1: %\documentclass[11pt, preprint, onecolumn]{aastex}
2: %\documentclass{aastex}
3: %\usepackage{emulateapj5,epsfig}
4: \documentclass[useAMS, usenatbib, onecolumn]{mn2e}
5: \usepackage{epsfig}
6: \voffset -1.8cm
7: \def\be{\begin{equation}}
8: \def\ee{\end{equation}}
9: \def\ba{\begin{eqnarray}}
10: \def\ea{\end{eqnarray}}
11: \def\go{\mathrel{\raise.3ex\hbox{$>$}\mkern-14mu
12: \lower0.6ex\hbox{$\sim$}}}
13: \def\lo{\mathrel{\raise.3ex\hbox{$<$}\mkern-14mu
14: \lower0.6ex\hbox{$\sim$}}}
15:
16: \def\tomega{{\tilde{\omega}}}
17: \def\cR{{\cal R}}
18:
19: \begin{document}
20: \title[Corotational Damping of C-Modes]
21: {Corotational Damping of Diskoseismic C-modes in Black Hole Accretion Discs}
22: \author[D. Tsang and D. Lai]
23: {David Tsang$^{1,2}$\thanks{Email:
24: dtsang@astro.cornell.edu; dong@astro.cornell.edu}
25: and Dong Lai$^{1}$\footnotemark[1] \\
26: $^1$Department of Astronomy, Cornell University, Ithaca, NY 14853, USA \\
27: $^2$Department of Physics, Cornell University, Ithaca, NY 14853, USA \\}
28:
29: \label{firstpage}
30: \maketitle
31:
32: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
33: \begin{abstract}
34: Diskoseismic c-modes in accretion discs have been
35: invoked to explain low-frequency variabilities observed
36: in black-hole X-ray binaries. These modes are
37: trapped in the inner-most region of the disc and
38: have frequencies much lower than the rotation frequency
39: at the disc inner radius. We show that because
40: the trapped waves can tunnel through the evanescent barrier to the corotational
41: wave zone, the c-modes are damped due to wave
42: absorption at the corotation resonance. We calculate the corotational
43: damping rates of various c-modes using the WKB approximation.
44: The damping rate varies widely depending on the mode frequency, the black hole
45: spin parameter and the disc sound speed, and is generally much less than $10\%$ of
46: the mode frequency. A sufficiently strong excitation mechanism is needed
47: to overcome this corotational damping and make the mode observable.
48: \end{abstract}
49:
50: \begin{keywords}
51: accretion, accretion discs -- hydrodynamics -- waves --
52: -- black hole physics -- X-rays: binaries
53: \end{keywords}
54: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
55: \section{Introduction}
56:
57: Diskoseismic oscillations of accretion discs around relativistic
58: objects have been studied for over two decades (e.g., Kato \& Fukue
59: 1980; Okazaki et al. 1987; Nowak \& Wagoner 1991, see Wagoner 1999,
60: Kato 2001 for reviews), and have been used as models for the time
61: variability and quasi-periodic oscillations (QPOs) in X-ray emissions
62: from black-hole X-ray binaries.
63:
64: The c-modes (or so-called corrugation waves) were first proposed to
65: explain low-frequency variabilities as their oscillation frequencies
66: are lower than the associated g-modes and p-modes (see section
67: 3). Kato (1983) and Okazaki \& Kato (1985) showed the existence of
68: one-armed ($m=1$), low-frequency modes in nearly Keplerian (Newtonian) discs,
69: while later work (Kato 1989; Silbergleit et al.~2001) demonstrated the
70: presence of low-frequency c-modes in relativistic accretion discs,
71: particularly the one-armed corrugation waves with a single node
72: ($n=1$; see section 3 below) in the
73: vertical direction, which oscillate at (approximately) the
74: Lense-Thirring precession frequency evaluated at the outer edge of the
75: trapping region.
76:
77: In a recent paper (Lai \& Tsang 2008), we studied the global
78: corotational instability of non-axisymmetric p-modes ($n=0$) in black hole
79: accretion discs. The mode is trapped inside the corotation resonance
80: radius $r_c$ (where the wave pattern rotation speed $\omega/m$ equals
81: the disc rotation rate $\Omega$) and carries a negative energy. We
82: showed that when the mode frequency $\omega$ is sufficiently high,
83: positive wave energy is absorbed at the corotation resonance, leading
84: to the growth of mode amplitude. The mode growth is further enhanced
85: by wave transmission beyond the corotation barrier.
86: Non-axisymmetric g-modes, on the other hand, may contain
87: a corotation resonance in the wave zone.
88: Kato (2003) and Li, Goodman \& Narayan (2003) showed that
89: such g-modes are heavily damped as the wave propagate through the corotation
90: resonance (see also Zhang \& Lai 2006).
91:
92: Diskoseismic c-modes are trapped in the inner most regions of black
93: hole accretion discs. Although their primary wave zones are
94: separated from the corotation resonance, the wave can tunnel through
95: the evanescent barrier and propagate again around the corotation.
96: In this paper we calculate the analytic damping rate of c-modes due to
97: wave absorption at the corotation resonance. In section 2 we briefly
98: review the basic properties of perturbations in a thin isothermal disc, and
99: present the linear perturbation equations. In section 3 we
100: discuss the propagation regions associated with various
101: diskoseismic modes, while in section 4 we demonstrate the effect of
102: the corotation resonance on wave propagation. In section 5 the effect of the
103: corotation on the c-mode is studied and the
104: c-mode damping rates are calculated for different disc parameters.
105: Section 6 contains our conclusion.
106:
107: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
108: \section{Basic Setup and Equations}
109:
110: Consider a thin isothermal disc with the unperturbed velocity
111: ${\bf u}_o = (0, r\Omega, 0)$ in the cylindrical coordinates.
112: The vertical density profile is given by (for small $z \ll r$)
113: \be
114: \rho_o(r,z) = \frac{\Sigma(r)}{\sqrt{2\pi} H} \exp\left(-z^2/2H^2\right).
115: \ee
116: Here $\Sigma(r)$ is the (vertically integrated) surface density and
117: $H=c_s/\Omega_\perp$ is the vertical scale height, where
118: $\Omega_\perp$ is the vertical oscillation frequency and $c_s$ is the
119: isothermal sound.
120:
121: Perturbing the mass and momentum conservation equations gives
122: \ba
123: \frac{\partial}{\partial t}\delta \rho + \nabla\cdot(\rho_o \delta
124: {\bf u} + {\bf u}_o \delta \rho) &=& 0~,\\
125: \frac{\partial}{\partial t}\delta {\bf u} + ({\bf u}_o \cdot \nabla)
126: \delta {\bf u} + (\delta {\bf u} \cdot \nabla) {\bf u}_o &=& -\nabla
127: \delta h~,
128: \ea
129: with the enthalpy perturbation $\delta h \equiv \delta P/\rho = c_s^2
130: \delta \rho/\rho$, where we assume that the perturbations are also
131: isothermal.
132: Assuming perturbations of the form $\delta P, \delta {\bf u}, \delta
133: \rho \propto \exp(im\phi - i\omega t)$, we have
134: \ba
135: -i\tomega \frac{\rho_o}{c_s^2}\delta h +
136: \frac{1}{r}\frac{\partial}{\partial r}(r\rho_o \delta u_r) +
137: \frac{im}{r}\rho_o \delta u_\phi + \frac{\partial}{\partial z}(\rho
138: \delta u_z) &=& 0 \label{fullmomentumpert}\\
139: -i\tomega \delta u_r - 2\Omega \delta u_\phi &=& -\frac{\partial}{\partial r} \delta h
140: \label{eq:dur}\\
141: -i\tomega \delta u_\phi + \frac{\kappa^2}{2\Omega} \delta u_r &=& -\frac{im}{r}\delta h
142: \label{eq:duphi}\\
143: -i\tomega \delta u_z &=& -\frac{\partial}{\partial z}\delta h\label{zpert}
144: \ea
145: where $\tomega = \omega - m\Omega$.
146: %Using (\ref{zpert}) the last term of (\ref{fullmomentumpert}) can be expressed as
147: %\ba \frac{\partial}{\partial z}(\rho \delta u_z) &=&
148: %\frac{\partial}{\partial
149: %z}\left(\frac{\rho_o}{i\tomega}\frac{\partial}{\partial z}\delta
150: %h\right)\\ &=& \frac{\rho_o}{i\tomega} \frac{\partial^2}{\partial^2
151: %z}\delta h - \frac{z}{H}\frac{\rho_o}{i\tomega}
152: %\frac{\partial}{\partial z}\delta h \ea
153: Following Okazaki et al.~(1987), we assume a $z$-dependence of the
154: perturbations such that $\delta h, \delta u_r, \delta u_\phi \propto
155: {\rm H}_n(z/H)$ where ${\rm H}_n(z/H)$ is the Hermite polynomial
156: of order $n$.
157: Then equation (\ref{fullmomentumpert}) reduces to
158: \ba
159: -i\tomega \frac{\rho_o}{c_s^2}\delta h + \frac{1}{r}\frac{\partial}{\partial r}(r\rho_o \delta u_r) + \frac{im}{r}\rho_o \delta u_\phi - \frac{n\rho_o}{i\tomega} \delta h= 0
160: \label{eq:dh}
161: \ea
162: Neglecting terms proportional to $dH/dr \sim O(1/r)$, and eliminating
163: the velocity perturbations $\delta u_r$ and $\delta u_\phi$ from equations
164: (\ref{eq:dur}), (\ref{eq:duphi}) and (\ref{eq:dh}), we obtain
165: (see eq.~[29] in Zhang \& Lai [2006])
166: \be
167: \frac{d^2}{dr^2}\delta h - \left(\frac{d}{dr}\ln \frac{D}{r\Sigma}
168: \right)\frac{d}{dr}\delta h +
169: \left[\frac{2m\Omega}{r\tomega}\frac{d}{dr}\ln \frac{D}{\Omega\Sigma}
170: - \frac{m^2}{r^2} - \frac{D(\tomega^2 - n\Omega_\perp^2)}{c_s^2
171: \tomega^2}\right] \delta h = 0~, \label{mastereq}
172: \ee
173: where $D = \kappa^2 - \tomega^2$. This is our basic working equation.
174:
175: The approach above is Newtonian. More rigorous fully relativistic
176: derivations of the dispersion relation have been given by Ipser (1994,
177: 1996), Perez et al.~(1997), and Silbergleit et al.~(2001). Some
178: aspects of the general relativistic effects can be incorporated
179: into our analysis by using the Paczynski-Witta psuedo-Newtonian
180: potential, which gives $\kappa<\Omega=\Omega_\perp$.
181: For our purposes of estimating the c-mode damping rates,
182: it suffices to employ equation (\ref{mastereq})
183: but with the relevant fully general relativistic frequencies
184: (e.g., Aliev \& Gal'tsov 1981; Okazaki et al.~1987)
185: \ba
186: \Omega &=& \frac{1}{r^{3/2} + a},\label{Omega}\\
187: \Omega_\perp &=& \Omega\left( 1 - \frac{4a}{r^{3/2}}+ \frac{3a^2}{r^2}\right)^{1/2},\\
188: \kappa &=& \Omega\left(1 - \frac{6}{r} + \frac{8a}{r^{3/2}} -
189: \frac{3a^2}{r^2}\right)^{1/2},\label{kappa},
190: \ea
191: where the frequencies are in units of $c^3/GM$, $r$ in units of $GM/c^2$, and
192: $a$ is the spin parameter of the black hole.
193:
194: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
195: \section{Propagation Diagram and C-Modes}
196:
197: \begin{figure}
198: \centering
199: \includegraphics[width=13cm]{modes.eps}
200: \vskip -2.5cm
201: \caption{The propagation diagram for non-axisymmetric g-modes and
202: c-modes. Note that the curves $m\Omega-\kappa$ and
203: $m\Omega + \kappa$ meet at $r=r_{\rm ISCO}$ since
204: $\kappa(r_{\rm ISCO}) = 0$ due to general relativistic effects.
205: The g-modes are trapped between the inner Lindblad resonance
206: (where $\omega=m\Omega-\kappa$) and the outer Lindblad resonance
207: (where $\omega=m\Omega+\kappa$), or around the peak of the $m\Omega+\kappa$
208: curve.}
209: \label{fig1}
210: \end{figure}
211:
212: There are three possible critical resonant points in the disc:
213: the Lindblad resonances (LRs) where $D = 0$, the vertical resonances (VR) where
214: $\tomega^2 = n\Omega_\perp^2$, and the corotation resonance (CR) where
215: $\tomega = 0$. Far from these critical points, the WKB dispersion relation
216: [for $\delta h\propto \exp(i\int k\, dr)$] takes the form (Okazaki et al.~1987)
217: \be
218: c_s^2k^2 = \frac{(\kappa^2 - \tomega^2)(n\Omega_\perp^2 -
219: \tomega^2)}{\tomega^2}.
220: \label{dispersion} \ee
221: %(Okazaki et al. 1987, Nowak \&
222: %Wagoner 1991, 1992, Ipser 1994, and Kato 2001 for review).
223: The modes with $n=0$ have no vertical structure, and are referred to
224: as p-modes -- their stability properties are studied in Lai \& Tsang
225: (2008). We focus on the modes with $n\ge 1$ and $m\ge 1$ in this
226: paper. Wave propagation is allowed in regions where
227: $\tomega^2<\kappa^2<n\Omega_\perp^2$ or
228: $\tomega^2>n\Omega_\perp^2>\kappa^2$ (note that $\kappa<\Omega_\perp$
229: in GR). The former defines the g-mode propagation zone: the mode is trapped
230: in the region where $m\Omega-\kappa<\omega<m\Omega+\kappa$; the latter leads to
231: c-modes, for which the wave zone is specified by
232: $\omega < m\Omega - \sqrt{n}\Omega_\perp$ (see Fig.~1). Clearly, the c-modes
233: exist only when $m\Omega - \sqrt{n}\Omega_\perp>0$ and wave reflection
234: occurs at $r=r_{\rm in}=r_{\rm ISCO}$,
235: the radius of the Inner-most Stable Circular Orbit.
236: For Newtonian discs, since $\Omega_\perp = \kappa = \Omega$ the
237: c-modes can only exist if $m > \sqrt{n}$; for relativistic
238: discs, however, $\Omega>\Omega_\perp$ (for spinning black holes), we can obtain
239: modes for $m = \sqrt{n}$ that may have very low frequencies.
240: The ``fundamental'' c-mode, with $m=n=1$ is of particular interest (Kato 1990),
241: since it corresponds to the Lense-Thirring precession of the inner, tilted disc.
242: Note that for extremely low mode frequencies, the terms of order $1/r$
243: previously ignored may become important, and care must be taken to
244: obtain the real eigenfrequencies for trapped modes. Here we ignore
245: these complications and refer to Silbergleit et al.~(2001) for a more
246: thorough relativistic analysis.
247:
248: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
249: \section{Wave Absorption at the Corotation Resonance}
250:
251: The Lindblad resonances ($D=0$) are apparent singularities of the
252: master equation (\ref{mastereq}) [this can be seen easily by writing
253: (\ref{mastereq}) as two coupled first-order differential equations],
254: and no wave absorption occurs at the Lindblad
255: resonances (e.g., Goldreich \& Tremaine 1979; Li et al.~2003;
256: Zhang \& Lai 2006). The vertical resonances (where $\tomega^2=n\Omega_\perp^2$)
257: act purely as turning points, and no wave absorption occurs there either.
258: However, the corotation resonance must be treated more carefully
259: (Kato 2003; Li et al.~2003; Zhang \& Lai 2006).
260:
261: Here we follow the analysis of Zhang \& Lai (2006).
262: Near the corotation ($r=r_c$, where $\omega=m\Omega$),
263: equation (\ref{mastereq}) can be written as
264: \be
265: \frac{d^2}{dr^2}\delta h -\frac{D(\tomega^2 - n\Omega_\perp^2)}{c_s^2 \tomega^2} \delta h \simeq 0,\label{nearcr}
266: \ee
267: since for a thin disc the sound speed $c_s \ll r\Omega$ and
268: the last term in (\ref{mastereq}) dominates the other terms.
269: Defining $x\equiv (r-r_c)/r_c$ and expanding
270: (\ref{nearcr}) around $x=0$, we have
271: \be
272: \frac{d^2}{dx^2}\delta h + \frac{C}{(x+i\epsilon)^2}\delta h = 0,\label{nearcrC}
273: \ee
274: where
275: \be
276: C \equiv \frac{n}{m^2}\left(\frac{\kappa\Omega_\perp}
277: {c_s d\Omega/dr}\right)^2_{r_c} \gg 1.
278: \ee
279: In equation (\ref{nearcrC}), we have inserted a small imaginary part $i\epsilon$
280: (with $\epsilon>0$) in $1/x^2$ because we consider the response of the
281: disc to a slowly growing perturbation.
282:
283: The two independent solutions to equation (\ref{nearcrC}) are
284: \be
285: \delta h_\pm=z^{1/2}z^{\pm i\nu}
286: =z^{1/2}e^{\pm i\nu \ln z}
287: \label{c10}\ee
288: where $\nu=\sqrt{C-{1\over4}}\gg 1$ and $z=x+i\epsilon$.
289: The solution $z^{1/2}z^{i\nu}$ has a local wavenumber
290: $k=d(\nu\ln z)/dr=\nu/(r_c x)$, with the group velocity
291: $v_g=d\omega/dk=-\tomega/k=-q r_c x^2\omega/\nu<0$
292: (where we have assumed $\Omega\propto r^{-q}$, with $q>0$),
293: thus it represents waves propagating toward small $r$. Similarly,
294: the solution $z^{1/2}z^{-i\nu}$ has $v_g>0$ and represents waves propagating
295: toward large $r$.
296:
297: As shown in Zhang \& Lai (2006), waves with $n\ge 1$ can propagate into
298: the corotation region and be absorbed there. Consider an incident
299: wave propagating from the $x<0$ (or $r<r_c$) region toward $x=0$, with
300: the amplitude (up to a constant factor)
301: \be
302: \delta h(x<0)= z^{1/2}e^{-i\nu\ln z}=i e^{\pi\nu}(-x)^{1/2}e^{-i\nu\ln(-x)},
303: \qquad ({\rm incident~wave}).
304: \label{eq:incident2}\ee
305: The transmitted wave is simply
306: \be
307: \delta h(x>0)= x^{1/2}e^{-i\nu\ln x},\qquad
308: ({\rm transmitted~wave}),
309: \ee
310: and there is no reflection.
311: Thus the amplitude of the transmitted wave is decreased by a factor
312: of $e^{-\pi \nu}$ (Zhang \& Lai 2006).
313: Similarly, a wave incident from the $r>r_c$ region toward $r_c$
314: encounters the same attenuation. Since $\nu \gg 1$ for thin discs,
315: We readily conclude that all waves incident upon the corotation will be
316: absorbed (Kato 2003; Li et al.~2003; see Zhang \& Lai 2006 and Lai \& Zhang 2008
317: for applications of this result to the problem of wave excitation by
318: external forces).
319:
320: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
321: \section{Corotational Damping of C-modes}
322:
323: \begin{figure}
324: \centering
325: \includegraphics[width=13cm]{modepot.eps}
326: \caption{The wave propagation diagram (upper panel) and effective potential
327: (lower panel) for c-modes. The modes are trapped in region I (between $r_{\rm in}=
328: r_{\rm ISCO}$ and the inner vertical resonance $r_{\rm IVR}$), but can tunnel
329: through an evanescent barrier (region II, between
330: the IVR and the inner Lindblad resonance $r_{\rm ILR}$)
331: and propagate into the corotation ($r_c$), where they are absorbed.}
332: \label{fig2}
333: \end{figure}
334:
335: The result of section 4 shows that waves propagating through the
336: corotation are heavily damped. Since g-modes trapped between the inner
337: and outer Lindblad resonances must cross the corotation, they are
338: damped very quickly by corotation absorption, as shown by Kato (2003) and
339: Li et al.~(2003). Only higher frequency g-modes which
340: are trapped around the peak of $m\Omega+\kappa$ (see Fig.~1)
341: can avoid such corotational damping; these modes have frequencies
342: $\omega\simeq m\Omega(r_{\rm ISCO})$.
343:
344: In this section we calculate the damping rate of the c-mode. The damping
345: mechanism is illustrated in Fig.~2, where we also plot the effective potential:
346: \be
347: V_{\rm eff}(r)= \frac{D(\tomega^2 - n\Omega_\perp^2)}{c_s^2 \tomega^2} + \frac{m^2}{r^2} +\frac{2m\Omega}{r\tomega}\left(\frac{d}{dr}\ln\frac{\Omega \Sigma}{D}\right).
348: \ee
349: Based on the result of section 4, we will adopt the approximation that
350: waves transmitted through the barrier between the inner vertical
351: resonance (IVR) and the inner Lindblad resonance (ILR) are absorbed at
352: the corotation. We will also assume that the inner disc boundary is
353: completely reflective (see section 6).
354:
355:
356: \subsection{Reflection Coefficient}
357:
358: We first calculate the reflection coefficient ${\cal R}$ when a wave
359: in region I (see Fig. 2) propagates outward and is reflected back at the IVR.
360:
361: From the dispersion relation (\ref{dispersion}), the group velocity
362: of the wave is given by
363: \be
364: v_g\equiv {d\omega\over
365: dk}={kc_s^2\over \tomega \Bigl[
366: 1-(\kappa/\tomega)^2(n\Omega_\perp^2/\tomega^2)\Bigr]}.
367: \label{gv}\ee
368: The relative sign of $v_g$ and the phase velocity $v_p=\omega/k$
369: is important. In region III ($r_{\rm ILR} < r < r_{c}$), $v_g$ and $v_p$
370: have the same sign, thus waves propagating outwards
371: correspond to $k > 0$. In region I ($r < r_{\rm IVR}$),
372: $v_g$ and $v_p$ have opposite signs, so that
373: waves propagating outwards have $k < 0$ and waves propagating inwards
374: have $k > 0$.
375:
376: As shown in the previous section we can, to good approximation, assume
377: that waves propagating into the corotation are completely damped.
378: Thus, only an outward-propagating wave exists in region III (see Fig.~2), with
379: the wave amplitude (up to a constant prefactor) given by
380: \be
381: \delta h =A\exp\left( i \int_{r_{\rm ILR}}^r k\,dr
382: + \frac{\pi}{4}\right),
383: \label{eq:outside}\ee
384: where $k>0$ is given by equation (\ref{dispersion}) and
385: and $A \equiv \sqrt{D/r k \Sigma}$ is the WKB amplitude.
386: The connection formulae for the ILR (Tsang \& Lai 2008) are
387: \be
388: \delta h_1 \sim
389: \Biggl\{\begin{array}{ll}
390: \frac{1}{2}A\, \exp\left(-\int_r^{r_{\rm ILR}}\!|k| \,dr \right)
391: & \qquad \textrm{for } r \ll r_{\rm ILR}\\
392: A\, \cos\left(\int_{r_{\rm ILR}}^r\! k\, dr + \pi/4\right)
393: &\qquad \textrm{for } r\gg r_{\rm ILR}
394: \end{array}
395: \label{eq:con1}\ee
396: \be
397: \delta h_2 \sim
398: \Biggl\{\begin{array}{ll}
399: A \exp\left(\int_r^{r_{\rm ILR}}\! |k|\, dr \right)
400: & \qquad \textrm{for } r \ll r_{\rm ILR}\\
401: A\, \sin\left(\int_{r_{\rm ILR}}^r\! k\, dr + \pi/4\right)
402: &\qquad \textrm{for } r\gg r_{\rm ILR}~.
403: \end{array}
404: \label{eq:con2}\ee
405: These then give for the evanescent zone (region II in Fig.~2):
406: \ba
407: \delta h &\simeq& \frac{A}{2}\exp\left(-\int_r^{r_{\rm ILR}}|k|\,dr
408: \right) + iA\exp\left(\int_r^{r_{\rm ILR}} |k| dr\right) \nonumber\\
409: &=& \frac{A}{2}
410: \exp(-\Theta_{\rm{II}})
411: \exp\left(\int_{r_{\rm IVR}}^{r}|k| dr \right) + iA
412: \exp(+\Theta_{\rm{II}})\exp\left(-\int_{r_{\rm IVR}}^r |k| dr\right)
413: \ea
414: where
415: \be
416: \Theta_{\rm{II}} = \int_{r_{\rm IVR}}^{r_{\rm ILR}} |k|\, dr.
417: %= \int_{r_{\rm IVR}}^{r_{\rm ILR}} \sqrt{(\kappa^2 -
418: %\tomega^2)(n\Omega_\perp^2 - \tomega^2)\over \tomega^2c_s^2}\, dr.
419: \ee
420:
421: %%%%%%%%%%%%%%
422: \begin{figure}
423: \centering
424: \includegraphics[width=17cm]{R2m1.eps}
425: \vskip -2cm
426: \caption{The reflection coefficient, ${\cal R}$, of waves incident
427: upon the inner vertical resonance (IVR), as a function of the (real)
428: wave frequency $\omega_r$ (in units of $m\Omega_{\rm in}$, where $\Omega_{\rm in}$
429: is the disc rotation rate at $r=r_{\rm in}$), for
430: various black hole spin parameters ($a$) and sound speeds $c_s =
431: \beta r \Omega_{\rm in}$. The upper panels are for $m=2,~n=1$ and
432: the lower panels for $m=n=1$.}
433: \label{fig3}
434: \end{figure}
435: %%%%%%%%%%%%%%%%
436:
437: The connection formulae at the IVR can be similarly derived; they are
438: \footnote{Note that around the IVR, the differential equation is matched
439: by the Airy functions, rather than by the Airy function derivatives
440: as in the case of the ILR.}
441: \be
442: \delta h_1 \sim
443: \Biggl\{\begin{array}{ll}
444: \frac{1}{2}A\, \exp\left(-\int^r_{r_{\rm IVR}}\!|k| \,dr \right)
445: & \qquad \textrm{for } r \gg r_{\rm IVR}\\
446: A\, \sin\left(\int^{r_{\rm IVR}}_r\! k\, dr + \pi/4\right)
447: &\qquad \textrm{for } r\ll r_{\rm IVR}
448: \end{array}
449: \label{eq:conn1}\ee
450: \be
451: \delta h_2 \sim
452: \Biggl\{\begin{array}{ll}
453: A \exp\left(\int^r_{r_{\rm IVR}}\! |k|\, dr \right)
454: & \qquad \textrm{for } r \gg r_{\rm IVR}\\
455: A\, \cos\left(\int^{r_{\rm IVR}}_r\! k\, dr + \pi/4\right)
456: &\qquad \textrm{for } r\ll r_{\rm IVR}
457: \end{array}
458: \label{eq:conn2}\ee
459: (Abramowitz \& Stegun 1964).
460: Thus, we find that for $r < r_{\rm IVR}$ (region I in Fig.~2),
461: \be
462: \delta h \simeq \frac{A}{2}e^{-\Theta_{\rm{II}}}
463: \cos\left(\int_r^{r_{\rm IVR}} k\,dr +\frac{\pi}{4}\right) + i2A
464: e^{\Theta_{\rm{II}}}\sin\left(\int_r^{r_{\rm IVR}} k\,dr
465: +\frac{\pi}{4}\right)~.
466: \ee
467: Expressing this in terms of traveling waves and defining
468: $y = \int_{r_{\rm IVR}}^r k dr - \pi/4$, we have
469: \be
470: \delta h \simeq iA \left[ e^{-iy} \left(e^{+\Theta_{\rm{II}}}
471: + \frac{1}{4}e^{-\Theta_{\rm{II}}}\right) - e^{+iy} \left(
472: e^{\Theta_{\rm{II}}} - \frac{1}{4}e^{-\Theta_{\rm{II}}}\right)\right],
473: \label{eq:inside}\ee
474: where the first term ($\propto e^{-iy}$) corresponds to the incident (outgoing) wave
475: and the second term ($\propto e^{iy}$) the inward going wave reflected from the IVR.
476: Thus the reflection coefficient is
477: \be
478: {\cal R} = -\frac{e^{\Theta_{\rm{II}}} - \frac{1}{4}e^{-\Theta_{\rm{II}}}}{e^{\Theta_{\rm{II}}} +
479: \frac{1}{4}e^{-\Theta_{\rm{II}}}},
480: \ee
481: and the transmission coefficient (through region II) is
482: \be
483: {\cal T} = \frac{i}{e^{\Theta_{\rm{II}}} +
484: \frac{1}{4}e^{-\Theta_{\rm{II}}}}.
485: \ee
486: We can clearly see that $|{\cal R}|^2 < 1$.
487:
488: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
489: \subsection{Trapped C-modes and their Damping Rates}
490:
491: Assuming a reflective boundary exists at $r = r_{\rm in} < r_{\rm IVR}$,
492: we can develop trapped c-modes in the inner disc, between $r_{\rm in}$ and
493: the IVR. To illustrate this we consider a simple boundary condition at $r=r_{\rm in}$:
494: \be
495: \delta h(r_{{\rm in}}) = 0. \label{boundary}
496: \ee
497: From section 5.1, the wave in region I can be written as
498: \be
499: \delta h=A\exp(-iy)+{\cal R}A\exp(iy),\qquad {\rm with}~~
500: y=\int_{r_{\rm IVR}}^{r} \!k\,dr-\pi/4.
501: %\delta h = A\exp\left(-i\int_{r_{\rm IVR}}^{r} \!k\,
502: %dr-\pi/4\right)+ {\cal R}A
503: %\exp\left(i\int_{r_{\rm IVR}}^{r}\!k\,dr-\pi/4\right),
504: \label{eq:rinside} \ee
505: where $k=k_r+ik_i$ is complex.
506: Applying the boundary condition (\ref{boundary}) to
507: equation (\ref{eq:rinside}) yields the eigenvalue condition:
508: \be
509: \exp(2i\Theta)=-i|{\cal R}|,\quad {\rm with}~~
510: \Theta=\int_{r_{\rm in}}^{r_{\rm IVR}}\!k\,dr=\Theta_r+i\Theta_i,
511: \label{eq:eigcond}
512: \ee
513: where $\Theta_r$ and $\Theta_i$ are real.
514: The real eigen frequency $\omega_r$ is given by
515: \be
516: \Theta_r=\int_{r_{\rm in}}^{r_{\rm IVR}}\!k_r\,dr
517: =\int_{r_{\rm in}}^{r_{\rm IVR}}
518: {\sqrt{(\kappa^2 - \tomega_r^2)(n\Omega_\perp^2 - \tomega_r^2)}\over
519: c_s |\tomega_r|}\,dr
520: =\mu \pi +{3\pi\over 4},
521: \label{omegarcond}\ee
522: where $\mu=0,1,2,\cdots$ is an integer and
523: $\tomega_r = \omega_r - m\Omega < 0$ in the trapping region
524: between $r_{\rm in}$ and $r_{\rm IVR}$.
525: The imaginary part of the frequency $\omega_i$ is determined by
526: $|{\cal R}|=\exp(-2\Theta_i)$, or
527: \be
528: \tanh\Theta_i=-\left({|{\cal R}|-1\over |{\cal R}|+1}\right).
529: \ee
530: Note that $\Theta_i=\int_{r_{\rm in}}^{r_{\rm IVR}}k_i\,dr\ll 1$ and
531: \be
532: k_i=\omega_i {dk\over d\omega}\Bigr|_{\omega_r}=
533: \frac{\omega_i\tomega_r}{k_r c_s^2}\left(1-
534: \frac{n\kappa^2\Omega_\perp^2}{\tomega_r^4} \right).
535: \ee
536: We then obtain
537: \be
538: \omega_i \simeq -\frac{1}{4}\exp(-2\Theta_{\rm II})
539: \left[\int_{r_{\rm in}}^{r_{\rm IVR}}\!\frac{|\tomega_r|^2}
540: {\sqrt{(\kappa^2 -\tomega^2)(n\Omega_\perp^2 - \tomega^2)}}
541: \left(1 -\frac{n\kappa^2\Omega_\perp^2}{\tomega_r^4}\right)
542: {dr\over c_s}\right]^{-1}.
543: \label{omegai}
544: \ee
545: Note that in the trapping region (between $r_{\rm in}$ and $r_{\rm IVR}$),
546: $n\kappa^2\Omega_\perp^2/\tomega^4 < 1$. Thus the mode is always damped
547: ($\omega_i<0$).
548:
549: %%%%%%%%%%%%%%%%%%%%%
550: \begin{figure}
551: \centering
552: \begin{tabular}{ccc}
553: %\epsfig{file=wf1.eps,width=0.5\linewidth,clip=} &
554: %\epsfig{file=wf2.eps,width=0.5\linewidth,clip=}
555: \epsfig{file=omegavam2n1.eps,width=0.5\linewidth,clip=} &
556: \epsfig{file=omegavam1n1.eps,width=0.5\linewidth,clip=}
557: \end{tabular}
558: \caption{The real and imaginary frequencies for the primary trapped
559: c-mode with the $\delta h(r_{\rm in}) = 0$ boundary condition and
560: $r_{\rm in} = r_{\rm ISCO}$ for various sound speeds versus the black
561: hole spin parameter. Note that there is no mode for the one-armed
562: corrugation wave (the right panel) for sound speed $c_s = 0.1
563: r\Omega_{\rm in}$.}
564: \label{fig4}
565: \end{figure}
566:
567: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
568: \subsection{Numerical Results}
569:
570: Figure 3 depicts the reflection coefficients for waves impinging
571: upon the inner vertical resonance (IVR) from $r < r_{\rm IVR}$
572: as a function of the wave frequency. We consider both the
573: $m=2,~n=1$ and the $m=1,~n=1$ modes. The real frequency ranges from
574: $0$ to $(m\Omega-\sqrt{n}\Omega_\perp)|_{r_{\rm in}}$. %Obviously,
575: %waves with higher frequencies are ``protected'' by a wider evanescent zone
576: %(see Figs.~1-2) and have $|{\cal R}|^2$ closer to unity.
577: Waves with higher frequencies are ``protected'' by a larger potential barrier in the evanescent zone (larger $\Theta_{\rm II}$) and have $|{\cal R}|^2$ closer to unity.
578:
579: Figures 4-5 show the real and imaginary c-mode frequencies, computed
580: using the WKB expressions derived in section 5.2. Waves with
581: higher frequencies have smaller damping rate $|\omega_i|$, consistent with
582: the reflectivity results shown in Fig.~3. The damping rate is also smaller for
583: cooler (smaller $\beta$) discs. %For the $m=1, n=1$ case the range for possible c-mode frequencies is bounded from above by the Lense-Thirring precession frequency at the inner boundary. Thus as $a\rightarrow 0$, $\omega_r \rightarrow 0$. The damping rate increases as $\omega_r$ decreases, as the width of the region between $r_{\rm ILR}$ and $r_{\rm IVR}$ falls with the perturbation frequency (see Fig.~2). However, increasing the spin parameter $a$ also tends to reduce this width, and thus increases the damping rate (see Fig.~5). The combination of these two competing effects produces a minimum in the damping rate as a function of $a$.
584: For the $m > n$ case (e.g. Fig. 4, left panel), the dimensionless damping rate $|\omega_i|/m\Omega_{\rm in}$ increases with increasing $a$ (while $\omega_r/m\Omega_{\rm in}$ remans approximately constant), because the width of the region between $r_{\rm IVR}$ and $r_{\rm ILR}$ is smaller for larger $a$. For the $m = n = 1$ case (Fig. 4, right panel) the range of possible c-mode frequencies is bounded from above by the Lense-Thirring precession frequency at the inner boundary. Thus $\omega_r \rightarrow 0$ when $a \rightarrow 0$. This leads to larger $|\omega_i/\omega_r|$ for small $a$ since the potential barrier in the evanescent zone decreases with decreasing $\omega_r/m\Omega_{\rm in}$.
585: Note that although we use the relativistic
586: frequencies (\ref{Omega})-(\ref{kappa}) our calculations not fully
587: relativistic, and the real frequencies calculated shown in Fig.~4, particularly for
588: the $m=n=1$ mode with large black hole spin parameter $a$, are
589: correct only in orders of magnitude [Fully relativistic calculation of the real
590: frequencies of these trapped modes was done
591: by Silbergleit et al.~(2001).] Also note that adopting different inner disc boundary
592: conditions would lead to different real mode eigenfrequencies than presented in
593: Fig.~4.
594:
595: Finally, we note that equation (\ref{omegai}) is valid only when there
596: is no loss of wave energy at the inner disc boundary $r_{\rm in}$.
597: In the presence of the rapid radial inflow at the ISCO, we would expect
598: additional mode damping due to the leakage of waves into the plunging region
599: of the disc (see Lai \& Tsang 2008 for a study of such leakage for p-modes).
600: Alternatively, a sufficiently strong excitation mechanism is needed to
601: overcome the corotational damping and make the c-modes grow.
602:
603: %%%%%%%%%%%%%%%%%%
604: \begin{figure}
605: \centering
606: \includegraphics[width=17cm]{oivor.eps}
607: \vskip -2cm
608: \caption{The damping rate $|\omega_i|$ for various sound speeds and
609: black hole spin parameters for c-modes with purely reflective boundary
610: conditions at $r_{in}$ as a function of the real mode frequencies $\omega_r$.
611: The upper panels show the results for the $m=2,n=1$ modes and the lower panels
612: show the $m=n=1$ modes. The sound speeds are taken to be constant,
613: $c_s = \beta r\Omega_{\rm in}$.
614: }\label{fig5}
615: \end{figure}
616:
617: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
618: \section{Conclusion}
619:
620: In this paper, we have shown that diskoseismic c-modes suffer
621: corotational damping due to wave absorption at the corotation
622: resonance. These modes are trapped between the inner disc edge and the inner
623: vertical resonance (where $\omega-m\Omega=-\sqrt{n}\Omega_\perp$,
624: with $m,n\ge 1$), but can tunnel through the evanescent zone and leak
625: out to the corotation zone where wave absorption occurs.
626: The mode damping rates are generally much smaller than the mode
627: frequencies, and depend sensitively on the disc sound speed and
628: the black hole spin parameters.
629:
630: With this paper, we now have in hand a complete picture of how the
631: corotation resonance affects various diskoseismic modes
632: in black-hole accretion discs, at least in the linear regime:
633: Non-axisymmetric g-modes are heavily damped at the corotation resonance
634: (Kato 2003; Li et al.~2003), while p-modes (inertial-acoustic modes)
635: can be overstable due to the corotational wave absorption (see Lai \& Tsang
636: 2008 and references therein). The corotational damping rates of c-modes
637: are much smaller than those of g-modes.
638:
639: Diskoseismic c-modes have been invoked to explain low-frequency
640: variabilities in black-hole X-ray binaries (van der Klis 2006;
641: Remillard \& McClintock 2006). Fu \& Lai (2008) showed that the basic
642: properties of c-modes are largely unaffected by the disc magnetic
643: fields and thus these modes are present in real discs. The results
644: presented in this paper show that in order for the c-modes to be
645: observable, a sufficiently strong excitation mechanism is needed to
646: overcome the corotational damping.
647:
648:
649: \section*{Acknowledgments}
650:
651: This work has been supported in part by NASA Grant NNX07AG81G, NSF
652: grants AST 0707628, and by {\it Chandra} grant TM6-7004X
653: (Smithsonian Astrophysical Observatory).
654:
655:
656: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
657: \begin{thebibliography}{99}
658:
659: \bibitem[]{}
660: Abramowitz M. \& Stegun I.A., 1964, Handbook of Mathematical Functions. Dover Press, New York
661:
662: \bibitem[]{}
663: Aliev, A.N., Gal'tsov, D.V. 1981, General Relativity and Gravitation,
664: 13, 899
665:
666: %\bibitem[]{}
667: %Ferreira, B. T. \& Ogilvie, G. I., 2008, astro-ph/08031671
668:
669: \bibitem[]{}
670: Fu, W., Lai, D. 2008, ApJ, in press (arXiv:0806.1938)
671:
672: \bibitem[]{}
673: Goldreich, P., Tremaine, S. 1979, ApJ, 233, 857
674:
675: \bibitem[]{}
676: Ipser, J.R., 1994, ApJ, 435, 767
677:
678: \bibitem[]{}
679: Ipser, J.R., 1996, ApJ, 458, 508
680:
681: \bibitem[]{}
682: Kato, S. 1983, PASJ, 35, 249
683: %Low-frequency, one-armed oscillations of Keplerian gaseous disks
684:
685: \bibitem[]{}
686: Kato, S., 1989, PASJ, 41, 745
687: %Low-frequency, one-armed corrugation waves in relativistic accretion disks
688:
689: \bibitem[]{}
690: Kato, S., 1990, PASJ, 42, 99
691: %Trapped one-armed corrugation waves and QPOs
692:
693: \bibitem[]{}
694: Kato, S., 2001, PASJ, 53, 1
695:
696: \bibitem[]{}
697: Kato, S., 2003, PASJ, 55, 257
698: % g-mode damping
699:
700: %\bibitem[]{}
701: %Kato, S., 2003b, PASJ, 55, 801
702: % Excitation by disk deformation
703:
704: %\bibitem[]{}
705: %Kato, S., 2008, PASJ, 60, 111
706:
707: \bibitem[]{}
708: Kato, S. \& Fukue, J., 1980, PASJ, 32, 377
709:
710: \bibitem[]{}
711: Lai, D. \& Tsang, D., 2008, MNRAS, submitted
712:
713: \bibitem[]{}
714: Lai, D., Zhang, H. 2008, ApJ, 683, 949
715:
716: \bibitem[]{}
717: Lee, W. H., Abramowicz, M. A. \& Kluzniak, W., 2004, ApJ, 603, L93
718:
719: \bibitem[]{}
720: Li, L. X., Goodman, J. \& Narayan, R., 2003, ApJ, 593, 980
721:
722: \bibitem[]{}
723: Nowak, M. A. \& Wagoner, R. V., 1992, ApJ, 393, 697
724:
725: \bibitem[]{}
726: Okazaki, A.T. \& Kato, S., 1985, PASJ, 37, 683
727: %One-armed adiabatic oscillations of a nonself-gravitating polytropic disk.
728: %I - A disk with constant thickness
729:
730: \bibitem[]{}
731: Okazaki, A.T., Kato, S., \& Fukue, J., 1987, PASJ, 39, 457
732:
733: \bibitem[]{}
734: Perez, C.A., Silbergleit, A.S., Wagoner, R.V. \& Lehr, D.E., 1997, ApJ, 476, 589
735:
736: \bibitem[]{}
737: Paczynki, B. \& Wiita, P.J. 1980, A\&A 88, 23
738:
739: \bibitem[]{}
740: Remillard, R. A. \& McClintock, J. E., 2006, ARAA, Vol. 44, pp. 49-92
741:
742: \bibitem[]{}
743: Silbergleit, A. S., Wagoner, R. V. \& Ortega-Rodriguez, M., 2001, ApJ, 548, 335
744:
745: \bibitem[]{}
746: Tsang, D. \& Lai, D., 2008, MNRAS, 387, 446
747:
748: \bibitem[]{}
749: van der Klis, M. 2006, in Compact Stellar X-ray Sources, ed.
750: W.H.G. Lewin and M. van der Klis (Cambridge Univ. Press)
751: (astro-ph/0410551)
752:
753: \bibitem[]{}
754: Wagoner, R. V., 1999, Phys. Rep., 311, 259
755:
756: \bibitem[]{}
757: Zhang, H. \& Lai, D., 2006, MNRAS, 368, 917
758:
759: \end{thebibliography}
760: \end{document}