1: \documentclass{article}

2: \usepackage{graphicx}

3: %\usepackage{caption2}

4: \usepackage{amssymb}

5: \usepackage{times}

6: \usepackage{emulateapj}

7: \usepackage{psfig}

8: 

9: \def\AaA{{\em Astr.~Astrophys.}}

10: \def\AJ{{\em Astr.~J.}}

11: \def\AN{{\em Astron.~Nachr.}}

12: \def\ApJ{{\em Astrophys.~J.}}

13: \def\ApJL{{\em Astrophys.~J.~Lett.}}

14: \def\ApJS{{\em Astrophys.~J.~Suppl.}}

15: \def\ARAA{{\em Ann.~Rev.~Astr.~Astrophys.}}

16: \def\MN{{\em Mon.~Not.~R.~astr.~Soc.}}

17: \def\Nat{{\em Nature}}

18: \def\PASJ{{\em Publ.~astr.~Soc.~Japan}}

19: \def\PASP{{\em Publ.~astr.~Soc.~Pacif.}}

20: \def\PhD{{\em PhD thesis}}

21: \def\Pre{Preprint}

22: \def\PS{{\em Phys.~Scr.}}

23: \def\RMP{{\em Rev.~Mod.~Phys.}}

24: \def\SSR{{\em Space Sci.~Rev.}}

25: \def\etal{{et al.\thinspace}}

26: \def\cf{{\em cf.\ }}

27: \def\eg{{\em e.g.\ }}

28: \def\etc{{\em etc.\ }}

29: \def\ie{{\em i.e.\ }}

30: \def\spose#1{\hbox to 0pt{#1\hss}}

31: 

32: \let\approxlt=\lesssim

33: \let\approxgt=\gtrsim

34: \def\multcent#1{\halign {\centerline{##}\cr #1}}

35: \def\multleft#1{\hbox to size{\vbox {\halign {\lft{##}\cr #1}}\hfill}\par}

36: \def\multright#1{\hbox to size{\vbox {\halign {\rt{##}\cr #1}}\hfill}\par}

37: \def\Mdot{\hbox{$\dot M$}}

38: \def\degmark{^\circ}

39: \def\boxit#1{\vbox{\hrule\hbox{\vrule\kern3pt\vbox{\kern3pt

40:           #1 \kern3pt}\kern3pt\vrule}\hrule}}

41: \font\big=cmr10 scaled\magstep2

42: \font\bigbf=cmbx10 scaled\magstep2

43: \font\bigit=cmti10 scaled\magstep2

44: 

45: %       Simple units

46: 

47: \def\cm{{\rm\thinspace cm}}

48: \def\dyn{{\rm\thinspace dyn}}

49: \def\erg{{\rm\thinspace erg}}

50: \def\eV{{\rm\thinspace eV}}

51: \def\g{{\rm\thinspace g}}

52: \def\ga{{\rm\thinspace gauss}}

53: \def\K{{\rm\thinspace K}}

54: \def\keV{{\rm\thinspace keV}}

55: \def\km{{\rm\thinspace km}}

56: \def\kpc{{\rm\thinspace kpc}}

57: \def\Lsun{\hbox{$\rm\thinspace L_{\odot}$}}

58: \def\m{{\rm\thinspace m}}

59: \def\Mpc{{\rm\thinspace Mpc}}

60: \def\Msun{\hbox{$\rm\thinspace M_{\odot}$}}

61: \def\pc{{\rm\thinspace pc}}

62: \def\ph{{\rm\thinspace ph}}

63: \def\s{{\rm\thinspace s}}

64: \def\yr{{\rm\thinspace yr}}

65: \def\sr{{\rm\thinspace sr}}

66: \def\Hz{{\rm\thinspace Hz}}

67: \def\chisq{\hbox{$\chi^2$}}

68: \def\delchi{\hbox{$\Delta\chi$}}

69: 

70: %       Compound units

71: 

72: \def\cmps{\hbox{$\cm\s^{-1}\,$}}

73: \def\cmsq{\hbox{$\cm^2\,$}}

74: \def\cmcu{\hbox{$\cm^3\,$}}

75: \def\pcmcu{\hbox{$\cm^{-3}\,$}}

76: \def\pcmcuK{\hbox{$\cm^{-3}\K\,$}}

77: \def\dynpcmsq{\hbox{$\dyn\cm^{-2}\,$}}

78: \def\ergcmcups{\hbox{$\erg\cm^3\ps\,$}}

79: \def\ergpcmps{\hbox{$\erg\cm^{-3}\s^{-1}\,$}}

80: \def\ergpcmsqps{\hbox{$\erg\cm^{-2}\s^{-1}\,$}}

81: \def\ergpcmsqpspA{\hbox{$\erg\cm^{-2}\s^{-1}$\AA$^{-1}\,$}}

82: \def\ergps{\hbox{$\erg\s^{-1}\,$}}

83: \def\gpcm{\hbox{$\g\cm^{-3}\,$}}

84: \def\gpcmps{\hbox{$\g\cm^{-3}\s^{-1}\,$}}

85: \def\gps{\hbox{$\g\s^{-1}\,$}}

86: \def\kmps{\hbox{$\km\s^{-1}\,$}}

87: \def\Lsunppc{\hbox{$\Lsun\pc^{-3}\,$}}

88: \def\Msunpc{\hbox{$\Msun\pc^{-3}\,$}}

89: \def\Msunpkpc{\hbox{$\Msun\kpc^{-1}\,$}}

90: \def\Msunppc{\hbox{$\Msun\pc^{-3}\,$}}

91: \def\Msunppcpyr{\hbox{$\Msun\pc^{-3}\yr^{-1}\,$}}

92: \def\Msunpyr{\hbox{$\Msun\yr^{-1}\,$}}

93: \def\pcm{\hbox{$\cm^{-3}\,$}}

94: \def\pcmsq{\hbox{$\cm^{-2}\,$}}

95: \def\pcmK{\hbox{$\cm^{-3}\K$}}

96: \def\phpcmsqps{\hbox{$\ph\cm^{-2}\s^{-1}\,$}}

97: \def\pHz{\hbox{$\Hz^{-1}\,$}}

98: \def\pmpc{\hbox{$\Mpc^{-1}\,$}}

99: \def\pmpccu{\hbox{$\Mpc^{-3}\,$}}

100: \def\ps{\hbox{$\s^{-1}\,$}}

101: \def\psqcm{\hbox{$\cm^{-2}\,$}}

102: \def\psr{\hbox{$\sr^{-1}\,$}}

103: \def\pyr{\hbox{$\yr^{-1}\,$}}

104: \def\kmpspMpc{\hbox{$\kmps\Mpc^{-1}$}}

105: \def\Msunpyrpkpc{\hbox{$\Msunpyr\kpc^{-1}$}}

106: 

107: \begin{document}

108: 

109: \title{A variable efficiency for thin disk black hole accretion}

110: 

111: \author{Christopher~S.~Reynolds}

112: 

113: \affil{JILA, Campus Box 440, University of Colorado,

114: Boulder CO~80309\altaffilmark{1}}

115: 

116: \and

117:  

118: \author{Philip~J.~Armitage}

119: 

120: \affil{School of Physics and Astronomy, University of St.Andrews,

121: Fife, KY16~9SS, UK.}

122: 

123: \altaffiltext{1}{Present address: Dept. of Astronomy, University of Maryland, College Park, MD~20742}

124: 

125: 

126: 

127: \begin{abstract}

128: We explore the presence of torques at the inner edges of geometrically-thin

129: black hole accretion disks using 3-dimensional

130: magnetohydrodynamic (MHD) simulations in a pseudo-Newtonian potential.

131: By varying the saturation level of the magnetorotational instability 

132: that leads to angular momentum transport, we show that the dynamics 

133: of gas inside the radius of marginal stability varies depending upon 

134: the magnetic field strength just {\em outside} that radius. Weak 

135: fields are unable to causally connect material within the plunging region 

136: to the rest of the disk, and zero torque is an approximately correct 

137: boundary condition at the radius of marginal stability. Stronger fields, 

138: which we obtain artificially but which may occur physically within more 

139: complete disk models, are able to couple at least some parts of the 

140: plunging region to the rest of the disk. In this case, angular momentum 

141: (and implicitly energy) is extracted from the material in the plunging 

142: region. Furthermore, the magnetic coupling to the plunging region can 

143: be highly time dependent with large fluctuations in the torque at the 

144: radius of marginal stability.  This implies 

145: varying accretion efficiencies, both across systems and within a given system 

146: at different times. The results suggest a possible link between changes in 

147: X-ray and outflow activity, with both being driven by transitions between 

148: weak and strong field states.

149: \end{abstract}

150: 

151: \keywords{accretion, accretion disks -- black hole physics -- MHD --

152:  hydrodynamics -- instabilities}

153: 

154: 

155: \section{Introduction}

156: 

157: In a wide class of black hole systems, including QSOs, Seyfert galaxies,

158: and high-state Galactic Black Hole Candidates, the accretion disk is almost

159: certainly geometrically thin and radiatively efficient.  The radiation

160: emitted from an annulus of such a disk does not simply equal the change in

161: gravitational binding energy as gas flows across the annulus, but rather

162: includes a contribution from energy transported into the annulus from

163: elsewhere in the disk. As a consequence, the disk structure, along with

164: related quantities such as the radiative efficiency (the fraction of rest

165: mass energy which is radiated during the accretion process), depends upon

166: the location and nature of the boundary conditions at the inner disk edge.

167: 

168: For a black hole accretion disk, a natural location to place the inner

169: boundary condition is at the radius of marginal stability, $r=r_{\rm

170: ms}$. This is the radius inside of which circular orbits are no longer

171: stable, and is at $6GM/c^2$ for a non-rotating (Schwarzschild) black

172: hole. In the standard model for black hole accretion (Novikov \& Thorne

173: 1973; Paczynski \& Bisnovatyi-Kogan 1981; Abramowicz \& Kato 1989;

174: Paczynski 2000), one assumes that there is no angular momentum transport

175: across $r=r_{\rm ms}$. The motivation for this zero torque boundary

176: condition (ZTBC) is that the radial flow rapidly becomes supersonic once

177: inside $r=r_{\rm ms}$, and so the material loses causal contact with the

178: rest of the disk. Thereafter, material simply spirals ballistically into

179: the black hole. For this reason, we will refer to the region $r<r_{\rm ms}$

180: as the plunging region.

181: 

182: The ZTBC limits the accretion efficiency to around 6\% (for a non-rotating

183: black hole) -- already a substantial figure. However, it is possible that

184: even higher efficiencies may arise. The suggestion is that magnetic fields,

185: which are generated in the disk by the magnetorotational instability

186: (Balbus \& Hawley 1991) may become dynamically significant within the

187: plunging region. These fields could then causally connect the plunging

188: region to the rest of the accretion disk, allowing energy and angular

189: momentum to be extracted from gas as it executes its final spiral into the

190: black hole (Krolik 1999; Gammie 1999; Agol \& Krolik 2000). Subsequent

191: numerical simulations (none of which are General Relativistic) have

192: confirmed the presence of magnetic torques at the marginally stable orbit,

193: but have reached different conclusions as to their impact on the dynamics

194: of the flow.  Global MHD simulations of adiabatic accretion tori (Hawley

195: 2000; Hawley \& Krolik 2001) showed outward angular momentum transport

196: continuing within $r_{\rm ms}$, while unstratified simulations suggested

197: more modest effects (Armitage, Reynolds \& Chiang 2001; Hawley 2001). The

198: important parameters or numerical differences that cause this different

199: behavior remain to be determined.

200: 

201: In this {\it letter} we explore the suggestion (Charles Gammie, private

202: communication) that the dynamics of the flow within the

203: plunging region may depend upon the strength of magnetic fields 

204: in the disk at $r > r_{\rm ms}$. We extend our earlier work by 

205: presenting `boosted' simulations in which the strength of the magnetic 

206: field in the saturated turbulence is enhanced by starting with an initially 

207: large seed field. This is a numerical trick, though 

208: it may have a physical counterpart if the inner disk is 

209: threaded by a large-scale magnetic field. For the purposes of 

210: this letter, however, the aim is solely to allow us to 

211: study how the dynamics of the plunging region vary with field 

212: strength in a controlled fashion. We show that 

213: the dynamics of gas in the plunging region, and the validity 

214: of the ZTBC, do indeed depend upon the saturation field strength, 

215: and discuss the implications for observations of 

216: accreting black holes.

217: 

218: \section{The simulations}

219: 

220: \begin{table*}

221: \begin{center}

222: \begin{tabular}{ccccccc}

223: Run & $\phi_{\rm max}$ & $z_{\rm max} / r_{\rm Sch}$ & $n_r$ & $n_\phi$ & $n_z$ &

224: $\beta_{\rm i}$ \\\hline

225: 1  &  $\pi/4$ & $0.5$ & $200(40)$ & $60$ & $40$ & $5000(z)$ \\

226: 2  &  $\pi/4$ & $0.5$ & $200(40)$ & $60$ & $40$ & $500(z)$ \\

227: 3  &  $\pi/4$ & $0.5$ & $200(40)$ & $60$ & $40$ & $100(\phi)$ \\

228: 4  &  $\pi/2$ & $0.5$ & $200(40)$ & $120$ & $40$ & $5000(z)$ \\\hline

229: \end{tabular}

230: \end{center}

231: \caption{The set of simulations.  Columns show (1) Run number, (2)

232: $\phi$-domain, (3) $z$-domain, (4) number of radial cells (and number

233: inside $r=r_{\rm ms}$), (5) number of $\phi$ cells, (6) number of vertical

234: cells, and (7) initial $\beta$ of plasma (and direction).}

235: \end{table*}

236: 

237: We use the ZEUS code to solve the equations of ideal MHD (Stone \&

238: Norman 1992a, 1992b; Clarke, Norman \& Fielder 1994; Norman 2000) in

239: cylindrical co-ordinates, using a setup very similar to that described

240: previously (Armitage, Reynolds \& Chiang 2001). Stated briefly, the

241: computational domain is the wedge given by $r\in(r_{\rm in},r_{\rm

242: out})$, $\phi\in(0,\phi_{\rm max})$ and $z\in(-z_{\rm max},z_{\rm

243: max})$. The effects of General Relativity (in particular the existence

244: of an innermost stable orbit) are modelled in this inherently

245: non-relativistic code by using a pseudo-Newtonian potential (Paczynski

246: \& Wiita 1980), $\Phi=-\frac{GM}{r-r_{\rm Sch}}$, where $r_{\rm

247: Sch}=2GM/c^2$. The vertical component of gravity is neglected, so the

248: disk has no vertical structure. There is therefore no time-averaged

249: variation of density or magnetic field strength with $z$.  The

250: equation of state is isothermal, with a sound speed $c_{\rm s}$ which

251: is assumed uniform and constant. At $r_{\rm ms}$, the ratio of the

252: sound speed to the Keplerian velocity is $c_s / v_\phi = 0.065$,

253: slightly cooler than our previous simulations.

254: 

255: The initial conditions for density and velocity were obtained by 

256: relaxing a Gaussian radial density profile, using ZEUS in its 1D

257: hydrodynamic mode, to produce a density profile which is in

258: accurate numerical equilibrium.  The initial configurations of the 

259: 3D MHD simulations were then produced by adding a seed magnetic field.

260: Table~1 shows the computational domain, resolution, initial value of 

261: $\beta$, and seed field configuration for the simulations. As

262: usual, $\beta$ is defined as the ratio of thermal energy density to

263: magnetic energy density. In all cases, we set $r_{\rm in}=4GM/c^2$ and

264: $r_{\rm out}=20GM/c^2$. The radial boundary conditions were such as to

265: permit outflow, and all other boundaries were made periodic. These

266: simulations were then evolved until the magnetorotational instability (MRI)

267: became non-linear and produced fully developed MHD turbulence 

268: (in the sense that the MRI has saturated and there is no further 

269: long-term growth of the magnetic energy).  This turbulence drives 

270: accretion of material through the inner radial boundary, and the run is 

271: continued until an appreciable amount of material has left

272: the computational domain..

273: 

274: \section{The dynamics of the flow within the plunging region}

275: 

276: \begin{figure*}[t]

277: \hbox{

278: \psfig{figure=f1a.ps,width=0.48\textwidth}

279: \psfig{figure=f1b.ps,width=0.48\textwidth}

280: }

281: \caption{Time-averaged radial profiles of $\beta^{-1}$ (the ratio of

282: magnetic energy density to thermal energy density) and $l$ (specific

283: angular momentum) for our high resolution simulations.  The runs are

284: denoted by: Run~1 (solid line), Run~2 (dotted line), Run~3 (dashed line),

285: Run~4 (dot-dashed line).}

286: \end{figure*}

287: 

288: \begin{figure*}[t]

289: \hbox{

290: \psfig{figure=f2a.ps,width=0.48\textwidth}

291: \psfig{figure=f2b.ps,width=0.48\textwidth}

292: }

293: \caption{Radial profiles of $\beta^{-1}$ and $l$ for individual timeslices

294: from the boosted simulation, Run~2.  In particular, note the anomalous

295: event (shown with a thick line) during which the plunging region becomes

296: magnetically dominated and there is efficient angular momentum transport

297: out of this region.}

298: \end{figure*}

299: 

300: \subsection{Time averaged properties}

301: 

302: We consider first the time-averaged properties of our runs.

303: Figure~1 shows time-averaged radial profiles for $\beta^{-1}$ (i.e.,

304: the ratio of magnetic field energy density to thermal energy

305: density) and the specific angular momentum $l$. As intended, the 

306: magnetic field in Run~2, which starts with a relatively strong vertical 

307: seed field ($\beta_z = 500$), saturates with significantly higher 

308: relative magnetic field energy than those runs which start with either 

309: a weaker initial field or a comparable toroidal field.  This is due to 

310: the continued boosting of the MRI from the conserved vertical flux 

311: that accompanies the initial vertical field (Hawley, Gammie \& 

312: Balbus 1995). Increasing the size of the computational domain also 

313: increases the saturation value of the magnetic field, but by a 

314: smaller factor. 

315: 

316: The dynamics of the flow -- diagnosed using the specific angular

317: momentum -- correlates with the saturation field strength.  Firstly, 

318: consider the unboosted runs (Runs~1,3 and 4; shown by the solid, dashed 

319: and dot-dashed lines respectively).  As one approaches and enters the 

320: plunging region, the radial profile of $l$ flattens. 

321: If the ZTBC is strictly valid, $l$ should be constant within the

322: plunging region. For these runs, we obtain the same result as 

323: before (Armitage, Reynolds \& Chiang 2001), namely that the ZTBC yields a 

324: good but not perfect description of the inner disk dynamics. Indeed, 

325: examination of the $r\phi$ component of the magnetic stress tensor 

326: shows that the Maxwell stresses at $r=r_{\rm ms}$ do not vanish but 

327: are less than the peak stress, which occurs at $r\sim 10-12\,GM/c^2$, 

328: by a factor of 5--8.

329: 

330: On the other hand, angular momentum transport within the plunging region is

331: much more significant in the boosted run (Run~2; dotted line in Fig.~1).

332: The time averaged stress at $r=r_{\rm ms}$ is still less than the peak value

333: (which occurs at $r\sim 8\,GM/c^2$), but only by a factor of 2. 

334: Significant stresses continue to operate within the region $r<r_{\rm ms}$, 

335: and there is evidently a clear violation of the ZTBC.

336: 

337: \subsection{Variability within individual simulations}

338: 

339: Turbulent accretion disks are inherently variable systems and so it is

340: also interesting to examine variability during our simulations.  For

341: this purpose, we will compare and contrast Run~1 and Run~2.  Both runs

342: display order unity temporal and spatial fluctuations in $\beta^{-1}$.

343: For the unboosted run these fluctuations have only a small effect on

344: $l$. For the boosted run shown in Fig.~2, however, there are much

345: larger fluctuations in $l$. In particularly, during one timeslice from

346: Run~2, a large part of the plunging region becomes magnetically

347: dominated ($\beta<1$).  During this period (shown with a thick line in

348: the bottom panels of Fig.~2) $l$ declines within the plunging region

349: with the same slope as in the rest of the disk, with the Maxwell

350: stresses staying fairly constant well into the plunging region.

351: Events as dramatic as this are fairly rare (caught in 1 out of the 12

352: independent timeslices that we analyzed), but it is clear that the

353: stress at $r=r_{\rm ms}$ is a rapidly varying function of time, with

354: much more frequent occurrences of moderate stress.

355: 

356: Extraction of energy and angular momentum from within the plunging

357: region can only occur if the material can remain causally connected to

358: the rest of the disk. Analysis of the simulations shows that large

359: stresses occur within the plunging region when the magnetic fields are

360: strong enough for the {\it peak} radial MHD wave speed to exceed the

361: radial inflow velocity throughout a large part of the plunging region.

362: It is worth noting that even during the magnetically dominated event

363: seen in Run~2, the {\it azimuthally averaged} MHD wave speeds are

364: always slower that the inflow speeds.  The angular momentum transport

365: within this highly inhomogeneous plasma therefore appears to be

366: mediated by low-$\beta$ (i.e., higher relative magnetic field)

367: filaments.

368: 

369: \section{Discussion}

370: 

371: \subsection{Radiative efficiency}

372: 

373: Calculating the implied change to the radiative efficiency as 

374: a consequence of the magnetic torque at $r_{\rm ms}$ is not 

375: straightforward (see the discussion in Hawley \& Krolik 2001). 

376: A simple approach is to note that the dissipation rate is

377: \begin{equation} 

378:  Q(r) = \vert { {{\rm d} \Omega} \over {{\rm d} \ln r } } \vert 

379:  \int T^{r \phi} dz,

380: \end{equation} 

381: where $T^{r \phi}$ is the local stress. We can then estimate the 

382: change in the radiative efficiency by comparing $Q(r)$, computed 

383: using the stress measured in the simulations, with the form 

384: expected in a steady disk if there were no stress at $r_{\rm ms}$. 

385:  

386: In practice, the non-steady nature of the flow, and the 

387: limited radial extent of the simulations, creates 

388: difficulties. An indicative estimate, however, follows from noting 

389: that in run 2 the mean magnetic stress is roughly constant 

390: between $r_{\rm ms}$ and $2 r_{\rm ms}$. Compared to a 

391: standard disc model, this implies a 50\% increase in 

392: the dissipation in the region $r_{\rm ms} < r < 2r_{\rm ms}$, 

393: and a 10\% increase over $r_{\rm ms} < r < 10r_{\rm ms}$. 

394: Significantly larger effects are seen in some individual 

395: timeslices.

396: 

397: \subsection{Comparison with previous work}

398: 

399: The simulations presented here suggest that the different 

400: dynamics within the plunging region, obtained by Hawley \& 

401: Krolik (2001) and Armitage, Reynolds \& Chiang (2001), are 

402: plausibly due to different values in the saturation levels 

403: of magnetic fields in the disk. The relatively strong 

404: magnetic fields in the former simulations, which obtained $\beta\sim 5-10$, 

405: are more comparable to the boosted simulations in the current 

406: work, and lead to substantial angular momentum transport within the 

407: plunging region. Weaker disk fields, with $\beta\sim 20$ and 

408: a Shakura-Sunyaev (1973) $\alpha$ parameter of $\alpha\sim

409: 0.02-0.04$, produce much less striking effects. Although there 

410: are undoubtably other significant differences between the 

411: simulations, all work to date is broadly consistent with 

412: the hypothesis that significant magnetic coupling to 

413: the plunging region requires disk fields with 

414: $\alpha\sim 0.1$ (and $\beta\sim 10$). This applies 

415: for disks that are moderately thin, with relative 

416: thickness $h/r\sim 0.1$, which are the only systems that 

417: have been simulated so far.

418: 

419: \subsection{Scaling to thinner disks}

420: 

421: Very thin disks are difficult to simulate, and it is not obvious how

422: the threshold $\beta$ seen in the current work scales with $h/r$. We

423: note, however, that the Alfven speed in the disk scales as, $v_A

424: \propto \left( {h \over r} \right) \beta^{-1/2}$.  Unless $\beta$ in

425: the disk decreases rapidly with decreasing $h/r$, the Alfven speed in

426: the gas immediately outside the plunging region will decrease with

427: decreasing disk thickness. Conversely, the inflow speed $v_r$ within

428: the plunging region is a fixed function of $r$, and is independent of

429: $h/r$. Hence, a larger increase in the Alfven speed is needed if

430: thinner disks are to maintain causal contact into the plunging

431: region. It is therefore probable that it becomes increasingly more

432: difficult to extract energy from the plunging region as the disk

433: becomes thinner, although further simulation work is required before

434: drawing robust conclusions.

435: 

436: \section{Conclusions}

437: 

438: In this letter, we have used MHD simulations to investigate the effect

439: of the magnetic field strength in the body of the accretion disk on

440: the dynamics of the material within the plunging region. For a disk

441: with a sound speed corresponding to an $h/r \approx 0.1$, we find

442: that plausible saturation values of the magnetic field are within a

443: factor of a few of a threshold between two regimes. In the runs which

444: saturated with a fairly low magnetic field ($\beta\sim 20$), there was

445: only modest extraction of angular momentum from material within the

446: plunging region, and the ZTBC was a reasonable approximation.  In our

447: `boosted' runs, where the field saturated at a higher level

448: ($\beta\sim 5-10$), we found significantly stronger angular momentum

449: transport within the plunging region.  Furthermore, our boosted run

450: showed sporadic intervals of very efficient angular momentum transport

451: (with stresses that remained fairly constant well within the plunging

452: region). We note that stratified simulations by other authors (Hawley

453: 2000; Miller \& Stone 2000; Hawley \& Krolik 2001) suggest that

454: magnetic fields could be stronger (i.e. lower $\beta$) above the disk

455: midplane. Such fields, which are not included in our simulations,

456: could contribute to more significant transport within the plunging

457: region.

458: 

459: The extraction of energy and angular momentum from material in the

460: plunging region can significantly increase the efficiency of accretion

461: above the standard value, especially in the innermost regions. This

462: raises the possibility of varying efficiencies, both across systems,

463: and within a given system at different times.

464: 

465: In this paper, we have used a large scale flux threading the accretion

466: disk as a numerical device for altering the saturation level of the

467: MRI. Large scale fields could also, of course, be present in real

468: systems, and are often invoked in models for the formation of jets and

469: outflows (e.g. Ouyed, Pudritz \& Stone 1997, and references therein).

470: This raises the interesting possibility that periods of increased

471: luminosity (due to a high radiative efficiency) may be related to

472: periods during which strong outflows occur.  Such associations have

473: been reported for Galactic Black Hole Candidates such as GRS~1915+105

474: (Mirabel et al. 1998).

475: 

476: \section*{Acknowledgments}

477: 

478: We extend a special thanks to Charles Gammie for suggesting this

479: problem.  We are grateful to the developers of ZEUS and ZEUS-MP for

480: making these codes available as community resources.  CSR acknowledges

481: support from Hubble Fellowship grant HF-01113.01-98A.  PJA thanks JILA

482: for hospitality.

483: 

484: \begin{references}

485: \reference{}Abramowicz, M.A., Kato, S., 1989, ApJ, 336, 304

486: \reference{}Agol E., Krolik J.H., 2000, ApJ, 528, 161

487: \reference{}Armitage P.J., Reynolds C.S., Chiang J., 2001, ApJ, 548, 868

488: \reference{}Balbus S.A., Hawley J.F., 1991, ApJ, 376, 214

489: \reference{}Clarke D.A., Norman M.L., Fiedler R.A., 1994,

490:  National Center for Supercomputing Applications Technical Report 15 

491: \reference{}Gammie C.F., 1999, ApJ, 522, L57

492: \reference{}Hawley, J.F., 2000, ApJ, 528, 462 

493: \reference{}Hawley, J.F., 2001, ApJ, 554, 534

494: \reference{}Hawley, J.F., Gammie, C.F., Balbus, S.A., 1995, ApJ, 440, 742

495: \reference{}Hawley, J.F., Krolik, J.H., 2001, ApJ, 548, 348

496: \reference{}Krolik J.H., 1999, ApJ, 515, L73 

497: \reference{}Miller K.A., Stone J.M., 2000, 534, 398

498: \reference{}Mirabel, I.F., Dhawan, V., Chaty, S., Rodriguez, L.F., Marti, J., 

499:  Robinson, C.R., Swank, J., Geballe, T., 1998, A\&A, 330, L9

500: \reference{}Norman, M.L., 2000, in Astrophysical Plasmas: Codes, Models \&

501:  Observations, eds J. Arthur, N. Brickhouse and J. Franco, Rev. Mex. Astron. Astrophys.

502:  (Conf. Ser.), 9, 66, astro-ph/0005109

503: \reference{}Novikov I.D., Thorne K.S., 1973, in Black Holes --- Les Astres

504: Occlus, ed. S. DeWitt, B. DeWitt, (New York, London: Gordon and Breach), 345

505: \reference{}Ouyed, R., Pudritz, R.E., Stone, J.M., 1997, Nature, 385, 409

506: \reference{}Paczynski, B., 2000, astro-ph/0004129

507: \reference{}Paczynski, B., Bisnovatyi-Kogan, G., 1981, Acta. Astron., 31, 283

508: \reference{}Paczynski B., Wiita P.J., 1980, A\&A, 88, 23

509: \reference{}Shakura N.I., Sunyaev R.A., 1973, A\&A, 24, 337 

510: \reference{}Stone J.M., Norman M.L., 1992a, ApJS, 80, 791 

511: \reference{}Stone J.M., Norman M.L., 1992b, ApJS, 80, 819

512: \end{references}

513: 

514: 

515: \end{document}

516: 

517: 

518: 

519: 

520: 

521: 

522: 

523: