Precision Microstate Counting of Small Black Rings

Atish Dabholkar,
1
Norihiro Iizuka,
1
Ashik Iqubal,
1
and Masaki Shigemori
2
1
Department of Theoretical Physics, Tata Institute of Fundamental Research, Mumbai 400 005, India
2
California Institute of Technology 452-48, Pasadena, California 91125, USA
(Received 4 December 2005; published 23 February 2006)
We examine certain two-charge supersymmetric states with spin in ve-dimensional string theories
which can be viewed as small black rings when the gravitational coupling is large. Using the 4D-5D
connection, these small black rings correspond to four-dimensional nonspinning small black holes. Using
this correspondence, we compute the degeneracy of the microstates of the small black rings exactly and
show that it is in precise agreement with the macroscopic degeneracy to all orders in an asymptotic
expansion. Furthermore, we analyze the ve-dimensional small black ring geometry and show qualita-
tively that the Regge bound arises from the requirement that closed timelike curves be absent.
DOI: 10.1103/PhysRevLett.96.071601 PACS numbers: 11.25.Mj, 04.70.Dy, 11.25.Wx
Introduction. One of the important successes of string
theory has been that for a special class of supersymmetric
black holes with a large classical area in the supergravity
approximation, one can explain the Bekenstein-Hawking
entropy in terms of statistical counting of their microstates
[1]. Recently, it has become possible to extend these results
to certain small black holes that have vanishing classical
entropy but for which the quantum corrected entropy is
nonzero and in precise agreement with the microscopic
counting [28]. In fact, for these small black holes, a far
more detailed comparison is possible and even the sub-
leading corrections to the entropy are found to be in
agreement with the state counting to all orders in a pertur-
bative expansion in large charges. For this comparison to
work, it is essential to include the quantum corrections to
the Bekenstein-Hawking area formula itself in a systematic
way [9,10] using the attractor mechanism [1116] and a
specic statistical ensemble [17].
The black hole entropy thus provides a valuable clue
about the microscopic structure of the theory. It is remark-
able how tightly constrained the structure of string theory
is. Various terms in the string effective action have to be of
a very denite form with the right coefcients in order that
the resulting macroscopic entropy matches with the count-
ing of the microscopic quantum states of the theory to all
orders.
In this Letter we generalize these results to states that
carry spin. Spin introduces a number of qualitatively new
features in the analysis of spacetime geometry especially in
conjunction with supersymmetry. Supersymmetry requires
that the angular velocity at the horizon be zero because
otherwise there would be an ergoregion and energy can be
extracted from the system for xed charge in conict with
Bogomolnyi-Prasad-Sommerfeld (BPS) stability. Hence
for supersymmetric states, the angular momentum is typi-
cally swirling around outside the horizon. The nonspinning
horizons, however, can have nontrivial geometry or topol-
ogy. For example, adding spin to the D1-D5-P large black
hole in ve dimensions [1] deforms the round S
3
horizon
into an ellipsoid [18,19]. The area of the deformed horizon
then correctly accounts for the modication of the entropy
due to spin. For the same system, in a different regime of
charge assignment, more exotic ringlike horizons with
S
2
S
1
topology are possible [2024]. It is of interest
therefore to know which of these possibilities would be
realized for states that correspond to small black holes once
spin is introduced and to ask if the counting still continues
to be in agreement with the macroscopic entropy.
We consider spinning Dabholkar-Harvey (DH) states
[25,26] with two charges in toroidally compactied heter-
otic string theory in ve dimensions. Using the chain of
dualities and 4D-5D connection [2730], it was shown in
Ref. [31] that we can relate them to BPS states in four
dimensions with four charges but no spin. This relation
enables us to use the well-developed technology of the
attractor equations including certain quantum corrections
in four dimensions. One then nds that the resulting con-
guration has nite entropy after including the quantum
corrections [31,32]. Moreover, as we will show, the micro-
scopic and macroscopic entropy, in fact, agree to all orders
in an asymptotic expansion.
Small black rings and 4D-5D connection. Consider
heterotic string compactied on T
4
S
1
where T
4
is a
4-torus in 6789 directions and S
1
is a circle along the 5
direction. Consider nowa string state with winding number
valong the X
5
direction. In a given winding sector, there is
a tower of BPS states each in the right-moving ground state
but carrying arbitrary left-moving oscillations subject to
the Virasoro constraint N
L
1 +nv, where N
L
is the left-
moving oscillation number and n is the quantized momen-
tum along X
5
[25,26]. Note that N
L
is positive and hence a
BPS state that satises this constraint has positive n for
positive v for large N
L
. This state can carry angular
momentum 1, say in the 34 plane. The angular momen-
tum operator 1 is given by 1

n1
jo
|
n
o
n
o
|
n
o
n
| where
o
n
and o
n
are the oscillator modes with frequency n of the
coordinates (X
3
+|X
4
) and (X
3
|X
4
) respectively, nor-
malized as jo
m
, o
|
n
| j o
m
, o
|
n
| u
mn
.
Following the 4D-5D connection explained in Ref. [31]
we can map this state to a conguration in Type IIA
PRL 96, 071601 (2006)
P HYS I CAL RE VI E W L E T T E RS
week ending
24 FEBRUARY 2006
0031-9007,06,96(7),071601(4)$23.00 071601-1 2006 The American Physical Society
compactied on K3 T
2
with charges D2-D2-D0-D4
which in turn is dual to the DH states in heterotic string
on T
4
S
1

~
S
1
with momentum and winding (n, v) and
(~ n, ~ v) along the circles S
1
and
~
S
1
, respectively, with all
integers n, v, ~ n, ~ v positive. It is useful to state the 4D-5D
connection entirely in the heterotic language. The basic
idea following Refs. [2730] is to make use of the
TaubNewman-Unti-Tamburino (Taub-NUT) geometry.
For a Taub-NUT space with unit charge, the geometry
near the origin is R
4
whereas at asymptotic innity it is
R
3

~
S
1
. Thus a contractible circle at the origin of Taub-
NUT turns into a noncontractible circle
~
S
1
at asymptotic
innity. The angular momentum 1 at the origin then turns
into momentum ~ n along the circle at innity [28].
Consider now DH states with spin in heterotic string
theory on T
4
S
1
. Spinning strings that are wrapping
along the circle S
1
in 5 direction and rotating in the
34 plane have a helical prole [3336]. The helix goes
around a contractible circle S
1

of radius R

along an
angular coordinate in the 34 plane as the string wraps
around the noncontractible circle S
1
. Let us denote the
pitch of the helix by , which is the winding number of
the projection of the helix onto the contractible circle.
Macroscopically it corresponds to a dipole charge. We
can now embed this system in Taub-NUT space with
very large Taub-NUT radius R
TN
>R

and regard S
1

in
R
4
as being situated at the origin of a Taub-NUT geometry.
Varying the radius of Taub-NUT, which is a modulus, we
can smoothly go to the regime R
TN
<R

. Then the con-

tractible S
1

effectively turns into the noncontractible circle

~
S
1
at asymptotic innity. The entropy of BPS states is not
expected to change under such an adiabatic change of
moduli. We can dimensionally reduce the system to 4D
along
~
S
1
and obtain a 4D DH state with four charges
(n, v, ~ n, ~ v) with the identication that ~ n 1 and ~ v
. This system has a string scale horizon in 4D [28]
which suggests that the original spinning DH system in
5D is a small black ring [31].
Since we have unit Taub-NUT charge to begin with, we
do not have a purely electric conguration in 4D but
instead have a Kaluza-Klein (KK) monopole of unit charge
in addition to the 4-charge purely electric small black hole.
However, since the helix is far away from the origin of
Taub-NUT space in 5D before dimensional reduction, the
KK monopole is sitting far away from the 4-charge black
hole in 4D. The separation is determined by Denefs con-
straint [37] and is determined by 1 and the asymptotic
values of the moduli and can thus be made arbitrarily large.
The local microscopic counting therefore does not depend
on the addition of the KK monopole and is given by the
counting of DH states.
Four-dimensional counting. In the 4D description, the
state is specied by the charge vector Q in the Narain
charge lattice
2,2
of the S
1

~
S
1
factor with four integer
entries. The norm of this vector is
Q
2
2

1
2
n v ~ n ~ v
_ _
0 1 0 0
1 0 0 0
0 0 0 1
0 0 1 0
_
_
_
_
_
_
_
_
_
_
n
v
~ n
~ v
_
_
_
_
_
_
_
_
_
_
. (1)
The degeneracy of these perturbative DH states can be
computed exactly and the asymptotic degeneracy for large
Q
2
is given as in Refs. [4,5] by

micro
(n, v, ~ n, ~ v)
^
I
13
_
4r
......
Q
2
2
_
_

^
I
13
(4r
.................... .
nv ~ n ~ v
_
),
(2)
where
^
I
13
(z) is the modied Bessel function dened in
Ref. [5].
Turning to the macroscopic degeneracy, we compute it
using the Ooguri-Strominger-Vafa (OSV) relation [17]
between topological string partition function and the mac-
roscopic degeneracy, for the nonspinning four-dimensional
conguration. For this purpose we use the Type IIA de-
scription, viewing this state as a collection of D2-D2-D0-
D4 branes. There are n D2-branes wrapping a 2-cycle n
1
in
K3 and v D2-branes wrapping a 2-cycle n
2
such that the
intersection matrix of n
1
and n
2
is as in the upper-left 2
2 block of the matrix in (1). We therefore identify the
charges as (q
2
, q
3
) (n, v). Similarly, we identify the
D0-D4 charges as (q
0
,
1
) (~ n, ~ v) so in the labeling of
charges used in Ref. [5], we have Q (q
2
, q
3
, q
0
,
1
) and
all other charges zero. Using the formula (2.26) in [5] we
then see that the macroscopic degeneracy is given by

macro
(n, v, 1) (
1
)
2
^
I
13
(4r
...........................
q
2
q
3

1
q
0
_
)
(
1
)
2
^
I
13
(4r
.................... .
nv ~ n ~ v
_
). (3)
Therefore, up to the overall (
1
)
2
factor, the microscopic
(2) and macroscopic (3) degeneracies match precisely to
all orders in an asymptotic expansion for large Q
2
.
Five-dimensional microscopic counting. We now
would like to count the degeneracy of the spinning DH
system from the 5D side. The nontrivial issue is to deter-
mine the correct ensemble. The relevant states correspond
to quantum uctuations around a specic coherent oscil-
lating state which is essentially Bose-Einstein condensate
on the world sheet and describes the helical geometry with
pitch [31]. As we will argue below, the precise micro-
states turn out to be of the form
(o
|

)
1
....
microscopicoriginof
~ n1and ~ vof
thering

n1
_

|1,2,,5...24
(n
|
n
)
N
n|
_
}0)
........................ ........................ ..
fluctuation: all possiblestates
withlevel N
eff
N1. Angular
momentum1isnotfixed.
(4)
for 1, >0. Namely, we consider the states with the world
sheet energy N
eff
N 1 and chemical potential
PRL 96, 071601 (2006)
P HYS I CAL RE VI E W L E T T E RS
week ending
24 FEBRUARY 2006
071601-2
conjugate to 1 set to zero, and multiply all those states by
(o
|

)
1
. Note, in particular, that the o
|

and o
|

oscillators are
included in the uctuation part. The degeneracy of the
states (4) is the same as that of the DHsystemwith Q
2
,2
N
eff
and proportional to
^
I
13
(4r
................ .
N 1
_
), in precise
agreement with (2) and (3). If 1, <0, (o
|

)
1
in (4) must
be replaced by ( o
|
}}
)
}1}
.
This separation between the classical coherent conden-
sate that describes the large helix and the small quantum
uctuations around it that account for the entropy is similar
to the one used in Refs. [3842]. It is valid in the regime
when R

is much larger than the amplitude of uctuation.

Therefore, we conclude that in this regime, the states of the
form (4) are the states that account for the microstates of
the ring. This in turn agrees with (2) and (3) in 4D through
the 4D-5D connection.
Note that (4) means that the microscopic counting in 5D
must not be done for xed angular momentum 1. Fixing 1
would impose an additional constraint on the uctuation
part in (4). From the 4D point of view, it would correspond
to imposing a constraint on the world sheet oscillators of
the DH system, which would lead to a result contradictory
to the 4D degeneracy (2) and (3). To demonstrate this, let
us count the degeneracy of spinning DHstates with xed 1.
The degeneracy (N, 1) is summarized in the partition
function
Z(, )

N,1
(N, 1)q
N
c
1
, q e

, c e

,
(5)
where N nv N
L
1. can be thought of as the
inverse temperature on the world sheet for a 1 +1 gas of
left-moving 24 bosons conjugate to the total energy N and
can be thought of as the chemical potential conjugate to
the quantum number 1 of this gas. Since N
L
is the oscil-
lation number for the 24 left-moving transverse bosons,
using the expression 1

n1
jo
|
n
o
n
o
|
n
o
n
|, the parti-
tion function can be readily evaluated [43] and is given by
Z(, )
_
q

n1
(1 q
n
)
22
(1 cq
n
)(1 c
1
q
n
)
_
1

1
,
21
(e

)
2| sinh(,2)
0
11
(,2r|, |,2r)
, (6)
in terms of the standard Dedekind eta function and theta
function with characteristics.
The number of states with given N and 1 is then given
by the inverse Laplace transform: (N, 1)
1
(2r|)
2

_
C

Je
N
_
C

Je
1
Z(, ), where the contour
C

runs from |r + to +|r + with >0 to avoid

singularities on the imaginary axis. Similarly, C

goes
from r|, +e to +r|, +e with 1 <e <1 to
avoid poles. To nd the asymptotic degeneracy at large
N, we want to take the high temperature limit, or -0.
Using the modular properties of the Dedekind eta and the
theta functions we can write the degeneracy at high tem-
perature as in [43] as
(N, 1)
1
2r|
_
C

Je
N+(2r)
2
,
_

2r
_
12
I(, 1), (7)
where I(, 1) is dened by
I(, 1)
1
2r|
_
C

Je

2
,21
sinh(,2)
sin(r)
. (8)
To arrive at (7), we dropped terms that are exponentially
suppressed for small as e
(2r)
2
,
. This is justied
although is still to be integrated over, because the saddle
point around 1,
.............
N 1
_
<1 will make the leading
contribution, as we will see below.
Now we evaluate (8) using the method of residues.
Deform the contour C

into sum of three intervals C

1

jr|,+e, r|, +K|, C
2
jr|, +K, r|, +
K|, and C
3
jr|, +K, r|, +e|, with K >1. One
can readily show that the contour integral along C
1,2,3
vanishes due to the periodicity of the original integrand
(6) in the small and large K limit. In the process of
deforming the contour, we pick up poles at m, m
1, 2, . . . . In the end, we obtain
I(, 1) Res
1
_
e

2
,21
sinh(,2)
sin(r)
_
+O(e
21
)
(1e

)e
1
+O(e
21
). (9)
Here O(e
21
) comes from the poles at 2, 3, . . . and is
negligible when 1 O(N) since 1 O(N
1,2
). We
interpret the term e
1
as the contribution from the
1 sector. Substituting this back into (7), we conclude
that the degeneracy (N, 1) is
1
2r|
_
J(,2r)
12
(

2
2
+ )e
(2r)
2
,+(N1)
. Each term in the integral is
of the Bessel type as discussed in Ref. [5], and thus the
nal result is (N, 1)
^
I
14
(4r
............
N1
_
) (2r,2)
^
I
15
(4r
............
N1
_
) +, which agrees with (2) and (3) with
1 only in the leading exponential but disagrees in the
subleading corrections. This demonstrates that micro-
scopic counting in 5D must be done not for xed 1 but
for the states (4). One can show that states with xed
0 also lead to degeneracy in disagreement with (2) and (3).
In general, subleading corrections to thermodynamic
quantities depend on the choice of the statistical ensembles
and are different for different ensembles. For example,
even for nonspinning black holes the ensemble with xed
angular momentum 1 0 differs from the ensemble with
xed chemical potential 0 in subleading corrections.
It was noted in [5] that the correct microscopic ensemble
that is consistent with the OSV conjecture is the one with
0. In our case, the description of small black rings
requires that we also x the pitch of the helix as an addi-
tional requirement and that we are counting states around
this classical coherent condensate on the world sheet.
PRL 96, 071601 (2006)
P HYS I CAL RE VI E W L E T T E RS
week ending
24 FEBRUARY 2006
071601-3
Five-dimensional macroscopic geometry. Finally, we
comment on the geometry of this 5D small black ring,
which can be determined in the near ring limit by using the
4D-5D uplift. Exact uplift is possible near the horizon,
since the near-horizon geometry of 4D black hole can be
determined precisely by using string-corrected attractor
equations. Let us consider the case where 1 O(N).
We consider a helix with general pitch even though the
contribution from the 1 subsector dominates entropy.
In 4D heterotic string theory, the near-horizon black hole
geometry is determined by the attractor equations to be
e
2
4
1,
................ .
N 1
_
,
........
g

_

.........
1,
_
, and the horizon radius
is r
S
2 l
s
. Using e
2
5
e
2
4
........
g

_
and l
s
e
2
5
,3
l
(5)
l
,
the scale of S
1

along ring is given by

r
S
1
........
g

_
l
s

1,3
1
1,3
(N 1)
1,6
l
(5)
l
,
r
S
2 l
s

1,6
1
1,6
(N 1)
1,6
l
(5)
l
,
(10)
such that S A,4G
5
r
S
1 (r
S
2 )
2
,(l
(5)
l
)
3

................ .
N 1
_
. One
important qualitative feature of the solution is that when
1 exceeds N, the solution develops closed timelike
curves. When N 1 is saturated, g

0 is saturated
at the ring horizon. Hence, the Regge bound N 1 on
the angular momentum of the underlying microstates can
be understood from the macroscopic solution as a conse-
quence of the physical requirement that closed timelike
curves be absent. The details will be presented elsewhere
[44].
We are very grateful to Ashoke Sen for numerous illu-
minating discussions and collaboration on Ref. [44]. A. D.
would like to thank the high energy group at ASICTP,
Trieste for their hospitality where part of this work was
completed. The work of M. S. was supported in part by
Department of Energy Grant No. DE-FG03-92ER40701
and the Sherman Fairchild Foundation.
[1] A. Strominger and C. Vafa, Phys. Lett. B 379, 99
(1996).
[2] A. Dabholkar, Phys. Rev. Lett. 94, 241301 (2005).
[3] A. Dabholkar, R. Kallosh, and A. Maloney, J. High Energy
Phys. 12 (2004) 059.
[4] A. Dabholkar, F. Denef, G. W. Moore, and B. Pioline,
J. High Energy Phys. 08 (2005) 021.
[5] A. Dabholkar, F. Denef, G. W. Moore, and B. Pioline, hep-
th/0507014.
[6] A. Sen, Mod. Phys. Lett. A 10, 2081 (1995).
[7] A. Sen, J. High Energy Phys. 05 (2005) 059.
[8] A. Sen, hep-th/0504005.
[9] R. M. Wald, Phys. Rev. D 48, R3427 (1993).
[10] V. Iyer and R. M. Wald, Phys. Rev. D 50, 846 (1994).
[11] S. Ferrara, R. Kallosh, and A. Strominger, Phys. Rev. D
52, R5412 (1995).
[12] G. Lopes Cardoso, B. de Wit, and T. Mohaupt, Phys.
Lett. B 451, 309 (1999).
[13] G. Lopes Cardoso, B. de Wit, and T. Mohaupt, Fortschr.
Phys. 48, 49 (2000).
[14] G. Lopes Cardoso, B. de Wit, and T. Mohaupt, Classical
Quantum Gravity 17, 1007 (2000).
[15] G. Lopes Cardoso, B. de Wit, J. Kappeli, and T. Mohaupt,
J. High Energy Phys. 12 (2000) 019.
[16] J. M. Maldacena, A. Strominger, and E. Witten, J. High
Energy Phys. 12 (1997) 002.
[17] H. Ooguri, A. Strominger, and C. Vafa, Phys. Rev. D 70,
106007 (2004).
[18] J. C. Breckenridge, R. C. Myers, A. W. Peet, and C. Vafa,
Phys. Lett. B 391, 93 (1997).
[19] J. P. Gauntlett, R. C. Myers, and P. K. Townsend, Classical
Quantum Gravity 16, 1 (1999).
[20] H. Elvang, R. Emparan, D. Mateos, and H. S. Reall, Phys.
Rev. Lett. 93, 211302 (2004).
[21] H. Elvang, R. Emparan, D. Mateos, and H. S. Reall, Phys.
Rev. D 71, 024033 (2005).
[22] J. P. Gauntlett and J. B. Gutowski, Phys. Rev. D 71,
025013 (2005).
[23] J. P. Gauntlett and J. B. Gutowski, Phys. Rev. D 71,
045002 (2005).
[24] I. Bena and N. P. Warner, hep-th/0408106.
[25] A. Dabholkar and J. A. Harvey, Phys. Rev. Lett. 63, 478
(1989).
[26] A. Dabholkar, G. W. Gibbons, J. A. Harvey, and F. Ruiz
Ruiz, Nucl. Phys. B340, 33 (1990).
[27] D. Gaiotto, A. Strominger, and X. Yin, hep-th/0503217.
[28] H. Elvang, R. Emparan, D. Mateos, and H. S. Reall,
J. High Energy Phys. 08 (2005) 042.
[29] D. Gaiotto, A. Strominger, and X. Yin, hep-th/0504126.
[30] I. Bena, P. Kraus, and N. P. Warner, Phys. Rev. D 72,
084019 (2005).
[31] N. Iizuka and M. Shigemori, J. High Energy Phys. 08
(2005) 100.
[32] K. Behrndt, G. L. Cardoso, and S. Mahapatra, Nucl. Phys.
B732, 200 (2006).
[33] A. Dabholkar, J. P. Gauntlett, J. A. Harvey, and D. Wal-
dram, Nucl. Phys. B474, 85 (1996).
[34] C. G. Callan, J. M. Maldacena, and A. W. Peet, Nucl. Phys.
B475, 645 (1996).
[35] O. Lunin and S. D. Mathur, Nucl. Phys. B610, 49 (2001).
[36] O. Lunin and S. D. Mathur, Phys. Rev. Lett. 88, 211303
(2002).
[37] F. Denef, J. High Energy Phys. 08 (2000) 050.
[38] B. C. Palmer and D. Marolf, J. High Energy Phys. 06
(2004) 028.
[39] D. Marolf and B. C. Palmer, Phys. Rev. D 70, 084045
(2004).
[40] I. Bena and P. Kraus, J. High Energy Phys. 12 (2004) 070.
[41] P. Kraus and F. Larsen, J. High Energy Phys. 09 (2005)
034.
[42] D. Bak, Y. Hyakutake, S. Kim, and N. Ohta, Nucl. Phys.
B712, 115 (2005).
[43] J. G. Russo and L. Susskind, Nucl. Phys. B437, 611
(1995).
[44] A. Dabholkar, N. Iizuka, A. Iqubal, A. Sen, and
M. Shigemori (to be published).
PRL 96, 071601 (2006)
P HYS I CAL RE VI E W L E T T E RS
week ending
24 FEBRUARY 2006
071601-4