1: % ****** Start of file apssamp.tex ******
2: %
3: % This file is part of the APS files in the REVTeX 4 distribution.
4: % Version 4.0 of REVTeX, August 2001
5: %
6: % Copyright (c) 2001 The American Physical Society.
7: %
8: % See the REVTeX 4 README file for restrictions and more information.
9: %
10: % TeX'ing this file requires that you have AMS-LaTeX 2.0 installed
11: % as well as the rest of the prerequisites for REVTeX 4.0
12: %
13: % See the REVTeX 4 README file
14: % It also requires running BibTeX. The commands are as follows:
15: %
16: % 1) latex apssamp.tex
17: % 2) bibtex apssamp
18: % 3) latex apssamp.tex
19: % 4) latex apssamp.tex
20: %
21: %\documentclass[twocolumn,showpacs,preprintnumbers,amsmath,amssymb,floatfix,eqsecnum]{revtex4}
22: %\documentclass[preprint,showpacs,preprintnumbers,amsmath,amssymb,floatfix]{revtex4}
23: %\documentclass[aps,prd,showpacs,eqsecnum,floatfix]{revtex4}
24: \documentclass[12pt,preprint,usenatbib,eqsecnum]{aastex}
25: % Some other (several out of many) possibilities
26: %\documentclass[preprint,aps]{revtex4}
27: %\documentclass[preprint,aps,draft]{revtex4}
28: %\documentclass[prb]{revtex4}% Physical Review B
29:
30: \usepackage{graphicx}% Include figure files
31: \usepackage{dcolumn}% Align table columns on decimal point
32: \usepackage{bm}% bold math
33: \usepackage[]{natbib}
34: %\bibpunct{(}{)}{,}{a}{}{,}
35: %\nofiles
36: \bibliographystyle{apj}
37:
38:
39: \begin{document}
40:
41: %\preprint{APS/}
42:
43: \title{Perturbative Analysis of Universality and Individuality\\ in Gravitational Waves from Neutron Stars}
44: % Force line breaks with \\
45:
46: \author{L.K. Tsui and P.T. Leung\footnote{Email:
47: ptleung@phy.cuhk.edu.hk} }
48: \affil{%
49: Physics Department and Institute of Theoretical Physics,\\ The
50: Chinese University of Hong Kong,\\ Shatin, Hong Kong SAR, China.
51: }%
52:
53:
54: \date{\today}% It is always \today, today,
55: % but any date may be explicitly specified
56: \def\tomega{ \tilde{\omega} }
57: \def\tOmega{ \tilde{\omega} }
58: \def\tr{ \tilde{r} }
59: \def\tx{ \tilde{r}_* }
60: \def\tV{ \tilde{V} }
61: \def\tpsi{ \tilde{\psi} }
62: \def\trho{ \tilde{\rho} }
63: \def\tP{ \tilde{P} }
64: \def\tR{ \tilde{R} }
65: \def\tX{ \tilde{R}_* }
66: \def\tm{ \tilde{m} }
67: \def\tphi{ \tilde{\nu} }
68: \def\tlam{ \tilde{\lambda}}
69: \def\tepsilon { \tilde{\epsilon}}
70: \def\tU{ \tilde{U} }
71: \def\tphi{ \tilde{\nu} }
72: \def\tlam{ \tilde{\lambda}}
73: \def\tepsilon { \tilde{\epsilon}}
74: \def\cc{{\cal C}}
75: \def\a{ {\rm a}}
76: \def\c{ {\rm c}}
77: \def\e{ {\rm e}}
78: \def\p{ {\rm p}}
79: \def\f{ {\rm f}}
80: \def\d{ {\rm d}}
81: \def\i{ {\rm i}}
82: \def\r{ {\rm r}}
83: \begin{abstract}
84: The universality observed in gravitational wave spectra of
85: non-rotating neutron stars is analyzed here. We show that the
86: universality in the axial oscillation mode can be reproduced with
87: a simple stellar model, namely the centrifugal barrier
88: approximation (CBA), which captures the essence of the Tolman VII
89: model of compact stars. Through the establishment of scaled
90: co-ordinate logarithmic perturbation theory (SCLPT), we are able
91: to explain and quantitatively predict such universal behavior. In
92: addition, quasi-normal modes of individual neutron stars
93: characterized by different equations of state can be obtained from
94: those of CBA with SCLPT.
95: \end{abstract}
96:
97: %\pacs{04.40.Dg, 04.30.Db, 97.60.Jd, 95.30.Sf}
98: % PACS, the Physics and Astronomy Classification Scheme.
99: \keywords{gravitational waves --- stars: neutron --- stars:
100: oscillations (including pulsations) --- equation of state ---
101: relativity
102: }
103: %Use showkeys class option if keyword display desired
104:
105: \maketitle
106: \section{Introduction}
107: As is well known, neutron stars of high densities are ideal
108: extra-terrestrial test beds for theories of nuclear matters, quark
109: matters and high energy physics \citep[see e.g.][and references
110: therein]{ComStar}. For instance, the possible existence of the
111: quark star \citep[see e.g.][]{Cheng}, a variant of the neutron
112: star, could lend direct support to the theory of quark matter
113: \citep{MITBM,Witten,AFO,PBP}. Besides, the relation between the
114: mass and the radius of a neutron star could determine, up to
115: certain accuracy, the equation of state (EOS) of relevant nuclear
116: matter through the process of inversion \citep{Lindblom_invert}.
117: Therefore, a comprehensive study of observed data and statistics
118: of neutron stars is likely to lead to fruitful results in the
119: fields mentioned above.
120:
121: Traditional means to gather information about neutron stars have
122: so far relied on electromagnetic waves emitted from them. However,
123: with the advent of gravitational wave detectors of various designs
124: \citep[see e.g.][and references therein]{Hughes_03,Grishchuk},
125: including resonant antennas (e.g. EXPLOPER and NIBOE),
126: ground-based interferometers (e.g. LIGO and VIRGO), and the
127: space-based interferometer LISA, it is generally believed that
128: neutron stars, as promising gravitational wave emitters, can be
129: observed and analyzed in the gravitational wave channel within one
130: or two decades. For example, it has been argued that the
131: frequency of detection of gravitational waves emitted in the
132: mergers of binary neutron stars could be as high as several
133: hundreds per year in the near future
134: \citep{Belczynski:2001uc,Hughes_03}. Besides, gravitational waves
135: could also be emitted during the formation of neutron stars
136: \citep{Lindblom:1998wf,Fryer:2001zw}. Being spurred on by the
137: possibility of inferring the internal structure of neutron stars
138: from the gravitational wave signals emitted by them, researchers
139: have been actively examining the peculiarities embedded in such
140: signals
141: \citep{Andersson_1996,Andersson1998,Ferrari,Kokkotas_2001}.
142:
143: Despite that the evaluation of gravitational waves emitted in
144: violent stellar activities such as binary mergers and asymmetric
145: core collapse has indeed posed a grand challenge to researchers in
146: numerical relativity, linearized theory of pulsating neutron stars
147: pioneered by \citet{Thorne} can still provide useful insight into
148: such complex situations. By analogy of radiating electric circuits
149: \citep{ToyModel}, linearized gravitational waves are analyzed in
150: terms of quasi-normal modes (QNMs), which are damped harmonic
151: pulsations with time dependence $\exp(\i\omega t)$ and are
152: characterized by complex eigenfrequencies
153: $\omega=\omega_\r+\i\omega_\i$
154: \citep{Press_1971,Leaver_1986,Ching,Kokkotas_rev,Nollert_rev}. It
155: is believed that the QNM frequencies of a pulsating neutron star
156: can reflect the physical characteristics of the star, such as its
157: mass $M$, radius $R$ and EOS as well
158: \citep{Andersson_1996,Andersson1998,Ferrari,Kokkotas_2001}.
159: However, \citet{Andersson1998} and \citet{Ferrari} also noted that
160: the frequency $\omega_\r$ and the damping time $\tau\equiv
161: 1/\omega_\i$ of the leading axial and polar $w$-modes of
162: non-rotating neutron stars follow approximately the equations:
163: \begin{eqnarray}
164: \omega_\r & \approx & \frac{1}{R}\left[a_\r
165: \left(\frac{M}{R}\right)+b_\r
166: \right] ,\label{BBF1}\\
167: \frac{1}{\tau}& \approx & \frac{1}{M}\left[a_\i
168: \left(\frac{M}{R}\right)^2+b_\i\left(\frac{M}{R}\right)+c_\i
169: \right],\label{BBF2}
170: \end{eqnarray}
171: where $a_\r$, $b_\r$, $a_\i$, $b_\i$ and $c_\i$ are
172: model-independent constants determined from curve fitting. These
173: universal behaviors seem to undermine the possibility of inferring
174: the characteristics of a neutron star from its $w$-mode QNMs.
175: Instead, they are indicative of a common feature shared among
176: neutron stars with different EOS's.
177:
178: To seek the physical mechanism underlying the universality in the
179: gravitational wave of neutron stars, we showed in a recent paper
180: \citep{preprint} that (i) The scaled complex eigenfrequencies
181: $\tomega \equiv M \omega$ of $w$-mode oscillations, to a good
182: approximation, depend only on the compactness $\cc \equiv M/R$.
183: (ii) Eqs.~(\ref{BBF1}) and (\ref{BBF2}) are equivalent to a single
184: formula of $\tomega$:
185: \begin{eqnarray}
186: \tomega \approx a
187: \left(\frac{M}{R}\right)^2+b\left(\frac{M}{R}\right)+c
188: ,\label{TLA}
189: \end{eqnarray}
190: where $a$, $b$ and $c$ are complex constants. (iii) The
191: universality indicated by (\ref{TLA}) is a direct reflection of
192: the fact that the the mass distribution inside neutron stars with
193: different EOS's can be nicely approximated by the Tolman VII model
194: (TVIIM) proposed by \citet{Tolman:1939jz}.
195:
196: In this paper we further pinpoint the crux of the mechanism
197: leading to the universal behavior (\ref{TLA}) for axial $w$-mode
198: pulsations. Inspired by the success of TVIIM, we propose here yet
199: another simple approximation, namely the centrifugal barrier
200: approximation (CBA), in which the potential in the axial
201: gravitational wave equation \citep{Chandrasekhar1} is replaced by
202: a standard centrifugal barrier in the tortoise radius plus a
203: constant determined from continuity of the potential at the
204: stellar surface. The CBA potential yields the correct asymptotic
205: behavior of the actual potential at the center of the star and is
206: a good global approximation as well. The wave function of the
207: axial gravitational wave equation in CBA then becomes exactly
208: solvable. It is remarkable that the universal behavior (\ref{TLA})
209: can still be reproduced with CBA using the tortoise radius of
210: TVIIM star as an input. Hence, the observed universality in
211: neutron star axial pulsations is in fact ascribable to (i) the
212: centrifugal potential at the center of the star; (ii) the
213: continuity of the potential; and (iii) the tortoise radius of
214: TVIIM star, which will be given explicitly in Sect.~2 of our
215: paper.
216:
217: To understand quantitatively the universal behavior (\ref{TLA}),
218: we establish here a scaled co-ordinate logarithmic perturbation
219: theory (SCLPT), which is a generalization of the logarithmic
220: perturbation theory (LPT) formulated previously to study QNMs of
221: ``dirty black holes" \citep{dirty1,dirty2}. Applying SCLPT to CBA
222: and using the tortoise radius of TVIIM star, we obtain from first
223: principle the numerical values of $a$, $b$ and $c$ in
224: (\ref{TLA}).
225: The universal behavior
226: (\ref{TLA}) is therefore substantiated analytically.
227:
228: While the CBA star itself clearly displays universality as
229: mentioned above, we can also obtain accurate QNM frequencies of
230: individual realistic neutron stars from those of CBA with SCLPT.
231: As a consequence, we are able to consider the EOS-dependence of
232: the scaled QNM frequency $\tomega$ from the static structure of a
233: star (e.g. mass distribution), thereby paving the way for
234: inferring the EOS of a star from its gravitational wave spectra.
235: Thus, the universality and individuality in QNMs of neutron stars
236: constructed with different EOS's are fully explored with SCLPT
237: developed in this paper.
238:
239:
240: The organization of our paper is as follows. In Sect.~2, after
241: briefly reviewing the scaling behavior of the axial gravitational
242: wave equation, we introduce CBA and show that under such
243: approximation QNMs still manifest the generic behavior summarized
244: in (\ref{TLA}). In Sect.~3 we formulate SCLPT to study how QNM
245: frequencies are affected by changes in the potential and the
246: tortoise radius of the star. In Sect.~4 we apply SCLPT to expand
247: the scaled QNM frequency $\tomega$ as a quadratic function of the
248: compactness of the star and in turn corroborate the universal
249: behavior in (\ref{TLA}). In Sect.~5 SCLPT is applied to
250: investigate the EOS-dependence of the scaled QNM frequency
251: $\tomega$. We then conclude our paper in Sect.~6 with a discussion
252: studying the applicability of CBA and SCLPT to $w_{\rm II}$ and
253: trapped modes \citep[see e.g.][]{Kokkotas_rev,trap1}. Unless
254: otherwise stated, geometrized units in which $G=c=1$ are adopted
255: in the present paper and numerical results are shown for the least
256: damped mode of quadrupole gravitational waves.
257:
258: \section{Generic behavior of QNMs and CBA}
259: The eigenvalue equation for QNMs of axial oscillations of neutron
260: stars is given by a Regge-Wheeler-type equation
261: \citep{Chandrasekhar1}:
262: \begin{equation}
263: \label{KG_eq} \left[\frac{\d^2}{\d
264: r_{*}^2}+\omega^2-V_{}(r_{*})\right]\psi_{}(r_{*}) = 0.
265: \end{equation}
266: Here the tortoise coordinate $r_{*}$ is related to the
267: circumferential radius $r$ by
268: \begin{equation}\label{r*_in}
269: r_{*}=\int_{0}^r \e^{(-\nu+\lambda)/2} \d r,
270: \end{equation}
271: where $e^{\nu(r)}$ and $e^{\lambda(r)}$ are metric coefficients
272: defined by the line element $\d s$ as follows:
273: \begin{equation}
274: \d s^2=-e^{\nu(r)}\d t^2+e^{\lambda(r)} \d r^2+r^2(\d
275: \theta^2+\sin^2\theta \d \varphi^2).
276: \end{equation}
277: In addition, the metric coefficient $e^{-\lambda(r)}$ is given
278: explicitly by:
279: \begin{equation}
280: e^{-\lambda(r)}=1-\frac{2m(r)}{r},
281: \end{equation}
282: with $m(r)$ being the mass-energy inside circumferential radius
283: $r$. The potential $V$ inside the star is given by:
284: \begin{equation}\label{RW_V}
285: V_{}(r_{*})=\frac{\e^\nu}{r^3}[l(l+1)r+4\pi r^3(\rho-P)-6m(r)],
286: \end{equation}
287: where $\rho(r)$ and $P(r)$ are the mass-energy density and the
288: pressure at a point $(r,\theta,\varphi)$, respectively. Outside
289: the star, the tortoise radial coordinate reduces to
290: \begin{equation}\label{r*_out}
291: r_{*}=r+2M \ln\Big(\frac{r}{2M}-1\Big)+C,
292: \end{equation}
293: where $C$ is a constant that can be obtained by matching
294: (\ref{r*_in}) with (\ref{r*_out}) at $r=R$, and the potential $V$
295: is given by the well known Regge-Wheeler potential \citep{RWeq}.
296: \begin{equation}\label{RWP_out}
297: V_{\rm
298: rw}(r_{*})=\left(1-\frac{2M}{r}\right)\left[\frac{l(l+1)}{r^2}-
299: \frac{6M}{r^3}\right].
300: \end{equation}
301:
302:
303: Eq.~(\ref{KG_eq}), referred to as neutron star Regge-Wheeler
304: equation (NSRWE) in the following discussion, and the outgoing
305: wave boundary condition at spatial infinity together determine the
306: QNM frequency of axial $w$-mode oscillations. We show in
307: Fig.~\ref{f1} the QNM frequencies obtained numerically for neutron
308: stars with different EOS's, including models A \citep{modelA} and
309: C \citep{modelC} proposed by Pandharipande, three models (AU, UU
310: and UT) proposed by \citet{AU}, models APR1 and APR2 proposed by
311: \citet{APR}, and model GM${24}$ \citep[][p. 244]{ComStar}. It is
312: clearly shown that both the real and imaginary parts of the scaled
313: frequency $\tomega$ are well approximated by quadratic functions
314: in compactness $\cc$, as indicated by the dotted lines in the
315: corresponding figures. This in turn leads to the universality
316: discussed in Sect.~1.
317:
318: To investigate the physical origin underlying the universality, we
319: have shown that the axial mode wave equation displays scaling
320: behavior \citep{preprint}, and can be rewritten as follows:
321: \begin{equation}
322: \label{KG_sc}
323: \left[\frac{\d^2}{\d\tx^2}+\tomega^2-\tV_{}(\tx)\right]\tpsi_{}(\tx)
324: = 0,
325: \end{equation}
326: where
327: \begin{equation}\label{SP}
328: \tV_{}(\tx)\equiv M^2 V_{}(\tx)=\left\{%
329: \begin{array}{ll}
330: {\e^{\tphi}}\tr^{-3}\left[l(l+1)\tr+4\pi
331: \tr^3(\trho-\tP)-6\tm(\tr)\right], & \hbox{$\tr \leq R/M$;} \\
332: \left(1-{2}{\tr}^{-1}\right)\left[{l(l+1)}{\tr^{-2}}-
333: {6}{\tr^{-3}}\right], & \hbox{$\tr > R/M$;} \\
334: \end{array}%
335: \right.
336: \end{equation}
337: %\end{widetext}
338: and
339: \begin{eqnarray}
340: \tr &=& \frac{r}{M}, \\
341: \tx &=& \frac{r_*}{M}, \\
342: \tm(\tr) &=& \frac{m(r)}{M}, \\
343: \tP(\tr) &=& M^2 P(r), \\
344: \trho(\tr) &=& M^2 \rho(r), \\
345: \tilde{\nu}(\tr) &=& \nu(r).
346: \end{eqnarray}
347: Besides, we also let $\tR \equiv R/M$ and $\tX \equiv R_*/M$,
348: where $R_*=r_{*}(r=R)$ is the tortoise radius of the star.
349:
350: From (\ref{KG_sc}) it is then apparent that the mass-independence
351: of the scaled frequency $\tomega$ follows directly from that of
352: the scaled potential $\tV$. Motivated by this and noting that the
353: following mass distribution:
354: \begin{equation}
355: m_\c(r)= M \left[ \frac{5}{2}\left(\frac{r}{R}\right)^3
356: -\frac{3}{2}\left(\frac{r}{R}\right)^5 \right],
357: \end{equation}
358: conventionally termed as TVIIM in the literature
359: \citep{Tolman:1939jz,Lattimer:2001}, is indeed a good
360: approximation of $m(r)$ for neutron stars with various EOSs (see
361: Fig.~\ref{f2}), we demonstrated that QNMs of TVIIM (solid circles
362: in Fig.~\ref{f1}) display universal behavior observed in realistic
363: stars and interpreted such universality
364: as the consequence of the mass-independence of the
365: scaled mass distribution of TVIIM \citep{preprint}:
366: \begin{equation}
367: \tm_\c(\tr)\equiv \frac {m_\c(r)}{M}=\frac{5}{2}\cc^3\tr^3
368: -\frac{3}{2}\cc^5\tr^5.
369: \end{equation}
370:
371:
372: In addition to approximating the mass distribution in realistic
373: neutron stars, TVIIM is a solvable model whose metric coefficients
374: $e^{\lambda}$ and $e^{\nu}$ can be obtained in closed forms
375: \citep{Tolman:1939jz}:
376: \begin{eqnarray}
377: e^{-\lambda}&=&1-{\cal C} \xi^2(5-3\xi^2)\ , \label{e_lam_eq}\\
378: e^{\nu}&=&(1-5{\cal C}/3)\cos^2\phi \ . \label{e_phi_eq}
379: \end{eqnarray}
380: Here $\xi=r/R$,
381: \begin{eqnarray}
382: \phi&=&({w}_1-{w})/2+\phi_1 \ , \\
383: {w}&=&\log\left[\xi^2-\frac{5}{6}+\sqrt{\frac{e^{-\lambda}}{3\cal{C}}}\right],\\
384: \phi_1&=&\phi(\xi^2=1)=\arctan\sqrt{\frac{\cal{C}}{3(1-2\cal{C})}}\ ,\\
385: w_1&=&w(\xi^2=1)\ ,
386: \end{eqnarray}
387: and the pressure is given by:
388: \begin{eqnarray}
389: P=\frac{1}{4\pi
390: R^2}\left[\sqrt{3\cc
391: e^{-\lambda}}\tan\phi-\frac{\cc}{2}(5-3\xi^2)\right],
392: \end{eqnarray}
393: By using (\ref{e_lam_eq}), (\ref{e_phi_eq}) and the definition of
394: $r_*$, we obtain
395: \begin{eqnarray}\label{intI_eq}
396: r_{*}&=&R\int^{\xi}_0
397: I(\xi',\cc) \d \xi',
398: \end{eqnarray}
399: where
400: \begin{equation}
401: I(\xi,\cc)=\left\{\left[1-{\cal C} \xi^2(5-3\xi^2)\right](1-5{\cal
402: C}/3)\cos^2\phi \right\}^{-1/2}.
403: \end{equation}
404: Hence, the scaled tortoise radius of the star is
405: \begin{eqnarray}\label{STR}
406: \tX = \frac{1}{\cc}\int^{1}_0 I(\xi,\cc) \d \xi.
407: \end{eqnarray}
408: Since TVIIM has a mass profile similar to those of realistic
409: neutron stars, we expect differences in $\tX$ between TVIIM and
410: realistic neutron stars would be small. In fact, our conjecture is
411: verified by Fig.~\ref{f3} where $R_*/M$ is plotted against $M/R$
412: for TVIIM and other realistic stars.
413:
414: As the integral in (\ref{intI_eq}) could not be evaluated
415: analytically, we are not able to express $V(r_*)$ in terms of
416: simple functions of $r_*$ and hence it is not possible to find the
417: analytical form of the wave function in NSRWE even inside the
418: star. However, we could expand $V(r_{*})$ as a power series around
419: the center of the star:
420: \begin{equation}\label{VCEA}
421: \tilde{V}(\tx)=\frac{l(l+1)}{\tx^2}+\tilde{\alpha}+O(\tx^2),
422: \end{equation}
423: where $\tilde{\alpha}$ is a constant dependent on $\cc$ and $l$.
424: It is worthwhile to note that the leading term in the expansion of
425: $\tilde{V}(\tilde{r}_*)$ is nothing but a centrifugal barrier in
426: $\tilde{r}_*$.
427:
428: Despite that the leading two terms of the expansion in
429: (\ref{VCEA}) can nicely approximate the potential at the center of
430: the star, it usually produces a discontinuity across the stellar
431: surface. As QNMs are sensitive to discontinuities, we have to
432: adopt another potential $\tilde{V}_{\c}(\tx)$ given by
433: \begin{equation}\label{CBAV}
434: \tilde{V}_{\c}(\tx)=\frac{l(l+1)}{\tx^2}-\frac{l(l+1)}{\tX^2}+\left(1-\frac{2}{\tR}\right)
435: \left(\frac{l(l+1)}{\tR^2}-\frac{6}{\tR^3}\right), \quad \tx
436: \le \tX ,
437: \end{equation}
438: where $R_*$ is obtained from (\ref{STR}). This potential is
439: similar to the one in (\ref{VCEA}) around $\tilde{r}_*=0$ and, in
440: addition, is continuous across the stellar surface. This scheme is
441: termed the centrifugal barrier approximation (CBA) in our paper.
442: It is worthwhile to note that under CBA the wave function of NSRWE
443: inside the star is simply given by $\tilde{k}\tx
444: j_l(\tilde{k}\tx)$, e.g. for $l=2$:
445: \begin{equation}
446: \tpsi_{}(\tx)=\frac{3\sin(\tilde{k}\tx)}{\tilde{k}^2\tx^2}-\frac{3\cos(\tilde{k}
447: \tx)}{\tilde{k}\tx}-\sin(\tilde{k}\tx),\quad \tx
448: \le \tX.
449: \end{equation}
450: Here
451: \begin{equation}\label{}
452: \tilde{k}=\left[\tomega^2+\frac{l(l+1)}{\tX^2}-\left(1-\frac{2}{\tR}\right)
453: \left(\frac{l(l+1)}{\tR^2}-\frac{6}{\tR^3}\right)\right]^{1/2}
454: \end{equation}
455: is the scaled effective wave number.
456:
457: In addition to its simplicity, as shown in Fig.~\ref{f1}, CBA also
458: yields QNMs (denoted by the solid line in the figure) close to
459: those of TVIIM and realistic neutron stars. Therefore, QNMs of CBA
460: manifest the universal behavior (\ref{TLA}). This remarkable
461: discovery is indeed the crux of the present paper. As the scaled
462: potential in (\ref{CBAV}) is completely characterized by
463: $\tR=1/\cc$ and $\tX$, which is also a function of $\cc$, the
464: cause of the observed universality in neutron star axial
465: pulsations then becomes obvious and understandable. In a nutshell,
466: the universality is attributable to (i) the centrifugal potential
467: at the center of the star; (ii) the continuity of the potential;
468: and (iii) the tortoise radius $\tX$ of TVIIM star.
469: \section{Scaled-Coordinate Logarithmic Perturbation}
470: After locating the physical origin of the universality in QNMs of
471: neutron stars, we further consider the following questions: (i)
472: Can the coefficients $a$, $b$ and $c$ in (\ref{TLA}) be obtained
473: analytically from CBA? (ii) Can QNMs of realistic stars be
474: obtained from those of CBA? We hold positive views on these
475: challenging issues. In this paper, we aim to develop a
476: perturbative study to consider how changes in compactness and EOS
477: could affect the frequencies of the QNMs of a star.
478: In fact, the simplicity and generality of CBA
479: directly leads to a feasible perturbative analysis for QNMs of
480: neutron stars and in turn provides appropriate solutions to these
481: two questions.
482:
483: As discussed above, axial oscillations of neutron stars are
484: described by NSRWE with a potential term dependent on the
485: distribution of energy-mass density and the pressure inside a
486: star. QNMs of these oscillations are the eigen-solution to
487: (\ref{KG_eq}) and, in addition, they are regular at the origin and
488: satisfy the outgoing boundary condition at spatial infinity. It
489: is well known that the wave function of QNMs diverges at spatial
490: infinity and is not amenable to standard perturbation theory
491: \citep[see, e.g.][]{dirty2}. To this end, we will generalize the
492: logarithmic perturbation theory for QNMs, previously formulated by
493: \citet{dirty1,dirty2} to study how the QNMs of a black hole
494: respond to static perturbations such as a static mass shell, to
495: consider QNMs of neutron stars. In this case, the scaled NSRWE
496: (\ref{KG_sc}) that makes use of the scaled coordinate $\tr$ is
497: obviously more amenable to perturbative expansion since outside
498: the star $\tV(\tr)$ depends only $\tr$ and is independent of the
499: stellar mass. Therefore, the method developed here is referred to
500: as the Scaled-Coordinate Logarithmic Perturbation Theory (SCLPT).
501: In the subsequent discussion, we will show that shifts in scaled
502: QNM frequencies can be expressed in terms of integrals with finite
503: domains of integration.
504:
505: \subsection{Perturbative expansion}
506: To formulate a perturbative expansion for the QNM frequency, we
507: introduce a formal perturbation parameter, $\mu$, to measure the
508: departure of the stellar configuration from an unperturbed one,
509: which has a scaled circumferential (tortoise) radius
510: $\tilde{R}_{0}$ ($\tilde{R}_{*0}$), a scaled potential
511: $\tilde{V}_0(\tilde{r}_{*})$, and a QNM with a scaled frequency
512: $\tomega_0$. Analogous to other perturbation theories, we will
513: assume that the QNM wave function is known. This assumption does
514: not pose any problems to our current study because we will use the
515: CBA potential as the unperturbed system and, as mentioned above,
516: the wave function can be obtained analytically. Besides,
517: $\tilde{R}_{*0}$ and $\tilde{V}_0(\tilde{r}_{*})$ can be obtained
518: from (\ref{STR}) and (\ref{CBAV}), respectively. The subject of
519: interest in this paper is to obtain the QNM frequency of a
520: perturbed star whose potential function, as a function of $\tx$,
521: can be expressed as:
522: \begin{eqnarray}\label{Vexp}
523: \tV(\tx)&=&\tilde{V}_0(\tx)+\Delta \tV \nonumber \\
524: &=&\tilde{V}_0(\tx)+\mu\tilde{V}_1(\tx)+\mu^2\tilde{V}_2(\tx)+...~,
525: \,
526: \end{eqnarray}
527: for $\tx \leq \tX(\mu)$. Here $\tilde{R}(\mu)$ and
528: $\tilde{R}_{*}(\mu)$ are the circumferential and tortoise radii of
529: the perturbed star, respectively, which are also functions of
530: $\mu$, with $\tilde{R}(\mu=0)=\tilde{R}_{0}$ and
531: $\tilde{R}_*(\mu=0)=\tilde{R}_{*0}$.
532:
533: Outside the star, it is clear from (\ref{SP}) that $\tV$, as a
534: function of $\tilde{r}$, is independent of $\mu$ and hence
535: $\Delta\tV=0$ for $\tr \geq \tR(\mu)$. It is therefore more
536: convenient to adopt the scaled tortoise coordinate $\tx$ and the
537: scaled circumferential radius $\tr$ to describe wave propagation
538: inside and outside the star, respectively. Each of these radial
539: coordinates has its own advantage. The scaled tortoise coordinate
540: description casts the NSRWE into the standard Klein-Gordon
541: equation form, whereas the use of the scaled circumferential
542: radius simplifies the form of $\tV$ outside the star. In our
543: formalism, we will apply these coordinates in different regions
544: of interest.
545:
546: In the following we seek a power series expansion for the QNM
547: frequency $\tomega(\mu)$ of the perturbed star:
548: \begin{equation}\label{Fexp}
549: \tomega(\mu)=\tomega_0+\mu\tomega_1+\mu^2\tomega_2+...~.
550: \end{equation}
551: We will find a general expression for $\tomega_n$ ($n=1,2,3,...$)
552: and explicitly determine the leading two expansion coefficients,
553: $\tomega_1$ and $\tomega_2$.
554:
555:
556:
557:
558:
559: \subsection{Connection formula}
560: Inside the star we adopt the scaled tortoise coordinate $\tx$ as
561: the independent variable and consider the solution to
562: (\ref{KG_sc}) as a function of three independent variables $\tx$,
563: $\tomega$ and $\mu$, namely:
564: \begin{equation}
565: \tpsi=\tpsi_-(\tx,\tomega,\mu).
566: \end{equation}
567: This solution is specified by the regularity condition at the
568: origin and holds for $\tx < \tX$. Outside the star ($\tr \geq
569: \tR$), the scaled circumferential radius is used instead and we
570: regard the solution to (\ref{KG_sc}) there as a function of $\tr$
571: and $\tomega$. In particular, $\tpsi_+(\tr,\tomega)$ is a solution
572: that satisfies the outgoing wave boundary condition at spatial
573: infinity. Outside the star, NSRWE is identical to the
574: Regge-Wheeler equation of black holes and therefore
575: $\tpsi_+(\tr,\tomega)$ can be obtained from the Leaver's series
576: solution originally developed for determination of QNMs of black
577: holes \citep{Leaver_series,Liu,dirty2}.
578:
579: For each $\mu$, we can evaluate the QNM frequency $\tomega$ by
580: matching the logarithmic derivatives of $\tpsi_+$ and $\tpsi_-$ at
581: $\tr=\tR(\mu)=\tR_0+\mu\tR_1+\mu^{2}\tR_{2}+\cdots$, i.e.
582: \begin{equation}\label{connect}
583: f_-(\tX(\mu),\tomega(\mu),\mu) = f_+(\tR(\mu),\tomega(\mu)),
584: \end{equation}
585: where
586: \begin{eqnarray}
587: f_-(\tx,\tomega,\mu)&=&\frac{1}{\tpsi_-(\tx,\tomega,\mu)}\frac{\partial\tpsi_-}{\partial\tx}, \\
588: f_+(\tr,\tomega)&=&\frac{1}{\tpsi_+(\tr,\tomega)}\frac{\partial\tpsi_+}{\partial\tx},
589: \nonumber \\
590: &=&\frac{(\tr-2)}{\tr\tpsi_+(\tr,\tomega)}\frac{\partial\tpsi_+}{\partial\tr}
591: .
592: \end{eqnarray}
593: It is interesting to note that $f_{\pm}$ are matched at the scaled
594: radius $\tR(\mu)$ (or $\tX(\mu)$) that would vary with $\mu$.
595:
596: \subsection{Expansion of logarithmic derivative}
597: To generate an expansion for the QNM frequency $\tomega(\mu)$, we
598: firstly expand $f_-$ as a power series in the perturbation
599: parameter $\mu$:
600: \begin{eqnarray}
601: f_-(\tx,\tomega(\mu),\mu)&=&f_0(\tx)+\mu f_1(\tx)+\mu^2
602: f_2(\tx)+...~,\label{fminus}
603: \end{eqnarray}
604: where $f_-(\tx,\tomega_0,\mu=0)=f_0(\tx)$ and we have suppressed
605: the dependence on $\tomega_0$ in the expansion coefficients
606: $f_i(\tx)$ ($i=1,2,3,\ldots$). By transforming (\ref{KG_sc}) into
607: the form of the Riccati equation:
608: \begin{equation}\label{riccati_eqn}
609: f_-'+f_-^2+\tomega^2- \tilde{V}(\tx)=0,
610: \end{equation}
611: making use of the expansions (\ref{Vexp}) and (\ref{Fexp}), and
612: comparing equal powers of $\mu$, we show that
613: \begin{equation}\label{riccati_eqn_2}
614: f_n'+2f_0f_n+2\tomega_0\tomega_n=\tU_n, \quad n=1,2,...\,.
615: \end{equation}
616: Here $\tU_1(\tx)=\tV_1(\tx)$ and for $n \ge 2$,
617: \begin{equation}
618: \tU_n(\tx)=\tV_n(\tx)-\sum_{i=1}^{n-1}[f_i(\tx)f_{n-i}(\tx)+\tomega_i\tomega_{n-i}],
619: \end{equation}
620: and we have used the prime to symbolize differentiation with
621: respect to $\tx$. Eq.~(\ref{riccati_eqn_2}) can be solved by
622: introducing the integrating factor
623: $\exp[2\int^{\tx}dyf_0(y)]=\tpsi_0^2(\tx)$, resulting in an
624: integral expression for $f_n(\tilde{R}_{*0})$:
625: \begin{equation}
626: f_n(\tilde{R}_{*0})\tpsi_0^2(\tilde{R}_{*0})=\int_{0}^{\tilde{R}_{*0}}\d\tx[\tU_n(\tx)-2\tomega_0\tomega_n]\tpsi_0^2(\tx).
627: \end{equation}
628:
629: Secondly, as the logarithmic derivatives are to be matched at
630: $\tx=\tX(\mu)=\tR_{*0}+\mu\tR_{*1}+\mu^{2}\tR_{*2}+\cdots$, we
631: expand $f_-(\tX(\mu),\tomega(\mu),\mu)$ and
632: $f_+(\tR(\mu),\tomega(\mu))$ in power series of $\mu$, yielding:
633: \begin{eqnarray}\label{fminusmu}
634: f_-(\tX(\mu),\tomega(\mu),\mu)&=&f_0(\tX(\mu))+\mu
635: f_1(\tX(\mu))+\mu^2f_2(\tX(\mu))+...\nonumber\\
636: &=&\left[f_0(\tilde{R}_{*0})+\mu\frac{\d\f_0}{\d\mu}+\frac{\mu^2}{2!}\frac{\d^2f_0}{\d\mu^2}+...\right]\nonumber\\
637: &&+\mu\left[f_1(\tilde{R}_{*0})+\mu\frac{\d f_1}{\d\mu}+\frac{\mu^2}{2!}\frac{\d^2f_1}{\d\mu^2}+...\right]\nonumber\\
638: &&+\mu^2\left[f_2(\tilde{R}_{*0})+\mu\frac{\d
639: f_2}{\d\mu}+\frac{\mu^2}{2!}\frac{\d^2f_2}{\d\mu^2}+...\right]
640: +...~,
641: \end{eqnarray}
642: and
643: \begin{eqnarray}\label{fplusmu}
644: f_+(\tR(\mu),\tomega(\mu))&=&f_+(\tR(\mu),\tomega(\mu))\nonumber\\
645: &=&f_+(\tR_0,\tomega_0)+\mu\frac{\d
646: f_+}{\d\mu}+\frac{\mu^2}{2!}\frac{\d^2f_+}{\d\mu^2}+...~.
647: \end{eqnarray}
648: Here it is understood that all derivatives with respective to
649: $\mu$ are evaluated at $\mu=0$. Hence, by comparing equal powers
650: of $\mu$ in (\ref{fminusmu}) and (\ref{fplusmu}),
651: $f_n(\tilde{R}_{*0})$ can be expressed in terms of derivatives of
652: $f_{i}$ ($i<n$) and $f_+$ as follows:
653: \begin{eqnarray}
654: f_n(\tilde{R}_{*0})&=&\frac{1}{n!}\frac{\d^nf_+}{\d\mu^n}-\left[\frac{1}{n!}\frac{\d^nf_0}{\d\mu^n}
655: +\frac{1}{(n-1)!}\frac{\d^{n-1}f_1}{\d\mu^{n-1}}+\cdots+\frac{\d f_{n-1}}{\d\mu}\right]\nonumber\\
656: &=&\frac{1}{n!}\frac{\d^nf_+}{\d\mu^n}-\Theta_n,
657: \end{eqnarray}
658: where, by definition, $\Theta_n$ is equal to the sum in the square
659: bracket. Furthermore, by introducing another expression $\Delta_n$
660: defined by:
661: \begin{equation}\label{}
662: \Delta_n=
663: \left[f_+(\tR_0+\mu\tR_1+\cdots+\mu^{n-1}\tR_{n-1},\tomega_0+\mu\tomega_1
664: +\cdots+\mu^{n-1}\tomega_{n-1})-f_+(\tR_0,\tomega_0)\right]_n
665: \end{equation}
666: where $[F(\mu)]_n$ indicates the $n$th-order term of a function
667: $F(\mu)$, it is readily shown that
668: \begin{eqnarray}
669: \frac{1}{n!}\frac{\d^nf_+}{\d\mu^n}&=
670: &\Delta_n+\left(\tomega_n\frac{\partial}{\partial\tomega_0}
671: +\tR_n\frac{\partial}{\partial\tr}\right)f_+(\tomega_0,\tR_0).
672: \end{eqnarray}
673: As a result, we obtain a formal expansion for the scaled frequency
674: $\tomega$, which reads:
675: \begin{equation}\label{freq1}
676: \tomega_n=\frac{\langle\tpsi_0|\tU_n|\tpsi_0\rangle}{2\tomega_0\langle\tpsi_0|\tpsi_0\rangle},
677: \end{equation}
678: with
679: \begin{equation}\label{freq2}
680: \langle\tpsi_0|\tU_n|\tpsi_0\rangle=\int_{0}^{\tilde{R}_{*0}}\d\tx\tU_n(\tx)\tpsi_0^2(\tx)
681: +\left(\Theta_n-\Delta_n-\tR_n\frac{\partial
682: f_+}{\partial\tr}\right)\tpsi_0^2(\tilde{R}_{*0}),
683: \end{equation}
684: and
685: \begin{equation}\label{freq3}
686: \langle\tpsi_0|\tpsi_0\rangle=\int_0^{\tilde{R}_{*0}}\tpsi_0^2(\tx)\d\tx+\frac{\tpsi_0^2(\tilde{R}_{*0})}{2\tomega_0}
687: \frac{\partial f_+}{\partial\tomega_0}.
688: \end{equation}
689:
690: Despite the simplicity of (\ref{freq1}), which is akin to standard
691: perturbation formulas in quantum mechanics with
692: $\langle\tpsi_0|\tpsi_0\rangle$ playing the role of the norm
693: squared of a quantum state, the emergence of terms like $\tU_n$,
694: $\Delta_n$ and $\Theta_n$ reveals the achievement underlying this
695: formula. It is worthy of remark that in (\ref{freq2}) there are
696: three different contributions to the frequency shift $\tomega_n$,
697: namely an integral over the interior of the star and two surface
698: terms originating from $f_+$ and $f_-$, respectively. In the
699: following discussion, we will work out explicit expressions for
700: the first and second order results.
701: \subsection{First and second order frequency shifts}
702: For the case $n=1$, the expression of $\tU_1(\tx)$ is trivial,
703: while $\Theta_n$ and $\Delta_n$ are given by:
704: \begin{eqnarray}
705: \Theta_1&=&\frac{\d f_0}{\d\mu}=\frac{\d f_0}{\d\tx}\frac{\d\tX}{\d\mu} \\
706: \Delta_1&=&0.
707: \end{eqnarray}
708: As a result, it is clear that
709: \begin{equation}\label{U_1}
710: \langle\tpsi_0|\tU_1|\tpsi_0\rangle=\int_{0}^{\tilde{R}_{*0}}\d\tx\tV_1(\tx)\tpsi_0^2(\tx)
711: +\left(\tR_{*1}\frac{d f_0}{d\tx}-\tR_1\frac{\partial
712: f_+}{\partial\tr}\right)\tpsi_0^2(\tilde{R}_{*0}),
713: \end{equation}
714: and hence $\tomega_1$ can be readily obtained from (\ref{freq1})
715: and (\ref{freq3}).
716:
717: %\subsection{Second order perturbation}\label{2ndOrderPerb}
718: For the case $n=2$, it is straightforward to show that
719: \begin{eqnarray}
720: \tU_2(\tx)&=&\tV_2(\tx)-f_1^2(\tx)-\tomega_1^2\;;\\
721: \Theta_2&=&\frac{1}{2}\frac{\d^2f_0}{\d\mu^2}+\frac{\d f_1}{\d\mu}
722: \nonumber\\
723: &=&\frac{\d f_0}{\d\tx}\tR_{*2}+\frac{1}{2}\frac{\d^2f_0}{\d\tx^2}
724: \tR_{*1}^2+\frac{\d f_1}{\d\tx}\tR_{*1}\;;\\
725: \Delta_2&=&\left[f_+(\tR_0+\mu\tR_1,\tomega_0+\mu\tomega_1)-
726: f_+(\tR_0,\tomega_0)\right]_2\nonumber\\
727: &=&\frac{1}{2}\tomega_1^2\frac{\partial^2f_+}{\partial\tomega^2}+
728: \tR_1\tomega_1\frac{\partial
729: f_+}{\partial\tr\partial\tomega}
730: +\frac{1}{2}\tR_1^2\frac{\partial^2f_+}{\partial\tr^2}.
731: \end{eqnarray}
732: Direct substitution of these results into (\ref{freq2}) leads to
733: the second-order term:
734: \begin{eqnarray}\label{U_2}
735: &&\langle\tpsi_0|\tU_2|\tpsi_0\rangle \nonumber
736: \\&=&\int^{\tR_{*0}}_0\left[\tV_2(\tx)-f^2_1(\tx)-\tomega_1^2\right]\tpsi^2_0(\tx)\d\tx
737: \nonumber \\&&+\tpsi^2_0(\tR_{*0})\left[\frac{\d
738: f_0}{\d\tx}\tR_{*2}+\frac{\tR_{*1}^2}{2}
739: \frac{\d^2 f_0}{\d\tx^2}+\frac{\d f_1}{\d\tx}\tR_{*1}\right]_{\tx=\tR_{*0}}\nonumber\\
740: &&-\tpsi^2_0(\tR_{*0})\left[\frac{\tomega_1^2}{2}\frac{\partial^2f_+}
741: {\partial\tomega^2}+\tR_1\tomega_1\frac{\partial
742: f_+}{\partial\tr\partial\tomega}+\frac{\tR_1^2}{2}\frac{\partial^2f_+}{\partial\tr^2}
743: +\tR_2\frac{\partial f_+}{\partial\tr}\right]_{\tr=\tR_0}\,,
744: \end{eqnarray}
745: and the second-order frequency change follows directly from
746: (\ref{freq1}).
747:
748: After succeeding in deriving the first and second order shifts in
749: the eigenfrequency of QNMs of neutron stars, we will apply
750: relevant formulas to answer the questions posed in the beginning
751: of this section.
752:
753: \section{Universality in QNMs}
754: Being a good global approximation to realistic stars, CBA also
755: demonstrates the universal behavior summarized by (\ref{TLA}). In
756: fact, the solid line in Fig.~\ref{f1}, representing the QNM
757: frequencies of CBA, is close to the best quadratic fit to those of
758: realistic stars (the dotted line). Therefore, we expect that the
759: universal behavior displayed by realistic neutron stars can be
760: understood from the QNMs of CBA and Eq.~(\ref{TLA}) can be deduced
761: from CBA as well. Motivated by this conjecture, we evaluate the
762: QNM frequency of CBA with SCLPT.
763:
764: In SCLPT, the QNM frequencies for CBA stars with different
765: compactness can be obtained by considering compactness $\cc$ as
766: the perturbation parameter. In this case, the formal expansion
767: parameter $\mu=\cc-\cc_0$, where $\cc_0$ is the compactness of a
768: reference CBA star whose QNMs are known. Hence, to second order in
769: $\cc-\cc_0$, $\tomega(\cc)$ is approximately given by:
770: \begin{equation}\label{f_exp}
771: \tomega(\mu)=\tomega_0+(\cc-\cc_0)\tomega_1+(\cc-\cc_0)^2\tomega_2
772: \,,
773: \end{equation}
774: where $\tomega_0$ is the QNM frequency of the reference CBA star.
775: The first and second order shifts in QNM frequencies are
776: proportional to $\cc-\cc_0$ and $(\cc-\cc_0)^2$, respectively.
777: Under this approximation, the QNM frequency is expressible in
778: terms of a quadratic function of $\cc$:
779: \begin{equation}\label{omega_quad}
780: \tomega=a\cc^2+b\cc+c\,,
781: \end{equation}
782: with
783: \begin{eqnarray}
784: a&=&\tomega_2;\label{a}\\
785: b&=&\tomega_1-2\cc_0\tomega_2;\label{b} \\
786: c&=&\tomega_0-\cc_0\tomega_1+\cc_0^2\tomega_2.\label{c}
787: \end{eqnarray}
788: It is obvious that Eq.~(\ref{omega_quad}) is in prefect agreement
789: with the universal behavior (\ref{TLA}) discovered numerically by
790: \citet{Andersson1998,Ferrari,preprint}.
791:
792: To gauge the accuracy of the second-order SCLPT mentioned above,
793: we apply it to a reference CBA star with compactness $M/R=0.2$. As
794: shown in Fig.~\ref{f6}, the results obtained from SCLPT
795: (represented by the solid line) are good approximation of the
796: exact QNMs (represented by the stars). This clearly demonstrates
797: the validity of SCLPT. In addition, the perturbative results also
798: faithfully demonstrate the universal behavior displayed by
799: realistic stars (represented by the dotted line), and the
800: numerical values of $a$, $b$ and $c$ obtained from
801: (\ref{a}),(\ref{a}) and (\ref{a}) respectively are in nice
802: agreement with the those obtained from the best quadratic fit to
803: the QNMs of the realistic stars (see Table~1 for reference).
804: Hence, the universality in axial pulsations of neutron stars is
805: fully understood and predicted analytically.
806:
807:
808:
809: The technical details of the perturbation scheme yielding
810: $\tomega_1$, $\tomega_2$ and hence
811: the constants $a$, $b$ and $c$ are as follows. To obtain the
812: first and second order shifts, we have to evaluate the
813: all quantities appearing in (\ref{U_1}) and (\ref{U_2}).
814: Specifically, $\tilde{R}_1$ and $\tilde{R}_2$ are given by:
815: \begin{eqnarray}\label{}
816: \tR_1&=&\left(\frac{\d\tR}{\d\cc}\right)_{\cc=\cc_0}=-\frac{1}{\cc_0^2},
817: \\
818: \tR_2&=&\frac{1}{2}\left(\frac{\d^2\tR}{\d\cc^2}\right)_{\cc=\cc_0}=\frac{1}{\cc_0^3}.
819: \end{eqnarray}
820: Analogously, $\tR_{*1}$ and $\tR_{*2}$ can be obtained:
821: \begin{eqnarray}\label{}
822: \tR_{*1}&=&\left(\frac{\d\tX}{\d\cc}\right)_{\cc=\cc_0}\nonumber
823: \\
824: &=&\frac{1}{\cc_0}\int^{1}_0 \left[\frac{\partial
825: I(\xi,\cc_0)}{\partial
826: \cc_0 }-\frac{I(\xi,\cc_0)}{\cc_0}\right] \d\xi , \\
827: \tR_{*2}&=&
828: \frac{1}{2}\left(\frac{\d^2\tX}{\d\cc^2}\right)_{\cc=\cc_0}
829: \nonumber
830: \\ &=& \frac{1}{2\cc_0}\int^{1}_0 \left[\frac{\partial^2
831: I(\xi,\cc_0)}{\partial \cc_0^2 }-\frac{2}{\cc_0}\frac{\partial
832: I(\xi,\cc_0)}{\partial \cc_0
833: }+\frac{2I(\xi,\cc_0)}{\cc_0^2}\right] \d\xi .
834: \end{eqnarray}
835: On the other hand, as $\tilde{V}_c$ contains explicit dependence
836: on $\tX$ and $\tR$, $\tV_1$ and $\tV_2$ can be found:
837: \begin{eqnarray}\label{}
838: \tV_1(\tx)&=&\left(\frac{\partial\tV_c}{\partial\cc}\right)_{\cc=\cc_0}\nonumber
839: \\
840: &=& \frac{2l(l+1)\tR_{*1}}{\tR_{*0}^3}+
841: 2l(l+1)\cc_0-3(l^2+l+3)\cc_0^2+12\cc_0^3 \nonumber
842: \\ \tV_2(\tx)&=&\left(\frac{\partial^2\tV_c}{\partial\cc^2}\right)_{\cc=\cc_0}\nonumber
843: ,\\
844: &=& -\frac{6l(l+1)\tR_{*1}^2}{\tR_{*0}^4}+\frac{4
845: l(l+1)\tR_{*2}}{\tR_{*0}^3}+ 2l(l+1)-6(l^2+l+3)\cc_0+36\cc_0^2,
846: \end{eqnarray}
847: which are constants independent of $\tx$.
848:
849: It is easy to see that $\tV_n(\tx)$ ($n=1,2,3,\ldots$) are indeed
850: all $\tx$-independent under CBA. This fact greatly simplifies the
851: perturbation calculation in CBA and renders the integrals involved
852: in the evaluation of $\tomega_1$ and $\tomega_2$ exactly solvable.
853: For example, noting that $\tpsi_0(\tx)=\tilde{k}\tx
854: j_l(\tilde{k}\tx)$, we have
855: \begin{equation}\label{}
856: \int_{0}^{\tilde{R}_{*0}}\d\tx\tV_1(\tx)\tpsi_0^2(\tx) =
857: \frac{\tV_1
858: \tR_{*0}}{2}\left\{x^2\left[j_l^2(x)+j_{l-1}^2(x)\right]-
859: (2l+1)xj_l(x)j_{l-1}(x)\right\}_{x=\tilde{k} \tR_{*0}}
860: \end{equation}
861: and
862: \begin{eqnarray}
863: f_1(\tilde{r}_{*})&=&\frac{1}{\tpsi_0^2(\tilde{r}_{*})}
864: \int_{0}^{\tilde{r}_{*}}\d\tx'[\tV_1(\tx')-2\tomega_0\tomega_1]
865: \tpsi_0^2(\tx') \nonumber \\
866: &=&\frac{(\tV_1-2\tomega_0\tomega_1)\tR_{*0}}{2}\left[1+\frac{j_{l-1}^2(x)}{j_l^2(x)}-
867: \frac{(2l+1)j_{l-1}(x)}{xj_l(x)}\right]_{x=\tilde{k} \tR_{*0}}\,.
868: \end{eqnarray}
869:
870:
871:
872: Besides, we have also apply SCLPT to evaluate QNMs of other
873: realistic stars. As an example, we show here the result for a star
874: constructed with APR2 EOS. As demonstrated in Fig.~\ref{f7}, the
875: second order result of SCLPT is again a good approximation to the
876: exact numerical results. This provides independent corroboration
877: to the validity of SCLPT.
878:
879: \section{Individuality of QNMs}
880: After establishing the universality in axial pulsations of neutron
881: stars, we turn our attention to the individuality of such QNMs. As
882: clearly shown in Figs.~\ref{f1} and \ref{f4}, QNM frequencies of
883: individual realistic neutron stars in general deviate slightly
884: from the universal curve (\ref{TLA}). In the present paper we aim
885: to evaluate QNMs of realistic stars from those of CBA star with
886: SCLPT developed here. For a realistic star whose compactness $\cc$
887: and potential $\tV(\tx)$ are known, we compare it with a CBA star
888: with the same compactness $\cc$. Therefore, we have
889: $\tV_0=\tV_\c$, $\tV=\tV_\c+\tV_1$, and $\tX=\tR_{*0}+\tR_{*1}$,
890: where $\tR_{*0}$ is the tortoise radius of the CBA star. However,
891: in this case the scaled circumferential radii of the two stars in
892: consideration are the same. In other words, $\tR_n=0$ for
893: $n=1,2,\ldots$. Besides, it is convenient to take the formal
894: expansion parameter $\mu=1$ in the current situation. As both
895: $\tV_n=0$ and $\tR_{*n}=0$ for $n=2,3\ldots$, the matrix elements
896: $\langle\tpsi_0|\tU_1|\tpsi_0\rangle$ and
897: $\langle\tpsi_0|\tU_2|\tpsi_0\rangle$ can be simplified and
898: respectively take the form:
899: \begin{equation}
900: \langle\tpsi_0|\tU_1|\tpsi_0\rangle=
901: \int_{0}^{\tilde{R}_{*0}}\d\tx\tV_1(\tx)\tpsi_0^2(\tx)
902: +\tR_{*1}\frac{\d f_0}{\d\tx}\tpsi_0^2(\tilde{R}_{*0}),
903: \end{equation}
904: \begin{eqnarray}
905: &&\langle\tpsi_0|\tU_2|\tpsi_0\rangle\nonumber\\
906: &=&-\int^{\tR_{*0}}_0\left[f^2_1(\tx)+\tomega_1^2\right]\tpsi^2_0(\tx)\d\tx
907: \nonumber
908: \\&&+\tpsi^2_0(\tR_{*0})\left[\left(\frac{\tR_{*1}^2}{2} \frac{\d^2 f_0}{\d\tx^2}+
909: \tR_{*1}\frac{\d f_1}{\d\tx}\right)_{\tx=\tR_{*0}}
910: -\frac{\tomega_1^2}{2}\left(\frac{\partial^2f_+}
911: {\partial\tomega^2}\right)_{\tr=\tR_0}\right]\,.
912: \end{eqnarray}
913: The first and second order shifts in QNM frequency then follow
914: directly from (\ref{freq1}), (\ref{freq2}) and (\ref{freq3}).
915:
916: Figure \ref{f4} shows QNMs obtained from the perturbation scheme
917: outlined above for stars constructed from various EOSs with a
918: common compactness $\cc=0.20$. The unfilled, dark, and grey
919: symbols respectively
920: indicate the exact, the first and the second order results,
921: while the star is the unperturbed frequency of CBA. It is clearly
922: shown that the second order perturbation results can nicely
923: approximate the exact ones. We have also applied SCLPT
924: to stable neutron stars with a larger (or smaller) compactness and found that
925: the second-order results are indeed reliable. In fact, the first
926: order formula readily suffices to yield accurate prediction of the
927: shift in ${\rm Re}\,\tomega$.
928:
929: On the other hand, we have to point out the limitations
930: of SCLPT. As in other cases, perturbation theory is likely to
931: break down in the presence of instabilities. Therefore, we expect
932: that the accuracy of perturbative results obtained from SCLPT gets
933: worse when the star becomes unstable against perturbation. We
934: verify this point by applying SCLPT to neutron stars constructed
935: with GM24 EOS \citep[][p.~244]{ComStar}, which are known to be
936: unstable if the compactness is greater than $0.21$. As clearly
937: demonstrated in Fig.~\ref{f5}, where QNMs of CBA and GM24 stars
938: with $\cc=0.17,0.19,0.20$ are shown, the deviation of the
939: second-order result from the exact value increases as the
940: compactness approaches the border of stability. However, as far as
941: ${\rm Re}\,\tomega$ is concerned, both the first and second order
942: results are satisfactory.
943: \section{Conclusion and discussion}
944: The main achievement in the present paper is the discovery that
945: the universality in the QNM frequency of axial pulsations of
946: neutron stars in fact originates from the CBA, under which the
947: potential term in the scaled NSRWE is essentially a centrifugal
948: barrier. There are two parameters in the scaled potential, namely
949: the scaled circumferential radius $\tR$ and tortoise radius $\tX$
950: of the star. While the former is just the inverse of the
951: compactness $\cc$, the latter can be determined from that of
952: TVIIM. These two parameters completely determine the QNMs of CBA
953: and in turn give rise to the observed universality.
954:
955: In order to consider how the the physical characteristics of a
956: neutron star could affect the frequencies of its QNMs, we have
957: also developed a systematic perturbative scheme SCLPT, which is
958: designed to cope with the divergence in QNM wavefunction at
959: spatial infinity.
960: Direct application of SCLPT to
961: CBA then successfully predicts the universality in the QNM
962: frequency of realistic neutron stars. The advantage of application
963: of the scaled coordinates ($\tr$ and $\tx$) in our study becomes
964: manifest in this regard because the potential $\tV$ outside a star
965: is independent of its detailed internal structure save its
966: compactness, leading to the observed universality
967: \citep{Andersson1998,Ferrari,preprint}.
968:
969: On the other hand, SCLPT can also predict small deviations in QNM
970: frequencies from those of CBA for individual realistic neutron
971: stars. This opens possibilities for researchers to link the
972: QNM frequencies observed from a distant star with its
973: internal structure and EOS as well. We are currently working on
974: the inverse problem of SCLPT, namely inferring the internal
975: structure of a pulsating neutron star from its gravitational wave
976: spectrum. Relevant progress in this direction will be reported
977: elsewhere in due course.
978:
979: So far we have used the least-damped $w$-mode to illustrate the
980: principle and accuracy of our method. However, the results
981: mentioned in the present paper are general and hold for modes of
982: higher orders (i.e. modes with frequencies higher than that of the
983: least-damped one). More interestingly, we note that the $w_{\rm
984: II}$-mode \citep[see e.g.][and references therein]{Kokkotas_rev}
985: also displays similar universality. As shown in Fig.~\ref{f8},
986: where the real and imaginary parts of the scaled QNM frequencies
987: of the $w_{\rm II}$-modes of various realistic stellar models,
988: TVIIM and CBA are plotted against the compactness, the
989: universality summarized in (\ref{TLA}) is clearly exhibited. It is
990: remarkable that the corresponding values obtained from application
991: of second-order SCLPT to CBA
992: (the solid line) successfully capture the essence of
993: such universal behavior. To further examine the validity of CBA
994: and SCLPT in this case, we show in Fig.~\ref{f9} the scaled QNM
995: frequency of a $w_{\rm II}$-mode of a CBA star (the star symbol)
996: and other realistic stars (the unfilled symbols) with a common
997: compactness
998: $\cc= 0.2$, the first and second order
999: results obtained from applying SCLPT to the CBA star
1000: (respectively denoted by the corresponding dark and grey symbols). It
1001: is obvious that SCLPT can indeed reproduce accurate eigenfrequencies for
1002: the $w_{\rm II}$-mode.
1003:
1004: Despite that we have demonstrated here the validity of CBA and
1005: SCLPT for ordinary $w$-modes and $w_{\rm II}$ modes; we have to
1006: caution that the trapped mode is an exception to our method. As is
1007: well known, trapped modes usually exist in ultra-compact stars
1008: with compactness greater than 1/3 and have small decaying rates
1009: \citep{trap1}. The physical origin of such modes is the
1010: development of a local minimum in the potential $V(r_*)$ inside
1011: the star when $\cc > 1/3$. Hence, gravitational waves are trapped
1012: in the potential minimum and acquire much longer lifetime. The
1013: analysis proposed in the present paper relies on the validity of
1014: the CBA model whose potential term $V(r_*)$ is obviously a
1015: monotonic function inside the star. In fact, the CBA potential
1016: deviates significantly from those of TVIIM (or other ultra-compact
1017: stars). As shown in Fig.~\ref{f10}, trapped modes of CBA are no
1018: longer good approximation to those of TVIIM if $\cc > 1/3$ and
1019: therefore it is not possible to obtain the trapped modes of TVIIM
1020: (or other ultra-compact stars) from application of SCLPT to CBA.
1021: We conclude our paper with this remark.
1022:
1023:
1024: \begin{acknowledgments}
1025: We thank J Wu for discussions. We also express our gratitude to
1026: an anonymous referee for drawing
1027: our attention to the $w_{\rm II}$ and the trapped modes. Our work
1028: is supported in part by the Hong Kong Research Grants Council
1029: (grant No: CUHK4282/00P and 401905) and a direct grant (Project
1030: ID: 2060260) from the Chinese University of Hong Kong.
1031: \end{acknowledgments}
1032: %\appendix
1033:
1034:
1035: \newpage
1036: %\bibliography{1st}
1037: \newcommand{\noopsort}[1]{} \newcommand{\printfirst}[2]{#1}
1038: \newcommand{\singleletter}[1]{#1} \newcommand{\switchargs}[2]{#2#1}
1039: \begin{thebibliography}{37}
1040: \expandafter\ifx\csname
1041: natexlab\endcsname\relax\def\natexlab#1{#1}\fi
1042:
1043: \bibitem[{Akmal {et~al.}(1998)Akmal, Pandharipande, \& Ravenhall}]{APR}
1044: Akmal, A., Pandharipande, V.~R., \& Ravenhall, D.~G. 1998, Phys.
1045: Rev. C, 58,
1046: 1804
1047:
1048: \bibitem[{Alcock {et~al.}(1986)Alcock, Farhi, \& Olinto}]{AFO}
1049: Alcock, C., Farhi, C.~E., \& Olinto, A. 1986, ApJ, 310, 261
1050:
1051: \bibitem[{Andersson \& Kokkotas(1996)}]{Andersson_1996}
1052: Andersson, N., \& Kokkotas, K.~D. 1996, Phys. Rev. Lett., 77, 20
1053:
1054: \bibitem[{Andersson \& Kokkotas(1998)}]{Andersson1998}
1055: ---. 1998, MNRAS, 299, 1059
1056:
1057: \bibitem[{Belczynski {et~al.}(2001)Belczynski, Kalogera, \&
1058: Bulik}]{Belczynski:2001uc}
1059: Belczynski, K., Kalogera, V., \& Bulik, T. 2001, ApJ, 572, 407
1060:
1061: \bibitem[{Benhar {et~al.}(1999)Benhar, Berti, \& Ferrari}]{Ferrari}
1062: Benhar, O., Berti, E., \& Ferrari, V. 1999, MNRAS, 310, 797
1063:
1064: \bibitem[{Chandrasekhar \& Ferrari(1991a)}]{Chandrasekhar1}
1065: Chandrasekhar, S., \& Ferrari, V. 1991a, Proc. R. Soc. A, 432, 247
1066:
1067: \bibitem[{Chandrasekhar \& Ferrari(1991b)}]{trap1}
1068: ---. 1991b, Proc. R. Soc. A, 434, 449
1069:
1070: \bibitem[{Cheng {et~al.}(1998)Cheng, Dai, Wei, \& Lu}]{Cheng}
1071: Cheng, K.~S., Dai, Z.~G., Wei, D.~M., \& Lu, T. 1998, Science,
1072: 280, 407
1073:
1074: \bibitem[{Ching {et~al.}(1996)Ching, Leung, Suen, \& Young}]{Ching}
1075: Ching, E.~S.~C., Leung, P.~T., Suen, W.~M., \& Young, K. 1996,
1076: Phys. Rev. D,
1077: 54, 3778
1078:
1079: \bibitem[{Chodos {et~al.}(1974)Chodos, Jaffe, Johnson, Thorne, \&
1080: Weisskopf}]{MITBM}
1081: Chodos, A., Jaffe, R.~L., Johnson, K., Thorne, C.~B., \&
1082: Weisskopf, V.~F. 1974,
1083: Phys. Rev. D, 9, 3471
1084:
1085: \bibitem[{Fryer {et~al.}(2002)Fryer, Holz, \& Hughes}]{Fryer:2001zw}
1086: Fryer, C.~L., Holz, D.~E., \& Hughes, S.~A. 2002, ApJ, 565, 430
1087:
1088: \bibitem[{Glendenning(1997)}]{ComStar}
1089: Glendenning, N.~K. 1997, Compact Stars - Nuclear Physics, Particle
1090: Physics, and
1091: General Relativity (Springer, NY)
1092:
1093: \bibitem[{Hughes(2003)}]{Hughes_03}
1094: Hughes, S. 2003, Ann. Phys., 303, 142
1095:
1096: \bibitem[{Kokkotas {et~al.}(2001)Kokkotas, Apostolatos, \&
1097: Andersson}]{Kokkotas_2001}
1098: Kokkotas, K.~D., Apostolatos, T.~A., \& Andersson, N. 2001, MNRAS,
1099: 320, 307
1100:
1101: \bibitem[{Kokkotas \& Schmidt(1999)}]{Kokkotas_rev}
1102: Kokkotas, K.~D., \& Schmidt, B.~G. 1999, Living Rev. Rel., 2, 2
1103:
1104: \bibitem[{Kokkotas \& Schutz(1986)}]{ToyModel}
1105: Kokkotas, K.~D., \& Schutz, B.~F. 1986, Gen. Relativ. Gravitation,
1106: 18, 913
1107:
1108: \bibitem[{Lattimer \& Prakash(2001)}]{Lattimer:2001}
1109: Lattimer, J.~M., \& Prakash, M. 2001, ApJ, 550, 426
1110:
1111: \bibitem[{Leaver(1986{\natexlab{a}})}]{Leaver_1986}
1112: Leaver, E.~W. 1986{\natexlab{a}}, Phys. Rev. D, 34, 384
1113:
1114: \bibitem[{Leaver(1986{\natexlab{b}})}]{Leaver_series}
1115: ---. 1986{\natexlab{b}}, J. Math. Phys., 27, 1238
1116:
1117: \bibitem[{Leung {et~al.}(1997)Leung, Liu, Suen, Tam, \& Young}]{dirty1}
1118: Leung, P.~T., Liu, Y.~T., Suen, W.~M., Tam, C.~Y., \& Young, K.
1119: 1997, Phys.
1120: Rev. Lett., 78, 2894
1121:
1122:
1123: \bibitem[{Leung {et~al.}(1999)Leung, Liu, Suen, Tam, \& Young}]{dirty2}
1124: ---. 1999, Phys. Rev. D, 59, 044034
1125:
1126: \bibitem[{Lindblom(1992)}]{Lindblom_invert}
1127: Lindblom, L. 1992, ApJ, 398, 56
1128:
1129: \bibitem[{Lindblom {et~al.}(1998)Lindblom, Owen, \& Morsink}]{Lindblom:1998wf}
1130: Lindblom, L., Owen, B.~J., \& Morsink, S.~M. 1998, Phys. Rev.
1131: Lett., 80, 4843
1132:
1133: \bibitem[{Liu(1997)}]{Liu}
1134: Liu, Y.~T. 1997, MPhil thesis, The Chinese University of Hong Kong
1135:
1136: \bibitem[{Mason(2004)}]{Grishchuk}
1137: Mason, J.~W., ed. 2004, Astrophysics Update, Vol.~I
1138: (Springer-Praxis), 281--310
1139:
1140: \bibitem[{Nollert(1999)}]{Nollert_rev}
1141: Nollert, H.-P. 1999, Class. Quantum Grav., 16, R159
1142:
1143: \bibitem[{Pandharipande(1971{\natexlab{a}})}]{modelA}
1144: Pandharipande, V. 1971{\natexlab{a}}, Nucl. Phys A, 174, 641
1145:
1146: \bibitem[{Pandharipande(1971{\natexlab{b}})}]{modelC}
1147: ---. 1971{\natexlab{b}}, Nucl. Phys A, 178, 123
1148:
1149: \bibitem[{Prakash {et~al.}(1990)Prakash, Baron, \& Prakash}]{PBP}
1150: Prakash, M., Baron, E., \& Prakash, M. 1990, Phys. Lett. B, 243,
1151: 175
1152:
1153: \bibitem[{Press(1971)}]{Press_1971}
1154: Press, W.~H. 1971, ApJ, 170, L105
1155:
1156: \bibitem[{Regge \& Wheeler(1957)}]{RWeq}
1157: Regge, T., \& Wheeler, J.~A. 1957, Phys. Rev., 108, 1063
1158:
1159: \bibitem[{Thorne \& Campolattaro(1967)}]{Thorne}
1160: Thorne, K.~S., \& Campolattaro, A. 1967, ApJ, 149, 591
1161:
1162: \bibitem[{Tolman(1939)}]{Tolman:1939jz}
1163: Tolman, R.~C. 1939, Phys. Rev., 55, 364
1164:
1165: \bibitem[{Tsui \& Leung(2004)}]{preprint}
1166: Tsui, L.~K., \& Leung, P.~T. 2004, MNRAS, 357, 1029
1167:
1168: \bibitem[{Wiringa {et~al.}(1988)Wiringa, Fiks, \& Fabrocini}]{AU}
1169: Wiringa, R.~B., Fiks, V., \& Fabrocini, A. 1988, Phys. Rev. C, 38,
1170: 1010
1171:
1172: \bibitem[{Witten(1984)}]{Witten}
1173: Witten, E. 1984, Phys. Rev. D, 30, 272
1174:
1175: \end{thebibliography}
1176:
1177: \newpage
1178:
1179: \begin{figure}%Figure 6.7
1180: \includegraphics[angle=270,width=8.5cm]{f1.eps}
1181: \caption{The real and imaginary parts of the scaled QNM
1182: frequencies of various realistic stellar models (including APR1,
1183: APR2, AU, GM24, Models A and B, UT and UU), TVIIM and CBA (solid
1184: line) are plotted against the compactness. The dotted line is the
1185: best quadratic fit to those of the realistic stars} \label{f1}
1186: \end{figure}
1187:
1188:
1189:
1190: \begin{figure}%Figure 6.2
1191: \includegraphics[angle=270,width=8.5cm]{f2.eps}
1192: \caption{The mass distributions of various realistic neutron stars
1193: and TVIIM with a common compactness $\cc=0.2$ are plotted against
1194: $r/R$.} \label{f2}
1195: \end{figure}
1196: \begin{figure}%Figure 6.6
1197: \includegraphics[angle=270,width=8.5cm]{f3.eps}
1198: \caption{The figure shows the relationship between $R_*/M$ and the
1199: compactness for TVIIM (solid line) and other realistic neutron
1200: stars.}\label{f3}
1201: \end{figure}
1202:
1203:
1204:
1205:
1206:
1207:
1208:
1209: \begin{figure}%Figure 6.7
1210: %\vspace{10cm}
1211: \includegraphics[angle=270,width=8.5cm]{f4.eps}
1212: \caption{The scaled QNM frequency of the CBA star is plotted
1213: against the compactness $\cc$. The stars and the solid line
1214: represent the exact values and the ones obtained from applying the
1215: second-order SCLPT to a CBA star with $\cc=0.2$, respectively. For
1216: purpose of comparison, we also show the universal curve (the
1217: dotted line), i.e. the best quadratic fit to QNM frequencies of
1218: the realistic stars considered in this paper.} \label{f6}
1219: \end{figure}
1220: \begin{figure}%Figure 6.7
1221: %\vspace{10cm}
1222: \includegraphics[angle=270,width=8.5cm]{f5.eps}
1223: \caption{The scaled QNM frequency of the APR2 star is plotted
1224: against the compactness $\cc$. The stars and the solid line are
1225: the exact values and the ones obtained by applying the
1226: second-order SCLPT to an APR2 star with $\cc=0.2$, respectively.}
1227: \label{f7}
1228: \end{figure}
1229: \begin{figure}%Figure 6.8
1230: \includegraphics[angle=270,width=8.5cm]{f6.eps}
1231: %\vspace{10 cm}
1232: \caption{The star symbol shows the scaled QNM frequency of a CBA star with compactness
1233: $\cc= 0.2$.
1234: By applying SCLPT to the CBA star, the first and second order
1235: results, respectively denoted by dark and grey symbols,
1236: for QNM frequencies of realistic
1237: stars with the same compactness are obtained. For comparison,
1238: exact numerical QNM frequencies of realistic
1239: stars are indicated by the corresponding unfilled symbols.}
1240: \label{f4}
1241: \end{figure}
1242:
1243: \begin{figure}%Figure 6.7
1244: %\vspace{10cm}
1245: \includegraphics[angle=270,width=8.5cm]{f7.eps}
1246: \caption{The figure shows the scaled QNM frequencies of GM24 stars
1247: with $\cc=0.17,0.19,0.20$. The star symbol indicates the QNM
1248: frequencies of the corresponding CBA stars.
1249: The first and second order
1250: results obtained from SCLPT are denoted by dark and grey symbols,
1251: respectively.
1252: For comparison
1253: exact numerical QNM frequencies of GM24
1254: stars are indicated by the corresponding unfilled symbols.} \label{f5}
1255: \end{figure}
1256: \begin{figure}%Figure 6.7
1257: \includegraphics[angle=270,width=8.5cm]{f8.eps}
1258: \caption{The real and imaginary parts of the scaled QNM
1259: frequencies of a $w_{\rm II}$-mode of various realistic stellar
1260: models, TVIIM and CBA are plotted against the compactness. The
1261: solid line shows the corresponding values obtained from the
1262: second-order SCLPT.} \label{f8}
1263: \end{figure}
1264: \begin{figure}%Figure 6.8
1265: \includegraphics[angle=270,width=8.5cm]{f9.eps}
1266: %\vspace{10 cm}
1267: \caption{The star symbol shows the scaled QNM frequency of a $w_{\rm II}$-mode
1268: of a CBA star with compactness
1269: $\cc= 0.2$.
1270: By applying SCLPT to the CBA star, the first and second order
1271: results, respectively denoted by dark and grey symbols,
1272: for QNM frequencies of realistic
1273: stars with the same compactness are obtained. For comparison,
1274: exact numerical QNM frequencies of realistic
1275: stars are indicated by the corresponding unfilled symbols.}
1276: \label{f9}
1277: \end{figure}
1278: \begin{figure}%Figure 6.8
1279: \includegraphics[angle=270,width=8.5cm]{f10.eps}
1280: %\vspace{10 cm}
1281: \caption{The real and imaginary parts of the scaled QNM
1282: frequencies of a trapped-mode of TVIIM and CBA are plotted against
1283: the compactness.} \label{f10}
1284: \end{figure}
1285: \clearpage %Just because of unusual number of tables stacked at end
1286: % Produces the bibliography via BibTeX.
1287: \begin{table}
1288: \centering
1289: \begin{tabular}{|c|c|c|c|}
1290: \hline
1291: % after \\: \hline or \cline{col1-col2} \cline{col3-col4} ...
1292: Data & $a$ & $b$ & $c$\\
1293: \hline
1294: % CEA & $-8.46-i15.42$ & $6.48+i4.81$ & $-0.20+i0.12$ \\
1295: Best fit to realistic stars & $-3.9-\i5.6$ & $2.8+\i1.6$ & $-0.03+\i0.125$ \\
1296: Perturbative result of CBA& $-4.63-\i4.86$ & $3.15+\i1.30$ & $-0.063+\i0.155$ \\
1297: \hline
1298: \end{tabular}
1299: \caption{The complex coefficients $a$, $b$ and $c$ in (\ref{TLA}) are shown.
1300: In the first row the values are obtained from the best quadratic fit to
1301: the QNMs of the realistic stars considered in Fig.~\ref{f1},
1302: while in the second row those obtained from the second
1303: order result of SCLPT for
1304: a reference CBA star with compactness $M/R=0.2$ are presented.
1305: }
1306: \label{coef_cea_cba}
1307: \end{table}
1308: \clearpage
1309: \end{document}
1310: %
1311: % ****** End of file apssamp.tex ******
1312: