gr-qc0308077/ms.tex
1: \documentclass[twocolumn,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
2: %\documentclass[onecolumn,showpacs,preprintnumbers,amssymb]{revtex4}
3: %\documentclass[twocolumn,showpacs,preprintnumbers,amssymb]{revtex4}
4: %\documentclass[preprint,showpacs,preprintnumbers,amssymb,nofootinbib]{revtex4}
5: 
6: \usepackage{graphicx}% Include figure files
7: \usepackage{dcolumn}% Align table columns on decimal point
8: \usepackage{bm}
9: 
10: \begin{document}
11: 
12: \title{Numerical analysis of quasinormal modes in nearly extremal Schwarzschild-de 
13: Sitter spacetimes}
14: 
15: \author{Shijun Yoshida}
16: \email{yoshida@fisica.ist.utl.pt}
17: \affiliation{
18: Centro Multidisciplinar de Astrof\'{\i}sica - CENTRA, 
19: Departamento de F\'{\i}sica, Instituto Superior T\'ecnico,
20: Av. Rovisco Pais 1, 1049-001 Lisboa, Portugal}
21: 
22: \author{Toshifumi Futamase}
23: \email{tof@astr.tohoku.ac.jp}
24: \affiliation{Astronomical Institute, Graduate School of Science,
25:            Tohoku University, Sendai 980-8578, Japan}
26: 
27: \date{\today}
28: 
29: \begin{abstract}
30: 
31: We calculate high-order quasinormal modes with large imaginary frequencies 
32: for electromagnetic and gravitational perturbations in nearly extremal 
33: Schwarzschild-de Sitter spacetimes. Our results show that for low-order quasinormal 
34: modes, the analytical approximation formula in the extremal limit derived by Cardoso 
35: and Lemos is a quite good approximation for the quasinormal frequencies as long as the 
36: model parameter $r_1\kappa_1$ is small enough, where $r_1$ and $\kappa_1$ are the 
37: black hole horizon radius and the surface gravity, respectively. For high-order 
38: quasinormal modes, to which corresponds quasinormal frequencies with large imaginary 
39: parts, on the other hand, this formula becomes inaccurate even for small 
40: values of $r_1\kappa_1$.  We also find that the real parts of the quasinormal 
41: frequencies have oscillating behaviors in the limit of highly damped modes, which are 
42: similar to those observed in the case of a Reissner-Nordstr{\" o}m black hole. 
43: The amplitude of oscillating ${\rm Re(\omega)}$ as a function of
44: ${\rm Im}(\omega)$ approaches a non-zero constant
45: value for gravitational perturbations and zero for electromagnetic
46: perturbations in the limit of highly damped modes, where $\omega$ denotes the 
47: quasinormal frequency. This means that for gravitational perturbations, the 
48: real part of quasinormal modes of the nearly extremal Schwarzschild-de Sitter 
49: spacetime appears not to approach any constant value in the limit of highly damped 
50: modes.  On the other hand, for electromagnetic perturbations, the real part of 
51: frequency seems to go to zero in the limit. 
52: 
53: \end{abstract}
54: 
55: \pacs{04.70.Bw, 04.70.-s, 04.30.-w}
56: 
57: \maketitle
58: 
59: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
60: \section{Introduction}
61: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
62: The quasinormal modes (QNMs) of spacetimes containing black holes 
63: have been studied since the pioneering work of Vishveshwara \cite{vish}, who 
64: first observed quasinormal ringing of a Schwarzschild spacetime in his 
65: numerical calculations. 
66: The main motivation to study the QNM of a black hole is twofold: 
67: One is to answer the question of whether the spacetime is stable or not, and the other 
68: to know what kind of oscillations will be excited in the spacetime when  
69: some perturbations are given. In an astrophysical point of view, the latter is 
70: quite important from the observational point of view because we could 
71: determine fundamental parameters of a black hole, such as the mass or 
72: the angular momentum, through the information of the QNMs. Thus, a large number 
73: of studies on the QNMs of spacetimes containing black holes 
74: have been done (for review, see, e.g., \cite{kokkotas99,nollert99}). 
75: 
76:  
77: There is another interesting aspect of the QNMs in a black hole which 
78: is related to the quantum theory of gravity. Bekenstein and Mukhanov 
79: discussed the relationship between the fundamental area unit in the quantum 
80: theory of gravity and a Bohr transition frequency, applying Bohr's 
81: correspondence principle with a hydrogen atom to the quantum theory of 
82: a black hole \cite{MB95,B97}. For a Schwarzschild black hole, they then 
83: derived the Bohr transition frequency $\omega$, given by
84: %
85: \begin{equation}
86: \omega = \ln k / 8\pi M \,,
87: \label{bekenstein} 
88: \end{equation}
89: %
90: where $k=2, 3, 4\cdots$, and $M$ stands for the mass of a black hole. 
91: Furthermore they predicted the value of $k$ to be 
92: $k=2$, and suggested that this frequency should be equal to classical 
93: oscillation frequencies of the black hole. 
94: %
95: A few years ago, Hod however noticed that if $k=3$, the frequency 
96: given by formula (\ref{bekenstein}) is in quite good agreement with the 
97: asymptotic frequency of the QNM of a Schwarzschild black hole in the 
98: limit of highly damped modes, and proposed to apply Bohr's correspondence 
99: principle in order to determine the value of the fundamental area unit 
100: in the quantum theory of gravity, namely the value of $k$ \cite{hod} (see, also 
101: \cite{dreyer}). Since Hod's
102: proposal, the QNMs with large imaginary frequencies of spacetimes
103: including black holes have been attracted much attention, and a lot of
104: papers related to this subject have appeared in order to see
105: whether Hod's conjecture is applicable not only for a Schwarzschild
106: black hole but also for other black hole spacetimes. For example, Motl
107: analytically obtained an asymptotic constant value of the QNM
108: frequencies of a Schwarzschild black hole \cite{motl}, which had
109: been obtained numerically by Nollert \cite{nollert}. Motl and
110: Neitzke \cite{motl03}, Berti and Kokkotas \cite{berti}, Neitzke \cite{an03}, 
111: and Andersson and Howls
112: \cite{andersson_howls} studied the asymptotic behaviors of the QNMs in a
113: Reissner-Nordstr{\" o}m black hole. Berti et al. \cite{bertietal}, and Hod
114: \cite{hod03} discussed the QNMs of a Kerr black hole in the limit of highly damped
115: modes. 
116: As for a Schwarzschild-de Sitter (SdS) black hole, Cardoso and Lemos
117: \cite{cardoso03} and Maassen van den Brink \cite{brink} analytically obtained
118: asymptotic form of the QNM frequencies in almost the extremal limit, in which
119: the cosmological horizon becomes very close to the black hole horizon 
120: (for the case of a Schwarzschild black hole in an anti-de Sitter spacetime, 
121: see, e.g., \cite{CKL03,K02}).   
122: 
123: 
124: A large number of papers related to Hod's conjecture, which have recently 
125: appeared, suggest that Hod's conjecture is not universal, at least as it 
126: stands, even though it is applicable for Schwarzschild black holes in four 
127: and higher dimensions (for the higher dimensional case, see, e.g.,  
128: \cite{motl03,db03,an03,cly03}). For example, a 
129: real part of the quasinormal frequencies in a Reissner-Nordstr{\" o}m black hole 
130: appears not to go to any constant value in the limit of highly damped mode, 
131: but shows some periodic behaviors as the imaginary part is increased 
132: \cite{motl03,berti,an03,andersson_howls}. This means that Hod's conjecture is not 
133: applicable in the Reissner-Nordstr{\" o}m case because a real part of the 
134: quasinormal frequencies does not have a limit as the imaginary part goes to infinity. 
135: In such a situation, it is necessary to explore a problem whether or not there 
136: is another black hole spacetime in which Hod's conjecture is applicable. 
137: The purpose of this paper is to improve our understanding of this problem 
138: and we are concerned here with the QNMs of SdS spacetimes.  
139: SdS spacetime has no spatial infinity but has cosmological horizon, and if 
140: Hod's conjecture is applicable, it is interesting to see whether Hod's conjecture 
141: depends only on the black hole horizon, but not the cosmological horizon.
142: Furthermore, recent observations show that the universe does have a non-zero positive 
143: cosmological constant. Therefore, SdS spacetimes are considered to be a good simple 
144: model of a black hole in the universe.
145: 
146: In this study, in particular, we calculate numerically the QNMs of nearly 
147: extremal SdS spacetimes for reason we describe below. 
148: We therefore assume the surface gravity $\kappa_1$ at the 
149: black hole horizon to be $\kappa_1\le 10^{-2}\ r_1^{-1}$,
150: where $r_1$ stands for the coordinate radius of the black hole horizon. Note that 
151: the extremal limit of the SdS space corresponds to the limit of $\kappa_1\rightarrow 0$. 
152: In a nearly extremal SdS black hole, as mentioned before, an analytical formula for 
153: quasinormal frequencies can be derived \cite{cardoso03,brink}. One of the aims of this 
154: paper is to examine whether this analytical formula is correct in the limit of highly 
155: damped modes. 
156: 
157: The paper is organized as follows. 
158: In \S 2 we present the basic equations for obtaining QNMs 
159: in the SdS spacetime using Leaver's continued fraction technique \cite{leaver}, 
160: which was extended to the case 
161: of the SdS spacetime by Moss and Norman \cite{moss}.  
162: Numerical results are given in \S 3, and \S 4 is devoted for 
163: conclusion. 
164: 
165: 
166: 
167: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
168: \section{Method of Solutions}
169: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
170: In order to examine the QNMs of the SdS spacetime, we make use of the same 
171: formalism as that derived by Moss 
172: and Norman \cite{moss}, who obtained low-order quasinormal frequencies of the SdS 
173: spacetime for a wide range of the model parameter. The line element of the SdS 
174: spacetime is given by
175: %
176: \begin{equation}
177: ds^2=-{\Delta\over r^2}\,dt^2+{r^2\over\Delta}\,dr^2+
178: r^2(d\theta^2+\sin^2\theta\,d\phi^2)\,, 
179: \label{metric}
180: \end{equation}
181: %
182: where 
183: %
184: \begin{equation}
185: \Delta=r^2-2Mr-{1\over 3}\Lambda r^4\,. 
186: \label{def-delta}
187: \end{equation}
188: %
189: Here, $M$ and $\Lambda$ stand for the mass of the black hole and the cosmological 
190: constant, respectively. If a non-negative cosmological constant is assumed, 
191: namely de Sitter spacetime, there are two horizons, whose radial coordinates are 
192: given as positive solutions of $r^{-1}\Delta=r-2M-{1\over 3}\Lambda r^3=0$, 
193: one is the black hole horizon $r=r_1$ and the other the cosmological horizon 
194: $r=r_2$, where $r_2>r_1>0$. Note that the equation $r-2M-{1\over 3}\Lambda r^3=0$ 
195: has one negative solution $r=r_3<0$ for the SdS space. One of the important model 
196: parameters of the SdS spacetime is the surface gravity $\kappa_1$, defined by
197: %
198: \begin{equation}
199: \kappa_1=\lim_{r\rightarrow r_1}{1\over 2r^2}{d\Delta\over dr}\,. 
200: \label{def-kappa}
201: \end{equation}
202: %
203: In terms of the non-dimensional surface gravity $r_1\kappa_1$, the mass and the 
204: cosmological constant can be written as 
205: %
206: \begin{equation}
207: M={1\over 3}r_1(r_1\kappa_1+1)\,,\quad \Lambda=r_1^{-2}(1-2r_1\kappa_1)\,, 
208: \end{equation}
209: %
210: which shows that $0<r_1\kappa_1<1/2$ for the SdS spacetime. In this study, 
211: we employ the non-dimensional parameter $r_1\kappa_1$ to specify the SdS 
212: spacetime. 
213: 
214: 
215: By virtue of the symmetry properties of the SdS spacetime, the master equations 
216: for the scalar ($s=0$), electromagnetic ($s=1$) and gravitational perturbations ($s=2$) 
217: can be cast into a wave equation of the simple form, given by \cite{mm90,of91,cl01}
218: %
219: \begin{equation}
220: {d^2\phi(r)\over dr_*^2}+[\omega^2-V(r)]\phi(r)=0\,,
221: \label{wave-equation}
222: \end{equation}
223: %
224: where $r_*$ denotes the tortoise coordinate, defined by $dr_*=r^2\Delta^{-1}dr$, 
225: and $\omega$ is the oscillation frequency of the perturbations.  Depending on the 
226: type of perturbations, here, the effective potential is explicitly given by  
227: %
228: \begin{equation}
229: {r^4V\over\Delta}=\left\{
230: \begin{array}{ll}
231: l(l+1)+{2M\over r}-{2\Lambda r^2\over 3}&{\rm for\ } s=0 \\ 
232: l(l+1)&{\rm for\ } s=1 \\
233: l(l+1)-{6M\over r}&{\rm for\ } s=2\,, 
234: \end{array}
235: \right.
236: \label{def-potential}
237: \end{equation}
238: %
239: where $l$ means the angular quantum number of perturbations. Here, only the 
240: axial parity perturbations have been considered for the gravitational case because 
241: quasinormal frequencies of the polar parity perturbations are the same as those 
242: of the axial parity perturbations (for the proof, see Appendix). 
243: 
244: 
245: The QNMs of the SdS spacetime are characterized by the boundary 
246: conditions of incoming waves at the black hole horizon and outgoing waves at 
247: the cosmological horizon, given by 
248: %
249: \begin{equation}
250: \phi(r)\rightarrow\left\{
251: \begin{array}{ll}
252: e^{-i\omega r_*}&{\rm as}\ r_*\rightarrow \infty\\ 
253: e^{ i\omega r_*}&{\rm as}\ r_*\rightarrow-\infty\,,
254: \end{array}
255: \right.
256: \label{def-bc}
257: \end{equation}
258: % 
259: where the time dependence of perturbations has been assumed to be $e^{i\omega t}$. 
260: In general, it is impossible to adapt the boundary condition (\ref{def-bc}) in 
261: a straightforward numerical integration to obtain quasinormal frequencies. Some 
262: special technique is therefore required for computations of QNMs. 
263: In the present investigation, we employ a standard technique devised by Leaver, namely 
264: the continued fraction method \cite{moss,leaver}.  
265: 
266: 
267: To apply the continued fraction method to the SdS spacetime, it is convenient to 
268: introduce a new independent variable, defined by $x=r^{-1}$. With this new variable 
269: $x$, the asymptotic form of the perturbations as $r_*\rightarrow\infty$ can be 
270: rewritten as
271: %
272: \begin{equation}
273: e^{-i\omega r_*}=(x-x_1)^{-\rho_1}(x-x_2)^{-\rho_2}(x-x_3)^{-\rho_3}\,,
274: \label{expwr}
275: \end{equation}
276: %
277: where $x_i=r_i^{-1}$ and $\rho_i=i\omega/(2\kappa_i)$ for $i=1,2,3$, where 
278: %
279: \begin{eqnarray}
280: \kappa_1&=&M(x_1-x_2)(x_1-x_3)\,,\nonumber \\
281: \kappa_2&=&M(x_2-x_1)(x_2-x_3)\,, \\
282: \kappa_3&=&M(x_3-x_1)(x_3-x_2)\,.\nonumber 
283: \end{eqnarray}
284: %
285: The perturbation function $\phi$ is expanded around the black hole horizon as 
286: %
287: \begin{equation}
288: \phi=(x-x_1)^{\rho_1}(x-x_2)^{-\rho_2}(x-x_3)^{-\rho_3}\sum_{n=0}^\infty 
289: a_n\left({x-x_1\over x_2-x_1}\right)^n\,,
290: \label{expansion}
291: \end{equation}
292: %
293: where $a_0=1$ and $a_n$'s for $n\ge 1$ are determined by the three term recurrence 
294: relation, given by 
295: %
296: \begin{equation}
297: \alpha_na_{n+1}+\beta_na_n+\gamma_na_{n-1}=0\,,
298: \end{equation}
299: %
300: where 
301: %
302: \begin{eqnarray}
303: \alpha_n&=&2M(x_1-x_3)\{n^2+2(\rho_1+1)n+2\rho_1+1\}\,,\\
304: \beta_n&=&-2M(2x_1-x_2-x_3)\times\nonumber\\ 
305: &&\{n^2+(4\rho_1+1)n+4\rho_1^2+2\rho_1\}-l(l+1)\nonumber\\
306: &&+2Mx_1(s^2-1)\,,\\
307: \gamma_n&=&2M(x_1-x_2)(n^2+4\rho_1n+4\rho_1^2-s^2)\,. 
308: \end{eqnarray}
309: %
310: Here, $s=2$ for the gravitational perturbations and $s=1$ for the electromagnetic 
311: perturbations. Note that in the present study, we do not consider the scalar 
312: perturbations because Leaver's method cannot be directly applied to the scalar case. 
313: Comparing the expanded eigenfunction (\ref{expansion}) with equation (\ref{expwr}), 
314: we can see that the eigenfunction (\ref{expansion}) satisfies the QNM  
315: boundary condition (\ref{def-bc}) if the power series in equation (\ref{expansion}) 
316: converges for $x_2\le x\le x_1$. This convergence condition is equivalent to the 
317: condition written in terms of continued fractions \cite{gautschi}, which is given by 
318: %
319: \begin{eqnarray}
320: 0=\beta_0-{\alpha_0\gamma_1\over\beta_1-}{\alpha_1\gamma_2\over\beta_2-}\cdots
321: \label{a-eq1}
322: \end{eqnarray}
323: %
324: Therefore, we solve this algebraic equation to obtain the quasinormal frequency.  
325: 
326: 
327: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
328: \section{Numerical Results}
329: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
330: In this study, we are concerned with asymptotic behaviors of high-order QNMs, 
331: namely quasinormal frequencies with large imaginary parts, in almost the  
332: extremal limit, in which two horizons are quite near. We therefore consider only the 
333: case of $r_1\kappa_1\le 10^{-2}$. For these small values of $r_1\kappa_1$, the continued 
334: fractions (\ref{a-eq1}) converge very quickly and Leaver's method works quite well even 
335: when quasinormal frequencies have quite large imaginary parts. 
336: Note that for moderate values of $r_1\kappa_1$, however, the convergence of the continued 
337: fractions gets worse and Leaver's method is applicable only for modes with smaller imaginary 
338: frequencies. For those cases, thus, some other techniques such as Nollert's method \cite{nollert} 
339: or a phase integral method \cite{andersson} should be used to obtain high-order QNMs.
340: 
341: 
342: In order to check our numerical code, we have calculated fundamental frequencies of 
343: the QNMs for several values of $r_1\kappa_1$ and have fitted the mode 
344: frequencies as a function of $r_1\kappa_1$ with the polynomials defined by 
345: %
346: \begin{eqnarray}
347: {\rm Re}(\omega r_1)&=&r_1\kappa_1 b_0(1-b_1r_1\kappa_1)\,,\nonumber \\
348: {\rm Im}(\omega r_1)&=&r_1\kappa_1 c_0(1-c_1r_1\kappa_1)\,. 
349: \label{w-exp}
350: \end{eqnarray}
351: %
352: Recently, Cardoso and Lemos \cite{cardoso03} and Maasse van den Brink \cite{brink} 
353: analytically obtained the expansion coefficients, which are given by
354: %
355: \begin{equation}
356: b_0=\left\{
357: \begin{array}{ll}
358: \sqrt{l(l+1)-{1\over 4}}&{\rm for\ } s=0,\ 1\\
359: \sqrt{l(l+1)-{9\over 4}}&{\rm for\ } s=2 \,,
360: \end{array}
361: \right. 
362: \label{ana-coef1}
363: \end{equation} 
364: %
365: \begin{equation}
366: c_0=n+{1\over 2}\,,\quad b_1=c_1={2\over 3}\,,
367: \label{ana-coef2}
368: \end{equation}
369: %
370: where $n$ is the mode number. It is found that numerically obtained coefficients are 
371: in good agreement with coefficients given by (\ref{ana-coef1}) and (\ref{ana-coef2}). 
372: Note that our numerical results for the mode frequencies are consistent with those 
373: obtained by Moss and Norman \cite{moss}, who studied low-order QNMs 
374: of the gravitational perturbations for full range of the parameter $r_1\kappa_1$.
375:   
376: 
377: First, let us discuss properties of the quasinormal frequencies for the low-order 
378: modes. In Figure 1, we show the 
379: real parts of the frequencies for the low-order QNMs versus the imaginary 
380: parts of the frequencies for the gravitational perturbations with $l=2$. In this figure, 
381: non-dimensional frequencies ${\rm Re}(\omega/\kappa_1)$ have been plotted as a function of 
382: ${\rm Im}(\omega/\kappa_1)$ for $r_1\kappa_1=10^{-3}$ and $r_1\kappa_1=5\times 10^{-3}$,  
383: and the dashed curve indicates the approximate frequency for $r_1\kappa_1\rightarrow 0$, 
384: derived by Cardoso and Lemos \cite{cardoso03} (see, also \cite{brink}). For low-order modes,  
385: Figure 1 illustrates how the frequencies of the QNMs behaves when the mode number and/or 
386: the value of $r_1\kappa_1$ is altered. Basic properties of the low-order QNMs 
387: are summarized as follows: For the modes associated with a small mode number, the real parts of 
388: the frequencies decrease with the increase of the imaginary parts of the frequency, even though 
389: the real parts of the frequency are constant in the approximation formula 
390: (\ref{w-exp})-(\ref{ana-coef2}). In other words, the 
391: analytical approximation formula in the extremal limit (\ref{w-exp})-(\ref{ana-coef2}) is quite 
392: good for the fundamental modes as long as $r_1\kappa_1$ is small enough, while, as expected in 
393: \cite{brink}, this approximation formula gets worse as the mode number is increased even for 
394: small values of $r_1\kappa_1$. This means that formula 
395: (\ref{w-exp})-(\ref{ana-coef2}) 
396: does not give a correct asymptotic value of the QNM frequencies in the limit of large imaginary 
397: frequencies. Similar properties can be seen for other perturbations 
398: having different $s$ and $l$. In Figure 2, we show the same results as those in 
399: Figure 1 but for the electromagnetic perturbations having $l=1$. It is observed  
400: that similar behaviors of the QNM frequencies are seen in the case of electromagnetic 
401: perturbations, too. 
402: 
403: 
404: Now, we explain our numerical results for the asymptotic behavior of the QNM 
405: of nearly extremal SdS spacetimes in the limit of large imaginary 
406: frequencies. 
407: In Figure 3, we plot the imaginary parts of the non-dimensional 
408: QNM frequencies, $\omega/\kappa_1$, of the gravitational perturbations 
409: with $l=2$ as a function of the mode number $n$. 
410: In this figure, the model parameter is $r_1\kappa_1=10^{-3}$. 
411: The figure shows that an asymptotic form of ${\rm Im}(\omega/\kappa_1)\sim
412: n$, 
413: which is similar to the analytical 
414: formula (\ref{w-exp})-(\ref{ana-coef2}), is a good approximation for the imaginary 
415: parts in the limit of large imaginary frequencies. 
416: The same asymptotic form is inferred in all other QNM's 
417: we have calculated in the present study, 
418: regardless of the values of $r_1\kappa_1$, $l$, and $s$. 
419: 
420: 
421: Let us next focus on the behaviors of the real part of the QNM
422: frequencies in the limit of highly damped modes. 
423: In Figures 4 and 5, the real parts of the non-dimensional mode frequencies, 
424: $\omega/\kappa_1$ are plotted 
425: as a function of the imaginary parts of the frequencies up to
426: sufficiently high-order modes for the $l=2,\ 3$ gravitational and 
427: for the $l=1,\ 2$ electromagnetic perturbations, respectively. 
428: The results for the model parameter of $r_1\kappa_1=10^{-3}$ are shown
429: in both figures. It is found that the real parts of the frequencies show 
430: oscillating behaviors as the imaginary parts of the frequencies are increased. 
431: It is important to note that similar oscillating behaviors have been 
432: observed in the QNMs of a Reissner-Nordstr{\" o}m black hole. (The quasinormal 
433: frequencies with large imaginary frequency of a Reissner-Nordstr{\" o}m black 
434: hole can be given in terms of a solution of the algebraic equation 
435: \cite{motl03,an03,andersson_howls}, 
436: %
437: \begin{equation}
438: e^{\beta\omega}+2+3e^{k^2\beta\omega}=0\,,
439: \label{rne}
440: \end{equation}
441: %
442: where $\beta$ and $k$ are contents determined with the mass and charge of the 
443: black hole (see, \cite{motl03}). As shown by Neitzke \cite{an03} and Anderson 
444: and Howls \cite{andersson_howls}, equation (\ref{rne}) has an infinite number 
445: of solutions and some solutions of equation (\ref{rne}) show periodicity.) 
446: The behavior of the amplitude of the oscillating 
447: ${\rm Re}(\omega/\kappa_1)$ as a function of ${\rm Im}(\omega/\kappa_1)$  
448: resembles that of QNM 
449: frequencies of a Schwarzschild black hole. For the gravitational perturbations, 
450: the amplitude decreases for small values of 
451: ${\rm Im}(\omega/\kappa_1)$, approaches the imaginary axis of the complex 
452: frequency plan, increases again, and finally approaches some constant value. 
453: The asymptotic value of the amplitude in the limit 
454: of highly damped modes seems to be non-zero constant, which is inferred as 
455: $\sim 0.4$. Therefore the limit of ${\rm Re}(\omega/\kappa_1)$ as 
456: ${\rm Im}(\omega/\kappa_1)\rightarrow\infty$ appears not to exist for the gravitational 
457: perturbations. For the electromagnetic perturbations, on the other hand, the 
458: amplitude of oscillating ${\rm Re}(\omega/\kappa_1)$ decreases monotonically as 
459: ${\rm Im}(\omega/\kappa_1)$ is increased. Its asymptotic value in the limit of 
460: highly damped modes seems to be zero. This means that for the electromagnetic 
461: perturbations, the limit of ${\rm Re}(\omega/\kappa_1)$ as ${\rm Im}(\omega/\kappa_1)$ 
462: goes to infinity seems to exist and to be zero. It is important to note that in a 
463: nearly extremal SdS black hole, the asymptotic behaviors of the quasinormal 
464: frequencies in the 
465: limit of highly damped modes are independent of the angular quantum number 
466: $l$ of the perturbations. Although we do not show the results for other 
467: values of $r_1\kappa_1$ and $l$, the asymptotic behavior of the QNM frequencies 
468: is not highly dependent on these parameters. 
469: %
470: In summary, our numerical results suggest that for the gravitational 
471: perturbations, the real parts of the QNM frequencies of nearly extremal 
472: SdS spacetimes do not go to any constant value in the limit of 
473: large imaginary frequencies because they show oscillating behaviors in 
474: the limit. For the electromagnetic perturbations, on the other hand, 
475: the real parts of the QNM frequencies seem to go to zero in the limit 
476: of large imaginary frequencies.  
477: 
478: 
479: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
480: \section{Conclusions}
481: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
482: We have calculated the high-order QNMs with large imaginary frequencies 
483: for the electromagnetic and gravitational perturbations in nearly
484: extremal SdS spacetimes using  Leaver's continued 
485: fraction method \cite{leaver}. 
486: Our results show that for low-order QNMs, analytical formulas in 
487: the extremal limit derived by Cardoso and Lemos \cite{cardoso03} 
488: and Maassen van den Brink \cite{brink} is a quite good approximation 
489: for the QNM frequencies as long as the model parameter $r_1\kappa_1$ 
490: is small enough. 
491: For high-order QNMs, whose imaginary frequencies are 
492: sufficiently large, on the other hand, this formula becomes inaccurate 
493: even for small values of $r_1\kappa_1$. 
494: Therefore, the approximation derived by Cardoso and Lemos cannot give 
495: correct asymptotic behaviors of the QNMs in the limit of large
496: imaginary frequencies (see, also \cite{brink}). 
497: We also found that the real parts of the quasinormal frequencies 
498: have oscillating behaviors in the limit of highly damped modes.  
499: (Similar behaviors have been found in the quasinormal frequencies in a
500: Reissner-Nordstr{\" o}m black hole \cite{motl03,berti,an03,andersson_howls}.) 
501: The amplitude of oscillating ${\rm Re(\omega)}$ approaches a non-zero 
502: constant value for the gravitational perturbations and zero for the 
503: electromagnetic perturbations in the limit of highly damped modes, 
504: regardless of values of $l$ and $r_1\kappa_1$. This means that for the 
505: gravitational perturbations, the real parts of the quasinormal frequencies 
506: of nearly extremal SdS spacetimes appear not to go to any constant value 
507: in the limit of highly damped modes. Therefore our numerical results suggest 
508: that Hod's conjecture is not applicable for nearly extremal SdS black holes 
509: because the the limit of ${\rm Re(\omega)}$ as 
510: ${\rm Im}(\omega)\rightarrow\infty$ does not exist. 
511: 
512: 
513: Although we computed high-order QNMs whose damping rates are quite large, all the QNMs 
514: we obtained in this study are still associated with a finite mode number but not infinity, 
515: because we investigated the properties of the QNMs with straightforward numerical approach. 
516: Thus, we cannot exclude the possibility that our numerical results do not show correct 
517: asymptotic behaviors. 
518: Other approaches to examine asymptotic behaviors of QNMs in the highly damping 
519: limit are necessary, in order to confirm our results of the asymptotic behaviors. 
520: As for high-order QNMs with large imaginary frequencies for moderate values of 
521: $r_1\kappa_1$, Leaver's method cannot be applied straightforwardly. 
522: Therefore, other numerical techniques are needed to obtain QNMs with large imaginary 
523: frequencies. 
524: 
525: 
526: \begin{acknowledgments} 
527: The authors are grateful to V. Cardoso for fruitful discussions and a 
528: careful reading of the manuscript. 
529: They are also grateful to J. Bekenstein for correspondence regarding 
530: the early history of the black hole thermodynamics.  S.Y. 
531: acknowledges financial support from Funda\c c\~ao para a  Ci\^encia e a 
532: Tecnologia (FCT) through project SAPIENS 36280/99.
533: \end{acknowledgments}
534: 
535: 
536: \appendix* 
537: 
538: \section{Superpartner and iso-spectral relationship between axial and polar perturbations in 
539: SdS spacetime} 
540: 
541: Non-radial gravitational perturbations of a SdS spacetime obey a Schr\"odinger-type wave 
542: equation, given by  
543: %
544: \begin{equation}
545: {d^2\phi^{(\pm)}\over dr_*^2}+[\omega^2-V^{(\pm)}(r)]\phi^{(\pm)}=0\,, 
546: \label{a1}
547: \end{equation}
548: %
549: where $\phi^{(+)}$ and $V^{(+)}$ ($\phi^{(-)}$ and $V^{(-)}$) are the gauge invariant perturbation 
550: function and the effective potential for polar (axial) parity perturbations, respectively. 
551: The effective potentials are given by 
552: %
553: \begin{eqnarray}
554: V^{(+)}&=&{2\Delta\over r^5(cr+3M)^2}\nonumber\\
555: &\times&[9M^3+9M^2cr+3c^2Mr^2\nonumber\\
556: &&\ \ +c^2(c+1)r^3-3M^2\Lambda r^3]\,, 
557: \label{a2}
558: \end{eqnarray}
559: % 
560: \begin{eqnarray}
561: V^{(-)}={2\Delta\over r^5}\,[(c+1)r-6M]\,,
562: \label{a3}
563: \end{eqnarray}
564: %
565: where $c=(l+2)(l-1)/2$ (for detailed derivations of the master equation (\ref{a1}), see, e.g., 
566:  \cite{mm90,cl01}). As shown first by Cardoso and Lemos \cite{cl01}, two potentials $V^{(+)}$ 
567: and $V^{(-)}$ are simply related through the relation, given by
568: %
569: \begin{eqnarray}
570: V^{(\pm)}=\pm\beta {df\over dr_*}+\beta^2 f^2+\kappa f\,,
571: \label{a4}
572: \end{eqnarray}
573: % 
574: where $\beta=6M$, $\kappa=4c(c+1)$, and 
575: %
576: \begin{equation}
577: f={\Delta\over 2r^3(cr+3M)}\,. 
578: \label{a5}
579: \end{equation}
580: %
581: This relation between two potentials are called the superpartner relationship. 
582: By virtue of the superpartner relationship, the perturbation function $\phi^{(+)}$ 
583: ($\phi^{(-)}$) can be written in term of $\phi^{(-)}$ ($\phi^{(+)}$) and its first 
584: derivative \cite{ch83}, given by
585: %
586: \begin{eqnarray}
587: (\kappa\pm 2i\omega\beta)\phi^{(\pm)}=(\kappa+2\beta^2f)\,\phi^{(\mp)}\pm2\beta\,
588: {d\phi^{(\mp)}\over dr_*}\,. 
589: \label{a6}
590: \end{eqnarray} 
591: %
592: It is worthwhile to note that the superpartner relationship (\ref{a4}) and (\ref{a5}) and 
593: the relations between two functions $\phi^{\pm}$ (\ref{a6}) in a SdS spacetime have the same 
594: functional form as those in a Schwarzschild spacetime except for the definition of the 
595: function $\Delta$ \cite{ch83}. Since $\Delta=\Lambda(r-r_1)(r_2-r)(r-r_3)r/3$, the 
596: asymptotic form of the function $f$ in the limit of $r_*\rightarrow\pm\infty$ is 
597: given by 
598: %
599: \begin{equation}
600: f\rightarrow\left\{
601: \begin{array}{ll}
602: {\Lambda(r_2-r_1)(r_1-r_3)\over 6r_1^2(cr_1+3M)}\,e^{2|\kappa_1|r_*}&
603: {\rm as}\ r_*\rightarrow -\infty\\
604: {\Lambda(r_2-r_1)(r_2-r_3)\over 6r_2^2(cr_2+3M)}\,e^{-2|\kappa_2|r_*}&
605: {\rm as}\ r_*\rightarrow\infty\,. 
606: \end{array}
607: \right.
608: \end{equation}
609: %
610: Then, it is easy to see that the function $f$ has three properties; i) smooth for 
611: $-\infty< r_*< \infty$, ii) $f$ and its derivatives of all orders vanish as 
612: $r_*\rightarrow\pm\infty$, iii) an integral $\int^\infty_{-\infty}fdr_*$ exists.  
613: If the function $f$ appearing in (\ref{a4}) satisfies three conditions above, as shown in 
614: \cite{ch83}, two potentials $V^{(\pm)}$ give the same transmission amplitude and the 
615: same quasinormal frequencies. In a SdS spacetime, therefore, axial and polar perturbations 
616: yield the same set of quasinormal mode frequencies. This iso-spectral properties in a SdS 
617: spacetime is attributed to the fact that gravitational perturbations associated with 
618: a spin $s=-2$ in a SdS spacetime can be described with a single Weyl scalar $\Psi_4$ 
619: \cite{of91}. 
620: 
621: 
622: 
623: In a Schwarzschild-anti-de Sitter spacetime, exactly the same relations between polar 
624: and axial perturbations (\ref{a4})--(\ref{a6}) obviously hold \cite{cl01}. Yet, 
625: there is no iso-spectral property between polar and axial perturbations in a 
626: Schwarzschild-anti-de Sitter spacetime. In a Schwarzschild-anti-de Sitter spacetime, 
627: the master equation (\ref{a1}) does not 
628: have an asymptotic solution given by $e^{\pm i\omega r_*}$ as $r\rightarrow\infty$ 
629: and, furthermore, $r_*$ has a finite range. The boundary condition at spatial 
630: infinity therefore must be modified.  One of the plausible boundary conditions is 
631: that perturbation functions vanish at spatial infinity, even though there are other 
632: options for the boundary conditions \cite{cl01}. If this boundary condition is taken 
633: at spatial infinity, in general, the transformation (\ref{a6}) between polar and 
634: axial perturbations cannot hold this boundary condition. Therefore, the set of the 
635: quasinormal frequencies of polar perturbations in a Schwarzschild-anti-de Sitter 
636: spacetime is not the same as that of axial perturbations. 
637: 
638: 
639: 
640: 
641: \begin{thebibliography}{99}
642: %
643: \bibitem{vish} C. V. Vishveshwara, Nature {\bf 227}, 936 (1970). 
644: %
645: \bibitem{kokkotas99} K. D. Kokkotas and B. G. Schmidt, 
646: Living Rev. Relativity {\bf 2}, 2 (1999).
647: %
648: \bibitem{nollert99} H. P. Nollert, 
649: Class. Quant. Grav. {\bf 16}, R159 (1999).
650: %
651: \bibitem{MB95} J. D. Bekenstein and V. F. Mukhanov, Phys. Lett {\bf B 360},
652:  7 (1995). 
653: %
654: \bibitem{B97} J. D. Bekenstein, in {\it Proceedings of the Eight Marcel 
655: Grossmann  Meeting on General Relativity}, eds. T. Piran and R. Ruffini 
656: (World Scientific, Singapore 1999), pp. 92-111.
657: %
658: \bibitem{hod} S. Hod, Phys. Rev. Lett. {\bf 81}, 4293 (1998).
659: %
660: \bibitem{dreyer} O. Dreyer, Phys. Rev. Lett. {\bf 90}, 081301 (2003).
661: %
662: \bibitem{motl} L. Motl, Adv. Theor. Math. Phys. {\bf 6}, 1135 (2003).
663: %
664: \bibitem{nollert} H.-P. Nollert, Phys. Rev. D {\bf 47}, 5253 (1993).
665: %
666: \bibitem{motl03} L. Motl and A. Neitzke, Adv. Theor. Math. Phys. {\bf 7}, 
667: 307 (2003). 
668: %
669: \bibitem{berti} E. Berti and K. D. Kokkotas, Phys. Rev. {\bf D} 68, 
670: 044027 (2003).
671: %
672: \bibitem{an03} A. Neitzke, preprint (hep-th/0304080)
673: %
674: \bibitem{andersson_howls} N. Andersson and C. J. Howls, 
675: preprint (gr-qc/0307020). 
676: %
677: \bibitem{bertietal}E. Berti, V. Cardoso, K. D. Kokkotas, and 
678: H. Onozawa, Phys. Rev. D {\bf 68}, 124018 (2003).
679: %
680: \bibitem{hod03} S. Hod, Phys. Rev. D {\bf 67}, 081501 (2003). 
681: %
682: \bibitem{cardoso03} V. Cardoso and J. P. S. Lemos, Phys. Rev. D {\bf 67}, 
683: 084020 (2003).
684: %
685: \bibitem{brink} A. Maassen van den Brink, Phys. Rev. D {\bf 68}, 
686: 047501 (2003).
687: %
688: \bibitem{K02} R. A. Konoplya, Phys. Rev. D {\bf 66}, 044009 (2002).
689: %
690: \bibitem{CKL03} V. Cardoso, R. A. Konoplya, and J. P. S. Lemos,
691: Phys. Rev. D {\bf 68}, 044024 (2003).
692: %
693: \bibitem{leaver} E. W. Leaver,
694: Proc. R. Soc. London {\bf A402}, 285 (1985).
695: %
696: \bibitem{moss} I. G. Moss and J. P. Norman,
697: Class. Quant. Grav. {\bf 19}, 2323 (2002).
698: %
699: \bibitem{mm90}
700: F. Mellor and I. Moss, Phys. Rev. D {\bf 41}, 403 (1990); 
701: %
702: \bibitem{of91}
703: H. Otsuki and T. Futamase, Prog. Theor. Phys. {\bf 85}, 771 (1991); 
704: %
705: \bibitem{cl01}
706: V. Cardoso and J. P. S. Lemos, Phys. Rev. D {\bf 64}, 084017 (2001). 
707: %
708: \bibitem{gautschi} W. Gautschi, SIAM Rev. {\bf 9}, 24 (1967). 
709: %
710: \bibitem{andersson} N. Andersson, 
711: Proc. R. Soc. London {\bf A439}, 47 (1992).
712: %
713: \bibitem{cly03} V. Cardoso, J. P. S. Lemos, and S. Yoshida, Phys. 
714: Rev. D, in press (gr-qc/0309112); V. Cardoso, J. P. S. Lemos, and S. 
715: Yoshida, JHEP {\bf 0312}, 041 (2003). 
716: %
717: \bibitem{db03} D. Birmingham, Phys. Lett. {\bf B 569}, 199 (2003). 
718: %
719: \bibitem{ch83}
720: S. Chandrasekhar, {\it The Mathematical Theory of Black holes} 
721: (Oxford University, New York, 1983). 
722: \end{thebibliography}
723: 
724: %\newpage
725: 
726: %Figures
727: 
728: \begin{figure}[ht]
729: \centering
730: \includegraphics[height=5cm,clip]{f1.eps}
731: \caption{Real parts of the non-dimensional QNM frequencies, $\omega/\kappa_1$, 
732: given as a function of the imaginary parts of the frequencies for the l=2 
733: gravitational perturbations. The model parameters $r_1\kappa_1$ are taken to be
734: $r_1\kappa_1=10^{-3}$ and $r_1\kappa_1=5\times 10^{-3}$. The frequencies 
735: obtained with the approximation formula in the limit of $r_1\kappa_1\rightarrow 0$ 
736: are also shown as the dashed curve.}
737: \end{figure}
738: %
739: \begin{figure}[ht]
740: \centering
741: \includegraphics[height=5cm,clip]{f2.eps}
742: \caption{Same as Figure 1 but for the $l=1$ electromagnetic perturbations.}
743: \end{figure}
744: %
745: \begin{figure}[ht]
746: \centering
747: \includegraphics[height=5cm,clip]{f3.eps}
748: \caption{Imaginary parts of the non-dimensional QNM frequencies, $\omega/\kappa_1$, 
749: given as a function of the mode number, $n$, for the gravitational perturbations 
750: associated with $l=2$. The model parameter $r_1\kappa_1$ is taken to be 
751: $r_1\kappa_1=10^{-3}$.}  
752: \end{figure}
753: %
754: \begin{figure}
755: \centering
756: \includegraphics[width=10cm]{f4.eps}
757: \includegraphics[width=10cm]{f5.eps}
758: \caption{Real parts of the non-dimensional QNM frequencies, $\omega/\kappa_1$,
759: given as a function of the imaginary parts of the frequencies for the 
760: gravitational perturbations having $l=2$ and $l=3$. The frequencies of the QNMs 
761: associated with different $l$, $l=2,\ 3$, are shown in each panel. The model 
762: parameters $r_1\kappa_1$ is taken to be $r_1\kappa_1=10^{-3}$.}
763: \end{figure}
764: %
765: \begin{figure}
766: \centering
767: \includegraphics[width=10cm]{f6.eps}
768: \includegraphics[width=10cm]{f7.eps}
769: \caption{Same as Figure 3 but for the electromagnetic perturbations having 
770: $l=1$ and $l=2$.}  
771: \end{figure}
772: %
773: 
774: \end{document}
775: 
776: 
777: