hep-th0507100/GB.tex
1: \documentclass[twocolumn,english,showpacs]{revtex4}
2: %\documentclass[preprint,english,showpacs,floatfix]{revtex4}
3: \usepackage{babel,amsmath,amssymb,dcolumn}
4: \usepackage[dvips]{graphics}
5: 
6: 
7: 
8: \begin{document}
9: 
10: \title{Scalar field evolution in Gauss-Bonnet black holes}
11: 
12: 
13: \author{E. Abdalla}
14: \email{eabdalla@fma.if.usp.br}
15: %
16: \author{R.A. Konoplya}
17: \email{konoplya@fma.if.usp.br}
18: %
19: \affiliation{Instituto de F\'{\i}sica, Universidade de S\~{a}o Paulo \\
20: C.P. 66318, 05315-970, S\~{a}o Paulo-SP, Brazil}
21: 
22: 
23: \author{C. Molina}
24: \email{cmolina@usp.br}
25: %
26: \affiliation{Escola de Artes, Ci\^{e}ncias e Humanidades, Universidade de
27:   S\~{a}o Paulo\\ Av. Arlindo Bettio 1000, CEP 03828-000, S\~{a}o
28:   Paulo-SP, Brazil} 
29: 
30: 
31: \pacs{04.30.Nk,04.50.+h}
32: 
33: 
34: \begin{abstract}
35: It is presented a thorough analysis of scalar perturbations in the
36: background of Gauss-Bonnet, Gauss-Bonnet-de Sitter and
37: Gauss-Bonnet-anti-de Sitter black hole spacetimes. The 
38: perturbations are considered both in frequency and time
39: domain. The dependence of the scalar field evolution  
40: on the values of the cosmological constant $\Lambda$
41: and the Gauss-Bonnet coupling $\alpha$ is investigated.
42: For Gauss-Bonnet and  Gauss-Bonnet-de Sitter black holes, 
43: at asymptotically late times either power-law or exponential tails
44: dominate, while for Gauss-Bonnet-anti-de Sitter black hole, the
45: quasinormal modes govern the scalar field decay at all times. 
46: The power-law tails at asymptotically late times 
47: for odd-dimensional Gauss-Bonnet black holes does not depend on $\alpha$,  
48: even though the black hole metric contains $\alpha$ as a new
49: parameter. The corrections to quasinormal spectrum due to Gauss-Bonnet
50: coupling is not small and should not be neglected. For the limit of
51: near extremal value of the (positive) cosmological constant
52: and pure de Sitter and anti-de Sitter modes in Gauss-Bonnet gravity we
53: have found analytical expressions.     
54: \end{abstract}
55: 
56: 
57: \maketitle
58: 
59: 
60: 
61: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
62: \section{Introduction}
63: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
64: 
65: 
66: Black holes in more than four spacetime dimensions are of
67: considerable interest recently  due to the two main reasons: they naturally appear
68: in string theory, and in extra dimensional brane-world scenarios  \cite{rs}.
69: According to some of these  scenarios it is possible that
70: the small higher dimensional black holes can be produced in particles
71: collisions in Large Hadron Collider. At the same time quantum gravity
72: may show itself  already at TeV-energy scale. Yet, the  effects of
73: quantum gravity then may be observed as corrections to classical
74: General Relativity.     
75:   
76: String theory predicts quantum corrections to classical General
77: Relativity, and the Gauss-Bonnet terms is the first and  dominating
78: correction among the others. 
79: Several higher other theories of gravity sustain black hole
80: solutions. The solution for neutral black hole in Gauss-Bonnet
81: gravity was obtained by Boulware and Deser \cite{deser} and  
82: Wheeler \cite{wheeler} . More generally, Lovelock gravity 
83: \cite{lovelock} has been studied and shown to possess black hole 
84: solutions with interesting thermodynamical properties
85: \cite{zanelli,abdcorrea}. 
86: 
87: Thus, the problem of black hole production in transplanckian particle
88: collisions has attracted considerable interest recently in the context
89: of large extra dimensions scenarios of TeV-scale gravity. It was
90: observed that the classical spacetime has large curvature along the
91: transverse collision plane, as signaled by the curvature invariant
92: $(R_{ijkl} R^{ijkl}$  and thereby quantum gravity effects, and  higher
93: curvature corrections to the Einstein gravity, cannot be ignored
94: \cite{Rychkov}. At the same time we know that  after formation of
95: such a black hole its evolution has three stages: first, it looses its
96: ``hairs'' coming into Kerr-like phase, then looses angular momentum
97: transforming to Schwarzschild-like black hole, and finally exerts
98: strong Hawking evaporation what results in loosing mass (see for
99: instance \cite{Kanti2005} and references therein). The stage when
100: black hole perturbations  decay, transforming perturbed black hole
101: into unperturbed one, is governed by quasinormal modes and is the aim
102: of our present research.     
103: 
104: In this paper we consider quasinormal perturbations of Gauss-Bonnet
105: black holes including a non-vanishing cosmological
106: constant. Quasinormal modes are a very useful tool to uncover
107: properties of the intrinsic geometry, since the modes characterizes
108: well the geometry and does not depend on further extrinsic properties,
109: independent of the geometry itself \cite{Kokkotas-99}. They have been
110: used successfully in a large class of astrophysical questions, from
111: black holes to stars. In addition, it has been argued that the
112: Gauss-Bonnet gravity in asymptotically anti-de Sitter (AdS)
113: spacetimes may be analyzed through anti-de Sitter/conformal field
114: theory (AdS/CFT) correspondence within next-to-leading order
115: \cite{OdintsovJHEP}. In this case the quasinormal modes of the large
116: Gauss-Bonnet-AdS black hole could find a holographic interpretation
117: in conformal field theory, as is the cases for the AdS black hole in
118: Einstein gravity \cite{Horowitz00}.   
119:    
120: In \cite{Konoplya05} the quasinormal modes for a charged asymptotically
121: flat black hole in Gauss-Bonnet gravity were found with the help of
122: the WKB approach \cite{yer}. The tensor-type gravitational
123: perturbations for Gauss-Bonnet black hole has been considered recently
124: in \cite{Dotti}.
125: 
126: To obtain the quasinormal modes we use numerical analysis as well as a
127: semi-analytical WKB-type treatment. Such an approach is based on the fact
128: that the wave equation is similar to a Schr\"odinger equation, and
129: depending on the kind of potential, it makes sense to borrow the methods
130: used in quantum mechanics in order to define a semi-classical
131: approximation. This vein has been followed and an approximation for the
132: quasinormal frequencies has been obtained to a high WKB order \cite{yer}.  
133: In addition to the frequency domain, we analyze the evolution of 
134: scalar perturbations in the time domain and find good agreement
135: between the results found by the two approaches.
136: 
137: We have observed that at asymptotically late
138: time, power-law tails do not depend on the Gauss-Bonnet coupling and
139: are the same as for the $d$-dimensional Schwarzschild black hole,
140: when $d$ is odd. For several simpler particular cases, namely, for
141: pure de Sitter and anti-de Sitter space-time (without black hole), 
142: and for extremal Gauss-Bonnet-de Sitter black hole we have found exact 
143: analytical formulas for (quasi)normal modes. The QNMs for 
144: Gauss-Bonnet black holes, with coupling $\alpha \sim 1$ predicted by 
145: string theory, is seemingly different from those of Schwarzschild
146: black hole. Therefore the GB-corrections to the QN spectrum should not 
147: be ignored, when considering Tev-scale of quantum gravity scenarios.  
148: All found modes are damping, what implies stability of Gauss-Bonnet 
149: black holes against scalar field perturbations.    
150: 
151: The paper is organized as follows: Sec. II represents the
152: preliminaries of the Gauss-Bonnet-(A)dS metric and its scalar
153: perturbations. Sec. III is devoted to the methods used in the paper,
154: namely the WKB method (in the frequency domain), and the
155: characteristic integration method (in the time domain).
156: Sec. IV discuss the quasinormal behavior of the Gauss-Bonnet (GB),
157: Gauss-Bonnet-de Sitter (GBdS) and  Gauss-Bonnet-anti-de Sitter (GBAdS)
158: black holes. In Sec. V we discuss the future perspective and some
159: unsolved questions in this field.  
160: 
161: 
162: 
163: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
164: \section{Gauss-Bonnet Black Hole Solutions}
165: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
166: 
167: The Einstein-Gauss-Bonnet action in the \mbox{$d$-dimensional} spacetime
168: model has the form
169: %
170: $$
171: I = \frac{1} {16 \pi G_{d} } \int d^{d} x \sqrt{-g} R
172: $$
173: \begin{equation}
174: + \alpha\prime \int d^{d} x \sqrt{-g} (R_{abcd}R^{abcd} - 4 R_{cd}
175: R^{cd}
176: + R^{2} - 2 \Lambda)  
177: \end{equation}
178: %
179: where $R$ and $\Lambda$ are the $d$-dimensional Ricci scalar and the
180: cosmological constant, respectively.  The parameter
181: $\alpha$ represents the (positive) Gauss-Bonnet coupling constant,
182: which is related to the Regge slope parameter or string scale.    
183: 
184: The Gauss-Bonnet Lagrangian ${\cal L}_{GB}$ is given by
185: %
186: \begin{equation}
187: {\cal L_{GB}} = R^2 - 4 R_{\mu\nu} R^{\mu\nu}+ 
188: R_{\mu\nu\rho\sigma} R^{\mu\nu\rho\sigma}\quad .
189: \label{GB_term}
190: \end{equation}
191: %
192: One should note that in four dimensions the Gauss-Bonnet term
193: (\ref{GB_term}) is a total divergency, and yields upon integration 
194: a topological invariant, namely the genus of the hypersurface
195: defining the Gauss-Bonnet action (but even in four dimensions there
196: are interest in the GB correction, as seen  in \cite{olea} for
197: example).
198: 
199: A metric obtained as a solution of the field equations is given
200: by
201: %
202: \begin{equation}
203: ds^{2}=-h(r)dt^{2} + h(r)^{-1}dr^{2}  +r^{2} d\Omega^{2}_{d-2}
204: \quad ,
205: \label{metric_form}
206: \end{equation}
207: %
208: where the function $h(r)$ is given by the expression
209: %
210: \begin{equation}
211: h(r) = 1 + \frac{r^2}{2\alpha} - \frac{r^2}{2\alpha} 
212:        \sqrt{1 + \frac{8\alpha\mu}{r^{d-1}} 
213:        + \frac{8\alpha \Lambda}{(d-1)(d-2)}   } \, .
214: \label{metric}
215: \end{equation}
216: %
217: The constant $\mu$ is proportional to the black hole mass and
218: $d\Omega_{d-2}^{2}$ is the line element of the
219: ($d-2$)-dimensional unit sphere.  The constants 
220: $\alpha$ and $\alpha\prime$ are connected by the relation:
221: \begin{equation}
222: \alpha = 16 \pi G_{d} (d-3) (d-4) \alpha\prime
223: \end{equation}
224: 
225: We set up a scalar field $\Phi$ on such a background obeying the
226: Klein-Gordon equation 
227: %
228: \begin{equation}
229: \Box \Phi = \frac{1}{\sqrt{-g}} \partial_{\mu} (\sqrt{-g} g^{\mu\nu}
230: \partial_{\nu} \Phi) = 0 \quad .
231: \end{equation}
232: %
233: In order to separate the wave function in terms of eigenpotential we first
234: separate the variables as  
235: %
236: \mbox{$\Phi (t,r, \{\theta_{i}\}) = R_{\ell}(t,r)
237: Y_{\ell m}(\{\theta_{i}\})/r$}. 
238: %
239: As usual we obtain a simple equation for
240: $R_{\ell}(t,r)$, which is given by the expression
241: %
242: \begin{equation}
243: 4 \frac{\partial^2 R_{\ell}(u,v)}{\partial u\partial v} + 
244: V(r(u,v)) R_{\ell}(u,v)=0 
245: \label{simpledalembertian}
246: \end{equation}
247: %
248: where $u= t-r_\star$, $v=t+r_\star $ and the tortoise coordinate
249: $r_\star $ is defined by the relation
250: %
251: \begin{equation}
252: \frac{dr_\star}{dr}=\frac{1}{h(r)}\quad .
253: \end{equation}
254: %
255: The variables $u$ and $v$ are the light cone coordinates corresponding to the
256: time and tortoise coordinate. The effective potential for the
257: scalar field in (\ref{simpledalembertian}) is  
258: %
259: \begin{eqnarray}
260: V(r) = h(r) \left[ \frac{(d-2)(d-4)}{4 r^{2}} h(r) \right. 
261:                                   \nonumber \\
262: %
263:                 \left. \mbox{} + \frac{d-2}{2r} \frac{d h(r)}{dr}
264:                   + \frac{\ell (\ell + d - 3)}{r^{2}}  \right] \quad .
265: \end{eqnarray}
266: 
267: The effective potential is positive definite potential barrier for any
268: $l$ for GB black hole, and, for $l>0$ for GBdS black hole (For $l=0$
269: GbdS case, the negative pitch appears). For GBAdS case the potential
270: diverges at infinity. 
271: 
272: 
273: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
274: \section{Numerical and Semi-analytical approaches}
275: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
276: 
277: 
278: \subsection{Characteristic integration}
279: 
280: In \cite{Gundlach-94} a simple but very efficient way of dealing with 
281: two-dimensional d'Alembertians  has been set up. Along the general lines
282: of the pioneering work \cite{Price-72},  light-cone variables have been
283: introduced, leading to (\ref{simpledalembertian}). 
284: 
285: In the characteristic initial value problem, initial data are specified
286: on the two null surfaces $u = u_{0}$ and $v = v_{0}$. The basic
287: aspects of the field decay are independent of the initial conditions
288: (as confirmed by simulations), so we use Gaussian initial conditions.
289: 
290: Since we do not have analytic solutions to the
291: time-dependent wave equation with the effective potentials introduced,
292: one approach is to discretize the equation (\ref{simpledalembertian}), and
293: then implement a finite differencing scheme to solve it
294: numerically. One possible discretization, used for example in
295: \cite{Wang-01,Brady-97,Brady-99},  is  
296: %
297: \begin{eqnarray}
298: \lefteqn{R_{\ell}(N) = R_{\ell}(W) + R_{\ell}(E) -
299: R_{\ell}(S) }  \nonumber \\  
300: %
301: & & \mbox{} - \Delta^2 V(S) \frac{ R_{\ell}(W) + R_{\ell}(E)}{8} \ , 
302: \label{discrete1}
303: \end{eqnarray}
304: %
305: where we have used the definitions for the points: $N = (u + \Delta, v
306: + \Delta)$, $W = (u + \Delta, v)$, $E = (u, v + \Delta)$ and $S =
307: (u,v)$.
308: Another possible scheme is
309: %
310: \begin{gather}
311: \left[ 1 - \frac{\Delta^{2}}{16}V(S)\right] R_{\ell}(N) = 
312: R_{\ell}(E)+ R_{\ell}(W) - R_{\ell}(S)\nonumber \\
313: -\frac{\Delta^{2}}{16}
314:        \left[V(S)R_{\ell}(S)+V(E)R_{\ell}(E)+V(W) R_{\ell}(W)\right] \,\,.   
315: \label{discrete2}
316: \end{gather}   
317: %
318: Although the second discretization (\ref{discrete2}) is more time consuming
319: than (\ref{discrete1}), it was observed in \cite{Wang-04}  that
320: (\ref{discrete2}) is more stable for fields in asymptotically AdS
321: geometries. With the use of expression  (\ref{discrete1}) or
322: (\ref{discrete2}), the basic algorithm will cover the  region of
323: interest in the $u-v$ plane, using the value of the field at three
324: points  in order to calculate  it at a forth one.  After the
325: integration is completed, the values of $R_{\ell}$ in the regions of
326: interest are extracted.   
327: 
328: 
329: \subsection{WKB analysis}
330: 
331: Considering the Laplace transformation of the
332: Eq. (\ref{simpledalembertian}) (in terms of $t$ and $r_{\star}$), one
333: gets the ordinary differential equation     
334: %
335: \begin{equation}
336: \frac{d^2 \psi_{\ell} (r_\star)}{d r_\star^2} - \left[s^2 + V(r(r_\star))
337:   \right]\psi_{\ell}(r_\star)=0 \quad .  
338: \end{equation}
339: %
340: One finds that there is a discrete set of possible values of $s$
341: such that the function $\psi_{\ell}(r_\star)$ satisfies both
342: boundary conditions:
343: %
344: \begin{equation}
345: \lim_{r_{\star} \rightarrow -\infty}\psi_{\ell} \, e^{s r_{\star}}=1 \ ,
346: \end{equation}
347: %
348: \begin{equation}
349: \lim_{r_{\star}\rightarrow +\infty}\psi_{\ell} \, e^{-s r_{\star}}=1 \ .
350: \end{equation}
351: %
352: By making the formal replacement $s=i\omega$, we have the usual quasinormal
353: mode boundary conditions. The frequencies $\omega$ (or $s$) are
354: the quasinormal frequencies.
355: 
356: The semi-analytic approach used in this work \cite{yer}
357: is a very efficient algorithm to calculate the quasinormal
358: frequencies, which have been applied in a variety of situations
359: \cite{WKB-ap}. 
360: 
361: 
362: Under the choice of the positive sign of the real part of
363: $\omega$, QNMs of Gauss-Bonnet and Gauss-Bonnet-de Sitter black holes
364: satisfy the following boundary conditions 
365: %
366: \begin{equation}
367: \psi (r_{\star}) \approx C_\pm \exp( \pm i \omega r_{\star}) \quad
368: \textrm{as} \quad r_{\star} \rightarrow \pm \infty \quad ,
369: \end{equation}
370: %
371: corresponding to purely in-going waves at the event horizon and
372: purely out-going waves at null infinity (or cosmological horizon, if
373: $\Lambda>0$). For the Gauss-Bonnet-anti-de Sitter geometries, the
374: effective potential is divergent at spatial infinity (which
375: corresponds to a finite value of $r_{\star}$, here taken as 0). In the
376: present work, we assume Dirichlet boundary conditions, setting
377: $\psi_{\ell} (r_\star = 0)=0$.   
378: 
379: To find the quasinormal modes of the black hole whose  effective
380: potential has the form of a potential barrier (GB and GBdS black holes)
381: one can use a high order WKB approach, finding   
382: %
383: \begin{equation}
384: i \frac{\omega^2 - V_0}{\sqrt{-2V_0^{\prime\prime}}} - L_2 - L_3 - L_4 -
385: L_5 - L_6 = n + \frac{1}{2} \quad ,
386: \end{equation}
387: %
388: where $V_0$ is the height and $V_0^{\prime\prime}$ is the second
389: derivative with respect to the tortoise coordinate of the potential at
390: the maximum. $L_2$, $L_3$  $L_4$, $L_5$ and $L_6$ are presented in
391: \cite{yer}. Thus we are able to use this formula for finding the
392: quasinormal modes of Gauss-Bonnet and Gauss-Bonnet de Sitter black
393: holes. Yet, for Gauss-Bonnet anti-de Sitter it cannot be applied as
394: the corresponding potential is divergent at spatial infinity.   
395: 
396: 
397: 
398: Accuracy of WKB approach may be bad for some cases of higher
399: dimensional black holes. 
400: We think that it is mainly not because of second  small peak in higher
401: dimensional case \cite{konoplya03}, \cite{A}: the WKB inaccuracy is
402: limited by the case $\ell=n$ or $\ell<n$. To judge about accuracy of
403: WKB method one has to compare the WKB results with results obtained by
404: an accurate Frobenius procedure. This was done for a $d$-dimensional Schwarzschild
405: black hole in a paper \cite{B}, where it was shown that for low
406: overtones ($\ell>n$) the difference between 6th order WKB  and
407: Frobenius method results  is less then one per-cent. We believe this
408: signifies the relialability of WKB formulas for $\ell>n$ modes, even for
409: higher dimensional black holes. After all, for $l>0$ modes, and for
410: scalar field perturbations considered in this paper, there is no
411: negative pitch in the potential.        
412: 
413: 
414: 
415: 
416: 
417: 
418: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
419: \section{Evolution of perturbations: time and frequency domain}
420: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
421: 
422: In this section we shall discuss the quasinormal and late-time
423: behavior for scalar field perturbations in the exterior of
424: Gauss-Bonnet black holes, generally, with a null, positive and
425: negative $\Lambda$-term, and therefore one has to consider the
426: correlation of the scalar field evolution with ``global''  parameters:
427: GB-coupling $\alpha$, $\Lambda$-term, spacetime dimensionality, and
428: ``local'' parameters such as black hole mass $\mu$ and  multipole
429: number $\ell$.  
430:    
431: 
432: 
433: 
434: 
435: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
436: \begin{figure}
437: %
438: \resizebox{1\linewidth}{!}{\includegraphics*{fig1.eps}}
439: %
440: \caption{Field decay in the Gauss-Bonnet
441:     black holes, for \mbox{$d=5$} and $d=7$. It is observed a quasinormal mode
442:     dominated region. Asymptotically, the field decays as a power-law
443:     tail (dashed lines).  The parameters in this graph are $\alpha=1$,
444:     $\mu=1.0$ and $\ell=0$.}
445: %
446: \label{fig_GB1}
447: %
448: \end{figure}
449: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
450: 
451: 
452: 
453: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
454: \begin{figure}
455: %
456: \resizebox{1\linewidth}{!}{\includegraphics*{fig2.eps}}
457: %
458: \caption{Power-law tails in the Gauss-Bonnet
459:     black holes. The estimated power-law
460:     are $R_{\ell} \propto t^{-5.076}$,  $R_{\ell} \propto t^{-5.082}$ and
461:     $R_{\ell} \propto t^{-5.068}$ for $\alpha = 0$,  $\alpha = 5$ and
462:     $\alpha = 10$ respectively. The predicted power for $\alpha=0$ is
463:     $-5$. The parameters in this graph are $d=7$, $\mu=1.0$ and
464:     $\ell=0$.}   
465: %
466: \label{fig_GB2}
467: %
468: \end{figure}
469: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
470: 
471: 
472: 
473: \subsection{Gauss-Bonnet black holes}
474: 
475: As seen in a previous work \cite{Konoplya05}, the WKB method allows
476: very accurate calculations of the quasinormal modes associated with
477: the field evolution. A complementary analysis can be performed within
478: the time-dependent picture. For this purpose, we use here a
479: characteristic initial value algorithm.   
480: 
481: The scenario presented by the WKB calculations is consistent with the
482: results obtained with time evolution approach. From the wave-functions
483: calculated with  the characteristic integration routine, it is
484: observed that, after an initial transient phase, the decay is
485: dominated by the quasinormal mode ringing. It is possible
486: to estimate with high precision the oscillatory and exponential decay
487: parameters using a non-linear fitting based in a $\chi^2$
488: analysis. We emphasize that the numerical concordance is excellent, as
489: seen in Tables \ref{GB_QNM_1} - \ref{GB_QNM_4}. 
490: The results, as compared between WKB approximation and characteristic 
491: integration agree to an accuracy within a few per-cents for $l>n$ case. 
492: This small difference must exist, because we compare the data for the
493: fundamental overtones in frequency domain with time domain data 
494: where the contribution from all overtones is taken into consideration. 
495: Unfortunately the WKB accuracy for $l=n=0$ is not satisfactory what
496: results in large difference between frequency and time domain data for
497: that case.  
498: 
499: Strictly speaking, the WKB technique we used here converges only
500: asymptotically. Practically, WKB formula shows good convergence within
501: several few orders after eikonal approximation. Yet,  
502: the worse convergence of the WKB method takes place when we deal with 
503: the intermediate values of $\alpha \approx 1$. That is why, in this
504: regime, the agreement between the WKB and the characteristic
505: integration results is the worst. 
506: 
507: 
508: 
509: The imaginary part of the frequency does not
510: show too much dependence on $\alpha$, yet slightly decrease when
511: $\alpha$ is increasing. On the other hand, the real part
512: increases with $\alpha$, though not significantly either. This might be
513: showing that a quasinormal mode is much more an effect connected with the
514: local geometry containing the black hole rather than with the global
515: effect of the geometry, namely, the effect of the existence of an event
516: horizon matters much more than a detailed dependence on the parameter
517: $\alpha$. Yet, large enough values of $\alpha$ certainly affect the
518: quasinormal spectrum: the QNMs are proportional to $\alpha$ in the
519: regime of large $\alpha$ \cite{Konoplya05}. As $\alpha$ approaches
520: zero, the QNMs go to those of ordinary $d$-dimensional Schwarzschild 
521: black hole described by the Tangherlini metric.
522: 
523: However, it does not mean that GB corrections are negligible.
524: On the contrary, according to  string theory 
525: the Gauss-Bonnet coupling $\alpha$ should be around $1$. Let us compare 
526: the results, for instance, for $\ell=2$, $n=0$ QNMs for Schwarzschild and
527: $\alpha = 1$ GB black holes for $d=6$: for Schwarzschild we have 
528: $\omega = 1.5965 - 0.3967 i$ (6th order WKB),  for $\alpha =1$ (6th
529: order WKB) we get $\omega = 1.69654-0.31929 i$ (Note that for this
530: case convergence is good and the 3th order WKB value is not much
531: different  $1.67624 - 0.323698 i$). Thus the effect of GB coupling is
532: about $6.3\%$ in real and more then $20\%$ in the imaginary
533: part here. For larger values of $\alpha$ it is certainly larger. We
534: have the same order of difference for other values of $n$ and $\ell$.  
535:   
536: 
537: 
538: 
539:    
540: \begin{table}
541: \caption{Values for the quasinormal frequencies for the fundamental
542:   mode in the Gauss-Bonnet geometry, obtained from sixth order WKB
543:   method and directly from characteristic data, for
544:   $d = 5$, $\ell=0,1$ and several values of $\alpha$.} 
545: %
546: \label{GB_QNM_1} 
547: %
548: \begin{ruledtabular}
549: \begin{tabular}{cccccc}
550: %
551: \multicolumn{2}{l}{ d = 5 }         &
552: \multicolumn{2}{c}{WKB}&
553: \multicolumn{2}{c}{Characteristic Integration}\\ 
554: %
555: $\ell$ & $\alpha$ & Re($\omega_0$) & -Im($\omega_0$)  &
556: Re($\omega_0$) & -Im($\omega_0$)
557: \\
558: \hline
559: \\
560: %
561: %
562: 0 & 0.1 & 0.389935 & 0.256159 &  0.379 & 0.282 \\
563: %
564: 0 & 0.2 &  0.396034 & 0.250548 &  0.383 & 0.272  \\ 
565: % 
566: 0 & 0.5 &  0.429741 & 0.208293 &  0.391 & 0.246 \\ \\
567: % 
568: %
569: %
570: 1 & 0.1 &  0.720423 & 0.255506 &  0.7234 & 0.250  \\
571: %
572: 1 & 0.2 &  0.723177 & 0.252684 &  0.7284 & 0.245  \\
573: % 
574: 1 & 0.5 &  0.739205 & 0.236885 &  0.7443 & 0.228 \\
575: %
576: \end{tabular} 
577: \end{ruledtabular}
578: \end{table}
579: 
580: \begin{table}
581: \caption{Values for the quasinormal frequencies for the fundamental
582:   mode in the Gauss-Bonnet geometry, obtained from sixth order WKB
583:   method and directly from characteristic data, for
584:   $d = 6$, $\ell=0,1$ and several values of $\alpha$.} 
585: %
586: \label{GB_QNM_2}
587: %
588: \begin{ruledtabular}
589: \begin{tabular}{cccccc}
590: %
591: \multicolumn{2}{l}{ d = 6 }         &
592: \multicolumn{2}{c}{WKB}&
593: \multicolumn{2}{c}{Characteristic Integration}\\ 
594: %
595: $\ell$ & $\alpha$ & Re($\omega_0$) & -Im($\omega_0$)  &
596: Re($\omega_0$) & -Im($\omega_0$)
597: \\
598: \hline
599: \\
600: %
601: %
602: 0 & 0.1 & 0.735854 & 0.402416 & 0.7109 & 0.412  \\
603: %
604: 0 & 0.2 & 0.748053 & 0.391049 & 0.7144 & 0.402 \\
605: % 
606: 0 & 0.5 & 0.83530  & 0.304837 & 0.7225 &  0.375  \\
607: % 
608: 0 & 5   & 0.906661 & 0.144820 & 0.9526 & 0.226 \\
609: % 
610: 0 & 10  & 1.513624 & 0.456935 & 1.5182 & 0.423 \\
611: % 
612: 0 & 20  & 2.961675 & 0.927128 & 2.878  & 0.952 \\ \\
613: % 
614: %
615: %
616: 1 & 0.1 & 1.139007 & 0.415034 & 1.153  & 0.395  \\
617: %
618: 1 & 0.2 & 1.136193 & 0.415706 & 1.159  & 0.386  \\
619: % 
620: 1 & 0.5 & 1.158635 & 0.391339 & 1.176  & 0.363  \\
621: % 
622: 1 & 5   & 1.790103 & 0.258382 & 1.791  & 0.253  \\
623: % 
624: 1 & 10  & 3.260415 & 0.498426 & 3.253  & 0.504  \\
625: % 
626: 1 & 20  & 6.437139 & 0.991374 &  -     &   -    \\
627: %
628: \end{tabular} 
629: \end{ruledtabular}
630: \end{table}
631: 
632: 
633: 
634: 
635: 
636: 
637: 
638: \begin{table}
639: \caption{Values for the quasinormal frequencies for the fundamental
640:   mode in the Gauss-Bonnet geometry, obtained from sixth order WKB
641:   method and directly from characteristic data, for
642:   $d = 7$, $\ell=0,1$ and several values of $\alpha$.} 
643: %
644: \label{GB_QNM_3}
645: %
646: \begin{ruledtabular}
647: \begin{tabular}{cccccc}
648: %
649: \multicolumn{2}{l}{ d = 7 }         &
650: \multicolumn{2}{c}{WKB}&
651: \multicolumn{2}{c}{Characteristic Integration}\\ 
652: %
653: $\ell$ & $\alpha$ & Re($\omega_0$) & -Im($\omega_0$)  &
654: Re($\omega_0$) & -Im($\omega_0$)
655: \\
656: \hline
657: \\
658: %
659: %
660: 0 & 0.1 & 1.11738 & 0.546056 & 1.092 & 0.532 \\
661: %
662: 0 & 0.2 & 1.13699 & 0.529543 & 1.092 & 0.520 \\
663: % 
664: 0 & 0.5 & 1.29469 & 0.395111 & 1.092 &  0.493 \\
665: % 
666: 0 & 5   & 1.39823 & 0.574472 & 1.275 & 0.351 \\
667: % 
668: 0 & 10  & 1.48053 & 0.385616 & 1.515 & 0.358 \\
669: % 
670: 0 & 20  & 2.00368 & 0.56458  & 2.001 & 0.567 \\ \\
671: % 
672: %
673: %
674: 1 & 0.1 & 1.54573 & 0.577608 & 1.587 & 0.527 \\
675: %
676: 1 & 0.2 & 1.53194 & 0.58701 & 1.530 & 0.517 \\
677: % 
678: 1 & 0.5 & 1.56277 & 0.554551 & 1.609 &  0.489 \\
679: % 
680: 1 & 5   & 2.01379 & 0.308316 & 1.982 &  0.337 \\
681: % 
682: 1 & 10  & 2.47824 & 0.243737 & 2.475 &  0.423 \\
683: % 
684: 1 & 20  & 3.39094 & 0.59452 & 3.387 &  0.597 \\
685: \end{tabular}
686: \end{ruledtabular}
687: \end{table}
688: 
689: 
690: 
691: 
692: \begin{table}
693: \caption{Values for the quasinormal frequencies for the fundamental
694:   mode in the Gauss-Bonnet geometry, obtained from sixth order WKB
695:   method and directly from characteristic data, for
696:   $d = 8$, $\ell=0,1$ and several values of $\alpha$.} 
697: %
698: \label{GB_QNM_4}
699: %
700: \begin{ruledtabular}
701: \begin{tabular}{cccccc}
702: %
703: \multicolumn{2}{l}{ d = 8 }         &
704: \multicolumn{2}{c}{WKB}&
705: \multicolumn{2}{c}{Characteristic Integration}\\ 
706: %
707: $\ell$ & $\alpha$ & Re($\omega_0$) & -Im($\omega_0$)  &
708: Re($\omega_0$) & -Im($\omega_0$)
709: \\
710: \hline
711: \\
712: %
713: %
714: 0 & 0.1 & 1.51702 & 0.694245 & 1.461 & 0.676 \\
715: %
716: 0 & 0.2 & 1.54463 & 0.673566 & 1.463 & 0.658 \\
717: % 
718: 0 & 0.5 & 1.7854  & 0.488086 & 1.469 & 0.616  \\
719: % 
720: 0 & 5   & 1.47941 & 0.868859 & 1.647 & 0.420 \\
721: % 
722: 0 & 10  & 1.800   & 0.410252 & 1.838 & 0.421 \\
723: % 
724: 0 & 20  & 2.15426 & 0.556954 & 2.154 & 0.544 \\ \\
725: % 
726: %
727: %
728: 1 & 0.1 & 1.93407 & 0.750046 & 2.021 & 0.652  \\
729: %
730: 1 & 0.2 & 1.90391 & 0.773947 & 2.024 &  0.637 \\
731: % 
732: 1 & 0.5 & 1.94463 & 0.734591 & 2.035 &  0.602\\
733: % 
734: 1 & 5   & 2.44511 & 0.455166 & 2.348 & 0.434 \\
735: % 
736: 1 & 10  & 2.6644 & 0.463966 &  2.670 & 0.458 \\
737: % 
738: 1 & 20  & 3.2276 & 0.581024 & 3.209 & 0.590 \\ 
739: \end{tabular} 
740: \end{ruledtabular}
741: \end{table}
742: 
743: 
744: In the time domain the signal has three stages: the initial pulse
745: dependent on the source of perturbations, the quasinormal ringing
746: dominating period, and the power-law tail (see Fig.\ref{fig_GB1}).
747: The bigger GB-coupling is, the larger the quasinormal dominated region, 
748: i.e. at later times the tails start dominating. 
749: As can be seen from Fig.\ref{fig_GB2}, the power-law
750: tails do not show dependence on the Gauss-Bonnet coupling
751: $\alpha$ and are the same as for the d-dimensional Schwarzschild 
752: black hole in Einstein general relativity, when $d$ is odd. 
753: That is, the fields always shows a power-law falloff: 
754: for odd $d>3$ the field behaves as 
755: %
756: \begin{equation}
757: R_{\ell} \propto t^{-(2\ell + d - 2)} 
758: \end{equation}
759: %
760: at late times, where  $\ell$ is the multipole number. This behavior is
761: entirely due to $d$ being odd and  does not depend on the presence of
762: a black hole \cite{Vitor-Tail}. It is known, that in Einstein gravity,
763: for even $d>4$, the field decays as  \cite{Vitor-Tail}  
764: %
765: \begin{equation}
766: R_{\ell} \propto t^{-(2\ell + 3d - 8)} \quad ,
767: \end{equation}
768: %
769: and for the latter case there is no contribution from the flat background. 
770: This power-law tail is entirely due to the presence of the black hole
771: \cite{Vitor-Tail}. At the same time, the Gauss-Bonnet black hole
772: metric (\ref{metric_form}-\ref{metric}) goes to pure Minkowskian
773: metric when the black hole mass equals zero, i.e. in a space-time
774: without a black hole. In other words, empty space-time in Gauss-Bonnet
775: gravity ``does not see'' the $\alpha$. That is why we do not observe
776: the $\alpha$-dependence of tails in odd space-time dimensions.  
777: 
778: Thus, if the Gauss-Bonnet term changes late-time behavior, it must show itself
779: only for even-dimensional space-time. In the numerical procedure
780: developed with the characteristic integration scheme, no
781: tails (power-law or otherwise) were observed in even dimensions
782: Gauss-Bonnet spherical black holes. Yet, it should be pointed out that the
783: integration of the scalar field equation in the GB background is a
784: much more demanding numerical problem than the same integration with
785: the usual Einstein coupling. In the latter case there are auxiliary
786: analytical results, such as explicit expression for the tortoise
787: coordinate function. Therefore, the GB codes are less precise and more
788: time consuming, and eventual tails could be hidden. The possible
789: absence of tails with even $d$ deserves further consideration.     
790: 
791: 
792: 
793: 
794: 
795: 
796: \subsection{Gauss-Bonnet-de Sitter black holes}
797: 
798: 
799: For the GBdS black holes the quasinormal ringing stage becomes correlated
800: with a new parameter: a positive $\Lambda$-term. When the $\Lambda$-term
801: is growing, both real oscillation frequency and the damping rate are
802: decreasing. Yet, real part of $\omega$ is more sensitive to the changes
803: of  $\Lambda$-term. 
804: 
805: Qualitatively this resembles the quasinormal
806: oscillations of $d$-dimensional  Schwarzschild-de Sitter black hole
807: \cite{konoplya03}. In the limit of extremal value of the
808: $\Lambda$-term, i.e. when the cosmological horizon ($r_{c}$) is very
809: close to the event horizon ($r_{+}$), it is possible to generalize the
810: formulas found in \cite{Cardoso-03} for four dimensional black hole
811: and in \cite{Molina} for $d$-dimensional case.  Namely, the
812: quasinormal frequencies for the near extreme Gauss-Bonnet
813: asymptotically de Sitter black holes are given by  
814: %
815: \begin{gather}
816: \frac{\omega_{n}}{\kappa_{+}} =  \sqrt{ \frac{\ell(\ell+d-3)}{r_{+}^{2}}
817:   \frac{r_{c} - r_{+}}{2\kappa_{+}} - \frac{1}{4}} -i
818:   \left(n+\frac{1}{2}\right)  \quad , \nonumber \\  
819: %
820: n = 0, 1, 2, \ldots \quad ,
821: %
822: \label{QNM_near_extreme}
823: \end{gather}
824: %
825: where $\kappa_{+}$ (a function of $\alpha$ in this generalized
826: context) is the surface gravity at the event horizon. The above
827: formula is well confirmed numerically: Fig.\ref{near_extreme} shows a
828: comparison of the values obtained by direct numerical calculation and
829: from Eq. (\ref{QNM_near_extreme}). Thus the quasinormal modes are
830: proportional to the surface gravity $\kappa_{+}$, at least for lower
831: overtones. It should be pointed that for the usual Schwarzschild-de
832: Sitter black holes, numerical and analytical investigations
833: \cite{SdShigh} suggest that the high overtone behavior does not obey
834: the formula (\ref{QNM_near_extreme}).       
835: 
836: 
837: 
838: 
839: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
840: \begin{figure}
841: %
842: \resizebox{1\linewidth}{!}{\includegraphics*{fig3.eps}}
843: %
844: \caption{Graph of $\textrm{Re} (\omega)^2$ as a function of $\ell
845:     (\ell + d - 3)$, in the near-extreme limit positive $\Lambda$
846:     limit. The bullets are the values calculated from the
847:     time-evolution profiles, and the dashed lines are values obtained
848:     from the expression (\ref{QNM_near_extreme}). The parameters in
849:     this graph are $\kappa_{+}=0.01$, $\alpha=1$, $\mu=1.0$, and the
850:     differences between the analytical and numerical results are under
851:     2\%.} 
852: %
853: \label{near_extreme}
854: %
855: \end{figure}
856: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
857: 
858:   
859: When the mass parameter $\mu$ is set to zero, we have the case of pure
860: de Sitter spacetime in the Gauss-Bonnet gravity. The metric function
861: (\ref{metric}) then reduces to the following form:
862: %
863: \begin{equation}
864: h(r) = 1 + \frac{r^2}{2\alpha} - \frac{r^2}{2\alpha} 
865:        \sqrt{1 + \frac{8\alpha \Lambda}{(d-1)(d-2)} } \quad .
866: \end{equation}
867: %
868: Repeating the analysis of \cite{Abdalla-02,natario}, we
869: come to the conclusion that quasinormal modes exist only in odd
870: spacetime dimensions and are given by the formula:
871: %
872: \begin{gather}
873: \omega_{n} = i\left[\frac{1} {2\alpha}\left(1 - \sqrt{1 + \frac{8\alpha
874: \Lambda}{(d-1)(d-2)}}\right)\right]^{1/2} (2 n + \ell) \nonumber \\
875: %
876: n = 0, 1, 2, \ldots \quad .
877: \end{gather}
878: %
879: Note that pure Gauss-Bonnet-de Sitter quasinormal modes are purely
880: imaginary, which corresponds to exponential decaying without
881: oscillations.   
882: 
883: 
884: It is well-known  that the late-time tails of black holes in
885: asymptotically de Sitter space-time for
886: zero multipole and for higher multipoles are qualitatively different.
887: For the zero multipole field ($\ell=0$), the time domain picture is
888: the following: after a transient part, a
889: quasinormal mode dominated region is best observed. Following the
890: quasinormal mode dominated region, a late-time decay region settles.
891: In this latter phase, the  wave-functions decay asymptotically to a
892: constant value, as has been the case in the Schwarzschild de Sitter
893: black hole which was studied before
894: \cite{Brady-97,Brady-99,Molina-04}. 
895: This is illustrated in Fig.\ref{fig_GBdS1}.  
896: 
897: For first and higher multipoles ($\ell>0$) at late times we observe
898: exponential tails in vicinity of GBdS black hole. This is also an
899: expected result, since in Einstein gravity the exponential tails are
900: observed as well in usual de Sitter black holes
901: \cite{Brady-97,Brady-99,Molina-04}.  This is illustrated in
902: Fig.\ref{fig_GBdS2}. The dependence of the quasinormal modes 
903: on $\Lambda$-term and Gauss-Bonnet coupling can be learnt 
904: from Tables X-XII for different space-time dimensionality.
905:    
906: 
907: 
908: 
909: 
910: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
911: \begin{figure}
912: %
913: \resizebox{1\linewidth}{!}{\includegraphics*{fig4.eps}}
914: %
915: \caption{Field decay in the Gauss-Bonnet-de Sitter
916:     black holes, with $\ell=0$.  A quasinormal mode
917:     dominated region is observed (above), and asymptotically the field
918:     decays to a constant (below). The fundamental mode, calculated with the WKB
919:     method and directly from the characteristic data are $0.8356 - 0.2935i$ and
920:     $0.8011 - 0.2608i$. The parameters in this graph are $d=6$, $\alpha=3.0$,
921:     $\Lambda=0.1$ and $\mu=1.0$.}
922: %
923: \label{fig_GBdS1}
924: %
925: \end{figure}
926: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
927: 
928: 
929: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
930: \begin{figure}
931: %
932: \resizebox{1\linewidth}{!}{\includegraphics*{fig5.eps}}
933: %
934: \caption{Field decay in the Gauss-Bonnet-de Sitter 
935:     black holes, for $d=5,6,7,8$. It is observed a quasinormal mode
936:     dominated region. Asymptotically, the field decays as an
937:     exponential tail.  The parameters in this graph are $\alpha=0.1$,
938:     $\mu=1.0$ and $\ell=1$.}
939: %
940: \label{fig_GBdS2}
941: %
942: \end{figure}
943: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
944: 
945: 
946: The numerical simulations developed for Gauss-Bonnet-de Sitter
947: black hole indicate that the massless scalar perturbation in
948: this geometry behave asymptotically as 
949: %
950: \begin{eqnarray}
951: R_{\ell}\approx \exp \left[ \ell \left(\kappa_{c} + c^{quad}
952: \kappa_{c}^{2}\right) t \right] & \textrm{as} & t\rightarrow\infty
953: \quad , \nonumber \\ 
954: %
955: d \ge 4  \quad \textrm{and} \quad \alpha \ge 0  \quad ,
956: %
957: \label{tails-GBdS}
958: \end{eqnarray}
959: %
960: where $\kappa_{c}$ is the surface gravity at the cosmological
961: horizon and $c^{quad}$ is an adjustment parameter for the
962: $\kappa_{c}^{2}$ correction. The expression (\ref{tails-GBdS}) shows
963: that, although the exponential tail in the GBdS background is
964: dependent on the Gauss-Bonnet coupling $\alpha$ (since $\kappa_{c}$ is
965: a function of $\alpha$), the form of the dependence is identical to
966: the null $\alpha$ case. Eq. (\ref{tails-GBdS}) generalizes the
967: analogous expression found in \cite{Molina-04} for the usual
968: Schwarzschild black holes.  
969: 
970: 
971: 
972: 
973: \begin{table}
974: \caption{Values for the quasinormal frequencies for the fundamental
975:   mode in the Gauss-Bonnet-de Sitter geometry, obtained from third and
976:   sixth order WKB method, for
977:   $d = 5$, $\ell=0$ and several values of $\alpha$ and $\Lambda$.} 
978: %
979: \label{GB-dS-L0-1} 
980: %
981: \begin{ruledtabular}
982: \begin{tabular}{cccccc}
983: %
984: \multicolumn{2}{l}{ d = 5}         &
985: \multicolumn{2}{c}{WKB (3th order)}&
986: \multicolumn{2}{c}{WKB (6th order)}\\ 
987: %
988: $\alpha$ & $\Lambda$ & Re($\omega_0$) & -Im($\omega_0$)  &
989: Re($\omega_0$) & -Im($\omega_0$)
990: \\
991: \hline
992: \\
993: %
994: %
995: 0.1 & 1/8 & 0.30485   & 0.278407 & 0.334981   & 0.26202 \\
996: %
997: 0.1 & 1/2 & 0.14922   & 0.217542 & 0.152444   & 0.2164  \\ 
998: % 
999: 0.1 & 2/3 & 0.0677006 & 0.14792  & 0.00666957 & 0.150988  \\ \\
1000: %
1001: %
1002: 1   & 1/5 & 0.301869  & 0.218752  & 0.378865  & 0.159611 \\
1003: %
1004: 1   & 1   & 0.112613  & 0.148749  & 0.11619   & 0.145465 \\ 
1005: % 
1006: 1   & 7/5 & 0.0240051 & 0.0683652 & 0.0242716 & 0.0709775  \\
1007: % 
1008: %
1009: \end{tabular} 
1010: \end{ruledtabular}
1011: \end{table}
1012: 
1013: 
1014: \begin{table}
1015: \caption{Values for the quasinormal frequencies for the fundamental
1016:   mode in the Gauss-Bonnet-de Sitter geometry, obtained from third and
1017:   sixth order WKB method, for $d = 6$, $\ell=0$ and several values of
1018:   $\alpha$ and $\Lambda$.}  
1019: %
1020: \label{GB-dS-L0-2} 
1021: %
1022: \begin{ruledtabular}
1023: \begin{tabular}{cccccc}
1024: %
1025: \multicolumn{2}{l}{ d = 6}         &
1026: \multicolumn{2}{c}{WKB (3th order)}&
1027: \multicolumn{2}{c}{WKB (6th order)}\\ 
1028: %
1029: $\alpha$ & $\Lambda$ & Re($\omega_0$) & -Im($\omega_0$)  &
1030: Re($\omega_0$) & -Im($\omega_0$)
1031: \\
1032: \hline
1033: \\
1034: %
1035: %
1036: 0.1 & 1/4 & 0.580472   & 0.428367  & 0.652079  & 0.411337 \\
1037: %
1038: 0.1 & 1   & 0.379931   & 0.392027  & 0.401925  & 0.384688 \\ 
1039: % 
1040: 0.1 & 2   & 0.0653851  & 0.168544  & 0.0636328 & 0.173794  \\ \\
1041: % 
1042: %
1043: 1   & 1/2 & 0.582293   & 0.358225  & 0.736989  & 0.235326 \\
1044: %
1045: 1   & 2   & 0.318697   & 0.307084  & 0.331268  & 0.282461  \\
1046: %
1047: 1   & 4   & 0.00390515 & 0.0487306 & 0.0574555 & 0.0529852 \\ \\
1048: % 
1049: %
1050: 10  & 100 & 1.5033     & 0.574511  & 1.5741    & 0.599309 \\
1051: %
1052: \end{tabular} 
1053: \end{ruledtabular}
1054: \end{table}
1055: 
1056: 
1057: \begin{table}
1058: \caption{Values for the quasinormal frequencies for the fundamental
1059:   mode in the Gauss-Bonnet-de Sitter geometry, obtained from third and
1060:   sixth order WKB method, for $d = 7$, $\ell=0$ and several values of
1061:   $\alpha$ and $\Lambda$.}  
1062: %
1063: \label{GB-dS-L0-3} 
1064: %
1065: \begin{ruledtabular}
1066: \begin{tabular}{cccccc}
1067: %
1068: \multicolumn{2}{l}{ d = 7}         &
1069: \multicolumn{2}{c}{WKB (3th order)}&
1070: \multicolumn{2}{c}{WKB (6th order)}\\ 
1071: %
1072: $\alpha$ & $\Lambda$ & Re($\omega_0$) & -Im($\omega_0$)  &
1073: Re($\omega_0$) & -Im($\omega_0$)
1074: \\
1075: \hline
1076: \\
1077: %
1078: %
1079: 0.1 & 1 & 0.769109   & 0.556647 & 0.852395  & 0.551612 \\
1080: %
1081: 0.1 & 1 & 0.561644   & 0.517639 & 0.598484  & 0.511069 \\ 
1082: % 
1083: 0.1 & 1 & 0.00963007 & 0.222564 & 0.0939457 & 0.228191  \\ \\
1084: % 
1085: %
1086: 1   & 1 & 0.868246   & 0.48557  & 1.12603   & 0.2999562 \\
1087: %
1088: 1   & 4 & 0.461426   & 0.419016 & 0.483098  & 0.381285  \\
1089: %
1090: 1   & 7 & 0.113334   & 0.215329 & 0.111663  & 0.218394 \\
1091: % 
1092: \end{tabular} 
1093: \end{ruledtabular}
1094: \end{table}
1095: 
1096: 
1097: \begin{table}
1098: \caption{Values for the quasinormal frequencies for the fundamental
1099:   mode in the Gauss-Bonnet-de Sitter geometry, obtained from third and
1100:   sixth order WKB method, for $d = 8$, $\ell=0$ and several values of
1101:   $\alpha$ and $\Lambda$.}  
1102: %
1103: \label{GB-dS-L0-4} 
1104: %
1105: \begin{ruledtabular}
1106: \begin{tabular}{cccccc}
1107: %
1108: \multicolumn{2}{l}{ d = 8}         &
1109: \multicolumn{2}{c}{WKB (3th order)}&
1110: \multicolumn{2}{c}{WKB (6th order)}\\ 
1111: %
1112: $\alpha$ & $\Lambda$ & Re($\omega_0$) & -Im($\omega_0$)  &
1113: Re($\omega_0$) & -Im($\omega_0$)
1114: \\
1115: \hline
1116: \\
1117: %
1118: %
1119: 0.1 & 1 & 1.1449   & 0.669364 & 1.29739  & 0.695376 \\
1120: %
1121: 0.1 & 4 & 0.644892 & 0.598472 & 0.682218 & 0.594199 \\
1122: % 
1123: \end{tabular} 
1124: \end{ruledtabular}
1125: \end{table}
1126: 
1127: 
1128: \begin{table}
1129: \caption{Values for the quasinormal frequencies for the fundamental
1130:   mode in the Gauss-Bonnet-de Sitter geometry, obtained from third and
1131:   sixth order WKB method and directly from characteristic data, for
1132:   $d = 5$, $\ell=1$ and several values of $\alpha$ and $\Lambda$.} 
1133: %
1134: \label{GBdS_L1_1}
1135: %
1136: \begin{ruledtabular}
1137: \begin{tabular}{cccccc}
1138: %
1139: \multicolumn{2}{l}{ d = 5 }        &
1140: \multicolumn{2}{c}{WKB (6th order)}&
1141: \multicolumn{2}{c}{Characteristic Integration}\\
1142: %
1143:   $\alpha$ & $\Lambda$ & 
1144:   Re($\omega_0$) & -Im($\omega_0$) & 
1145:   Re($\omega_0$) & -Im($\omega_0$) \\ 
1146: %
1147: \hline \\
1148: %
1149: %
1150: 0.1 & 1/8  & 0.64317  & 0.241723 & 0.6451 & 0.2382 \\
1151: %
1152: 0.1 & 1/2  & 0.379265 & 0.165101 & 0.3856 & 0.1564 \\ 
1153: % 
1154: 0.1 & 2/3  & 0.231024 & 0.103394 & 0.2247 & 0.1051 \\ \\
1155: %
1156: %
1157: 1 & 1/5    & 0.698655 & 0.181854  & 0.6926 & 0.1877  \\
1158: %
1159: 1 & 1     & 0.365998 & 0.111668  & 0.3704 & 0.1133 \\ 
1160: % 
1161: 1 & 7/5  & 0.153506 & 0.0465747 & 0.1518 & -0.042631 \\ 
1162: %
1163: \end{tabular} 
1164: \end{ruledtabular}
1165: \end{table}
1166: 
1167: 
1168: 
1169: \begin{table}
1170: \caption{Values for the quasinormal frequencies for the fundamental
1171:   mode in the Gauss-Bonnet-de Sitter geometry, obtained from third and
1172:   sixth order WKB method and directly from characteristic data, for
1173:   $d = 6$, $\ell=1$ and several values of  $\alpha$ and $\Lambda$.} 
1174: %
1175: \label{GBdS_L1_2}
1176: %
1177: \begin{ruledtabular}
1178: \begin{tabular}{cccccc}
1179: %
1180: \multicolumn{2}{l}{ d = 6 }        &
1181: \multicolumn{2}{c}{WKB (6th order)}&
1182: \multicolumn{2}{c}{Characteristic Integration}\\
1183: %
1184:   $\alpha$ & $\Lambda$ & 
1185:   Re($\omega_0$) & -Im($\omega_0$) & 
1186:   Re($\omega_0$) & -Im($\omega_0$) \\
1187: %
1188: \hline \\
1189: %
1190: %
1191: 0.1 & 1/4  & 1.04698  & 0.39859  & 1.0574 & 0.3828  \\
1192: %
1193: 0.1 & 1    & 0.74644  & 0.323514 & 0.7479 & 0.3183  \\ 
1194: % 
1195: 0.1 & 2    & 0.238066 & 0.114496 & 0.2387 & 0.1212  \\ \\
1196: %
1197: %
1198: 1 & 1/2   & 1.09756   & 0.301244  & 1.053   & 0.2948 \\
1199: %
1200: 1 & 2     & 0.6962    & 0.244326  & 0.6962  & 0.2307 \\ 
1201: % 
1202: 1 & 4 & 0.0847573 &0.0315107&0.08965&0.03323 \\ \\
1203: %
1204: %
1205: 10 & 100  & 3.21004 & 0.52559 & 3.1940 & 0.5360 \\
1206: %
1207: 10 & 2000         & -       & -       & 2.1711  & 0.4900 \\ 
1208: % 
1209: 10 & 5000         & -       & -       & 0.6228 & 0.2018 \\   
1210: \end{tabular} 
1211: \end{ruledtabular}
1212: \end{table}
1213: 
1214: 
1215: 
1216: 
1217: \begin{table}
1218: \caption{Values for the quasinormal frequencies for the fundamental
1219:   mode in the Gauss-Bonnet-de Sitter geometry, obtained from third and
1220:   sixth order WKB method and directly from characteristic data, for
1221:   $d = 7$, $\ell=1$, $\alpha=0.1$ and several values of $\Lambda$.} 
1222: %
1223: \label{GBdS_L1_3}
1224: %
1225: \begin{ruledtabular}
1226: \begin{tabular}{cccccc}
1227: %
1228: \multicolumn{2}{l}{ d = 7 }        &
1229: \multicolumn{2}{c}{WKB (6th order)}&
1230: \multicolumn{2}{c}{Characteristic Integration}\\
1231: %
1232:   $\alpha$ & $\Lambda$ & 
1233:   Re($\omega_0$) & -Im($\omega_0$) & 
1234:   Re($\omega_0$) & -Im($\omega_0$) \\
1235: %
1236: \hline \\
1237: %
1238: %
1239: 0.1 & 1  & 1.28621  & 0.519229 & 1.304  & 0.4910  \\
1240: %
1241: 0.1 & 2  & 1.00125  & 0.440876 & 1.008  & 0.4306 \\ 
1242: % 
1243: 0.1 & 4  & 0.303405 & 0.153651 & 0.3091 & 0.1416 \\ \\
1244: %
1245: %
1246: 1 & 1  & 1.48546  & 0.411346 & 1.467  & 0.3623 \\
1247: %
1248: 1 & 4  & 0.899964 & 0.338928 & 0.8944 & 0.3224 \\ 
1249: % 
1250: 1 & 7  & 0.361602 & 0.152214 & 0.3638 &0.1398 \\ 
1251: \end{tabular} 
1252: \end{ruledtabular}
1253: \end{table}
1254: 
1255: 
1256: 
1257: \begin{table}
1258: \caption{Values for the quasinormal frequencies for the fundamental
1259:   mode in the Gauss-Bonnet-de Sitter geometry, obtained from third and
1260:   sixth order WKB method and directly from characteristic data, for
1261:   $d = 8$, $\ell=1$, $\alpha=0.1$ and several values of $\Lambda$.} 
1262: %
1263: \label{GBdS_L1_4}
1264: %
1265: \begin{ruledtabular}
1266: \begin{tabular}{cccccc}
1267: %
1268: \multicolumn{2}{l}{ d = 8 }        &
1269: \multicolumn{2}{c}{WKB (6th order)}&
1270: \multicolumn{2}{c}{Characteristic Integration}\\
1271: %
1272:   $\alpha$ & $\Lambda$ & 
1273:   Re($\omega_0$) & -Im($\omega_0$) & 
1274:   Re($\omega_0$) & -Im($\omega_0$) \\
1275: %
1276: \hline \\
1277: %
1278: %
1279: 0.1 & 1 & 1.75036 & 0.697491 & 1.803 & 0.6250 \\
1280: %
1281: 0.1 & 4 & 1.10954 & 0.508301 & 1.103 & 0.4964 \\ 
1282: % 
1283: \end{tabular} 
1284: \end{ruledtabular}
1285: \end{table}
1286: 
1287: 
1288: 
1289: 
1290: 
1291: \subsection{Gauss-Bonnet-anti-de Sitter black holes}    
1292: 
1293: The quasinormal and late-time behavior of black holes 
1294: in anti-de Sitter spacetime is significantly different from those 
1295: in asymptotically de Sitter or flat spacetimes. The key difference is
1296: stipulated by the effective potential behavior, which is divergent at
1297: spacial infinity. Thus the anti-de Sitter space acts as an effective
1298: confining box. Therefore the Dirichlet boundary conditions are
1299: natural. These boundary conditions  are required also by AdS/CFT
1300: correspondence for scalar field perturbations \cite{Horowitz00}. Yet, for 
1301: higher spin perturbations the true boundary conditions may be
1302: different \cite{moss}.  
1303:  
1304: In the usual Schwarzschild-anti-de Sitter black holes,
1305: the quasinormal modes govern the decay at all times and thereby no power-law or
1306: exponential tails appear \cite{Wang-01,Wang-04}. We observe a similar
1307: behavior the scalar field perturbations in the Gauss-Bonnet-anti-de
1308: Sitter black holes.  
1309: 
1310: It is not possible to use WKB method to find the quasinormal modes in
1311: Gauss-Bonnet-AdS case because the effective potential is not a
1312: potential barrier anymore. The Horowitz-Hubeny method
1313: \cite{Horowitz00} is not applicable either, because the Taylor expansion
1314: of the effective potential has infinite number of terms. That is why
1315: we were limited only by time domain analysis, which is free of the
1316: above problems. From Fig.\ref{fig_GBAdS1} we see that, indeed, the
1317: quasinormal modes are dominating even at sufficiently late times. We
1318: also have learnt  from Fig.\ref{fig_GBAdS1}  that  the
1319: quasinormal mode dominated region grows, as the  multipole index
1320: $\ell$ grows.       
1321: 
1322: As is known from Einstein action case, as the radius of the AdS black
1323: hole goes to zero, the quasinormal modes of the black hole approach
1324: its pure anti-de Sitter values \cite{konoplya02}. Repeating the
1325: calculations of \cite{natario}, we find the exact
1326: expression for the normal modes in GB gravity: 
1327: %
1328: \begin{gather}
1329: \omega_{n} = \left[ \frac{1} {2\alpha}\left(1 - 
1330: \sqrt{1 + \frac{8\alpha \Lambda}{(d-1)(d-2)}}\right)\right]^{1/2} (2 n +
1331: \nonumber \\
1332: %
1333: \ell + d - 1) \quad ,
1334: \nonumber \\
1335: %
1336: n = 0, 1, 2, \ldots \quad .
1337: \end{gather}
1338: %
1339: The pure GB-AdS modes, unlike GB-dS modes, exist in any any spacetime
1340: dimension.
1341: 
1342: 
1343: 
1344: 
1345: 
1346: 
1347: \begin{table}
1348: \caption{Values for the quasinormal frequencies for the fundamental
1349:   mode in the Gauss-Bonnet-anti-de Sitter geometry, estimated from the
1350:   characteristic data, for $d = 5$, $\ell=0$, $\mu = 1.0$, $\Lambda=
1351:   -0.1$ and several values of $\alpha$.}  
1352: %
1353: \label{GBAdS_L01_d5}
1354: %
1355: \begin{ruledtabular}
1356: \begin{tabular}{ccc}
1357: %
1358:   $\alpha$ & 
1359:   Re($\omega_0$) & -Im($\omega_0$) \\ 
1360: %
1361: \hline \\
1362: %
1363: %
1364: 0.1 & 0.4923 & -0.01585 \\
1365: %
1366: 0.1 & 0.4920 & -0.01593 \\
1367: %
1368: 0.5 & 0.4904 & -0.01634 \\ 
1369: %
1370: 1.0 & 0.4885 & -0.01702 \\ 
1371: % 
1372: 1.5 & 0.4866 & -0.01766 \\ 
1373: %
1374: %
1375: \end{tabular} 
1376: \end{ruledtabular}
1377: \end{table}
1378: 
1379: 
1380: 
1381: 
1382: 
1383: 
1384: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1385: \begin{figure}
1386: %
1387: \resizebox{1\linewidth}{!}{\includegraphics*{fig6.eps}}
1388: %
1389: \caption{Field decay in the Gauss-Bonnet anti-de Sitter
1390:     black holes, for several values of the Gauss-Bonnet coupling. It
1391:     is observed a quasinormal mode dominated region. Asymptotically,
1392:     the field decays in  quasinormal modes. The parameters in this
1393:     graph are $d=5$, $\mu=1.0$ $\Lambda= -0.1$ and $\ell=0$.}
1394: %
1395: \label{fig_GBAdS1}
1396: %
1397: \end{figure}
1398: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1399: 
1400: 
1401: 
1402: 
1403: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1404: \section{Conclusions}
1405: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1406: 
1407: We have considered here frequency and time domain description of
1408: evolution of scalar field perturbations in the exterior of black holes in
1409: Gauss-Bonnet theory of gravity, generally with a $\Lambda$-term. The
1410: quasinormal behavior even though being corrected by
1411: a new parameter, Gauss-Bonnet coupling $\alpha$, are qualitatively
1412: dependent mainly on the $\Lambda$-term and black hole parameters such
1413: as mass $\mu$ and  multipole number $\ell$.  The late-time tails 
1414: for asymptotically flat Gauss-Bonnet black holes, 
1415: do not depend on the Gauss-Bonnet coupling in odd space-time
1416: dimensions, and therefore are the same
1417: as those for d-dimensional Schwarzschild black hole in Einstein
1418: gravity. Moreover, in the case of Gauss-Bonnet-de Sitter black holes, 
1419: the late-time tails, though dependent on $\alpha$, yet, rather
1420: trivially, i.e. only through dependence of the surface gravity at the
1421: cosmological radius on $\alpha$. Thus, the Gauss-Bonnet coupling shows
1422: itself ``minimally'' in late-time behavior. The most interesting 
1423: problem  which remains unsolved is, to find late time tails for 
1424: even dimensions, and thereby, to know whether the power-law tails depend
1425: upon the Gauss-Bonnet term.   
1426: At the same time, we have shown that corrections to the quasinormal
1427: frequencies due to GB-term are not negligible: they may reach $20\%$
1428: for string theory predicted values of $\alpha \approx 1$. 
1429:      
1430:  
1431: 
1432: Even though our analysis can easily be extended to the massive scalar
1433: field, we were limited here by the massless case. We expect that the
1434: influence of the massive term upon the QNMs will be similar to that
1435: found in \cite{konoplyaPLB}, i.e. the lower overtones should be
1436: corrected by the field mass, infinitely high overtone asymptotic will
1437: be unchanged no matter the value of the massive term. 
1438: Also we did not consider the high overtone behavior of the GB black holes.
1439: Generally, the high overtone asymptotics must be studied by totally
1440: different methods \cite{high-n} and deserves separate investigation.   
1441: 
1442: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1443: 
1444: \begin{acknowledgments}
1445:  This work was partially supported by \emph{Funda\c{c}\~{a}o de Amparo
1446: \`{a} Pesquisa do Estado de S\~{a}o Paulo (FAPESP)} and \emph{Conselho
1447:  Nacional de Desenvolvimento Cient\'ifico e Tecnol\'ogico (CNPq)}, Brazil.
1448: \end{acknowledgments}
1449: 
1450: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1451: 
1452: 
1453: 
1454: 
1455: 
1456: 
1457: 
1458: 
1459: 
1460: 
1461: 
1462: \begin{thebibliography}{80}
1463: 
1464: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1465: %1
1466: \bibitem{rs}L. Randall and R. Sundrum,  Phys. Rev. Lett. \textbf{83}
1467:    3370, 4690 (1999).
1468: 
1469: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1470: %2
1471: \bibitem{deser} D. G. Boulware and S. Deser,
1472:   Phys. Rev. Lett. \textbf{55}, 2656 (1985).  
1473: 
1474: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1475: %3
1476: \bibitem{wheeler} J. Wheeler, Nucl. Phys. B\textbf{268}, 737 (1986).
1477: 
1478: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1479: %4
1480: \bibitem{lovelock} D. Lovelock, J. Math. Phys. \textbf{12}, 498 (1971).
1481: 
1482: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1483: %5
1484: \bibitem{zanelli} J. Crisostomo, R. Troncoso and J. Zanelli,
1485:     Phys. Rev. D \textbf{62}, 084013 (2000).
1486: 
1487: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1488: %6
1489: \bibitem{abdcorrea} Elcio Abdalla and L. Alejandro Correa-Borbonet,
1490:   Phys. Rev. D \textbf{65}, 124011 (2002). 
1491: 
1492: 
1493: 
1494: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1495: %7
1496: \bibitem{Rychkov} V. S. Rychkov,  Phys. Rev. D \textbf{70}, 044003 (2004).
1497: 
1498: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1499: %8
1500: \bibitem{Kanti2005}  G. Duffy, C. Harris, P. Kanti, E. Winstanley, 
1501: hep-th/0507274 (2005). 
1502:  
1503: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1504: %9
1505: \bibitem{Kokkotas-99}  K. D. Kokkotas and B. G. Schmidt,
1506: Living Rev. Relativity \textbf{2}, 2 (1999). 
1507: 
1508: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1509: %10
1510: \bibitem{OdintsovJHEP}  S. Nojiri, S. D. Odintsov, J. of High
1511:   Energy Phys. \textbf{0007}, 049 (2000).
1512: 
1513: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1514: %11
1515: \bibitem{Horowitz00} G. T. Horowitz and V. E. Hubeny, Phys. Rev. D
1516:  \textbf{62}, 024027 (2000);  A. Nunez, A. O. Starinets,
1517: Phys. Rev. D \textbf{67}, 124013 (2003).
1518: 
1519: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1520: %12
1521: \bibitem{Konoplya05} R. Konoplya, Phys. Rev. D \textbf{71}, 024038
1522:   (2005) [hep-th/0410057]. 
1523: 
1524: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1525: %13
1526: \bibitem{yer} B. F. Schutz and C. M. Will, Astrophys. J. \textbf{291},
1527:   L33 (1985); S. Iyer and C. M. Will, Phys. Rev. D \textbf{35}, 3621 (1987); 
1528:   R. A. Konoplya, Phys. Rev. D \textbf{68}, 024018 (2003) [gr-qc/0303052].
1529: 
1530: 
1531: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1532: %14
1533: \bibitem{Dotti} G. Dotti, R. J. Gleiser,  gr-qc/0503117.
1534: 
1535: 
1536: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1537: %15
1538: \bibitem{olea} Rodrigo Olea, J. of High Energy Phys. \textbf{0506},
1539:   023 (2005).
1540: 
1541: 
1542: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1543: %16
1544: \bibitem{Gundlach-94} C. Gundlach, R. Price and J. Pullin, Phys. Rev. D
1545:  \textbf{49}, 883 (1994). 
1546: 
1547: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1548: %17
1549: \bibitem{Price-72} R. Price,  Phys. Rev. D  \textbf{5}, 2419 (1972).
1550: 
1551: 
1552: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1553: %18
1554: \bibitem{Brady-97} P. R. Brady, C. M. Chambers,  W. Krivan and
1555:   P. Laguna, Phys. Rev. D \textbf{55}, 7538 (1997).
1556: 
1557: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1558: %19
1559: \bibitem{Brady-99}  P. R. Brady, C. M. Chambers,
1560:   W. G. Laarakkers and E. Poisson, Phys. Rev. D \textbf{60}, 064003
1561:   (1999).
1562: 
1563: \bibitem{moss} I.G. Moss and J.P. Norman, Class. Quant. Grav. 19, 2323 (2002)
1564: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1565: %20
1566: \bibitem{Wang-01} B. Wang, C. Molina and E. Abdalla, Phys. Rev. D
1567: \textbf{63}, 084001 (2001).
1568: 
1569: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1570: %21
1571: \bibitem{Wang-04}  Bin Wang, Chi-Yong Lin, C. Molina, Phys. Rev. D
1572: \textbf{70},  064025 (2004).
1573: 
1574: 
1575: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1576: %22
1577: \bibitem{WKB-ap}  E. Berti, K. D. Kokkotas,  Phys. Rev. \textbf{D} 71,
1578:   124008 (2005); 
1579: R. A. Konoplya, Phys. Rev. D \textbf{66}, 084007 (2002) [gr-qc/0207028]; 
1580: S. Fernando,  Gen. Rel. Grav.  \textbf{37}, 585 (2005);
1581: R. A. Konoplya, E. Abdalla, Phys. Rev. \textbf{D} 71, 084015 (2005) [hep-th/0503029];
1582: R. A. Konoplya, C. Molina,  Phys. Rev. \textbf{D} 71, 124009 (2005) [gr-qc/0504139];
1583: H. Nomura, T. Tamaki, Phys. Rev. \textbf{D} 71, 124033 (2005).
1584: 
1585: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1586: %23
1587: \bibitem{Vitor-Tail} V. Cardoso, S. Yoshida, O.J.C. Dias,
1588:  J. P.S. Lemos,  Phys. Rev.\textbf{D} 68,  061503 (2003). 
1589: 
1590: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1591: %24
1592: \bibitem{konoplya03} R. A. Konoplya,  Phys. Rev. \textbf{D} 68, 124017
1593:   (2003) [hep-th/0309030]. 
1594: 
1595: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1596: %25
1597: \bibitem{Cardoso-03} V. Cardoso and J. P. S. Lemos, Phys. Rev. D \textbf{67}
1598:  084020 (2003).
1599: 
1600: 
1601: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1602: %26
1603: \bibitem{Molina} C. Molina, Phys.Rev. D \textbf{69}, 104013 (2004). 
1604: 
1605: 
1606: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1607: %27
1608: \bibitem{SdShigh} S. Yoshida and T. Futamase, PRD 69, 064025 (2004); 
1609: V. Cardoso, J. Natario, R. Schiappa, J. Math. Phys. \textbf{45} 4698 (2004);
1610: R. A. Konoplya and A. Zhidenko, J. of High Energy Phys. \textbf{037}, 06 (2004) [hep-th/0402080].
1611: 
1612: 
1613: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1614: %28
1615: \bibitem{Abdalla-02} E. Abdalla, K. H. C. Castello-Branco,
1616:   A. Lima-Santos,  Phys. Rev. D \textbf{66}, 104018 (2002).  
1617: 
1618: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1619: %29
1620: \bibitem{natario}  J. Natario, R. Schiappa,  hep-th/0411267 (to be
1621: published in Adv. Theor. Math. Phys.).
1622: 
1623: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1624: %30
1625: \bibitem{Molina-04} C. Molina, D. Giugno, E. Abdalla,
1626:   A. Saa, Phys.Rev. D \textbf{69}, 104013 (2004). 
1627: 
1628: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1629: %31
1630: \bibitem{konoplya02} R. A. Konoplya,  Phys. Rev. \textbf{D} 66, 044009
1631:   (2002)  [hep-th/0205142]. 
1632: 
1633: 
1634: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1635: %32
1636: \bibitem{konoplyaPLB}  R. A. Konoplya,
1637: Phys. Lett. \textbf{B} 550, 117 (2002) [gr-qc/0210105]; R. A. Konoplya, A. V. Zhidenko,
1638: Phys. Lett. \textbf{B} 609, 377 (2005) [gr-qc/0411059].
1639: 
1640: 
1641: 
1642: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1643: %33
1644: \bibitem{high-n} H-P Nollert, Phys. Rev. D \textbf{47}, 5253 (1993);
1645: V. Cardoso, R. Konoplya, J. P.S. Lemos, Phys. Rev. D \textbf{68},
1646: 044024 (2003) [gr-qc/0305037];  L. Motl and A. Neitzke,
1647: Adv. Theor. Math. Phys. \textbf{7}, 307 (2003). 
1648: 
1649: \bibitem{A} E. Berti, M. Cavaglia and L. Gualtieri, Phys. Rev. D69,
1650: 124011 (2004); H. Yoshino, T. Shiromizu and M. Shibata, gr-qc/0508063.
1651: \bibitem{B} V.Cardoso, J.Lemos, S. Yoshida, Phys. Rev. D 69, 044004 (2004).
1652: %\bibitem{D} S. Hollands and A. Ishibashi, J. Math. Phys. 46, 022503 (2005).
1653: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1654: \end{thebibliography}
1655: 
1656: 
1657: 
1658: 
1659: 
1660: \end{document}
1661: