1: \documentclass[12pt,aps,floats,showpacs,amssymb,tightenlines]{revtex4}
2:
3:
4:
5: \usepackage{amsmath}
6: \usepackage{amsfonts}
7: \usepackage{amscd}
8: \usepackage{epsfig}
9: \usepackage{amssymb}
10: \usepackage{tabularx}
11:
12: \renewcommand{\a}{\alpha}
13: \renewcommand{\b}{\beta}
14: \newcommand{\g}{\gamma}
15: \newcommand{\e}{\epsilon}
16: \newcommand{\m}{\mu}
17: \renewcommand{\d}{\delta}
18: \newcommand{\w}{\omega}
19: \renewcommand{\t}{\tau}
20: \newcommand{\p}{\partial}
21: \newcommand{\A}{\bar{A}}
22: \newcommand{\R}{{\cal{R}}}
23: \newcommand{\s}{ Schwarzschild }
24: \newcommand{\x}{{x_h}}
25: \newcommand{\h}{{x_h}}
26: \renewcommand{\r}{{\hat r_h}}
27:
28:
29: \begin{document}
30: \title{Gravitational instability of static spherically
31: symmetric Einstein-Gauss-Bonnet black holes in five and six
32: dimensions}
33: \author{Mart\'{\i}n Beroiz, Gustavo Dotti and Reinaldo J. Gleiser}
34: \affiliation{Facultad de Matem\'atica, Astronom\'{\i}a y
35: F\'{\i}sica (FaMAF), Universidad Nacional de C\'ordoba, Ciudad
36: Universitaria, (5000) C\'ordoba, Argentina}
37: \email{gdotti@famaf.unc.edu.ar}
38:
39:
40:
41: \begin{abstract}
42: Five and six dimensional static, spherically
43: symmetric, asymptotically Euclidean black holes, are unstable under gravitational perturbations if
44: their mass is lower than a critical value set by the string tension.
45: The instability is due to the Gauss-Bonnet correction to Einstein's
46: equations, and was found in a previous
47: work on linear stability of Einstein-Gauss-Bonnet black holes with
48: constant curvature horizons in arbitrary dimensions.
49: We study the
50: unstable cases and calculate the
51: values of the critical masses. The results are relevant to the issue
52: of black hole production in high energy collisions.
53: \end{abstract}
54: \pacs{04.50.+h,04.20.Jb,04.30.-w,04.70.-s}
55:
56: \maketitle
57:
58: \section{Introduction}
59: The most conservative approach to gravity in higher dimensions is the
60: one due to Lovelock \cite{Lovelock}, in which the LHS of Einstein's
61: equation $G_{ab} + \Lambda g_{ab} = 8 \pi G \; T_{ab}$ is replaced with
62: ${\cal G}_{ab}$, the most general symmetric, divergency free rank $(0,2)$ tensor
63: than can be constructed out of the metric and its
64: first two derivatives. Lovelock's tensor is
65: \begin{equation} \label{l}
66: {\cal G}_{ab} = \sum_{n=0}^{[(D-1)/2]} c_ n G^{(n)}{}_{ab}
67: \end{equation}
68: where $D$ is the spacetime dimension, $[z]$ the highest integer
69: satisfying $[z] \leq z$, and $G^{(n)}{}_{ab}$ is obtained by
70: making appropriate contractions on a tensor product of $n$ copies of the
71: Riemman tensor, contractions that trivially vanish if $n > [(D-1)/2]$.
72:
73:
74:
75:
76: The first few $G^{(n)}{}_{ab}$'s are the spacetime metric
77: ${G^{(0)}}_{ab} =
78: g_{ab}$, Einstein's tensor $ {G^{(1)}}_{ab} = R_{ab}
79: -\frac{1}{2} R g_{ab}$, and the Gauss-Bonnet tensor
80: \begin{multline} \label{g2}
81: {G^{(2)}}_b{}^a = R_{cb}{}^{de} R_{de}{}^{ca} -2 R_d{}^c
82: R_{cb}{}^{da} -2 R_b{}^c R_c{}^a + R R_b{}^a
83: -\frac{1}{4} g^a_b \left(
84: R_{cd}{}^{ef}R_{ef}{}^{cd} - 4 R_c{}^d R_d{}^c + R^2 \right),
85: \end{multline}
86: If $D=4$, $G^{(n)}{}_{ab}$ vanishes for all $n>1$ and Lovelock
87: theory reduces to Einstein theory with a cosmological constant
88: $c_0$. Starting with $D=5$, we may add the ${G^{(2)}}_b{}^a$ term,
89: and the resulting theory, usually referred to as {\em
90: Einstein-Gauss-Bonnet theory} (EGB, for short), is the most general
91: Lovelock theory in five and six dimensions:
92: \begin{equation} \label{egb}
93: \Lambda {G_{(0)}}_b{}^a + {G_{(1)}}_b{}^a + \a {G_{(2)}}_b{}^a =
94: 8 \pi G \; T_b{}^a,
95: \end{equation}
96: As is well known, EGB theory arises in the low energy limit of
97: heterotic string theories
98: \cite{str1,bd}, $\alpha > 0 $ being proportional to the inverse string
99: tension, thus string related treatments of BHs in higher dimension
100: should use the EGB equations.
101: Spherically symmetric, asymptotically
102: Euclidean vacuum black hole solutions of the EGB equations
103: (\ref{egb}) with $\Lambda=0$ are well known since the eighties
104: \cite{bd,w1,w2}. They are given by
105: \begin{equation} \label{m1}
106: ds^2 = -f(r) dt^2 + f(r)^{-1} dr^2 + r^2 \bar g _{ij} dx^i dx^j,
107: \end{equation}
108: $\bar g _{ij} dx^i dx^j$ the line element of $S^n$,
109: $n=D-2$, and
110: \begin{equation} \label{m2}
111: f(r) = 1 +\frac{r^2}{\a(n-1)(n-2)} \left( 1 -
112: \sqrt{1+\frac { 4 \a \m \left( n-1 \right) \left( n-2 \right)}{n r^{n+1}}
113: } \; \right) .
114: \end{equation}
115: $\m$ above is an integration constant, related to the mass $M$ of the black
116: hole through \cite{m1,m2}
117: \begin{equation} \label{mass}
118: M = \frac{\mu}{8 \pi G } \left[ \frac{2 \pi^{\frac{n+1}{2}}}
119: {\Gamma \left( \frac{n+1}{2} \right) } \right] =: \frac{\mu {\cal
120: A}_n}{8 \pi G },
121: \end{equation}
122: ${\cal A}_n$ being the area of the $n-$sphere. For positive $\mu$
123: and $\alpha$, the case we are interested in, there is a single
124: horizon $r_h$ located at the only positive root of (note the missing
125: factor of $1/4$ in \cite{dg1a})
126: \begin{equation} \label{h1}
127: \mu = \frac{n r^{(n-3)}}{4} \left[ \alpha (n-1)(n-2)+ 2 r^2
128: \right],
129: \end{equation}
130: then
131: \begin{equation} \label{masa}
132: M = \frac{n r_h{}^{(n-3)} {\cal A}_n}{32 \pi G }
133: \left[ \alpha (n-1)(n-2)+ 2 r_{h}{}^2
134: \right].
135: \end{equation}
136: The temperature and entropy of the black hole (\ref{m1})-(\ref{m2}) are \cite{m1}
137: \begin{eqnarray} \label{temp}
138: T &=& \left[ \frac{(n-1)}{8 \pi r_h} \right] \left( \frac{2 r_h{}^2 + \alpha (n-2)(n-3)}{r_h{}^2 + \alpha (n-1)(n-2)} \right), \\ \label{ent}
139: S &=& \frac{r_h{}^n {\cal A}_n}{4 G } \left[ 1 + \frac{\alpha n (n-1)}{r_h{}^2} \right]
140: \end{eqnarray}
141: The specific heat can be obtained from (\ref{masa}) and (\ref{temp}) using
142: \begin{equation} \label{sh1}
143: C = \frac{\p M }{\p T} = \left(\frac{\p M }{\p r_h}\right)
144: \left(\frac{\p T}{\p r_h}\right)^{-1}.
145: \end{equation}
146: Introducing $\r := \frac{r_h}{\sqrt{\alpha}}$ we obtain
147: \begin{equation} \label{sh2}
148: C = - \left( \frac{n {\cal A}_n \alpha^{n/2}}{4 G} \right) \left[ \frac{
149: \r ^{n-2} ( \r^2+(n-1)(n-2))^2 (2 \r^2+(n-2)(n-3))}{2 \r^4 + (n-2)(n-7) \r^2 + (n-1)(n-2)^2(n-3)} \right].
150: \end{equation}
151: Note that (\ref{m1})-(\ref{m2}) reduces to the $n+2$
152: dimensional Schwarzschild-Tangherlini \cite{st} black hole of
153: Einstein's theory in the $\a \to 0$ limit, since
154: \begin{equation} \label{fst}
155: f(r) = 1 -\frac{2 \m}{n r^{n-1}} + {\cal O} (\a).
156: \end{equation}
157: The thermodynamic functions (\ref{temp})-(\ref{sh1}) reduce to the
158: Schwarzschild-Tangherlini (ST) ones in the $\alpha \to 0$ limit.
159: Note, however, that some solutions to the EGB equations are found to
160: diverge as $\alpha \to 0$,
161: an example being the solution (\ref{m1})-(\ref{m2})
162: with a plus sign in front of the square root
163: in (\ref{m2}). Other crucial issues strongly depend
164: on $\alpha$ being nonzero (we will restrict to $\alpha >0$, as in string theory).
165: Consider first five dimensional ($n=3$) EGB black holes. From
166: (\ref{h1}) follows that there is a minimum mass $\mu = \frac{3}{2} \alpha$
167: for black hole formation, otherwise, (\ref{m1})-(\ref{m2}) has a naked singularity.
168: This does not happen for five dimensional Schwarzschild-Tangherlini (ST) black holes. The temperature
169: $$ T_{5D} = \frac{r_h}{2 \pi (r_h^2 + 2 \alpha)}$$
170: goes to infinity as $r_h \to 0^+$ ($\mu \to 0^+$) for ST holes, whereas
171: it tends to zero as $r_h \to 0^+$ ($\mu \to \frac{3 \alpha}{2} ^+$) in the EGB case.
172: The specific heat is always negative in the ST case, whereas it has a pole in the
173: EGB case at $r_h{} = \sqrt{2 \alpha} $, with $C > 0 $ for $ r_h < \sqrt{2 \alpha}$,
174: and $C < 0 $ for $ r_h > \sqrt{2 \alpha}$, i.e, small five dimensional EGB black
175: holes can be in equilibrium with a heat bath, contrary to what happens for ST holes.
176: Six dimensional EGB black holes behave more like ST black holes, their temperature decreasing monotonically from infinity in the interval $0 < r_h < \infty$, and their specific heat being always negative. However, both five and six dimensional {\em low mass} EGB black holes were found to be unstable under (linear) gravitational perturbations \cite{dg1a,dg1b,dg2}, whereas all $D>4$ ST black holes are well known to be stable under linear gravitational perturbations \cite{gh}. \\
177: The purpose of this work is to find the values for the
178: critical mass below which
179: five and six dimensional EGB black holes become unstable under linear gravitational perturbations. The perturbation
180: treatment in \cite{dg1a,dg1b,dg2} is based in the decomposition in
181: tensor, vector and scalar modes given in \cite{koda}, which is a
182: higher dimensional generalization of the axial an polar modes found in the
183: Regge-Wheeler treatment of Schwarzschild perturbations \cite{rw}.
184: The metric perturbation in the tensor modes are made from symmetric,
185: divergency free tensor fields $T_{ij}$ on $S^n$ satisfying $D_k D^k
186: T_{ij} = -k_T{}^2 T_{ij}$, $D_j$ the covariant derivative on $S^n$.
187: Similarly, vector (scalar) mode perturbations are made from vector
188: (scalar) fields satisfying $D_k D^k T_{i} = -k_V{}^2 T_{i}$ ($D_k
189: D^k T = -k_S{}^2 T$). A detailed exposition of the construction of these modes
190: can be found in \cite{koda}. The spectrum of the Laplacian acting on
191: divergency free, rank $p$ symmetric tensors on $S^n$ is
192: \cite{higu}
193: \begin{equation} \label{spc}
194: k_p{}^2 = \ell (\ell+n-1) - p, \; \; \ell=0,1,2,...
195: \end{equation}
196: $k_S, k_V$ and $k_T$ correspond to $p=0,1$ and $2$ respectively.
197:
198:
199:
200:
201: \section{Scalar mode instability of five dimensional black holes}
202:
203: In five dimensions there is a single horizon located at (see (\ref{m2})-(\ref{h1}))
204: \begin{equation}
205: r_h = \sqrt{\frac{2}{3} \mu - \alpha},
206: \end{equation}
207: as long as $\mu$ is greater than $3 \alpha /2$, the minimum value
208: required for black hole formation. It is convenient to adopt
209: the dimensionless variables from Section 4a of reference \cite{dg2}
210: \begin{equation} \label{dv}
211: x := r \alpha^{-1/2},
212: \;\;\; m := \frac{\mu}{\alpha},
213: \end{equation}
214: then $x_h = \sqrt{\frac{2}{3} m - 1}$ and the dimensionless tortoise
215: coordinate
216: \begin{equation} \label{star}
217: x^*(x) := \int_{2 \, x_h}^x \frac{dx'}{f(x')}, \hspace{.5cm} f(x) =
218: 1 + \frac{x^2}{2} \left[ 1 - \sqrt{1 + \frac{8m}{3x}} \right]
219: \end{equation}
220: extends from minus to plus infinity.\\
221: Scalar perturbations in five dimensions ($n=3$) of harmonic number
222: $k_S{}^2 =\ell (\ell + 2), \ell = 2,3,..., $ (the modes $\ell=0,1$
223: are trivial \cite{koda}) are entirely described by a single
224: function $\hat \phi(t,x)$ governed by an equation $\left({\cal
225: H}_{k_S} + \alpha \p^2/\p t^2 \right) \hat \phi =0$, which admits
226: separation of variables $\hat \phi(t,x) = \phi(x) e^{\omega t}$,
227: giving ${\cal H}_{k_S} \phi
228: = -\alpha \omega^2 \phi \equiv \alpha E \phi$, with $\phi$ satisfying
229: appropriate boundary conditions (reference \cite{dg2}, eqs.
230: (61)-(66)). The ``Hamiltonian"
231: \begin{equation} \label{hs}
232: {\cal H}_{k_S} = -\frac{\p ^2}{\p x^*{}^2} + \alpha V_{k_S}
233: \end{equation}
234: can be constructed following Section V in \cite{dg2}.
235: A negative eigenvalue of ${\cal H}_{k_S}$ -real $\omega$- implies
236: that this mode grows exponentially with time, i.e., is unstable.
237: Generic perturbations have projections on each harmonic
238: (tensor, vector or scalar) mode. Since 5D
239: BHs were found to be stable under tensor and vector perturbations \cite{dg1a,dg1b,dg2}, they will be
240: unstable
241: if and only if a $k_S$
242: is found such that the spectrum of ${\cal H}_{k_S}$ is not positive.
243: The boundary conditions defining the space of functions on which
244: ${\cal H}_{k_S}$ acts determine its spectrum, $L^2(x^*,dx^*)$ being
245: an appropriate function space for black hole spacetimes (see
246: however, the discussion in \cite{dgns} regarding nakedly singular
247: spacetimes). The problem of stability is then entirely equivalent
248: to the quantum mechanical problem of determining the sign of the
249: lowest eigenvalue for each member of the family of Hamiltonians
250: ${\cal H}_{k_S}, k_S = \sqrt{\ell (\ell+2)}, \ell=2,3,...$
251: Our strategy to prove instability
252: consists in showing that, if $\mu/ \alpha$ is small enough, then
253: for sufficiently high $k_S$, there exists
254: a wave function with a negative expectation value of ${\cal
255: H}_{k_S}$ (numerical evidence of this fact was given in Section IVa of ref \cite{dg2}). This implies that the ground state of ${\cal H}_{k_S}$
256: has negative energy, from where the instability
257: follows.\\
258: From the results in Section IV of \cite{dg2}, we find,
259: after a long calculation, that, after introducing
260: \begin{equation} x_o :=
261: \sqrt{\x^2+1}, \;\;\; y := \sqrt{x^4+4 x_o{}^2},
262: \end{equation}
263: the
264: potential can be conveniently split as
265: \begin{equation}
266: U_{k_S} := \frac{\alpha V_{k_S}}{f} = {k_S}^2 \; q_{\infty} + q_0 +
267: \frac{ {k_S}^2 q_1 + q_2}{{\cal{D}}}
268: \end{equation}
269: where ${\cal D}$ is a quartic polynomial in ${k_S}$:
270: \begin{equation} \label{D}
271: {\cal D} = 2 \; x^2\; y^4 \; \left[({k_S}^2-3) y+6 x_o{}^2) \right]^2,
272: \end{equation}
273: and the $q'$s {\em do not depend} on $k_S$:
274: \begin{eqnarray}
275: q_{\infty} &=& \frac{(x^4-4 x_o{}^2)}{x^2 y^2} \\
276: q_0 &=& \frac{(x^8 + 120 x^4 {x_o}^2- 240 {x_o}^4)}{8 x^2 y^3 } \nonumber \\
277: && - \left[\frac{ x^{10} - 6 x^8 + 200 {x_o}^2 x^6 + 528 {x_o}^2 x^4
278: - 560 {x_o}^4 x^2 -480 {x_o}^4}{8
279: x^2 y^4} \right] \\
280: q_1 &=& 24 x^2 {x_o}^2 \left[ \left( 24 {x_o}^2 + 20 {x_o}^2 x^2 - 6
281: x^4-x^6 \right) y - 48 {x_o}^4 -
282: 8 {x_o}^2x^4+x^8 \right] \\
283: q_2 &=& 72 x_o^2 x^4 \left[ \left(
284: x^6-(2x_o^2-6)x^4-20x_o^2x^2-24x_o^4-24x_o^2 \right) y \right. \nonumber \\
285: &&\left. - x^8+2x_o^2 x^6+4x_o^2 x^4+40x_o^4 x^2+96x_o^4 \right]
286: \end{eqnarray}
287: Note that $q_{\infty}$ is negative in the range
288: $ |x| < x_c := \sqrt{2} \;
289: (1+\x^2)^{1/4}$, and that $0 < \x < x_c $ if and only if
290: $3/2 < m < 9/2 + 3
291: \sqrt{2}$. Suppose this is the case and let $\psi(x)$ be a real
292: ${\mathbb C}^{\infty}$ function vanishing outside $(\x,x_c)$,
293: normalized such that $$ 1 = \int_{-\infty}^{\infty} \bar \psi \psi
294: \; dx^* = \int_{\x}^{x_c} \frac{\psi^2}{f} dx$$ Using $\psi$ as a
295: test function, the expectation value of the kinetic piece of
296: (\ref{hs}) is
297: \begin{equation}
298: < - \p^2 / \p x^*{}^2 > = - \int_{-\infty}^{\infty} \bar \psi
299: \frac{\p^2 \psi}{\p x^*{}^2} \; dx^*
300: = \int_{\x}^{x_c} f \left[ \frac{\p}{\p x} \left( f \frac{\p \psi}{\p x} \right) \right]^2
301: dx
302: \end{equation}
303: and that of the scalar potential is
304: \begin{equation}
305: < \alpha V_{k_S} > = \alpha \int_{-\infty}^{\infty} \bar \psi V_{k_S} \psi dx^*
306: = \int_{\x}^{x_c} \psi^2 \;U_{k_S} dx
307: = {k_S}^2 Q_{\infty} + Q_{0} + Q({k_S})
308: \end{equation}
309: where
310: \begin{equation}
311: Q_{\infty} \equiv \int_{\x}^{x_c} \psi^2 q_{\infty} dx < 0
312: \end{equation}
313: and
314: \begin{equation}
315: Q_{0}
316: \equiv \int_{\x}^{x_c} \psi^2 q_{0} dx
317: \end{equation}
318: do not depend on $k_S$, and
319: \begin{equation} \label{qs}
320: Q({k_S}) = \int_{\x}^{x_c} \psi^2
321: \left(\frac{ {k_S}^2 q_1 + q_2}{{\cal{D}}} \right) dx
322: \end{equation}
323: Note that the integrand in (\ref{qs}) converges uniformly to zero in
324: the interval $x \in [\x,x_c]$ as $k_S \to \infty$. This follows from
325: the fact that ${\cal{D}}$ is strictly positive in $[\x,x_c]$ (see
326: (\ref{D})) and is a quartic polynomial in $k_S$. As a consequence,
327: $\lim_{k_S \to \infty} Q(k_S) = 0$ and thus $ < {\cal H}_{k_S}
328: > $ is negative for the given test function and large values of
329: $k_S$. Since the above construction is possible if
330: \begin{equation}
331: 3/2 < m < m_{c(5D)} = 9/2 + 3 \sqrt{2} \simeq 8.743,
332: \end{equation}
333: we conclude that, in this mass range, all (static,
334: spherically symmetric, asymptotically Euclidean) 5D black holes have
335: a high harmonic scalar instability.\\
336: Although solving the quantum mechanical
337: problem (\ref{hs}) analytically is out of consideration, in some
338: cases we were able to spot the fundamental energy using a shooting
339: algorithm to numerically integrate (\ref{hs}). This was done in the
340: standard coordinate
341: $x$ (instead of $x^*$), for which (\ref{hs}) reduces to an equation of the form
342: $\phi'' + P \phi ' + Q \phi =0$ with a regular singular point at the horizon.
343: The first few terms of the Frobenius series around the horizon were used to generate appropriate
344: initial conditions for the shooting algorithm.
345: As an example, we exhibit in fig.1
346: the scalar potential vs. $x$, together with the ground state wave
347: function corresponding to $m = 1.7$, $\ell=2$. We also remark that
348: no bound state was found for $m > m_{c(5D)}$.
349:
350: \begin{figure}[h]
351: \includegraphics[width=5in]{f1.EPS}
352: \caption{ The scalar potential (arbitrary scale) in five dimensions
353: for $\ell = 2$ and $\mu / \a = 1.69$ is shown together with the
354: ground state wave function and energy, found numerically using a
355: shooting algorithm. The origin of $x^*$ was chosen as in
356: (\ref{star}), $x^*=0$ for $x=2 x_h$.}
357: \end{figure}
358:
359:
360: \section{Tensor mode instability of six dimensional black holes}
361:
362: As in \cite{dg1a}, we find it convenient to introduce
363: dimensionless variables
364: \begin{equation} \label{dv2}
365: m := \mu \a^{-3/2}, \;\;\; x := r /(\mu \alpha)^{1/5} = r
366: \alpha^{-1/2} m^{-1/5} , \;\;\; dx^* /dx := 1/f
367: \end{equation}
368: The spectrum of the Laplacian on symmetric divergency free tensors
369: on $S^4$ is $k_T{}^2 = \ell (\ell + 3) - 2, \ell \in {\mathbb Z}$,
370: only $\ell > 1$ tensors being required to construct non trivial
371: tensor perturbations of 6D black holes. These perturbations are
372: entirely described by a single function $\hat \phi(t,x)$ governed by
373: an equation that, after separation of variables $\hat \phi(t,x) =
374: \phi(x) e^{\omega t}$, assumes the form ${\cal H}_{k_T} \psi
375: = -\alpha m^{2/5} \omega^2 \phi \equiv \alpha m^{2/5} E \phi$ (\cite{dg1a},
376: equation (16)),
377: with ``Hamiltonian"
378: \begin{equation} \label{ht}
379: {\cal H}_{k_T} = -\frac{\p ^2}{\p x^*{}^2} + \alpha
380: m^{\frac{2}{5}} V_{k_T},
381: \end{equation}
382: $V_{k_T}$ being the RHS of eq. (18) in \cite{dg1a}. From \cite{dg1a}
383: we can readily construct the potential, the result is
384: \begin{equation} \label{p6d1}
385: U := \frac{\a V_{k_T}}{f} = (k_T{}^2+2)m^{-\frac{2}{5}} \; U_o + m^{-\frac{2}{5}}\;
386: U_1 + U_2 - U_3
387: \end{equation}
388: where the $U_j$'s depend {\em only} on $x$:
389: \begin{eqnarray}
390: U_0 &=& \frac{2 (x^5 + 6)^2-75}{2 x^2(x^5+1) (x^5+6)}\\
391: U_1 &=& \frac{8 \;x^{20} + 72 \; x^{15} + 1218 \; x^{10} + 1752\; x^5 -27}{4 x^2 (x^5+1)^2 (x^5+6)^2} \\
392: U_2 &=& \frac{24 \; x^{20} + 336 \; x^{15}+ 2414 \; x^{10} + 2916 \; x^5 + 189}{24 (x^5+1)^2 (x^5+6)^2} \\
393: U_3 &=& \frac{ 24 \; x^{20} + 216 \; x^{15} + 1154 \; x^{10} + 1506 x^5 -81}{24 x^5 (x^5+1)^2 (x^5+6) \sqrt{1+ 6/x^5}}
394: \end{eqnarray}
395: Let $x_c = (\sqrt{75/2}-6)^{1/5} \simeq 0.658$ be the only positive
396: root of $U_o$, note that $U_0 < 0$ for $0< x < x_c$. The $x$
397: coordinate of the horizon is
398: \begin{equation} \label{h}
399: x_h = \frac{z^2-4}{2 m^{\frac{1}{5}} \; z}, \;\; z = \left( 2 m + 2 \sqrt{16 + m^2} \right) ^{1/3}.
400: \end{equation}
401: $x_h$ is a monotone increasing function of $m$, and $x_h = x_c$ at
402: $m=m_{c(6D)}$ given by
403: \begin{equation} \label{mc6d}
404: m_{c(6D)} = \frac{72 \sqrt{6} \left( 5 \sqrt{6} -12 \right)
405: ^{3/2}}{\left( 12 - 5 \sqrt{6} + 6^{1/4} \; \sqrt{5} \sqrt{5
406: \sqrt{6}-12} \right) ^{5/2}} \simeq 7.965
407: \end{equation}
408: If $m<m_{c(6D)}$, then $x_h < x_c$ and we can take a test function
409: supported in $(x_h,x_c)$, so that the expectation value of the
410: $U_0$ piece of the potential is negative. Note from (\ref{p6d1})
411: that this term is proportional to the harmonic $\ell (\ell +3)$, and
412: no other term of ${\cal H}_{k_T}$ depends on $\ell$, thus the
413: expectation value of ${\cal H}_{k_T}$ for such a test function will
414: be negative for sufficiently high harmonic number. We conclude that
415: 6D BHs are unstable if $m < m_c$ above.\\
416: Now we prove
417: stability for $m > m_{c(6D)}$: $U_o,U_1$ (and $U_2$) are positive if
418: $x>x_c$, whereas $U_2-U_3 > 0$ if $x > x_c' \simeq 1.176$. Since
419: $x_h = x_c'$ for $m = m' \simeq 48.927$, stability will follow if
420: we prove that $U
421: > 0$ for $\ell=2,3...$, $m_{c(6D)} < m < m'$ and $x > x_h$ given in (\ref{h}). A lower bound
422: for $U$ in this region of parameter space is given by the minimum of
423: the single variable function $U_L := (10 U_0+U_1) 50^{-2/5} + U_2
424: -U_3$ in the interval $x \in (x_c,\infty)$. After some work $U_L$
425: can be seen to be positive in this interval, thus proving stability.
426: We conclude that 6D BHs are linearly unstable if and only if $ \mu /
427: \alpha^{3/2} =m < m_{c(6D)}$.
428:
429: Fig.2 exhibits the potential and fundamental state
430: (found numerically) corresponding to $\ell=2$, $\mu \a^{-3/2} =
431: 1.85$.
432:
433: \begin{figure}[h]
434: \includegraphics[width=5in]{f2.EPS}
435: \caption{\label{fig4} The tensor potential (arbitrary scale) in six
436: dimensions for $\ell = 2$ and $\mu \a^{-3/2} = 1.85$ is shown
437: together with the ground state wave function and energy, found
438: numerically using a shooting algorithm. The origin $x^*=0$
439: corresponds to $x=2 x_h$.}
440: \end{figure}
441:
442: \section{Conclusions}
443:
444: Gauss-Bonnet corrections to Einstein's equations in higher
445: dimensions have been considered in many different models, and
446: naturally arise in the low energy effective action of certain string
447: theories. However, their
448: effects on black hole formation has long been disregarded. The instability
449: found in \cite{dg1a,dg1b,dg2} and this paper implies that the
450: simplest EGB black holes (asymptotically Euclidean, static,
451: spherically symmetric), which are the closest analogue of \s black
452: holes, cannot actually be formed in five space time dimensions if
453: their mass parameter $\mu$ (see (\ref{m2})-(\ref{mass})) is less
454: than $\sim 8.743 \alpha$. The Gauss-Bonnet term also prevents the
455: formation of these black holes in six dimensions unless $\mu$ is
456: greater than $\sim 7.965 \alpha^{3/2}$. The implications of these
457: figures depend on the context where (\ref{m1})-(\ref{m2}) is used.
458: As an example, the $n$-dimensional EGB black hole
459: (\ref{m1})-(\ref{m2}) is an approximate EGB solution if we
460: periodically identify one of the asymptotically Euclidean
461: coordinates with a period much larger than the horizon radius, and
462: our perturbative analysis should be valid in this setting. The large
463: extra dimensions scenario (suitable only for $D \geq 6$,
464: \cite{alexeyev}) is of interest because it allows
465: $\alpha$ to be in the TeV scale \cite{alexeyev, otros},
466: and so mini black holes could be produced in high energy collisions and be eventually
467: detected at LHC.
468: In view of our results, the probability of these events may be severely
469: limited due to low mass black hole instabilities. As far as we know, this fact has not been taken into account
470: in previous calculations on black hole production rates
471: in high energy collisions.
472: In theories where the EGB equations simply arise as a low energy effective theory
473: of some quantum gravity model, $\alpha$
474: is of the order of the Planck scale and the bounds we obtained
475: for small black hole masses are much more stringent.
476:
477: \section*{Acknowledgments}
478:
479: This work was supported in part by grants of the Universidad
480: Nacional de C\'ordoba and CONICET (Argentina).
481: GD and RJG are supported by
482: CONICET (Argentina).
483:
484: \begin{thebibliography}{9}
485: \bibitem{Lovelock} D. Lovelock, {\it Jour. Math. Phys.} {\bf 12} (1971) 498.
486: \bibitem{str1} M.~J.~Duff, B.~E.~W.~Nilsson and C.~N.~Pope,
487: %``GAUSS-BONNET FROM KALUZA-KLEIN,''
488: Phys.\ Lett.\ B {\bf 173}, 69 (1986);
489: D. J. Gross and E. Witten, Nucl. Phys. {\bf B277}, 1 (1986);
490: B. Zumino, Phys. Rep. {\bf 137}, 109 (1985); B. Zwiebach, Phys.
491: Lett. {\bf 156B}, 315 (1985); D. Friedan, {\it Phys. Rev. Lett.}
492: {\bf 45} (1980) 1057; I. Jack, D.Jones and N. Mohammedi, Nucl. Phys.
493: {\bf B322}, 431 (1989); C. Callan, D. Friedan, E. Martinec and M.
494: Perry, Nucl. Phys. {\bf B262}, 593 (1985).
495: \bibitem{bd}
496: D.~G.~Boulware and S.~Deser,
497: %``String Generated Gravity Models,''
498: Phys.\ Rev.\ Lett.\ {\bf 55}, 2656 (1985).
499: \bibitem{w1} J. T. Wheeler, {\it Nuc. Phys.} {\bf B268} (1986) 737
500: \bibitem{w2} J. T. Wheeler, {\it Nuc. Phys.} {\bf B273} (1986) 732.
501: \bibitem{m1} R.~C.~Myers and J.~Z.~Simon,
502: %``BLACK HOLE THERMODYNAMICS IN LOVELOCK GRAVITY,''
503: Phys.\ Rev.\ D {\bf 38} (1988) 2434.
504: \bibitem{m2} R.~C.~Myers and M.~J.~Perry,
505: %``Black Holes In Higher Dimensional Space-Times,''
506: Annals Phys.\ {\bf 172}, 304 (1986).
507: \bibitem{dg1a}
508: %``Gravitational instability of Einstein-Gauss-Bonnet black holes under
509: %tensor mode perturbations,''
510: G.~Dotti and R.~J.~Gleiser, Class.\ Quant.\ Grav.\ {\bf 22}
511: (2005) L1
512: [arXiv:gr-qc/0409005].
513: \bibitem{st} F.R. Tangherlini, {\it Nuovo Cimento} {\bf 27}, 365
514: (1963).
515: \bibitem{dg1b}
516: G.~Dotti and R.~J.~Gleiser,
517: %``Linear stability of Einstein-gauss-bonnet static spacetimes. part. I:
518: %Tensor perturbations,''
519: Phys.\ Rev.\ D {\bf 72} (2005) 044018
520: [arXiv:gr-qc/0503117].
521: \bibitem{dg2}
522: R.~J.~Gleiser and G.~Dotti,
523: %``Linear stability of Einstein-Gauss-Bonnet static spacetimes. II: Vector
524: %and scalar perturbations,''
525: Phys.\ Rev.\ D {\bf 72} (2005) 124002
526: [arXiv:gr-qc/0510069].
527: \bibitem{gh}
528: G.~Gibbons and S.~A.~Hartnoll,
529: %``A gravitational instability in higher dimensions,''
530: Phys.\ Rev.\ D {\bf 66} (2002) 064024
531: [arXiv:hep-th/0206202].
532: \bibitem{koda} H. Kodama, A. Ishibashi and O. Seto, Phys. Rev. {\bf D62} (2000)
533: 064022.\\H. Kodama and A. Ishibashi {\it Prog.Theor.Phys.} {\bf 111}
534: (2004) 29, hep-th/0308128;
535: H. Kodama and A. Ishibashi, gr-qc/0312012.
536: \bibitem{rw} T. Regge and J. A. Wheeler, Phys. Rev. 108 (1957)
537: 1063.
538: \bibitem{higu} A. Higuchi, J. Math. Phys. 28 (1987) 1553;
539: J. Math. Phys. 43 (2002) 6385 (erratum);
540: M.~A.~Rubin and C.~R.~Ordo\~nez,
541: %``Symmetric Tensor Eigen Spectrum Of The Laplacian On N Spheres,''
542: J.\ Math.\ Phys.\ {\bf 26} (1985) 65.
543: \bibitem{dgns}
544: R.~J.~Gleiser and G.~Dotti,
545: %``Instability of the negative mass Schwarzschild naked singularity,''
546: Class.\ Quant.\ Grav.\ {\bf 23}, 5063 (2006)
547: [arXiv:gr-qc/0604021].
548: \bibitem{alexeyev}
549: S.~Alexeyev, A.~Barrau and J.~Grain,
550: %``Gauss-Bonnet black holes at the LHC: beyond the dimensionality of space,''
551: %\href{http://www.slac.stanford.edu/spires/find/hep/www?irn=5868734}{SPIRES entry}
552: {\it Proceedings of Quarks-2004, 13th International Seminar on High Energy Physics, Pushkinskie Gory, Russia, May 24-30, 2004}; S.Alexeyev, N.Popov, A.Barrau and J.Grain,
553: Journal of Physics: Conference Series {\bf 33} (2006) 343–348.
554: \bibitem{otros} N. Arkani-Hamed, S. Dimopoulos and G.R. Dvali, Phys. Lett. {\bf B 429} (1998) 257
555: I. Antoniadis et al., Phys. Lett. {\bf B 436} (1998) 257
556: N. Arkani-Hamed, S. Dimopoulos and G.R. Dvali, Phys. Rev. {\bf D 59} (1999) 086004.
557:
558:
559: \end{thebibliography}
560:
561: \end{document}
562: