1: \documentclass[prl,aps,twocolumn,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
2: \usepackage{graphicx,epsfig}
3: \usepackage{bm}% bold math
4:
5: \def\d{{\rm d}}
6: \def\bea{\begin{eqnarray}}
7: \def\eea{\end{eqnarray}}
8:
9: \begin{document}
10: \title{Exactly Soluble BPS Black Holes in Higher Curvature $N=2$ Supergravity}
11: \author{Dongsu Bak, Seok Kim, Soo-Jong Rey}
12: \email{dsbak@mach.uos.ac.kr, seok@kias.re.kr, sjrey@phya.snu.ac.kr}
13: \affiliation{
14: Physics Department, University of Seoul, Seoul 130-743 KOREA\\
15: School of Physics, Korea Institute for Advanced Study, Seoul 130-012 KOREA\\
16: School of Physics, Seoul National University, Seoul 151-747 KOREA}
17: \begin{abstract}
18: We find a class of d=4, N=2 supergravity with $R^2$-interactions
19: that admits exact BPS black holes. The prepotential contains
20: quadratic, cubic and chiral curvature-squared terms. Black hole
21: geometry realizes stretched horizon, and consists of anti-de Sitter,
22: intermediate and outermost flat regions. Mass and entropy depends on
23: charges and are modified not only by higher curvature terms but also
24: by quadratic term in the prepotential. Consequently, even for large
25: charges, entropy is no longer proportional to mass-squared.
26: \end{abstract}
27: \pacs{04.70.-s 11.25.-w 04.65.+e }
28: \keywords{black hole, string theory, supergravity}
29: \maketitle
30: Recently, in string theory, black holes in higher curvature gravity received renewed interest. For one thing, an interesting connection \cite{osv} was found between the partition function of Bogomolnyi-Prasad-Sommerfield (BPS) black holes in quantum Type II string theory compactified on a Calabi-Yau 3-fold and the product of partition sums for topological and anti-topological strings. It is claimed that black hole entropy, which is modified by higher curvature and higher genus effects, is interpretable as a Legendre
31: transform of free energies for the topological and anti-topological strings. For another, for BPS black holes with fewer charges, it was discovered \cite{dabholkar} that the stretched horizon as envisaged by Sen \cite{sen} emerges naturally by the higher curvature and
32: higher genus effects. See \cite{osvrecent}, \cite{horizonrecent} for subsequent works for these issues.
33:
34: In $d=4, N\ge 2$ supergravity, part of higher curvature effects is encoded into so-called gravitational $F$-terms. Bosonic part of these terms takes a form of square of Riemann curvature $R^2$ times a compactification-dependent (local) function involving graviphoton field strengths and scalar fields in the vector and hyper multiplets. It was established in \cite{dewit} that universal near-horizon behavior of the black hole configuration, called attractor mechanism \cite{attrac}, remains valid even after including gravitational $F$-terms. Aforementioned works \cite{osv, dabholkar} are both built essentially on this observation: black hole entropy encodes information of gravitational $F$-terms. Despite such development, no analytic solution for BPS black holes is currently available once higher curvature terms are
35: incorporated. In this Letter, we report discovery of a class of exactly soluble $1\over 2$-BPS black holes in $d=4, N \ge 2$ supergravity with gravitational $F$-terms. They are built upon a simple variant of the supergravity prepotential, and hence are expected to serve as invaluable reference for better understanding of topological string, black hole entropy, stretched horizon and beyond.
36:
37: We shall adopt the method of superconformal multiplet calculus in constructing $N=2$ Poincar\'e supergravity coupled with matter, and follow closely the formalism developed in \cite{dewit}. Denote the vector multiplet as $(X^I, F^{+I}_{ab})$, the hypermultiplet
38: potential as $\chi$, the auxiliary multiplet as $(D, T_{ab})$, the Weyl curvature-squared multiplet as $(\hat{A}, \hat{C})$, and the prepotential as $F=F(X^I, \hat{A})$ with notations $F_I \equiv (\partial F / \partial X^I)$, $F_A \equiv \partial F / \partial \hat{A}$. Then relevant bosonic part of the Lagrangian reads
39: %
40: \bea && 8 \pi e^{-1} {\cal L} \nonumber \\
41: &=& i {\cal D}_\mu F_I {\cal D}^\mu \overline{X}^I - i F_I \overline{X}^I ({1 \over 6}R - D) + {\chi \over 4} ({1 \over 3}R + D) \nonumber \\
42: &+& {i \over 4} F_{IJ} (F^{-I}_{ab} -{1 \over 4} \overline{X}^I \overline{T}_{ab})(F^{-Jab} - {1 \over 4} \overline{X}^J \overline{T}^{ab}) \nonumber \\ &-& {i \over 8} F_I (F^{+I}_{ab} -{1 \over 4} X^I T_{ab})T^{ab} -{i \over 32} F (T_{ab})^2 \nonumber \\
43: &+& {i \over 2} \hat{F}^{-ab} F_{AI} (F^{-I}_{ab} -{1 \over 4} \overline{X}^I \overline{T}_{ab}) + {i \over 2} F_A \hat{C} + {\rm h.c.}\,. \eea
44: %
45:
46: The $1\over 2$-BPS black hole is constructable as follows \cite{dewit}. Denote local $N=2$ supersymmetry parameters as $\varepsilon_A = (\varepsilon^A)^*$ where $A=1,2$ are SU(2) indices. The $1 \over 2$-BPS condition reads $e^{ i \alpha (x)}
47: \varepsilon_A(x) = \epsilon_{AB} \gamma_0 \varepsilon^B(x)$, where $\alpha(r)$ is a phase-factor to be determined. This puts the spacetime metric for a single-centered black hole in Tod's form:
48: %
49: \begin{eqnarray}
50: \d s^2 = -e^{2 g(r)} \d t^2 + e^{-2g(r)} (\d r^2 + r^2 \d \Omega^2_2), \label{metric}
51: \end{eqnarray}
52: %
53: sets all hypermultiplet scalars and hence the potential $\chi$ to constant, and imposes all vector multiplet scalars to obey to so-called ``generalized stabilization conditions":
54: %
55: \begin{eqnarray} {\rm Im} Y^I &=& \widetilde{H}^I \equiv {1\over2}\left(\widetilde{h}^I
56: + {p^I \over r} \right) \nonumber \\
57: {\rm Im} F_I &=& H_I \equiv {1\over2}\left(h_I + {q_I \over r}\right). \end{eqnarray}
58: %
59: Here, $Y^I= e^{-g} X^I$ and $\Upsilon = e^{-2g} \hat{A}$ are scale and U(1) invariant variables of the vector and the Weyl curvature-squared multiplets, and $F = F(Y, \Upsilon)$ is the prepotential of degree 2, where $Y^I$ and $\Upsilon$ carry degree 1
60: and 2, respectively. Other equations to be solved together are
61: %
62: \begin{eqnarray} && \Upsilon = - 64 (\nabla_p g)^2 \label{upsilon} \\
63: && {\rm Im} (Y^I \overline{F}_I) + {\chi \over 4} e^{- 2g} = - 128
64: e^g \nabla_p ({\rm Im} F_{\Upsilon} \nabla_p e^{-g}) \label{g} \\
65: && \widetilde{H}^I \nabla_p H_I - H_I \nabla_p \widetilde{H}^I = 0, \end{eqnarray}
66: %
67: where we made use of rotational symmetry of the ansatz eq.(\ref{metric}).
68: All hypermultiplet scalars being constant, $\chi$ takes a negative constant value and sets an overall mass scale in the theory.
69:
70: We shall consider a theory defined by prepotential:
71: %
72: \bea F (Y, \Upsilon) = - \left( C {Y^1 Y^2 Y^3 \over Y^0} + { i \tau \over 2} (Y^3)^2 + k {Y^3 \over Y^0} \Upsilon \right), \label{prepotential} \eea
73: %
74: and look for a BPS black hole solution. Specifically, we shall look for a solution carrying both electric and magnetic charges, but not of dyonic type. Choosing these charges along $0$ and $3$ directions in vector multiplets, we find the solution is obtained by taking an ansatz that $Y^0$ is purely real, $Y^1=Y^2=0$, and $Y^3$ is purely imaginary. In addition, we shall fix the phase and the scale in the superconformal symmetry by the gauge $\alpha = 0, g(\infty) = 0$. The generalized stabilization equations are solved by
75: %
76: \bea -i Y^3 &=& \widetilde{H} \equiv \Big(\widetilde{h} + {p \over r}\Big) = \widetilde{h} \Big(1 + {\widetilde{Q} \over r} \Big) \qquad (\widetilde{Q} > 0) \label{y3} \\
77: +i F_0 &=& H \equiv \Big(h + {q \over r} \Big) = h \Big(1 + {Q \over 4}\Big) \qquad (Q > 0). \nonumber \eea
78: %
79: Using (\ref{upsilon}), $Y^0$ is then solved from $F_0$ in terms of $H, \tilde{H}, g$:
80: %
81: \bea Y^0 = 8 \sqrt{k} g'(r) \sqrt{\widetilde{H} / H}. \label{y0} \eea
82: %
83: We also have Im$F_\Upsilon = - (\sqrt{k}/8g') \sqrt{H \widetilde{H}}$ and
84: %
85: \bea {\rm Im} (Y^I \overline{F}_I) &\equiv& {e^{-K} \over 2} = \tau
86: \widetilde{H}^2 + 16 \sqrt{k} g'(r) \sqrt{H \widetilde{H}} \nonumber \eea
87: %
88: Using these expressions, eq.(\ref{g}) becomes
89: %
90: \bea \tau \widetilde{H}^2 &+& 16 \sqrt{k} g' \sqrt{H \tilde{H}} + {\chi\over 4} e^{-2g} \nonumber \\ &=& 16 \sqrt{k} e^g \, \nabla \cdot \left({1 \over g'} \sqrt{H \widetilde{H}} \nabla e^{-g} \right). \eea
91: %
92: The equation is further simplifiable to
93: %
94: \bea \tau \widetilde{H}^2 +{1 \over 4} \chi e^{-2g(r)} = - 16 {
95: \sqrt{k} \over r^2} \partial_r (r^2 \sqrt{H \widetilde{H}}). \label{eg} \eea
96: %
97: This equation determines the metric factor $e^{g(r)}$. Notice that derivatives of $g(r)$ all dropped out in eq.(\ref{eg}). This is because in our ansatz we took spherical symmetry and turned off $Y^1, Y^2$ completely. Later, we will revisit this issue when
98: considering more general solution.
99:
100: Now, in the gauge $g(\infty) = 0$ adopted, the negative constant $\chi$ is determined as $\chi = - 4 \tau \tilde{h}^2$ (Notice that we have taken $\tau > 0$). The final form of the solution yields eq.(\ref{y3}, \ref{y0}) along with the metric factor
101: %
102: \bea e^{-2g(r)} = \Big(1+{\widetilde{Q} \over r}\Big)^2 + {Q \varepsilon \over r^2} \left[r^2 \sqrt{\Big(1+{Q \over r}\Big)\Big(1+{\tilde Q \over r}\Big)}\right]' \nonumber\eea
103: %
104: where $\varepsilon = (16 \sqrt{k}/Q \tau \widetilde{h}^2) \sqrt{ h \widetilde{h}}$. The metric factor is remarkably simple and behaves similar to that of extremal charged black hole. To see this, choose $Q = \widetilde{Q} $. Then,
105: %
106: \bea e^{-2g(r)} = (1 + \varepsilon) \Big(1 + {Q \over r} \Big)^2 - \varepsilon. \eea
107: %
108: Near the horizon, $r \sim 0$, the spacetime metric becomes
109: %
110: \bea \d s^2 = (1 + \varepsilon) Q^2 \Big( {\d r^2 \over r^2} + \d \Omega_2^2 \Big) - {r^2 \d t^2 \over (1 + \varepsilon) Q^2}, \eea
111: %
112: and yields the Bertotti-Robinson geometry except rescaling by $(1 + \varepsilon)$. This near-horizon behavior is in fact universal for all values of $p, q$, equivalently, $Q, \widetilde{Q}$.
113:
114: These black holes carry nontrivial central charge $Z$, defined by the charge coupled to the graviphoton:
115: %
116: \bea Z(r) \equiv e^{{1 \over 2} K(r)} \left(p^3 F_3(r) - q_0 Y^0(r) \right). \eea
117: %
118: The BPS mass is then determined by $Z$ at spatial infinity:
119: %
120: \bea M_{\rm BPS} = |Z|_{r=\infty} = \sqrt{2 \tau} p . \label{mass}\eea
121: %
122: The entropy as defined by surface integral of Noether current at the horizon yields
123: %
124: \bea S_{\rm BH} &=& \pi \Big[ |Z|^2 -256 {\rm Im} \partial_{\Upsilon} F(Y, \Upsilon) \Big]_{r=0} \nonumber \\ &=& \pi\Big[ ( 2 \tau p^2 + 32 \sqrt{k} \sqrt{pq}) + (32 \sqrt{k}
125: \sqrt{pq} )\Big]. \label{entropy}\eea
126: %
127: The first term is given by the area of the horizon, while the second term, which is not manifestly of geometric origin, is due to Weyl curvature-squared terms. This shows clearly that the black hole has a finite horizon area even for a single magnetic charge along $1$-component of the vector multiplet, and is set by the quadratic coupling $\tau$ in the prepotential. In $\tau \rightarrow 0$ limit, the entropy is the same as the one studied in \cite{dabholkar}. From the solution, this limit appears singular, but one can take a scaling limit that $\tau \sim \delta^2, \tilde{h} = 1/\delta, h = \delta$. In this limit, the black hole becomes massless, but the geometry and the entropy remains finite so long as $k$ is nonzero. Notice also that mass-entropy relation of these black holes now reads
128: %
129: \bea S_{\rm BH} = \pi M_{\rm BPS}^2 + 64 \pi \Big[ {k q \over \sqrt{2 \tau} } M_{\rm BPS}\Big]^{1/2 }, \eea
130: %
131: deviating from familiar one by higher curvature effects.
132:
133: In obtaining the exact solution, we took $Y^1 = Y^2 = 0$. Actually, we can relax this and take more general ansatz $Y^1 = i \epsilon_1, Y^2 = i \epsilon_2$, where $\epsilon_{1,2}$ are real-valued constants. The stabilization equations are again solved similarly:
134: %
135: \bea
136: Y^0 &=& \sqrt{ C \epsilon_1 \epsilon_2 + 64 k (g')^2} \sqrt{\widetilde{H}/H} \nonumber \\
137: Y^1 &=& i \epsilon_1, \,\qquad Y^2 = i \epsilon_2, \,\qquad Y^3 = i \widetilde{H} \nonumber \\
138: F_0 &=& - i H, \quad F_1 = C \widetilde{H} \epsilon_2 / Y^0, \quad F_2 = C \widetilde{H} \epsilon_1 / Y^0 \nonumber \\
139: F_3 &=& \sqrt{C \epsilon_1 \epsilon_2 + 64 k (g')^2} \sqrt{H/\widetilde{H}} + \tau \widetilde{H}. \eea
140: %
141: The equation determining $g(r)$ is now given by
142: %
143: \bea -{1 \over 4} \chi e^{-2 g} &=& \tau \widetilde{H}^2 + 4 \sqrt{H \widetilde{H}} {C \epsilon_1 \epsilon_2 \over \sqrt{C \epsilon_1 \epsilon_2 + 64 k (g')^2}} \nonumber \\
144: &+& 128 k {1 \over r^2} \partial_r \Big({ r^2 \sqrt{H \widetilde{H}} g' \over \sqrt{C \epsilon_1 \epsilon_2 + 64k (g')^2}}\Big). \label{e-g}\eea
145: %
146: Now, this is a complicated {\sl differential} equation for $g(r)$. Still, it clearly shows that the solution behaves essentially the same as the one we have found above in the region where $(g')^2 \gg C \epsilon_1 \epsilon_2$. Indeed, in the near-horizon region, it is
147: straightforward to check that $g'(r) \sim 1/r$ is a consistent solution. In the region where $(g')^2 \ll C \epsilon_1 \epsilon_2$, the first two terms on the right-hand side are of order ${\cal O}(r^0)$, and the equation is reduced to a 3-dimensional Liouville type equation. It is easy to see that the solution is {\sl oscillatory} (related observations were made in
148: \cite{horizonrecent}):
149: %
150: \bea g(r) \rightarrow g_0 +{1 \over 2r} (g_1 + g_2 \sin (\omega r + \varphi
151: ) + {\cal O}(r^{-2}). \nonumber \eea
152: %
153: Expanding (\ref{e-g}) in $1/r$-series, we see that the oscillation frequency is set by
154: %
155: \bea \omega^2 = {1 \over 16 k}\left(C \epsilon_1 \epsilon_2 +{\tau \tilde{h}^2 \sqrt{C \epsilon_1 \epsilon_2}\over 4 \sqrt{h\tilde{h}}} \right) \eea
156: %
157: The coefficients $g_2, \varphi$ are undetermined at this order, but are
158: determinable at order $1/r^2$ and higher.
159:
160: Black hole's global geometry thus consists of three regions: near-horizon region of $AdS_2 \times S^2$, intermediate flat region, and outermost flat region.
161: See fig.\ref{fig1} for a cartoon view.
162: By turning on $Y^1, Y^2$ to nonzero constant values, we have made the black hole to open up a new asymptotic region patched as the outermost region, where the metric factor exhibits oscillatory behavior. See \ref{fig2}. Actually, the oscillatory behavior can be absorbed by redefining the metric as $g_{mn} \rightarrow (1+cR) g_{mn}$. At asymptotic infinity, by linearizing the metric as $g_{mn} = \eta_{mn} + h_{mn}$, it is straightforward to see that $h_{mn} \rightarrow h_{mn} + cR \eta_{mn}$ for $c = -g_2/12 \omega^2$ eliminates the oscillatory part completely. Notice that it can be made effective only at the outermost region.
163:
164: Behavior of the central charge $Z(r)$ is now complicated, but we will record it for reference:
165: %
166: \bea Z(r) = e^{K/2} \left( p^3 F_3 - q_0 Y^0 \right) \eea
167: %
168: where
169: %
170: \bea e^{-K} &=& 2 \tau \tilde{H}^2+ 4 \sqrt{H \widetilde{H}} \label{ek} \\
171: &\times&\Big( \sqrt{ C \epsilon_1 \epsilon_2 + 64 k (g')^2} + {C \epsilon_1 \epsilon_2 \over \sqrt{C \epsilon_1 \epsilon_2 + 64 k (g')^2}}\Big). \nonumber\eea
172: %
173: The BPS mass is then given again by $Z$ at spatial infinity:
174: %
175: \bea M_{\rm BPS} = |Z|_{r=\infty} = {\sqrt{2}\left(\tau p + \beta (p
176: +{\tilde{h} \over h} q) \right)\over \sqrt{\tau + 4 \beta}}, \eea
177: %
178: where $\beta = \sqrt{C \epsilon_1 \epsilon_2 h\tilde{h}} \slash \tilde{h}^2$. On the other hand, the entropy remains the same as (\ref{entropy}), since, near the horizon $r=0$, $\epsilon_1, \epsilon_2$ are completely negligible and the black hole behaves the same as the analytic solution found above.
179: %%%%%%%%%%%%%%%%%%%%
180: \begin{figure}
181: \includegraphics[width=2.in,height=2.2in,angle=0]{bh}
182: \caption{ \label{fig1} Cartoon view of the BPS black hole.}
183: \end{figure}
184:
185: Causal structure of the black hole is quite similar to Bertotti-Robinson geometry. This is most clearly seen for the special case $Q = \widetilde{Q}$. Introduce Eddington-Finkelstein coordinate $v = t + x$ where $x = r + 2Q(1 + \varepsilon) \log r - Q^2 (1 + \varepsilon)/r$. For $v=$ constant and $r \rightarrow 0$, the metric is nonsingular. Hence, the spacetime is extendible beyond the future horizon to $r<0$. We find that the
186: extended spacetime exhibits a timelike singularity at $r = r_\star = - (1 + \sqrt{\varepsilon / (1 + \varepsilon)})^{-1} Q$, where $g_{vv} = e^{2g(r)}$ diverges. Similar analysis and conclusion follow for the past horizon defined with coordinate $u = t - x$.
187: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
188: \begin{figure}
189: \includegraphics[width=2.5in,height=3.2in,angle=0]{osc1}
190: \caption{\label{fig2} $e^{-g(r)}$ for near-horizon, intermediate, and outermost regions. We set $h = \widetilde{h}=1, p=q=1, \tau=2, 64k=1, c \epsilon_1 \epsilon_2 = 10^{-3}$, so $64k (g')^2 \sim c \epsilon_1 \epsilon_2$ around $r \sim 8$.}
191: \end{figure}
192:
193: The exact black hole found above is based on the specific choice of the prepotential $F(Y, \Upsilon)$ as in (\ref{prepotential}). We shall now motivate how such a form might arise from string theory. The gravitational $F$-term, $\partial_\Upsilon^n F(Y, 0)$ originates from from one-loop $n$-point amplitudes for heterotic string theory compactified on $\mathbf{K}_3 \times \mathbf{T}^2$ or genus-$n$ amplitudes for Type II string theories compactified on a Calabi-Yau 3-fold $X$ with a structure of $\mathbf{K}_3$ fibration over $P^1(\mathbf{C})$. Each amplitude can be represented as topological partition function of topologically twisted string sigma model. It is of interest whether the prepotential (\ref{prepotential}) may be motivated from appropriate string theory compactifications. As is well-known, the special geometry of the corresponding vector multiplet moduli space ${\cal M}_{\rm V}$ is characterized by the period vector $(X^I, F_I)$ ($I=0,\cdots, n)$. The corresponding inhomogeneous special coordinates are $S = X^1/X^0$ and $T^A = X^A/X^0$ ($A=1,2,\cdots, n$). In heterotic string description, the prepotential ought to respect S-duality symmetry. Requiring covariance under the S-duality group, the prepotential becomes restricted to the form \cite{qbh}:
194: %
195: \bea F(X) = -{X^1 \over X^0} \eta_{AB} X^A X^B - G (X) \eea
196: %
197: where $G(X)$ is of degree 2 in $X^I$, invariant under the shift $S \rightarrow S + \mathbb{Z}$, and regular in the weak coupling limit $S \rightarrow + i \infty$. This implies that $G(X)$ is expandable in $m$-instanton contributions:
198: %
199: \bea G(X) = \sum_{m=0}^\infty e^{ 2 \pi i m S} g^{(m)}(X^0, X^A) \nonumber \eea
200: %
201: where again each $g^{(m)}$ is of degree 2 in $X^0, X^A$. We are interested in weak coupling limit, so set $g^{(m)}=0$ except for $m=0$. $g^{(m)}$ is also constrained to exhibit requisite singularity structure over ${\cal M}_{\rm V}$. For $g^{(0)}$, the singularity is of trilogarithmic type at gauge symmetry enhancement points. As is also well-known \cite{qbh}, T-duality symmetry imposes further constraints on $F(X)$. Such constraints are fairly complicated due to quantum corrections to the transformation laws of $X^I$'s. In particular, the dilaton $S = X^1/X^0$ is generically not invariant under T-duality. Rather, T-duality invariant dilaton $S_{\rm inv}$ takes the form
202: %
203: \bea S^{\rm inv} = S + {1 \over 2 (n+1)} \left( (\partial_2^2 + \cdots + \partial_n^2) g^{(0)} - {4 \over 2 \pi i} + K \right), \nonumber \eea
204: %
205: where $K$ is a function of degree {\sl zero}, determined up to an additive constant, and has no singularity over ${\cal M}_{\rm V}$. It thus shows that, at large $S, T^A$ far away from singular points, a given choice of $g^{(0)}$, which is of degree two, approximates specific form of quantum effects. Clearly, the simplest form is that $f^{(0)} = i \tau \delta_{AB} X^A X^B$. Its origin is somewhat trivial but is also known to be generic \cite{qbh}. By appropriate redefintion of $X^I$'s, we thus have the structure of
206: $F(X)$ in (\ref{prepotential}) we started with. As for the gravitational part of $F(X)$, it typically takes the form
207: %
208: \bea F_{\rm grav} = {\Upsilon} \sum_{A=1}^n{ \beta_A \over 4 \pi i}
209: \log \Delta(e^{4 \pi i Y^A/Y^0}), \label{gravpart} \eea
210: %
211: where $\Delta(q) = \eta(q)^{24}$ and $\eta(q)$ is the Dedekind eta function. for a given string compactification, threshold corrections gives rise these terms and $\beta_A (A=1,2,3)$ define gravitational $\beta$-functions. Again, at large $S, T^A$, eq.(\ref{gravpart}) is approximated by terms proportional to $Y^A/Y^0$. Our choice in (\ref{prepotential}) amounts to $\beta_1 = \beta_2 = 0$ and $\beta_3 = k$, but one can relaxed it and still find exact black holes.
212:
213: We thank Gungwon Kang and Ho-Ung Yee for useful discussions. DB was
214: supported by KOSEF ABRL R14-2003-012-01002-0 and KOSEF R01-2003-000-10319-0. SJR was supported in part by KOSEF Leading Scientist Grant and by W.F. Bessel Award of Alexander von Humboldt Foundation.
215:
216: \begin{thebibliography}{*99}
217: \bibitem{osv} H.~Ooguri, A.~Strominger and C.~Vafa,
218: Phys.\ Rev.\ D {\bf 70}, 106007 (2004) [arXiv:hep-th/0405146].
219: %%CITATION = HEP-TH 0405146;%%
220:
221: \bibitem{dabholkar}
222: A.~Dabholkar, arXiv:hep-th/0409148;
223: %%CITATION = HEP-TH 0409148;%%
224: A.~Dabholkar, R.~Kallosh and A.~Maloney, arXiv:hep-th/0410076.
225: %%CITATION = HEP-TH 0410076;%%
226:
227: \bibitem{sen}
228: A.~Sen, Mod.\ Phys.\ Lett.\ A {\bf 10}, 2081 (1995) [arXiv:hep-th/9504147].
229: %%CITATION = HEP-TH 9504147;%%
230:
231: \bibitem{osvrecent}
232: C.~Vafa, arXiv:hep-th/0406058; %%CITATION = HEP-TH 0406058;%%
233: U.~H.~Danielsson, M.~E.~Olsson and M.~Vonk, JHEP {\bf 0411}, 007 (2004) [arXiv:hep-th/0410141]; %%CITATION = HEP-TH 0410141;%%
234: M.~Aganagic, H.~Ooguri, N.~Saulina and C.~Vafa, arXiv:hep-th/0411280;
235: %%CITATION = HEP-TH 0411280;%%
236: E.~Verlinde, arXiv:hep-th/0412139.
237: %%CITATION = HEP-TH 0412139;%%
238:
239: \bibitem{horizonrecent}
240: A.~Sen, arXiv:hep-th/0411255;
241: %%CITATION = HEP-TH 0411255;%%
242: V.~Hubeny, A.~Maloney and M.~Rangamani, arXiv:hep-th/0411272.
243: %%CITATION = HEP-TH 0411272;%%
244:
245: \bibitem{dewit} We follow closely the formalism developed in G.~Lopes Cardoso, B.~de Wit, J.~Kappeli and T.~Mohaupt, JHEP {\bf 0012}, 019 (2000) [arXiv:hep-th/0009234]. For earlier works, see references cited therein.
246: %%CITATION = HEP-TH 0009234;%%
247:
248: \bibitem{attrac}
249: S.~Ferrara, R.~Kallosh and A.~Strominger, Phys.\ Rev.\ D {\bf 52}, 5412 (1995)
250: [arXiv:hep-th/9508072];
251: %%CITATION = HEP-TH 9508072;%%
252: S.~Ferrara and R.~Kallosh, Phys.\ Rev.\ D {\bf 54}, 1514 (1996)
253: [arXiv:hep-th/9602136];
254: %%CITATION = HEP-TH 9602136;%%
255: S.~Ferrara and R.~Kallosh,
256: Phys.\ Rev.\ D {\bf 54}, 1525 (1996)
257: [arXiv:hep-th/9603090].
258: %%CITATION = HEP-TH 9603090;%%
259:
260: \bibitem{noether}
261: J.~Lee and R.~M.~Wald, J.\ Math.\ Phys.\ {\bf 31} (1990) 725;
262: %%CITATION = JMAPA,31,725;%%
263: R.~M.~Wald, Phys.\ Rev.\ D {\bf 48}, 3427 (1993) [arXiv:gr-qc/9307038];
264: %%CITATION = GR-QC 9307038;%%
265: T.~Jacobson, G.~Kang and R.~C.~Myers, Phys.\ Rev.\ D {\bf 49}, 6587 (1994) [arXiv:gr-qc/9312023]; %%CITATION = GR-QC 9312023;%%
266: V.~Iyer and R.~M.~Wald, Phys.\ Rev.\ D {\bf 50}, 846 (1994) [arXiv:gr-qc/9403028];
267: %%CITATION = GR-QC 9403028;%%
268: T.~Jacobson, G.~Kang and R.~C.~Myers, Phys.\ Rev.\ D {\bf 52}, 3518 (1995) [arXiv:gr-qc/9503020]; %%CITATION = GR-QC 9503020;%%
269: V.~Iyer and R.~M.~Wald, Phys.\ Rev.\ D {\bf 52}, 4430 (1995) [arXiv:gr-qc/9503052].
270: %%CITATION = GR-QC 9503052;%%
271:
272: \bibitem{qbh} See, for example, K.~Behrndt, G.~Lopes Cardoso, B.~de Wit, R.~Kallosh, D.~Lust and T.~Mohaupt, Nucl.\ Phys.\ B {\bf 488}, 236 (1997) [arXiv:hep-th/9610105] %%CITATION = HEP-TH 9610105;%%
273: ; S.~J.~Rey, Nucl.\ Phys.\ B {\bf 508}, 569 (1997) [arXiv:hep-th/9610157].
274: %%CITATION = HEP-TH 9610157;%%
275:
276: \end{thebibliography}
277: \end{document}
278: