1: \documentclass[12pt,preprint]{aastex}

2: 

3: \usepackage{amsmath}

4: \usepackage{amssymb}

5: \usepackage{amsmath}

6: \usepackage{mathrsfs}

7: \usepackage{natbib}

8: 

9: \shorttitle{Jets, Disks and Field Topology}

10: \shortauthors{Beckwith, Hawley, Krolik}

11: 

12: \begin{document}

13: 

14: \title{The Influence of Magnetic Field Geometry on the Evolution

15: of Black Hole Accretion Flows: Similar Disks, Drastically Different Jets}

16: 

17: 

18: \author{Kris Beckwith and John F. Hawley}

19: \affil{Astronomy Department\\

20: University of Virginia\\ 

21: P.O. Box 400325\\

22: Charlottesville, VA 22904-4325}

23: 

24: \email{krb3u@virginia.edu; jh8h@virginia.edu}

25: 

26: \and

27: \author{Julian H. Krolik}

28: \affil{Department of Physics and Astronomy\\

29: Johns Hopkins University\\

30: Baltimore, MD 21218}

31: 

32: \email{jhk@pha.jhu.edu}

33: 

34: \begin{abstract}

35: 

36: Because the magnetorotational instability is capable of exponentially

37: amplifying weak preexisting magnetic fields, it might be hoped that the

38: character of the magnetic field in accretion disks would be independent

39: of the nature of the seed field.  However, the divergence-free

40: nature of magnetic fields in highly conducting fluids ensures that

41: their large-scale topology is preserved, no matter how greatly the

42: field intensity is changed.  By performing global two-dimensional

43: and three-dimensional general relativistic magnetohydrodynamic disk

44: simulations with several different topologies for the initial magnetic

45: field, we explore the degree to which the character of the flows around

46: black holes depends on the initial topology.  We find that while the

47: qualitative properties of the accretion flow are nearly independent

48: of field topology, jet-launching is very sensitive to it: a sense of

49: vertical field consistent for at least an inner disk inflow time is

50: essential to the support of strong jets.

51: 

52: \end{abstract}

53: 

54: \keywords{Black holes - magnetohydrodynamics - stars:accretion}

55: 

56: \section{Introduction}\label{intro}

57: 

58:  Magnetic fields are now generally acknowledged to be essential

59: to accretion \citep{Balbus:1998}.  Amplified exponentially from weak seed

60: fields by the magneto-rotational instability (MRI), they are stretched

61: in a consistent direction so as to produce a net stress $\langle -B_r

62: B_\phi\rangle/4\pi$ capable of carrying enough angular momentum outward

63: through the disk to support sizable accretion rates.   Because the

64: nonlinear state of the MRI is fully-developed magnetohydrodynamic (MHD)

65: turbulence, it is often thought (or at least hoped) that its properties

66: have little to do with the character of the seed magnetic field brought

67: to the disk with the accretion flow.

68: 

69:  However, there are certain properties of the magnetic field

70: that are preserved despite the development of nonlinear turbulence.

71: Because magnetic fields are divergence-free and the plasma in accretion

72: flows has very little electrical resistivity, the global topology

73: of the field persists, no matter what happens.  It is therefore

74: natural to ask whether any interesting properties of the accretion

75: flow may depend on the global field topology.  For example, could

76: the field topology alter the relation between surface density and

77: accretion rate?   Or how the stress varies through the marginally

78: stable region?

79: 

80:   In addition to swallowing matter, accreting black holes often

81: expel relativistic jets.  Although the mechanisms by which these jets are

82: launched have long been the subject of investigation and speculation,

83: the current consensus focuses on a combination of large-scale magnetic

84: fields and the rotation of either the central hole, the accretion

85: disk, or both.  Two specific MHD models have received the greatest

86: attention: the Blandford-Znajek mechanism \citep{Blandford:1977}

87: and the Blandford-Payne mechanism \citep{Blandford:1982}.  In the

88: Blandford-Payne model a large-scale vertical magnetic field is anchored

89: in the disk, rotating with the orbital frequency.  Above and below the

90: disk magnetic tension dominates, and the field lines force the matter to

91: rotate with this same frequency.  If the fieldlines are angled outward

92: sufficiently with respect to the disk, there can be a net outward force,

93: accelerating the matter along the rotating fieldlines.

94: The Blandford-Znajek model also depends upon rotating field lines, but

95: the source of that rotation (and the source of the power in the jet)

96: is the black hole rather than the disk: any field lines connecting the

97: immediate vicinity of the black hole to infinity are forced to rotate

98: by frame-dragging, so that Poynting flux carries off the energy.

99: 

100:  Given the strong dependence of both the Blandford-Znajek and the

101: Blandford-Payne mechanisms on the nature of the magnetic field in the

102: disk and near the black hole, it is equally natural to inquire whether

103: the topology of the field embedded in the accretion flow has any influence

104: on their operation.  

105: 

106:  It is the goal of this paper to begin to investigate these questions,

107: both as they relate to accretion disk dynamics and as they relate to

108: jet properties.  Unfortunately, it is difficult even to scratch the

109: surface by purely analytic techniques \cite[although there have been some

110: attempts using highly-simplified models:][]{Krolik:1999a,Gammie:1999}.

111: On the other hand, direct numerical simulations provide a means by which

112: aspects of these problems can be investigated in detail.  \cite{HK:2002}

113: made a first exploration of the dependence of accretion properties on

114: field topology in a highly-simplified approximation: Newtonian dynamics

115: in a pseudo-Newtonian potential.   They found that a purely toroidal

116: initial field led to a systematically smaller accretion rate than one

117: whose initial field was dipolar.

118: 

119: More recently, fully general relativistic MHD  (GRMHD) simulations have

120: become possible \citep{McKinney:2004, De-Villiers:2005,Hawley:2006}.

121: These works have focused primarily on a single initial field structure:

122: a set of nested dipole field loops contained entirely within the

123: initial gas torus.  Other field geometries have been examined, but

124: only in rather limited fashion.  For example, \cite{De-Villiers:2005}

125: simulated the evolution of a torus with an initial toroidal field around

126: a Schwarzschild black hole and noted the lack of any resulting jet,

127: but did not consider why no jet formed.  \cite{McKinney:2004}, using

128: axisymmetric simulations, compared the electromagnetic luminosity

129: of the standard dipole jet with that resulting from a few alternate

130: magnetic field topologies.  In particular, they studied cases in

131: which the initial field was arranged in loops of alternating sign

132: arranged either vertically or horizontally within the initial torus.

133: The principal effect they reported was a factor of 2--3 reduction

134: in the ratio of the time-averaged jet Poynting flux to the mean mass

135: accretion rate.  They also noted that an initial field that is purely

136: vertical led to significant enhancements in both the mean mass accretion

137: rate and the time-averaged jet Poynting flux.  Axisymmetric simulations

138: of tori threaded by net vertical field led \cite{De-Villiers:2006} to

139: concur that such a field can greatly enhance jet power.  On the basis

140: of these results, \cite{McKinney:2007} proposed that black holes

141: inevitably become threaded by an organized vertical field if the initial

142: disk contains a large-scale poloidal field with few parity changes.

143: The intriguing character of these results motivates a more systematic

144: study, one that employs three-dimensional simulations, considers the

145: accretion flow as well as the jet, and investigates time-dependence.

146: 

147: The question immediately arises: how can one best attempt to describe the

148: very complex mixture of topologies likely to be found in real accretion

149: flows?  The magnetic field near a black hole can be thought of as simply

150: what the accretion process itself has self-consistently brought there

151: from whatever source supplies it.   In order to gain understanding of

152: what, in realistic circumstances, must be a truly complex combination of

153: structures, in this paper we do not try to construct a single ``realistic"

154: field geometry; instead, we study the effects due to individual members

155: of a ``basis set" of field topologies.  Our hope is that by understanding

156: the action of pure types, we can gain some understanding of how realistic

157: mixtures behave.

158: 

159: To be specific, we will begin with a carefully chosen set of different

160: kinds of closed field loops that we expect to be generic.  Closed field

161: loops have the advantage of requiring many fewer free parameters

162: to describe them fully within a simulation than do field geometries

163: with net flux.  Net flux geometries may well exist in Nature and have

164: important effects; we will consider them in the second stage of this

165: project.   We begin with the conjecture that, at a qualitative level,

166: the different sorts of closed field loop behavior can all be represented

167: by a combination of those due to: dipolar loops, or fieldlines initially

168: entirely poloidal and which cross the equatorial plane at well-separated

169: radii; toroidal loops, which encircle the black hole; and quadrupolar

170: loops, which (again, in the initial state) are poloidal, but close

171: without crossing the equatorial plane.  For each of these cases, we have

172: run a lengthy three-dimensional GRMHD simulation whose results we hope

173: characterize the accretion flow structure created by that field topology.

174: 

175: We have also run a series of high resolution axisymmetric simulations of

176: each of the initially poloidal topologies (i.e. dipole and quadrupole). In

177: addition, in order to understand better the implications of the finite

178: lifetimes of closed loops, we have performed an axisymmetric simulation

179: whose initial magnetic field structure was a series of dipolar loops,

180: each contained within a span of radii narrow enough that the difference

181: in inflow times across a loop is comparable to the equilibration time

182: for the magnetic field in the jet.

183: 

184: The rest of this work is structured as follows. In \S\ref{numerics}

185: we briefly review the numerical scheme employed to solve the

186: equations of GRMHD and give a detailed description of the initial

187: conditions used to generate each topology. In \S\ref{flowstructure}

188: we examine the late time accretion flow structure that evolves from

189: each of the initial field topologies, paying attention to each of the

190: structures (main disk body, magnetized corona) that were identified by

191: \cite{De-Villiers:2003b} as arising from the evolution of the dipole

192: field topology.  In \S\ref{jets} we examine the properties of the third

193: structure identified by \cite{De-Villiers:2003b}, the unbound outflow,

194: and examine how the initial evolution of the field during the linear

195: growth phase of the MRI gives rise to the components of the accretion

196: flow at late times. Finally in \S\ref{conclusion} we summarize our

197: results and highlight their importance for jets in astrophysical settings.

198: 

199: \section{Numerical Details}\label{numerics}

200: 

201: In this work we undertake a study of both two- and three-dimensional

202: simulations to investigate the influence of magnetic field topology

203: on both the evolved accretion flow structure and relativistic jet

204: formation.  In doing so we continue a program of black hole accretion

205: disk simulations begun in \cite{De-Villiers:2003b}, \cite{Hirose:2004},

206: \cite{De-Villiers:2005}, \cite{Krolik:2005} and \cite{Hawley:2006}.

207: Three-dimensional simulations, such as those presented in the previous

208: work, are the only means by which one can gain insight into the

209: evolved accretion flow properties due to the fundamental restriction posed by

210: the anti-dynamo theorem \cite[]{Moffatt:1978}.  Further, toroidal field

211: models cannot be studied in axisymmetry, since the toroidal MRI requires

212: nonzero azimuthal wavevectors.  On the other hand, because jet formation

213: occurs promptly

214: during the evolution of the flow \cite[]{Hawley:2006}, restrictions due

215: to the anti-dynamo theorem are less critical for jet studies.  In this

216: case the higher resolution afforded by the two-dimensional simulations

217: is particularly valuable.

218: 

219: The simulation code we use is

220: described in \cite{De-Villiers:2003}.  This code solves the equations of

221: ideal non-radiative MHD in the static Kerr metric of a rotating black hole

222: using Boyer-Lindquist coordinates.  Values are expressed in gravitational

223: units $(G = M = c = 1)$ with line element $ds^{2} = g_{tt} dt^{2} +

224: 2g_{t \phi} dt d\phi + g_{rr} dr^{2} + g_{\theta \theta} d \theta^{2}

225: + g_{\phi \phi} d\phi^{2}$ and signature $(-,+,+,+)$.  Since we are

226: focusing in this paper on the influence of the magnetic field topology,

227: we fix the black hole spin to $a/M =0.9$ for all of the new simulations

228: reported here.

229: 

230: The relativistic fluid at each grid zone is described by its density

231: $\rho$, specific internal energy $\epsilon$, $4$-velocity $U^{\mu}$, and

232: isotropic pressure $P$. The relativistic enthalpy is $h = 1 + \epsilon

233: + P / \rho$. The pressure is related to $\rho$ and $\epsilon$ via the

234: equation of state for an ideal gas, $P = \rho \epsilon ( \Gamma - 1)$. The

235: magnetic field is described by two sets of variables.  The first is the

236: constrained transport magnetic field $\mathcal{B}^{i} = [ijk] F_{jk}$,

237: where $[ijk]$ is the completely anti-symmetric symbol, and $F_{jk}$ are

238: the spatial components of the electromagnetic field strength tensor.

239: From the constrained transport (CT) magnetic field components, we

240: derive the magnetic field four-vector,

241: $(4\pi)^{1/2}b^{\mu} = *F^{\mu \nu} U_{\nu}$,

242: and the magnetic field scalar,

243: $||b^{2}|| = b^{\mu} b_{\mu}$. The electromagnetic component of the

244: stress-energy tensor is $T^{\mu \nu}_{\mathrm{(EM)}} = \frac{1}{2}g^{\mu

245: \nu} ||b||^{2} + U^{\mu} U^{\nu} ||b||^{2} - b^{\mu} b^{\nu}$.

246: 

247: In all of the simulations, the initial condition for the matter

248: is an isolated, hydrostatic torus orbiting near the black hole.

249: The initial state of the three-dimensional simulations was chosen

250: so as to be consistent with that used in earlier simulations in this

251: series \cite[see][]{De-Villiers:2003b}.  The matter's adiabatic index,

252: $\Gamma$ was $5/3$ and the angular momentum distribution was slightly

253: sub-Keplerian, with a specific angular momentum at the inner edge of

254: the torus (located at $r=15$M) $\ell_{\rm in} = 4.567$ (we define $\ell

255: \equiv U_\phi/U_t$).  For this choice of $\ell_{\rm in}$, the pressure

256: maximum is at $r \approx 25M$.

257: 

258: In the two-dimensional simulations, our fiducial torus was chosen

259: to have a constant angular momentum distribution with $\ell = 5.2$,

260: $\Gamma$ was 4/3, and the inner edge was again located at $r=15$M.

261: This torus is significantly thicker than the nearly-Keplerian torus

262: used for the three-dimensional simulations; its density scale height

263: at the pressure maximum, measured along surfaces of constant radius,

264: is a factor $\sim1.3$ greater than that of the three-dimensional torus.

265: This thicker configuration makes it easier to resolve the multiple field

266: loop model, where several field loops must fit into a restricted space.

267: This specific combination of $\ell$ and $\Gamma$ was also necessary to

268: manipulate the shape of the torus so that there was an appropriate balance

269: in magnetic flux between the different components of the multiple dipole

270: loop simulation.  In the quadrupolar field case, in order to study how

271: the evolution of the field depends on its detailed geometry, we examined

272: what happened as $\ell$ varied from 4.9 to 5.2 in steps of 0.05.

273: 

274: The initial torus is overlaid with a variety of magnetic field configurations,

275: all with zero net flux.

276: The initial poloidal magnetic field is determined from the curl of the

277: four-vector potential, i.e., $F_{\mu\nu} = \partial_{\mu} A_{\nu} - 

278: \partial_{\nu} A_{\mu}$ with only $A_\phi \neq 0$.  For a single

279: dipolar loop, we set

280: \begin{equation}

281: A_{\phi} = A_0 \left(\rho - \rho_{cut} \right).

282: \end{equation}

283: For a quadrupole, the azimuthal component of the vector potential is

284: \begin{equation}

285: A_{\phi} = A_0 \left(\rho - \rho_{cut} \right) r \cos \theta.

286: \end{equation}

287: In both of these cases $A_\phi$ is set to zero where the density is less

288: than $\rho_{cut}$; this device confines the initial field to lie well

289: inside the torus. In all the poloidal simulations, the initial field

290: intensity (i.e., the parameter $A_0$) is determined by setting

291: the volume-averaged ratio of gas to magnetic pressure $\beta = 100$.

292: 

293: The vector potential for $n$ dipolar loops (used only for the 

294: two-dimensional simulations) is more complicated.  It is given by

295: \begin{equation}

296: A_{\phi} = A_0 \left(\rho - \rho_{cut} \right) \displaystyle\sum_{\mathrm{j} = 1}^{n}

297: \frac{J_\mathrm{j} R_\mathrm{j}}{\sqrt{R^{2}_\mathrm{j} + r^{2}

298: + 2 R_\mathrm{j} r \sin \theta}} \left[\frac{\left( 2- k_j^{2} \right)

299: K(k_j) - 2E(k_j)}{k_j^{2}} \right],

300: \end{equation}

301: where the total vector potential is the superposition of the

302: (non-relativistic) vector potentials from $n$ current loops

303: \cite[see][Pg. 182, Eqn. 5.37]{Jackson:1975}.  Here $J_{\mathrm{j}}$

304: is the current in the $j$th loop, located at radius $R_{\mathrm{j}}$

305: in the equatorial plane.  $K$ and $E$ are the complete elliptic

306: integrals of the first and second kinds, and $k_j = 4 R_\mathrm{j}

307: r \sin \theta / (R^{2}_\mathrm{j} + r^{2} + 2 R_\mathrm{j} r \sin

308: \theta )$.  The $R_\mathrm{j}$ are chosen so that the difference

309: between the inflow times of adjacent loops is approximately equal to

310: one jet equilibration time, $\sim 700$--$1000M$.  The signs of the $J_j$

311: alternate with increasing radius so that the senses of the dipolar

312: field loops alternate.  The magnitudes of the currents were chosen with the goal of

313: ensuring that the evolution of the two radially innermost loops was governed by the field

314: local to those loops.  The purpose of this simulation was to study the interaction

315: of the two inner loops as they accrete; in that context, the only function of the

316: two outer loops is to supply the outer part of the torus with magnetic field, so

317: that it could accept the angular momentum brought to it from the true accretion flow.

318: Thus, the most appropriate outer field structure for this particular numerical experiment

319: is one that has the least influence on dynamics inside the pressure maximum.

320: After some experimentation, we found that this goal could be achieved only by

321: giving the two outer loops relatively small currents.    When that was done,

322: the outer two loops moved outward with the outer portion of the disk and were quickly

323: destroyed (i.e. forced to reconnect) as the second loop from the inside expanded

324: toward them.    We make no pretense that this choice is a fair representation of generic

325: four-loop evolution, but it does support our study of what happens when two complete

326: loops of opposite sense are brought into the inner  disk one after the other.

327: 

328: For a purely toroidal initial field, all that is necessary is to set

329: ${\mathcal B}^r = {\mathcal B}^\theta = 0$ and determine ${\mathcal

330: B}^\phi$ by a choice of volume-averaged $\beta$.  Here we use $\beta=10$,

331: as, in the purely toroidal case, accretion can occur only once MRI-induced

332: turbulence on the large scales has grown to sufficiently large amplitude,

333: and long wavelength modes grow comparatively slowly.  Beginning with a

334: comparatively strong field ensured that accretion activity started soon

335: after the simulation began.  After saturation, the field (in relative

336: terms) actually weakened somewhat: the volume-averaged $\beta$ grew to

337: $\approx 30$ (see Section \ref{diskbody}). As before, the initial

338: magnetic field is set to zero for densities below $\rho_{cut}$.

339: 

340: We take the dipole simulation KDPg \cite[originally presented

341: in][where further discussion of our boundary conditions and grids

342: may be found]{Hawley:2006} as our fiducial three-dimensional run.

343: The three-dimensional quadrupole simulation is designated QDPa

344: and the three-dimensional toroidal simulation is TDPa.  

345: KDPg and QDPa were run for $10^4\ M$ in time, which corresponds to

346: approximately $12$ orbits at the initial pressure maximum.  The toroidal

347: field model, TDPa, evolves initially at a slower rate; it was run

348: to $2.9\times 10^4 M$ in time.  For each simulation, the time step

349: $\Delta t$ was determined by the minimal light crossing time for a zone

350: on the spatial grid and remained constant for the entire simulation

351: \cite[]{De-Villiers:2003}.

352: 

353: Each of the three-dimensional simulations used $192 \times 192 \times

354: 64$ $(r,\theta,\phi)$ grid zones.  The radial grid extends from an inner

355: boundary located at $r_{in} = 1.5 M$, which lies just outside the black

356: hole event horizon, to an outer boundary located at $r_{out} = 120 M$.

357: For the two poloidal topologies, the radial grid was graded using a

358: hyperbolic cosine distribution; for the toroidal case a logarithmic

359: distribution was used.  An outflow condition is applied at both the inner

360: and outer radial boundary.  The $\theta$-grid spans the range $0.045

361: \pi \le \theta \le 0.955 \pi$ using an exponential distribution that

362: concentrates zones near the equator.  A reflecting boundary condition is

363: used along the conical cutout surrounding the coordinate axis.  

364: The $\phi$-grid spans the quarter plane, $0 \le \phi \le \pi / 2$, with

365: periodic boundary conditions applied in $\phi$. The use of this restricted

366: angular domain significantly reduces the computational requirements of

367: the simulation.

368: 

369: The two-dimensional simulations were performed on a $1024\times 1024$

370: $(r,\theta)$ grid in which the radial cells are graded logarithmically,

371: and the polar angle cells are concentrated toward the equator in the

372: same way as for the three-dimensional simulations.  The $\theta$ grid,

373: however, goes much closer to the axis than in the three-dimensional

374: simulations, with a cut-out of only $10^{-5}\pi$~radians.  In addition

375: to allowing higher poloidal resolution at a reasonable computational

376: cost, two-dimensional simulations have the additional advantage that

377: field-lines can be easily visualized as level surfaces of the vector

378: potential component $A_\phi$.  Using this approach, we illustrate the

379: magnetic field topologies for the three different poloidal configurations

380: in Figure \ref{topologies}.

381: 

382: \section{The Accretion Flow}\label{flowstructure}

383: 

384: We begin by examining the average properties of the accretion flows in

385: simulations KDPg, QDPa and TDPa.  \cite{De-Villiers:2003b} described the

386: accretion flow structure in terms of three distinct regions:

387: the main disk body, the coronal envelope, and the unbound outflow.

388: In this section, we study the first two, paying special attention to

389: contrasts (or lack of contrast) when the magnetic topology changes.

390: We will discuss the unbound outflow in the following section.

391: 

392: Much of the analysis will be in terms of various time- and space-averaged

393: quantities.  The radial profile of the time-averaged shell integral of $\cal Q$

394: is defined as

395: \begin{equation}

396: \langle{\cal Q}(r)\rangle_{F} = 

397: \frac{2}{\pi T} \int \sqrt{-g} {\cal Q}(t,r,\theta,\phi) dt d\theta d\phi

398: \end{equation}

399: Similarly, the shell average of a quantity $\cal Q$ at a radius $r$ is defined to be:

400: \begin{equation}

401: \langle{\cal Q}(r)\rangle_{S} = 

402: \frac{ \int \sqrt{-g} {\cal Q} (t,r,\theta,\phi) d\theta d\phi} {\int \sqrt{-g} d\theta d\phi}

403: \end{equation}

404: The angular profile of the time-average of $\cal Q$ at radius $r$ 

405: is defined by

406: \begin{equation}

407: \langle{\cal Q}(\theta;r)\rangle_{A} = 

408: \frac{2}{\pi T}\int \sqrt{-g} {\cal Q} (t,r,\theta,\phi) dt  d\phi.

409: \end{equation}

410: Lastly, the time- and volume-average of $\cal Q$ is:

411: \begin{equation}

412: \langle{\cal Q}\rangle_{V} = \frac{\int \sqrt{-g} {\cal Q} (t,r,\theta,\phi) dt dr d\theta d\phi}

413: {\int \sqrt{-g} dt dr d\theta d\phi}.

414: \end{equation}

415: In these equations $T$ is the time over which the integral is computed, and

416: $g$ is the usual metric determinant.  Typically $T=6000M$; for KDPg

417: and QDPg this is the last $6000M$ of the full evolution, while for

418: TDPa we choose a $6000M$ window in the middle of the simulation after

419: the accretion flow is established.  The spatial extent of the shell

420: integration is the full $\theta$ and $\phi$ computational domain.

421: During a given simulation, various shell integrals and radial fluxes are

422: computed and stored every $M$ in time.  This data can then be integrated

423: over time to obtain quantities such as the total or time-averaged jet

424: outflow or accretion rate.

425: 

426: It is also useful to divide the shell and volume

427: integrals into two parts, one for bound and one for unbound flow.

428: For simplicity we define a particular zone to be ``unbound" if $-h U_t >

429: 1$.  Unbound outflow can further be defined as those unbound cells

430: with $U^r > 0$.  In these simulations only the outflow near the axis

431: (the jet outflow) is unbound; the coronal backflow from the disk itself

432: remains bound.

433: 

434: \subsection{Disk body}\label{diskbody}

435: 

436: The initial evolution of the accretion disks in KDPg and QDPa is

437: qualitatively similar.   Both field configurations begin with considerable

438: radial field within the torus.  This is sheared out, generating toroidal

439: field which, by $t\sim 500M$, is sufficiently strong that the

440: resulting poloidal gradient in $||b||^2$ begins to drive the inner edge

441: of the torus (initially located at $r=15M$) inward.  The inner edge of the

442: disk arrives at the black hole at $t\sim 1000M$.  Within the disk body, the

443: MRI generates the turbulence that will determine the subsequent evolution

444: of the disk, and by $t\sim 4000M$ a statistically stationary

445: turbulent accretion flow has been established inside the radius of

446: the inner edge of the initial torus.

447: 

448: The toroidal field model TDPa evolves more slowly than the two poloidal

449: field cases, consistent with the results from earlier toroidal field

450: pseudo-Newtonian simulations \citep{HK:2002}.  As discussed at length in

451: \cite{HK:2002}, this behavior stems both from the absence of an initial

452: radial field (which means that there is no toroidal field amplification

453: due to shear) and from the fact that long-wavelength modes---which

454: are the most effective in driving accretion---grow relatively slowly.

455: Inflow can begin only when the MRI has produced turbulence of sufficient

456: amplitude, which occurs by $t=4000M$, corresponding to about 5 orbits at

457: the radius of the torus pressure maximum.  The accretion rate into the

458: hole increases until about $t=1.5\times 10^4 M$, after which it shows

459: large fluctuations without an overall trend.

460: 

461: Figure~\ref{bndavgplt1} shows time-averaged, shell-integrated radial profiles of a

462: number of quantities relevant to the accretion flow: accretion rate

463: $\dot{M} = \langle \rho U^{r} (r) \rangle_{F}$, surface density $\Sigma(r) = \langle \rho \rangle_F / \int \sqrt{g_{rr}g_{\phi\phi}} d\phi$, the net accreted angular momentum per unit rest mass, $L = \langle T^{r}_{\phi \; \mathrm{(FL)}} (r) + T^{r}_{\phi \;\mathrm{(EM)}} (r)\rangle_{F} / \dot{M}$, $\langle ||b||^2 (r) \rangle_{F}$, the electromagnetic Poynting flux $\langle |T^r_{t} |_{\; \mathrm{(EM)}} (r) \rangle_{F}$ and the EM angular momentum flux $\langle |T^r_{\phi} |_{\; \mathrm{(EM)}}  (r) \rangle_{F}$.  The subscripts FL and EM denote the fluid and electromagnetic contributions to the stress-energy tensor, respectively.

464: All quantities are computed in the coordinate frame after the turbulent

465: accretion flow is established, and in all cases the volume integral

466: was restricted to cells where the matter was bound.  The poloidal field

467: simulations were averaged over time $t=4000$--$10000M$ while the toroidal

468: field data were averaged over $t=12500$--$18500M$.

469: 

470: These six plots may be divided into two groups: those with little

471: dependence on initial field topology (surface density and accreted angular

472: momentum per unit mass), and those with a stronger dependence (accretion

473: rate, magnetic field strength, Poynting flux, and electromagnetic angular

474: momentum flux).  Although the radial distribution of the surface density

475: in all three cases is fairly similar, there is a slight distinction

476: between the two poloidal models on the one hand and the toroidal on

477: the other.  The surface density in the toroidal model rises somewhat

478: more steeply with radius, with lower density near the black hole and

479: higher values outside of $r/M=7$.  For all three topologies, the specific

480: angular momentum carried into the hole is close to the angular momentum

481: of the marginally stable orbit.  For TDPa, the value is essentially

482: equal to that of the ISCO, while both QDPa and KDPa have smaller values,

483: indicative of continued stress operating near or inside the ISCO.

484: 

485: In all three cases, the time-averaged accretion rate is constant with

486: radius out to $r/M \simeq 7$--10; in this sense, all three have achieved a

487: quasi-steady state within $\simeq 4 r_{\rm ISCO}$, where $r_{ISCO}$ is the

488: radial coordinate of the innermost stable circular orbit, $2.32M$

489: (the rise in $\dot{M}$ and $L$ at large radius in TDPa is likely an

490: artifact of the stronger magnetic field with which that simulation

491: was begun; equilibration takes place on timescales proportional to

492: the local orbital period, which is, of course, $\propto r^{3/2}$, so

493: inflow equilibration is slower at large radius).

494: However, the actual value of the accretion rate is significantly

495: different in the three simulations.

496: The time-averaged mass accretion rate is highest in QDPa, exceeding that

497: of KDPg by about $30\%$, while $\dot M$ is lower in TDPa by more than

498: a factor of 2 compared to QDPa. 

499: 

500: Accretion is, of course, driven by angular momentum transport, which

501: is largely due to magnetic torques.  To support these torques, the

502: shell-integrated electromagnetic angular momentum flux must increase

503: outward, as the net torque

504: on the matter is the divergence of the EM angular momentum flux. The

505: dependence of the electromagnetic angular momentum flux on initial

506: field topology is therefore similar to that of the mass accretion rate:

507: it is consistently greater for KDPg and QDPa than for TDPa.  The most

508: significant difference between the KDPg and QDPa curves is that in

509: the region near the ISCO, the electromagnetic angular momentum flux

510: in KDPg becomes almost constant with decreasing radius; this effect

511: has previously been seen in other simulations that begin with dipole

512: magnetic field loops \citep{Krolik:2005}.  In the other two cases,

513: the EM angular momentum flux continues to decline inward.

514: 

515: The mean magnetic field strength and Poynting flux show a very similar

516: dependence on initial field topology: QDPa and KDPg closely resemble

517: each other, TDPa is somewhat weaker.  In the toroidal case, the disk body

518: average magnetic energy density is weaker by a factor of several than in

519: the other two simulations, and decreases inward toward the horizon rather

520: more rapidly.  Correspondingly, the time- and volume-averaged $\beta$

521: parameter was 11 in QDPa, 16 in KDPg, and 30 in TDPa (because this is

522: a volume-average, the corona is weighted heavily in this statistic).

523: The electromagnetic energy flux shows the greatest variation between

524: models.   $\langle |T^r_{t} |_{\; \mathrm{(EM)}}  (r) \rangle_{F}$ in KDPg is consistently larger than the value in QDPa: by a factor of two at $r/M=10$ and a factor of

525: six at $r/M=1.65$.   The value in TDPa is smaller still by a factor of

526: $\sim 2-4$ at these same radii.

527: 

528: All these relations can be understood in a straightforward fashion:

529: there is a nearly-constant ratio between the EM angular momentum

530: flux (and therefore the accretion rate), the Poynting flux, and the

531: magnetic field intensity.  In Figure~\ref{ratio}, we plot the ratios of

532: $\dot{M}$, $\langle |T^r_{t} |_{\; \mathrm{(EM)}}  (r) \rangle_{F}$ and$\langle |T^r_{\phi} |_{\; \mathrm{(EM)}}  (r) \rangle_{F}$

533: to $\langle ||b||^2 (r) \rangle_{F}$.  The inter-simulation contrast in the means of these

534: quantities falls by a factor of several from the plot of their absolute

535: values (Fig.~\ref{bndavgplt1}) to the plot of their ratios to $||b||^2$

536: (Fig.~\ref{ratio}).   Thus, we see that their primary dependence is on

537: the magnetic field strength.  Only in the ratio of the electromagnetic

538: energy flux to magnetic field strength do we see any global variation

539: between the field topologies: $\langle |T^r_{t} |_{\; \mathrm{(EM)}}  (r) \rangle_{F}$ in KDPg is consistently a factor of two above QDPa and TDPa.  This contrast arises

540: from the strong Poynting flux in KDPg associated with the slower-moving,

541: bound portion of the funnel-wall outflow \cite[as can be seen from Figure

542: 10 of][see \S\ref{jets}]{Krolik:2005}.

543: 

544: The behavior of the fluid-frame stress can be seen in Figure~\ref{maxwell},

545: which depicts its time-averaged vertically-integrated value as a function of

546: radius.  To compute this quantity, we project the stress tensor onto a local

547: tetrad system in each cell and multiply by the fluid-frame cell volume,

548: which is computed by a similar projection \citep{Krolik:2005}.  After

549: integrating over the bound matter on that spherical shell, we normalize

550: the result to the surface area that this finite-thickness shell occupies

551: in the equatorial plane.  We present the result in this form in order to

552: compare with the conventional representation of the vertically-integrated

553: stress given by \cite{Novikov:1973}, which is presented in these terms.

554: The Novikov-Thorne fluid-frame stress, unlike ours, is {\it assumed}

555: to be zero at and within the ISCO.   In the figure, the Novikov-Thorne

556: stress is normalized by the time-averaged value of the accretion rate in

557: the simulation.

558: 

559: Outside the ISCO, the radial profile of the Maxwell stress is similar

560: to the prediction of the Novikov-Thorne model, but in no case is there a

561: roll-over in the value near the ISCO.  As has been noted previously for

562: the initial dipole field topology \citep{Krolik:2005}, the $r-\phi$ stress

563: does not go to zero at the ISCO, but is continuous down to the event

564: horizon.  This is true for all three field topologies, at least in a

565: time-averaged sense, although the amplitude of the stress declines as

566: one goes from dipole to quadrupole to toroidal field topology.  In the

567: dipole case, the logarithmic gradient of the stress as a function of

568: radius actually steepens as the accretion flow approaches the horizon.

569: 

570: As can be seen from Figure~\ref{bndavgplt1}, this extra stress does

571: reduce the specific angular momentum carried into the hole by the

572: accretion flow.  At the ISCO the accretion flow in the toroidal field

573: simulation has $L \sim L_{ISCO}$ (where $L_{ISCO} = 2.1$ is the specific

574: angular momentum of the ISCO) and the accretion flows

575: in the dipole and quadrupole simulations have $L \sim 2.05$ (a change of

576: $\sim 2.5\%$ compared to $L_{ISCO}$). Below the ISCO, $L$ in the quadrupole

577: and toroidal cases remains constant with decreasing $r$.  In the dipole

578: simulation, $L$ decreases to $\sim 1.98$ at $r/M = r_{in}$, a decrease of

579: $\sim6\%$ compared to $L_{ISCO}$. The fall off in the dipole case below

580: the ISCO is presumably due to the steepening of the Maxwell stress in

581: this region.  

582: 

583: As has been remarked in earlier papers \cite{Krolik:2005},

584: \cite{Gammie:2004}, when $a/M = 0.9$, the magnitude of the Maxwell stress

585: (i.e., the electromagnetic angular momentum flux) is smaller than the

586: angular momentum flux associated directly with matter ($\dot M u_\phi$)

587: in the inner disk.  It follows (as shown in Figure~\ref{ratio}), that the

588: mean accreted angular momentum per unit rest-mass is only slightly below

589: the Novikov-Thorne value (i.e., $u_\phi$ at the ISCO).  Nonetheless,

590: this continuity in the stress is important for other reasons: because

591: the dissipation is rather more centrally-concentrated than the stress,

592: this additional stress can be significant in the disk heat budget; and

593: the relative importance of the additional stress increases sharply for

594: higher black hole spins.

595: 

596:  Thus, we see that the initial magnetic field geometry makes relatively

597: little qualitative difference to the character of the accretion flow.

598: The sense of its (weak) influence is that the simulation beginning with

599: a purely toroidal magnetic field results in a saturated field amplitude

600: a factor of two smaller than those resulting from poloidal seed fields.

601: The Poynting flux, electromagnetic angular momentum flux, and accretion

602: rate scale in direct proportion to the magnetic field energy density.

603: 

604: In their study based on axisymmetric simulations, \cite{McKinney:2007} also found

605: that certain disk properties are largely independent of the initial field configuration.  They

606: found that the shell-averages of magnetic field quantities within the bound portion of

607: the accretion flow (strictly speaking, where $||b||^2 < \rho$) follow

608: power-law distributions that are the same regardless of the initial field  topology.

609: For example, they found that the field amplitude $\langle ||b||(r) \rangle_{S} \propto r^{-1.3}$  from the ISCO

610: out to about $r=10M$.  Integrating the same magnetic field quantities (magnetic field strength

611: and the lab-frame projections of the CT-fields) over the bound portion of the flow

612: in our simulations reveals that these quantities do scale roughly as power-laws in

613: radius, but with shallower gradients than as determined by \cite{McKinney:2007}.   In

614: addition, we see a slow trend with field topology, as well as significant fluctuations in

615: time:  $\langle ||b||(r) \rangle_{S} \propto  r^{-0.75\pm0.09}$ in simulation KDPg (where $\pm$ denotes 1 std.

616: deviation about the time-average), $\langle ||b||(r) \rangle_{S} \propto r^{-0.67\pm0.08}$ in  simulation QDPa,

617: and $\langle ||b||(r) \rangle_{S} \propto r^{-0.58\pm0.09}$ in simulation TDPa.  The contrast between our results and \cite{McKinney:2007} could result from a number of considerations: the initial torus configurations differed in their radial pressure and angular momentum profiles, and our simulations are

618: fully three-dimensional, rather than axisymmetric.    These scalings may or may not be

619: universal, but they appear to have little to do with the jet.  As we will shortly show in

620: \S\ref{jets}, jet properties differ strongly from one initial magnetic topology to another; it

621: must be, then, that jets depend on other properties in addition to the current distribution in the disk.

622: 

623: \subsection{Corona}

624: 

625: \cite{De-Villiers:2003b} noted that the main disk body is surrounded by

626: a low density region where the magnetic and thermal energies are in

627: approximate equipartition.  This region was designated as the coronal

628: envelope.  In contrast to the main disk body, where poloidal fluid

629: velocities are dominated by turbulence, the poloidal motions in the

630: coronal envelope are smooth on small scales and create an outgoing,

631: but bound, backflow.

632: 

633: Figure \ref{beta} displays the spatial distribution of the time- and

634: azimuthally-averaged $\beta$ parameter overlaid with contours of the

635: gas density for each of the three different topologies.  KDPg shows the

636: familiar strongly-magnetized axial funnel and a corona with relatively

637: large regions of $\beta < 1$.  QDPa shows a lower level of magnetization

638: in the funnel region, while TDPa has much larger regions with $\beta >

639: 1$ overall.  To quantify these impressions, we adopt a more precise

640: definition of the corona than that given in \cite{De-Villiers:2003b}.

641: We define it as those regions of {\it bound} material that lie outside

642: of $\sqrt{2}$ density-scale heights from the mid-plane, that is those

643: cells that satisfy both $-h U_t \le 1$ and $\rho / \rho_{eq} \le e^{-2}$,

644: where $\rho_{eq}$ is the equatorial density and we 

645: measure along curves of constant $r$.

646: 

647: Using this definition, we integrate over the corona and over time

648: to compute the average total volume and mass fraction in the corona,

649: the volume-averaged $\beta$ parameter, and the mean fluid-frame

650: thermal and magnetic energy per unit mass.  These data are given in

651: Table~\ref{corona}.

652: 

653: The coronal regions for all three simulations take up half or more of the

654: total simulation volume but contain only $5\%$ of the mass.  The volume

655: assigned to the corona for KDPg is a smaller percentage compared to the

656: other models, but KDPg also contains an unbound jet along the funnel

657: axis, which is excluded from the corona by definition.

658: Within the coronal volume, the average thermal and magnetic

659: energies per mass are similar, with volume averaged $\beta$ values

660: of $\simeq 2$--4.  

661: 

662: Figure \ref{angcor} plots the angular profiles of density $\langle \rho (\theta; r) \rangle_{A}$, gas pressure $\langle P (\theta; r) \rangle_{A}$, 

663: magnetic pressure $\langle \frac{1}{2} ||b||^{2} (\theta; r) \rangle_{A}$ and $\beta$ $\langle 2 P / ||b||^{2} (\theta; r) \rangle_{A}$ at $r=10M$.  All three magnetic topology cases

664: have very similar density profiles in the disk body, but appear to differ

665: in the corona.  Although our averaging period of $6000M$ is 30

666: orbital periods at the radius shown ($r = 10M$), it is not long

667: enough to erase completely fluctuations in the coronal density profile:

668: observe that the quadrupolar case is the most extended in the ``southern

669: hemisphere" corona while the toroidal case is the most extended in

670: the ``northern hemisphere".

671: 

672: Consistent with the general behavior we have already emphasized,

673: the gas and magnetic pressure profiles in the disk body are very

674: similar in the dipolar and quadrupolar cases, but both pressures are

675: rather smaller in the toroidal field case.  Nonetheless, all three cases

676: share one important property: the magnetic pressure is almost flat

677: for several scale-heights around the midplane.   This finding is

678: consistent with results from shearing-box simulations with much

679: better vertical resolution \citep{Hirose:2005,Hirose:2007}.

680: In the corona, the magnetic pressure drops rapidly with height in

681: QDPa and TDPa, but remains nearly constant in KDPg.

682: In all three cases, the magnetic and gas pressures track each other closely,

683: so that the $\beta$ profiles in the three runs are very similar.

684: 

685: Figure \ref{angtrptrt} plots the angular profiles of the electromagnetic

686: contributions to  $T^r_\phi$ and $T^r_t$ ($\langle |T^r_{t} |_{\; \mathrm{(EM)}} (\theta; r) \rangle_{A}$ and $\langle |T^r_{\phi} |_{\; \mathrm{(EM)}} (\theta; r) \rangle_{A}$ respectively) at two different radii,

687: $r/M=1.65$ and $r/M=10$.  At the larger radius, it should come

688: as no surprise that the profile of magnetic pressure shown at that

689: radius in Figure~\ref{angcor} accurately predicts the profile of these

690: two electromagnetic quantities.  At the smaller radius, as we have

691: already shown (Fig.~\ref{bndavgplt1}), the magnetic field is considerably

692: more intense in the dipole case than in the other two, so both

693: the electromagnetic angular momentum flux and energy flux follow

694: suit.

695: 

696: To summarize this section, we have found that, just as for the main

697: disk body, the character of disk coronae is only weakly dependent upon

698: the initial field topology.  There are modest quantitative contrasts

699: of the same sort as seen in the disk body---dipolar and quadrupolar

700: initial fields both create somewhat stronger magnetic fields on average

701: than does initial toroidal field---but that is all.

702: 

703: \section{Outflows}\label{jets}

704: 

705: \subsection{Global measures}

706: 

707: The formation of unbound outflows, or jets, is one of the most striking

708: features of the KD models \cite[]{De-Villiers:2003b}, all of which were

709: initialized with single-loop dipolar magnetic fields.  These outflows

710: generically have two components:  a Poynting flux jet inside the axial

711: funnel formed by the centrifugal barrier, and a sheath of much denser

712: unbound gas moving outward at sub-relativistic speed  along the funnel's

713: outer edge.  As described in \cite{Hawley:2006}, energy for the Poynting

714: flux comes from the rotation of the black hole, in a manner closely

715: related to the classical Blandford-Znajek mechanism.  In the case of

716: the Schwarzschild hole, the absence of black hole rotation means that

717: the magnetic field in the funnel, although relatively strong, is mostly

718: radial and carries no Poynting flux.  The funnel wall jet, on the other

719: hand, appears to be powered by the significant gas and magnetic pressure

720: near the ISCO.  The gas retains enough angular momentum to be excluded

721: from the funnel and is squeezed outward along the centrifugal barrier.

722: 

723: In contrast, the three-dimensional quadrupole and toroidal

724: field simulations  have substantially weaker unbound outflows, if any.

725: Figure \ref{ubndavgplt1} shows the time-averaged shell-integrals of the

726: mass outflow rate $\langle\dot{M}(r)\rangle_{F}$, magnetic field strength$\langle ||b||^2 (r) \rangle_{F}$; along with the electromagnetic energy flux, $\langle |T^r_{t} |_{\; \mathrm{(EM)}} (r) \rangle_{F}$ and angular momentum flux $\langle |T^r_{\phi} |_{\; \mathrm{(EM)}} (r) \rangle_{F}$ in unbound

727: material between $r=10$M and $r=100$M. The values in KDPg are 10--100

728: times larger than those in QDPa, whose outflow quantities are, in turn,

729: an order of magnitude larger than those in TDPa.  For example, the unbound

730: mass outflow rate at $r=100M$ in KDPg is $\simeq 25$ times that of QDPa,

731: while the mass outflow rate in TDPa is 10 times smaller than in QDPa.

732: The strength of the outflow is directly related to the strength of the

733: funnel field.  In KDPg the time-averaged funnel magnetic energy density

734: is 20--$400$

735: times greater than in the quadrupole model, while the toroidal model is

736: weaker by another factor of 10.  The mean Poynting flux in KDPg is therefore

737: $\sim 100$ times the Poynting flux for QDPa.  In other words, unlike

738: the qualitative similarity of the accretion flows produced by these

739: different field topologies, the outflows differ dramatically.  To

740: understand better the origin of this sharp contrast, in the next section

741: we focus more closely on exactly how field is introduced into the funnel,

742: and how its magnitude is controlled.

743: 

744: \subsection{Funnel Field Creation}\label{earlytimes}

745: 

746: We have seen that the strength of the Poynting flux jet is very sensitive

747: to the initial magnetic field topology.  Because the establishment of

748: large-scale field within the funnel appears to be critical to driving

749: a jet with substantial Poynting flux, we devote the remainder of this

750: section to a close examination of just how different magnetic topologies

751: lead to different funnel magnetic fields, leaning heavily on high-resolution

752: 2-d axisymmetric simulations.

753: 

754: \subsubsection{Dipole Topology}\label{dipole}

755: 

756: We begin by analyzing the axisymmetric simulation whose initial

757: condition has already been shown to produce a significant jet, the

758: dipole.  Figure~\ref{DFPevln} shows a time series of color plots

759: of the $\beta$ parameter overlaid with white contours depicting

760: poloidal field lines.  As the radial field is sheared, a toroidal

761: field (visible in Fig.~\ref{DFPevln} through its effect on $\beta$) is

762: created and amplified, leading to poloidal magnetic pressure gradients

763: \cite[]{De-Villiers:2003a}.  Because the initial radial field has opposite

764: sign on opposite sides of the equatorial plane, the toroidal field created

765: by the shear also changes sign across the equator.  The inner edge of

766: the torus begins to move inward toward the event horizon, dragging the

767: magnetic field with it.  As the gas moves inward, its rotation velocity

768: increases, further strengthening the toroidal field.  As the magnetic

769: pressure increases, gas is forced toward the equator where the field is

770: comparatively weak, bringing the regions of oppositely-directed toroidal

771: field closer together (see, e.g., the panel illustrating $t = 800M$).

772: As a result, a strong current sheet forms along the equatorial plane.

773: 

774: Once the inner edge of the accretion flow arrives at the black hole

775: ($t \sim 700M$), field lines attach to the event horizon, and gas

776: rapidly drains off them.  Field lines slide toward the poles as they

777: expand outward at relativistic speed, driven by the contrast between

778: the high magnetic pressure near the equatorial plane and the much lower

779: pressure higher up along the axis.   This flow moves outward so swiftly

780: because it is almost entirely unencumbered by the inertia of matter.

781: As discussed in \cite{Hawley:2006}, a magnetic tower forms, similar

782: to the one predicted on analytic grounds \citep{Lynden-Bell:2003} and

783: seen in other simulations \citep{Kato:2004}.  However, in contrast to

784: these models, tightly-wrapped field lines persist only briefly; the large

785: vertical gradient in magnetic pressure propels the upper loops so rapidly

786: outward that in the steady-state the field lines are better described as

787: loosely helical than as a tightly-wrapped coil.  The radial field inside

788: the funnel has the same sign as the vertical component of the initial

789: dipole loop inside the initial pressure maximum; these field lines close

790: outside of $r=100$M, reentering the problem volume throughout the corona.

791: 

792: A time history of the unbound Poynting flux at $r=100M$, normalized by

793: the instantaneous (coordinate frame) magnetic field energy within the

794: bound accretion flow, is shown in Figure~\ref{2dTrtdU} for both the

795: three-dimensional simulation KDPg and the high-resolution axisymmetric

796: model.  In both cases, once the jet is established

797: (a process taking $\lesssim 1000M$), the Poynting flux jet persists.

798: It is frequently perturbed, and the amplitude undergoes fluctuations,

799: but its basic properties persist throughout the simulation.  Excluding

800: initial transients (i.e. $t<1500M$), the jet histories of the two

801: simulations are very similar to one another, despite the differences

802: in dimensionality, (poloidal)

803: resolution, torus configuration and equation of state: in both cases,

804: the ratio of jet EM luminosity to energy density in magnetic energy

805: is $\simeq 0.1$--0.2 per time-unit. This similarity demonstrates

806: that, for this topology at least, the properties of the funnel field

807: depend primarily on the field strength within the initial accretion flow,

808: and that funnel field creation is a primarily axisymmetric process.

809: 

810: Although the initial magnetic field had structure only on the scale

811: of the disk thickness, the {\it essentially laminar} dynamics of initial

812: inflow lead quickly to the formation of truly large-scale field.  On the

813: other hand, because the accretion flow takes longer to be established

814: than the jet does, the magnetic field in the funnel acts in effect

815: as a permanent boundary condition for the accretion flow.  In other

816: words, these simulations mimic the long-term behavior of an accretion

817: flow in which there is large-scale net magnetic flux threading the

818: event horizon and filling a cone around the axis, but no large-scale

819: flux piercing the disk proper.  As both the three-dimensional and

820: two-dimensional simulations demonstrate, it is entirely possible for

821: long-term quasi-stationary accretion to coexist with a strong Poynting

822: flux jet whose foundation is magnetic flux trapped long before.

823: 

824: 

825: \subsubsection{Quadrupole Topology}\label{quadrupole}

826: 

827: We next consider the evolution of an initial quadrupolar field in high

828: resolution axisymmetry.  Figure \ref{QFPevln} shows a time series of plots

829: illustrating $\beta$ and the poloidal field line structure for this case.

830: Initially, the loops of weak poloidal magnetic field enclose two regions,

831: located above and below the pressure maximum (see Figure \ref{topologies},

832: center panel). In a similar fashion to the dipole case, these loops are

833: sheared out, creating ever growing toroidal field and producing strong

834: poloidal magnetic pressure gradients. In contrast to the dipole case,

835: however, two current sheets are formed, one above the equatorial plane,

836: one below.  The magnetic pressure forces gas toward these current sheets,

837: where the magnetic pressure has a local minimum.

838: 

839: As the inner edge of the torus moves radially inward toward the black

840: hole, the current sheets maintain their structure.  When the field

841: reaches the black hole, the current sheets are located above and below

842: the equator.  As with the dipole, gas drains off the field lines, which

843: anchor themselves in the event horizon,  and the field lines between the

844: current sheets and the axis slide toward the poles, where the density and

845: pressure are very low.  These field lines expand outward relativistically,

846: filling the funnel region and establishing a large-scale quadrupole field.

847: 

848: While this field is in place it supports a Poynting-flux jet, but the power

849: in the jet is significantly less than that in the axisymmetric dipolar

850: simulation.  Averaged from $t=2800M$ to $t^\prime = 3800M$, the EM flux

851: in the dipolar case is twice as strong as in the quadrupolar simulation

852: when measured at $r=10M$, and 10 times stronger measured at $r=100M$ (we

853: will discuss later why the contrast was more than an order of magnitude

854: greater in the 3-d simulations).  Figure~\ref{TrtdUcomp} compares

855: the dipole jet Poynting flux with that of the quadrupole simulation as a

856: function of time at $r=100M$, normalized by the total magnetic field

857: energy in the disk.  In terms of this normalized power, the jet in

858: the quadrupole simulation is a factor of

859: 2--10 weaker than that associated with dipole model; in other words,

860: much of the actual power contrast between the dipolar and quadrupolar

861: cases can be explained in terms of a reduction in this ``magnetic

862: efficiency".  Unlike the situation for accretion properties, contrasts

863: in jet luminosity from one topology to another cannot be described

864: solely in terms of contrasts in mean magnetic field strength.

865: 

866:    \cite{McKinney:2004} reported that replacing the dipole

867: with a quadrupole field led to a jet

868: that is  weaker by a factor of 2--3, but for several reasons one should

869: expect only qualitative, not quantitative agreement with our results.

870: They measured the jet luminosity at $r=40M$ and their initial pressure

871: maximum was much closer to the black hole ($r=12M$), so that the radial

872: dynamic range of their genuine accretion flow was considerably smaller.

873: Their disk was also geometrically thicker, partly because their code

874: solved a total energy equation, and partly because their initial state

875: was hotter.

876: 

877: Simple geometry accounts for the qualitative contrast between the jet

878: produced by an initially quadrupolar magnetic geometry and that driven

879: by an initially dipolar field.  In the quadrupolar case, relatively small

880: amplitude vertical oscillations in the inner region of the accretion flow

881: can lead to reconnection between magnetic field in the funnel and magnetic

882: field in the disk because vertical field of both signs is available in

883: the same hemisphere.  The result is to break the magnetic connection

884: between the black hole horizon and large radius.  Put another way, the

885: presence of closed field loops on the same side of the equator offers

886: the opportunity for easy collapse of these loops to zero.  In the dipole

887: case, however, the sign of the vertical magnetic field  passing through

888: the equator is the same everywhere inside $r=25M$.

889: %Although the accreting gas is similarly squeezed toward the

890: %equatorial plane, in that case gas pressure inhibits the reconnection

891: %of both radial and toroidal fields across the equator.

892: 

893: The net result of this geometric contrast is that, while quadrupolar

894: field configurations can pump magnetic field into the outflow funnel and

895: support a jet, they can also equally easily destroy such a jet by reconnection

896: with oppositely-directed field.  Jets in this case are consequently very

897: unsteady, and exhibit strong variability in the Poynting flux, as shown

898: in Figure~\ref{TrtdUcomp}.  This variability is a major factor in

899: depressing the time-averaged EM luminosity of jets driven by quadrupolar

900: fields.

901: 

902: Because the code simulates the equations of ideal MHD, reconnection is

903: numerical and occurs at the grid scale. As such, the reconnection rate

904: is resolution dependent (lower resolution promotes reconnection). We

905: cannot therefore make specific predictions regarding the variability

906: characteristics of jets in Nature, as this is determined (in our

907: simulations at least) by competition between processes that are

908: physical in character (e.g., orbital dynamics, magnetic forces) and

909: purely numerical (the reconnection rate).

910: 

911: The rate of reconnection is also determined by details of the field

912: geometry, in addition to its basic topology.  To investigate this sort

913: of dependence, we have conducted a series of lower resolution ($256^2$)

914: axisymmetric quadrupolar simulations in which we varied the size of the

915: initial torus, and hence the scale of the initial field loops.  We kept

916: the inner edge of the torus fixed at $r=15M$, but reduced the angular

917: momentum $\ell$ in order to reduce the vertical thickness of the

918: torus, and consequently reduce the vertical extent of the quadrupolar

919: loops. Simulations were carried out for a variety of values of $\ell$,

920: ranging from $\ell = 4.9$ to $\ell=5.2$ in steps of $\ell=0.05$.

921: 

922: For the purposes of our discussion here, we focus on the $\ell=4.9$

923: torus, which serves to illustrate the impact of vertical scale height on

924: the evolution of the quadrupole topology. Specifically, setting $\ell =

925: 4.9$ results in a torus which is initially a factor of two thinner (in

926: terms of the density scale height, measured along surfaces of constant

927: radius at the location of the pressure maximum) than the $\ell = 5.2$

928: torus previously discussed. The vertical thickness of the resulting

929: quadrupolar field loops is then very similar to that of loops in

930: the sub-Keplerian torus used in QDPa.

931: 

932: The initial evolution of this system is shown in Figure \ref{QNPevln}. The

933: evolution proceeds as before: current sheets form above and below

934: the equatorial plane as the inner edge of the torus moves toward the

935: black hole.  This time, however, the two current sheets are much closer to

936: the equatorial plane (contrast Fig.~\ref{QNPevln} with Fig.~\ref{QFPevln}).

937: Around $t=800M$ the current sheets buckle,

938: apparently due to an instability, and the coherent poloidal field

939: structure is destroyed.  A similar phenomenon was observed in QDPa,

940: which had a similar scale-height:

941: the current sheets destabilized far from the black hole, destroying the

942: coherent poloidal field structure. As a result, no large scale poloidal

943: field was dragged into the near horizon region by the torus, and the

944: resulting field inside the funnel region is substantially reduced.

945: We note that in three-dimensions the presence of non-axisymmetric MRI

946: modes are likely to enhance the rate of reconnection, further decreasing

947: the likelihood that an organized quadrupole configuration will survive

948: to create a significant funnel field.

949: 

950: These additional simulations show that the strength of Poynting-dominated

951: jets depends on field geometry in two senses: it is topology-dependent

952: and shape-dependent.  Quadrupolar magnetic topology leads to easier

953: reconnection, and therefore weaker funnel fields and jets, than does

954: dipolar topology, and geometrically thinner quadrupole loops make

955: reconnection still more rapid.

956: 

957: \subsubsection{Multiple Loop Topology}\label{multipleloop}

958: 

959: We have shown how large dipolar loops can create strong jets, but what

960: happens if the oppositely-directed vertical field on the far side

961: of the loop is accreted?  This did not occur in KDPg, in part because

962: it did not run long enough and in part because the finite mass in our

963: torus meant that the matter outside the pressure maximum moves outward

964: over time, not inward.   In a real accretion disk, however, continuing

965: mass re-supply will force the far sides of even large dipolar loops

966: to accrete.

967: 

968: To explore this sort of event, we ran a two-dimensional

969: simulation whose initial state contained a series of four narrow dipolar

970: loops (described in \S\ref{numerics}). The initial evolution of

971: this topology is shown in Figures \ref{MFPevln1} and \ref{MFPevln2}.

972: The first of these figures shows the descent of the innermost radial

973: loop into the near-horizon region and the establishment of a Poynting

974: flux jet. The second figure shows the descent of the second (moving out

975: in radius) of these loops and its interaction with the field structures

976: established by the first loop. 

977: 

978: Over the timescale considered, the evolution of this model is determined

979: by the two innermost loops, both initially inside the pressure maximum

980: (see Fig.~\ref{topologies}).  As for a single dipole loop, the first loop

981: forms a funnel dipole field by $t=700M$.  However, its shape differs

982: significantly from the single loop because the second loop begins its

983: descent toward the inner disk even before the inner edge of the first

984: loop nears the event horizon (Fig.~\ref{MFPevln1}).  As the second loop

985: moves inward, it displaces the field lines of the first loop, forcing

986: the midplane portions of the outer half of the first loop inward, and

987: the off-plane portions of the first loop to higher altitude at their

988: original radius.  The result is that the first loop is very nearly

989: transformed into a pair of loops, one on each side of the midplane.

990: After arriving in the inner disk, the field of the second loop expands

991: rapidly toward the funnel region.  It pushes on the almost closed

992: sections of the first loop, triggering rapid reconnection where its

993: oppositely-directed segments pass very close to one another at small

994: radius (Fig.~\ref{MFPevln2}).  At this point, the transformation of the

995: first loop into a disconnected quadrupolar loop pair is complete, and the

996: loops fly off to large distance, emptying the funnel of magnetic field.

997: The funnel is then refilled with flux from the second

998: loop.  These dramatic events are mirrored in the very large fluctuations

999: in normalized Poynting luminosity shown

1000: in Figure~\ref{TrtdUcomp}.  Jumps in output even greater in relative

1001: terms than those displayed by the quadrupolar simulation are frequent,

1002: and can be associated with these episodes of funnel field establishment

1003: and annihilation.  Note that while the multiple-loop jet is highly

1004: variable, similar to the quadrupolar jet, the instantaneous

1005: strength of the Poynting-flux jet (when it is strong) is similar to

1006: that of model KDPg when normalized for disk magnetic field strength.

1007: In this sense, the ``efficiency" of these systems in using magnetic

1008: field generated in the disk for jets appears to be a function of

1009: poloidal field character, declining as the relative importance

1010: of higher multipole moments increases.

1011: 

1012: In interpreting this simulation, it is important to bear in mind that,

1013: unlike any of the others reported in this paper, its accretion flow never

1014: achieves anything resembling a quasi-stationary state.  The multiple

1015: loop structure also requires high resolution to prevent reconnection from

1016: occurring at too rapid a rate.  It is for this reason that, unlike

1017: the large dipolar loop and quadrupolar caess, we report no 3-d

1018: simulation results for this field geometry.  Thus, we can be less

1019: confident than

1020: in the other cases that the specific phenomenology we observe may

1021: be associated with real systems in a state of continuing accretion.

1022: What it {\it does} demonstrate, however, is that accretion of a series

1023: of closed dipolar loops can lead to a sequence of field establishment

1024: and destruction rather similar to what was seen in the quadrupolar case,

1025: and through mechanisms that also resemble those seen when the initial

1026: field is quadrupolar.  It also supports our qualitative inference that

1027: maintenance of a consistent sign of vertical field is essential to

1028: the long-term support of a Poynting-dominated jet.  The lifetime for

1029: an episode of strong outflow is roughly the inflow time for the outer

1030: anchor point of the loop.

1031: 

1032: \subsubsection{Toroidal Field Topology}\label{toroidal}

1033: 

1034: In the poloidal cases, magnetic field is injected into the funnel by

1035: strong toroidal field pressure in the plunging region.  One might expect

1036: that this process could act equally efficiently when the magnetic field

1037: is toroidal from the start.  If the toroidal field has a consistent sign,

1038: the radial component created when it expands outward would also have a

1039: consistent sign, satisfying the coherent vertical field criterion we

1040: have inferred from study of the poloidal simulations.  If so, the toroidal

1041: field case should also drive a powerful jet.  However, it does

1042: nothing of the sort.

1043: 

1044: Figures~\ref{TDPevln1} -- \ref{TDPevln2} are a series of snapshots of

1045: the $\beta$ parameter and the radial field structure in

1046: the evolving torus of simulation TDPa.  In contrast to the poloidal

1047: loop simulations, there is no initial radial field to generate toroidal

1048: field through shear.  Instead, the evolution is governed entirely by

1049: the development of the nonaxisymmetric MRI.  The fastest-growing modes

1050: for the MRI in this context have large $k_z$ and azimuthal wavelength

1051: $\sim v_A/\Omega$, which is much shorter than the disk thickness when the

1052: field is weak \citep{Balbus:1992}.  The magnetic structures that develop

1053: are then predominantly on short lengthscales.  Note the rapid spatial

1054: variation in sign in the radial field in Figure~\ref{TDPevln2}.  With

1055: time, the radial field features become somewhat more radially elongated,

1056: but field lines do not connect extended regions;  no large-scale coherent

1057: poloidal field is generated from the initial toroidal field.  Magnetic

1058: field amplification is less than in the poloidal field initial conditions,

1059: and $\beta$ never becomes less than 1 near the hole.  Matter tends to

1060: remain with the field rather than drain either into the hole or toward

1061: a localized region such as the equator, as seen in the dipole case.

1062: Any vertical expansion of the field would therefore have to overcome the

1063: inertia of the matter and, as can be seen from Figure \ref{TDPevln2},

1064: this does not occur.

1065: 

1066: The absence of a funnel-field in the toroidal topology follows directly

1067: from the specific properties of the toroidal field MRI: the poloidal

1068: fields created by the MRI are typically small scale.  No radially-extended

1069: poloidal field connects the near horizon region to the disk.  What radial

1070: field there is is accreted faster than it can tap into the differential

1071: rotation and generate strong near-hole fields.  The main disk-body in

1072: the toroidal case remains gas pressure-dominated at all times, and there

1073: are no strong magnetic pressure gradients to drive an expansion of the

1074: field into the funnel region.  The absence of a funnel field means there

1075: is no Poynting flux jet.

1076: 

1077: It appears, then, that creation of a strong funnel field, and therefore

1078: a strong Poynting jet when the black hole spins rapidly, depends

1079: essentially on the existence of some coherent poloidal field in the

1080: initial state.  Purely toroidal field initial conditions can lead to

1081: short coherence-length poloidal loops with rapidly alternating sign,

1082: but no large-scale poloidal field components.

1083: 

1084: \section{Summary and Conclusions}\label{conclusion}

1085: 

1086: Through a series of numerical studies, we have examined the influence

1087: of the initial field topology on the properties of magnetized black hole

1088: accretion flows. In particular, we have performed full three-dimensional

1089: simulations of three different initial field topologies: a large-scale

1090: dipolar loop, a pair of quadrupolar loops separated by the equatorial

1091: plane, and a purely toroidal field.  Apart from the imposed initial field

1092: topologies, the initial conditions used in each of these simulations

1093: were identical: a $\Gamma = 5/3$ hydrostatic torus with a slightly

1094: sub-Keplerian distribution of angular momentum around a rotating

1095: ($a/M=0.9$) black hole.  The two poloidal field topologies were evolved

1096: to $t=10000$M, while the toroidal case was evolved to $t=29000$M.

1097: 

1098: \cite{De-Villiers:2003b} described the late-time structure of the

1099: resulting accretion flow in terms of distinct components: the main

1100: disk body, the coronal envelope, and an unbound outflow which

1101: itself could be separated into a matter-dominated funnel-wall jet

1102: and a Poynting-flux jet in the funnel interior.  We have found that

1103: the late-time averaged properties of the main disk body exhibit only

1104: modest quantitative contrasts attributable to different initial field

1105: topologies: the dipolar and quadrupolar cases are roughly similar, and

1106: both create somewhat stronger magnetic fields than the toroidal case.

1107: So long as there is some weak magnetic field present, the MRI generates

1108: turbulence and drives accretion, although the character of the purely

1109: toroidal MRI is somewhat different from that produced by poloidal fields.

1110: Because the Maxwell stress controls the surface density required to

1111: support a given accretion rate, and the cascade of energy to smaller

1112: scales in the MHD turbulence leads ultimately to disk heating, we expect

1113: that the radiation emitted by the disk will likewise be only weakly

1114: dependent on magnetic topology.  It is possible that different magnetic

1115: topologies could produce different vertical profiles of dissipation

1116: (which these simulations are not designed to study), but the similarity

1117: in magnetic pressure profiles (Fig.~\ref{angcor}) suggests that even

1118: those differences are likely to be minor.

1119: 

1120: In regard to the accretion flow, we found only two dependences on field

1121: topology worth noting: for fixed outer torus mass, the initial toroidal

1122: field simulation gives a somewhat lower accretion rate \cite[a result

1123: foreshadowed in][] {HK:2002}.  In addition, the magnetic field in the

1124: plunging region is, in relative terms, somewhat stronger in the large

1125: dipolar loop case than for quadrupolar or toroidal fields, so that

1126: stresses near the ISCO are somewhat stronger when the field topology

1127: is dipolar.  In other words, the short-wavelength behavior of the

1128: MRI-driven MHD turbulence in the disk produces near-universal behavior,

1129: independent of magnetic field structure.

1130: 

1131: Jets are quite different.  {\it Laminar} flow dynamics can, given the

1132: right conditions (e.g., the right field topology), inflate field with no

1133: pre-existing structure on scales larger than a disk scale-height into

1134: a truly global field.  Simulations that begin with a large dipole loop

1135: in the initial torus produce a strong Poynting-flux jet in the axial

1136: funnel and a comparably strong matter-dominated funnel-wall outflow.

1137: The inner field lines of the dipolar loop move quickly inward to fill the

1138: funnel, and remain there until the opposite side of the loop is accreted,

1139: an event which may not happen until quite a long time later. Thus,

1140: this case mimics a situation in which net magnetic flux has been trapped

1141: against the black hole event horizon by previous accretion.  When this

1142: trapped flux creates a coherent poloidal field in the funnel, the black

1143: hole's spin forces nearby field lines into rotation and drives a large

1144: outward-directed Poynting flux.

1145: 

1146: This sequence of events is not as easily achieved with any

1147: of the other field geometries we explored.   Unlike the dipole

1148: configuration, the quadrupole

1149: features a pair of current sheets located above and below the equator.

1150: These current sheets are subject to instabilities that can lead to

1151: reconnection in the magnetic field and loss of extended field coherence.

1152: Even when the quadrupole is successful in establishing a funnel field,

1153: the jet may live only briefly because fluctuations in the accretion

1154: flow can generate reconnection events.  By the same token, quadrupolar

1155: field can also restore field to an empty funnel.   Consequently, any

1156: jet based on quadrupolar field tends to be unstable and episodic, with

1157: the timescale and amplitude of changes set both by details of the field

1158: geometry and the rate of reconnection.  Unfortunately, it is very hard

1159: to obtain any information about the former in real systems, and in our

1160: simulations the latter is controlled by numerical effects and

1161: resolution, so it is impossible for us to make predictive statements

1162: about the timescale or amplitude of jet variability in this case.

1163: 

1164: All of the models reported on above used a black hole with a spin $a/M

1165: = 0.9$.  To test our conclusions against other black hole spins we

1166: also ran two three-dimensional initially quadrupolar simulations

1167: with black hole spins $a/M=0.0$ and, $a/M =0.998$.  The effect of a wide

1168: range of black hole spins (including a counter-rotating hole) on dipolar

1169: models was studied by \cite{Hawley:2006}.  No Poynting

1170: flux jet is present for a Schwarzschild hole regardless of the initial

1171: field topology.  The energy to drive the Poynting flux originates with

1172: the spin of the hole.  However, the funnel is still filled with radial

1173: field, which is considerably stronger in the dipole case compared to

1174: the quadrupole.  On the other hand, the funnel field strength increases

1175: with black hole spin for a given field topology.

1176: For the rapidly-spinning hole, the Poynting flux jet is stronger than

1177: in the $a/M = 0.9$ simulations for both topologies, but as we have discussed,

1178: it is considerably weaker when the initial field is quadrupolar than when

1179: it is dipolar.  Specifically, the ratio between the time-averaged Poynting

1180: luminosity with dipolar field to the Poynting luminosity with quadrupolar

1181: field is $\sim 350$ for $a/M = 0.998$ and $\sim

1182: 335$ for $a/M = 0.9$.  Thus, the accretion dynamics control what sort of

1183: field is brought to the vicinity of the black hole and fill the funnel,

1184: and the black hole spin determines how energetic the resulting jet is.

1185: 

1186: The purely toroidal field topology does not generate any

1187: jet at all.  The only way for toroidal field to create any poloidal field

1188: is by initiating non-axisymmetric perturbations of alternating sign.

1189: The lengthscales of these perturbations must be relatively short, and

1190: the portions of opposite sign must balance in magnitude.  Fields of this

1191: sort can do very little to fill the funnel region because they are even

1192: more subject to reconnection than is the quadrupolar field.

1193: 

1194: When the initial field is a single dipolar loop that wraps around the

1195: initial pressure maximum of the simulation, all the vertical field brought

1196: down to the funnel by the initial accretion has the same direction.

1197: In real disks containing closed dipolar loops, oppositely directed

1198: vertical field will, sooner or later, arrive in the inner disk, and

1199: it may cancel the funnel field.  To investigate this possibility,

1200: we ran a high-resolution axisymmetric simulation that began with

1201: four narrow dipolar loops within the initial torus.  Its subsequent evolution

1202: qualitatively supports this conjecture.  The innermost loop produces a

1203: funnel field as in the single dipole loop model.  When the second loop,

1204: oriented oppositely from the first one, descends, it displaces the field

1205: established by the first loop.  The funnel field established by the first

1206: loop reconnects, detaches from the event horizon, and is rapidly expelled

1207: from the grid.  The second loop then expands poloidally and reestablishes

1208: the funnel field.  In other words, the accretion of dipolar field

1209: loops leads to jet creation, but its structure is highly-variable:

1210: the narrower the dipolar loops present in the accretion flow, the more

1211: variability we can expect in the jet.

1212: 

1213: When a simulation is initialized with (possibly multiple) dipolar field

1214: loop(s) that cross the equatorial plane, the strength of the resulting

1215: Poynting-flux jet depends on the strength of the magnetic field within

1216: the disk itself and on the spin of the hole.  Poynting-flux jets that

1217: are formed from quadrupolar fields are generically much weaker than

1218: their dipolar counterparts, even when magnetic field strength

1219: in the disk body is taken into account.  How much weaker depends on

1220: other factors as well, including disk thickness, location of the inner

1221: edge of the initial torus, and numerical resolution.  In the limited

1222: range of numerical experiments reported here, we have seen suppression

1223: factors ranging from several to several hundred.

1224: 

1225: Real accretion flows are likely to have a mixture, possibly time-varying,

1226: of all these field topologies.  Real jets from black holes will exhibit

1227: behavior combining the sorts of events we have seen in this paper:

1228: stability over longer timescales when the field is mostly made of

1229: large dipolar

1230: loops, greater variability and occasional jet quenching when the field

1231: has a larger contribution from higher multipoles, jet elimination when

1232: the field is mostly toroidal.  It is possible that in addition

1233: to the sort of coherent funnel field spontaneously created by large

1234: closed-loop poloidal structures, there is a large-scale field imposed

1235: by external boundary conditions.  We plan to investigate the impact of

1236: such an externally-imposed field in future work.

1237: 

1238: This study is only one small step in the process of exploring

1239: the mechanisms to create large scale field configurations within

1240: accretion flows.  Here the large scale field formation process we have

1241: observed results from an interaction between poloidal field components

1242: in the accretion flow and the boundary represented by the black hole.

1243: The mechanism operates in axisymmetry as well as in three dimensions, so

1244: it is not a dynamo if that word is defined as in the antidynamo theorem.

1245: Many authors have conjectured that an inverse cascade within the three

1246: dimensional turbulent disk itself might generate large scale fields; we

1247: see no evidence in these simulations of any disk field with structure on

1248: scales more than a few scale-heights, but we caution that the resolution

1249: limitations of three dimensional global simulations make this observation

1250: less than definitive.

1251: 

1252: The results we have reported also stimulate speculations about whether

1253: the contrasts in jet phenomenology from one object, or class of objects,

1254: to another is due in part to differences in the mixture of different

1255: field topologies, or the interaction of different topologies in the flow.

1256: For example, could the jet outbursts sometimes seen in Galactic black

1257: hole binaries when they cross the ``jet line" \citep{RM:2006} be due to

1258: rapid field annihilation and loop expulsion of the sort observed in the

1259: multiple dipolar loop simulation?  Perhaps the radio-weakness of most AGN

1260: \cite[see e.g.][]{White:2007} is due to the predominance of small-scale

1261: multipolar field in their accretion flows as a result of accreting from a

1262: turbulent interstellar medium?  As we learn more about how it acts, we may

1263: find that magnetic topology can, in fact, be inferred from phenomenology.

1264: 

1265: \acknowledgements{ This work was supported by NSF grant PHY-0205155 and

1266: NASA grant NNG04GK77G (JFH), and by NSF grant AST-0507455 (JHK). KB thanks

1267: Sean Matt for useful discussions. We acknowledge Jean-Pierre De~Villiers

1268: for continuing collaboration in the development of the algorithms used

1269: in the GRMHD code.  The simulations were carried out on the DataStar

1270: system at SDSC.}

1271: 

1272: \begin{thebibliography}{28}

1273: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi

1274: 

1275: \bibitem[{{Balbus} \& {Hawley}(1992)}]{Balbus:1992}

1276: {Balbus}, S.~A., \& {Hawley}, J.~F. 1992, \apj, 400, 610

1277: 

1278: \bibitem[{{Balbus} \& {Hawley}(1998)}]{Balbus:1998}

1279: ---. 1998, Reviews of Modern Physics, 70, 1

1280: 

1281: \bibitem[{{Blandford} \& {Payne}(1982)}]{Blandford:1982}

1282: {Blandford}, R.~D., \& {Payne}, D.~G. 1982, \mnras, 199, 883

1283: 

1284: \bibitem[{{Blandford} \& {Znajek}(1977)}]{Blandford:1977}

1285: {Blandford}, R.~D., \& {Znajek}, R.~L. 1977, \mnras, 179, 433

1286: 

1287: \bibitem[{{De Villiers} \& {Hawley}(2003{\natexlab{a}})}]{De-Villiers:2003}

1288: {De Villiers}, J., \& {Hawley}, J.~F. 2003{\natexlab{a}}, \apj, 589, 458

1289: 

1290: \bibitem[{{De Villiers} \& {Hawley}(2003{\natexlab{b}})}]{De-Villiers:2003a}

1291: ---. 2003{\natexlab{b}}, \apj, 592, 1060

1292: 

1293: \bibitem[{{De Villiers} {et~al.}(2003){De Villiers}, {Hawley}, \&

1294:   {Krolik}}]{De-Villiers:2003b}

1295: {De Villiers}, J., {Hawley}, J.~F., \& {Krolik}, J.~H. 2003, \apj, 599, 1238

1296: 

1297: \bibitem[{{De Villiers} {et~al.}(2005){De Villiers}, {Hawley}, {Krolik}, \&

1298:   {Hirose}}]{De-Villiers:2005}

1299: {De Villiers}, J., {Hawley}, J.~F., {Krolik}, J.~H., \& {Hirose}, S. 2005,

1300:   \apj, 620, 878

1301: 

1302: \bibitem[{{De Villiers}(2006)}]{De-Villiers:2006}

1303: {De Villiers}, J.-P. 2006, ArXiv Astrophysics e-prints, astro-ph

1304: 

1305: \bibitem[{{Gammie}(1999)}]{Gammie:1999}

1306: {Gammie}, C.~F. 1999, \apjl, 522, L57

1307: 

1308: \bibitem[{{Gammie} {et~al.}(2004){Gammie}, {Shapiro}, \&

1309:   {McKinney}}]{Gammie:2004}

1310: {Gammie}, C.~F., {Shapiro}, S.~L., \& {McKinney}, J.~C. 2004, \apj, 602, 312

1311: 

1312: \bibitem[{{Hawley} \& {Krolik}(2002)}]{HK:2002}

1313: {Hawley}, J.~F., \& {Krolik}, J.~H. 2002, \apj, 566, 164

1314: 

1315: \bibitem[{{Hawley} \& {Krolik}(2006)}]{Hawley:2006}

1316: ---. 2006, \apj, 641, 103

1317: 

1318: \bibitem[{{Hirose} {et~al.}(2004){Hirose}, {Krolik}, {De Villiers}, \&

1319:   {Hawley}}]{Hirose:2004}

1320: {Hirose}, S., {Krolik}, J.~H., {De Villiers}, J., \& {Hawley}, J.~F. 2004,

1321:   \apj, 606, 1083

1322: 

1323: \bibitem[{{Hirose} {et~al.}(2005){Hirose}, {Krolik}, \& {Stone}}]{Hirose:2005}

1324: {Hirose}, S., {Krolik}, J.~H., \& {Stone}, J.~M. 2005, ArXiv Astrophysics

1325:   e-prints

1326: 

1327: \bibitem[{{Jackson}(1975)}]{Jackson:1975}

1328: {Jackson}, J.~D. 1975, {Classical electrodynamics} (92/12/31, New York: Wiley,

1329:   1975, 2nd ed.)

1330: 

1331: \bibitem[{{Kato} {et~al.}(2004){Kato}, {Mineshige}, \& {Shibata}}]{Kato:2004}

1332: {Kato}, Y., {Mineshige}, S., \& {Shibata}, K. 2004, \apj, 605, 307

1333: 

1334: \bibitem[{{Krolik}(1999)}]{Krolik:1999a}

1335: {Krolik}, J.~H. 1999, \apjl, 515, L73

1336: 

1337: \bibitem[{{Krolik} {et~al.}(2005){Krolik}, {Hawley}, \& {Hirose}}]{Krolik:2005}

1338: {Krolik}, J.~H., {Hawley}, J.~F., \& {Hirose}, S. 2005, \apj, 622, 1008

1339: 

1340: \bibitem[{{Krolik} {et~al.}(2007){Krolik}, {Hirose}, \& {Blaes}}]{Hirose:2007}

1341: {Krolik}, J.~H., {Hirose}, S., \& {Blaes}, O. 2007, \apj, 664, 1045

1342: 

1343: \bibitem[{{Lynden-Bell}(2003)}]{Lynden-Bell:2003}

1344: {Lynden-Bell}, D. 2003, \mnras, 341, 1360

1345: 

1346: \bibitem[{{McKinney} \& {Gammie}(2004)}]{McKinney:2004}

1347: {McKinney}, J.~C., \& {Gammie}, C.~F. 2004, \apj, 611, 977

1348: 

1349: \bibitem[{{McKinney} \& {Narayan}(2007{\natexlab{a}})}]{McKinney:2007}

1350: {McKinney}, J.~C., \& {Narayan}, R. 2007{\natexlab{a}}, \mnras, 375, 513

1351: 

1352: \bibitem[{{McKinney} \& {Narayan}(2007{\natexlab{b}})}]{McKinney:2007a}

1353: ---. 2007{\natexlab{b}}, \mnras, 375, 531

1354: 

1355: \bibitem[{{Moffatt}(1978)}]{Moffatt:1978}

1356: {Moffatt}, H.~K. 1978, {Magnetic field generation in electrically conducting

1357:   fluids} (Cambridge, England, Cambridge University Press, 1978.~353 p.)

1358: 

1359: \bibitem[{{Novikov} \& {Thorne}(1973)}]{Novikov:1973}

1360: {Novikov}, I.~D., \& {Thorne}, K.~S. 1973, in {Black Holes: Les Astres Occlus},

1361:   ed. C.~{DeWitt} \& B.~{DeWitt} (New York; Gordon and Breach)

1362: 

1363: \bibitem[{{Remillard} \& {McClintock}(2006)}]{RM:2006}

1364: {Remillard}, R.~A., \& {McClintock}, J.~E. 2006, \araa, 44, 49

1365: 

1366: \bibitem[{{White} {et~al.}(2007){White}, {Helfand}, {Becker}, {Glikman}, \& {de

1367:   Vries}}]{White:2007}

1368: {White}, R.~L., {Helfand}, D.~J., {Becker}, R.~H., {Glikman}, E., \& {de

1369:   Vries}, W. 2007, \apj, 654, 99

1370: 

1371: \end{thebibliography}

1372: 

1373: \begin{figure*}

1374: \leavevmode

1375: \begin{center}

1376: \includegraphics{f1}

1377: \end{center}

1378: \caption[]{Initial configurations of dipole (left panel), quadrupole

1379: (center panel) and multiple loop (right panel) field topologies. The torus for

1380: the multiple loop topology is shown slightly zoomed to illustrate better the

1381: initial field structure.  White contours denote magnetic field lines, color contours

1382: the gas $\beta$ parameter.  Solid versus dashed lines indicate field polarity; solid

1383: lines denote current into the page, dashed lines current out of the page.

1384: }

1385: \label{topologies} 

1386: \end{figure*}

1387: 

1388: \begin{figure*}

1389: \begin{center}

1390: \leavevmode

1391: \includegraphics[width=0.44\textwidth]{f2a}

1392: \includegraphics[width=0.44\textwidth]{f2b}

1393: \includegraphics[width=0.44\textwidth]{f2c}

1394: \includegraphics[width=0.44\textwidth]{f2d}

1395: \includegraphics[width=0.44\textwidth]{f2e}

1396: \includegraphics[width=0.44\textwidth]{f2f}

1397: \end{center}

1398: \caption[]{Time-averaged, shell-integrals of the accretion rate, $\dot{M}$ (top left panel), surface density $\Sigma(r)$ \cite[top right panel, dot-dash lines show the surface density distribution predicted by][]{Novikov:1973}, the net accreted angular momentum

1399: per unit rest mass, $L$, (center left panel; dot-dashed line indicates value at the ISCO), magnetic energy density $\langle ||b||^2 (r) \rangle_{F}$ (center right panel); the electromagnetic contribution to the energy flux $\langle |T^r_{t} |_{\; \mathrm{(EM)}} (r) \rangle_{F}$ (bottom left panel) and angular

1400: momentum flux $\langle |T^r_{\phi} |_{\; \mathrm{(EM)}}  (r) \rangle_{F}$ (bottom right panel) in bound material.  Each figure shows data from KDPg (black lines), QDPa

1401: (blue lines) and TDPa (magenta lines) time-averaged over $4000-10000$M

1402: for KDPg and QDPa, and $12500-18500$M for TDPa. Solid lines denote the

1403: time average, dashed lines (where shown) $\pm1$ standard deviation from the

1404: average. Note that the radial coordinate of the ISCO is $2.32M$ for the black

1405: hole spin $a/M = 0.9$.}

1406: \label{bndavgplt1} 

1407: \end{figure*}

1408: 

1409: \begin{figure*}

1410: \begin{center}

1411: \leavevmode

1412: \includegraphics[width=0.32\textwidth]{f3a}

1413: \includegraphics[width=0.32\textwidth]{f3b}

1414: \includegraphics[width=0.32\textwidth]{f3c}

1415: \end{center}

1416: \caption[]{Time-averaged ratio of the accretion rate $\dot{M}$ (left

1417: panel), the electromagnetic contribution to the energy

1418: flux $\langle |T^r_{t} |_{\; \mathrm{(EM)}} (r) \rangle_{F}$ (center panel) and angular

1419: momentum flux $\langle |T^r_{\phi} |_{\; \mathrm{(EM)}}  (r) \rangle_{F}$ (right panel) to the magnetic field strength $\langle ||b||^2 (r) \rangle_{F}$; all are restricted to bound material.  Each figure shows data from KDPg (black lines), QDPa

1420: (blue lines) and TDPa (magenta lines).  The time-averaging is over the

1421: period $4000-10000$M for KDPg and QDPa, and $12500-18500$M for TDPa. }

1422: \label{ratio} 

1423: \end{figure*}

1424: 

1425: \begin{figure*}

1426: \begin{center}

1427: \leavevmode

1428: \includegraphics[width=0.32\textwidth]{f4a}

1429: \includegraphics[width=0.32\textwidth]{f4b}

1430: \includegraphics[width=0.32\textwidth]{f4c}

1431: \end{center}

1432: \caption[]{Time-averaged fluid-frame Maxwell stress, ${\cal W}^{(r)}_{(\phi)}$

1433: (solid lines) for simulations KDPg (left panel), QDPa (center panel) and

1434: TDPa (right panel). Solid lines denote the time average, dotted lines

1435: $\pm1$ standard deviation from the average. The Novikov-Thorne prediction

1436: of the fluid-frame stress for a thin disk with an accretion rate

1437: equal to the time-average for that simulation is shown with a dashed line. KDPg and QDPa are averaged over 8000--$10000 M$ and TDPa is

1438: averaged over 2--$2.2\times 10^4 M$.}

1439: \label{maxwell} 

1440: \end{figure*}

1441: 

1442: \begin{figure*}

1443: \begin{center}

1444: \leavevmode

1445: \includegraphics{f5}

1446: \end{center}

1447: \caption[]{Color contours of the time- and azimuthally-averaged

1448: $\beta$ parameter for simulations KDPg (left panel), QDPa (center panel)

1449: and TDPa (right panel), overlaid with white contours of gas

1450: density.  KDPg and QDPa are averaged over 8000--$10000 M$ and TDPa is

1451: averaged over 2--$2.2\times 10^4 M$.}

1452: \label{beta} 

1453: \end{figure*}

1454: 

1455: \begin{deluxetable}{cccccccc}

1456: %\rotate

1457: \tablecolumns{9}

1458: \tablewidth{0pc}

1459: \tablecaption{Coronal Diagnostics}

1460: \tablehead{\colhead{Model}          &

1461: \colhead{$\langle M_\mathrm{c} \rangle_{V} / \langle M_\mathrm{tot} \rangle_{V}$} &

1462: \colhead{$\langle \beta_{\mathrm{c}} \rangle_{V}$} &

1463: \colhead{$\langle E^\mathrm{therm}_{\mathrm{c}} \rangle_{V} / \langle M_\mathrm{c} \rangle_{V}$} &

1464: \colhead{$\langle E^\mathrm{mag}_{\mathrm{c}} \rangle_{V}  / \langle M_\mathrm{c}\rangle_{V}$} &

1465: \colhead{$\langle\mathrm{Vol_{c}}\rangle_{V} / \langle\mathrm{Vol_{tot}}\rangle_{V}$}

1466: }

1467: \startdata

1468: KDPg & 0.055 &  3.6 & $2.9\times 10^{-3}$&$1.1\times 10^{-3}$& 0.48 \\

1469: QDPa & 0.050 &  2.1 & $2.7\times 10^{-3}$&$1.8\times 10^{-3}$& 0.59 \\

1470: TDPa & 0.054 &  3.4 & $2.5\times 10^{-3}$&$1.0\times 10^{-3}$& 0.61

1471: \enddata

1472: \label{corona}

1473: \tablecomments{Subscript ``c'' denotes volume integral over the coronal region (as defined in the text), whereas subscript ``tot'' denotes volume integral over the total simulation region.}

1474: \end{deluxetable}

1475: 

1476: \begin{figure*}

1477: \begin{center}

1478: \leavevmode

1479: \includegraphics[width=0.48\textwidth]{f6a}

1480: \includegraphics[width=0.48\textwidth]{f6b}

1481: \includegraphics[width=0.48\textwidth]{f6c}

1482: \includegraphics[width=0.48\textwidth]{f6d}

1483: \end{center}

1484: \caption[]{Time-averaged angular profiles of density $\langle \rho (\theta; r=10M) \rangle_{A}$, gas

1485: pressure $\langle P (\theta; r=10M) \rangle_{A}$, magnetic pressure $\langle \frac{1}{2} ||b||^{2} (\theta; r=10M) \rangle_{A}$ and $\beta$ parameter $\langle 2 P / ||b||^{2} (\theta; r=10M) \rangle_{A}$ for simulations

1486: KDPg (black lines), QDPa (blue lines) and TDPa (magenta lines).

1487: Dotted lines denote the (approximate) boundary of the disk and corona.

1488: Note that in the case of TDPa and QDPa, the region of bound matter 

1489: extends all the way to the axial cutout. KDPg and QDPa are averaged over 8000--$10000 M$ and TDPa is

1490: averaged over 2--$2.2\times 10^4 M.$}

1491: \label{angcor} 

1492: \end{figure*}

1493: 

1494: \begin{figure*}

1495: \begin{center}

1496: \leavevmode

1497: \includegraphics[width=0.48\textwidth]{f7a}

1498: \includegraphics[width=0.48\textwidth]{f7b}

1499: \includegraphics[width=0.48\textwidth]{f7c}

1500: \includegraphics[width=0.48\textwidth]{f7d}

1501: \end{center}

1502: \caption[]{Time-averaged angular profiles of $\langle |T^r_{t} |_{\; \mathrm{(EM)}} (\theta; r) \rangle_{A}$ (top row) and $\langle |T^r_{\phi} |_{\; \mathrm{(EM)}} (\theta; r) \rangle_{A}$ (bottom row) for

1503: simulations KDPg (black lines), QDPa (blue lines) and TDPa (purple

1504: lines) at $r=1.65$M (left column) and $r=10$M (right column). Dashed

1505: lines denote the (approximate) boundary between the disk body and

1506: the corona. KDPg and QDPa are averaged over 8000--$10000 M$ and TDPa is

1507: averaged over 2--$2.2\times 10^4 M.$

1508: }

1509: \label{angtrptrt} 

1510: \end{figure*}

1511: 

1512: \begin{figure*}

1513: \begin{center}

1514: \leavevmode

1515: \includegraphics[width=0.48\textwidth]{f8a}

1516: \includegraphics[width=0.48\textwidth]{f8b}

1517: \includegraphics[width=0.48\textwidth]{f8c}

1518: \includegraphics[width=0.48\textwidth]{f8d}

1519: \end{center}

1520: 

1521: \caption[]{Time-average shell integrals of data from unbound outflows as a function of radius for models KDPg (black lines), QDPa (blue lines), and TDPa (magenta lines).

1522: Shown are mass outflow rate $\dot{M}$ (top left panel), $\langle ||b||^2 (r) \rangle_{F}$

1523: (top right panel); along with the electromagnetic energy flux, $\langle |T^r_{t} |_{\; \mathrm{(EM)}} (r) \rangle_{F}$

1524: (bottom left panel) and angular momentum flux $\langle |T^r_{\phi} |_{\; \mathrm{(EM)}} (r) \rangle_{F}$ (bottom right panel).  KDPg and QDPa are time-averaged over $4000-10000$M while TDPa is averaged over $12500-18500$M.  Dashed lines show $\pm1$ standard deviation from the average.}

1525: 

1526: \label{ubndavgplt1} 

1527: \end{figure*}

1528: 

1529: \clearpage

1530: 

1531: \begin{figure*}

1532: \begin{center}

1533: \leavevmode

1534: \includegraphics{f9}

1535: \end{center}

1536: \caption[]{Evolution of the magnetic field (white contours) and gas

1537: $\beta$ parameter (filled contours) at $t = 500,700,750,800,850,900$M

1538: for the high-resolution, axisymmetric simulation of the dipole field

1539: topology. Dark red denotes regions of the simulation that are gas

1540: pressure dominated and dark blue denotes regions that are magnetic

1541: field dominated.}

1542: \label{DFPevln} 

1543: \end{figure*}

1544: 

1545: \begin{figure}

1546: \begin{center}

1547: \leavevmode

1548: \includegraphics[width=0.5\textwidth]{f10}

1549: \end{center}

1550: \caption[]{Time history of the Poynting flux crossing the $r=100M$

1551: surface normalized to the total magnetic field strength within the disk

1552: for simulation KDPg (black lines) and the high resolution axisymmetric

1553: simulation of the dipole (magenta line).}

1554: \label{2dTrtdU} 

1555: \end{figure}

1556: 

1557: \begin{figure*}

1558: \begin{center}

1559: \leavevmode

1560: \includegraphics{f11}

1561: \end{center}

1562: \caption[]{Field evolution in the high-resolution,

1563: axisymmetric simulation of the quadrupole field topology. Solid white

1564: contours denote field lines that have the same parity as the dipole,

1565: dashed contours show field lines with the opposite parity. Times

1566: shown are the same as in Figure~\ref{DFPevln}.}

1567: \label{QFPevln} 

1568: \end{figure*}

1569: 

1570: \begin{figure}

1571: \begin{center}

1572: \leavevmode

1573: \includegraphics[width=0.5\textwidth]{f12}

1574: \end{center}

1575: \caption[]{Time

1576: history of the Poynting flux crossing the $r=100M$ surface 

1577: normalized by the total magnetic field strength within the disk.

1578: The KDPg dipole simulation is shown in black , the quadrupole

1579: simulation in blue, and the multiple loop simulation is green.}

1580: \label{TrtdUcomp} 

1581: \end{figure}

1582: 

1583: \begin{figure*}

1584: \begin{center}

1585: \leavevmode

1586: \includegraphics{f13}

1587: \end{center}

1588: \caption[]{As in Figure \ref{QFPevln} for the quadrupole field topology 

1589: overlaid on the narrow $\ell=4.9$ constant angular momentum torus.}

1590: \label{QNPevln} 

1591: \end{figure*}

1592: 

1593: \begin{figure*}

1594: \begin{center}

1595: \leavevmode

1596: \includegraphics{f14}

1597: \end{center}

1598: 

1599: \caption[]{ Axisymmetric simulation of a series of four narrow dipolar

1600: loops at $t = 400,500,600,700$M, showing the initial infall of the

1601: innermost loop and subsequent formation of a large-scale dipole

1602: field. Solid white contours denote field loops of the same polarity

1603: as used in the dipole and quadrupole topologies, dashed white contours

1604: denote loops of opposite polarity.}

1605: 

1606: \label{MFPevln1} 

1607: \end{figure*}

1608: 

1609: \begin{figure*}

1610: \begin{center}

1611: \leavevmode

1612: \includegraphics{f15}

1613: \end{center}

1614: \caption[]{As in Figure \ref{MFPevln1} at $t =

1615: 750,800,850,900$M, showing the interaction of the second

1616: field loop with the dipole field established by the innermost

1617: field loop.  The initial dipole field is destroyed but a new

1618: large-scale field is subsequently established.}

1619: \label{MFPevln2} 

1620: \end{figure*}

1621: 

1622: \begin{figure*}

1623: \begin{center}

1624: \leavevmode

1625: \includegraphics{f16}

1626: \end{center}

1627: \caption[]{Azimuthally averaged gas $\beta$ parameter in the toroidal

1628: field simulation at (from left to right) $t=2000,4000,5600,7200$M. The

1629: equatorial region remains gas pressure dominated throughout the initial

1630: evolution, in marked contrast to all of the poloidal topologies.}

1631: \label{TDPevln1} 

1632: \end{figure*}

1633: 

1634: \begin{figure*}

1635: \begin{center}

1636: \leavevmode

1637: \includegraphics{f17}

1638: \end{center}

1639: \caption[]{Poloidal structure of ${\cal B}^r$ for $\phi = 0.25\pi$ in the toroidal

1640: field simulation at (from left to right) $t=2000,4000,5600,7200$M. Red denotes strong, positive field

1641: strength, blue strong negative field strength. Grey denotes zero field.}

1642: \label{TDPevln2} 

1643: \end{figure*}

1644: 

1645: \end{document}