International
Centre for Theoretical Physics
Strada Costiera 11, Trieste 34151 Italy
arXiv:1905.04060v1 [hep-th] 10 May 2019

2 Sorbonne Université, CNRS,

Laboratoire de Physique Théorique et Hautes Énergies,
LPTHE, F-75005 Paris, France
3 CNRS, UMR 7589, LPTHE, Paris, F-75005 France

Abstract: Explorations of quantum black holes in string theory have led to fascinating
connections with the work of Ramanujan on partitions and mock theta functions, which in
turn relate to diverse topics in number theory and enumerative geometry. This article aims to
explain the physical significance of these interconnections.

Contribution to ‘Encyclopedia of Srinivasa Ramanujan and His Mathematics’

Based on ‘Nag Memorial Lectures’ delivered at the Institute for Mathematical Sciences

Keywords: black holes, partitions, mock modular forms, string theory

Contents
1 Quantum Black Holes 1

2 Colored Partitions 2

3 Hardy-Ramanujan Formula 3

4 Mock Jacobi Forms 4

1 Quantum Black Holes

A classical black hole is the region of spacetime which cannot send signals to faraway observers.
It is black because even light cannot escape its strong gravity. To make these notions precise,
consider the prototypical Schwarzschild spacetime with line element [1]

2M 2 2M −1 2
ds2 = −(1 − )dt + (1 − ) dr + r2 (dθ2 + sin2 θdφ2 ) , (1.1)
r r
where t is time and r, θ, φ are spherical coordinates. This pseudo-Riemannian metric is Ricci flat
and hence satisfies Einstein equations without matter. The region r ≤ 2M is the Schwarzschild
black hole of mass M . The boundary of this region at r = 2M is called the event horizon. It
is a peculiar 3-surface that is both stationary (independent of time) and lightlike (its conormal
dr has vanishing norm). Light emanating from inside the black hole cannot escape to faraway
region at large r because it cannot overtake the event horizon moving at the speed of light.
One consequence of this unusual causal structure is the discovery of Bekenstein and Hawking
[2, 3] that a black hole carries entropy S(M ) given by

A(M )
S(M ) = . (1.2)
4
This remarkable formula, valid in the limit of large area, implies a surprising connection between
thermodynamics, geometry, and quantum mechanics. It poses two important questions.

1. In quantum theory, a system with entropy S(M ) and mass M corresponds to an ensemble
of vectors in an eigensubspace H(M ) of dimension d(M ) with mass eigenvalue M . Can
one associate such a subspace with a black hole in the Hilbert space of quantum gravity
and compute its dimension?

–1–
 2. Can one define and compute [4, 5] a quantum generalization of (1.2) that is valid even
for small area? Schematically, it is expected to have the form
a2
S = a0 A + a1 log(A) + + . . . b0 e−c0 A + . . . (1.3)
A
with some coefficients (a0 , b0 , c0 . . .).
Both questions have rich implications and have provided invaluable clues in the search for
quantum gravity.
String theory offers the most promising approach to a consistent quantum theory of gravity.
Equations of motion of string theory are generalizations of Einstein equations. Some of the
‘supersymmetric’ solutions of these equations are ten-dimensional product manifolds M10 =
M4 (Q) × Σ6 where Σ6 is a compact Ricci-flat six-dimensional manifold. The manifold M4 (Q)
contains a black hole with a metric analogous to (1.1) but now specified by a vector of integral
charges Q so that the mass is determined by Q. The entropy S(Q) of this special class of black
holes is determined entirely in terms of the charges and geometric properties of M4 (Q) by a
formula analogous to (1.3).
The fundamental physical significance of entropy of these black holes stems from the Boltz-
mann relation
d(Q) = exp[S(Q)] (1.4)
which links a macroscopic geometric property of spacetime to the underlying microscopic
Hilbert space of quantum gravity. It provides a window into the quantum structure of spacetime
in much the same way entropy of gases revealed the quantum structure of matter.
Within the framework of string theory, a black hole in M4 (Q) corresponds to a multi-
dimensional ‘membrane’ wrapping a homology cycle in Σ6 . The Hilbert space H(Q) is the
space of states of this membrane. The integers d(Q), sometimes called the degeneracy, are
then given by certain enumerative invariants of Σ6 , which in turn are related to problems in
combinatorics and number theory. This line of enquiry leads naturally to the work of Ramanujan
as we illustrate below.

2 Colored Partitions
A simple example is Σ6 = K3 × T 2 where K3 is a Ricci-flat 4-manifold (‘Kummer surface’)
and T 2 is a 2-torus. The integer d(Q) in this case equals the Euler character of the symmetric
product of n copies of K3, where n is a particular integral norm of the vector Q. Recall that
the Euler character χ1 of a single copy K3 is 24 and K3 has only even cohomology, that is, it
admits only even harmonic forms. Given this topological data, the problem of computing the
Euler character χn of Symn (K3 ) is equivalent [6, 7] to the problem of finding the number of
partitions p24 (n) of a positive integer n using integers of 24 different colors, a problem close to
Ramanujan’s work. The solution is given [8] in terms of q-coefficients of a partition function:
1
Z(τ ) = , (q := e2πiτ ) (2.1)
∆(τ )

–2–
where ∆(τ ) is the Ramanujan cusp form of weight 12. The partition function is thus a modular
form of weight −12:
aτ + b
Z( ) = (cτ + d)−12 Z(τ ) (2.2)
cτ + d
for all  
ab
∈ SL(2, Z) , (2.3)
cd
and admits a Fourier expansion
∞
X
Z(τ ) = C(n)q n−1 . (2.4)
n=0

It is easy combinatorics to see that

p24 (n) = C(n) (n > 0) . (2.5)

3 Hardy-Ramanujan Formula
The modular properties of Z(τ ) imply that C(n) admits the Hardy-Ramanujan-Rademacher
expansion [9] and equals
∞  14
X 2π z 
K(n, −1, c) I13 (3.1)
c=1
c c
√
with z = 4π n, where Z +i∞
1 dt z2
I13 (z) := exp[t + ]
2πi −i∞ t14 4t
is a modified Bessel function and
d a
e2πi(n c +m c )
X
K(n, m, c) :=
d∈Z/cZ
da=1 mod c

is the Kloosterman sum for n, m, c ∈ Z.

For a large class of other black holes, the integers d(Q) are similarly related [10] to more
complicated invariants such as the Donaldson-Thomas invariants. They are given in terms of
Fourier coefficients of a variety of modular objects such as Jacobi or Siegel forms, and admit
an expansion [11, 12] that generalizes (3.2).
Remarkably, the Hardy-Ramanujan formula is an exact convergent expansion of an integer
in terms of analytic functions. This is just what is needed to verify the Boltzmann relation
(1.4) where d(Q) is an integer given by a counting problem but S(Q) is an analytic function
determined by the geometry of spacetime. Its leading asymtoptics valid for large charges has
exponential growth. For example, the large z asymptotics gives
√
d(n) ∼ I(z) ∼ ez ∼ e4π n
, n1 (3.2)

–3–
This is in accordance with leading, large area entropy of black holes and the Boltzmann relation.
Using methods of supersymmetric localization it has become possible to compute both sides of
equation (1.4) in a number of examples [13–18] to find nontrivial agreement. This verification of
Boltzmann relation even beyond the large area approximation constitutes one of the important
successes of string theory.

4 Mock Jacobi Forms

In an interesting class of examples, the partition function Z(τ, z) depends on an additional
variable z. It is modular in τ and transforms like a meromorphic Jacobi form of index m under
the ‘elliptic’transformations
z → z + λτ + µ , λ, µ ∈ Z , (4.1)
with double poles at z = 0 and its images. The Fourier coefficients d(Q) are no longer uniquely
defined but depend on the choice of the z-contour. This seems to contradict (1.4) because
entropy of S(Q) of the corresponding black hole is uniquely defined and does not suffer from
any ambiguities.
The resolution of this puzzle has to do with the existence of multi-centered black holes.
Equations of string theory admit a two-centered black hole solution that depends on the mod-
uli of Σ6 . Locally, each center looks like a single black hole but the two centers are bound
together. The distance between them is fixed by the charges and the moduli, and diverges as
one approaches a co-dimension one ‘wall’ in the moduli space. The solution no longer exists
on the other side of the wall. For a given charge Q, the enumerative ‘invariants’ d(Q, µ) now
have a mild dependence on the moduli. They are invariant in a given chamber in the moduli
space but jump upon crossing a wall separating two chambers because on one side the counting
includes the two-centered black holes but on the other side it does not. This is known as the
wall-crossing phenomenon [19, 20].
The ambiguity in defining the Fourier coefficients of Z(τ, z) now has a nice physical in-
terpretation [21–23]. The choice of the Fourier contour depends on the moduli µ of Σ6 in a
specific way. Crossing a wall in the moduli space corresponds to crossing a pole of the Fourier
integrand. The residue at the pole gives the difference in the enumerative invariants d(Q, µ) on
the two sides of the wall.
There is a special attractive chamber in the moduli space which admits only single-centered
black holes as solutions. Let us denote the moduli in this chamber by µ∗ . One can now pose
a more refined question whether one can compute d(Q, µ∗ ) and if it satisfies (1.4). The answer
to this question naturally leads us into the realm of mock modular forms [24] and is provided
by the following theorem [25]. The partition function admits a unique decomposition
Z(τ, z) = Z F (τ, z) + Z P (τ, z) , (4.2)
such that the polar part has the form
p24 (m + 1)
Z P (τ, z) = Am (τ, z) (4.3)
∆(τ )

–4–
where
X q ms2 +s y 2ms+1
Am (τ, z) = (4.4)
s∈Z
(1 − q s y)2
is called an Appell-Lerch sum. It admits a completion obtained a correction term that is
nonholomorphic in τ :
Â(τ, z) = A(τ, z) + A∗m (τ, z) (4.5)
which transforms as a Jacobi form. The completion satisfies
r
8πi 3/2 ∂
τ Âm (τ, z) (4.6)
m 2 ∂ τ̄
X
=− ϑm,` (τ ) ϑm,` (τ, z) .
` mod 2m

It also implies that Z F (τ, z) by itself is not modular but admits a modular completion Ẑ F (τ, z)
which transforms like a Jacobi form and satisfies a ‘holomorphic anomaly equation’ similar to
(4.6). Such an object is called a (mixed) mock Jacobi form. It turns out that Ramanujan’s
mock theta functions are closely related to another type of mock Jacobi forms very similar to
the ones that appear in the context of quantum black holes.
Using these ingredients one obtains a beautifully consistent picture [25]. The degeneracy
d(Q, µ∗) of single-centered black holes is given by the Fourier coefficients of the mock Jacobi
form which are independent of the moduli. They are uniquely defined by the charges and
satisfy the Boltzmann relation (1.4). The degeneracy of multi-centered black holes is given by
the Fourier coefficients of Z P (τ, z). These do depend on the choice of the contour and hence
indirectly on the moduli, and jump upon crossing walls in the moduli space consistent with the
wall-crossing phenomenon.
It is remarkable that the mathematical ideas and tools created by Ramanujan a century
ago have now come to have deep applications in quantum gravity.

References
[1] R. M. Wald, General Relativity. Chicago Univ. Pr., Chicago, USA, 1984.
[2] J. D. Bekenstein, Black holes and entropy, Phys. Rev. D7 (1973) 2333–2346.
[3] S. W. Hawking, Particle creation by black holes, Commun. Math. Phys. 43 (1975) 199–220.
[4] R. M. Wald, Black hole entropy in the Noether charge, Phys. Rev. D48 (1993) 3427–3431,
[gr-qc/9307038].
[5] A. Sen, Quantum Entropy Function from AdS(2)/CFT(1) Correspondence, 0809.3304.
[6] A. Dabholkar and J. A. Harvey, Nonrenormalization of the superstring tension, Phys. Rev. Lett.
63 (1989) 478.
[7] C. Vafa and E. Witten, A Strong coupling test of S duality, Nucl. Phys. B431 (1994) 3–77,
[hep-th/9408074].

–5–
 [8] L. Goettsche, , Math. Ann. 286 (1990) 193.
[9] H. Rademacher, Lectures on Elementary Number Theory. Robert E. Krieger Publishing Co.,
1964.
[10] A. Sen, Black Hole Entropy Function, Attractors and Precision Counting of Microstates,
arXiv:0708.1270 [hep-th].
[11] G. W. Moore, Les Houches lectures on strings and arithmetic, hep-th/0401049.
[12] J. Manschot and G. W. Moore, A Modern Farey Tail, Commun.Num.Theor.Phys. 4 (2010)
103–159, [0712.0573].
[13] A. Sen, Extremal black holes and elementary string states, Mod. Phys. Lett. A10 (1995)
2081–2094, [hep-th/9504147].
[14] A. Strominger and C. Vafa, Microscopic Origin of the Bekenstein-Hawking Entropy, Phys. Lett.
B379 (1996) 99–104, [hep-th/9601029].
[15] J. M. Maldacena, G. W. Moore, and A. Strominger, Counting BPS black holes in toroidal Type
II string theory, hep-th/9903163.
[16] I. Mandal and A. Sen, Black Hole Microstate Counting and its Macroscopic Counterpart,
Nucl.Phys.Proc.Suppl. 216 (2011) 147–168, [1008.3801].
[17] A. Dabholkar, J. Gomes, and S. Murthy, Localization &; Exact Holography, JHEP 1304 (2013)
062, [1111.1161].
[18] A. Dabholkar, J. Gomes, and S. Murthy, Nonperturbative black hole entropy and Kloosterman
sums, JHEP 03 (2015) 074, [1404.0033].
[19] F. Denef and G. W. Moore, Split states, entropy enigmas, holes and halos, hep-th/0702146.
[20] M. Kontsevich and Y. Soibelman, Stability structures, motivic Donaldson-Thomas invariants
and cluster transformations, 0811.2435.
[21] A. Sen, Walls of Marginal Stability and Dyon Spectrum in N=4 Supersymmetric String Theories,
JHEP 05 (2007) 039, [hep-th/0702141].
[22] A. Dabholkar, D. Gaiotto, and S. Nampuri, Comments on the spectrum of CHL dyons, JHEP 01
(2008) 023, [hep-th/0702150].
[23] M. C. N. Cheng and E. P. Verlinde, Wall Crossing, Discrete Attractor Flow, and Borcherds
Algebra, 0806.2337.
[24] S. P. Zwegers, Mock theta functions, Thesis, Utrecht (2002).
[25] A. Dabholkar, S. Murthy, and D. Zagier, Quantum Black Holes, Wall Crossing, and Mock
Modular Forms, 1208.4074.

–6–