1: \documentclass[12pt,preprint]{aastex}
2:
3:
4:
5: \newcommand\be{\begin{equation}}
6: \newcommand\ee{\end{equation}}
7: \newcommand\ba{\begin{eqnarray}}
8: \newcommand\ea{\end{eqnarray}}
9: \newcommand\hatl{{\bf\hat l}}
10: \newcommand\bB{\textbf{\emph{B}}}
11: \newcommand\bT{\textbf{\emph{T}}}
12: \newcommand\bG{\textbf{\emph{G}}}
13: \newcommand\tomega{{\tilde\omega}}
14: \newcommand\bn{\begin{enumerate}}
15: \newcommand\en{\end{enumerate}}
16:
17: \newenvironment{inlinefigure}{
18: \medskip
19: \def\@captype{figure}
20: \noindent\begin{minipage}{0.999\linewidth}\begin{center}}
21: {\end{center}\end{minipage}\medskip}
22:
23: \begin{document}
24:
25: \title{Effects of Magnetic Fields on the Diskoseismic Modes of Accreting Black Holes}
26: \author{Wen Fu and Dong Lai}
27: \affil{Department of Astronomy, Cornell University, Ithaca, NY 14853\\
28: Email: wenfu, dong@astro.cornell.edu}
29:
30: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
31: \begin{abstract}
32: The origin of the rapid quasi-periodic variabilities observed in a number
33: of accreting black hole X-ray binaries is not understood. It has been
34: suggested that these variabilities are associated with diskoseismic oscillation
35: modes of the black hole accretion disk. In particular, in a disk with no magnetic
36: field, the so-called g-modes (inertial oscillations) can be self-trapped
37: at the inner region of the disk due to general relativistic effects. Real accretion
38: disks, however, are expected to be turbulent and contain appreciable magnetic fields.
39: We show in this paper that even a weak magnetic field (with the magnetic energy much
40: less than the thermal energy) can modify or ``destroy'' the self-trapping zone of disk
41: g-modes, rendering their existence questionable in realistic black hole accretion disks.
42: The so-called corrugation modes (c-modes) are also strongly affected when the poloidal
43: field approaches equal-partition. On the other hand, acoustic oscillations (p-modes),
44: which do not have vertical structure, are not affected qualitatively by the magnetic
45: field, and therefore may survive in a turbulent, magnetic disk.
46: \end{abstract}
47:
48: \keywords{accretion: accretion disks --- black hole physics --- magnetic fields --- X-ray: binaries
49: --- hydrodynamics --- MHD}
50:
51: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
52: \section{Introduction}
53:
54: In recent years, quasi-periodic variability has been observed from
55: a number of Galactic compact X-ray binary systems. Of particular
56: interest are several accreting black hole (BH) binaries which show
57: pairs of quasi-periodic oscillations (QPOs) of fixed frequencies
58: having ratios close to $2:3$ (for example, GRO J1655-40 shows
59: $f=300,~450$~Hz; see Remillard \& McClintock 2006 for a review).
60: The origin of these QPOs is not understood. The fixed frequencies
61: and frequency ratio led to the suggestion that these QPOs involve
62: certain nonlinear resonant phenomenon in the disk (e.g., coupling
63: between the radial and vertical epicyclic oscillations of the disk
64: fluid element; Kluzniak \& Abramowicz 2002), but so far no fluid
65: dynamical model producing these resonances has been developed (see
66: Rebusco 2008 and references therein). Alternatively, it has been
67: suggested that these QPOs may arise from acoustic oscillations in
68: an accretion torus (Rezzolla et al. 2003; Lee, Abramowicz \&
69: Kluziniak 2004), with the oscillation frequencies determined by
70: the (unknown) radial extent of the torus. Perhaps the
71: theoretically most developed model for QPOs is the relativistic
72: diskoseismic oscillation model (Kato \& Fukue 1980; Okazaki et
73: al.~1987; Nowak \& Wagoner 1991; see Wagoner 1999; Kato 2001 for
74: reviews), in which general relativistic (GR) effect produces
75: trapped g-mode (also called inertial mode or inertial-gravity
76: mode) oscillations in the inner region of the disk. Various
77: theoretical issues related to this model have been studied, such
78: as the role of corotational wave absorption (Kato 2003; Li,
79: Goodman \& Narayan 2003; Silbergleit \& Wagoner 2007) and resonant
80: mode excitations due to global disk deformation (Kato 2008;
81: Ferreira \& Ogilvie 2008).
82:
83: The studies of the oscillation modes of disks/torii, such as those
84: mentioned above, usually assume that the unperturbed flow is laminar
85: and has no magnetic field. Real accretion disks, on the other hand,
86: are highly turbulent due to the nonlinear development of
87: magnetorotational instability (MRI) (see Balbus \& Hawley 1998 for a
88: review). The question therefore arises as to how the MRI-driven
89: turbulence affects the oscillation modes obtained from hydrodynamical
90: models and to what extent these trapped modes remain ``valid'' in a
91: realistic situation. Arras, Blaes \& Turner (2006) attempted to
92: address this issue by carrying out MHD simulations in the shearing-box
93: geometry. They showed that axisymmetric standing sound waves give rise
94: to distinct peaks in the temporal power spectrum, while inertial waves
95: do not. The discrete frequencies obtained by them were due to the
96: imposed periodic boundary conditions adopted in the simulations, and
97: not due to any relativistic effect. Arras et al. suggested that their
98: result posses a serious problem for QPO models based on g-modes
99: (inertial waves). Recently, Reynolds \& Miller (2008) reported on the
100: results of global simulations of BH accretion disks (using
101: Paczynski-Wiita pseudo-Newtonian potential) and showed that, while
102: axisymmetric g-mode oscillations manifest in the hydrodynamic disk
103: with no magnetic field, they disappear in the magnetic disk where MHD
104: turbulence develops.
105:
106: In this paper, we study analytically the effects of magnetic
107: fields on the relativistic diskoseismic modes in accretion disks
108: around BHs. We consider both poloidal and toroidal fields and use
109: local analysis of the full MHD equations to examine how the
110: magnetic field changes the radial wave propagation diagrams for
111: various modes. We show that the trapping region of g-modes can be
112: easily ``destroyed'' even when the disk field strength is such
113: that the associated Alfv\'{e}n speed is much smaller than the
114: sound speed. On the other hand, the propagation characteristics of
115: p-modes (acoustic oscillations) and c-modes are largely unchanged.
116: Note that since we assume laminar flows for our unperturbed disks,
117: we do not directly address the effects of turbulence on disk
118: modes. However, we believe that our work is relevant to this
119: issue, since magnetic fields naturally arise in a turbulent disk.
120:
121: We summarize the basic MHD equations in \S 2 and
122: review the properties of diskoseismic modes important for our analysis
123: in \S 3. We examine in \S 4 and \S5 the effects of poloidal field and
124: toroidal field on those modes, respectively, and discuss the
125: implications of our result in \S 6.
126:
127:
128: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
129: \section{Basic Equations}
130:
131: We consider a non-self-gravitating accretion disk, satisfying
132: the usual ideal MHD equations:
133: \ba
134: && {\partial{\rho} \over \partial t}
135: +\nabla\cdot(\rho \textbf{\emph{v}})=0,\\
136: && {\partial{\textbf{\emph{v}}} \over \partial
137: t}+({\textbf{\emph{v}}}\cdot\nabla){\textbf{\emph{v}}}
138: =-\frac{1}{\rho}\nabla \Pi-{\nabla \Phi}+\frac{1}{\rho}\bT,\\
139: && {{\partial {\bB}} \over \partial t}=\nabla\times({\textbf{\emph{v}}}
140: \times{\bB}).
141: \ea
142: Here $\rho,\,P,\,\textbf{\emph{v}}$ are the fluid density, pressure and velocity,
143: $\Phi$ is the gravitational potential, and
144: \begin{equation}
145: \Pi \equiv P+\frac{B^2}{8\pi},\quad
146: \bT \equiv \frac{1}{4\pi}(\bB\cdot\nabla)\bB
147: \end{equation}
148: are the total pressure and
149: the magnetic tension, respectively. The magnetic field $\bB$ also
150: satisfies the equation $\nabla\cdot\bB=0$. We assume that the fluid obeys the barotropic
151: equation of state $P=P(\rho)$.
152:
153: We adopt the cylindrical coordinates $(r, \phi, z)$ which are
154: centered on the central BH and have the $z$-axis in the direction
155: perpendicular to the disk plane. The unperturbed background flow
156: is assumed to be axisymmetric with a velocity field
157: $\textbf{\emph{v}}=r\Omega(r)\hat \phi$, and magnetic field $\bB=
158: B_\phi(r)\hat \phi+B_z \hat z$, i.e., $B_z$ is constant while
159: $B_\phi$ has a radial dependance. Force balance in the unperturbed
160: flow implies \begin{equation} \bG\equiv
161: \frac{1}{\rho}\nabla\Pi-\frac{1}{\rho}\bT=\Omega^2 r
162: \hat{r}-\nabla \Phi. \end{equation}
163:
164: Consider perturbations of the form $e^{im\phi-i\omega t}$. The
165: linearized fluid equations are \ba && -i\tomega\delta
166: \rho+\frac{1}{r}\frac{\partial}{\partial r}(\rho r \delta v_r)
167: +\frac{im\rho}{r}\delta v_{\phi}+\frac{\partial}{\partial z}(\rho \delta v_z)=0,\\
168: && -i\tomega\delta v_r-2\Omega\delta v_{\phi}= G_r\frac{\delta
169: \rho}{\rho}-\frac{1}{\rho}\frac{\partial}{\partial r}\delta \Pi
170: +\frac{1}{\rho}(\delta \bT)_{r},\label{eq:linear2}\\
171: && -i\tomega\delta v_{\phi}+\frac{\kappa^2}{2\Omega}\delta v_r=
172: -\frac{im}{\rho r}\delta \Pi+\frac{1}{\rho}(\delta \bT)_{\phi},\\
173: && -i\tomega\delta v_z=G_z\frac{\delta
174: \rho}{\rho}-\frac{1}{\rho}\frac{\partial}{\partial z}\delta \Pi
175: +\frac{1}{\rho}(\delta \bT)_{z},\label{eq:linear4}\\
176: && -i\tomega\delta B_r=\left(\frac{imB_{\phi}}{r}
177: +B_z\frac{\partial}{\partial z}\right)\delta v_r,\\
178: && -i\tomega\delta B_{\phi}=-\frac{\partial}{\partial
179: r}(B_{\phi}\delta v_r)+B_z \frac{\partial}{\partial z}\delta
180: v_{\phi}-B_{\phi}\frac{\partial}{\partial z}\delta v_z+r\frac{d\Omega}{dr}\delta B_r,\\
181: && -i\tomega\delta B_z=-\frac{B_z}{r}\frac{\partial}{\partial
182: r}(r\delta v_r)-\frac{imB_z}{r}\delta
183: v_{\phi}+\frac{imB_{\phi}}{r}\delta v_z, \ea
184: where
185: \begin{equation} \tomega=\omega-m\Omega \end{equation} is the
186: comoving wave frequency, and
187: \begin{equation}
188: \kappa^2 \equiv \frac{2\Omega}{r}\frac{d(r^2\Omega)}{dr} \end{equation}
189: is
190: the radial epicyclic frequency.
191: $\delta\rho,~\delta\Pi,~\delta{\textbf{\emph{v}}},~\delta{\bf B}$
192: are Eulerian perturbations, and $\rho,~\bB$ refer to the
193: unperturbed flow variables. In addition, for barotropic fluid, we
194: have \begin{equation} \delta \rho=\frac{1}{c_s^2}\delta
195: P=\frac{1}{c_s^2}(\delta \Pi-\frac{1}{4\pi}\bB\cdot\delta \bB),
196: \end{equation} where $c_s$ is the sound speed.
197:
198: To perform local (WKB) analysis, we consider perturbations with
199: spatial dependence $e^{ik_r r+ik_z z}$. In the leading-order
200: approximation, we keep only the radial gradient of $\Omega(r)$ and
201: $B_\phi(r)$ while assuming that the variation scales of all the
202: other background quantities are much larger than the wavelength of
203: the perturbation, i.e., $k_r, k_z \gg 1/r$. The linearized MHD
204: equations then reduce to
205: \ba && -{i\tomega \over \rho c_s^2}\delta
206: \Pi + ik_r\delta v_r +ik_{\phi} \delta v_{\phi}+ik_z \delta v_z
207: +{i\tomega B_{\phi} \over 4\pi \rho c_s^2}\delta
208: B_{\phi} + {i\tomega B_z \over 4\pi \rho c_s^2}\delta B_z=0,\label{eq:wkb1}\\
209: && -{ik_r \over \rho}\delta \Pi +i\tomega \delta v_r +2\Omega
210: \delta v_{\phi}+ \left({ik_z B_z \over 4\pi \rho}+{ik_{\phi}
211: B_{\phi} \over 4\pi \rho}\right)\delta B_r - {B_{\phi} \over 2\pi
212: \rho
213: r}\delta B_{\phi}=0,\\
214: && -{ik_{\phi} \over \rho}\delta \Pi-{\kappa^2 \over
215: 2\Omega}\delta v_r+i\tomega\delta v_{\phi}+
216: \frac{(q+1)B_{\phi}}{4\pi\rho r}\delta B_r+\left({ik_z B_z \over
217: 4\pi \rho}+ {ik_{\phi}B_{\phi}
218: \over 4\pi \rho}\right)\delta B_{\phi}=0,\\
219: && -{ik_z \over \rho}\delta \Pi +i\tomega\delta v_z+\left({ik_z
220: B_z \over 4\pi \rho}+{ik_{\phi} B_{\phi} \over
221: 4\pi \rho}\right)\delta B_z=0,\\
222: &&
223: (ik_{\phi} B_{\phi} +ik_z B_z)\delta v_r +i\tomega\delta B_r=0,\\
224: && ik_r B_{\phi}\delta v_r -ik_z B_z \delta v_{\phi} +ik_z
225: B_{\phi} \delta v_z-p\Omega\delta
226: B_r-i\tomega\delta B_{\phi}=0,\\
227: && ik_r B_z\delta v_r +ik_{\phi}B_z \delta
228: v_{\phi}-ik_{\phi}B_{\phi}\delta v_z -i\tomega\delta
229: B_z=0,\label{eq:wkb7} \ea where $k_\phi \equiv m/r$. We have
230: assumed $B_\phi\sim r^q$ and $p \equiv d\ln \Omega/d\ln r$. Note
231: that in deriving eqs.~(16)-(22), we have dropped the terms
232: proportional to $G_r$ and $G_z$ in eqs.~(7) and (9): since
233: $G_r=(\Omega^2-\Omega_K^2)r \sim \Omega_K^2 r (H/r)^2$ and $G_z
234: \sim \Omega_K^2 z$ (where $\Omega_k$ is the Keplerian frequency,
235: i.e., the angular frequency in the absence of pressure force),
236: $G_r$ is much smaller than the other terms in eq.~(7) provided
237: that $k_r r \gg 1+v_{A\phi}^2/c_s^2$, and $G_z$ is also negligible
238: if we focus on the mid-plane of the disk.
239:
240: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
241: \section{Hydrodynamic Limit: Diskoseismic Modes}
242:
243: In the absence of magnetic fields, for $k_z, k_r \gg k_{\phi}$, the perturbed MHD equations (16)-(22)
244: lead to the dispersion relation:
245: \begin{equation}
246: (\tomega^2-\kappa^2)(\tomega^2-k_z^2 c_s^2)=k_r^2 c_s^2 \tomega^2.
247: \label{eq:disp1}\end{equation}
248: For $k_z=0$ (or $\tomega^2\gg k_z^2c_s^2$), this becomes
249: $\tomega^2=k_r^2c_s^2+\kappa^2$, the usual dispersion relation for
250: spiral density wave; for $\tomega^2\ll k_z^2 c_s^2$, this becomes
251: $\tomega=\pm \kappa k_z/(k_r^2+k_z^2)^{1/2}$, describing inertial oscillations
252: (e.g., Goodman 1993).
253:
254: For an accretion disk, with scale height $H\ll r$, the vertical
255: dependence of the perturbation is not well described by the plane
256: wave $e^{ik_z z}$ unless $k_zH\gg 1$. Okazaki et al.~(1987) showed
257: that for a thin disk, the perturbation equations can be
258: approximately separated in $r$ and $z$ (see also Nowak \& Wagoner
259: 1991, 1992; Ipser 1994). For example, for vertically isothermal disks with constant
260: scale height $H$, one finds $\delta P(r,z),\delta v_r(r,z),\delta
261: v_\phi(r,z) \propto H_n(z/H)$, while $\delta v_z(r,z)\propto
262: H'_n(z/H)$, where $H_n(Z)$ (with $n=0,1,\cdots$) is the Hermite
263: polynomials and $H_n'(Z)=dH_n(Z)/dZ$. With this separation of
264: variables, Okazaki et al. (1987) obtained the dispersion relation for a
265: given $n$: \begin{equation}
266: (\tomega^2-\kappa^2)(\tomega^2-n\Omega_\perp^2)=k_r^2 c_s^2
267: \tomega^2, \label{eq:disp2}\end{equation} where $\Omega_\perp$ is the
268: vertical epicyclic frequency and is related to $H$ by
269: $H=c_s/\Omega_\perp$. An important property of relativistic
270: disks around BHs is that $\kappa$ is non-monotonic.
271: Three types of trapped modes can be identified (see Fig.~1; see
272: also Wagoner 1999, Kato 2001 and Ortega-Rodriguez et al. 2006 for
273: reviews):
274:
275: (i) P-modes. For $n=0$, waves can propagate in the region where $\tomega^2 > \kappa^2$.
276: These are acoustic waves (modified by disk rotation), and have also been termed inertial-acoustic
277: modes. If waves can be
278: reflected at the disk inner radius ($r_{\rm ISCO}$, the inner-most stable circular
279: orbit), discrete p-modes can
280: be self-trapped at the inner-most region of the disk (see Fig.~1a,~1c).
281:
282: (ii) G-modes. For $n\geq 1$, waves can propagate in the region where
283: $\tomega^2 < \kappa^2 < n\Omega_\perp^2$ or $\tomega^2>\Omega_\perp^2 >\kappa^2$
284: (note that $\kappa< \Omega_\perp$ in GR). The former specifies
285: the g-mode propagation zone: self-trapped g-modes can be maintained in the
286: region where $\kappa$ peaks (for $m=0$: see Fig.~1b) or in the region where
287: $\Omega-\kappa/m < \omega/m < \Omega+\kappa/m$ (for $m\neq 0$: see Fig.~1d).\footnote
288: {Note that non-axisymmetric g-modes with $\omega/m < \Omega(r_{\rm ISCO})$
289: contain corotation resonance in the wave zone, leading to strong damping of the mode
290: (Kato 2003; Li et al. 2003; Zhang \& Lai 2006). On the other hand, modes with $\Omega(r_{\rm ISCO}) < \omega/m
291: < {\rm max}(\Omega+\kappa/m)$ do not suffer corotational damping, and are therefore of great
292: interest.}
293: Because these discrete, self-trapped modes do not require special boundary conditions
294: (e.g., wave reflection at $r=r_{\rm ISCO}$), they have been the focus of most studies of
295: relativistic diskoseismology. Note that although we call these g-modes (following the terminology
296: of Kato 2001 and Wagoner 1999), they have no relation to gravity waves, which are driven
297: by buoyancy. Instead, these modes describe inertial oscillations, and have also been termed
298: inertial modes (or inertial-gravity modes).
299:
300: \begin{inlinefigure}
301: \scalebox{1}{\rotatebox{0}{\plotone{f1.ps}}} \figcaption{Wave
302: propagation diagram showing various trapped modes in BH accretion
303: disks: (a) axisymmetric p-mode; (b) axisymmetric g-mode;
304: (c)non-axisymmetric p-mode; (d) non-axisymmetric g-mode and
305: c-mode. The curves depict various critical frequencies ($\kappa$,
306: $\sqrt{n}\Omega_\perp$, $\Omega$, $\Omega \pm \kappa/m$, $\Omega
307: \pm \sqrt{n}\Omega_\perp/m$), the vertical dotted lines denote the
308: inner-most stable circular orbit (ISCO). The curvy horizontal
309: lines specify wave propagation zones and the height of the line is
310: $\omega$ (for the top panels) or $\omega/m$ (for the bottom
311: panels) of the mode.}
312: \end{inlinefigure}
313:
314: (iii) C-modes. For $n\geq 1$ and $m\geq 1$, the wave propagation condition $\tomega^2 >n\Omega_\perp^2 >\kappa^2$
315: leads to an additional wave trapping region, where $\omega/m < \Omega-\sqrt{n}\Omega_\perp/m$ (see Fig.~1d).
316: Note that for spinning BHs, $\Omega_\perp <\Omega$. Clearly, these modes exist only when
317: $\Omega-\sqrt{n}\Omega_\perp/m >0$ and wave reflection occurs at $r=r_{\rm ISCO}$. Following the previous
318: works (e.g., Kato 1990 and Silbergleit et al. 2001, who focused on the ``fundamental''
319: $n=m=1$ mode, corresponding to the Lense-Thirring precession of the inner disk), we call these (``corrugation'') c-modes.
320:
321:
322: Comparing eqs.~(23) and (24), we see that we can
323: obtain the radial dispersion relation of different modes by adopting the
324: vertical ``quantization'' condition $k_z=\sqrt{n}/H$ in
325: eq.~(23), with $k_z=0$ specifying p-modes. In a generic disk
326: (e.g., when the disk is not isothermal vertically),
327: the same ``quantization'' condition would not hold, but we still expect
328: $k_z\sim 1/H\sim \Omega_\perp/c_s$ for the (vertically) lowest-order g-mode or c-mode.
329: In the next sections, we will adopt $k_z=\sqrt{\eta}\Omega_\perp/c_s$, with $\eta$ of order
330: unity, when we study how magnetic fields modify low-order g-modes and c-modes.
331:
332: Our approach in this paper is based on Newtonian theory. GR effect can be
333: incorporated into our analysis by using the Paczynski-Witta pseudo-Newtonian potential,
334: $\Phi=-M/(r-2M)$. Alternatively, we could simply replace the Newtonian $\Omega,\,\Omega_\perp,\,\kappa$
335: by their exact general relativistic counterparts (e.g., Okazaki et al. 1987):
336: \ba
337: &&
338: \Omega={{\sqrt {M/r^3}} \over {1+ a\sqrt {M/r^3}}},\\
339: &&
340: \Omega_{\perp}=\Omega\left[1 - {{4a M^{1/2}} \over r^{3/2}}+{{3a^2} \over r^2}\right]^{1/2},\\
341: &&
342: \kappa=\left[{{M(r^2-6Mr + 8aM^{1/2}r^{1/2}-3a^2)} \over {r^2(r^{3/2} + aM^{1/2})^2}}\right]^{1/2}
343: \ea
344: (in geometric units such that $G=c=1$),
345: where $a = J_s/M$ is the spin parameter of the black hole. In general, $\Omega \geq \Omega_\perp > \kappa$.
346: In the case of a Schwarzschild BH, $\Omega=\Omega_\perp > \kappa,$ with $\kappa$ peaks at $r=8{\rm M}$
347: and becomes zero at $r_{\rm ISCO}=6{\rm M}$.
348: This non-monotonic behavior of the radial epicyclic frequency is preserved for Kerr BHs, and,
349: as discussed above, is the key ingredient for the existence of trapped diskoseismic modes.
350:
351:
352: \section{Effect of Poloidal Fields}
353: We first consider the case of a pure poloidal field, with $B_{\phi}=0$. For $k_z, k_r \gg k_{\phi}$, equations~(16)-(22)
354: then lead the dispersion relation:
355: \ba
356: &&\tomega^6-[(k_z^2+k_r^2)(c_s^2+v_{Az}^2)+k_z^2v_{Az}^2+\kappa^2]\tomega^4\nonumber\\
357: &&+\left\{k_z^2 v_{Az}^2\left[(k_z^2+k_r^2)(2c_s^2+v_{Az}^2)
358: +\frac{d\Omega^2}{d\ln r}\right]+\kappa^2k_z^2c_s^2\right\}\tomega^2\nonumber\\
359: &&-k_z^4v_{Az}^2c_s^2\left[(k_z^2+k_r^2)v_{Az}^2+\frac{d\Omega^2}{d\ln r}\right]=0,\label{eq:mri}
360: \ea
361: where $v_{Az}\equiv B_z/\sqrt{4\pi \rho}$. In the incompressible limit, this reduces to the dispersion relation found in,
362: e.g., Balbus \& Hawley (1991).
363: For a given ${\bf k}=(k_r, k_{\phi}, k_z)$, equation~(28) admits three branches, corresponding to fast,
364: slow magnetosonic waves and Alfv\'{e}n wave, all modified by differential rotation. For $k_z \gg k_r$, the Alfv\'{e}n
365: branch can become unstable when $k_z^2 v_{Az}^2 < -{d\Omega^2}/{d\ln r}.$ This is the well-known MRI (e.g., Balbus \&
366: Hawley 1998).
367:
368:
369: \subsection{P-modes}
370: If $k_z=0$, equation (28) reduces to
371: \begin{equation}
372: \tomega^2=\kappa^2+k_r^2(c_s^2+v_{Az}^2).
373: \end{equation}
374: This is almost the same expression as in pure hydrodynamic case ($\tomega^2=\kappa^2+k_r^2 c_s^2$), the only
375: difference being that the sound speed is replaced by fast magnetosonic wave speed, $\sqrt{c_s^2+v_{Az}^2}$.
376: Thus the basic property of p-modes is not affected by poloidal magnetic fields.
377:
378: \subsection{G-modes}
379: For a fixed $k_z=\sqrt{\eta} /H=\sqrt{\eta} \Omega_\perp/c_s$ (see \S3), we can rewrite eq.~(28) as
380: an expression for $k_r^2$:
381: \begin{equation}
382: (c_s^2+v_{Az}^2)k_r^2={{(\tomega^2-\omega_1^2)(\tomega^2-\omega_2^2)(\tomega^2-\omega_5^2)} \over {(\tomega^2-\omega_3^2)(\tomega^2-\omega_4^2)}}.
383: \end{equation}
384: The five critical frequencies are given by
385: \ba
386: &&\omega_1^2=\eta (\Omega_{\perp})^2,\\
387: &&\omega_2^2={1 \over 2}\left[\kappa^2+2\eta(\Omega_{\perp})^2 b^2+\sqrt{\kappa^4+16\eta (\Omega_{\perp}\Omega)^2 b^2}\right]\label{eq:omega2}\\
388: &&\omega_3^2=\eta (\Omega_{\perp})^2 b^2,\\
389: &&\omega_4^2=\eta (\Omega_{\perp})^2 {b^2 \over {1+b^2}},\\
390: &&\omega_5^2={1 \over 2}\left[\kappa^2+2\eta (\Omega_{\perp})^2 b^2-\sqrt{\kappa^4+16\eta (\Omega_{\perp}\Omega)^2 b^2}\right]
391: \ea
392: where $b\equiv v_{Az}/c_s$.
393:
394: Equation~(30) allows us to identify various wave propagation regions ($k_r^2 > 0$). We first consider subthermal
395: fields, with $b<1$. When $b\lesssim 0.4$ (and with $\eta =1$ for the lowest order g-modes),
396: the five critical frequencies satisfy
397: $\omega_1^2 > \omega_2^2 > \omega_3^2 > \omega_4^2 > 0 > \omega_5^2$ in the inner region of the disk.
398: Thus there are three wave propagation regions:
399:
400: \ba
401: &&{\rm Region ~I}:~~ \tomega^2 > \omega_1^2,\\
402: &&{\rm Region ~II}: \omega_3^2 < \tomega^2 < \omega_2^2,\\
403: &&{\rm Region ~III}: \tomega^2 < \omega_4^2.
404: \ea
405:
406: \begin{inlinefigure}
407: \scalebox{1}{\rotatebox{0}{\plotone{f2.ps}}} \figcaption{The
408: effect of poloidal magnetic field on the g-mode propagation zone
409: for $m=0, \eta =1$. In each panel, the upper three curves are
410: $\omega_2$ (eq.~[32]) and the lower two curves are $\omega_3$
411: (eq.~[33]). Axisymmetric g-modes of frequency $\omega$ can
412: propagate in the region where $\omega_3 < \omega < \omega_2$. The
413: solid line refers to the case of $b=0$, the short-dashed line
414: $b=0.03$ and the long-dashed line $b=0.1$, where $b\equiv
415: v_{Az}/c_s$, with $v_{Az}=B_z/\sqrt{4\pi \rho}$ (Alfv\'{e}n speed)
416: and $c_s$ the sound speed. The vertical dotted lines correspond to
417: the inner disk radius at ISCO. The curvy horizontal lines specify
418: the wave propagation zones, and the height of the line is $\omega$
419: of the mode. The lower and upper panels are for the case of a
420: Schwarzschild BH ($a=0$) and a Kerr BH ($a=0.8{\rm M}$),
421: respectively. The angular frequencies are in units of
422: $M^{-1}=c^3/(GM)$.}
423: \end{inlinefigure}
424:
425:
426: Region II corresponds to the original g-mode cavity modified by
427: the magnetic field; in the zero field limit, $\omega_3=0$, $\omega_2=\kappa$
428: and eq.~(37) reduces to $\tomega^2 < \kappa^2$. Fig.~2 depicts the
429: critical frequencies $\omega_2$ and $\omega_3$ for several values of $b$. This also
430: serves as the propagation diagram for $m=0$ g-modes (wave can
431: propagate in region where $\omega_3 < \omega <\omega_2$). We see
432: that as the magnetic field increases, the g-mode self-trapping zone gradually shrinks
433: and disappears even when the magnetic field is still very
434: subthermal (for a Schwarzschild BH, this occurs for $b
435: \gtrsim 0.08$). More precisely, the g-mode cavity can still exist for large b, but
436: it now requires a reflection boundary at $r_{\rm ISCO}$.
437: This behavior can be easily understood by
438: inspecting eq.~(32):
439: While $\kappa$ peaks at some radius $r_{\rm max}$, $\Omega_\perp$ and
440: $\Omega$ both increase monotonically with decreasing $r$. Since
441: $\Omega_\perp$ and $\Omega$ are much larger than $\kappa$ in the
442: inner region of the disk, the $2(\Omega_\perp b)^2$ term or
443: the $4\Omega_\perp \Omega b$ term can dominate over $\kappa^2$
444: even when $b$ is still small, therefore making the self-trapping zone disappear.
445: Roughly, this occurs at $b \gtrsim b_{\rm crit} \sim
446: (\kappa^2/2\Omega_\perp \Omega)_{r_{\rm max}}$.
447:
448: \begin{inlinefigure}
449: \scalebox{1}{\rotatebox{0}{\plotone{f3.ps}}} \figcaption{The
450: effect of poloidal magnetic field on the g-mode propagation zone
451: for $m \neq 0, \eta =1, a=0$. The three panels are for
452: $b=v_{Az}/c_s =0$, $0.005$ and $0.015$. The solid line,
453: short-dashed lines and the long-dashed lines show $\Omega$,
454: $\Omega \pm \omega_2$ and $\Omega \pm \omega_2/2$, respectively.
455: In the bottom panel, the dotted lines show $\Omega \pm \omega_3$
456: (In the upper and middle panels, $\Omega \pm \omega_3$ almost
457: concide with $\Omega$, since $\omega_3=0$ for $b=0$ and $\omega_3
458: \ll \Omega$ for $b \ll 1$). Non-axisymmetric g-modes can propagate
459: in the region where $\Omega-\omega_2/m < \omega/m <
460: \Omega-\omega_3/m$ or $\Omega+\omega_3/m < \omega/m <
461: \Omega+\omega_2/m$. The vertical dotted lines correspond to the
462: inner disk radius at ISCO. The curvy horizontal lines in top and
463: middle panels specify wave propagation zones and the height of the
464: line is $\omega/m$ of the mode. Note that the self-trapping zone
465: (depicted in the upper and middle panels) disappears as b
466: increases. The angular frequencies are in units of
467: $M^{-1}=c^3/(GM)$.}
468: \end{inlinefigure}
469:
470:
471:
472: For non-axisymmetric perturbations ($m \neq 0$), the wave propagation region II is determined by
473: (i) $\Omega-\omega_2/m < \omega/m < \Omega+\omega_2/m,$ and
474: (ii) $\omega/m >\Omega+\omega_3/m$ or $\omega/m < \Omega-\omega_3/m$.
475: Fig.~3 shows the propagation diagram. As mentioned before (see Footnote~1), for $b=0$,
476: only the modes with $\omega> m\Omega(r_{\rm ISCO})$
477: are of interest, since otherwise there is a corotation resonance in the wave zone,
478: leading to strong mode damping (Kato 2003; Li, Narayan \& Goodman 2003; Zhang \& Lai 2006). Thus,
479: self-trapped g-modes reside around the radius where $\Omega+\omega_2/m$ is the maximum
480: (and this maximum arises because $\kappa$ depends nonmonotonically on $r$).
481: We see from Fig.~3 that this g-mode self-trapping region
482: disappears as $b$ increases. The larger $m$ is, the more fragile is the cavity. For example, the $m=1$ cavity
483: disappears for $b \gtrsim 0.015$, while for $m=2$, this occurs for $b \gtrsim 0.005$.
484:
485:
486: \subsection{C-modes}
487: For $m \neq 0$ and $\eta \sim 1$, equation (30) also describes trapped c-modes. When $b \lesssim 0.4$,
488: $\omega_1^2$ is the largest among all the critical frequencies and the c-mode propagation zone corresponds to
489: Region I (see eq.~[36]). Note that since $\omega_1$ is not affected by the magnetic field, the trapping region is
490: determined by $\omega/m < \Omega-\omega_1/m =\Omega-\Omega_\perp/m$ (for $\eta =1$, see the upper panel of Fig.~4; cf.
491: Fig.~1d).
492: When $b \gtrsim 0.4$, the ordering between $\omega_1$ and $\omega_2$ switches and c-modes propagate in the region
493: where $\tomega^2 > \omega_2^2$, with the trapping zone determined by $\omega/m < \Omega -\omega_2/m$ (see the
494: bottom panel of Fig.~4). Thus, in the presence of a reflection boundary at $r_{\rm ISCO}$, c-modes
495: are not affected by the poloidal magnetic field when $b\lesssim 0.4$, but can be appreciably modified
496: when $b \gtrsim 0.4$.
497:
498: \begin{inlinefigure}
499: \scalebox{1}{\rotatebox{0}{\plotone{f4.ps}}} \figcaption{The
500: effect of poloidal magnetic field on c-modes with $m=1$ and
501: $a=0.2\rm{M}$. The upper panel shows the original c-modes since
502: $b$ is small and $\omega_1=\Omega_\perp$; in the bottom panel,
503: with a large $b$, the c-mode trapping zone is instead bounded by
504: the inner reflection boudary and $\Omega-\omega_2$. The vertical
505: dotted line refers to the inner disk radius at ISCO. The curvy
506: horizontal lines specify wave propagation zones and the height of
507: the line is $\omega$ of the mode.}
508: \end{inlinefigure}
509:
510: \begin{inlinefigure}
511: \scalebox{1}{\rotatebox{0}{\plotone{f5.ps}}} \figcaption{Wave
512: propagation diagram for non-axisymmetric g-modes and c-modes, for
513: $b=v_{Az}/c_s=0.7$, $m=2$, $\eta =1$, and $a=0$. The solid line,
514: dot-short dashed lines, long-dashed lines, dotted lines and
515: short-dashed lines show $\Omega$, $\Omega \pm \omega_4/2$, $\Omega
516: \pm \omega_3/2$, $\Omega \pm \omega_1/2$ and $\Omega \pm
517: \omega_2/2$ (eqs.~[31]-[34]), respectively. The vertical dotted
518: line shows the inner disc radius. The curvy horizontal lines
519: specify wave propagation zones and the height of the line is
520: $\omega/2$ of the mode. The singular points (where $k_r
521: \rightarrow \infty$) are indicated by filled circles. The angular
522: frequencies are in units of $M^{-1}=c^3/(GM)$.}
523: \end{inlinefigure}
524:
525: From eq.~(30) we can identify other wave propagation zones (see
526: eq.~[38]). Fig.~5 gives an example, for $m=2$, $b=0.7$. Note that,
527: except for the c-mode trapping zone discussed above, all the
528: propagation zones are bounded by at least one ``singular point''
529: (where $k_r \rightarrow \infty$). Unlike the turning point ($k_r
530: \rightarrow 0$) associated with wave reflection, wave absorption
531: is expected to occur at these singular points (see Zhang \& Lai
532: 2006; Tsang \& Lai 2008a and references therein). Thus, the new
533: wave trapping regions given by eq.~(38) will not lead to
534: interesting global oscillation modes (Note that in the case of
535: $b=0.7$, the ordering of five critical frequencies is different
536: from the one described in \S4.2. However, our conclusion still
537: holds true, i.e., there is no chance to form a wave trapping zone
538: bounded by two reflection points other than the c-mode oscillation
539: region, which is bounded by a reflection point and the ISCO).
540:
541: \section{Effect of Toroidal Fields}
542: In this section, we consider the effect of a pure toroidal field,
543: with $B_z=0$. Various instabilities may exist for such field
544: geometry, depending upon the rotation profile $\Omega(r)$ and the
545: magnetic field profile $B_{\phi}(r)$ (e.g., Acheson \& Gibbons 1978; Terquem \&
546: Papaloizou 1996). Here we focus on how $B_{\phi}$ affects the
547: diskoseismic modes.
548:
549: \subsection{P-modes}
550: With $k_z =0$, equations~(16)-(22) reduces to
551: \begin{equation}
552: \tomega^2=\kappa^2+k_r^2(c_s^2+v_{A\phi}^2),
553: \end{equation}
554: where $v_{A\phi} \equiv B_{\phi}/\sqrt{4\pi\rho}$.
555: Thus, the toroidal field affects p-modes in the same way as the poloidal field does (see \S4.1).
556:
557: \subsection{G-modes}
558: Since the general dispersion relation for $m \neq 0$ is quite complicated, here we focus on
559: axisymmetric perturbations.\footnote{Since c-modes necessarily require $m>0$, our analysis here cannot be
560: applied to c-modes.} With $m=0$, Equations~(16)-(22) lead to
561: \ba
562: &&\omega^4-[\kappa^2+(k_z^2+k_r^2)(c_s^2+v_{A\phi}^2)]\omega^2+\kappa^2k_z^2(c_s^2+v_{A\phi}^2)\nonumber\\
563: &&+2(1-q)v_{A\phi}^2 c_s^2 k_z^2/r^2=0,
564: \ea
565: where $q=d\ln B_{\phi}/d\ln r$. Solving for $k_r^2$, we have
566: \begin{equation}
567: k_r^2=\frac{(\omega^2-\omega_+^2)(\omega^2-\omega_-^2)}{(c_s^2+v_{A\phi}^2)\omega^2},
568: \end{equation}
569: with the two critical frequencies given by
570: \ba
571: &&\omega_\pm^2=\frac{\kappa^2+\eta \Omega_\perp^2(1+b_{\phi}^2)}{2}\pm\nonumber\\
572: &&\frac{1}{2} \sqrt{[\kappa^2-\eta \Omega_\perp^2(1+b_{\phi}^2)]^2
573: -8(1-q)\eta v_{A\phi}^2 \Omega_\perp^2/r^2},
574: \ea
575: where $b_{\phi} \equiv v_{A\phi}/c_s$ and we have used $k_z =\sqrt{\eta} /H=\sqrt{\eta} \Omega_\perp/c_s$
576: as in \S4. Clearly, for $b_{\phi}=0$, eq.~(42) reduces to eq.~(24).
577:
578: When $q=1$ (i.e., $B_{\phi} \propto r$), eq.~(42) gives
579: $\omega_+^2=\eta(\Omega_\perp)^2 (1+b_{\phi}^2),$ and $\omega_-^2=\kappa^2.$
580: Since $\omega_-^2$ is independent of $B_{\phi}$, the g-mode propagation zone is unaffected no matter
581: how strong the field is. When $q \neq 1$, as long as $v_{A\phi} \ll \Omega_\perp r$ (which is
582: valid in most disk situations), the $8(1-q)\eta v_{A\phi}^2\Omega_\perp^2/r^2$ term in eq.~(42) represents only
583: a small correction, i.e., $\omega_-^2$ is still very close to $\kappa^2$. Thus for
584: general toroidal field satisfying $v_{A\phi} \ll \Omega_\perp r$, the axisymmetric
585: g-mode propagation zone is not affected by the magnetic field.
586:
587: \section{Summary and Discussion}
588:
589: In this paper we have studied the effects of both poloidal and
590: toroidal magnetic fields on the diskoseismic modes in BH accretion
591: disks. Previous works by Kato, Wagoner and others have been based on
592: hydrodynamic disks with no magnetic field. The key finding of our
593: paper is that the g-mode self-trapping zone (which arises from GR
594: effect) disappears when the disk contains even a small poloidal
595: magnetic field, corresponding to $v_{Az}/c_s =0.01-0.1$ (see Fig.~2-3;
596: $v_{Az}$ is the Alfv\'{e}n speed and $c_s$ is the sound speed). It is
597: well-known that the combination of a weak poloridal field and
598: differential rotation gives rise to MRI, making real astrophysical
599: disks turbulent.
600: Earlier numerical simulations indicated that the magnetic
601: field grows as MRI develops, until it saturates at $v_{Az}/c_s \sim 0.1-1$,
602: %(i.e., the magnetic energy reaches equal-partition with the thermal
603: %energy;
604: with the toroidal field stronger than the poloidal field by a
605: factor of a few (see, e.g., Hawley et al. 1996; Balbus \& Hawley
606: 1998). Recent simulations showed that the turbulent state depends
607: strongly on the net magnetic flux through the disk (e.g., Fromang
608: \& Papaloizou 2007; Simon, Hawley \& Beckwith 2008). In any case,
609: it is likely that the magnetic field in a turbulent disk is large
610: enough to ``destroy'' the g-mode self-trapping zone.
611:
612: Thus, the g-mode properties (including the frequencies and
613: excitations) derived from hydrodynamical models are unlikely to be
614: applicable to real BH accretion disks. The disappearance of the g-mode
615: trapping zone might also explain why Arras et al. (2006) and Reynolds
616: \& Miller (2008) did not see any global g-modes in their MHD
617: simulations.
618:
619: As mentioned in \S1, g-mode oscillations have been considered a
620: promising candidate to explain QPOs in BH X-ray
621: binaries. Theoretically, these modes are appealing because in
622: hydrodynamic disks their existence depends on general relativistic
623: effect and does not require special disk boundary conditions. Our
624: analytical results presented in this paper, together with recent
625: numerical simulations (Arras et al. 2006; Reynolds \& Miller 2008),
626: suggest that magnetic fields and turbulence associated with real
627: accretion disks can change this picture significantly.
628:
629: While g-modes can be easily modified or ``destroyed'' by magnetic
630: fields, our analysis showed that p-modes are not affected
631: qualitatively. The magnetic field simply changes the sound speed to
632: the fast magnetosonic wave speed and leaves the p-mode propagation
633: diagram unchanged. We also showed that a weak poloidal field
634: ($v_{Az}/c_s \ll 1$) does not affect the c-mode propagation zone,
635: although a stronger field modifies it. Our results therefore suggests
636: that global p-mode oscillation is robust and may exist in real BH
637: accretion disks, provided that partial wave reflection at the disk
638: inner edge can be achieved.\footnote{Kato (2001) has discussed why
639: such reflection may be possible.} Of particular interest is the
640: non-axisymmetric p-modes, since they may be excited by instabilites
641: associated with corotation resonance (Tsang \& Lai 2008a, b).
642:
643: \begin{acknowledgements}
644: We thank David Tsang for useful discussion. This work has been supported in part by NASA Grant
645: NNX07AG81G and by NSF grant AST 0707628.
646: \end{acknowledgements}
647:
648: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
649: \begin{thebibliography}{}
650: \bibitem[]{}
651: Acheson, D. J. \& Gibbsons, M. P., 1978, Phil. Trans. Roy. Soc. Lond. A 289, 459
652:
653: \bibitem[]{}
654: Arras, P., Blaes, O. \& Turner, N. J., 2006, \apj, 645, L65
655:
656: \bibitem[]{}
657: Balbus, S. A. \& Hawley, J. F., 1991, \apj, 376, 214
658:
659: \bibitem[]{}
660: Balbus, S. A. \& Hawley, J. F., 1998, Rev. Mod. Phys., 70, 1
661:
662: \bibitem[]{}
663: Ferreira, B. T. \& Ogilvie, G. I., 2008, arXiv:0803.1671
664:
665: \bibitem[]{}
666: Fromang, S. \& Papaloizou, J., 2007, A\&A, 476, 1113
667:
668: \bibitem[]{}
669: Goodman, J., 1993, \apj, 406, 596
670:
671: \bibitem[]{}
672: Hawley, J. F., Gammie, C. F. \& Balbus, S. A., 1996, \aj, 464, 690
673:
674: \bibitem[]{}
675: Ipser, J. R., 1994, \apj, 435, 767
676:
677: \bibitem[]{}
678: Kato, S. \& Fukue, J., 1980, PASJ, 32, 377
679:
680: \bibitem[]{}
681: Kato, S., 1990, PASJ, 42, 99
682:
683: \bibitem[]{}
684: Kato, S., 2001, PASJ, 53, 1
685:
686: \bibitem[]{}
687: Kato, S., 2003, PASJ, 55, 257
688:
689: \bibitem[]{}
690: Kato, S., 2008, PASJ, 60, 111
691:
692: \bibitem[]{}
693: Kluzniak, W. \& Abramowicz, M. A., 2002, astro-ph/0203314
694:
695: \bibitem[]{}
696: Lee, W. H., Abramowicz, M. A. \& Kluziniak, W., 2004, \apj, 603, L93
697:
698: \bibitem[]{}
699: Li, L. X., Goodman, J. \& Naryan, R., 2003, \apj, 593, 980
700:
701: \bibitem[]{}
702: Nowak, M. A. \& Wagoner, R. V., 1991, \apj, 378, 656
703:
704: \bibitem[]{}
705: Nowak, M. A. \& Wagoner, R. V., 1992, \apj, 393, 697
706:
707: \bibitem[]{}
708: Okazaki, A. T., Kato, S. \& Fukue, J., 1987, PASJ, 39, 457
709:
710: \bibitem[]{}
711: Ortega-Rodriguez, M., Silbergleit, A. S. \& Wagoner, R. V., 2006, astro-ph/0611101
712:
713: \bibitem[]{}
714: Rebusco, P., 2008, arXiv:0801.3658
715:
716: \bibitem[]{}
717: Remillard, R. A. \& McClintock, J. E., 2006, ARAA, Vol. 44, Issue 1, pp. 49-92
718:
719: \bibitem[]{}
720: Reynolds, C. S. \& Miller, M. C., 2008, arXiv:0805.2950
721:
722: \bibitem[]{}
723: Rezzolla, L., Yoshida, S'i., Maccarone \& Zanotti, O., 2003, \mnras, 344, L37-L41
724:
725: \bibitem[]{}
726: Silbergleit, A. S., Wagoner, R. V. \& Ortega-Rodriguez, M., 2001, \apj, 548, 335
727:
728: \bibitem[]{}
729: Silbergleit, A. S. \& Wagoner, R. V., 2007, arXiv:0711.4848
730:
731: \bibitem[]{}
732: Simon, J. B., Hawley, J. F. \& Beckwith, K., 2008, arXiv:0806.3969
733:
734: \bibitem[]{}
735: Terquem, C. \& Papaloizou, J. C. B., 1996, \mnras, 279, 767
736:
737: \bibitem[]{}
738: Tsang, D. \& Lai, D., 2008a, \mnras, 387, 446
739:
740: \bibitem[]{}
741: Tsang, D. \& Lai, D., 2008b, \mnras, to be submitted
742:
743: \bibitem[]{}
744: Wagoner, R. V., 1999, Phys. Rep., 311, 259
745:
746: \bibitem[]{}
747: Zhang, H. \& Lai, D., 2006, \mnras, 368, 917
748: \end{thebibliography}
749:
750:
751: \end{document}
752: