1: %From - Tue Sep 23 14:21:18 2003
2: %Return-Path: <lemos@kelvin.ist.utl.pt>
3: %Delivered-To: vcardoso@einstein.fisica.ist.utl.pt
4: \documentclass[twocolumn,showpacs,preprintnumbers,amssymb]{revtex4}
5: %\documentclass[preprint,showpacs,preprintnumbers,amssymb,nofootinbib]{revtex4}
6:
7: \usepackage{graphicx}% Include figure files
8: \usepackage{bm}
9: %\topmargin 4mm
10: \begin{document}
11:
12: \centerline{}
13: %\vskip 1cm
14: \title{Quasinormal modes of Schwarzschild black holes in four and higher
15: dimensions}
16:
17: \author{Vitor Cardoso}
18: \email{vcardoso@fisica.ist.utl.pt}
19: \author{Jos\'e P. S. Lemos}
20: \email{lemos@kelvin.ist.utl.pt}
21: \author{Shijun Yoshida}
22: \email{yoshida@fisica.ist.utl.pt}
23: \affiliation{
24: Centro Multidisciplinar de Astrof\'{\i}sica - CENTRA,
25: Departamento de F\'{\i}sica, Instituto Superior T\'ecnico,
26: Av. Rovisco Pais 1, 1049-001 Lisboa, Portugal,}
27:
28:
29:
30:
31:
32:
33: \date{\today}
34:
35: \begin{abstract}
36:
37: We make a thorough investigation of the asymptotic quasinormal modes
38: of the four and five-dimensional Schwarzschild black hole for scalar,
39: electromagnetic and gravitational perturbations. Our numerical
40: results give full support to all the analytical predictions by Motl
41: and Neitzke, for the leading term. We also compute the first order
42: corrections analytically, by extending to higher dimensions,
43: previous work of Musiri and
44: Siopsis, and find excellent agreement with the
45: numerical results. For generic spacetime dimension number $D$ the
46: first-order corrections go as $\frac{1}{n^{(D-3)/(D-2)}}$. This means
47: that there is a more rapid convergence to the asymptotic value
48: for the five dimensional case than for the four dimensional case,
49: as we also show numerically.
50:
51: \end{abstract}
52:
53: %\pacs{}
54:
55: \maketitle
56: \newpage
57: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
58: \section{Introduction}
59: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
60: The study of quasinormal modes of black holes began more than thirty
61: years ago, when Vishveshwara \cite{vish} noticed that the signal from
62: a perturbed black hole is, for most of the time, an exponentially
63: decaying ringing signal. It turns out that the ringing frequency and
64: damping timescale are characteristic of the black hole, depending only
65: on its parameters (like the mass and angular momentum). We call these
66: characteristic oscillations the quasinormal modes (QNMs) and the
67: associated frequencies are termed quasinormal frequencies (QN
68: frequencies), because they are really not stationary perturbations.
69: Not surprisingly, QNMs play an important role in the
70: dynamics of black holes, and consequently in gravitational wave
71: physics. In fact, it is possible \cite{echeverria,nakanoringing} to
72: extract the parameters of the black hole simply by observing these
73: QN frequencies, using for example gravitational wave
74: detectors. The discovery that QNMs dominate the answer of a black
75: hole to almost any exterior perturbation was followed by a great
76: effort to find, numerically and analytically, the QN frequencies.
77: For excellent reviews on the status of QNMs, prior to 2000,
78: we refer the reader to Kokkotas and Schmidt
79: \cite{kokkotas} and Nollert \cite{nollert}.
80: It is important to note that on the astrophysical aspect, the most
81: important QN frequencies are the lowest ones, i.e., frequencies
82: with smaller imaginary part, and the most important spacetimes are the
83: asymptotically flat
84: and perhaps now the asymptotically de Sitter.
85: However, three years ago Horowitz and Hubeny \cite{horowitz}
86: pointed out that QNMs of black holes in anti-de Sitter space have a
87: different importance. According to the AdS/CFT correspondence
88: \cite{maldacena}, a black hole in anti-de Sitter space may be viewed
89: as a thermal state in the dual theory. Perturbing this black hole
90: corresponds to perturbing the thermal state, and therefore the typical
91: timescale of approach to thermal equilibrium (which is hard to compute
92: directly in the dual theory) should as well be governed by the lowest
93: QN frequencies. That this is indeed the case was proved by Birmigham,
94: Sachs and Solodukhin \cite{birmingham2} for the BTZ black hole,
95: taking advantage that this is one of the few spacetimes where one can
96: compute exactly its QN frequencies, as showed by Cardoso and Lemos
97: \cite{cardosoqnmbtz}. A similar study was made by Kurita and Sakagami
98: \cite{kurita} for the D-3 brane.
99: This interpretation for the imaginary part of the QN frequencies
100: in terms of timescales of approach
101: to thermal equilibrium in the dual conformal field theory has
102: motivated a generalized search for the quasinormal modes of different
103: black holes in anti-de Sitter spacetime, over the last three years
104: \cite{horowitz,cardosoqnmbtz,qnmads}.
105:
106: Recently, the motivation for studying QN modes of black holes has grown
107: enormously with the conjectures \cite{hod,dreyer,oppenheim,ling}
108: relating the highly damped QNMs (i.e., QN frequencies with large
109: imaginary parts) to black hole area quantization and to the Barbero
110: Immirzi parameter appearing in Loop Quantum Gravity. The seeds of that idea
111: were planted some
112: time ago by Bekenstein \cite{bek0}. A semi-classical reasoning of
113: the conjecture \cite{bek1,bek2} that the black hole area spectrum
114: is quantized leads to
115: \begin{equation}
116: A_{\,n}=\gamma\, l_{P}^2\, n\,,\quad n=1,2,...\quad.
117: \label{areaspectrum}
118: \end{equation}
119: Here $l_{P}$ is the Planck length and $\gamma$ is an undetermined constant.
120: However, statistical physics arguments impose a constraint on $\gamma$
121: \cite{bek1,bek3}:
122: \begin{equation}
123: \gamma=4 \log k\,,
124: \label{areaspectrum1}
125: \end{equation}
126: where $k$ is an integer.
127: The integer $k$ was left undetermined (although there were some suggestions
128: for it \cite{bek2}), until Hod \cite{hod}, supported
129: by some of Bekenstein's ideas, put forward the proposal to determine
130: $k$ via a version of Bohr's correspondence principle, in which one admits
131: that the real part of QN frequencies with a large imaginary
132: part plays a fundamental role.
133: It was seen numerically by Nollert \cite{nollert,nollert2}
134: that QN frequencies with a large imaginary part behave,
135: in the Schwarzschild geometry as
136: \begin{equation}
137: \omega M= 0.0437123+\frac{i}{8}\,(2n+1)\,,
138: \label{asymptQNnollert}
139: \end{equation}
140: where $M$ is the black hole mass, and $n$ the mode number.
141: Hod first realized that $ 0.0437123 \sim \frac{\ln3}{8\pi}$, and
142: went on to say that, if one supposes that the emission of a quantum
143: with frequency $\frac{\ln3}{8\pi}$ corresponds to the least possible energy
144: a black hole can emit, then the change in surface area will be
145: (using $A=16\pi M^2$)
146: \begin{equation}
147: \Delta A= 32\pi MdM=32\pi M{\hbar} \omega= 4{\hbar}\ln 3 \,.
148: \label{surfacearea}
149: \end{equation}
150: Comparing with (\ref{areaspectrum}) we then get $k=3$ and therefore the area
151: spectrum is fixed to
152: \begin{equation}
153: A_{\rm n}=4\ln 3 l_{P}^2 n\,;\,\,n=1,2,...
154: \label{areaspectrumfinal}
155: \end{equation}
156: It was certainly a daring proposal to map $0.0437123$ to $\ln
157: 3$, and even more to use the QN frequencies to quantize the black hole
158: area, by appealing to ``Bohr's correspondence principle''. The
159: risk paid off: recently, Dreyer \cite{dreyer} put forward the
160: hypothesis that if one uses such a correspondence it is possible to
161: fix a formerly free parameter, the Barbero-Immirzi parameter, appearing
162: in Loop Quantum Gravity. Dreyer's work implied that
163: the entropy has contributions from $J=1$ edges mainly.
164: One interpretation (proposed by Dreyer), is that one changes
165: the gauge group from $SU(2)$, which is the most natural one, to
166: $SO(3)$.
167: There are other explanations, though. For example, Corichi
168: \cite{corichi} presents an argument based on simple conservation
169: principles to suggest that one can keep $SU(2)$ and still have
170: consistency with QNM, while Ling and Zhang \cite{ling} propose to
171: consider the supersymmetric extension of the theory.
172:
173: All of these proposals and conjectures could be useless
174: if one could not prove that the real part of the QN frequencies
175: do approach $\frac{\ln3}{8\pi}$, i.e., that the number $0.0437123$ in
176: Nollert's paper was exactly $\frac{\ln3}{8\pi}$. This
177: was accomplished by Motl \cite{motl1} some months ago, using an
178: ingenious technique, working with the continued fraction
179: representation of the wave equation. Subsequently, Motl and Neitzke
180: \cite{motl2} used a more flexible and powerful approach, called
181: the monodromy method, and were not only able to rederive the four
182: dimensional Schwarzschild value $\frac{\ln3}{8\pi}$, but also to
183: compute the asymptotic value for generic $D$-dimensional Schwarzschild
184: black holes thereby confirming a previous conjecture by Kunstatter
185: \cite{kunstatter}.
186: They have also predicted the form of the asymptotic QN frequencies
187: for the Reissner-Nordstr\"om geometry, a prediction which was verified
188: by Berti and Kokkotas \cite{bertirn}.
189:
190: It is interesting to speculate why, in thirty years investigating QNMs
191: and QN frequencies, no one has ever been able to deduce analytically
192: the asymptotic value of the QN frequencies, nor even for the
193: Schwarzschild geometry, which is the simplest case. In our opinion,
194: there are three important facts that may help explain this: first,
195: the concern was mostly with the lowest QN frequencies, the ones
196: with smaller imaginary part, since they are the most important in the
197: astrophysical context where QNMs were inserted until three years
198: ago. In fact, most of the techniques to find QN frequencies, and there
199: were many, have been devised to compute the lowest QN frequencies (see
200: \cite{kokkotas,nollert} for a review). Secondly, there was no
201: suggestion for the asymptotic $\ln{3}$ value.
202: To know this value serves as
203: a strong stimulation. The third reason is obvious: this was a very difficult
204: technical problem.
205:
206: It is thus satisfying that Motl and Neitzke have proved exactly
207: that the limit is $\ln{3}$. In addition, Van den Brink \cite{brinkasympt}
208: used a
209: somewhat different approach to rederive the $\frac{\ln3}{8\pi}$
210: value, and at the same time computed the first order correction to
211: gravitational QN frequencies, confirming analytically Nollert's
212: \cite{nollert2} numerical results. However,
213: Motl and Neitzke's monodromy method is more flexible and powerful: it
214: has been easily adapted to other spacetimes. For example, Castello-Branco
215: and Abdalla \cite{karlucio} have generalized the results to the
216: Schwarzschild-de Sitter geometry, while Schiappa, Nat\'ario and
217: Cardoso \cite{ricardo} have used it to compute the asymptotic values
218: in the $D$-dimensional Reissner-Nordstr\"om geometry.
219: Moreover the first order corrections are also easily obtained, as has
220: been showed by Musiri and Siopsis \cite{musiri2} for the four
221: dimensional Schwarzschild geometry. Here we shall also generalize
222: these first order corrections so that they can encompass a general
223: $D$-dimensional Schwarzschild black hole.
224: A body of work has been growing on the subject of computing highly
225: damped QNMs, both in asymptotically flat and in de Sitter or anti-de
226: Sitter spacetimes \cite{qnmkerr,asymptall}.
227:
228: As an aside, we note that one of the most intriguing
229: features of Motl and Neitzke's technique is that the region beyond the
230: event horizon, which never enters in the definition of quasinormal
231: modes, plays an extremely important role. In particular, the
232: singularity at $r=0$ is decisive for the computation. At the
233: singularity the effective potential for wave propagation blows
234: up. Somehow, the equation, or the singularity, knows what we are
235: seeking! It is also worth of note the following: for high frequencies, or
236: at least frequencies with a large imaginary part, the important region
237: is therefore $r=0$ where the potential blows up, whereas for low
238: frequencies, of interest for late-time tails \cite{tails}, it is the
239: other limit, $r \rightarrow \infty$ which is important.
240:
241:
242:
243: The purpose of this paper is two-fold: first we want to enlarge and settle
244: the results for the four dimensional Schwarzschild black hole. We shall
245: first numerically confirm Nollert's results for the gravitational
246: case, and Berti and Kokkotas'\cite{bertirn} results for the scalar
247: case, and see that the match between these and the analytical
248: prediction is quite good. Then we shall fill-in a gap left behind in
249: these two studies: the electromagnetic asymptotic QN frequencies, for
250: which there are no numerical results, but a lot of analytical
251: predictions. This will put the monodromy method on a firmer
252: ground. One might say that in the present context (of black hole
253: quantization and relation with Loop Quantum Gravity) electromagnetic
254: perturbations are not relevant, only the gravitational ones are. This
255: is not true though, because when considering the Reissner-Nordstr\"om case
256: gravitational and electromagnetic perturbations are coupled
257: \cite{chandra} and it is therefore of great theoretical interest to
258: study each one separately, in the case they do decouple, i.e., the
259: Schwarzschild geometry.
260: We find that all the analytical predictions are correct: scalar and
261: gravitational QN frequencies asymptote to
262: $\frac{\omega}{T_H}=\ln{3}+i(2n+1)\pi$
263: with $\frac{1}{\sqrt n}$ corrections, where
264: $T_H=\frac{1}{8\pi M}$ is the Hawking temperature. The
265: corrections are well described by the existing analytical formulas.
266: On the other hand, electromagnetic QN frequencies asymptote to
267: $\frac{\omega}{T_H}=0+i2n\pi$ with no $\frac{1}{\sqrt n}$ corrections, as
268: predicted by Musiri and Siopsis \cite{musiri2}. Indeed we find that
269: the corrections seem to appear only at the $\frac{1}{n^{3/2}}$ level.
270:
271: The second aim of this paper is to establish numerically the validity of the
272: monodromy method by extending the numerical calculation to higher
273: dimensional Schwarzschild black holes, in particular to the
274: five dimensional one. We obtain numerically in the five dimensional case
275: the asymptotic
276: value of the QN frequencies, which turns out to be $\ln 3+i(2n+1)\pi$ (in
277: units of the black hole temperature), for scalar, gravitational tensor
278: and gravitational vector perturbations. Equally important, is that
279: the first order corrections appear as $\frac{1}{n^{2/3}}$ for these
280: cases. We shall generalize Musiri and Siopsis' method to higher
281: dimensions and find that (i) the agreement with the numerical data is
282: very good (ii) In generic $D$ dimensions the corrections appear at the
283: $\frac{1}{n^{(D-3)/(D-2)}}$ level. Electromagnetic perturbations in
284: five dimensions seem to be special: according to the analytical
285: prediction by Motl and Neitzke, they should not have any asymptotic QN
286: frequencies. Our numerics seem not to go higher enough in mode number
287: to prove or disprove this. They indicate that the real part approaches
288: zero, as the mode number increases, but we have not been able to
289: follow the QN frequencies for very high overtone number. We also give
290: complete tables for the lowest lying QN frequencies of the five
291: dimensional black holes, and confirm previous results
292: \cite{vitoroscarjose,konoplyawkb,roman2,ida} for the fundamental QN
293: frequencies obtained using WKB-like methods. We find no purely
294: imaginary QN frequencies for the gravitational QNMs, which may be
295: related to the fact that the potentials are no longer superpartners.
296:
297:
298:
299:
300:
301:
302:
303:
304: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
305: \section{QN frequencies of the four-dimensional Schwarzschild black hole }
306: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
307:
308: The first computation of QN frequencies of four-dimensional
309: Schwarzschild black holes was carried out by Chandrasekhar and
310: Detweiler \cite{chandradet}. They were only able to compute the
311: lowest lying modes, as it was and still is quite hard to find
312: numerically QN frequencies with a very large imaginary part. After
313: this pioneering work, a number of analytical, semi-analytical and
314: numerical techniques were devised with the purpose of computing higher
315: order QNMs and to establish the low-lying ones using different methods
316: \cite{kokkotas,nollert}. The most
317: successful numerical method to study QNMs of black holes was proposed
318: by Leaver \cite{Le85}. In a few words his method amounts to reducing the
319: problem to a continued fraction, rather easy to implement
320: numerically. Using this method Leaver was able to determine the first
321: 60 modes of the Schwarzschild black hole, which at that time was about
322: an order of magnitude better than anyone had ever gone before. The
323: values of the QN frequencies he determined, and we note he only
324: computed gravitational ones, suggested that there is an infinite
325: number of QN modes, that the real part of the QN frequencies approach
326: a finite value, while the imaginary part grows without bound.
327: That the real part of the QN frequencies approach
328: a finite value was however opened to debate \cite{guinn}: it seemed one still
329: had to
330: go higher in mode number in order to ascertain the true asymptotic
331: behaviour. Nollert \cite{nollert2} showed a way out
332: of these problems: he was able to improve Leaver's method in a very
333: simple fashion so as to go very much higher in mode number: he was
334: able to compute more than 2000 modes of the gravitational QNMs. His
335: results were clear: the real part of the gravitational QN frequencies
336: approaches a constant value $\sim 0.0437123/M $, and this value is
337: independent of $l$. The imaginary part grows without bound as
338: $(2n+1)/8$ where $n$ is any large integer. These asymptotic
339: behaviours have both leading corrections, which were also determined
340: by Nollert. He found that to leading order the asymptotic behaviour
341: of the gravitational QN frequencies of a Schwarzschild black hole is
342: given by Eq. (\ref{asymptQNnollert}) plus a leading correction,
343: namely,
344: \begin{equation}
345: \omega M=0.0437123+\frac{i}{8}\,(2n+1)+ \frac{a}{\sqrt{n}}\,,
346: \label{nono}
347: \end{equation}
348: where the correction term $a$ depends on $l$. About ten years
349: later, Motl \cite{motl1} proved analytically that this is indeed the
350: correct asymptotic behaviour. To be precise he proved that
351: \begin{equation}
352: \frac{\omega}{T_H}=\ln{3}+i\,(2n+1)\pi+\,{\rm corrections}\,.
353: \label{nonolubos}
354: \end{equation}
355: Since the Hawking temperature $T_H$ is
356: $T_H=\frac{1}{8\pi M}$, Nollert's
357: result follows.
358: He also proved that this same result also holds for scalar QN
359: frequencies, whereas electromagnetic QN frequencies asymptote to a
360: zero real part. Motl and Neitzke \cite{motl2} using a completely
361: different approach, have rederived this result whereas Van den Brink
362: \cite{brinkasympt} was able to find the leading asymptotic correction
363: term $a$ in (\ref{nono}) for the gravitational case. Very
364: recently, Musiri and Siopsis \cite{musiri2}, leaning on Motl and
365: Neitzke's technique, have found the leading correction term for any
366: field, scalar, electromagnetic and gravitational.
367: This correction term is presented in section \ref{apendice}, where
368: we also generalize to higher dimensional black holes.
369: In particular, for electromagnetic perturbations they find that the
370: first order
371: corrections are zero.
372:
373: Here we use Nollert's method to compute scalar, electromagnetic and
374: gravitational QN frequencies of the four-dimensional Schwarzschild
375: black hole. The details on the implementation of this method in four
376: dimensions are well known and we shall not dwell on it here any more.
377: We refer the reader to the original references \cite{nollert2,Le85}.
378: \begin{figure}
379: \centerline{\includegraphics[width=7 cm,height=7 cm]
380: {qnfall4d.eps}}
381: \caption{Scalar ($s=0$), electromagnetic ($s=1$) and gravitational ($s=2$)
382: QN frequencies of a four-dimensional Schwarzschild black hole.
383: Note that the dots indicate the QN frequencies there, and the lines
384: connecting the dots only help to figure out
385: to which multipole ($l$, $s$) they belong to.
386: We show
387: the lowest 500 modes for the lowest radiatable multipoles of each
388: field, i.e., $l=0$ for the scalar field, $l=1$ for the electromagnetic
389: and $l=2$ for gravitational. However, the asymptotic behaviour is
390: $l$-independent. It is quite clear from this plot that highly damped
391: electromagnetic QN frequencies have a completely different behaviour
392: from that of scalar or gravitational highly damped QN frequencies. In
393: fact, the electromagnetic ones asymptote to $\frac{\omega}{T_H} \rightarrow
394: i(2n)\pi\,\,,n\rightarrow \infty$, as predicted by Motl
395: \cite{motl1} and Motl and Neitzke \cite{motl2}. The scalar and
396: gravitational ones asymptote to $\frac{\omega}{T_H}\rightarrow
397: \ln{3}+i(2n+1)\pi\,\,,n\rightarrow \infty$.}
398: \label{fig:qnall4d}
399: \end{figure}
400:
401:
402:
403: Our results are summarized in Figures
404: \ref{fig:qnall4d}-\ref{fig:qnelec4d}, and Tables \ref{tab:corrd4j0}-
405: \ref{tab:corrd4j2}.
406: In Fig. \ref{fig:qnall4d} we
407: show the behaviour of the 500 lowest QN frequencies for the lowest
408: radiatable multipole of each field
409: (we actually have computed the first 5000 QN frequencies, but for
410: a better visualization we do not show them all): $l=0$ for scalar, $l=1$ for
411: electromagnetic and $l=2$ gravitational.
412: Note that in Fig. \ref{fig:qnall4d}, as well as in all other plots in this
413: work,
414: the dots indicate the QN frequencies there, and the lines
415: connecting the dots only help to figure out
416: to which multipole ($l$, $s$) they belong to.
417: In what concerns the asymptotic behaviour, our numerics show it is not
418: dependent on $l$, and that formula (\ref{nono}) holds. The results for
419: the scalar and gravitational QN frequencies are not new: the
420: gravitational ones have been obtained by Nollert \cite{nollert2}, as
421: we previously remarked, and the scalar ones have been recently arrived
422: at by Berti and Kokkotas \cite{bertirn}. Here we have verified both.
423: Both scalar and gravitational QN frequencies have a real part
424: asymptoting to $\ln{3}\times T_H$. Electromagnetic QN frequencies,
425: on the other hand, have a real part asymptoting to $0$. This is
426: clearly seen in Fig. \ref{fig:qnall4d}. For gravitational
427: perturbations Nollert found, and we confirm, that the correction term
428: $a$ in (\ref{nono}) is
429: \begin{eqnarray}
430: a= 0.4850\,,\,\,l=2
431: \nonumber \\
432: \!\!\!a= 1.067 \,,\,\,l=3
433: \label{corrtermsnum}
434: \end{eqnarray}
435: We have also computed the correction terms for the scalar case, and we have
436: obtained
437: results in agreement with Berti and Kokkotas' ones \cite{bertirn} .
438: Our results are summarized in Tables \ref{tab:corrd4j0}-\ref{tab:corrd4j2},
439: where
440: we also show the analytical prediction \cite{musiri2} (see section
441: \ref{apendice}).
442: To fix the conventions adopted throughout the rest of the paper,
443: we write
444: \begin{equation}
445: \frac{\omega}{T_H}=\ln{3}+i(2n+1)\pi+\frac{{\rm Corr}}{\sqrt{n}}\,.
446: \label{conv}
447: \end{equation}
448: Thus, the correction term $a$ in (\ref{nono}) is related to {\rm Corr}
449: by $a={\rm Corr}\times T_H$.
450:
451: \vskip 1mm
452: \begin{table}
453: \caption{\label{tab:corrd4j0} The correction coefficients for the
454: four-dimensional Schwarzschild black hole, both numerical, here
455: labeled as ``${\rm Corr_{4}^{N}}$'' and analytical, labeled as ``${\rm
456: Corr_{4}^{A}}$''. These results refer to scalar
457: perturbations.
458: The analytical results are extracted from \cite{musiri2}
459: (see also formula (\ref{corrd4} below). Notice the very good agreement
460: between the numerically extracted results and the
461: analytical prediction.These values were obtained using the first
462: 5000 modes.}
463: \begin{ruledtabular}
464: \begin{tabular}{l|ll} \hline
465: \multicolumn{1}{c}{} &
466: \multicolumn{2}{c}{ $s=0$}\\ \hline
467: $l$ &${\rm Corr_4^N}$:&${\rm Corr_4^A}$\\ \hline
468: 0 &1.20-1.20i &1.2190-1.2190i \\ \hline
469: 1 &8.52- 8.52i &8.5332-8.5332i \\ \hline
470: 2 &23.15-23.15i &23.1614-23.1614i \\ \hline
471: 3 &44.99-44.99i &45.1039-45.1039i \\ \hline
472: 4 &74.34-74.34i &74.3604-74.3605i \\ \hline
473: \end{tabular}
474: \end{ruledtabular}
475: \end{table}
476: \vskip 1mm
477: \vskip 1mm
478: \begin{table}
479: \caption{\label{tab:corrd4j2} The correction coefficients for the
480: four-dimensional Schwarzschild black hole, both numerical, here
481: labeled as ``${\rm Corr_{4}^{N}}$'' and analytical, labeled as ``${\rm
482: Corr_{4}^{A}}$''. These results refer to gravitational
483: perturbations.
484: The analytical results are extracted from \cite{musiri2}
485: (see also formula (\ref{corrd4} below). Notice the very good agreement
486: between the numerically extracted results and the
487: analytical prediction. These values were obtained using the first
488: 5000 modes.}
489: \begin{ruledtabular}
490: \begin{tabular}{l|ll} \hline
491: \multicolumn{1}{c}{} &
492: \multicolumn{2}{c}{ $s=2$}\\ \hline
493: $l$ &${\rm Corr_4^N}$:&${\rm Corr_4^A}$\\ \hline
494: 2 & 6.08-6.08i & 6.0951-6.0951i \\ \hline
495: 3 & 13.39-13.39i & 13.4092-13.4092i \\ \hline
496: 4 & 23.14-23.14i & 23.1614-23.1614i \\ \hline
497: 5 & 35.33-35.33i & 35.3517-35.3517i\\ \hline
498: 6 & 48.90-48.90i & 49.9799-49.9799i \\ \hline
499: \end{tabular}
500: \end{ruledtabular}
501: \end{table}
502: \vskip 1mm
503: As for the electromagnetic correction terms, we found it was not very easy
504: to determine
505: them. In fact, it is hard, even using Nollert's method, to go very high in mode
506: number for these perturbations (we have made it to $n=5000$).
507: The reason is tied to the fact that these frequencies asymptote to zero,
508: and it is hard to determine QN frequencies with a vanishingly small real part.
509: Nevertheless, there are some features one can be sure of.
510: We have fitted the electromagnetic data to the following form
511: $\frac{\omega}{T_H}=i(2n)\pi+\frac{b}{\sqrt{n}}\,,$
512: and found that this gave very poor results. Thus one saw numerically that
513: the first order correction term is absent, as predicted in \cite{musiri2}
514: (see also section \ref{apendice} where we rederive these corrections,
515: generalizing them
516: to arbitrary dimension).
517: Our data seems to indicate that the leading correction term is of the form
518: $\frac{b}{n^{3/2}}$. This is more clearly seen in Fig. \ref{fig:qnelec4d}
519: where we plot the first electromagnetic QN frequencies on a $\ln$ plot.
520: \begin{figure}
521: \centerline{\includegraphics[width=7 cm,height=7 cm]
522: {qnfelecd4.eps}}
523: \caption{Electromagnetic QN frequencies of a
524: four-dimensional Schwarzschild black hole.}
525: \label{fig:qnelec4d}
526: \end{figure}
527: For frequencies with a large imaginary part, the slope is about $-2/3$ as
528: it should if the asymptotic behaviour is of the form
529: \begin{equation}
530: \frac{\omega}{T_H}=i2n\pi+\frac{b}{n^{3/2}}\,.
531: \label{true}
532: \end{equation}
533: Our numerics indicate that $b$ has an $l$-dependence going like
534: \begin{equation}
535: b \sim c_1 l(l+1)+c_2 \left[l(l+1)\right]^2+c_3 \left[l(l+1)\right]^3\,,
536: \label{bcoef}
537: \end{equation}
538: where $c_1$, $c_2$, and$c_3$ are constants. This is also the expected
539: behaviour, since a $\frac{1}{n^{3/2}}$ dependence means we have to go
540: to third order perturbation theory, where we get corrections of the
541: form (\ref{bcoef}) as one can easily convince oneself.
542: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
543: \section{QN frequencies of $D$-dimensional Schwarzschild black holes }
544: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
545:
546: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
547: \subsection{Equations and conventions}
548: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
549: The metric of the $D$-dimensional
550: Schwarzschild black hole in ($t,r,\theta_1,\theta_2,..,\theta_{D-2}$)
551: coordinates is \cite{tangherlini,myersperry}
552: \begin{equation}
553: ds^2= -fdt^2+
554: f^{-1}dr^2
555: +r^2d\Omega_{D-2}^2\,,
556: \label{metrictang}
557: \end{equation}
558: with
559: \begin{equation}
560: f=1-\frac{m}{r^{D-3}}.
561: \label{fdef}
562: \end{equation}
563: The mass of the black hole is given by
564: $M=\frac{(D-2)\Omega_{D-2} m}{16\pi{\cal G}}$, where
565: $\Omega_{D-2}=\frac{2\pi^{(D-1)/2}}{\Gamma[(D-1)/2]}$ is the area of a
566: unit $(D-2)$ sphere, $d\Omega_{D-2}^{2}$ is the line element on
567: the unit sphere $S^{D-2}$, and ${\cal G}$ is Newton's constant in
568: $D$-dimensions.
569: In the following it will prove more useful to rescale variables so that
570: the form of the metric is (\ref{metrictang}) but with $f=1-\frac{1}{r^{D-3}}$,
571: i.e., we shall choose $m=1$, following the current fashion.
572: We will only consider the linearized
573: approximation, which means that we are considering wave fields outside
574: this geometry that are so weak they do not alter this
575: background.
576: The evolution equation for a massless scalar field follows
577: directly from the Klein-Gordon equation (see
578: \cite{cardosoradadsddim} for details). The gravitational evolution
579: equations have recently been derived by Kodama and Ishibashi
580: \cite{kodama}. There are three kinds of gravitational perturbations,
581: according to Kodama and Ishibashi's terminology: the scalar
582: gravitational, the vector gravitational and the tensor gravitational
583: perturbations. The first two already have their counterparts in
584: $D=4$, which were first derived by Regge and Wheeler \cite{regge} and
585: by Zerilli \cite{zerilli}. The tensor type is a new kind appearing in
586: higher dimensions. However, it obeys exactly the same equation as
587: massless scalar fields, as can easily be seen. Due to the complex
588: form of the gravitational scalar potential, we shall not deal with
589: it. Instead, we shall only consider the tensor and vector type of
590: gravitational perturbations. In any case, if the analytic results are correct,
591: then the gravitational scalar QN frequencies should have the same asymptotic
592: form as the gravitational vector and tensor QN frequencies.
593: The evolution equation for the electromagnetic field in the higher
594: dimensional Schwarzschild geometry was arrived at for the first time
595: by Crispino, Higuchi and Matsas \cite{matsas}. It has recently
596: been rederived by Kodama and Ishibashi, in a wider context of charged
597: black hole perturbations. We shall follow Kodama and Ishibashi's
598: terminology. According to them, there are two kinds of electromagnetic
599: perturbations: the vector and scalar type. If one makes the charge of
600: the black hole $Q=0$ in Kodama and Ishibashi's equations one recovers
601: the equations by Crispino, Higuchi and Matsas, although this seems to
602: have been overlooked in the literature.
603: The evolution equation for all kinds of
604: fields (scalar, gravitational and electromagnetic)
605: can be reduced to the second order differential equation
606: \begin{equation}
607: \frac{d^2\Psi}{dr_{*}^2}+(\omega^2-V)\Psi=0 \,,
608: \label{eveq}
609: \end{equation}
610: where $r$ is a function of the tortoise coordinate $r_*$, defined through
611: $\frac{\partial r}{\partial
612: r_*}=f(r)$, and the potential can be written in compact form as
613: \begin{eqnarray}
614: V &=& f(r)
615: {\biggl [}
616: \frac{l(l+D-3)}{r^2}+\frac{(D-2)(D-4)}{4r^2}+
617: \nonumber \\
618: & &
619: \frac{(1-j^2)(D-2)^2}{4r^{D-1}}
620: {\biggr ]}
621: \,.
622: \label{potentialj}
623: \end{eqnarray}
624: The constant $j$ depends on what kind of field one is studying:
625: \begin{equation}
626: j=\left\{ \begin{array}{llll}
627: 0\,, & {\rm \,scalar\, and\, gravitational\, tensor}\\
628: & {\rm perturbations}. \\
629: 1\,, &{\rm \, gravitational\, vector} \\
630: & {\rm perturbations}. \\
631: \frac{2}{D-2}\,, &{\rm \, electromagnetic \,vector} \\
632: & {\rm perturbations}. \\
633: 2- \frac{2}{D-2}\,, &{\rm \, electromagnetic\, scalar} \\
634: & {\rm perturbations}\,.
635: \end{array}\right.
636: \label{jdef}
637: \end{equation}
638: Notice that in four dimensions, $j$ reduces to the usual values
639: given by Motl and Neitzke \cite{motl2}.
640: According to our conventions, the Hawking temperature of a
641: $D$-dimensional Schwarzschild black hole is
642: \begin{equation}
643: T_H=\frac{D-3}{4\pi}
644: \label{hawktemp}
645: \end{equation}
646:
647: Our purpose here is to investigate the QN frequencies of higher
648: dimensional black holes. Of course one cannot study every $D$. We
649: shall therefore focus on one particular dimension, five, and make a
650: complete analysis of its QN frequencies. The results should be
651: representative. In particular, one feature that distinguishes
652: an arbitrary $D$ from the four dimensional case is that the correction
653: terms come with a different power, i.e., whereas in four dimensions
654: the correction is of the form $\frac{1}{\sqrt{n}}$ we shall find that in
655: generic
656: $D$ it is of the form $\frac{1}{n^{(D-3)/(D-2)}}$. Thus, if one verifies
657: this for
658: $D=5$ for example, one can ascertain it will hold for arbitrary $D$.
659: We shall now briefly sketch our numerical procedure for finding
660: QN frequencies of a five dimensional black hole. This is just Leaver's and
661: Nollert's technique with minor modifications.
662:
663: The QNMs of the higher-dimensional Schwarzschild black hole are characterized
664: by the boundary conditions of incoming waves at the black hole horizon and
665: outgoing waves at spatial infinity, written as
666: %
667: \begin{equation}
668: \Psi(r)\rightarrow\left\{
669: \begin{array}{ll}
670: e^{-i\omega r_*}&{\rm as}\ r_*\rightarrow \infty\\
671: e^{ i\omega r_*}&{\rm as}\ r_*\rightarrow-\infty\,,
672: \end{array}
673: \right.
674: \label{def-bc}
675: \end{equation}
676: %
677: where the time dependence of perturbations has been assumed to be
678: $e^{i\omega t}$.
679: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
680: \subsection{Perturbative calculation of QNMs of $D$-dimensional
681: Schwarzschild black holes}
682: \label{apendice}
683: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
684: In this section we shall outline the procedure for computing the
685: first order corrections to the asymptotic value of the QN frequencies
686: of a $D$-dimensional Schwarzschild black hole. This will be a
687: generalization of Musiri and Siopsis' method \cite{musiri2}, so we
688: adopt all of their notation, and we refer the reader to their paper
689: for further details.
690: The computations are however rather tedious and the final expressions are too
691: cumbersome, so we shall refrain from giving explicit
692: expressions for the final result.
693:
694: We start with the expansion of the potential $V$ near the singularity
695: $r=0$. One can easily show that near this point the potential
696: (\ref{potentialj})
697: may be approximated as
698: \begin{equation}
699: V \sim -\frac{\omega ^2}{4z^2}\left[1-j^2+\frac{A}{(-z
700: \omega)^{(D-3)/(D-2)}} \right]\,,
701: \label{potr0}
702: \end{equation}
703: where we adopted the conventions in \cite{musiri2} and therefore
704: $z=\omega r_*$.
705: The constant $A$ is given by
706: \begin{eqnarray}
707: \!\!\!\!A\!&=&\frac{4l(l+D-3)+(D-2)(D-4)-(D-2)^2(1-j^2)}{(D-2)^{(D-1)/(D-2)}}+
708: \nonumber \\
709: & &
710: \frac{2(1-j^2)(D-2)^{(2D-5)/(D-2)}}{2D-5} \,.
711: \label{constA}
712: \end{eqnarray}
713: In $D=4$ this reduces to the usual expression
714: \cite{musiri2,andy} for the potential near $r=0$. Expression
715: (\ref{potr0}) is a formal expansion in
716: $\frac{1}{\omega^{(D-3)/(D-2)}}$, so we may anticipate that indeed the
717: first order corrections will appear in the form
718: $\frac{1}{n^{(D-3)/(D-2)}}$. So now we may proceed in a direct
719: manner:
720: we expand the wavefunction to first order in $\frac{1}{\omega^{(D-3)/(D-2)}}$
721: as
722: \begin{equation}
723: \Psi= \Psi^{(0)}+\frac{1}{\omega^{(D-3)/(D-2)}}\Psi^{(1)}\,,
724: \label{expansionpsi}
725: \end{equation}
726: and find that the first-order correction obeys the equation
727: \begin{equation}
728: \frac{d^2\Psi^{(1)}}{dz^2}+(\frac{1-j^2}{4z^2}+1)\Psi^{(1)}=
729: \omega^{(D-3)/(D-2)}\delta V \Psi^{(0)}\,,
730: \label{firstordereq}
731: \end{equation}
732: with
733: \begin{equation}
734: \delta V=-\frac{A}{\omega^{(D-3)/(D-2)}(-z)^{(D-1)/(D-2)}}\,.
735: \label{deltaV}
736: \end{equation}
737: All of Musiri and Siopsis' expressions follow directly to the $D$-dimensional
738: case, if one makes the replacement $\sqrt{-\omega r_0} \rightarrow
739: \omega^{(D-3)/(D-2)}$
740: in all their expressions.
741: For example, the generalsolution of (\ref{firstordereq}) is
742: \begin{equation}
743: \Psi^{(1)}_{\pm}={\cal C} \Psi^{(0)}_{+} \int_{0}^z \Psi^{(0)}_{-}\delta V
744: \Psi^{(0)}_{\pm}-
745: {\cal C} \Psi^{(0)}_{-} \int_{0}^z \Psi^{(0)}_{+}\delta V \Psi^{(0)}_{\pm}
746: \,,
747: \label{solfirstorder}
748: \end{equation}
749: where
750: \begin{equation}
751: {\cal C}=\frac{\omega^{(D-3)/(D-2)}}{\sin{j\pi/2}}\,,
752: \label{calC}
753: \end{equation}
754: and the wavefunctions $\Psi^{(0)}_{\pm}$ are
755: \begin{equation}
756: \Psi^{(0)}_{\pm}=\sqrt{\frac{\pi z}{2}}J_{\pm j/2}(z)\,.
757: \label{psi0}
758: \end{equation}
759: The only minor modification is their formula (30).
760: In $D$-dimensions, it follows that
761: \begin{equation}
762: \Psi^{(1)}_{\pm}=z^{1 \pm j/2 +k} G_{\pm}(z)\,,
763: \label{psi1}
764: \end{equation}
765: where $k=\frac{D-4}{2(D-2)}$, and $G_{\pm}$ are even analytic functions of $z$.
766: So we have all the ingredients to construct the first order corrections in the
767: $D$-dimensional Schwarzschild geometry. Unfortunately the final expressions
768: turn out to be quite cumbersome, and we have not managed to simplify them.
769: We have worked with the symbolic manipulator {\it Mathematica}. It is
770: possible to
771: obtain simple expression for any particular $D$, but apparently not for a
772: generic
773: $D$. For generic $D$ we write
774: \begin{equation}
775: \frac{\omega}{T_H}=\ln{(1+2\cos{\pi j})}+i(2n+1)\pi+
776: \frac{{\rm Corr_D}}{\omega^{(D-3)/(D-2)}}
777: \label{xxx}
778: \end{equation}
779: So the leading term is
780: $\frac{\omega}{T_H}=\ln{3}+i(2n+1)\pi$, for scalar, gravitational tensor and
781: gravitational tensor perturbations (and the same holds for gravitational
782: scalar).
783: However, for electromagnetic perturbations (vector or scalar) in $D=5$,
784: $1+2\cos{\pi j}=1+2\cos{\frac{2\pi}{3}}=0$. So there seem to be no QN
785: frequencies
786: as argued by Motl and Neitzke \cite{motl2}.
787: We obtained for $D=4$ the following result for the correction coefficient
788: in (\ref{xxx}),
789: \begin{equation}
790: {\rm Corr_4}=- \frac {(1-i)\pi ^{3/2}(-1+j^2-3l(l+1))\frac{\cos{j\pi}}{2}
791: \Gamma[1/4]}
792: { 3(1+2\cos{j\pi})\Gamma[3/4]\Gamma[3/4-j/2]\Gamma[3/4+j/2] }\,,
793: \label{corrd4}
794: \end{equation}
795: which reduces to Musiri and Siopsis' expression. For $D=5$
796: it is possible also to find a simple expression for the
797: corrections:
798: \begin{eqnarray}
799: \!\!\!{\rm Corr_5}\! &=\!\!\!&\frac {i(9j^2-(24+20l(l+2)))\pi^{3/2}}
800: { 120\times 3^{1/3}\Gamma[2/3]\Gamma[2/3-j/2]\Gamma[2/3+j/2] }
801: \nonumber \\
802: & &
803: \times \Gamma[1/6]
804: \frac{ \frac{e^{ij\pi}}{\sin{((1/3+j/2)\pi)}}-
805: \frac{1}{\sin{((2/3+j/2)\pi)}}}{-1+e^{ij\pi}}\,.
806: \label{corrd5}
807: \end{eqnarray}
808: The limits $j \rightarrow 0,2$ are well defined and yield
809: \begin{equation}
810: {\rm Corr_5}^{j=0}= \frac {(1-i\sqrt{3})(6+10l+5l^2)\pi^{3/2}\Gamma[1/6]}
811: { 45\times 3^{1/3}\Gamma[2/3]^3} \,,
812: \label{corrd5j0}
813: \end{equation}
814: \begin{equation}
815: {\rm Corr_5}^{j=2}= \frac {(-1+i\sqrt{3})(-3+10l+5l^2)\pi^{3/2}\Gamma[1/6]}
816: { 45\times 3^{1/3}\Gamma[-1/3]\Gamma[2/3]\Gamma[5/3]} \,,
817: \label{corrd5j2}
818: \end{equation}
819: We list in Tables \ref{tab:corrj0}-\ref{tab:corrj2} some values of this
820: five dimensional
821: correction for some values of the parameter $j$ and $l$ and compare
822: them with the results extracted numerically. The agreement is quite
823: good. The code for extracting the generic $D$-dimensional corrections
824: is available from the authors upon request.
825: It should not come as a surprise that for $j=2/3$ the correction term blows up:
826: indeed already the zeroth order term is not well defined.
827: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
828: \subsection{Numerical procedure and results }
829: \label{numtechniqueresults}
830: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
831:
832: \subsubsection{Numerical procedure }
833: \label{numtechnique}
834:
835:
836: In order to numerically obtain the QN frequencies,
837: in the present investigation, we make use of Nollert's method
838: \cite{nollert}, since asymptotic behaviors of QNMs in the
839: limit of large imaginary frequencies are prime concern in the present study.
840:
841:
842: In our numerical study, we only consider the QNMs of the Schwarzschild black
843: hole in the five dimensional spacetime, namely the $D=5$ case.
844: As we have remarked, this should be representative.
845: The tortoise
846: coordinate is then reduced to
847: %
848: \begin{equation}
849: r_*=x^{-1}+{1\over 2x_1}\ln(x-x_1)-{1\over 2x_1}\ln(x+x_1)\,,
850: \end{equation}
851: %
852: where $x=r^{-1}$ and $x_1=r_h^{-1}$. Here, $r_h$ stands for the horizon
853: radius of the black hole. The perturbation function $\Psi$ may be expanded
854: around the horizon as
855: %
856: \begin{equation}
857: \Psi=e^{-i\omega x^{-1}}(x-x_1)^\rho(x+x_1)^\rho\sum_{k=0}^\infty
858: a_k\left({x-x_1\over-x_1}\right)^k\,,
859: \label{expansion}
860: \end{equation}
861: %
862: where $\rho=i\omega/2x_1$ and $a_0$ is taken to be $a_0=1$. The
863: expansion coefficients $a_k$ in equation (\ref{expansion}) are determined
864: via the four-term recurrence relation (it's just a matter of substituting
865: expression (\ref{expansion}) in the wave equation (\ref{eveq})), given by
866: %
867: \begin{eqnarray}
868: &&\alpha_0a_1+\beta_0a_0=0\,, \nonumber \\
869: &&\alpha_1a_2+\beta_1a_1+\gamma_1a_0=0\,, \\
870: &&\alpha_ka_{k+1}+\beta_ka_k+\gamma_ka_{k-1}+\delta_ka_{k-2}=0\,,
871: \ k=2,3,\cdots, \nonumber
872: \end{eqnarray}
873: %
874: where
875: %
876: \begin{eqnarray}
877: \alpha_k&=& 2(2\rho+k+1)(k+1)\,, \nonumber \\
878: \beta_k &=&-5(2\rho+k)(2\rho+k+1)-l(l+2)-{3\over 4} \nonumber \\
879: &&-{9\over 4}\,(1-j^2) \,, \nonumber \\
880: \gamma_k&=& 4(2\rho+k-1)(2\rho+k+1)+{9\over 2}\,(1-j^2)\,, \nonumber \\
881: \delta_k&=&-(2\rho+k-2)(2\rho+k+1)-{9\over 4}\,(1-j^2)\,. \nonumber
882: \end{eqnarray}
883: %
884: It is seen that since the asymptotic form of the perturbations as
885: $r_*\rightarrow\infty$ is written in terms of the variable $x$ as
886: %
887: \begin{equation}
888: e^{-i\omega r_*}=e^{-i\omega x^{-1}}(x-x_1)^{-\rho}(x+x_1)^{\rho}\,,
889: \label{expwr}
890: \end{equation}
891: %
892: the expanded perturbation function $\Psi$ defined by equation
893: (\ref{expansion}) automatically satisfy the QNM boundary conditions
894: (\ref{def-bc}) if the power series converges for $0\le x\le x_1$.
895: Making use of a Gaussian elimination \cite{Le90}, we can reduce the
896: four-term recurrence relation to the three-term one, given by
897: %
898: \begin{eqnarray}
899: &&\alpha'_0a_1+\beta'_0a_0=0\,, \nonumber \\
900: &&\alpha'_ka_{k+1}+\beta'_ka_k+\gamma'_ka_{k-1}=0\,,
901: \ k=1,2,\cdots,
902: \end{eqnarray}
903: %
904: where $\alpha'_k$, $\beta'_k$, and $\gamma'_k$ are given in terms of
905: $\alpha_k$, $\beta_k$, $\gamma_k$ and $\delta_k$ by
906: %
907: \begin{eqnarray}
908: \alpha'_k=\alpha_k,\quad \beta'_k=\beta_k,\quad \gamma'_k=\gamma_k,
909: \quad {\rm for\ } k=0,1,
910: \end{eqnarray}
911: %
912: and
913: %
914: \begin{eqnarray}
915: \alpha'_k&=&\alpha_k, \nonumber \\
916: \beta'_k&=&\beta_k-\alpha'_{k-1}\delta_k/\gamma'_{k-1}\,, \\
917: \gamma'_k&=&\gamma_k-\beta'_{k-1}\delta_k/\gamma'_{k-1}\,,\quad
918: {\rm for\ }k\ge 2\,. \nonumber
919: \end{eqnarray}
920: %
921: Now that we have the three-term recurrence relation for determining the
922: expansion coefficients $a_k$, the
923: convergence condition for the expansion (\ref{expansion}), namely the
924: QNM conditions, can be written in terms of the continued fraction
925: as \cite{Gu67,Le85}
926: %
927: \begin{eqnarray}
928: \beta'_0-{\alpha'_0\gamma'_1\over\beta'_1-}
929: {\alpha'_1\gamma'_2\over\beta'_2-}{\alpha'_2\gamma'_3\over\beta'_3-}
930: ...\equiv
931: \beta'_0-\frac{\alpha'_0\gamma'_1}{\beta'_1-
932: \frac{\alpha'_1\gamma'_2}{\beta'_2-\frac{\alpha'_2\gamma'_3}{\beta'_3-...}}}
933: =0 \,,
934: \label{a-eq}
935: \end{eqnarray}
936: %
937: where the first equality is a notational definition commonly
938: used in the literature for infinite continued
939: fractions.
940: Here we shall adopt such a convention.
941: In order to use Nollert's method, with which relatively high-order
942: QNM with large imaginary frequencies can be obtained, we have to know
943: the asymptotic behaviors of $a_{k+1}/a_k$ in the limit of
944: $k\rightarrow\infty$. According to a similar consideration as that by
945: Leaver \cite{Le90}, it is found that
946: %
947: \begin{eqnarray}
948: {a_{k+1}\over a_k}=1\pm2\sqrt{\rho}\,k^{-1/2}+
949: \left(2\rho-{3\over 4}\right)k^{-1}+\cdots\,,
950: \label{exp-R}
951: \end{eqnarray}
952: %
953: where the sign for the second term in the right-hand side is chosen so as
954: to be
955: %
956: \begin{eqnarray}
957: {\rm Re}(\pm2\sqrt{\rho}) < 0\,.
958: \end{eqnarray}
959: %
960: In actual numerical computations, it is convenient to solve the $k$-th
961: inversion of the continue fraction equation (\ref{a-eq}), given by
962: %
963: \begin{eqnarray}
964: &&\beta'_k-{\alpha'_{k-1}\gamma'_k\over\beta'_{k-1}-}
965: {\alpha'_{k-2}\gamma'_{k-1}\over\beta'_{k-2}-}\cdots
966: {\alpha'_0\gamma'_1\over\beta'_0} \nonumber \\
967: &=& {\alpha'_{k}\gamma'_{k+1}\over\beta'_{k+1}-}
968: {\alpha'_{k+1}\gamma'_{k+2}\over\beta'_{k+2}-}\cdots\,. \quad
969: \label{a-eq2}
970: \end{eqnarray}
971: %
972: The asymptotic form (\ref{exp-R}) plays an important role in Nollert's
973: method when the infinite continued fraction in the right-hand side of
974: equation (\ref{a-eq2}) is evaluated \cite{nollert}.
975:
976:
977: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
978: \subsubsection{Numerical results}
979: \label{numericalresults}
980: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
981: Using the numerical technique described in section \ref{numtechnique}
982: we have extracted the 5000 lowest lying QN frequencies for the five
983: dimensional Schwarzschild black hole, in the case of scalar, gravitational
984: tensor
985: and gravitational vector perturbations.
986: For electromagnetic vector perturbations the situation is different:
987: the real part rapidly approaches zero, and we have not been able to compute
988: more than the first $30$ modes.
989:
990: \bigskip\medskip
991: \centerline {\it (i) Low-lying modes}
992: \medskip
993:
994: Low lying modes are important since they govern the intermediate-time
995: evolution of any black hole perturbation. As such they may play a role
996: in TeV-scale gravity scenarios \cite{hamed}, and higher dimensional
997: black hole formation \cite{vitoroscarjose,bhprod}. For example, it is
998: known \cite{vitoroscarjose,cardosorad,bertihighd} that if one forms black holes
999: through the high energy collision of particles, then the fundamental
1000: quasinormal frequencies serve effectively as a cuttof in the energy
1001: spectra of the gravitational energy radiated away.
1002: In Tables \ref{tab:qnf5dj0}-\ref{tab:qnf5dj23} we list the five lowest lying
1003: QN frequencies for some values of the multipole index $l$.
1004: The fundamental modes are in excellent agreement with the ones presented
1005: by Konoplya \cite{konoplyawkb,roman2} using a high-order WKB approach.
1006: Notice he uses a different convention so one has to be careful when
1007: comparing the results. For example, in our units Konoplya obtains for
1008: scalar and tensorial perturbations ($j=0$) with $l=2$ a fundamental
1009: QN frequency $\omega _{0}/T_H=9.49089+2.24721i$, whereas we get, from
1010: Table \ref{tab:qnf5dj0} the number $9.4914+2.2462i$ so the WKB approach
1011: does indeed yield good results, at least for low-lying modes, since it
1012: is known it fails for high-order ones.
1013:
1014:
1015: \begin{table}
1016: \caption{\label{tab:qnf5dj0} The first lowest QN frequencies
1017: for scalar and gravitational tensor ($j=0$) perturbations
1018: of the five dimensional Schwarzschild black hole. The frequencies
1019: are normalized in units of black hole temperature, so the Table
1020: really shows $\frac{\omega}{T_H}$,
1021: where $T_H$ is the Hawking temperature of the black hole. }
1022: \begin{ruledtabular}
1023: \begin{tabular}{l|lll} \hline
1024: \multicolumn{1}{c}{} &
1025: \multicolumn{3}{c}{ $j=0$}\\ \hline
1026: $n$ &$l=0$: &$l=1$: &$l=2$\\ \hline
1027: 0&3.3539+2.4089i &6.3837+2.2764i &9.4914+2.2462i \\ \hline
1028: 1&2.3367+8.3101i &5.3809+7.2734i &8.7506+6.9404i \\ \hline
1029: 2&1.8868+14.786i &4.1683+13.252i &7.5009+12.225i \\ \hline
1030: 3&1.6927+21.219i &3.4011+19.708i &6.2479+18.214i \\ \hline
1031: 4&1.5839+27.607i &2.9544+26.215i &5.3149+24.597i \\ \hline
1032: \end{tabular}
1033: \end{ruledtabular}
1034: \end{table}
1035:
1036:
1037:
1038:
1039: \begin{table}
1040: \caption{\label{tab:qnf5dj2} The first lowest QN frequencies
1041: for gravitational vector ($j=2$) perturbations
1042: of the five dimensional Schwarzschild black hole. The frequencies
1043: are normalized in units of black hole temperature, so the Table
1044: really shows $\frac{\omega}{T_H}$,
1045: where $T_H$ is the Hawking temperature of the black hole. }
1046: \begin{ruledtabular}
1047: \begin{tabular}{l|lll} \hline
1048: \multicolumn{1}{c}{} &
1049: \multicolumn{3}{c}{ $j=2$}\\ \hline
1050: $n$ &$l=2$: &$l=3$: &$l=4$\\ \hline
1051: 0&7.1251+2.0579i &10.8408+2.0976i &14.3287+2.1364i \\ \hline
1052: 1&5.9528+6.4217i &8.8819+1.0929i &12.8506+1.0574i \\ \hline
1053: 2&3.4113+12.0929i &7.2160+16.3979i &11.5069+16.0274i \\ \hline
1054: 3&2.7375+19.6094i &5.5278+22.3321i &9.99894+21.5009i \\ \hline
1055: 4&2.5106+26.2625i &3.9426+28.6569i &8.56386+27.4339i \\ \hline
1056: \end{tabular}
1057: \end{ruledtabular}
1058: \end{table}
1059: \vskip 1mm
1060:
1061: \vskip 1mm
1062: \begin{table}
1063: \caption{\label{tab:qnf5dj23} The first lowest QN frequencies
1064: for electromagnetic vector ($j=2/3$) perturbations
1065: of the five dimensional Schwarzschild black hole. The frequencies
1066: are normalized in units of black hole temperature, so the Table
1067: really shows $\frac{\omega}{T_H}$,
1068: where $T_H$ is the Hawking temperature of the black hole. }
1069: \begin{ruledtabular}
1070: \begin{tabular}{l|lll} \hline
1071: \multicolumn{1}{c}{} &
1072: \multicolumn{3}{c}{ $j=2/3$}\\ \hline
1073: $n$ &$l=1$: &$l=2$: &$l=3$\\ \hline
1074: 0&5.9862+2.2038i &9.2271+2.2144i &12.4184+2.2177i\\ \hline
1075: 1&4.9360+7.0676i &8.4728+6.8459i &11.8412+6.7660i \\ \hline
1076: 2&3.6588+12.9581i&7.1966+12.0804i&10.7694+11.6626i\\ \hline
1077: 3&2.8362+19.3422i&5.9190+18.0343i&9.4316+17.1019i \\ \hline
1078: 4&2.3322+25.7861i&4.9682+24.3832i&8.1521+23.0769i \\ \hline
1079: \end{tabular}
1080: \end{ruledtabular}
1081: \end{table}
1082:
1083:
1084: In the four dimensional case, and for gravitational vector
1085: perturbations, there is for is each $l$, a purely
1086: imaginary QN frequency, which had been coined an ``algebraically
1087: special frequency'' by Chandrasekhar \cite{chandra}. For further
1088: properties of this special frequencies we refer the reader to
1089: \cite{brinkspecial}. The existence of these purely imaginary
1090: frequencies translates, in the four dimensional case, a relation
1091: between the two gravitational wavefunctions (i.e., between the
1092: Regge-Wheeler and the Zerilli wavefunction, or between the
1093: gravitational vector and gravitational scalar wavefunctions,
1094: respectively). It is
1095: possible to show for example that the associated potentials are
1096: related through supersymmetry. Among other consequences, this
1097: relation allows one to prove that the QN frequencies of both
1098: potentials are exactly the same \cite{chandra}.
1099:
1100: We have not spotted any purely imaginary QN frequency for this
1101: five-dimensional black hole. In fact the QN frequency with the
1102: lowest real part is a gravitational vector QN frequency with $l=4$ and
1103: overtone number $n=13$, $\omega=8.7560\times 10^{-2}+6.9934i$. This
1104: may indicate that in five dimensions, there is no relation between the
1105: wavefunctions, or put another way, that the potentials are no longer
1106: superpartner potentials. This was in fact already observed by Kodama
1107: and Ishibashi \cite{kodama} for any dimension greater than four. It
1108: translates also in Tables \ref{tab:qnf5dj0}-\ref{tab:qnf5dj2},
1109: which yield different values for the QN frequencies. Although we have not
1110: worked out the gravitational scalar QN frequencies, a WKB approach
1111: can be used for the low-lying QN frequencies, and also yields different values
1112: \cite{konoplyawkb,roman2}.
1113:
1114:
1115:
1116:
1117:
1118:
1119:
1120:
1121:
1122:
1123: \bigskip\medskip
1124: \centerline {\it (ii) Highly damped modes}
1125: \medskip
1126:
1127: Our numerical results for the highly damped modes, i.e., QN
1128: frequencies with a very large imaginary part, are summarized in
1129: Figures \ref{fig:qnall5d}-\ref{fig:qnvector5d} and in Tables
1130: \ref{tab:corrj0}-\ref{tab:corrj2}.
1131:
1132:
1133: \begin{figure}
1134: \centerline{\includegraphics[width=7 cm,height=7 cm]
1135: {qnfalld5.eps}}
1136: \caption{The QN frequencies of scalar ($j=0$), gravitational vector ($j=2$)
1137: and electromagnetic vector ($j=2/3$) perturbations
1138: of a five-dimensional Schwarzschild black hole. We only show the
1139: lowest radiatable multipoles,
1140: but the asymtptoic behaviour is $l$-independent.}
1141: \label{fig:qnall5d}
1142: \end{figure}
1143: \vskip 1mm
1144:
1145: In Fig. \ref{fig:qnall5d} we show our results for scalar,
1146: gravitational tensor ($j=0$), gravitational vector ($j=2$) and
1147: electromagnetic vector ($j=2/3$) QN frequencies. For $j=0\,,2$ the
1148: real part approaches $\ln{3}$ in units of the black hole temperature, a
1149: result consistent with the analytical result in \cite{motl2} (see also
1150: section \ref{apendice}). Electromagnetic vector ($j=2/3$) QN
1151: frequencies behave differently: the real part rapidly approaches zero,
1152: and this makes it very difficult to compute the higher modes.
1153: In fact, the same kind of problem appears in the Kerr geometry \cite{qnmkerr},
1154: and this a major obstacle to a definitive numerical characterization of the
1155: highly damped QNMs in this geometry.
1156: We have
1157: only been able to compute the first $30$ modes with accuracy. The
1158: prediction of \cite{motl2} for this case (see section \ref{apendice})
1159: is that there are no asymptotic QN frequencies.
1160: It is left as an open question whether this is true or not.
1161: Our results indicate that the real part rapidly approaches zero, but we
1162: cannot say whether the modes die there or not, or even if they perform
1163: some kind of oscillation. The imaginary part behaves in the usal manner,
1164: growing linearly with mode number.
1165:
1166:
1167:
1168: Let us now take a more detailed look at our numerical data for
1169: scalar and gravitational QN frequencies.
1170: In Fig. \ref{fig:qntensor5d} we show in a $\ln$ plot
1171: our results for scalar and gravitational tensor ($j=0$) QN frequencies.
1172: \begin{figure}
1173: \centerline{\includegraphics[width=7 cm,height=7 cm]
1174: {qnftensord5.eps}}
1175: \caption{The QN frequencies of
1176: scalar and gravitational tensor perturbations ($j=0$)
1177: of a five-dimensional Schwarzschild black hole.
1178: Notice that for frequencies with a large imaginary part
1179: the slope is $l$-independent and equals $3/2$.}
1180: \label{fig:qntensor5d}
1181: \end{figure}
1182: \vskip 1mm
1183: \begin{table}
1184: \caption{\label{tab:corrj0} The correction coefficients for the
1185: five-dimensional Schwarzschild black hole, both numerical, here
1186: labeled as ``${\rm Corr_{5}^{N}}$'' and analytical, labeled as ``${\rm
1187: Corr_{5}^{A}}$''. These results refer to scalar or gravitational tensor
1188: perturbations ($j=0$).
1189: The analytical results are extracted from the
1190: analytical formula (\ref{corrd5j0}). Notice the very good agreement, to
1191: within 0.5\% or less, between the numerically extracted results and the
1192: analytical prediction.}
1193: \begin{ruledtabular}
1194: \begin{tabular}{l|ll} \hline
1195: \multicolumn{1}{c}{} &
1196: \multicolumn{2}{c}{ $j=0$}\\ \hline
1197: $l$ &${\rm Corr_5^N}$:&${\rm Corr_5^A}$\\ \hline
1198: 0 & 1.155 - 1.993i & 1.15404-1.99886i \\ \hline
1199: 1 & 4.058 - 6.956i & 4.03915-6.99601i \\ \hline
1200: 2 & 8.894 - 15.26i & 8.84765-15.3246i \\ \hline
1201: 3 & 15.64 - 26.79i & 15.5796-26.9847i \\ \hline
1202: 4 & 24.33 - 41.67i & 24.2349-41.9761i \\ \hline
1203: \end{tabular}
1204: \end{ruledtabular}
1205: \end{table}
1206: \vskip 1mm
1207:
1208: Our numerical results are very clear: asymptotically the QN
1209: frequencies for scalar and gravitational tensor perturbations $(j=0)$ behave
1210: as
1211: \begin{equation}
1212: \frac{\omega}{T_H}=\ln{3}+i\,(2n+1)\pi+\frac{{\rm Corr_5}}{n^{2/3}}\,\,,
1213: \label{asympt5dScalar}
1214: \end{equation}
1215: So the leading terms $\ln{3}+(2n+1)i$ are indeed the ones predicted
1216: in \cite{motl2}. Interestingly, the first corrections do not appear as
1217: $\frac{1}{\sqrt{n}}$, but as $\frac{1}{n^{2/3}}$. This was shown to be
1218: the expected analytical result in section \ref{apendice}, where we
1219: generalized Musiri and Siopsis' \cite{musiri2} results to higher dimensions.
1220: In table
1221: \ref{tab:corrj0} we show the coefficient ${\rm Corr_5}$ extracted
1222: numerically along with the predicted coefficient (see expression
1223: (\ref{corrd5j0})). The table is very clear: the numerical values
1224: match the analytical ones. Another confirmation that the corrections
1225: appear as $\frac{1}{n^{2/3}}$ is provided by
1226: Fig. \ref{fig:qntensor5d}. In this figure the QN frequencies are plotted
1227: in a $\ln$ plot for an easier interpretation. One sees that for large
1228: imaginary parts of the QN frequencies, the slope of the plot is
1229: approximately $-3/2$, as it should be if the corrections are of the
1230: order $\frac{1}{n^{2/3}}$.
1231:
1232: In Fig. \ref{fig:qnvector5d} we show in a $\ln$ plot our results for
1233: gravitational vector ($j=2$) QN frequencies. Again, vector QN
1234: frequencies have the asymptotic behaviour given by expression
1235: (\ref{asympt5dScalar}), with a different correction term ${\rm
1236: Corr_5}$. Again, the $\frac{1}{n^{2/3}}$ corrections show themselves
1237: in the $\ln$ plot of Fig. \ref{fig:qnvector5d}: for very large imaginary
1238: parts, the slope is $-3/2$, as it should. The numerically extracted
1239: coefficient ${\rm Corr_5}$ for $j=2$ perturbations is listed in Table
1240: \ref{tab:corrj2}, along with the analytically predicted value (see
1241: section \ref{apendice}). the agreement is very good.
1242: \begin{figure}
1243: \centerline{\includegraphics[width=7 cm,height=7 cm]
1244: {qnfvectord5.eps}}
1245: \caption{The QN frequencies of gravitational vector perturbations
1246: ($j=2$) of a five-dimensional Schwarzschild black hole.}
1247: \label{fig:qnvector5d}
1248: \end{figure}
1249: \vskip 1mm
1250: \begin{table}
1251: \caption{\label{tab:corrj2} The correction coefficients for the
1252: five-dimensional Schwarzschild black hole, both numerical, here
1253: labeled as ``${\rm Corr_{5}^{N}}$'' and analytical, labeled as ``${\rm
1254: Corr_{5}^{A}}$'', for gravitational vector perturbations ($j=2$).
1255: The analytical results are extracted from the
1256: analytical formula (\ref{corrd5j2}). Notice the very good agreement, to
1257: within 0.5\% or less, between the numerically extracted results and the
1258: analytical prediction.}
1259: \begin{ruledtabular}
1260: \begin{tabular}{l|ll} \hline
1261: \multicolumn{1}{c}{} &
1262: \multicolumn{2}{c}{ $j=2$}\\ \hline
1263: $l$ &${\rm Corr_5^N}$:&${\rm Corr_5^A}$:\\ \hline
1264: 2 & 3.568 - 6.139i & 3.5583 -6.16316i \\ \hline
1265: 3 & 6.947 - 11.94i & 6.92425-11.9932i \\ \hline
1266: 4 & 11.29 - 19.38 i & 11.2519-19.4889i \\ \hline
1267: \end{tabular}
1268: \end{ruledtabular}
1269: \end{table}
1270: \vskip 1mm
1271:
1272: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1273: \section{Discussion of results and future directions}
1274: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1275: We have made an extensive survey of the QNMs of the four and five
1276: dimensional Schwarzschild black hole. The investigation presented
1277: here makes a more complete characterization of the highly damped QNMs in the
1278: Schwarzschild geometry. In the four-dimensional case, we confirmed
1279: previous numerical results regarding the scalar and gravitational
1280: asymptotic QN frequencies. We found that both the leading behaviour
1281: and the first order corrections for the scalar and gravitational
1282: perturbations agree extremely well with existing analytical formulas. We
1283: have presented new numerical results concerning electromagnetic QN frequencies
1284: of the four-dimensional Schwarzschild geometry. In particular, this is
1285: the first work dealing with highly damped electromagnetic QNMs.
1286: Again, we find that the leading behaviour and the first order
1287: corrections agree with the analytical calculations. The first order
1288: corrections appear at the $\frac{1}{n^{3/2}}$ level.
1289: In the five-dimensional case this represents the first study on highly
1290: damped QNMs. We have seen numerically that the asymptotic behaviour
1291: is very well described by $\frac{\omega}{T_H} = \ln{3}+\pi i(2n+1)$ for
1292: scalar and gravitational perturbations, and agrees with the predicted
1293: formula. Moreover the first order corrections appear at the level
1294: $\frac{1}{n^{2/3}}$, which is also very well described by the
1295: analytical calculations, providing one more consistency check on the
1296: theoretical framework.
1297: In generic $D$ dimensions the corrections appear at the
1298: $\frac{1}{n^{(D-3)/(D-2)}}$ level.
1299: We have not been able to prove or disprove the analytic result
1300: for electromagnetic QN frequencies in five dimensions ($j=2/3$).
1301: Other conclusions that can be taken from
1302: our work are: the monodromy method by Motl and Neitzke is correct.
1303: It is important to test it numerically, as we did, since there are some
1304: ambiguous
1305: assumptions in that method. Moreover, their method is highly flexible,
1306: since it allows an
1307: easy computation of the correction terms, as shown by Musiri and
1308: Siopsis, and generalized here for the higher dimensional Schwarzschild
1309: black hole.
1310: We have basically showed numerically that the monodromy method
1311: by Motl and Neitzke \cite{motl2}, and its extension by Musiri and
1312: Siopsis \cite{musiri2} to include for correction terms in overtone
1313: number, are excellent techniques to investigate the highly damped QNMs
1314: of black holes.
1315:
1316: Highly damped QNMs of black holes are the bed rocks that the recent
1317: conjectures
1318: \cite{hod,dreyer,ling} are built on, relating these to black hole area
1319: quantization. However, to put on solid ground the conjecture that
1320: QNMs can actually be of any use to black hole area quantization, one
1321: has to do much better then one has been able to do up to now. In
1322: particular, it is crucial to have a deeper understanding of the
1323: Kerr and Reissner-Nordstr\"om QNMs.
1324: These are, next to the Schwarzschild, the simpler asymptotically
1325: flat geometries. If these conjectures hold, they must to do also
1326: for these spacetimes.
1327: Despite the fact that there are convincing numerical results for the Kerr
1328: geometry
1329: \cite{qnmkerr}, there seems to be for the moment no serious analytical
1330: investigation of the highly damped QNMs for this spacetime. Moreover,
1331: even though we are in possession of an analytical formula for the
1332: highly damped QNMs of the Reissner-Nordstr\"om geometry \cite{motl2},
1333: and this has been numerically tested already \cite{bertirn}, we have
1334: no idea what it means! In particular, can one use it to quantize the
1335: black hole area? In this case the electromagnetic perturbations
1336: are coupled to the gravitational ones. How do we proceed with the
1337: analysis, that was so simple in the Schwarzschild geometry?
1338: Is the assumption of equally spaced eigenvalues a correct one, in this case?
1339: Can these ideas on black hole area quantization be translated to
1340: non-asymptotically
1341: flat spacetimes, as the
1342: de Sitter or anti-de Sitter spacetime?
1343: These are the fundamental issues that remain to be solved in this
1344: field. Should any of them be solved satisfactorily, then these
1345: conjectures would gain a whole new meaning. As they stand, it may just
1346: be a numerical coincidence that the real part of the QN frequency goes
1347: to $\ln{3}$.
1348:
1349:
1350:
1351:
1352:
1353:
1354:
1355:
1356:
1357:
1358:
1359:
1360:
1361:
1362: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1363: \section*{Acknowledgements}
1364: The authors would like to thank Emanuele Berti and Alejandro Corichi for very useful
1365: comments and suggestions. The authors thank also Ricardo Schiappa and Jos\'e Nat\'ario
1366: for very useful discussions, and are specially indebted to Roman
1367: Konoplya for sharing his numerical results prior to publication.
1368: This work was partially funded by Funda\c c\~ao para a Ci\^encia e
1369: Tecnologia (FCT) - Portugal through project PESO/PRO/2000/4014
1370: and CERN/FIS/43797/2001. SY
1371: acknowledges finantial support from FCT through project SAPIENS
1372: 36280/99. VC also acknowledge finantial support from FCT
1373: through PRAXIS XXI programme. JPSL
1374: thanks Observat\'orio Nacional do Rio de Janeiro
1375: for hospitality.
1376:
1377: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1378:
1379: \begin{thebibliography}{99}
1380:
1381: \bibitem{vish} C. V. Vishveshwara,
1382: Nature {\bf 227}, 936 (1970).
1383:
1384: \bibitem{echeverria} F. Echeverria,
1385: Phys. Rev. D {\bf 40}, 3194 (1989);
1386: L. S. Finn,
1387: Phys. Rev. D {\bf 46}, 5236 (1992).
1388:
1389: \bibitem{nakanoringing} H. Nakano, H. Takahashi, H. Tagoshi, M. Sasaki,
1390: gr-qc/0306082.
1391:
1392: \bibitem{kokkotas} K.D. Kokkotas, B.G. Schmidt,
1393: Living Rev. Relativ. {\bf 2}, 2 (1999).
1394:
1395: \bibitem{nollert} H.-P. Nollert,
1396: Class. Quant. Grav. {\bf 16}, R159 (1999).
1397:
1398: \bibitem{horowitz} G. T. Horowitz, and V. Hubeny,
1399: Phys. Rev. D {\bf 62}, 024027(2000).
1400:
1401: \bibitem{maldacena} J. M. Maldacena,
1402: Adv. Theor. Math. Phys. {\bf 2}, 253 (1998);
1403: E. Witten, Adv. Theor. Math. Phys.
1404: {\bf 2}, 505 (1998).
1405:
1406: \bibitem{birmingham2} D. Birmingham, I. Sachs and S. N. Solodukhin,
1407: Phys. Rev. Lett. {\bf 88}, 151301 (2002).
1408:
1409: \bibitem{cardosoqnmbtz} V. Cardoso and J. P. S. Lemos,
1410: Phys. Rev. D {\bf 63}, 124015 (2001).
1411:
1412: \bibitem{kurita} Y. Kurita and M. Sakagami,
1413: Phys. Rev. D {\bf 67}, 024003 (2003).
1414:
1415: \bibitem{qnmads} V. Cardoso and J. P. S. Lemos,
1416: Phys. Rev. D {\bf 64}, 084017 (2001);
1417: V. Cardoso and J. P. S. Lemos,
1418: Class. Quant. Grav. {\bf 18}, 5257 (2001);
1419: E. Berti and K. D. Kokkotas,
1420: Phys. Rev. D {\bf 67}, 064020 (2003);
1421: B. Wang, C. Y. Lin, and E. Abdalla,
1422: Phys. Lett. B {\bf481}, 79 (2000);
1423: B. Wang, C. M. Mendes, and E. Abdalla,
1424: Phys. Rev. D {\bf 63}, 084001 (2001);
1425: J. Zhu, B. Wang, and E. Abdalla,
1426: Phys. Rev. D {\bf 63}, 124004(2001);
1427: B. Wang, E. Abdalla and R. B. Mann,
1428: Phys.Rev. D {\bf 65}, 084006 (2002).
1429: W. T. Kim, J. J. Oh, Phys. Lett. B {\bf 514}, 155 (2001);
1430: S. F. J. Chan and R. B. Mann,
1431: Phys. Rev. D {\bf 55}, 7546 (1997);
1432: A. O. Starinets, Phys. Rev. D {\bf 66}, 124013 (2002);
1433: R. Aros, C. Martinez, R. Troncoso and J. Zanelli,
1434: Phys. Rev. D {\bf 67}, 044014 (2003);
1435: D. T. Son, A. O. Starinets,
1436: J.H.E.P. {\bf 0209}:042, (2002);
1437: T. R. Govindarajan, and V. Suneeta, Class. Quantum
1438: Grav. {\bf 18}, 265 (2001);
1439: S. Musiri and G. Siopsis,
1440: Phys. Lett. B {\bf 563}, 102 (2003);
1441: S. Fernando, hep-th/0306214.
1442:
1443: \bibitem{hod} S. Hod,
1444: Phys. Rev. Lett. {\bf 81}, 4293 (1998).
1445:
1446: \bibitem{dreyer} O. Dreyer,
1447: Phys. Rev. Lett. {\bf 90}, 081301 (2003).
1448:
1449: \bibitem{oppenheim} J. Oppenheim, gr-qc/0307089.
1450:
1451: \bibitem{corichi} A. Corichi,
1452: Phys.Rev. D {\bf 67}, 087502 (2003).
1453:
1454: \bibitem{ling} Y. Ling, H. Zhang, gr-qc/0309018.
1455:
1456: \bibitem{bek0} J. Bekenstein,
1457: Lett. Nuovo Cimento {\bf 11}, 467 (1974).
1458:
1459: \bibitem{bek1} J. Bekenstein and V. F. Mukhanov,
1460: Phys. Lett. B {\bf 360}, 7 (1995).
1461:
1462: \bibitem{bek2} J. Bekenstein, in
1463: {\it Cosmology and Gravitation}, edited by M. Novello
1464: (Atlasciences, France 2000), pp. 1-85, gr-qc/9808028 (1998).
1465:
1466: \bibitem{bek3} J. Bekenstein,
1467: Phys. Rev. D {\bf 7}, 2333 (1973).
1468:
1469: \bibitem{nollert2} H.-P. Nollert,
1470: Phys. Rev. D {\bf 47}, 5253 (1993).
1471:
1472: \bibitem{motl1} L. Motl,
1473: Adv. Theor. Math. Phys. {\bf 6}, 1135 (2003).
1474:
1475: \bibitem{motl2} L. Motl, A. Neitzke,
1476: Adv. Theor. Math. Phys. {\bf 7}, 2 (2003).
1477:
1478: \bibitem{kunstatter} G. Kunstatter,
1479: Phys. Rev. Lett. {\bf 90}, 161301 (2003).
1480:
1481: \bibitem{bertirn} E. Berti and K. D. Kokkotas,
1482: Phys. Rev. D {\bf 68}, 044027 (2003).
1483:
1484: \bibitem{brinkasympt} A. M. van den Brink, gr-qc/0303095.
1485:
1486: \bibitem{karlucio} K. H. C. Castello-Branco and E. Abdalla, gr-qc/0309090.
1487:
1488: \bibitem{ricardo} R. Schiappa, J. Nat\'ario and V. Cardoso,
1489: in preparation.
1490:
1491: \bibitem{musiri2} S. Musiri and G. Siopsis,
1492: hep-th/0308168.
1493:
1494: \bibitem{qnmkerr} E. Berti, V. Cardoso, K. D. Kokkotas, H. Onozawa
1495: hep-th/0307013.
1496:
1497: \bibitem{asymptall} V. Cardoso and J. P. S. Lemos,
1498: Phys. Rev. D {\bf 67}, 084020 (2003);
1499: C. Molina,
1500: Phys. Rev. D {\bf 68}, 064007 (2003);
1501: A. M. van den Brink,
1502: Phys. Rev. D {\bf 68}, 047501 (2003);
1503: V. Cardoso, R. Konoplya, J. P. S. Lemos,
1504: Phys. Rev. D {\bf 68}, 044024 (2003);
1505: D. Birmingham, S. Carlip and Y. Chen,
1506: hep-th/0305113;
1507: D. Birmingham,
1508: hep-th/0306004 (2003);
1509: N. Andersson, C. J. Howls,
1510: gr-qc/0307020;
1511: S. Yoshida and T. Futamase,
1512: gr-qc/0308077.
1513: S. Musiri and G. Siopsis, hep-th/0308196;
1514: E. Berti, K. D. Kokkotas, E. Papantonopoulos, gr-qc/0306106.
1515:
1516: \bibitem{tails} R. H. Price,
1517: Phys. Rev. D{\bf 5}, 2419 (1972);
1518: E. S. C. Ching, P. T. Leung, W. M. Suen and K. Young,
1519: Phys. Rev. Lett. {\bf 74}, 2414 (1995);
1520: E. S. C. Ching, P. T. Leung, W. M. Suen and K. Young,
1521: Phys. Rev. D{\bf 52}, 2118 (1995);
1522: V. Cardoso, S. Yoshida, O. J. C. Dias, J. P. S. Lemos,
1523: Phys. Rev. D (in press), hep-th/0307122.
1524:
1525: \bibitem{chandra} S. Chandrasekhar,
1526: in {\it The Mathematical Theory of Black Holes}
1527: (Oxford University, New York, 1983).
1528:
1529: \bibitem{vitoroscarjose} V. Cardoso, O. J. C. Dias and J. P. S. Lemos,
1530: Phys. Rev. D {\bf 67}, 064026 (2003).
1531:
1532: \bibitem{konoplyawkb} R. A. Konoplya,
1533: Phys. Rev. D {\bf 68}, 024018 (2003).
1534:
1535: \bibitem{roman2} R. A. Konoplya, hep-th/0309030.
1536:
1537: \bibitem{ida} D. Ida, Y. Uchida and Y. Morisawa,
1538: Phys. Rev. D {\bf 67}, 084019 (2003).
1539:
1540: \bibitem{chandradet} S. Chandrasekhar, and S. Detweiler,
1541: Proc. R. Soc. London A {\bf 344}, 441 (1975).
1542:
1543: \bibitem{Le85} E. W. Leaver, Proc. R. Soc. London {\bf A402}, 285 (1985).
1544:
1545: \bibitem{guinn} J. W. Guinn, C. M. Will, Y. Kojima and B. F. Schutz,
1546: Class. Quant. Grav. {\bf 7}, L47 (1990).
1547:
1548: \bibitem{tangherlini} F. R. Tangherlini,
1549: Nuovo Cim. {\bf 27}, 636 (1963).
1550:
1551: \bibitem{myersperry} R. C. Myers and M. J. Perry,
1552: Annals Phys. {\bf 172}, 304 (1986).
1553:
1554: \bibitem{cardosoradadsddim} V. Cardoso and J. P. S. Lemos,
1555: Phys. Rev. D {\bf 66}, 064006 (2002).
1556:
1557: \bibitem{kodama} H. Kodama, A. Ishibashi,
1558: hep-th/0305147; hep-th/0305185; hep-th/0308128.
1559:
1560: \bibitem{regge} T. Regge, J. A. Wheeler, Phys. Rev. {\bf 108},
1561: 1063 (1957).
1562:
1563: \bibitem{zerilli} F. Zerilli,
1564: Phys. Rev. Lett. {\bf 24}, 737 (1970).
1565:
1566: \bibitem{matsas} L. Crispino, A. Higuchi and G. Matsas,
1567: Phys. Rev. D {\bf 63}, 124008 (2001).
1568:
1569: \bibitem{andy} A. Neitzke, hep-th/0304080.
1570:
1571: \bibitem{Le90} E. W. Leaver, Phys. Rev. D {\bf 41}, 2986 (1990).
1572:
1573: \bibitem{Gu67} W. Gautschi, SIAM Rev. {\bf 9}, 24 (1967).
1574:
1575: \bibitem{hamed} N. Arkani-Hamed, S. Dimopoulos and G. Dvali,
1576: Phys. Lett. B {\bf 429}, 263 (1998); Phys. Rev. D {\bf 59}, 086004 (1999);
1577: I. Antoniadis, N. Arkani-Hamed, S. Dimopoulos and G. Dvali,
1578: Phys. Lett. B {\bf 436}, 257 (1998).
1579:
1580: \bibitem{bhprod} P. C. Argyres, S. Dimopoulos and J. March-Russell,
1581: Phys. Lett. B {\bf 441}, 96 (1998);
1582: S. Dimopoulos and G. Landsberg,
1583: Phys. Rev. Lett. {\bf 87}, 161602 (2001);
1584: S. B. Giddings and S. Thomas,
1585: Phys. Rev. D {\bf 65}, 056010 (2002);
1586: S. D. H. Hsu,
1587: Phys. Lett. B {\bf 555}, 92 (2003);
1588: H. Tu, hep-ph/0205024;
1589: A. Jevicki and J. Thaler, Phys. Rev. D {\bf 66}, 024041 (2002);
1590: Y. Uehara,
1591: hep-ph/0205199;
1592: V. Cardoso, J. P. S. Lemos and S. Yoshida,
1593: Phys. Rev. D, in press (2003); gr-qc/0307104;
1594: M. Cavaglia,
1595: Int. J. Mod. Phys. A {\bf 18}, 1843 (2003);
1596: P. Kanti and J. March-Russell,
1597: Phys. Rev. D {\bf 66}, 024023 (2002).
1598: P. Kanti and J. March-Russell, hep-ph/0212199.
1599: D. Ida, Kin-ya Oda and S. C. Park, hep-th/0212108;
1600: V. Frolov and D. Stojkovic,
1601: Phys. Rev. Lett. {\bf 89}, 151302 (2002);
1602: gr-qc/0211055; C. M. Harris and Panagiota Kanti,
1603: hep-ph/0309054;
1604: L. A. Anchordoqui, J. L. Feng, H. Goldberg, A. D.
1605: Shapere, hep-ph/0307228;
1606: L. A. Anchordoqui, J. L. Feng, H. Goldberg, A. D. Shapere,
1607: Phys. Rev. D {\bf 65}, 124027 (2002);
1608: J. L. Feng, A. D. Shapere,
1609: Phys. Rev. Lett. {\bf 88}, 021303 (2002).
1610:
1611: \bibitem{cardosorad} V. Cardoso and J. P. S. Lemos,
1612: Phys. Lett. B {\bf 538}, 1 (2002);
1613: V. Cardoso and J. P. S. Lemos,
1614: Gen. Rel. Grav. {\bf 35}, L327 (2003);
1615: V. Cardoso and J. P. S. Lemos,
1616: Phys. Rev. D {\bf 67}, 084005 (2003).
1617:
1618: \bibitem{bertihighd} E. Berti, M. Cavaglia and L. Gualtieri, hep-th/0309203.
1619:
1620: \bibitem{brinkspecial} P. T. Leung, A. Maassen van den Brink, K. W. Mak,
1621: K. Young,
1622: gr-qc/0301018 (2003);
1623: A. Maassen van den Brink,
1624: Phys. Rev. D {\bf 62}, 064009 (2000).
1625:
1626:
1627:
1628:
1629:
1630:
1631:
1632:
1633:
1634:
1635:
1636:
1637: \end{thebibliography}
1638: \end{document}
1639:
1640: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1641:
1642: tar -vzcvf QNMDdim.tgz QNMDdim qnfall4d.eps qnfelecd4.eps qnftensord5.eps qnfvectord5.eps qnfalld5.eps
1643:
1644:
1645: Comments: 12 pages, 5 figures, RevTeX4.
1646: Vitor Cardoso, Jose' P. S. Lemos, Shijun Yoshida
1647: Your user/password combination for this paper is
1648: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1649: User-ID: gr-qc/0309112
1650: Password: xwjb2
1651: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1652: Title: Quasinormal modes of Schwarzschild black holes in four and...
1653: Authors: Cardoso et al.
1654: Code: DJ9059
1655: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1656: