0704.3467/ms.tex
1: %\documentclass[aps,float,prl,psfig]{revtex4}
2: \documentclass[aps,twocolumn,float,prd,psfig]{revtex4}
3: \input epsf
4: \usepackage{graphicx}% Include figure files
5: \newcommand{\beq}{\begin{equation}}
6: \newcommand{\beqa}{\begin{eqnarray}}
7: \newcommand{\eeq}{\end{equation}}
8: \newcommand{\eeqa}{\end{eqnarray}}
9: \newcommand{\etal}{{\it et al. }}
10: \newcommand{\siml}{\lesssim}
11: \newcommand{\simg}{\gtrsim}
12: \newcommand{\lsim}{\lesssim}
13: \newcommand{\gsim}{\gtrsim}
14: \newcommand{\psim}{\mbox{\raisebox{-1.0ex}{$~\stackrel{\textstyle \propto}
15: {\textstyle \sim}~$ }}}
16: \newcommand{\vect}[1]{\mbox{\boldmath${#1}$}}
17: 
18: \begin{document}
19: \title{
20: Second and higher-order quasi-normal modes
21: in binary black hole mergers
22: } 
23: \author{Kunihito Ioka$^{1}$ and Hiroyuki Nakano$^{2}$
24: }
25: \affiliation{$^{1}$
26: Department of Physics, Kyoto University, Kyoto 606-8502, Japan \\
27: $^{2}$Center for Computational Relativity and Gravitation, 
28: School of Mathematical Sciences, \\
29: Rochester Institute of Technology, 
30: Rochester, New York 14623, USA
31: }
32: 
33: \begin{abstract}
34:  Black hole (BH) oscillations known as quasi-normal modes (QNMs) are one
35:  of the most important gravitational wave (GW) sources. We propose that
36:  higher perturbative order of QNMs, generated by nonlinear 
37:  gravitational interaction near the BHs, are detectable and worth
38:  searching for in observations and simulations of binary BH mergers. 
39:  We calculate the
40:  metric perturbations to second-order and explicitly regularize the
41:  master equation at the horizon and spatial infinity. 
42:  We find that the
43:  second-order QNMs have frequencies twice the first-order ones and the
44:  GW amplitude is up to $\sim 10\%$ that of the first-order one. 
45:  The QNM frequency would also shift blueward up to $\sim 1\%$.
46:  This provides a new test of general relativity as well as 
47:  a possible distance indicator.
48: \end{abstract}
49: \pacs{PACS number(s): 04.30.-w, 04.70.Bw, 95.55.Ym, 95.85.Sz, 98.80.Es}
50: \maketitle
51: 
52: %\baselineskip 8mm
53: 
54: \section{Introduction}
55: Direct detections of gravitational waves (GWs)
56: will become a reality in the near future with current and future
57: detectors, such as LIGO, LISA and DECIGO/BBO~\cite{Seto:2001qf}. 
58: GWs can not only provide a test of general relativity but also open 
59: a new window on the universe.
60: 
61: One of the most important GW sources are the quasi-normal modes (QNMs)
62: of black holes (BHs)~\cite{Kokkotas:1999bd}. QNMs are oscillations
63: of the BH metric perturbations, and they are damped by emitting GWs. QNM
64: frequencies are complex with the real part representing the oscillation
65: frequency and the imaginary part representing the damping. 
66: By observing the QNM frequencies, one can determine the mass and angular
67: momentum of spinning BHs.
68: 
69: The most promising sources that excite QNMs are binary BH mergers. For
70: these events QNMs will be detected with high signal-to-noise ratio (SNR)
71: (e.g., SNR $\sim 10^5$ for $\sim 10^8 M_{\odot}$ BH mergers at $\sim
72: 1$Gpc by LISA)~\cite{Flanagan:1997sx} since a large fraction of energy
73: ($E \sim 1\% \times M$ for equal mass mergers with a total mass $M$) 
74: goes into GWs. The merger rate is also estimated to be
75: large enough~\cite{Enoki:2004ew,Ioka:2005pm}.
76: In numerical relativity there has been recently breakthroughs
77: for calculating the entire phase of binary BH 
78: mergers~\cite{Pretorius:2005gq}. 
79: The result is that the $\ell=2$, $m=2$ QNM is dominant, carrying away 
80: $\sim1\%$ of the initial rest mass of 
81: the system~\cite{Berti:2007fi}.
82: 
83: In this paper we show that the higher-order QNMs are also prominent in
84: binary BH mergers and they are interesting to search for in the
85: observations and simulations. Here the ``higher-order'' is with respect to 
86: the metric perturbations. Although the dimensionless amplitude of the metric
87: perturbations are negligible small when we observe them at detectors, they
88: are relatively large near the BH, up to 
89: \beqa
90: \psi^{(1)}/M\sim \left(E/M\right)^{1/2} \sim 10\%
91: \eeqa
92: for QNMs with energy $E/M \sim 1\%$ 
93: [see Eq.~(\ref{eq:E1}) with $\omega \sim M^{-1}$. 
94: $\psi^{(1)}$ denotes a wave function to be discussed later].
95: Hence the generated
96: second-order perturbations would have the amplitude $\sim 10\%$ of the 
97: first-order ones, which may be detectable for high SNR events.
98: %(i.e., SNR $\sim 10 (s/100)$ for second-order QNMs if SNR $\sim s$ 
99: %for first-order ones).
100: 
101: Higher-order QNMs are essentially analogous to the anharmonic oscillations
102: discussed by Landau \& Lifshitz in ``Mechanics''~\cite{landau76}. In
103: general, an oscillation with small amplitude $x$ is described by an
104: equation,
105: %\beqa
106: $\ddot x+\omega^2 x=0$,
107: %\eeqa
108: which gives the first-order solution $x=a \cos(\omega t + \phi)$. 
109: Here $a$ and $\phi$ are integration constants. 
110: Including the second-order term with respect to the amplitude $\sim x^2$,
111: the equation becomes
112: \beqa
113: \ddot x+\omega^2 x=-\alpha x^2 \,.
114: \eeqa
115: With the right hand side as a source term, we obtain a successive solution 
116: $x=a \cos(\omega t +\phi)+x^{(2)}$ where
117: \begin{eqnarray}
118: x^{(2)}=\frac{\alpha a^2}{6 \omega^2} \cos 2\omega t 
119: - \frac{\alpha a^2}{2\omega^2} \propto a^2 \,.
120: \end{eqnarray}
121: Here an oscillation arises with a frequency $2\omega$.
122: The important point is that the
123: second-order oscillation $x^{(2)}$ is driven by the first-order one
124: and thereby always exists.
125: 
126: Our result is that the second-order QNMs would also have
127: frequencies twice the first-order ones
128: $\omega^{(2)}=2 \omega^{(1)}$ 
129: and the amplitude up to $\sim 10\%$ 
130: of the first-order ones.
131: Since higher-order QNMs always exist, we may test
132: the nonlinearity of general relativity.
133: The purpose of this paper is 
134: to outline the calculations of the second-order QNMs for the
135: Schwarzschild BH and clarify the order-of-magnitude estimates. 
136: More details will be given in the forthcoming papers.
137: We use the units $c=G=1$, and
138: arbitrarily set $M=1$ which we can always recover, if necessary. 
139: 
140: Although this paper is the first to study second-order QNMs,
141: the second-order analysis is pioneered by Tomita~\cite{Tomita},
142: and the $\ell=2$, $m=0$ case in the Schwarzschild spacetime
143: is studied by Gleiser et al.~\cite{Gleiser:1995gx}.
144: It is also extended to cosmology~\cite{Acquaviva:2002ud} 
145: and the Kerr case~\cite{Campanelli:1998jv}. 
146: 
147: \section{First-order}
148: Let us consider the metric perturbations to second-order,
149: \beqa
150: \tilde g_{\mu \nu}=g_{\mu \nu}+h_{\mu \nu}^{(1)}+h_{\mu \nu}^{(2)} \,,
151: \eeqa
152: where $g_{\mu \nu}$ is the Schwarzschild metric and the
153: superscripts denote a perturbative order. Expanding the Einstein's 
154: vacuum equation we can obtain basic equations order 
155: by order~\cite{Gleiser:1995gx,Bruni:1996im}.
156: 
157: For the first-order, we use the Regge-Wheeler-Zerilli 
158: formalism~\cite{rw57,z70}.
159: Separating angular variables with tensor harmonics of indices $(\ell,m)$,
160: the equations decouple to the even (or polar) part with parity
161: $(-1)^\ell$ 
162: under a transformation $(\theta,\phi) \to (\pi-\theta,\pi+\phi)$
163: and 
164: the odd (or axial) part with parity $(-1)^{\ell+1}$.
165: Seven equations for the even parity part are reduced to a single Zerilli
166: equation,
167: and the other three equations for the odd parity part to a single Regge-Wheeler equation.
168: The Zerilli equation is given by
169: \beqa
170: &&\left[-\frac{\partial^2}{\partial t^2}
171: +\frac{\partial^2}{\partial r_*^2}
172: -V_{Z}(r)\right] \psi^{(1)}_{\ell m}(t,r)=0 \,;
173: \\
174: &&V_{Z}(r)=\left(1-\frac{2}{r}\right)
175: \frac{2\lambda^2 (\lambda+1) r^3+6 \lambda^2 r^2+
176: 18 \lambda r +18}{r^3 (\lambda r+3)^2} \,,
177: \nonumber 
178: \eeqa
179: where $r_*=r+2 \ln \left(r/2-1\right)$ and
180: $\lambda=(\ell-1)(\ell+2)/2$. 
181: By Fourier transforming, 
182: $\psi^{(1)}_{\ell m}(t,r)=\int e^{-i\omega t} \psi^{(1)}_{\ell m
183: \omega}(r) d\omega$, the Zerilli equation gives a one-dimensional
184: scattering problem, which is familiar from quantum mechanics. The QNMs are
185: obtained by imposing the boundary conditions with purely ingoing waves
186: $\psi^{(1)}_{\ell m \omega} \sim e^{-i\omega r_*}$ at the horizon and purely
187: outgoing waves $\psi^{(1)}_{\ell m \omega} \sim e^{i\omega r_*}$ at
188: infinity. Such boundary conditions are
189: satisfied at discrete QNM frequencies $\omega^{(1)}_{\ell m n}$ 
190: that are complex with the
191: imaginary part representing the damping.
192: There is an infinite number of QNMs for each harmonics 
193: $(\ell,m)$~\cite{leaver85}.
194: 
195: All physical quantities for the first-order even parity part can be
196: reconstructed from $\psi^{(1)}_{\ell m}$. 
197: If we define 
198: \beqa
199: \psi^{(1)}_{\ell m}=\frac{r}{\lambda+1}
200: \left[K^{(1)}_{\ell m}+\frac{r-2M}{\lambda r+3M}
201: \left(H_{2\ell m}^{(1)}
202: -r \frac{\partial K_{\ell m}^{(1)}}{\partial r}\right)\right]
203: \eeqa
204: in the Regge-Wheeler (RW) gauge according to~\cite{Sago:2002fe,Martel:2005ir},
205: the GW power is given by
206: \beqa
207: \frac{dE^{(1)}}{dt}=\frac{1}{64\pi}\sum_{\ell m}
208: \frac{(\ell+2)!}{(\ell-2)!}
209: \left| \frac{\partial}{\partial t} \psi^{(1)}_{\ell m}(t,r) \right|^2 \,.
210: \label{eq:dEdt}
211: \eeqa
212: Note that in the RW gauge the gauge freedom is completely fixed
213: and the physical quantities can be expressed by
214: the gauge invariant functions in simple differential forms.
215: 
216: In this paper we focus on the most dominant modes in binary BH mergers,
217: i.e., the $\ell=2, m=2$ even parity mode for the first-order~\cite{Pretorius:2005gq}
218: and its driving second-order mode (the $\ell=4, m=4$ even parity mode as shown below).
219: Note that we have to specify $m$
220: since the degeneracy between $m$ breaks
221: at the second-order.
222: 
223: \section{Second-order}
224: For the second-order, we also separate angular
225: variables in terms of tensor harmonics. 
226: Instead of $\psi^{(2)}_{\ell m}$, 
227: we introduce a function
228: \begin{eqnarray}
229: \chi^{(2)}_{\ell m}=\frac{r-2M}{\lambda r+3M}
230: \left[\frac{r^2}{r-2M} 
231: \frac{\partial K^{(2)}_{\ell m}}{\partial t} - H_{1\ell m}^{(2)}\right]
232: \label{eq:defchi}
233: \end{eqnarray}
234: in the RW gauge for convenience, which is essentially the time-derivative of 
235: $\psi^{(2)}_{\ell m}$~\cite{Gleiser:1995gx}.
236: The first-order counterpart exactly
237: satisfies $\chi^{(1)}_{\ell m}=\partial
238: \psi^{(1)}_{\ell m}/\partial t$, and so the dimensions are
239: $\psi^{(i)}_{\ell m} \sim O(M)$ and 
240: $\chi^{(i)}_{\ell m} \sim O(M^0)$ ($i=1,\,2$).
241: The equations for the even parity part are reduced to the Zerilli equation with a
242: source term,
243: \begin{eqnarray}
244: \left[-\frac{\partial^2}{\partial t^2}
245: +\frac{\partial^2}{\partial r_*^2}
246: -V_{Z}(r)\right] \chi^{(2)}_{\ell m}(t,r)=S_{\ell m}(t,r) \,,
247: \label{eq:Zeqchi}
248: \end{eqnarray}
249: where the source term $S_{\ell m}$ is quadratic in $\psi^{(1)}_{\ell m}$.
250: 
251: The most dominant second-order mode is the $\ell=4$, $m=4$ even parity one.
252: This is because the dominant first-order mode is the 
253: $\ell=2$, $m=2$ even parity mode and hence
254: $S_{\ell m}$ is dominated by the product
255: $(\ell=2,m=2) \times (\ell=2,m=2)$ in Eq.~(\ref{eq:Zeqchi}).
256: This gives the $m=4$ mode that is $\ell=4$ and even parity.
257: Note that what we are considering is the particular solution.
258: Homogeneous solutions are just proportional to
259: the first-order ones and are therefore trivial.
260: 
261: \section{Regularization}
262: Although it is straightforward to calculate the
263: $\ell=4$, $m=4$ source term $S_{44}$ in terms of
264: $\psi^{(1)}_{22}$, the raw source term does not behave
265: well at infinity and is not suitable for calculations. We can
266: find $S_{44} \sim O(r^0)$ at infinity by using
267: the expansion
268: \begin{eqnarray}
269: \psi^{(1)}_{22}= 
270: \frac{1}{3}F_I''+\frac{1}{r}F_I'
271: +\frac{1}{r^2}\left(F_I-F_I'\right)+O(r^{-3}) \,,
272: \end{eqnarray}
273: where $F_I=F_I(T_{-})$ is some function of
274: $T_{-}=t-r_*$ and $F_I'$ denotes 
275: $dF_I(T_{-})/dT_{-}$. 
276: We need a regularized source $S_{44}^{reg} \sim O(r^{-2})$ at least 
277: [i.e., the same as the
278: potential $V_Z \sim O(r^{-2})$]. At the horizon
279: the source behaves well, $S_{44} \sim O[(r-2)]$, with
280: \begin{eqnarray}
281: \psi^{(1)}_{22}=F_H'+\frac{1}{4}F_H
282: +\frac{27}{56}(r-2)F_H
283:  +O[(r-2)^2] \,,
284: \end{eqnarray}
285: where $F_H=F_H(T_{+})$ is some function of $T_{+}=t+r_*$.
286: 
287: We can regularize the source term by using the regularized function,
288: \begin{eqnarray}
289: \chi^{(2) reg}_{44}
290: = \chi^{(2)}_{44}
291: -\frac{\sqrt{70}}{126\sqrt{\pi}}
292: \frac{(r-2)^2}{r}
293: \frac{\partial \psi^{(1)}_{22}}{\partial r}
294: \frac{\partial^2 \psi^{(1)}_{22}}{\partial r \partial t} \,,
295: \label{eq:chireg}
296: \end{eqnarray}
297: which satisfies the Zerilli equation (\ref{eq:Zeqchi})
298: with a well-behaved source term,
299: $S_{44}^{reg}\sim O(r^{-2})$ at infinity and 
300: $S_{44}^{reg}\sim O[(r-2)]$ at the horizon.
301: Thus we can remove an unphysical gauge-dependent divergence.
302: Note that such a regularization is not unique, 
303: and for example we can replace $\partial/\partial r$ with 
304: $\partial/\partial t$ in Eq.~(\ref{eq:chireg}).
305: The regularization is equivalent to
306: adding quadratic terms of the first-order gauge invariant
307: function to the second-order gauge invariant function,
308: so that it preserves the gauge 
309: invariance~\cite{Garat:1999vr}.
310: 
311: The explicit form of the regularized source term is
312: \begin{widetext}
313: \begin{eqnarray}
314: S^{reg}_{44}(t,r) &=&
315: \frac{r-2}{42} \frac{\sqrt{70}}{\sqrt {\pi }}
316: \Biggl\{  
317: {\frac { 228{r}^{7}+8{r}^{6}-370{r}^{5}+142{r}^{4}-384{r}^{3}
318: - 514{r}^{2}-273r-48 }
319: {{r}^{5} \left( 3r+1 \right) ^{2} \left( 2r+3 \right) ^{2}}}
320: \dot \psi'
321: \psi'
322: \cr
323: &-& {\frac {72{r}^{8}+3936{r}^{7}+2316{r}^{6}
324: - 2030{r}^{5}
325: -7744{r}^{4}-9512{r}^{3}-3540{r}^{2}
326: -1119r-144}
327: {{r}^{6} \left( 3r+1 \right) ^{2} 
328: \left( 2r+3 \right) ^{3}}}
329: \psi'
330: \dot \psi
331: \cr
332: &+& {\frac { \left( 7r+4 \right) r  }{3 \left( r-2 \right) ^{2}}}
333: \dddot \psi
334: \ddot \psi
335: +{\frac {24{r}^{7}+344{r}^{6}-872{r}^{5}-771{r}^{4}
336: +120{r}^{3}+77{r}^{2}-237r-48
337: }
338: {{r}^{3} \left( 3r+1
339:  \right) ^{2} \left( 2r+3 \right) ^{2} \left( r-2 \right) ^{2}}}
340: \ddot \psi
341: \dot \psi
342: \cr
343: &-& {\frac { 66{r}^{4}-106{r}^{3}-220{r}^{2}-156r-45 }
344: {r \left( 3r+1
345:  \right) ^{2} \left( 2r+3 \right)  \left( r-2 \right) }}
346: \ddot \psi'
347: \dot \psi
348: +{\frac { 
349: 198{r}^{5}-318{r}^{4}-664{r}^{3}-458{r}^{2}-127r-24 }
350: {3{r}^{2} \left( 3r+1 \right) ^{2}
351:  \left( 2r+3 \right)  \left( r-2 \right) }} 
352: {\dddot{\psi}}
353: \psi'
354: \cr
355: &-& \frac {3 \left(2160{r}^{9}+11760{r}^{8}
356: +30560{r}^{7}+41124{r}^{6}
357: + 31596{r}^{5}+11630{r}^{4}
358: -1296{r}^{3}-4182{r}^{2}-1341r-144 \right) }
359: {{r}^{7} \left( 2r+3 \right) ^{4} \left( 3r+1 \right) ^{2}}
360: \psi 
361: \dot \psi
362: \cr
363: &-& {\frac { 7r+4 }{3r}}
364: \ddot \psi'
365: \dot \psi'
366: - {\frac {
367:  2 \left( r-2 \right) }{3 \left( 3r+1 \right) ^{2}{r}^{2}}}
368: \ddot \psi
369: \dot \psi'
370: + {\frac { 252{r}^{6}+636{r}^{5}+674{r}^{4}
371: +730{r}^{3}+524{r}^{2}+171r+24 }
372: {{r}^{3} \left( 3r+1 \right) ^{2}
373:  \left( 2r+3 \right) ^{2} \left( r-2 \right) }} 
374: \psi 
375: \dddot \psi
376: \cr
377: &-& {\frac { 216{r}^{8}+4296{r}^{7}
378: +1992{r}^{6}
379: -3488{r}^{5}
380: - 8716{r}^{4}-9512{r}^{3} 
381: -3540{r}^{2}-1119r
382: -144 }
383: {{r}^{6} \left( 3r+1 \right) ^{2} \left( 2r+3 \right) ^{3}}} 
384: \psi 
385: \dot \psi'
386: \Biggr\} \,,
387: \label{eq:Sreg}
388: \end{eqnarray}
389: \end{widetext}
390: where $\dot \psi=\partial \psi^{(1)}_{22}/\partial t$ and 
391: $\psi'=\partial \psi^{(1)}_{22}/\partial r$. 
392: We can now solve the regularized Zerilli equation~(\ref{eq:Zeqchi}). 
393: With Fourier expansions, this provides a two-point
394: boundary value problem with purely ingoing boundary condition 
395: at the horizon and purely outgoing at infinity.
396: The numerical calculation will be presented 
397: in a forthcoming paper.
398: 
399: \section{Detectability}
400: Without solving Eq.~(\ref{eq:Zeqchi}) numerically,
401: we can find the essential properties of the solutions, 
402: i.e., the QNM frequency $\omega^{(2)}_{44 n}$
403: and the order-of-magnitude QNM amplitude. 
404: Since the source term $S^{(2) reg}_{44}$ 
405: is quadratic in the first-order function 
406: $\psi^{(1)}_{22} \propto e^{-i\omega^{(1)}_{22 n} t}$,
407: we have $\chi_{44}^{(2) reg} \propto e^{-2 i\omega^{(1)}_{22 n} t}$
408: from Eq.~(\ref{eq:Zeqchi}).
409: Therefore the second-order QNM frequencies are
410: twice the first-order ones,
411: \beqa
412: \omega^{(2)}_{44 n}=2 \omega^{(1)}_{22 n} \,,
413: \eeqa
414: which differ from any first-order QNM
415: frequencies $\omega^{(1)}_{\ell m n}$.
416: By matching the dimensions in both sides of Eq.~(\ref{eq:Zeqchi}),
417: we also estimate the order-of-magnitude amplitude as
418: \beqa
419: \left|\chi^{(2) reg}_{44}\right| 
420: \sim \left|\omega^{(1)}_{22 n} \psi^{(1)}_{22} \right|^2 \,.
421: \label{eq:ampli}
422: \eeqa
423: Then we can derive the second-order QNM energy $E^{(2)}$
424: from the first-order one $E^{(1)}$.
425: For the QNM waveform, $\psi^{(1)}_{22}=\psi^{(1)}_{22}(0) 
426: e^{-i\omega^{(1)}_{22 n} t}$,
427: the first-order energy $E^{(1)}$ in Eq.~(\ref{eq:dEdt}) is given by
428: \beqa
429: \frac{E^{(1)}}{M} = \frac{3}{8\pi} 
430: \frac{\left|\omega^{(1)}_{22 n}\right|^2 
431: \left|\psi^{(1)}_{22}(0)\right|^2}
432: {M \left|\Im \omega^{(1)}_{22 n}\right|} \,,
433: \label{eq:E1}
434: \eeqa
435: where $E^{(1)}/M \sim 1\%$ 
436: for equal-mass mergers \cite{Pretorius:2005gq}.
437: By noting $\chi^{(2)}_{44} \sim \partial \psi^{(2)}_{44}/\partial t 
438: \sim 2 \omega^{(1)}_{22 n} \psi^{(2)}_{44}$,
439: we have a similar expression for the second-order energy $E^{(2)}$ as
440: \beqa
441: \frac{E^{(2)}}{M} 
442: \sim \frac{45}{8\pi}
443: \frac{\left|\chi^{(2) reg}_{44}(0)\right|^2}
444: {M \left|2 \Im\omega^{(1)}_{22 n}\right|}
445: \sim M |\Im \omega^{(1)}_{22 n}|
446: \left(\frac{E^{(1)}}{M}\right)^2 \,,
447: \label{eq:E2}
448: \eeqa
449: where the second equality uses Eqs.~(\ref{eq:ampli}) and (\ref{eq:E1}).
450: In a next paper we will calculate
451: the coefficient as $E^{(2)}/M \sim 15 M |\Im \omega^{(1)}_{22 n}|
452: (E^{(1)}/M)^2$.
453: 
454: Once we know the QNM frequency $\omega$ and energy $E$,
455: we can obtain the GW energy 
456: spectrum~\cite{Flanagan:1997sx,Berti:2005ys}
457: \beqa
458: \frac{dE}{df} 
459: \simeq \frac{16\pi^2 E f^2 |\Im \omega|^3}
460: {|\omega|^2 \left[(2\pi f-\Re \omega)^2+(\Im \omega)^2\right]^2} \,,
461: \label{eq:dEdf}
462: \eeqa
463: and then the characteristic amplitude $h_{char}(f)$ 
464: with Eq.~(5.1) of Flanagan and Hughes \cite{Flanagan:1997sx},
465: \beqa
466: h_{char}(f)^2 = \frac{2(1+z)^2}{\pi^2 D(z)^2} \frac{dE}{df}[(1+z)f] \,,
467: \label{eq:hchar}
468: \eeqa
469: where $D(z)$ is the luminosity distance.
470: %In Fig.~\ref{fig:snr} we show $h_{char}(f)$ 
471: %for a binary BH merger of total mass $10^6 M_{\odot}$
472: %at redshift $z=5$ together with
473: %the rms noise amplitude $h_n(f)$ for 
474: %LISA \cite{Flanagan:1997sx,Berti:2005ys}
475: %and ultimate DECIGO \cite{Seto:2001qf}.
476: %We can see that the second and even higher-order QNMs
477: %are detectable.
478: 
479: \section{Discussions}
480: Our main results are summarized in Fig.~\ref{fig:snr} showing the
481: characteristic GW amplitudes $h_{char}(f)$ of the Schwarzschild BH
482: QNMs for two equal-mass binary BH mergers
483: of total mass $10^6 M_{\odot}$ at redshift $z=5$, together with the rms
484: noise amplitude $h_n(f)=\sqrt{5 f S_h(f)}$ for the space-based detector 
485: LISA~\cite{Flanagan:1997sx,Berti:2005ys}
486: and ultimate DECIGO~\cite{Seto:2001qf}. The first-order QNM has $\ell=2,
487: m=2$ and energy $E^{(1)}/M=1\%$. We estimate the second-order QNM with
488: Eqs.~(\ref{eq:E2})--(\ref{eq:hchar}) and the third-order QNM energy by
489: extrapolating the second-order equation (\ref{eq:E2}) as $E^{(3)}/M \sim
490: 15 (M \Im \omega)^2 (E^{(1)}/M)^3$. We can see that the second and
491: third-order QNMs appear at twice and three times the first-order
492: frequency, respectively, with detectable amplitudes.
493: 
494: \begin{figure}
495:   \begin{center}
496: \epsfxsize=8.cm
497: \begin{minipage}{\epsfxsize} 
498: \epsffile{f1.eps} 
499: \end{minipage}
500:   \caption{
501: The characteristic GW amplitudes $h_{char}(f)$
502: of the Schwarzschild black hole (BH) 
503: quasi-normal modes (QNMs)
504: for two equal-mass binary BH mergers 
505: of total mass $10^6 M_{\odot}$ at redshift $z=5$.
506: The first-order QNM has $\ell=2, m=2$ and energy $E^{(1)}/M=1\%$.
507: The second and third-order QNMs also appear
508: at twice and three times the first-order frequency,
509: respectively, with detectable amplitudes.
510: The dashed line is the rms noise amplitude 
511: $h_n(f)=\sqrt{5 f S_h(f)}$ for 
512: the space-based detector LISA and ultimate DECIGO.
513: The signal-to-noise ratio squared for a randomly
514: oriented source is given by 
515: $(SNR)^2=\int d(\ln f)[h_{char}(f)/h_n(f)]^2$.
516: We use a cosmology $(\Omega_m,\Omega_\Lambda,h)=(0.27,0.73,0.71)$.
517: We estimate the third-order QNM energy 
518: by extrapolating the second-order equation (\ref{eq:E2})
519: as $E^{(3)}/M \sim 15 (M \Im \omega)^2 (E^{(1)}/M)^3$.
520: }
521: \label{fig:snr}
522:  \end{center}
523: \end{figure}
524: 
525: We can in principle identify the higher-order QNMs
526: since their frequencies differ from any first-order ones.
527: However the actual identification
528: depends on the SNR of the observations~\cite{Berti:2005ys}
529: or the accuracy of the simulations.
530: In simulations with current accuracy
531: we may have already mistaken the second-order QNMs for
532: the first-order ones.
533: For example $2 M \omega^{(1)}_{2m0}=0.7473-0.1779i$ is close to
534: $M \omega^{(1)}_{4m0}=0.80918-0.09416 i$~\cite{leaver85}.
535: 
536: In order to proof that the second-order QNMs actually exist,
537: we have to find the second-order QNMs directly
538: in the numerical simulations.
539: Such simulations are challenging because the mesh size
540: should be less than $\sim 1\% \times M$
541: to resolve $\sim 1\%$ metric perturbations.
542: Even 1+1D spherical models for the fully self-gravitating case 
543: have not found the second-order QNMs~\cite{Zlochower:2003yh}.
544: Simulations of acoustic black holes may be an alternative
545: for this purpose~\cite{Okuzumi:2007hf}.
546: We also need a mathematically rigorous definition of second-order
547: QNMs like the first-order ones that use the Laplace transformation
548: rather than the Fourier transformation~\cite{leaver86,Kokkotas:1999bd}.
549: 
550: Future problems include the Kerr BH case, the odd parity mode case,
551: and a more solid third-order formulation.
552: Since the master equation for Kerr BHs also has a source term 
553: quadratic in the first-order function~\cite{Campanelli:1998jv}, 
554: we may expect similar results. When BHs have no spin before mergers,
555: the final spin is $a \sim 0.7$~\cite{Pretorius:2005gq}
556: and hence the Kerr effects may not be so large
557: (as inferred from the fact that the QNM frequencies shift 
558: by only a small factor).
559: The odd parity mode appears when BHs have spin before mergers.
560: At the third-order the QNM frequencies will also shift up to
561: $(\psi^{(1)}/M)^2\sim 1\%$
562: as suggested by the anharmonic oscillations in~\cite{landau76},
563: probably blueward because the GW carries away the BH mass.
564: 
565: The ratio between the first and higher-order QNM amplitude
566: include new information about the total GW energy $E$.
567: Since the observed GW amplitude is $h \sim (E/M)^{1/2}(M/r)$,
568: this could provide a distance indicator.
569: 
570: 
571: \acknowledgments
572: %We thank ... for useful comments.
573: We thank C.~O.~Lousto and Y.~Zlochower for a careful reading of the manuscript.
574: This work is supported in part by
575: the Grant-in-Aid (18740147) from the 
576: Ministry of Education, Culture, Sports, Science and Technology
577: (MEXT) of Japan (KI). 
578: This is also supported by JSPS Postdoctoral Fellowships 
579: for Research Abroad and in part by the NSF 
580: for financial support from grant PHY-0722315 (HN).
581: 
582: %\include{ref}
583: 
584: %%%%%%%%% references %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
585: \begin{thebibliography}{DUM}
586: 
587: %\cite{Seto:2001qf}
588: \bibitem{Seto:2001qf}
589:   N.~Seto, S.~Kawamura and T.~Nakamura,
590:   %``Possibility of direct measurement of the acceleration of the universe
591:   %using 0.1-Hz band laser interferometer gravitational wave antenna in
592:   %space,''
593:   Phys.\ Rev.\ Lett.\  {\bf 87}, 221103 (2001).
594:   %%CITATION = PRLTA,87,221103;%%
595: 
596: %\cite{Kokkotas:1999bd}
597: \bibitem{Kokkotas:1999bd}
598:   K.~D.~Kokkotas and B.~G.~Schmidt,
599:   %``Quasi-normal modes of stars and black holes,''
600:   Living Rev.\ Rel.\  {\bf 2}, 2 (1999).
601:   %%CITATION = 00222,2,2;%%
602: 
603: %\cite{Flanagan:1997sx}
604: \bibitem{Flanagan:1997sx}
605:   E.~E.~Flanagan and S.~A.~Hughes,
606:   %``Measuring gravitational waves from binary black hole coalescences.  I:
607:   %Signal to noise for inspiral, merger, and ringdown,''
608:   Phys.\ Rev.\  D {\bf 57}, 4535 (1998).
609:   %%CITATION = PHRVA,D57,4535;%%
610: 
611: %\cite{Enoki:2004ew}
612: \bibitem{Enoki:2004ew}
613:   M.~Enoki, K.~T.~Inoue, M.~Nagashima and N.~Sugiyama,
614:   %``Gravitational waves from supermassive black hole coalescence in a
615:   %hierarchical galaxy formation model,''
616:   Astrophys.\ J.\  {\bf 615}, 19 (2004).
617:   %%CITATION = ASJOA,615,19;%%
618: 
619: %\cite{Ioka:2005pm}
620: \bibitem{Ioka:2005pm}
621:   K.~Ioka and P.~Meszaros,
622:   %``Spatial Correlation of Massive Black Hole Mergers: Probing the Formation
623:   %Mechanism and the Reionization,''
624:   Astrophys.\ J.\  {\bf 635}, 143 (2005).
625: %  [arXiv:astro-ph/0502437].
626:   %%CITATION = ASJOA,635,143;%%
627: 
628: %\cite{Pretorius:2005gq}
629: \bibitem{Pretorius:2005gq}
630:   F.~Pretorius,
631:   %``Evolution of binary black hole spacetimes,''
632:   Phys.\ Rev.\ Lett.\  {\bf 95}, 121101 (2005);
633:   %%CITATION = PRLTA,95,121101;%%
634: %\cite{Campanelli:2005dd}
635: %\bibitem{Campanelli:2005dd}
636:   M.~Campanelli, C.~O.~Lousto, P.~Marronetti and Y.~Zlochower,
637:   %``Accurate evolutions of orbiting black-hole binaries without excision,''
638:   Phys.\ Rev.\ Lett.\  {\bf 96}, 111101 (2006);
639: %  [arXiv:gr-qc/0511048].
640:   %%CITATION = PRLTA,96,111101;%%
641: %\cite{Baker:2005vv}
642: %\bibitem{Baker:2005vv}
643:   J.~G.~Baker, J.~Centrella, D.~I.~Choi, M.~Koppitz and J.~van Meter,
644:   %``Gravitational wave extraction from an inspiraling configuration of  merging
645:   %black holes,''
646:   Phys.\ Rev.\ Lett.\  {\bf 96}, 111102 (2006).
647: %  [arXiv:gr-qc/0511103].
648:   %%CITATION = PRLTA,96,111102;%%
649: 
650: %\cite{Berti:2007fi}
651: \bibitem{Berti:2007fi}
652:   E.~Berti, V.~Cardoso, J.~A.~Gonzalez, U.~Sperhake, M.~Hannam, S.~Husa and B.~Bruegmann,
653:   %``Inspiral, merger and ringdown of unequal mass black hole binaries: A
654:   %multipolar analysis,''
655:   arXiv:gr-qc/0703053.
656:   %%CITATION = GR-QC/0703053;%%
657: 
658: \bibitem{landau76}
659:   L.~D.~Landau and E.~M.~Lifshitz,
660:   Mechanics (Butterworth, Heinemann, Oxford, 1976), 3rd ed.
661: 
662: \bibitem{Tomita}
663: K.~Tomita, Prog.\ Theor.\ Phys.\  {\bf 52}, 1188 (1974);
664: K.~Tomita and N.~Tajima, Prog.\ Theor.\ Phys.\  {\bf 56}, 551 (1976).
665: 
666: %\cite{Gleiser:1995gx}
667: \bibitem{Gleiser:1995gx}
668:   R.~J.~Gleiser, C.~O.~Nicasio, R.~H.~Price and J.~Pullin,
669:   %``Second order perturbations of a Schwarzschild black hole,''
670:   Class.\ Quant.\ Grav.\  {\bf 13}, L117 (1996);
671: %  [arXiv:gr-qc/9510049].
672:   %%CITATION = CQGRD,13,L117;%%
673: %\cite{Gleiser:1996yc}
674: %\bibitem{Gleiser:1996yc}
675:   R.~J.~Gleiser, C.~O.~Nicasio, R.~H.~Price and J.~Pullin,
676:   %``Colliding black holes: How far can the close approximation go?,''
677:   Phys.\ Rev.\ Lett.\  {\bf 77}, 4483 (1996);
678: %  [arXiv:gr-qc/9609022].
679:   %%CITATION = PRLTA,77,4483;%%
680: %\cite{Gleiser:1998rw}
681: %\bibitem{Gleiser:1998rw}
682:   R.~J.~Gleiser, C.~O.~Nicasio, R.~H.~Price and J.~Pullin,
683:   %``Gravitational radiation from Schwarzschild black holes: The second  order
684:   %perturbation formalism,''
685:   Phys.\ Rept.\  {\bf 325}, 41 (2000);
686:   %%CITATION = PRPLC,325,41;%%
687: %\cite{Nicasio:2000ge}
688: %\bibitem{Nicasio:2000ge}
689:   C.~O.~Nicasio, R.~Gleiser and J.~Pullin,
690:   %``Second order perturbations of a Schwarzschild black hole: inclusion of odd
691:   %parity perturbations,''
692:   Gen.\ Rel.\ Grav.\  {\bf 32}, 2021 (2000).
693:   %%CITATION = GRGVA,32,2021;%%
694: 
695: %\cite{Acquaviva:2002ud}
696: \bibitem{Acquaviva:2002ud}
697:   V.~Acquaviva, N.~Bartolo, S.~Matarrese and A.~Riotto,
698:   %``Second-order cosmological perturbations from inflation,''
699:   Nucl.\ Phys.\  B {\bf 667}, 119 (2003);
700: %  [arXiv:astro-ph/0209156].
701:   %%CITATION = NUPHA,B667,119;%%
702: %\cite{Matarrese:1997ay}
703: %\bibitem{Matarrese:1997ay}
704:   S.~Matarrese, S.~Mollerach and M.~Bruni,
705:   %``Second-order perturbations of the Einstein-de Sitter universe,''
706:   Phys.\ Rev.\  D {\bf 58}, 043504 (1998);
707: %  [arXiv:astro-ph/9707278].
708:   %%CITATION = PHRVA,D58,043504;%%
709: %\cite{Tomita:2005et}
710: %\bibitem{Tomita:2005et}
711:   K.~Tomita,
712:   %``Relativistic second-order perturbations of nonzero-Lambda flat
713:   %cosmological models and CMB anisotropies,''
714:   Phys.\ Rev.\  D {\bf 71}, 083504 (2005);
715: %  [arXiv:astro-ph/0501663].
716:   %%CITATION = PHRVA,D71,083504;%%
717: %\cite{Nakamura:2004rm}
718: %\bibitem{Nakamura:2004rm}
719:   K.~Nakamura,
720:   %``Second-order gauge invariant cosmological perturbation theory: Einstein
721:   %equations in terms of gauge invariant variables,''
722:   Prog.\ Theor.\ Phys.\  {\bf 117}, 17 (2007).
723: %  [arXiv:gr-qc/0605108].
724:   %%CITATION = PTPKA,117,17;%%
725: 
726: %\cite{Campanelli:1998jv}
727: \bibitem{Campanelli:1998jv}
728:   M.~Campanelli and C.~O.~Lousto,
729:   %``Second order gauge invariant gravitational perturbations of a Kerr  black
730:   %hole,''
731:   Phys.\ Rev.\  D {\bf 59}, 124022 (1999).
732: %  [arXiv:gr-qc/9811019].
733:   %%CITATION = PHRVA,D59,124022;%%
734: 
735: %\cite{Bruni:1996im}
736: \bibitem{Bruni:1996im}
737:   M.~Bruni, S.~Matarrese, S.~Mollerach and S.~Sonego,
738:   %``Perturbations of spacetime: Gauge transformations and gauge invariance  at
739:   %second order and beyond,''
740:   Class.\ Quant.\ Grav.\  {\bf 14}, 2585 (1997);
741:   %%CITATION = CQGRD,14,2585;%%
742: %\cite{Nakamura:2003wk}
743: %\bibitem{Nakamura:2003wk}
744:   K.~Nakamura,
745:   %``Gauge invariant variables in two-parameter nonlinear perturbations,''
746:   Prog.\ Theor.\ Phys.\  {\bf 110}, 723 (2003);
747:   %%CITATION = PTPKA,110,723;%%
748: %\cite{Brizuela:2007we}
749: %\bibitem{Brizuela:2007we}
750:   D.~Brizuela, J.~M.~Martin-Garcia and G.~A.~M.~Marugan,
751:   %``High-order gauge-invariant perturbations of a spherical spacetime,''
752:   arXiv:gr-qc/0703069.
753:   %%CITATION = GR-QC/0703069;%%
754: 
755: \bibitem{rw57}
756: T.~Regge and J.~Wheeler,
757: Phys.\ Rev.\ {\bf 108}, 1063 (1957).
758: 
759: \bibitem{z70}
760: F.~J.~Zerilli,
761: Phys.\ Rev.\ D {\bf 2}, 2141 (1970).
762: 
763: \bibitem{leaver85}
764: E.~W.~Leaver,
765: Proc.\ R.\ Soc.\ A\ {\bf 402}, 285 (1985).
766: 
767: %\cite{Sago:2002fe}
768: \bibitem{Sago:2002fe}
769:   N.~Sago, H.~Nakano and M.~Sasaki,
770:   %``Gauge problem in the gravitational self-force. I: Harmonic gauge  approach
771:   %in the Schwarzschild background,''
772:   Phys.\ Rev.\  D {\bf 67}, 104017 (2003).
773: %  [arXiv:gr-qc/0208060].
774:   %%CITATION = PHRVA,D67,104017;%%
775: 
776: %\cite{Martel:2005ir}
777: \bibitem{Martel:2005ir}
778:   K.~Martel and E.~Poisson,
779:   %``Gravitational perturbations of the Schwarzschild spacetime: A practical
780:   %covariant and gauge-invariant formalism,''
781:   Phys.\ Rev.\  D {\bf 71}, 104003 (2005).
782: %  [arXiv:gr-qc/0502028].
783:   %%CITATION = PHRVA,D71,104003;%%
784: 
785: %\cite{Garat:1999vr}
786: \bibitem{Garat:1999vr}
787:   A.~Garat and R.~H.~Price,
788:   %``Gauge invariant formalism for second order perturbations of  Schwarzschild
789:   %spacetimes,''
790:   Phys.\ Rev.\  D {\bf 61}, 044006 (2000).
791:   %%CITATION = PHRVA,D61,044006;%%
792: 
793: %\cite{Berti:2005ys}
794: \bibitem{Berti:2005ys}
795:   E.~Berti, V.~Cardoso and C.~M.~Will,
796:   %``On gravitational-wave spectroscopy of massive black holes with the  space
797:   %interferometer LISA,''
798:   Phys.\ Rev.\  D {\bf 73}, 064030 (2006).
799: %  [arXiv:gr-qc/0512160].
800:   %%CITATION = PHRVA,D73,064030;%%
801: 
802: %\cite{Zlochower:2003yh}
803: \bibitem{Zlochower:2003yh}
804:   Y.~Zlochower, R.~Gomez, S.~Husa, L.~Lehner and J.~Winicour,
805:   %``Mode coupling in the nonlinear response of black holes,''
806:   Phys.\ Rev.\  D {\bf 68}, 084014 (2003).
807: %  [arXiv:gr-qc/0306098].
808:   %%CITATION = PHRVA,D68,084014;%%
809: 
810: %\cite{Okuzumi:2007hf}
811: \bibitem{Okuzumi:2007hf}
812:   S.~Okuzumi and M.~a.~Sakagami,
813:   %``Quasinormal ringing of acoustic black holes in Laval nozzles: Numerical
814:   %simulations,''
815:   arXiv:gr-qc/0703070.
816:   %%CITATION = GR-QC/0703070;%%
817: 
818: \bibitem{leaver86}
819: E.~W.~Leaver,
820: Phys.\ Rev.\ D {\bf 34}, 384 (1986).
821: 
822: %\bibitem{chandra75}
823: %S.~Chandrasekhar,
824: %Proc.\ R.\ Soc.\ A\ {\bf 343}, 289 (1975).
825: 
826: \end{thebibliography}
827: 
828: \end{document}