1: %\documentclass[draft]{aipproc}

2: \documentclass[final]{aipproc}

3: \usepackage{amsmath,amssymb}

4: \usepackage{epsf}

5: 

6: \layoutstyle{6x9}

7: 

8: \newcommand\f{\frac}

9: \newcommand\p{\partial}

10: 

11: \begin{document}

12: 

13: \title{Warped and eccentric discs around black holes}

14: 

15: \classification{97.10.Gz, 97.60.Lf, 95.30.Lz}

16: 

17: \keywords{accretion discs -- black holes -- hydrodynamics}

18: 

19: \author{Gordon I. Ogilvie}{}

20: \author{B\'arbara T. Ferreira}

21: {address={Department of Applied Mathematics and Theoretical Physics,

22: University of Cambridge,\\

23: Centre for Mathematical Sciences, Wilberforce Road, Cambridge CB3 0WA, UK}}

24: 

25: \begin{abstract}

26: Accretion discs around black holes in X-ray binary stars are warped if

27: the spin axis of the black hole is not perpendicular to the binary

28: orbital plane.  They can also become eccentric through an instability

29: involving a resonance with the binary orbit.  Depending on the

30: thickness of the disc and the efficiency of dissipative processes,

31: these global deformations may be able to propagate into the innermost

32: part of the disc in the form of stationary bending or density waves.

33: We describe the solutions in the linear regime and discuss the

34: conditions under which a warp or eccentricity is likely to produce

35: significant activity in the inner region, which may include the

36: excitation of quasi-periodic oscillations.

37: \end{abstract}

38: 

39: \maketitle

40: 

41: \section{Introduction}

42: 

43: Accretion discs involve approximately Keplerian motion around a

44: massive central object, and the general solution allows for nested

45: orbits with smoothly varying inclination and eccentricity.  The shape

46: of a warped or eccentric disc evolves slowly under the action of

47: stresses in the disc and external forces that deviate from that of a

48: Newtonian point mass.  Coherent precession of discs is usually

49: possible in binary stars.

50: 

51: Warped and eccentric discs may be relevant to a wide variety of

52: phenomena in X-ray binaries.  Long-period modulations such as

53: superhumps and superorbital variability may be attributable to the

54: precession of global warping or eccentric modes that are not forced but

55: are excited by various instabilities.  Sufficiently large discs

56: encounter a resonance at the location where the angular velocity of

57: the disc is three times that of the binary orbit and may become

58: eccentric as a result \citep{W88,L91}.  This effect is well known in

59: cataclysmic variable stars of mass ratio $q\lesssim0.3$

60: \citep[e.g.][]{P05} where, as well as giving rise to superhumps, the

61: eccentricity enhances the viscous dissipation and significantly

62: affects the outburst dynamics.  Similar processes should occur in most

63: X-ray binaries with black hole primaries, and indeed superhumps are

64: reported in an increasing number of such systems

65: \citep{OC96,HKMC01,U02,NBC07,Z08}.  A warping instability involving

66: radiation forces \citep{P96} is more likely to occur with neutron star

67: primaries \citep{OD01} but may also be seen in GRS~1915+105, which has

68: a very large disc \citep{RGM03}.  Stationary warped discs also arise

69: whenever there is a misalignment between the orbital angular momentum

70: of the binary and the spin angular momentum of the central object

71: \citep{M02}, a knowledge of which is communicated to the disc through

72: general relativistic (gravitomagnetic) or magnetic torques

73: \citep{BP75,L99}.

74: 

75: The innermost parts of discs around black holes and neutron stars

76: depart significantly from Keplerian motion.  Indeed, the rapid

77: relativistic precession of elliptical or inclined orbits has often

78: been discussed in connection with quasi-periodic oscillations (QPOs)

79: in X-ray binaries, and may best be studied within the context of

80: eccentric or warped discs.  There is an apparent conflict between the

81: description of the outer part of the disc, which supports slowly

82: precessing global deformations, and that of the non-Keplerian inner

83: region.  In this paper we attempt to make this connection and to

84: describe how a global deformation of the disc in the form of a

85: stationary or slowly precessing warp or eccentricity may be able to

86: propagate inwards under some conditions, activating the inner region

87: and possibly exciting trapped oscillations that may explain

88: high-frequency QPOs in accreting black holes.  The excitation

89: mechanism itself has been studied by \citet{K03,K08} and by \citet[see

90: also this volume]{FO08}.

91: 

92: \section{Local analysis}

93: 

94: A small warp or eccentricity can be considered as a perturbation of a

95: standard (circular and coplanar) disc.  Such a disc supports a variety

96: of wave modes, having a dependence on time and azimuth of the form

97: $\exp(im\phi-i\omega t)$.  In the simplest case of a strictly

98: isothermal disc, a local dispersion relation

99: \begin{equation}

100:   k^2H^2=\f{(\hat\omega^2-\kappa^2)(\hat\omega^2-n\Omega_z^2)}{\hat\omega^2\Omega_z^2}

101: \end{equation}

102: can be derived \citep{OKF87}, which relates the Doppler-shifted wave

103: frequency $\hat\omega=\omega-m\Omega$ to the radial wavenumber $k$.

104: The integers $m$ and $n\ge0$ are the azimuthal and vertical mode

105: numbers, while $H$ is the vertical scaleheight of the disc and

106: $\Omega$, $\kappa$ and $\Omega_z$ are the orbital, epicyclic and

107: vertical oscillation frequencies characteristic of circular orbits in

108: the given potential or metric.  All of these quantities depend on the

109: radius $R$ at which the dispersion relation is evaluated.  The local

110: dispersion relation of a more general disc model can be calculated

111: numerically.  For this description to be accurate, the wavelength

112: $\lambda=2\pi/k$ should be much less than $R$.

113: 

114: Within this context, a warp corresponds to $(m,n)=(1,1)$ (vertical

115: motion independent of $z$) and an eccentricity to $(m,n)=(1,0)$

116: (horizontal motion independent of $z$).  The dispersion relation shows

117: that the warp takes the form of a propagating bending wave ($k^2>0$)

118: when $(\omega-\Omega)^2>\max(\kappa^2,\Omega_z^2)$ or

119: $<\min(\kappa^2,\Omega_z^2)$, while the eccentricity takes the form of

120: a propagating density wave when $(\omega-\Omega)^2>\kappa^2$.

121: 

122: \section{Secular theories}

123: 

124: A complementary description is provided by theories that consider a

125: warp or eccentricity that varies on a length-scale much longer than

126: $H$ and on a time-scale much longer than $\Omega^{-1}$.

127: Small-amplitude warps are governed by the equations

128: \begin{equation}

129:   \Sigma R^2\Omega\left[\f{\p W}{\p t}-i\left(\f{\Omega^2-\Omega_z^2}{2\Omega}\right)W\right]=\f{1}{R}\f{\p G}{\p R},

130: \end{equation}

131: \begin{equation}

132:   \f{\p G}{\p t}-i\left(\f{\Omega^2-\kappa^2}{2\Omega}\right)G+\alpha_W\Omega G=\f{PR^3\Omega}{4}\f{\p W}{\p R}

133: \end{equation}

134: \citep[e.g.][]{LOP02}, where $W$ describes the amplitude and phase of

135: the inclination of the disc at radius $R$ and time $t$, $G$ refers to

136: a horizontal torque communicated by Reynolds stresses, $\Sigma$ is the

137: surface density and $P=\Sigma H^2\Omega_z^2$ is the vertically

138: integrated pressure.  These equations describe propagating bending

139: waves with essentially the same local dispersion relation as in the

140: previous section.  (The theories overlap when $H\ll\lambda\ll R$,

141: which is possible in a thin disc.)  They also allow for viscous (i.e.\

142: turbulent) damping of the warp, parametrized using a dimensionless

143: number $\alpha_W$, which is equivalent to the usual Shakura--Sunyaev

144: parameter \citep{SS73} if the disc has an isotropic effective

145: viscosity.

146: 

147: The simplest equation describing a small eccentricity is

148: \begin{equation}

149:   -2i\Sigma R^2\Omega\f{\p E}{\p t}=\f{1}{R}\f{\p}{\p R}\left[(\gamma-i\alpha_E)PR^3\f{\p E}{\p R}\right]+R\f{\p P}{\p R}E+\Sigma R^2(\Omega^2-\kappa^2)E

150: \end{equation}

151: \citep[e.g.][]{GO06}, where $E$ describes the amplitude and phase of

152: the eccentricity of the disc at radius $R$ and time $t$, $\gamma$ is

153: the adiabatic exponent and $\alpha_E$ parametrizes the viscous

154: damping.  For short wavelengths, this equation also agrees with the

155: local dispersion relation of the isothermal disc when $\gamma=1$ and

156: $\alpha_E=0$.  This equation is based on a two-dimensional

157: approximation and neglects many of the complications of a shear

158: viscosity such as viscous overstability \citep{K78}.  More

159: sophisticated theories, including nonlinearity and all viscous or

160: viscoelastic effects, are available in the literature \citep{O00,O01}.

161: 

162: \section{Stationary deformations}

163: 

164: \begin{figure}

165: \label{f:lambda}

166: \centerline{\epsfysize=7.5cm\epsfbox{fig1a.eps}\epsfysize=7.5cm\epsfbox{fig1b.eps}}

167: \caption{Local wavelength of a stationary warp (\textit{left}) or

168: eccentricity (\textit{right}), in units of the local vertical

169: scaleheight $H$, plotted versus the radius in gravitational units, for

170: black holes with spin parameters $a=0.1$, $0.2$, $0.4$ and $0.8$.  The

171: disc is terminated at the marginally stable orbit in each case.}

172: \end{figure}

173: 

174: A global warp or eccentricity precesses at only a fraction of the

175: binary orbital frequency, so the wave frequency $\omega$ is completely

176: negligible compared to $\Omega$, $\kappa$ and $\Omega_z$ in the inner

177: part of the disc.  Setting $\omega=0$, we find the local radial

178: wavenumber to be given by

179: $k^2H^2=(\Omega^2-\kappa^2)(\Omega^2-\Omega_z^2)/\Omega^2\Omega_z^2$

180: in the case of a warp and

181: $k^2H^2=(\Omega^2-\kappa^2)/\gamma\Omega_z^2$ for an eccentricity.

182: Using the expressions for $\Omega$, $\kappa$ and $\Omega_z$ for the

183: Kerr metric \citep{K90} and taking $\gamma=5/3$, we find $\lambda/H$

184: as a function of the dimensionless radius $r=Rc^2/GM$ and the spin

185: parameter $a$ (see Fig.~\ref{f:lambda}).  For $a>0$, both $W$ and $E$

186: propagate at all radii, with wavelengths everywhere significantly

187: longer than $H$.  Of all the wave modes described by the local

188: dispersion relation, these are the most credible in a turbulent disc

189: because of their relatively long wavelengths.

190: 

191: It is also possible to predict how the amplitudes of the deformations

192: scale with radius.  In the absence of viscous damping, a WKB analysis

193: of the secular theories shows that $|W|\propto R^{1/8}(\Sigma

194: H)^{-1/2}$ and $|E|\propto R^{1/4}(\Sigma H)^{-1/2}$.  We apply these

195: results to a steady accretion disc in the regime dominated by gas

196: pressure and Thomson opacity, in which (assuming

197: $\alpha=\mathrm{constant}$) $\Sigma\propto f^{3/5}R^{-3/5}$ and

198: $H\propto f^{1/5}R^{21/20}$, where $f=1-(R_{in}/R)^{1/2}$

199: \citep{SS73}.  Then $|W|\propto f^{-2/5}R^{-1/10}$ and $|E|\propto

200: f^{-2/5}R^{1/40}$, implying a very mild dependence of the amplitude on

201: radius.  However, the gradients $dW/dR$ and $dE/dR$ do increase

202: sharply at small $R$, because of the rapidly decreasing wavelength.

203: 

204: \begin{figure}

205: \label{f:w}

206: \centerline{\epsfysize=8.5cm\epsfbox{fig2.eps}}

207: \caption{Stationary warp in a disc around a black hole with $a=0.5$,

208: $\epsilon=0.02$ and with $\alpha_W=0$ (\textit{top}) and $0.05$ (\textit{bottom}).  Real and imaginary parts of $W$ are

209: plotted as solid and dotted lines.  The amplitude is scaled such that $W\to1$ at large $r$.}

210: \end{figure}

211: 

212: \begin{figure}

213: \label{f:e}

214: \centerline{\epsfysize=8.5cm\epsfbox{fig3.eps}}

215: \caption{Stationary eccentricity in a disc around a black hole with $a=0.5$,

216: $\epsilon=0.02$ and with $\alpha_E=0$ (\textit{top}) and $0.05$ (\textit{bottom}).  Real and imaginary parts of $E$ are

217: plotted as solid and dotted lines.}

218: \end{figure}

219: 

220: When dissipation is taken into account, these solutions are modified

221: by viscous attenuation.  Our interest is in whether a deformation of

222: the outer part of the disc can propagate into the inner region with

223: non-negligible amplitude.  Noting that $H/R$ is almost independent of

224: $R$ in the above model, we define the constant parameter $\epsilon$ by

225: $H/R=\epsilon f^{1/5}r^{1/20}$.  Then the logarithm of the attenuation

226: factor for a complete crossing of the disc scales approximately with

227: $\alpha_W/\epsilon$ or $\alpha_E/\epsilon$.

228: 

229: The numerical solutions in Fig.~\ref{f:w} confirm this behaviour.  In

230: the absence of dissipation, the bending wave reflects perfectly from

231: the stress-free inner boundary and sets up a standing wave.

232: Significant attenuation is found when $\alpha_W$ exceeds $\epsilon$,

233: in which case only a wave with inward group velocity is seen.  If

234: $\alpha_W$ is increased much further, the oscillations are no longer

235: apparent.  Only in this case does the inner part of the disc lie in

236: the equatorial plane of the black hole as suggested by \citet{BP75}.

237: Note that $W$ tends to a constant at large $r$, which corresponds to

238: the inclination of the outer part of the disc with respect to the

239: black hole's equator.  The oscillatory structure of the warp has been

240: noted before \citep{II97,LOP02}.  Traces of it may already have been

241: detected in numerical simulations \citep{FBAS07}.

242: 

243: The behaviour of the eccentricity is similar (Fig.~\ref{f:e}) except

244: for the shorter wavelength (still everywhere much longer than $H$).

245: Again, unless $\alpha_E$ is several times greater than $\epsilon$, the

246: eccentricity can reach the inner region.

247: 

248: Several caveats accompany these solutions.  Radiation pressure, which

249: is more important at higher accretion rates, thickens the inner part

250: of the disc, increases the wavelength, and reduces the attenuation.

251: Viscous overstability may cause the eccentricity to grow, rather than

252: decay, as it propagates inwards.  Nonlinearity may be very important.

253: The relevant damping coefficients $\alpha_W$ and $\alpha_E$ are not

254: generally equal to $\alpha$ and remain relatively poorly understood.

255: In addition, the time for the warp or eccentricity to propagate into

256: the inner region and to establish the steady solutions shown here can

257: be long.

258: 

259: \section{Conclusions}

260: 

261: Accretion discs around black holes in X-ray binary stars may commonly

262: be warped or eccentric.  Depending on the thickness of the disc and

263: the efficiency of dissipative processes, these global deformations may

264: be able to propagate into the innermost part of the disc in the form

265: of stationary bending or density waves.  This is most likely to occur

266: when the disc is hotter and thicker, when the wavelengths are longer

267: and the viscous attenuation is less severe.  Under these conditions

268: the inner region may be activated and trapped oscillations may be

269: excited through nonlinear mode couplings \citep{K03,K08,FO08}.

270: 

271: \begin{theacknowledgments}

272: GIO acknowledges the support of STFC.  The work of BTF was supported

273: by FCT (Portugal) through grant no. SFRH/BD/22251/2005.

274: \end{theacknowledgments}

275: 

276: \begin{thebibliography}{99}

277: 

278: \bibitem[Whitehurst (1988)]{W88}

279: R.~Whitehurst, \emph{Mon. Not. R. Astron. Soc.} \textbf{232}, 35--51 (1988).

280: 

281: \bibitem[Lubow (1991)]{L91}

282: S.~H. Lubow, \emph{Astrophys. J.} \textbf{381}, 259--267 (1991).

283: 

284: \bibitem[Patterson et al. (2005)]{P05}

285: J.~Patterson et al., \emph{Publ. Astron. Soc. Pacific} \textbf{117},

286: 1204--1222 (2005).

287: 

288: \bibitem[O'Donoghue and Charles (1996)]{OC96}

289: D.~O'Donoghue and P.~A. Charles, \emph{Mon. Not. R. Astron. Soc.}

290: \textbf{282}, 191--205 (1996).

291: 

292: \bibitem[Haswell et al. (2001)]{HKMC01}

293: C.~A. Haswell, A.~R. King, J.~R. Murray and P.~A. Charles,

294: \emph{Mon. Not. R. Astron. Soc.} \textbf{321}, 475--480 (2001).

295: 

296: \bibitem[Uemura et al. (2002)]{U02}

297: M.~Uemura et al., \emph{Publ. Astron. Soc. Japan} \textbf{54}, 285--298 (2002).

298: 

299: \bibitem[Neil et al. (2007)]{NBC07} E.~T. Neil, C.~D. Bailyn and

300: B.~E. Cobb, \emph{Astrophys. J.} \textbf{657}, 409--414 (2007).

301: 

302: \bibitem[Zurita et al. (2008)]{Z08}

303: C.~Zurita et al., \emph{Astrophys. J.} \textbf{681}, 1458--1463 (2008).

304: 

305: \bibitem[Pringle (1996)]{P96}

306: J.~E. Pringle, \emph{Mon. Not. R. Astron. Soc.} \textbf{281}, 357--361 (1996).

307: 

308: \bibitem[Ogilvie and Dubus (2001)]{OD01}

309: G.~I. Ogilvie and G.~Dubus, \emph{Mon. Not. R. Astron. Soc.}

310: \textbf{320}, 485--503 (2001).

311: 

312: \bibitem[Rau et al. (2003)]{RGM03}

313: A.~Rau, J.~Greiner and M.~L. McCollough, \emph{Astrophys. J.}

314: \textbf{590}, L37--L40 (2003).

315: 

316: \bibitem[Maccarone (2002)]{M02}

317: T.~J. Maccarone, \emph{Mon. Not. R. Astron. Soc.} \textbf{336},

318: 1371--1376 (2002).

319: 

320: \bibitem[Bardeen and Petterson (1975)]{BP75}

321: J.~M. Bardeen and J.~A. Petterson, \emph{Astrophys. J.} \textbf{195},

322: L65--L67 (1975).

323: 

324: \bibitem[Lai (1999)]{L99}

325: D.~Lai, \emph{Astrophys. J.} \textbf{524}, 1030--1047 (1999).

326: 

327: \bibitem[Kato (2003)]{K03}

328: S.~Kato, \emph{Publ. Astron. Soc. Japan} \textbf{55}, 801--817 (2003).

329: 

330: \bibitem[Kato (2008)]{K08}

331: S.~Kato, \emph{Publ. Astron. Soc. Japan} \textbf{60}, 111--123 (2008).

332: 

333: \bibitem[Ferreira and Ogilvie (2008)]{FO08}

334: B.~T. Ferreira and G.~I. Ogilvie, \emph{Mon. Not. R. Astron. Soc.}

335: \textbf{386}, 2297--2310 (2008).

336: 

337: \bibitem[Okazaki et al. (1987)]{OKF87}

338: A.~T. Okazaki, S.~Kato and J.~Fukue, \emph{Publ. Astron. Soc. Japan}

339: \textbf{39}, 457--473 (1987).

340: 

341: \bibitem[Lubow et al. (2002)]{LOP02}

342: S.~H. Lubow, G.~I. Ogilvie and J.~E. Pringle,

343: \emph{Mon. Not. R. Astron. Soc.} \textbf{337}, 706--712 (2002).

344: 

345: \bibitem[Shakura and Sunyaev (1973)]{SS73}

346: N.~I. Shakura and R.~A. Sunyaev, \emph{Astron. Astrophys.}

347: \textbf{24}, 337--355 (1973).

348: 

349: \bibitem[Goodchild and Ogilvie (2006)]{GO06}

350: S.~Goodchild and G.~I. Ogilvie, \emph{Mon. Not. R. Astron. Soc.}

351: \textbf{368}, 1123--1131 (2006).

352: 

353: \bibitem[Kato (1978)]{K78}

354: S.~Kato, \emph{Mon. Not. R. Astron. Soc.} \textbf{185}, 629--642 (1978).

355: 

356: \bibitem[Ogilvie (2000)]{O00}

357: G.~I. Ogilvie, \emph{Mon. Not. R. Astron. Soc.} \textbf{317}, 607--622 (2000).

358: 

359: \bibitem[Ogilvie (2001)]{O01}

360: G.~I. Ogilvie, \emph{Mon. Not. R. Astron. Soc.} \textbf{325}, 231--248 (2001).

361: 

362: \bibitem[Kato (1990)]{K90}

363: S.~Kato, \emph{Publ. Astron. Soc. Japan} \textbf{42}, 99--113 (1990).

364: 

365: \bibitem[Ivanov and Illarionov (1997)]{II97}

366: P.~B. Ivanov and A.~F. Illarionov, \emph{Mon. Not. R. Astron. Soc.}

367: \textbf{285}, 394--402 (1997).

368: 

369: \bibitem[Fragile et al. (2007)]{FBAS07}

370: P.~C. Fragile, O.~M. Blaes, P.~Anninos and J.~D. Salmonson,

371: \emph{Astrophys. J.} \textbf{668}, 417--429 (2007).

372: 

373: \end{thebibliography}

374: 

375: \end{document}

376: