1: %\documentclass[manuscript]{aastex}

2: %\documentclass[preprint]{aastex}

3: 

4: %\documentclass[preprint2]{aastex}

5: %\usepackage{times}  % use times font, and makes pdf's really nice

6: 

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

8: 

9: \documentclass{emulateapj}

10: 

11: 

12: \shorttitle{Magnetic Core-Envelope Explosions}

13: \shortauthors{Matt, Frank, \& Blackman}

14: 

15: %\slugcomment{To submit to ApJ Letters.}

16: 

17: \begin{document}

18: 

19: \title{Astrophysical Explosions Driven by a Rotating, Magnetized,

20: Gravitating Sphere}

21: 

22: \author{

23: Sean Matt\altaffilmark{1},

24: Adam Frank\altaffilmark{2}, and

25: Eric G.\ Blackman\altaffilmark{2}}

26: 

27: %\affil{$^1$Department of Astronomy, University of Virginia, 

28: %Charlottesville VA, 22904\\

29: %$^2$ Physics \& Astronomy Department, University of Rochester, 

30: %Rochester NY, 14627}

31: 

32: %\altaffiltext{1}{Levinson/VITA Fellow; seanmatt@virginia.edu}

33: 

34: %\altaffiltext{2}{afrank@pas.rochester.edu, blackman@pas.rochester.edu}

35: 

36: \altaffiltext{1}{Dept. of Astronomy, U. of Virginia, 

37: Charlottesville VA, 22904; Levinson/VITA Fellow; seanmatt@virginia.edu}

38: 

39: \altaffiltext{2}{Physics \& Astronomy Dept., U. of Rochester, 

40: Rochester NY, 14627; afrank@pas.rochester.edu, blackman@pas.rochester.edu}

41: 

42: 

43: 

44: \begin{abstract}

45: 

46: We present the results of a numerical magnetohydrodynamic simulation

47: that demonstrates a mechanism by which magnetic fields tap rotational

48: energy of a stellar core and expel the envelope.  Our numerical setup,

49: designed to focus on the basic physics of the outflow mechanism,

50: consists of a solid, gravitating sphere, which may represent the

51: compact core of a star, surrounded by an initially hydrostatic

52: envelope of ionized gas.  The core is threaded by a dipolar magnetic

53: field that also permeates the envelope.  At the start of the

54: simulation, the core begins to rotate at 10\% of the escape speed.

55: The magnetic field is sufficiently strong to drive a

56: magneto-rotational explosion, whereby the entire envelope is expelled,

57: confirming the expectation of analytical models.  Furthermore, the

58: dipolar nature of the field results in an explosion that is enhanced

59: simultaneously along the rotation axis (a jet) and along the magnetic

60: equator.  While the initial condition is simplified, the simulation

61: approximates circumstances that may arise in astrophysical objects

62: such as Type II supernovae, gamma ray bursts, and proto-planetary

63: nebulae.

64: 

65: %When the field strength or rotation rate is sufficiently

66: %high, the result is a magnetic explosion whereby the entire envelope

67: %is expelled.  While our initial condition (in which a strong initial

68: %shear is present) is somewhat artificial, a similar situation may

69: %arise in astrophysical objects such as Type II supernovae, gamma ray

70: %bursts, and proto-planetary nebulae.

71: 

72: \end{abstract}

73: 

74: \keywords{MHD --- planetary nebulae: general --- stars: evolution ---

75: stars: mass loss --- stars: rotation --- supernovae: general}

76: 

77: 

78: 

79: \section{Introduction \label{sec_intro}}

80: 

81: 

82: 

83: Over the last two decades magnetic fields have been identified as the

84: principle, universal agent for creating collimated astrophysical

85: outflows. When a magnetic field is anchored in a rapidly rotating

86: object near the bottom of a gravitational potential, the field can act

87: as a drive-belt, tapping rotational kinetic energy and launching

88: plasma back up the potential well.  This magneto-rotational (MR)

89: scenario for jet launching has been explored by numerous authors, both

90: analytically \citep{blandfordpayne82, pelletierpudritz92} and

91: numerically \citep*{ouyed3ea97, krasnopolsky3ea99}.  These authors

92: showed this mechanism can produce steady flows of matter, in a sense

93: that the outflow engine still operates while the flow is

94: observable---as in jets from young stellar objects

95: \citep{reipurthbally01} and active galactic nuclei

96: \citep[][]{begelman3ea84}.

97: 

98: On the other hand, this mechanism can also operate in a transient

99: event, linked to a rapid evolution of the source and driving an

100: explosion.  In this case, the engine loses a significant fraction of

101: its power by the time the outflow (or its interaction with the

102: environment) is detected.  The proposed mechanism for gamma ray bursts

103: \citep[GRBs;][]{piran05} and supernovae (SNe), lie in the explosive

104: regime.  There is also growing evidence \citep{bujarrabalea01} that

105: collimated outflows from planetary nebulae (PNs) are explosive.

106: 

107: Transient MR explosions have not been explored in as much detail as

108: the steady-state models.  In a collapsing star, differential rotation

109: near the core may amplify the field linearly \citep[when turbulence is

110: unimportant;][]{kluzniakruderman98, wheeler3ea02} or exponentially

111: \citep[when turbulence is important;][]{akiyamaea03, blackman3ea06}.

112: When differential rotation twists a poloidal magnetic field,

113: generating toroidal field $B_\phi$, enough rotational energy may be

114: tapped to power a supernova explosion.  The role of toroidal field

115: pressure in then driving an outflow has been highlighted by

116: \citet{lyndenbell96}, who explored magneto-static ``magnetic towers''

117: to understand jet properties \citep[see also][]{liea01}, and

118: \citet{uzdenskymacfadyen06} has extended this work to exploding stars.

119: Numerical simulations by \citet{leblancwilson70} and more recently by

120: \citet*{ardeljan3ea05} support the idea that MR explosions can be

121: important for driving SNe and GRBs.

122: 

123: However, the inherent time-dependence and general complexity of the

124: system presents a significant challenge to our understanding of the

125: basic mechanism and identification of key parameters.  It is clear

126: that a MR explosion is ultimately driven by the rotational kinetic

127: energy extracted from the material that is left behind (the stellar

128: remnant).  But, notably, there is still some uncertainty as to whether

129: an accretion disk must form inside the star \citep[as

130: in][]{uzdenskymacfadyen06} or whether the rotation of the stellar core

131: \citep[e.g., a protopulsar;][]{wheeler3ea02} alone will drive a MR

132: explosion.

133: 

134: In this Letter, we present a magnetohydrodynamic (MHD) simulation of a

135: MR explosion.  While our setup is simple, similar to the analytical

136: ``protopulsar jet'' model described by \citet{wheeler3ea02}, our

137: simulation captures the nonlinear dynamics, as the magnetic field is

138: twisted at the shear layer between the core and envelope.  This

139: heuristic approach enables a better understanding of the basic MR

140: physics and complements the more phenomenological approach of

141: previous, more complex, numerical studies.

142: 

143: 

144: %In order to understand some of the basic physical processes at work,

145: %we adopt a heuristic approach, eliminating as many possible

146: %complications, while maintaining applicability to real stars.  The

147: %benefit of our approach is its simplicity, which allows us to

148: %interpret the results more readily than previous numerical studies.

149: 

150: %A detailed analysis of the dynamics, energetics, and morphology of the

151: %outburst will be published in forthcoming work.

152: 

153: 

154: 

155: \section{Physical Experiment and Numerical Method} \label{sec_setup}

156: 

157: 

158: 

159: 	\subsection{Physics \label{sub_basic}}

160: 

161: 

162: 

163: The model consists of a gravitating, magnetized sphere (the ``core,''

164: representing a stellar core) surrounded by an overlying envelope.  The

165: core is solid in a sense that no mass flows in or out of the surface,

166: and it rotates as a solid body.  The core is also a conductor with a

167: rotation-axis-aligned dipole magnetic field anchored to the surface.

168: This magnetic field couples to the envelope, whose behavior is

169: characterized by the ideal MHD equations.  This core-envelope model

170: approximates the conditions in the interior of a star where there is a

171: very steep density gradient, for example in the proximity of a nearly

172: degenerate core in an evolved star.  There is an implicit assumption

173: that the envelope self-gravity is negligible.

174: 

175: We assume an initial state in which the envelope is not rotating and

176: rests in hydrostatic equilibrium, and whose density falls of as

177: $r^{-2}$.  The initial dipole magnetic field is force-free and

178: permeates the envelope.  At a time defined as $t$ = 0 in the

179: simulations, the core begins rotating at a constant rate of 10\% of

180: the escape speed.  This initiates a shear at the boundary between the

181: surface of the core and the base of the envelope.

182: 

183: This strong initial shear and purely poloidal magnetic field

184: approximates the conditions in a real star immediately following an

185: abrupt change in structure, such as a core collapse (or envelope

186: expansion).  If the collapse is rapid enough, $B_\phi$ subsequently

187: amplified at the core-envelope interface will be much stronger than

188: $B_\phi$ generated elsewhere before the collapse, and the envelope

189: rotation will be negligible.  The envelope is unlikely to be at rest

190: following such a change, and any bulk motion could quantitatively (but

191: not qualitatively) change our results.  For example, the ram pressure

192: of an infalling envelope would change the magnetic energy required for

193: an explosion by approximately a factor of two.  Finally, the field

194: geometry is reasonable, since pulsars and magnetic white dwarfs are

195: thought to have dipolar fields (though tilted with respect to their

196: rotation axes).

197: 

198: The poloidal magnetic field connecting the core to the envelope is

199: twisted by the shear, generating an azimuthal component to the

200: magnetic field, $B_\phi$, near the interface, which increases

201: approximately linearly in time.  The magnetic pressure force

202: associated with $-\nabla B_\phi^2$ will be directed generally outward

203: from the core, so when $B_\phi$ becomes dynamically important, the

204: field will expand, driving the ionized envelope material in front of

205: it.  If the expansion is powerful enough, it can be explosive and

206: drive off the entire envelope.  We intentionally inhibit any sort of

207: wind from the surface of the core, so the outflow in this explosion is

208: composed of material that originally overlies the core (i.e., envelope

209: material).  The flow thus depends on the initial conditions existing

210: above the surface of the core and is inherently a transient

211: phenomenon.  This contrasts with any steady-state wind theory whose

212: solution depends only on the boundary conditions at the wind driving

213: source.

214: 

215: 

216: 

217:      \subsection{Parameters \label{sub_parms}}

218: 

219: 

220: 

221: The key parameters in the system are the initial magnetic, rotational

222: kinetic, gravitational potential, and thermal energy densities near

223: the core-envelope interface.  It is instructive to compare the

224: characteristic speeds associated with these energies, namely the

225: Alfv\'en speed [$v_{\rm A} = B / (4 \pi \rho)^{1/2}$, where $B$ is the

226: dipole field strength at the equator of the core and $\rho$ is the

227: mass density at the base of the envelope], rotation speed at the

228: equator ($v_{\rm rot}$), escape speed at the core surface ($v_{\rm

229: esc}$), and sound speed at the base of the envelope ($c_{\rm s}$),

230: respectively.

231: 

232: The assumption of initial hydrostatic equilibrium couples the thermal

233: energy to the potential energy, effectively reducing the number of

234: free parameters (by requiring the initial $c_{\rm s} \approx 0.53

235: v_{\rm esc}$).  Also, since the simulations can be scaled to any

236: system with similar ratios of energies, there are only two fundamental

237: parameters, which can be cast as the dimensionless velocity ratios

238: $v_{\rm A} / v_{\rm esc}$ and $v_{\rm rot} / v_{\rm esc}$.  Here, we

239: present results from a simulation with $v_{\rm A} / v_{\rm esc} = 1.0$

240: and $v_{\rm rot} / v_{\rm esc} = 0.1$.

241: 

242: The key parameter that determines whether or not the envelope will be

243: driven off is $\chi \equiv (v_{\rm A} v_{\rm rot})^{1/2} / v_{\rm

244: esc}$.  We have run a limited parameter study thus far (not presented

245: in this Letter), which suggests the threshold for explosion occurs

246: near $\chi \ga 0.2$.  The model presented here has $\chi \approx 0.3$.

247: 

248: 

249: 

250: \subsection{Numerics}

251: 

252: 

253: 

254: We employ the 2.5-dimensional MHD code of \citet{matt02}, and the

255: reader can find details of the code there and also in \citet{mattea02}

256: and \citet{mattbalick04}.  The code, solves the ideal (non-resistive)

257: MHD equations using a two-step Lax-Wendroff, finite difference scheme

258: \citep{richtmyermorton67} in cylindrical $(\varpi, \phi, z)$ geometry.

259: The formulation of the equations allows for a polytropic equation of

260: state (we adopt $\gamma = 5/3$), includes a source term in the

261: momentum and energy equations for point-source gravity, and assumes

262: axisymmetry ($\partial / \partial \phi$ = 0 for all quantities).

263: 

264: The computational domain consists of five nested grids (or ``boxes'')

265: in the cylindrical $\varpi$-$z$ plane.  Each box contains $401 \times

266: 400$ gridpoints (in $\varpi$ and $z$, respectively) with constant grid

267: spacing.  The boxes are nested concentrically, so that the inner

268: (first) box represents the smallest domain at the highest resolution.

269: The next outer box represents twice the domain size with half the

270: spatial resolution (an so on for other boxes).  The simulation ended

271: before the explosion propagated beyond the fourth box.  A circular

272: boundary with a radius of 30.5 gridpoints and centered on the origin

273: represents the surface of the core.  Thus, the innermost box has a

274: spatial resolution of 0.033 core radii, $R_{\rm c}$, and a domain size

275: of 13.1 $R_{\rm c}$, and the fourth box has a resolution and domain

276: size of 0.26 $R_{\rm c}$ and 105 $R_{\rm c}$, respectively.

277: 

278: 

279: %\begin{figure*}

280: \begin{figure}

281: %\epsscale{2.}

282: \epsscale{1.15}

283: %\epsscale{1.}

284: %\plotone{baseseq_sclose.eps}

285: \plotone{f1.eps}

286: 

287: \caption{Greyscale images of log density (black is highest density),

288: poloidal field lines, and velocity vectors show the evolution of the

289: system in the region near the core.  The data are from the first

290: simulation grid.  The core is in the lower left of each panel, and the

291: rotation axis is vertical.  The four panels represent a time-sequence

292: from 0, 1.5, 3.0, and 4.5 rotations of the core.

293: \label{fig_baseseq_close}}

294: 

295: %\end{figure*}

296: \end{figure}

297: 

298: 

299: We use standard outflow conditions on the outer box boundaries,

300: appropriate for wind studies.  We also use standard boundary

301: conditions on the rotation axis ($\varpi = 0$) and reflection symmetry

302: on the equator ($z =0$).  The simulation domain, then, consists of a

303: $\varpi$-$z$ slice through a single quadrant.  The surface of the core

304: is represented by a circular inner boundary, centered at the origin

305: ($\varpi = z = 0$).  Here, we enforce boundary conditions on a three

306: gridcell layer, and we have performed several tests to ensure that

307: this inner boundary behaves as appropriate for the surface of a solid,

308: rotating, conducting and magnetized sphere.

309: 

310: For the current study, we initialize the entire computational domain

311: with a spherically symmetric hydrostatic envelope, whose density falls

312: off as $r^{-2}$ (where $r$ is the spherical radius).  The envelope is

313: initially stagnant (no rotation or motion) and is threaded everywhere

314: by a dipole magnetic field that is anchored to the core, as shown in

315: the upper left panel of Figure \ref{fig_baseseq_close}.

316: 

317: 

318: 

319: \section{Results: Nature of the Explosion \label{sec_results}}

320: 

321: 

322: 

323: At the start of the simulations, the dipolar magnetic field begins to

324: wrap up.  For the chosen parameters, $B_\phi$ becomes dynamically

325: important after a few tenths of a rotation of the core.  The panels of

326: Figure \ref{fig_baseseq_close} show the evolution during the first 4.5

327: rotations of the core.  The expansion of newly generated $B_\phi$ is

328: evident, as it drives the envelope material outward.  This also

329: expands the poloidal magnetic field (lines), though some field lines

330: near the equator remain closed, where material is forced and held in

331: corotation with the core.  The region near the core is evacuated by

332: the expansion of the magnetic field, indicative of a transient event,

333: rather than the beginning of a steady wind.

334: 

335: 

336: 

337: 

338: \begin{figure}

339: %\epsscale{2.}

340: \epsscale{1.15}

341: %\epsscale{1.}

342: %\plotone{baseseq_1far.eps}

343: \plotone{f2.eps}

344: 

345: \caption{Greyscale image of log density (black is highest density) far

346: from the core, shown after 4.5 rotations of the core.  The data are

347: from the fourth simulation grid. The core is indicated by a white

348: circle at the center.  Vectors represent the flow velocity with the

349: maximum vector length corresponding to 4.1 times $v_{\rm rot}$.

350: \label{fig_baseseq_far}}

351: 

352: \end{figure}

353: 

354: 

355: 

356: Figure \ref{fig_baseseq_far} shows the density in the system after 4.5

357: rotations of the core, on a scale that is 10 times larger than in

358: Figure \ref{fig_baseseq_close}, and the data have been reflected about

359: the rotation axis and equator to better illustrate the flow.  In the

360: Figure, one sees a high-density swept-up shell of envelope material

361: that is moving outward from the core and bounded by a shock.

362: 

363: The shell has a quadrupolar shape.  This symmetry is due to the

364: coupling of the rotation to the dipole field, which produces a

365: $B_\phi$ that has a maximum strength at mid latitudes near the core,

366: so magnetic pressure forces direct outward from the core, and also

367: toward the pole and equator \citep[see][]{mattbalick04}.  This shape

368: can be qualitatively understood as an explosion driven by detonating

369: two doughnut-shaped charges placed at mid latitudes around the north

370: and south hemispheres, which results in a convergent blast wave on the

371: equator and pole.

372: 

373: By 4.5 rotations of the core, the shell material near the rotation

374: axis and equator is traveling faster than its own internal fast

375: magnetosonic wave speed, as well as that of the ambient envelope ahead

376: of the shell.  We expect the expansion speed of the shell to roughly

377: correlate with $(v_{\rm rot}^2 v_{\rm A})^{1/3}$.  At 4.5 rotations,

378: the speed of the shell near the rotation axis roughly equals this.

379: However, the flow speed of the shell material accelerates throughout

380: the simulation, and does not seem to reach an asymptotic velocity by

381: the end.  The simulations were stopped after 5.6 rotations of the

382: core, since the steepening shock front began to produce numerical

383: instabilities at that time.

384: 

385: The total energy (gravitational potential plus magnetic plus kinetic

386: plus thermal) integrated over the entire simulation grid (excluding

387: the core) increases monotonically in time and becomes positive at

388: around 2.5 rotations.  This means that there is enough energy for the

389: entire envelope to escape from the gravitational potential well of the

390: core.  Indeed, by the end of the simulations, the shell material near

391: the axis and equator is traveling faster than the local escape speed.

392: On the other hand, material in the shell at mid latitudes is

393: traveling substantially below its local escape speed.  However, since

394: this mid latitude material is still accelerating outward at the end of

395: the simulation, and since the total energy in the system is well above

396: zero, it appears likely that the entire envelope will escape from the

397: core (if the simulation were able to run long enough).  This confirms

398: the general picture of \citet{wheeler3ea02}.  In addition, the

399: structure of the outflow suggest that, for lower $\chi$, a partial

400: ejection of the envelope may occur (a ``failed'' explosion) in which

401: their is not only a jet \citep[as in][]{wheeler3ea02} but also an

402: equatorial flow, leaving behind the material at mid latitudes.

403: 

404: 

405: 

406: \begin{figure}

407: %\epsscale{1.}

408: \epsscale{1.15}

409: %\plotone{3dfar.eps}

410: \plotone{f3.eps}

411: 

412: \caption{Three-dimensional rendering revealing the explosion

413: mechanism, viewed from $\sim30$ degrees above the magnetic equator.

414: The two blue surfaces are contours of constant density, each at the

415: same density value.  The dense shell of swept up envelope material

416: (see Fig.\ \ref{fig_baseseq_far}) exists between the two surfaces.

417: Gold wires trace the magnetic field lines and illustrate that the

418: field is most highly twisted in the low density region interior to the

419: shell.

420: \label{fig_3dfar}}

421: 

422: \end{figure}

423: 

424: 

425: 

426: 

427: Finally, the magnetic field configuration after 4.5 rotations is

428: apparent in Figure \ref{fig_3dfar}, where we have exploited the

429: axisymmetry of the system to create a 3-dimensional rendering of data

430: from the fourth simulation grid.  For scale, the outer isodensity

431: surface corresponds to the outer edge of the shell in Figure

432: \ref{fig_baseseq_far}.  A comparison between Figures

433: \ref{fig_baseseq_far} and \ref{fig_3dfar} reveals that the ``hollow''

434: region interior to the dense shell is actually filled with magnetic

435: energy, mostly due to the $B_\phi$ component.  The shell is like a

436: magnetically inflated ``balloon,'' which expands as the spinning core

437: adds magnetic energy inside.  Thus the spin energy of the core is

438: ultimately being transferred to the kinetic energy of the envelope

439: (and the core should spin down, as a result).  The interior region is

440: magnetically (Poynting flux) dominated, as all material there flows

441: more slowly than the fast magnetosonic speed.  The flow along the axis

442: is a magnetic tower jet, similar to what is seen in other simulations

443: \citep*[e.g.,][]{kato3ea04}, while the flow near the equatorial plane

444: is more of a ``magnetic pancake,'' a feature of the rotating dipole

445: field and an interesting new result of this study.

446: 

447: 

448: 

449: %\section{Discussion and Conclusions \label{sec_discussion}}

450: \section{Application to Real Stars \label{sec_discussion}}

451: 

452: 

453: 

454: Our numerical setup is simple, which allows us to understand the basic

455: physical principles at work.  The requirements for the mechanism to

456: operate are a) a shear layer in which the inner region is rotating at

457: a fraction of breakup speed and b) a magnetic field that results in an

458: Alfv\'en speed comparable to the escape speed.  Because we focus on

459: the MR mechanism, and ignore other physics (e.g., neutrino heating),

460: we can scale our results to various astrophysical systems.  This is

461: instructive, as it determines what conditions would be necessary for a

462: MR explosion alone to drive off the envelope.  Here, we briefly apply

463: our results to the progenitor stars of SNe and PNs.

464: 

465: A simple model of a core-collapse SN suggests that core material

466: conserves magnetic flux and angular momentum during collapse and

467: naturally forms a rapidly-rotating, highly-magnetized proto-neutron

468: star.  We adopt a neutron star (``core'') mass of 1.4 $M_\odot$ and a

469: core radius of $R_{\rm c} = 10^6$ cm.  If the overlying envelope

470: contains 1 $M_\odot$ between $r = R_{\rm c}$ and $1 R_\odot$, our

471: simulation corresponds to a dipole field strength at the core surface

472: equal to $B = 3 \times 10^{15}$ G.  Note that the envelope mass is

473: comparable to the core mass, but the self-gravity of the envelope near

474: the core should not be important, and so the simulation still applies.

475: In our simulation, we find that the core loses rotational energy at a

476: constant rate of $B^2 R_{\rm c}^3 / 3 = 3 \times 10^{48}$ ergs per

477: rotation, which should extract most of the rotational energy of the

478: core ($\sim 10^{51}$ ergs) in 1 second.  These numbers are consistent

479: with previous work and should be sufficient to eject the envelope

480: \citep[e.g.,][]{wheeler3ea02, ardeljan3ea05, blackman3ea06}.

481: 

482: The evolutionary phase preceding PN formation is marked by high mass

483: loss rates, leading to an expansion of the stellar envelope, which

484: overlies a proto-white dwarf.  A shear layer at the core-envelope

485: boundary is expected \citep{blackmanea01}, and as the envelope density

486: decreases, the Alfv\'en speed should rise.  There may be a threshold

487: density, below which the remaining envelope may be expelled via a MR

488: explosion.  We adopt a white dwarf (``core'') mass of 0.5 $M_\odot$,

489: and $R_{\rm c} = 10^9$ cm.  At this time the envelope is essentially

490: an extension of the massive outflow, and if it contains 0.05 $M_\odot$

491: between $r = R_{\rm c}$ and $10^4$ AU, our simulation corresponds to

492: $B = 9 \times 10^6$ G.  Most of the rotational energy of the core

493: ($\sim 10^{48}$ ergs) should be transferred to the envelope in a few

494: hundred years.  The linear momentum of the swept up shell will be

495: $\sim 10^{39}$ g cm s$^{-1}$, which is consistent with the

496: observations of proto-PNs by \citet{bujarrabalea01}.

497: 

498: Finally, we note that the large-scale field in our simulations is

499: strong enough to suppress turbulence due to shear instabilities near

500: the core, and the field therefore is neither subject to decay nor the

501: exponential growth that is important in some other models.  We have

502: instead focused on the physics of how the field actually drives the

503: outflow.

504: 

505: 

506: %Note that $B^2 R_{\rm c}^3 / 3 = E_{\rm dip}$ is the initial magnetic

507: %field energy integrated from $r = R_{\rm c}$ to $\infty$.

508: 

509: 

510: 

511: \acknowledgements

512: 

513: This research was supported by NSERC of Canada, McMaster U., and

514: through a CITA National Fellowship and by the U. of Virginia through a

515: Levinson/VITA fellowship.

516: 

517: 

518: % ---------- BIBLIOGRAPHY -------------

519: 

520: %\bibliography{../../../references}

521: %\bibliographystyle{apj}

522: 

523: \begin{thebibliography}{}

524: 

525: \bibitem[\protect\citeauthoryear{{Akiyama} et~al.}{{Akiyama}

526:   et~al.}{2003}]{akiyamaea03}

527: {Akiyama}, S., {Wheeler}, J.~C., {Meier}, D.~L.,  \& {Lichtenstadt}, I. 2003,

528:   \apj, 584, 954

529: 

530: \bibitem[\protect\citeauthoryear{{Ardeljan}, {Bisnovatyi-Kogan}, \&

531:   {Moiseenko}}{{Ardeljan} et~al.}{2005}]{ardeljan3ea05}

532: {Ardeljan}, N.~V., {Bisnovatyi-Kogan}, G.~S.,  \& {Moiseenko}, S.~G. 2005,

533:   \mnras, 359, 333

534: 

535: \bibitem[\protect\citeauthoryear{{Begelman}, {Blandford}, \& {Rees}}{{Begelman}

536:   et~al.}{1984}]{begelman3ea84}

537: {Begelman}, M.~C., {Blandford}, R.~D.,  \& {Rees}, M.~J. 1984, Reviews of

538:   Modern Physics, 56, 255

539: 

540: \bibitem[\protect\citeauthoryear{{Blackman} et~al.}{{Blackman}

541:   et~al.}{2001}]{blackmanea01}

542: {Blackman}, E.~G., {Frank}, A., {Markiel}, J.~A., {Thomas}, J.~H.,  \& {Van

543:   Horn}, H.~M. 2001, \nat, 409, 485

544: 

545: \bibitem[\protect\citeauthoryear{{Blackman}, {Nordhaus}, \&

546:   {Thomas}}{{Blackman} et~al.}{2006}]{blackman3ea06}

547: {Blackman}, E.~G., {Nordhaus}, J.~T.,  \& {Thomas}, J.~H. 2006, New Astronomy,

548:   11, 452

549: 

550: \bibitem[\protect\citeauthoryear{{Blandford} \& {Payne}}{{Blandford} \&

551:   {Payne}}{1982}]{blandfordpayne82}

552: {Blandford}, R.~D.,  \& {Payne}, D.~G. 1982, \mnras, 199, 883

553: 

554: \bibitem[\protect\citeauthoryear{{Bujarrabal} et~al.}{{Bujarrabal}

555:   et~al.}{2001}]{bujarrabalea01}

556: {Bujarrabal}, V., {Castro-Carrizo}, A., {Alcolea}, J.,  \& {S{\' a}nchez

557:   Contreras}, C. 2001, \aap, 377, 868

558: 

559: \bibitem[\protect\citeauthoryear{{Kato}, {Hayashi}, \& {Matsumoto}}{{Kato}

560:   et~al.}{2004}]{kato3ea04}

561: {Kato}, Y., {Hayashi}, M.~R.,  \& {Matsumoto}, R. 2004, \apj, 600, 338

562: 

563: \bibitem[\protect\citeauthoryear{{Klu{\'z}niak} \& {Ruderman}}{{Klu{\'z}niak}

564:   \& {Ruderman}}{1998}]{kluzniakruderman98}

565: {Klu{\'z}niak}, W.,  \& {Ruderman}, M. 1998, \apjl, 505, L113

566: 

567: \bibitem[\protect\citeauthoryear{{Krasnopolsky}, {Li}, \&

568:   {Blandford}}{{Krasnopolsky} et~al.}{1999}]{krasnopolsky3ea99}

569: {Krasnopolsky}, R., {Li}, Z.-Y.,  \& {Blandford}, R. 1999, \apj, 526, 631

570: 

571: \bibitem[\protect\citeauthoryear{{LeBlanc} \& {Wilson}}{{LeBlanc} \&

572:   {Wilson}}{1970}]{leblancwilson70}

573: {LeBlanc}, J.~M.,  \& {Wilson}, J.~R. 1970, \apj, 161, 541

574: 

575: \bibitem[\protect\citeauthoryear{{Li} et~al.}{{Li} et~al.}{2001}]{liea01}

576: {Li}, H., {Lovelace}, R.~V.~E., {Finn}, J.~M.,  \& {Colgate}, S.~A. 2001, \apj,

577:   561, 915

578: 

579: \bibitem[\protect\citeauthoryear{{Lynden-Bell}}{{Lynden-Bell}}{1996}]{lyndenbe%

580: ll96}

581: {Lynden-Bell}, D. 1996, \mnras, 279, 389

582: 

583: \bibitem[\protect\citeauthoryear{{Matt} \& {Balick}}{{Matt} \&

584:   {Balick}}{2004}]{mattbalick04}

585: {Matt}, S.,  \& {Balick}, B. 2004, \apj, 615, 921

586: 

587: \bibitem[\protect\citeauthoryear{{Matt} et~al.}{{Matt} et~al.}{2002}]{mattea02}

588: {Matt}, S., {Goodson}, A.~P., {Winglee}, R.~M.,  \& {B{\" o}hm}, K. 2002, \apj,

589:   574, 232

590: 

591: \bibitem[\protect\citeauthoryear{{Matt}}{{Matt}}{2002}]{matt02}

592: {Matt}, S.~P. 2002, Ph.D.~Thesis, Astronomy, University of Washington

593: 

594: \bibitem[\protect\citeauthoryear{{Ouyed}, {Pudritz}, \& {Stone}}{{Ouyed}

595:   et~al.}{1997}]{ouyed3ea97}

596: {Ouyed}, R., {Pudritz}, R.~E.,  \& {Stone}, J.~M. 1997, \nat, 385, 409

597: 

598: \bibitem[\protect\citeauthoryear{{Pelletier} \& {Pudritz}}{{Pelletier} \&

599:   {Pudritz}}{1992}]{pelletierpudritz92}

600: {Pelletier}, G.,  \& {Pudritz}, R.~E. 1992, \apj, 394, 117

601: 

602: \bibitem[\protect\citeauthoryear{{Piran}}{{Piran}}{2005}]{piran05}

603: {Piran}, T. 2005, Reviews of Modern Physics, 76, 1143

604: 

605: \bibitem[\protect\citeauthoryear{{Reipurth} \& {Bally}}{{Reipurth} \&

606:   {Bally}}{2001}]{reipurthbally01}

607: {Reipurth}, B.,  \& {Bally}, J. 2001, \araa, 39, 403

608: 

609: \bibitem[\protect\citeauthoryear{{Richtmyer} \& {Morton}}{{Richtmyer} \&

610:   {Morton}}{1967}]{richtmyermorton67}

611: {Richtmyer}, R.~D.,  \& {Morton}, K.~W. 1967, {Difference Methods for

612:   Initial-Value Problems} (New York, NY: Wiley-Interscience)

613: 

614: \bibitem[\protect\citeauthoryear{{Uzdensky} \& {McFadyen}}{{Uzdensky} \&

615:   {McFadyen}}{2006}]{uzdenskymacfadyen06}

616: {Uzdensky}, D.~A.,  \& {McFadyen}, A.~I. 2006, \apj, p. submitted

617:   (astroph/0602419)

618: 

619: \bibitem[\protect\citeauthoryear{{Wheeler}, {Meier}, \& {Wilson}}{{Wheeler}

620:   et~al.}{2002}]{wheeler3ea02}

621: {Wheeler}, J.~C., {Meier}, D.~L.,  \& {Wilson}, J.~R. 2002, \apj, 568, 807

622: 

623: \end{thebibliography}

624: 

625: 

626: 

627: \end{document}

628: 

629: 

630: 

631: 

632: 

633: 

634: 

635: