1: %\documentclass[manuscript]{aastex}

2: %\documentclass[preprint2]{aastex}

3: %\documentclass[preprint]{aastex}

4: 

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

6: 

7: \documentclass{emulateapj}

8: 

9: % User commands

10: 

11: 

12: \shorttitle{Spin Equilibrium by APSWs}

13: \shortauthors{Matt \& Pudritz}

14: 

15: \begin{document}

16: 

17: \title{Accretion-Powered Stellar Winds III: Spin Equilibrium Solutions}

18: 

19: %\title{Accretion-Powered Stellar Winds III:\\Spin Equilibrium Solutions and Comparison to Disk Locking}

20: 

21: 

22: 

23: 

24: 

25: \author{Sean Matt\altaffilmark{1} and Ralph E. Pudritz\altaffilmark{2}}

26: 

27: 

28: \affil{$^1$Department of Astronomy, University of

29: Virginia, P.O. Box 400325, Charlottesville, VA 22904-4325; seanmatt@virginia.edu}

30: 

31: \affil{$^2$Physics and Astronomy Department, McMaster University,

32: Hamilton, ON L8S 4M1, Canada; pudritz@physics.mcmaster.ca}

33: 

34: 

35: %\altaffiltext{3}{Levinson/VITA Fellow, University of Virginia.}

36: 

37: 

38: 

39: 

40: \begin{abstract}

41: 

42: %Using the detailed calculations of the angular momentum carried in a

43: %magnetized stellar wind, from a previous work (Paper~II), we further

44: %develop the accretion-powered stellar wind (APSW) model for explaining

45: %the slow rotation observed in a large fraction of accreting

46: %pre-main-sequence stars.  In this paper

47: %For parameters typical of accreting pre-main-sequence

48: %stars, this predicts spin rates of $\sim 10$\% of breakup speed for

49: %$\dot M_{\rm w} / \dot M_{\rm a} \sim 0.1$, confirming the general

50: %APSW scenario.

51: 

52: We compare the stellar wind torque calculated in a previous work

53: (Paper~II) to the spin-up and spin-down torques expected to arise from

54: the magnetic interaction between a slowly rotating ($\sim 10$\% of

55: breakup) pre-main-sequence star and its accretion disk.  This analysis

56: demonstrates that stellar winds can carry off orders of magnitude

57: more angular momentum than can be transferred to the disk, provided

58: that the mass outflow rates are greater than the solar wind.  Thus,

59: the equilibrium spin state is simply characterized by a balance

60: between the angular momentum deposited by accretion and that extracted

61: by a stellar wind.  We derive a semi-analytic formula for predicting

62: the equilibrium spin rate as a function only of the ratio of $\dot

63: M_{\rm w} / \dot M_{\rm a}$ and a dimensionless magnetization

64: parameter, $\Psi \equiv B_*^2 R_*^2 (\dot M_{\rm a} v_{\rm

65:   esc})^{-1}$, where $\dot M_{\rm w}$ is the stellar wind mass outflow

66: rate, $\dot M_{\rm a}$ the accretion rate, $B_*$ the stellar surface

67: magnetic field strength, $R_*$ the stellar radius, and $v_{\rm esc}$

68: the surface escape speed.  For parameters typical of accreting

69: pre-main-sequence stars, this explains spin rates of $\sim 10$\% of

70: breakup speed for $\dot M_{\rm w} / \dot M_{\rm a} \sim 0.1$.

71: Finally, the assumption that the stellar wind is driven by a fraction

72: of the accretion power leads to an upper limit to the mass flow ratio

73: of $\dot M_{\rm w} / \dot M_{\rm a} \la 0.6$.

74: 

75: 

76: 

77: \end{abstract}

78: 

79: \keywords{accretion, accretion disks --- MHD --- stars: magnetic

80: fields --- stars: pre-main-sequence --- stars: rotation --- stars:

81: winds, outflows}

82: 

83: 

84: 

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

86: 

87: 

88: 

89: The slow rotation rates of low to intermediate mass ($\la 2 M_\odot$)

90: pre-main-sequence stars remains one of the most important aspects of

91: star formation that has, so far, resisted a generally accepted

92: explanation.  By the time they become optically visible as T Tauri

93: stars \citep[TTSs;][]{joy45}, approximately half of them are observed

94: to rotate at approximately 10\% of breakup speed \citep[the ``slow

95: rotators''; e.g.,][]{vogelkuhi81, bouvier3ea97, rebull3ea04,

96: herbstea07}.  This is a surprise because many TTSs (the Classical T

97: Tauri stars; CTTSs) are actively accreting material from surrounding

98: Keplerian disks \citep{lyndenbellpringle74, bertout3ea88,

99: calvetgullbring98, muzerolle3ea01}.  At a typical accretion rate of

100: $\dot M_{\rm a} \sim 10^{-8} M_\odot$ yr$^{-1}$, the angular momentum

101: deposited by accreting disk material should spin up a CTTS to near

102: breakup speed in $\sim 10^6$ years \citep{hartmannstauffer89,

103: mattpudritz07coolstars}.  Since the accretion phase lasts for $10^6$

104: -- $10^7$ years \citep{lyolawson05, jayawardhanaea06}, since the stars

105: accrete at much higher rates prior to the TTS phase, and since the

106: stars are still contracting \citep{rebullea02}, an efficient angular

107: momentum loss mechanism is required to explain the existence of the

108: slow rotators.

109: 

110: A few interesting and important ideas for explaining the TTS slow

111: rotators have been developed over the last two decades.  These have

112: resulted in the star-disk interaction model of \citet{ghoshlamb78},

113: applied to CTTSs by \citet[][and see \citealp{camenzind90}]{konigl91},

114: the X-wind model \citep{shuea94}, and the idea that stellar winds

115: provide strong torques \citep{hartmannstauffer89, toutpringle92,

116: paatzcamenzind96, ferreira3ea00, mattpudritz05l}.  Although both have

117: advanced our understanding of the magnetic star-disk interaction,

118: neither the Ghosh \& Lamb nor X-wind models are without problems

119: \citep{ferreira3ea00, uzdensky04, mattpudritz05}, and the idea that

120: stellar winds are important has not yet been worked out in sufficient

121: detail to compare to the other models.

122: 

123: %Prompted by some outstanding issues with the Ghosh \& Lamb and X-wind

124: %models, 

125: 

126: In \citet[][hereafter Paper~I]{mattpudritz05l}, we further explored

127: powerful stellar winds as a solution to the angular momentum problem

128: and suggested that a fraction of the accretion power provides the

129: energy necessary to drive the wind.  We showed that stellar winds are

130: capable of carrying off the accreted angular momentum, provided that

131: $\dot M_{\rm w} / \dot M_{\rm a} \sim 0.1$, where $\dot M_{\rm w}$ is

132: the outflow rate of material that is magnetically connected to the

133: star (the ``stellar wind'').  This analysis included a formulation for

134: the stellar wind torque that contained the Alfv\'en radius ($r_{\rm

135:   A}$), which is not easily determined a priori in the wind, and the

136: conclusions were based on a one-dimensional scaling estimate of this

137: important physical quantity.  Thus, while it is clear that

138: accretion-powered stellar winds (APSWs) can in principle provide the

139: necessary spin-down torque, this idea requires further development to

140: produce a more detailed model.

141: 

142: Toward this goal, \citet[][hereafter Paper~II]{mattpudritz08II} used

143: 2-dimensional (axisymmetric) magnetohydrodynamic simulations to solve

144: for $r_{\rm A}$ and calculate realistic stellar wind torques for a

145: range of parameters.  In the present paper, we use the stellar wind

146: solutions of Paper~II to compare the stellar wind torque to the

147: torques expected to arise from the star-disk interaction.

148: Furthermore, we find new solutions for stellar spins, based upon

149: torque balance between the accretion torque and the APSW spin-down

150: torque.  This paper begins with a brief description of the simulation

151: results of Paper~II (\S \ref{sec_simulations}).  We then compare the

152: stellar wind torque to the star-disk spin-down torque in section

153: \ref{sec_tdsd} and then to the star-disk spin-up torque in section

154: \ref{sec_equilibrium}, which contains spin-equilibrium solutions.

155: Section \ref{sec_discussion} contains a summary and discussion.

156: 

157: 

158: 

159: \section{Results of Stellar Wind Simulations} \label{sec_simulations}

160: 

161: 

162: 

163: This section contains a brief description of the simulation results of

164: Paper~II that we will use for our analysis, and the reader will find

165: details in that paper.  The primary purpose of the simulations was to

166: compute the spin-down torque on a star, due to the angular momentum

167: outflow in a wind.  We used numerical magnetohydrodynamic simulations

168: to directly calculate the torque $\tau_{\rm w}$ from steady-state, 2D

169: (axisymmetric) winds from isolated stars.  We adopted coronal

170: (thermal-pressure-driven) winds as a proxy for the unknown wind

171: driving mechanism.  In the simulations, the torque is entirely

172: determined by the seven key parameters listed in table

173: \ref{tab_parms}.  These are the stellar mass, $M_*$; stellar radius,

174: $R_*$; strength of the rotation-axis-aligned dipole magnetic field at

175: the surface and equator of the star, $B_*$; spin rate expressed as a

176: fraction of breakup speed,

177: \begin{eqnarray}

178: \label{eqn_f}

179: %

180: f \equiv \Omega_* R_*^{3/2} (G M_*)^{-1/2},

181: %

182: \end{eqnarray}

183: where $\Omega_*$ is the angular spin rate of the star; mass outflow

184: rate in the stellar wind, $\dot M_{\rm w}$; ratio of the thermal sound

185: speed to the escape speed, evaluated at the base of the wind (just

186: above the stellar surface), $c_{\rm s} / v_{\rm esc}$; and adiabatic

187: index, $\gamma$.

188: 

189: 

190: 

191: \begin{deluxetable}{ll}

192: \tablewidth{0pt}

193: \tablecaption{Fiducial Stellar Wind Parameters \label{tab_parms}}

194: \tablehead{

195: \colhead{Parameter} &

196: \colhead{Value}

197: }

198: 

199: \startdata

200: 

201: $M_*$                     & 0.5 $M_\odot$ \\

202: $R_*$                     & 2.0 $R_\odot$ \\

203: $B_*$ (dipole)            & 200 G         \\

204: $f$                       & 0.1           \\

205: %$\dot M_{\rm w}$\tablenotemark{a} & $1.9 \times 10^{-9} M_\odot$ yr$^{-1}$ \\

206: $\dot M_{\rm w}$          & $1.9 \times 10^{-9} M_\odot$ yr$^{-1}$ \\

207: $c_{\rm s} / v_{\rm esc}$ & 0.222          \\

208: $\gamma$                  & 1.05           \\

209: 

210: \enddata

211: 

212: %\tablenotetext{a}{This is the approximate wind mass loss rate

213: %resulting in the fiducial case.}

214: 

215: \end{deluxetable}

216: 

217: 

218: 

219: Table \ref{tab_parms} lists the value of each parameter adopted for a

220: fiducial case.  Paper~II contained a parameter study in which each of

221: the seven parameters were varied relative to the fiducial case, and 14

222: cases from the parameter study are listed in table

223: \ref{tab_torques}\footnote{In this paper, we do not discuss the cases

224: from Paper~II that include a quadrupole magnetic field, nor the

225: extremely slow rotator case (with $f = 0.004$).}.  In each case, six

226: of the parameters were held fixed at the fiducial value (as given in

227: table \ref{tab_parms}), and one parameter was varied as indicated by

228: the first column of table \ref{tab_torques}.

229: 

230: To compare with analytic theory, we also calculated the effective Alfv\'en

231: radius ($r_{\rm A}$), where the poloidal wind velocity equals the

232: poloidal Alfv\'en speed, using an analytic formula for the stellar

233: wind torque,

234: \begin{eqnarray}

235: \label{eqn_tw}

236: %

237: \tau_{\rm w} = - \dot M_{\rm w} \Omega_* \left< r_{\rm A}^2 \right>.

238: %

239: \end{eqnarray}

240: Since our simulations are multi-dimensional, we have used $\left<

241: r_{\rm A}^2 \right>$, which is the mass-loss-weighted average of

242: $r_{\rm A}^2$.  Hereafter, we'll refer to $\left< r_{\rm A}^2

243: \right>^{1/2}$ generically as $r_{\rm A}$.  Using the simulation

244: result for $\tau_{\rm w}$, equation \ref{eqn_tw} defines the value of

245: $r_{\rm A}$, which is tabulated for all cases in the second column of

246: table \ref{tab_torques}.

247: 

248: In this paper, we make use of the semi-analytic

249: formulation for the Alfv\'en radius from Paper~II,

250: \begin{eqnarray}

251: \label{eqn_rasim}

252: %

253: {r_{\rm A} \over R_*} = K

254:   \left({{B_*^2 R_*^2} \over {\dot M_{\rm w} v_{\rm esc}}}\right)^m,

255: \end{eqnarray}

256: where $K$ and $m$ are dimensionless constants fit to the simulation,

257: and $v_{\rm esc} = (2 G M_* / R_*)^{1/2}$ is the escape speed from the

258: stellar surface.  Paper~II showed that the values of $K \approx 2.11$

259: and $m \approx 0.223$ well-describe (to better than 1\%) the fiducial

260: case and those eight other cases with variations on $B_*$, $R_*$,

261: $\dot M_{\rm w}$, and $M_*$.  Although this is only approximately

262: valid for situations with different wind acceleration rates or

263: different rotation rates (in which case the values of $K$ and $m$ are

264: slightly different; see Paper~II), the formulation of equation

265: \ref{eqn_rasim} serves well as an indication of the approximate

266: dependence of the stellar wind on parameters, which will be important

267: for discussing a wide range of possible conditions.

268: 

269: The form of equation (\ref{eqn_rasim}) is similar to that derived by

270: (e.g.)\ \citet[][and see \citealp{ferreira97}]{pelletierpudritz92} for

271: the general theory of centrifugally driven disk winds.  The quantity

272: in brackets measures the magnetization of the wind.  By assuming that

273: the Alfv\'en speed $v_{r,A}$ (at the Alfv\'en radius) is directly

274: proportional to $\Omega_*r_A$, a relation of the kind given by

275: equation (\ref{eqn_rasim}) can be derived \citep[e.g., see equation

276:   2.27 of][]{pelletierpudritz92}.  In that case, the value of the

277: index is $m=1/3$.  While this value is not far from the results of our

278: numerical simulations, the difference is significant.  One key reason

279: for this may be that disk winds are in the regime of so-called fast

280: magnetic rotators, whereas the rather slowly rotating TTS are either

281: slow magnetic rotators (where wind-driving forces dominate over

282: centrifugal ones) or are intermediate between these two regimes (see

283: Paper~II).

284: 

285: For the discussion that follows, it is useful to highlight how the

286: lever arm ($r_{\rm A}$) and wind torque responds to changing the mass

287: load (the mass loss rate) of the wind.  The fact that the Alfv\'en

288: lever arm in a hydromagnetic wind gets smaller as the mass load of the

289: outflow increases, as is seen in equation (\ref{eqn_rasim}), seems to

290: suggest that the wind would become ineffective.  This is certainly not

291: true however, because equation (\ref{eqn_tw}) assures that an increase

292: in wind mass loss rate leads to a net increase in the torque that the

293: wind exerts upon the star (the net wind torque scales as $\dot

294: M_w^{1-2m}$).  This is the basic reason why, by having an outflow rate

295: that is a substantial fraction of the accretion rate, an

296: accretion-powered stellar wind can be effective in countering the

297: accretion torque.

298: 

299: 

300: 

301: %While this value is not far from the results of our numerical

302: %simulations, we note that the disk winds are in the regime of

303: %so-called fast magnetic rotators.  The rather slowly rotating TTS

304: %however, are either slow magnetic rotators (where wind-driving forces

305: %dominate over centrifugal ones) or are intermediate between these two

306: %regimes.  Thus, the difference between $m\approx0.223$ and the disk

307: %wind theory of $m\approx0.333$, although not large, is important in

308: %that it reflects the difference between the relative importance of

309: %pressure vs. fling in driving outflows in these two regimes.  For the

310: %same magnetic field strength, winds with higher mass loss rates exert

311: %stronger spin-down torques upon their stars. This is the basic reason

312: %why an accretion powered stellar wind is so effective in countering

313: %the accretion torque; it has at its disposal a large stellar wind

314: %mass loss rate that scales with the accretion rate.

315: 

316: 

317: 

318: \begin{deluxetable}{lcccccc}

319: \tablewidth{0pt}

320: \tablecaption{Stellar Wind Alfv\'en Radii and Comparison to Star-Disk Spin-Down Torques \label{tab_torques}}

321: \tablehead{

322: \colhead{Case} &

323: \colhead{$r_{\rm A} / R_*$} &

324: \colhead{$\tau_{\rm w} / \tau_{\rm dsd}$} &

325: \colhead{$\tau_{\rm w} / \tau_{\rm dsd}$} \\

326: \colhead{} &

327: \colhead{} &

328: \colhead{$(\beta = 0.1)$} &

329: \colhead{$(\beta = 0.01)$} 

330: }

331: 

332: %\hline

333: %\tableline

334: %\cline{}

335: 

336: % This is all assuming gammac = 1.0 and beta = 0.1

337: 

338: \startdata

339: 

340: 

341: fiducial                         &  6.97  &  59  &  490  \\

342: $f$ = 0.2                        &  6.26  &  23  &  200  \\

343: $f$ = 0.05                       &  7.65  &  140 &  1200 \\

344: $B_*$ = 400 G                    &  9.55  &  27  &  230  \\

345: $B_*$ = 2 kG                     &  19.3  &  4.6 &  39   \\

346: low $\dot M_{\rm w}$$^a$         &  11.8  &  17  &  140  \\

347: very low $\dot M_{\rm w}$$^a$    &  16.7  &  6.7 &  57  \\

348: $R_*$ = 1.5 $R_\odot$            &  5.96  &  86  &  730  \\

349: $R_*$ = 3 $R_\odot$              &  8.75  &  34  &  280  \\

350: $M_*$ = 0.25 $M_\odot$           &  7.52  &  49  &  410  \\

351: $M_*$ = 1 $M_\odot$              &  6.42  &  70  &  590  \\

352: $c_{\rm s}/v_{\rm esc}$ = 0.245  &  6.64  &  53  &  440  \\

353: $c_{\rm s}/v_{\rm esc}$ = 0.192  &  7.23  &  63  &  530  \\

354: $\gamma$ = 1.10                  &  7.79  &  72  &  610  \\

355: 

356: 

357: \enddata

358: 

359: \footnotetext{$^a$ The mass outflow rate in the low and very low $\dot

360: M_{\rm w}$ cases is $1.9 \times 10^{-10}$ and $3.8 \times 10^{-11}

361: M_\odot$ yr$^{-1}$, respectively.}

362: 

363: \end{deluxetable}

364: 

365: 

366: It is our goal here to compare the stellar wind torque to the torque

367: expected to arise from the star-disk interaction, and the latter has

368: only been determined thus far for a dipolar geometry.  So we only

369: consider here the cases from Paper~II with a dipole magnetic field.

370: We also adopt the following assumptions.  Paper~II indicated that the

371: details of the wind driving have a relatively small, but not entirely

372: negligible, effect on the stellar wind torque.  In the absence of a

373: detailed model for how APSWs are driven, we assume that the velocity

374: profile of an APSW does not differ substantially from our simulations

375: (Paper~II), so that the calculated torques are valid.  Secondly,

376: Paper~II considered winds from isolated stars.  Here, we will use the

377: computed torques to develop the APSW scenario, in which stellar winds

378: are accompanied by disk winds, accretion flows, and the general

379: star-disk interaction (see figure 1 of Paper~I).  In reality, the

380: accretion disk blocks a portion of the stellar wind, and it is not

381: clear how much this will affect the stellar wind torque.  For the

382: present study, we will assume that the presence of the disk and

383: accretion will not significantly influence the stellar wind torques as

384: computed in Paper~II.

385: 

386: 

387: 

388: %Our preliminary simulations with stellar winds and disks (not

389: %presented here) indicate that adding an accretion disk to the

390: %simulations significantly complicates things, as there is often no

391: %steady-state solution, there is mixing between stellar and disk

392: %material, and there are an increased number of parameters.

393: 

394: %As adding a disk to the simulations too greatly complicates our

395: %method, we have no choice but to assume that the presence of a disk

396: %does not drastically modify the wind.

397: 

398: %Our preliminary test simulations with stellar winds and accretion

399: %disks (not presented here) suggests that the presence of the disk

400: %affects the stellar wind torques at a level of a few times 10\%.

401: 

402: 

403: 

404: \section{Stellar Wind vs.\ Star-Disk Spin-Down Torque} \label{sec_tdsd}

405: 

406: 

407: 

408: The magnetic interaction between the star and disk results in angular

409: momentum transfer between the two.  All models that calculate the

410: torque on the star from this interaction are based on the framework

411: constructed by \citet{ghoshlamb78}.  In this general model, some of the

412: stellar magnetic dipole flux connects to the accretion disk and

413: conveys torques between the star and disk.  The net torque can be

414: separated into a spin-up part that adds angular momentum to the star

415: and a spin-down part that removes angular momentum from the star,

416: giving it back to the disk.  In the absence of a stellar wind, a

417: spin-down torque only arises when there is a magnetic connection

418: between the star and the region of the disk outside the corotation

419: radius,

420: \begin{eqnarray}

421: \label{eqn_rco}

422: %

423: R_{\rm co} \equiv f^{-2/3} R_*.

424: %

425: \end{eqnarray}

426: It is assumed that the disk is capable of transporting away the excess

427: angular momentum it receives from the star.  The goal of this section

428: is to compare the stellar wind torque to the spin-down torque arising

429: from the star-disk magnetic connection, to determine under which

430: circumstances each of these torques may be important and aid in the

431: angular momentum loss from the star.

432: 

433: To calculate the star-disk spin-down torque, $\tau_{\rm dsd}$, we

434: follow \citet[][hereafter MP05b]{mattpudritz05}, who formulated a

435: Ghosh \& Lamb type model that includes the effect of the opening of

436: magnetic field lines via the differential rotation between the star

437: and disk.  In this case, $\tau_{\rm dsd}$ is calculated by considering

438: only the magnetic flux that remains closed (connected), parametrized

439: as having an azimuthal twist of less than a critical angle,

440: $\tan^{-1}(\gamma_{\rm c})$.  Here we adopt the value of $\gamma_{\rm

441: c} = 1$ suggested by \citet{uzdensky3ea02}.  By combining equation (9)

442: and (22) of MP05b with equation (\ref{eqn_rco}), we find

443: \begin{eqnarray}

444: \label{eqn_tdsd}

445: %

446: \tau_{\rm dsd} = - {\chi(\beta) \over 3} f^2 B_*^2 R_*^3,

447: %

448: %\tau_{\rm d} \approx - B_*^2 R_*^3 (\chi/3) (R_* / R_{\rm co})^3,

449: \end{eqnarray}

450: where 

451: \begin{eqnarray}

452: \label{eqn_chi}

453: %

454: \chi(\beta) \equiv \beta^{-1} [1 + (1+\beta)^{-2} - 2(1+\beta)^{-1}]

455: %

456: \end{eqnarray}

457: is a dimensionless function of the strength of the effective magnetic

458: diffusion rate in the disk.  This is parametrized by $\beta$, which

459: for a standard $\alpha$-disk \citep{shakurasunyaev73} is

460: $\beta$~$\equiv$~$\alpha h (P_{\rm t} r)^{-1}$, where $\alpha$ has its

461: usual meaning, $h$ is the disk scale height at radius $r$, and $P_{\rm

462: t}$ is the turbulent magnetic Prandtl number\footnote{Note that this

463: $\beta$ has no relation to the usual ``plasma beta'' parameter that

464: often appears in MHD studies.}.  Small values of $\beta$ correspond to

465: strong coupling, and MP05b suggested that $\beta \sim 0.01$ was

466: appropriate for real disks.  For strong coupling (small $\beta$), the

467: magnetic field will be highly twisted azimuthally, leading to more

468: open flux and a weaker star-disk spin-down torque.  For small $\beta$,

469: $\chi(\beta) \approx \beta$, but $\chi(\beta)$ has a maximum value of

470: 0.25 when $\beta = 1$ (e.g., see fig.\ 7 of MP05b).

471: 

472: Equation (\ref{eqn_tdsd}) indicates that the star-disk spin-down

473: torque is completely independent of the accretion rate.  This torque

474: only requires that there exists a Keplerian disk outside $R_{\rm co}$,

475: to which the star can connect, and it does not matter whether or not

476: there is net accretion onto the star.  The dependence on the stellar

477: spin rate $f$ is due to the fact that when the star spins faster,

478: $R_{\rm co}$ is closer to the star, where the magnetic field is

479: stronger.

480: 

481: We compare the stellar wind torque computed in our simulations to the

482: star-disk spin-down torque by listing the ratio $\tau_{\rm w} /

483: \tau_{\rm dsd}$ in the last two columns of table \ref{tab_torques}.

484: We consider both a case with $\beta = 0.1$, resulting in $\chi \approx

485: 0.0826$, and a case with $\beta = 0.01$, resulting in $\chi \approx

486: 0.00980$.  It is apparent that the simulated stellar wind torques are

487: tens to hundreds of times greater than $\tau_{\rm dsd}$, the larger

488: difference existing for smaller values of $\beta$.

489: 

490: Thus, for the simulated winds, we see empirically that the stellar

491: wind is much more effective at spinning down the star than is the

492: star-disk connection.  This can be understood qualitatively as

493: follows.  In the stellar magnetosphere, any torque on the star is

494: primarily conveyed by the azimuthal twisting of its magnetic field.

495: In the case of the magnetic field connecting the star to the disk,

496: there is a limit to how much the magnetic field can be twisted before

497: the connection is lost \citep[][MP05b]{uzdensky3ea02}.  In the case of

498: a stellar wind flowing along the field, there is no such limit on the

499: twist.  The larger the mass outflow rate in the wind, the less capable

500: is the magnetic field to keep wind material corotating with the star,

501: and so the larger will be the twist of the field.

502: 

503: To show the dominance of the stellar wind torque more generally and to

504: identify the circumstances under which it might not dominate, we will

505: use the semi-analytic formulation of equation (\ref{eqn_rasim}).  By

506: combining equations (\ref{eqn_f}) -- (\ref{eqn_rasim}), one obtains

507: \begin{eqnarray}

508: \label{eqn_twind2}

509: %

510: \tau_{\rm w} = - {K^{1/m} \over \sqrt{2}}

511:      \left({R_* \over r_{\rm A}}\right)^{1/m - 2} f B_*^2 R_*^3

512: %

513: \end{eqnarray}

514: for the stellar wind torque.  At first it may seem unusual that

515: $\tau_{\rm w}$ is weaker when the magnetic lever arm length, $r_{\rm

516:   A}$, is larger (for fixed $f B_*^2 R_*^3$).  However this simply

517: indicates that the stellar wind torque increases with increasing $\dot

518: M_{\rm w}$, as noted in section \ref{sec_simulations}.  Also, note

519: that a weak $f$-dependence of $r_{\rm A}$ is not characterized in

520: equation (\ref{eqn_rasim}) (see Paper~II), so the exact dependence of

521: the torque on $f$ is not captured in equation (\ref{eqn_twind2}).

522: 

523: 

524: 

525: \begin{figure}

526: \epsscale{1.15}

527: \plotone{f1.eps}

528: %\plotone{figs/torqueratio.eps}

529: 

530: \caption{The ratio of the stellar wind torque to the spin-down portion

531: of the star-disk interaction torque versus the magnetic lever arm

532: length in the stellar wind.  When the ratio is much greater than one,

533: the stellar wind is most important for angular momentum loss from the

534: star.  The lines correspond to equation (\ref{eqn_ratio}), assuming a

535: stellar spin rate of $f = 0.1$ and three different values of the

536: factor $\chi(\beta)$ (see text), corresponding to $\beta$ = 0.01, 0.1,

537: and 1, as indicated.  The values listed in table \ref{tab_torques},

538: obtained by comparing simulated wind torques to the analytic disk

539: torques, are plotted as squares (for $\beta = 0.1$) and triangles

540: ($\beta = 0.01$).  The figure indicates that, unless the magnetic

541: lever arm length is very long (e.g., for very low stellar wind mass

542: loss rate), the spin-down torque from the disk is negligible.

543: \label{fig_torqueratio}}

544: 

545: \end{figure}

546: 

547: 

548: By combining equations (\ref{eqn_tdsd}) and (\ref{eqn_twind2}), using

549: $K = 2.11$ and $m = 0.223$ (see \S \ref{sec_simulations}), one finds

550: \begin{eqnarray}

551: \label{eqn_ratio}

552: {\tau_{\rm w} \over \tau_{\rm dsd}} \approx 6.0 \times 10^{4}

553:       \left({\chi(\beta) \over 10^{-2}}\right)^{-1} 

554:       \left({f \over 0.1}\right)^{-1} 

555:       \left({R_* \over r_{\rm A}}\right)^{2.48}.

556: %

557: \end{eqnarray}

558: The lines in figure \ref{fig_torqueratio} show equation

559: (\ref{eqn_ratio}) for $f=0.1$ and for three different values of

560: $\beta$.  This includes a line for $\beta=1$, which corresponds to the

561: strongest possible star-disk spin-down torque (as discussed by MP05b).

562: Even in this case, the fiducial stellar wind torque is $\sim 20$ times

563: stronger than the star-disk spin-down torque.  For smaller, more

564: realistic values of $\beta$, the star-disk spin-down torque only

565: becomes weaker, while the stellar wind torque is not affected.  It is

566: clear that for the parameters considered here, the angular momentum

567: extracted by the stellar wind completely dominates over that which can

568: be transferred from the star to the disk.

569: 

570: The stellar wind torque becomes weaker relative to $\tau_{\rm dsd}$

571: when $r_{\rm A}$ is larger (e.g., for smaller $\dot M_{\rm w}$) or for

572: more rapidly spinning stars (larger $f$).  For the case of $\beta =

573: 0.01$, favored by MP05b, $\tau_{\rm w}$ will be larger than $\tau_{\rm

574:   dsd}$ for a star with $f=0.1$, as long as $r_{\rm A} \la 84 R_*$.

575: This limiting value is much longer than any of the lever arm lengths

576: listed in table \ref{tab_torques}.  As an example, for all else being

577: equal to the fiducial case, equation (\ref{eqn_rasim}) suggests that

578: $\tau_{\rm w} > \tau_{\rm dsd}$, as long as $\dot M_{\rm w} \ga 3

579: \times 10^{-14} M_\odot$~yr$^{-1}$.  This limit is comparable to the

580: solar wind mass loss rate.  If the stellar dipole field strength is

581: instead $B_*=2$ kG, the limit becomes $\dot M_{\rm w} \ga 3 \times

582: 10^{-12} M_\odot$.

583: 

584: The squares and triangles in figure \ref{fig_torqueratio} represent

585: the data from table \ref{tab_torques}.  Note that nine of the data

586: points (for each $\beta$) match the line very well.  This is expected

587: since the results from these cases were used to obtain the value of

588: $K$ and $m$ used in equation (\ref{eqn_ratio}).  There are five cases

589: (for each $\beta$) that lie slightly off of the line.  Three of them

590: represent the last three cases listed in table \ref{tab_torques}, and

591: the other two are the $f=0.2$ and $f=0.05$ cases (which, for the plot,

592: we have scaled by a factor of $f/0.1$, to take into account the spin

593: dependence of equation (\ref{eqn_ratio})).  These five cases are not

594: expected to match exactly since equation (\ref{eqn_rasim}) is not

595: precise for cases with different wind driving or spin rates than the

596: fiducial case (see Paper~II).  Thus, the scatter of these five cases

597: around the line indicates a sort of uncertainty of the semi-analytic

598: formula for the stellar wind, due to variations in the wind driving

599: mechanism and stellar spin rate.  It is evident from the figure that

600: this uncertainty does not affect the main conclusion here that stellar

601: wind torques dominate the spin down of the star.

602: 

603: Thus, for the slow rotators ($f \sim 0.1$), we conclude that a stellar

604: wind will transport much more angular momentum from the star than will

605: a magnetic connection to the disk, as long as the stellar wind mass

606: outflow rate is substantially larger than the solar wind mass outflow

607: rate.  For the systems considered here, with $\do M_{\rm w} \sim

608: 10^{-9} M_\odot$~yr$^{-1}$, the stellar wind torque completely

609: dominates over any other spin-down torque felt by the star.  Since

610: $\tau_{\rm dsd}$ is negligible, the only important torques on these

611: stars are the spin-down torque from the stellar wind and the spin-up

612: portion of the star-disk interaction torque.  We compare these two

613: torques in the following section.

614: 

615: %$\dot M_{\rm w} \ga 2.7 \times 10^{-14} M_\odot$ yr$^{-1}$, comparable

616: 

617: 

618: 

619: \section{Spin Equilibrium by an APSW} \label{sec_equilibrium}

620: 

621: 

622: 

623: Section \ref{sec_tdsd} revealed that, for the slow rotators with

624: substantial stellar winds, the spin-down torque felt by the star from

625: the star-disk interaction is negligible.  Thus, the spin state of the

626: star is characterized as a competition between the spin-up component

627: of the star-disk interaction torque and the spin-down by the stellar

628: wind.  If a system's parameters are measured, the theory can be used

629: to determine the net torque on the star.

630: 

631: Given enough time ($\sim 10^5$--$10^6$ yr for CTTSs), the stellar spin

632: should approach an equilibrium spin state in which the net torque on

633: the star is zero \citep[see, e.g.,][]{cameroncampbell93,

634:   armitageclarke96}.  The variability observed in accreting systems

635: \citep[e.g.,][]{hartmann97} suggests that spin equilibrium may only

636: represent a time-averaged state, and the condition of net zero torque

637: simply identifies where the net torque changes sign.  In any case, it

638: is instructive to examine the conditions of spin-equilibrium.

639: 

640: In sections \ref{sec_balancing} -- \ref{sec_power}, we examine the

641: expected spin-equilibrium state of the specific cases of stellar winds

642: simulated in Paper~II.  In section \ref{sec_semianalytic}, we use the

643: semi-analytic formulation of equation \ref{eqn_rasim} to make more

644: general conclusions.

645: 

646: 

647: 

648: %     \subsection{Balancing the Accretion Torque} \label{sec_balancing}

649:      \subsection{Spin Equilibrium for Specific Cases} \label{sec_balancing}

650: 

651: 

652: 

653: The spin-up portion of the star-disk interaction torque comes

654: primarily from the accretion of material from the innermost part of

655: the disk onto the star.  In this section, we examine some specific

656: cases of spin equilibrium, by determining under what conditions this

657: spin-up torque balances the stellar wind torques for the simulations

658: listed in table \ref{tab_torques}.

659: 

660: When a star's magnetic field is strong enough, it will disrupt the

661: Keplerian disk at some radius, $R_{\rm t}$, the disk truncation

662: radius.  We calculate the location of $R_{\rm t}$ using the method and

663: equations contained in the Appendix, which follows MP05b.  From

664: $R_{\rm t}$, accreting material is channeled by the magnetic field to

665: the surface of the star.  There is a torque associated with the

666: truncation of the disk and the accretion of material from $R_{\rm t}$.

667: This torque, hereafter the ``accretion torque,'' is given by MP05b as

668: \begin{eqnarray}

669: \label{eqn_ta}

670: %

671: \tau_{\rm a} = \dot M_{\rm a} \sqrt{G M_* R_*} 

672:      \left[{\left({R_{\rm t} \over R_*}\right)^{1/2} - k^2 f}\right],

673: %

674: \end{eqnarray}

675: where $\dot M_{\rm a}$ is the mass accretion rate onto the stellar

676: surface and $k$ is the normalized radius of gyration of the star (we

677: assume $k^2 \approx 0.2$; \citealp{armitageclarke96}).  Equation

678: (\ref{eqn_ta}) assumes that all of the Keplerian specific angular

679: momentum of the disk material near $R_{\rm t}$ is transferred to the

680: star.  This naturally follows from the dynamical truncation of the

681: disk (e.g., \citealp{yi95, wang95}; MP05b) and is supported by

682: numerical simulations \citep[e.g.,][]{romanovaea02, long3ea05}.  It is

683: clear that the accretion torque depends both on $\dot M_{\rm a}$ and

684: $R_{\rm t}$.  At the same time, as shown in the Appendix, the location

685: of $R_{\rm t}$ itself depends on most of the parameters, including

686: $\dot M_{\rm a}$ and $\beta$.

687: 

688: 

689: 

690: %This is supported both by physical reasoning (e.g., MP05b) and by

691: %numerical simulations \citep[e.g.,][]{romanovaea02, long3ea05}.

692: 

693: 

694: 

695: \begin{deluxetable}{lcccc} 

696: \tablewidth{0pt}

697: \tablecaption{Spin-Equilibrium Results for $\gamma_{\rm c} = 1$ \& 

698:               $\beta = 0.1$ \label{tab_equilib0.1}}

699: \tablehead{

700: \colhead{Case} &

701: \colhead{$\dot M_{\rm w} / \dot M_{\rm a}$} &

702: \colhead{$R_{\rm t} / R_*$} &

703: \colhead{State} &

704: \colhead{$\epsilon_\infty$}

705: }

706: 

707: %\hline

708: %\tableline

709: %\cline{}

710: 

711: % This is all assuming gammac = 1.0 and beta = 0.1

712: 

713: \startdata

714: 

715: fiducial                         &  0.43  &  4.4  &  2  &  0.71   \\

716: $f$ = 0.2                        &  0.21  &  2.9  &  2  &  0.39   \\

717: $f$ = 0.05                       &  0.83  &  5.9  &  1  &  1.2    \\

718: $B_*$ = 400 G                    &  0.23  &  4.5  &  2  &  0.37   \\

719: $B_*$ = 2 kG                     &  0.057 &  4.6  &  2  &  0.055  \\

720: low $\dot M_{\rm w}$             &  0.15  &  4.6  &  2  &  0.23   \\

721: very low $\dot M_{\rm w}$        &  0.076 &  4.6  &  2  &  0.094  \\

722: $R_*$ = 1.5 $R_\odot$            &  0.58  &  4.3  &  2  &  0.96   \\

723: $R_*$ = 3 $R_\odot$              &  0.28  &  4.5  &  2  &  0.46   \\

724: $M_*$ = 0.25 $M_\odot$           &  0.37  &  4.5  &  2  &  0.61   \\

725: $M_*$ = 1 $M_\odot$              &  0.51  &  4.4  &  2  &  0.84   \\

726: $c_{\rm s}/v_{\rm esc}$ = 0.245  &  0.48  &  4.4  &  2  &  0.87   \\

727: $c_{\rm s}/v_{\rm esc}$ = 0.192  &  0.40  &  4.4  &  2  &  0.61   \\

728: $\gamma$ = 1.10                  &  0.34  &  4.4  &  2  &  0.48   \\

729: 

730: 

731: \enddata

732: 

733: \end{deluxetable}

734: 

735: 

736: 

737: The spin equilibrium state is defined by 

738: \begin{eqnarray}

739: \label{eqn_teq}

740: %

741: \tau_{\rm a} = - \tau_{\rm w}.

742: %

743: \end{eqnarray}

744: Each of our wind simulation cases represents a specific set of values

745: for $\tau_{\rm w}$, $\dot M_{\rm w}$, $B_*$, $M_*$, $R_*$, and

746: $\Omega_*$.  For each simulation case, we used equation (\ref{eqn_ta})

747: and the method in the Appendix to determine the values of $R_{\rm t}$

748: and $\dot M_{\rm a}$ such that the condition (\ref{eqn_teq}) is

749: satisfied. We consider both a case with $\beta = 0.1$ and a case with

750: $\beta = 0.01$ (see \S \ref{sec_tdsd}).  The results, given as $\dot

751: M_{\rm w} / \dot M_{\rm a}$ and $R_{\rm t} / R_*$, are listed in the

752: 2nd and 3rd columns of tables \ref{tab_equilib0.1} (for $\beta = 0.1$)

753: and \ref{tab_equilib0.01} (for $\beta = 0.01$).

754: 

755: A comparison between tables \ref{tab_equilib0.1} and

756: \ref{tab_equilib0.01} reveals that the disk magnetic coupling

757: parameter $\beta$ has little influence on the equilibrium values of

758: $\dot M_{\rm a}$ and $R_{\rm t}$.  This demonstrates that, although

759: $\beta$ has a large influence on the (negligible) spin-down part of

760: the star-disk interaction torque (as shown in \S \ref{sec_tdsd}),

761: $\beta$ has very little influence on the spin-up part.

762: 

763: For the specific cases of the simulated stellar winds, it is clear

764: that the equilibrium spin state is characterized by $\dot M_{\rm w} /

765: \dot M_{\rm a}$ of typically a few tens of percent.  This ratio is

766: smaller for cases with larger $r_{\rm A}$ (e.g., for larger field

767: strength or smaller $\dot M_{\rm w}$).  Thus, the cases listed in the

768: tables confirm the general conclusion of Paper~I, and represent valid

769: torque solutions for the spin equilibrium state.

770: 

771: 

772: 

773:      \subsection{Magnetic Connection State of the System} \label{sec_state}

774: 

775: 

776: 

777: As pointed out by MP05b (and see the Appendix), an accreting system

778: may exist in a state where the stellar magnetic field connects to the

779: disk outside the corotation radius, which they call ``state 2,'' and

780: which we implicitly assumed in section \ref{sec_tdsd}.  On the other

781: hand, if the disk truncation radius is sufficiently smaller than

782: $R_{\rm co}$, the star can lose its magnetic connection to all but the

783: very inner edge of the disk, which they call ``state 1.''  The

784: determination of $R_{\rm t}$ is different in the two states.  In the

785: absence of a stellar wind torque, a star in spin equilibrium must be

786: characterized by state 2 (MP05b).  Thus, having $R_{\rm t}$ very close

787: to $R_{\rm co}$ is a requirement of the ``disk locking'' models

788: \citep{konigl91, ostrikershu95, wang95}.  By contrast, this is not a

789: requirement of the APSW scenario.

790: 

791: 

792: 

793: \begin{deluxetable}{lcccc} 

794: \tablewidth{0pt}

795: \tablecaption{Spin-Equilibrium Results for $\gamma_{\rm c} = 1$ \& 

796:               $\beta = 0.01$ \label{tab_equilib0.01}}

797: \tablehead{

798: \colhead{Case} &

799: \colhead{$\dot M_{\rm w} / \dot M_{\rm a}$} &

800: \colhead{$R_{\rm t} / R_*$} &

801: \colhead{State} &

802: \colhead{$\epsilon_\infty$} 

803: }

804: 

805: %\hline

806: %\tableline

807: %\cline{}

808: 

809: % This is all assuming gammac = 1.0 and beta = 0.01

810: 

811: \startdata

812: 

813: 

814: fiducial                         &  0.44  &  4.6  &  2  &  0.73   \\

815: $f$ = 0.2                        &  0.21  &  2.9  &  2  &  0.39   \\

816: $f$ = 0.05                       &  0.83  &  5.9  &  1  &  1.2    \\

817: $B_*$ = 400 G                    &  0.24  &  4.6  &  2  &  0.39   \\

818: $B_*$ = 2 kG                     &  0.057 &  4.6  &  2  &  0.055  \\

819: low $\dot M_{\rm w}$             &  0.15  &  4.6  &  2  &  0.23   \\

820: very low $\dot M_{\rm w}$        &  0.077 &  4.6  &  2  &  0.096  \\

821: $R_*$ = 1.5 $R_\odot$            &  0.59  &  4.4  &  1  &  0.98   \\

822: $R_*$ = 3 $R_\odot$              &  0.28  &  4.6  &  2  &  0.46   \\

823: $M_*$ = 0.25 $M_\odot$           &  0.38  &  4.6  &  2  &  0.63   \\

824: $M_*$ = 1 $M_\odot$              &  0.52  &  4.6  &  2  &  0.86   \\

825: $c_{\rm s}/v_{\rm esc}$ = 0.245  &  0.49  &  4.6  &  2  &  0.88   \\

826: $c_{\rm s}/v_{\rm esc}$ = 0.192  &  0.41  &  4.6  &  2  &  0.62   \\

827: $\gamma$ = 1.10                  &  0.35  &  4.6  &  2  &  0.49   \\

828: 

829: 

830: \enddata

831: 

832: \end{deluxetable}

833: 

834: 

835: 

836: Following MP05b (using equation \ref{eqn_trans}), we determined the

837: magnetic connection state of the spin-equilibrium systems described

838: above and listed this in the 4th column of tables \ref{tab_equilib0.1}

839: and \ref{tab_equilib0.01}.  Note that MP05b only consider a loss of

840: magnetic connection via the differential twisting of field lines.  The

841: stellar wind should also influence the connectedness between the star

842: and disk \citep{safier98}, but we do not attempt to quantify this

843: here.

844: 

845: Tables \ref{tab_equilib0.1} and \ref{tab_equilib0.01} reveal that most

846: (though not all) of the simulated cases are in a magnetic connection

847: state 2, while in spin-equilibrium.  A characteristic of this state is

848: that $R_{\rm t}$ is very close to $R_{\rm co}$, which is also evident

849: in the tables ($R_{\rm co} / R_*$ = 7.4, 4.6, and 2.9 for $f$ = 0.05,

850: 0.1, and 0.2, respectively).  It appears that, when a stellar wind

851: torque balances the accretion torque, it may be common (though not

852: required) for the disc truncation radius to be close to the corotation

853: radius (unless the spin rate is substantially less than $f = 0.1$).

854: 

855: 

856: 

857:      \subsection{Accretion Power} \label{sec_power}

858: 

859: 

860: 

861: In the APSW scenario proposed in Paper~I, the energy that powers the

862: stellar wind ultimately comes from the energy released by the

863: accretion process.  In this section, we calculate what fraction of the

864: accretion power would be required to drive the wind, in the specific

865: cases for which we have determined the spin-equilibrium.

866: 

867: In order to tabulate the accretion power, the precise details of the

868: complicated interaction between the star and disk are not important.

869: The general behavior is that material from the Keplerian disk becomes

870: attached to the stellar magnetosphere near $R_{\rm t}$ and eventually

871: falls onto and becomes part of the star.  Energetically, this can be

872: treated as an inelastic process, wherein only the energy content

873: before and after the interaction needs to be specified.  Thus, the

874: rate of potential energy release is simply 1/2 $ \dot M_{\rm a} v_{\rm

875:   esc}^2 (1 - R_*/R_{\rm t})$.  Note that as accreting material piles

876: onto the stellar surface, there should be additional energy released

877: as material either ``sinks" into (convectively) or compresses the

878: star.  This is another potential energy source, but we neglect this

879: here.  The rate of (rotational) kinetic energy release is 1/4 $ \dot

880: M_{\rm a} v_{\rm esc}^2 (R_*/R_{\rm t} - k^2 f^2)$, where the last

881: term assumes that accreting material eventually achieves the same

882: specific angular momentum as the star.  The difference in thermal

883: energy density between material at the disk inner edge and material at

884: stellar photospheric temperature is negligible compared to the

885: potential and kinetic energy release.  Thus, by summing the potential

886: and kinetic energies, the rate of energy release in the vicinity of

887: the star is approximately

888: \begin{eqnarray}

889: \label{eqn_edota}

890: %

891: \dot E_{\rm a} = {1 \over 2} \dot M_{\rm a} v_{\rm esc}^2

892:        \left({1 - {1 \over 2} {R_* \over R_{\rm t}} - {1 \over 2} k^2 f^2}\right).

893: %\dot E_{\rm a} = {1 \over 2} \dot M_{\rm a} v_{\rm esc}^2

894: %       \left[{1 - {1\over2} {R_*\over R_{\rm t}} - 

895: %	      f \left({R_{\rm t} \over R_*}\right)^{1/2}}\right].

896: %\dot E_{\rm a} = {1 \over 2} \dot M_{\rm a} v_{\rm esc}^2

897: %       \left[{1 - {1\over2} {R_*\over R_{\rm t}} - 

898: %	      f \left({R_{\rm t} \over R_*}\right)^{1/2}

899: %	      + f^2 \left({k^2 - {1\over2}}\right)k^2}\right]

900: %

901: \end{eqnarray}

902: Since there is an accretion torque on the star, some of this energy is

903: added to the rotational energy of the star at a rate $\Omega_*

904: \tau_{\rm a}$.  The remaining energy ($\dot E_{\rm a} - \Omega_*

905: \tau_{\rm a}$) is available to power other accretion-related

906: activity\footnote{We follow a similar derivation of the accretion

907:   power to that of Paper~I, where equation (4) of that work

908:   corresponds to the remaining energy, $\dot E_{\rm a} - \Omega_*

909:   \tau_{\rm a}$, and neglects terms proportional to $f^2$.}.  In

910: particular, this remaining energy is responsible for powering the

911: observed excess continuum emission (such as the UV excess) and line

912: emission \citep[e.g.,][]{konigl91, calvetgullbring98, muzerolle3ea01},

913: in addition to driving an enhanced stellar wind (Paper~I).

914: 

915: Paper~I proposed that a fraction $\epsilon$ of this energy

916: specifically powers the thermal energy in the stellar wind.  Our

917: simulated winds are thermally driven, but the wind driving mechanism

918: at work in real systems is still uncertain (see Paper~II).  Thus, we

919: wish to calculate the power required to drive the wind, in a generic

920: form.  For this, we simply calculate the total energy in the wind far

921: from the star plus the potential energy required to lift the wind off

922: the stellar surface.  In this way, the power in a steady-state, 2.5D,

923: MHD wind can be obtained by \citep[see, e.g.,][]{ustyugovaea99,

924:   keppensgoedbloed00}

925: \begin{eqnarray}

926: \label{eqn_edotw}

927: %

928: \dot E_{\rm w} = 4 \pi R^2 \int_0^1 \rho v_{\rm R} E^\prime d(\cos \theta)

929:                  + {1 \over 2} \dot M_{\rm w} v_{\rm esc}^2,

930: %

931: \end{eqnarray}

932: where $\theta$ is the usual spherical coordinate and

933: \begin{eqnarray}

934: \label{eqn_e}

935: %

936: E^\prime \equiv {v_{\rm p}^2 + v_\phi^2 \over 2} + {B_\phi^2 \over 4

937:   \pi \rho} - {v_\phi B_\phi B_{\rm p} \over 4 \pi \rho v_{\rm p}}.

938: %

939: \end{eqnarray}

940: In equation (\ref{eqn_e}), we have neglected the thermal and

941: gravitational potential energy, so the integral in equation

942: (\ref{eqn_edotw}) should be evaluated at large $R$, where $E^\prime$

943: has reached an asymptotic value and these energies are negligible.

944: Thus, $\dot E_{\rm w}$ represents the total power required to lift

945: material off of the star, to accelerate it to the wind velocity, and

946: to provide the magnetic energy content carried with the wind.

947: 

948: For each of our simulated wind solutions, we evaluate the integral in

949: equation (\ref{eqn_edotw}) at a radius of $R = 50 R_*$, where

950: $E^\prime$ is within a few percent of its asymptotic value.  The spin

951: of the star does work on the wind at a rate $\Omega_* \tau_{\rm w}$.

952: This represents the power injected in the wind by magnetocentrifugal

953: processes.  We find that the ratio $\Omega_* \tau_{\rm w} / \dot

954: E_{\rm w}$ is 30\% in the fiducial case.  In most other cases, the

955: value of this ratio falls the range 10--60\%.  This indicates that, as

956: discussed by Paper~I and II \citep[and see][]{washimishibata93}, these

957: winds are in a regime where the magnetocentrifugal effects are of

958: nearly equal importance with the other source of wind driving.

959: 

960: It is this other source of wind driving that we propose is powered by

961: some fraction of the available accretion energy.  We define this

962: fraction as\footnote{This fraction is a more general definition than

963: $\epsilon$ in Paper~I, which assumes thermal wind driving. By

964: contrast, $\epsilon_\infty$ is the fraction of the accretion power

965: required to explain the energy in the wind at large distances from the

966: star, regardless of the driving mechanism.}

967: \begin{eqnarray}

968: \label{eqn_epsilon}

969: %

970: \epsilon_\infty \equiv {{\dot E_{\rm w} - \Omega_* \tau_{\rm w}} \over 

971:             {\dot E_{\rm a} - \Omega_* \tau_{\rm a}}}.

972: %

973: \end{eqnarray}

974: This represents the minimum fraction of the accretion power required

975: to drive the stellar wind, since whatever mechanism drives the wind

976: will not itself likely be 100\% efficient \citep{decampli81}.

977: 

978: In the last column of tables \ref{tab_equilib0.1} and

979: \ref{tab_equilib0.01}, we list the value of $\epsilon_\infty$ for each

980: case in spin-equilibrium.  In one case ($f = 0.05$), $\epsilon_\infty$

981: is greater than 100\%, indicating that there is not enough accretion

982: power in the spin equilibrium state to power the wind.  This case is

983: therefore not an acceptable solution for a system in spin-equilibrium

984: by an APSW.  All of the other cases have $\epsilon_\infty < 1$, and so

985: they are energetically viable solutions.

986: 

987: There is a relationship between $\epsilon_\infty$, the observed excess

988: emission, and the inferred mass accretion rate.  In particular, the

989: accretion rates are typically determined by measuring excess emission

990: and assuming that all of the accretion power is radiated

991: \citep[e.g.,][]{calvetgullbring98}.  In the APSW scenario, some of the

992: accretion power drives the stellar wind, so only a fraction $1 -

993: \epsilon_\infty$ of the accretion power can be radiated.  This means

994: that the true accretion rate ($\dot M_{\rm a}$) will be a factor of

995: $(1 - \epsilon_\infty)^{-1}$ larger than the observationally

996: determined value.  In this context, and since $\epsilon_\infty$ is the

997: miminum required fraction to drive the wind, a value of

998: $\epsilon_\infty \ga 0.5$ (as for several cases listed in tables

999: \ref{tab_equilib0.1} and \ref{tab_equilib0.01}) appears quite large.

1000: However, the observational determination of $\dot M_{\rm a}$ is

1001: uncertain by a factor of several, as exemplified by the large range of

1002: measurements compiled by \citet{johnskrullgafford02}.  Thus, while it

1003: is clear that $\epsilon_\infty < 1$ is a hard upper limit, it is not

1004: yet clear how close to unity $\epsilon_\infty$ can be.

1005: 

1006: As expected, the cases with lower values of $\dot M_{\rm w} / \dot

1007: M_{\rm a}$ (i.e., cases with larger field strength or smaller $\dot

1008: M_{\rm w}$) require a smaller fraction of the accretion power to drive

1009: the wind.  The cases in the table suggest approximately that

1010: $\epsilon_\infty \approx 1.6 \dot M_{\rm w} / \dot M_{\rm a}$.  Thus,

1011: it appears that $\dot M_{\rm w} / \dot M_{\rm a} \la 0.6$ represents a

1012: hard upper limit for APSWs.

1013: 

1014: It is important to note that the manner in which the accretion power

1015: transfers to the stellar wind is still unspecified in the APSW model.

1016: This will depend upon what is the wind driving mechanism, which is

1017: currently unknown.  In reality, the physics of the energy coupling

1018: will likely determine the value of $\epsilon_\infty$, which

1019: effectively sets the value for $\dot M_{\rm w} / \dot M_{\rm a}$.

1020: Then, given enough time, the stellar spin rate will evolve toward the

1021: equilibrium value set primarily by $\dot M_{\rm w} / \dot M_{\rm a}$

1022: and $r_{\rm A}/R_*$.  Thus the spin-equilibrium state of the star is

1023: ultimately determined by the power coupling and magnetic properties,

1024: and more work is needed to take this further.

1025: 

1026: 

1027: 

1028:      \subsection{Semi-Analytic Formulation for Spin Equilibrium} 

1029: 

1030: \label{sec_semianalytic}

1031: 

1032: 

1033: 

1034: In order to develop a more general, predictive theory, in this section

1035: we make use of the semi-analytic formulation of equation

1036: (\ref{eqn_rasim}).  As justified by the previous sections, we assume

1037: that the equilibrium spin rate of the star is simply determined by a

1038: balance between the spin-up torque from accretion and the spin-down

1039: torque from the stellar wind.  Using equations (\ref{eqn_f}) --

1040: (\ref{eqn_rasim}), (\ref{eqn_ta}), and (\ref{eqn_teq}), we can write

1041: the stellar wind equilibrium spin rate, expressed as a fraction of

1042: breakup spin, as

1043: \begin{eqnarray}

1044: \label{eqn_fsw}

1045: %

1046: %f_{\rm sw} = K^{-2}

1047: %               \left({R_{\rm t} \over R_*}\right)^{1/2}

1048: %               \left({\dot M_{\rm a} \over \dot M_{\rm w}}\right)^{1-2m}

1049: %               \left({{\dot M_{\rm a} v_{\rm esc}} \over 

1050: %                      {B_*^2 R_*^2}}\right)^{2m},

1051: f_{\rm sw} = K^{-2}

1052:                \left({R_{\rm t} \over R_*}\right)^{1/2}

1053:                \left({\dot M_{\rm a} \over \dot M_{\rm w}}\right)^{1-2m}

1054: 	       \Psi^{-2m},

1055: %

1056: \end{eqnarray}

1057: where

1058: \begin{eqnarray}

1059: \label{eqn_psi}

1060: %

1061: \Psi \equiv  {{B_*^2 R_*^2} \over {\dot M_{\rm a} v_{\rm esc}}}

1062: %

1063: \end{eqnarray}

1064: is a dimensionless magnetization parameter\footnote{The magnetization

1065:   parameter $\Psi$ is related to $\psi$ used by MP05b (see their

1066:   eq.\ 16) by a constant factor, $\psi = 2^{3/2} \Psi$.}.  Here, we

1067: have neglected the term proportional to $k^2 f$ in equation

1068: (\ref{eqn_ta}), since it is generally much smaller than the other

1069: term.  Again, note that a weak $f$-dependence of $r_{\rm A}$ is not

1070: included in equation (\ref{eqn_rasim}), so the dependence of $f_{\rm

1071:   sw}$ on some of the parameters is not precisely captured in equation

1072: (\ref{eqn_fsw}) (see Paper~II).

1073: 

1074: Equation (\ref{eqn_fsw}) includes a dependence on the truncation

1075: radius of the disk, $R_{\rm t}$.  This location itself has a

1076: dependence on the other parameters, and the determination of $R_{\rm

1077: t}$ depends on the magnetic connection state of the system (\S

1078: \ref{sec_state}; MP05b).  In general, $R_{\rm t}$ depends on $\Psi$,

1079: but if $R_{\rm t}$ is close to the corotation radius, $R_{\rm co}$,

1080: then $R_{\rm t}$ also depends on the stellar spin rate.  We will

1081: consider two cases that are expected to bracket reality.

1082: 

1083: The first case is one in which the system is in state 1 as defined by

1084: MP05b.  Here, $R_{\rm t}$ does not depend on the stellar spin rate,

1085: and it is simply proportional to the original calculations by

1086: \citet{lamb3ea73} and \citet{davidsonostriker73} used in most Ghosh \&

1087: Lamb type models.  Thus, using equation (\ref{eqn_rt1}) for $R_{\rm t}$,

1088: adopting $\gamma_{\rm c} = 1$, and plugging in to equation

1089: (\ref{eqn_fsw}), one finds

1090: \begin{eqnarray}

1091: \label{eqn_fsw1}

1092: %

1093: f_{\rm sw1} = {{2^{3/14}} \over {K^{2}}}

1094:                \left({\dot M_{\rm a} \over \dot M_{\rm w}}\right)^{1-2m}

1095: 	       \Psi^{1/7-2m}.

1096: %

1097: \end{eqnarray}

1098: This is the predicted equilibrium spin rate when the truncation radius

1099: is significantly smaller than the corotation radius (i.e., in magnetic

1100: connection state 1).

1101: 

1102: The second case to consider is where $R_{\rm t} \approx R_{\rm co}$,

1103: which is the requirement of all disk-locking models

1104: \citep[][MP05b]{shuea94, wang95}.  Using equation (\ref{eqn_rco}) and

1105: setting $R_{\rm t} = R_{\rm co}$ in equation (\ref{eqn_fsw}), one

1106: finds

1107: \begin{eqnarray}

1108: \label{eqn_fsw2}

1109: %

1110: f_{\rm sw2} = K^{-3/2}

1111:                \left({\dot M_{\rm a} \over \dot M_{\rm w}}\right)^{(3-6m)/4}

1112: 	       \Psi^{-3m/2}.

1113: %

1114: \end{eqnarray}

1115: This is the predicted equilibrium spin rate when the disk truncation

1116: occurs very close to $R_{\rm co}$.

1117: 

1118: Which case is more appropriate?  The first case is expected to occurs

1119: for relatively small values of $\Psi$ and low spin rates $f$ (MP05b),

1120: and the opposite is true for the second case.  For the parameter space

1121: we have considered thus far in this work, we found in section

1122: \ref{sec_state} that most (though not all) of the cases are expected

1123: to have $R_{\rm t} / R_{\rm co}$ near unity.  Thus, while it is not a

1124: formal requirement of APSW, it may often be the case that $R_{\rm t}

1125: \approx R_{\rm co}$ for systems in spin equilibrium, and we will focus

1126: on this second case for the remainder of this work.

1127: 

1128: 

1129: 

1130: \begin{figure}

1131: \epsscale{1.15}

1132: \plotone{f2.eps}

1133: %\plotone{figs/feqs1.eps}

1134: 

1135: \caption{The equilibrium spin rate predicted by a balance between the

1136: spin down from a stellar wind and the spin up from accretion, versus

1137: the dimensionless magnetization parameter $\Psi \equiv B_*^2 R_*^2

1138: (\dot M_{\rm a} v_{\rm esc})^{-1}$.  The solid lines show equation

1139: (\ref{eqn_fsw2}) for $K = 2.11$, $m = 0.223$ and several different

1140: possible values of $\dot M_{\rm w} / \dot M_{\rm a}$, as indicated.

1141: The squares show data from table \ref{tab_equilib0.1}, which indicates

1142: the range of parameters considered in our simulations and used to

1143: derive equation (\ref{eqn_fsw2}).  The shaded region corresponds

1144: approximately to where accretion power is not sufficient to drive the

1145: stellar wind.

1146: \label{fig_feqs1}}

1147: 

1148: \end{figure}

1149: 

1150: 

1151: 

1152: Figure \ref{fig_feqs1} shows the predicted spin rate of this case

1153: (eq.\ \ref{eqn_fsw2}) versus $\Psi$, for many different values of the

1154: ratio $\dot M_{\rm w} / \dot M_{\rm a}$.  The results of section

1155: \ref{sec_power} indicate that the accretion power is only capable of

1156: powering a wind with $\dot M_{\rm w} / \dot M_{\rm a} \la 0.6$, and

1157: this is a hard upper limit.  This ``forbidden'' region of the

1158: $f$-$\Psi$ space is indicated in figure \ref{fig_feqs1}.

1159: 

1160: The plot also shows the simulation results (squares).  The value of

1161: $\Psi$ for each case is determined mostly by input parameters but also

1162: by $\dot M_{\rm a}$.  The latter was set (in section

1163: \ref{sec_balancing}) by the condition that the equilibrium spin rate

1164: was equal to the value of $f$ used as the simulation input parameter.

1165: The squares indicate the range over which equation (\ref{eqn_fsw2}) is

1166: shown to be valid by the simulations.  Also, the plot shows the

1167: specific case of $K = 2.11$ and $m = 0.223$.

1168: 

1169: The results shown in figure \ref{fig_feqs1} and equation

1170: (\ref{eqn_fsw2}) provide the basis for predictions of the APSW model

1171: that can be observationally tested and constrained, and that can be

1172: compared to other models.  Different theories predict different power

1173: laws for $f$ vs.\ $\Psi$.  Specifically, equation (\ref{eqn_fsw2})

1174: predicts a power law index of $\approx -0.33$, the conditions

1175: appropriate for equation (\ref{eqn_fsw1}) predict $\approx -0.30$, and

1176: an index of $-3/7$ is predicted by the disk locking models

1177: \citep[e.g.,][]{konigl91, shuea94}.

1178: 

1179: 

1180: %they have quite similar slopes. On the one hand, this is unfortunate,

1181: %as it means that APSW does not predict drastically different trends of

1182: %the spin rate with stellar parameters, $\dot M_{\rm a}$, or magnetic

1183: %field strength.  So this makes it more difficult to observationally

1184: %discriminate between the models.  since APSW has slightly shallower

1185: %slopes, it works better for the cases with weaker fields or higher

1186: %accretion rates, relative to the disk locking model or X-wind.

1187: 

1188: %do we need to say somewhere what is typical range of $\Psi$ for T

1189: %Tauri stars?  I think plot range says it, and so does simulation

1190: %range, but maybe we could give some ``for example'' parameters that

1191: %give $\Psi$ of 1 or 1e5. (reality probably includes an even wider

1192: %range!) (fiducial parameters but 1e-8 msun/yr accretion rate gives psi

1193: %= 40.0 or for same but field strength of 2 kG gives psi = 4000.)

1194: 

1195: 

1196: 

1197: %\section{Summary and Conclusions} \label{sec_discussion}

1198: \section{Summary and Discussion} \label{sec_discussion}

1199: 

1200: 

1201: 

1202: In this work, we have further developed the accretion-powered stellar

1203: wind model proposed in Paper~I, where the stellar wind magnetic lever

1204: arm length, $r_{\rm A}$, was taken as a parameter to determine the

1205: stellar wind torque.  We employed the simulation results of Paper~II

1206: (see \S \ref{sec_simulations}) to obtain stellar wind torques for

1207: several cases representative of T Tauri systems.  We examined the

1208: total torque on the star arising from the stellar wind plus the

1209: magnetic interaction between the star and its accretion disk.  Our

1210: results can be summarized as follows.

1211: \begin{enumerate}

1212: 

1213: \item{We found that the spin-down torque from a stellar wind can be

1214:   orders of magnitude stronger than the spin-down portion of the

1215:   star-disk interaction torque, for slowly rotating stars with mass

1216:   loss rates substantially larger than the solar wind outflow rate

1217:   (see \S \ref{sec_tdsd}).  This confirms the assumption of Paper~I

1218:   that the condition for net zero torque on the star (spin

1219:   equilibrium) is simply determined by a balance between the stellar

1220:   wind torque and the accretion torque.}

1221: 

1222: \item{Using the computed stellar wind torques for several cases, we

1223:   looked at the conditions for spin equilibrium (\S

1224:   \ref{sec_balancing}).  We found that a rotation rate of 10\% of

1225:   breakup speed typically requires $\dot M_{\rm w} / \dot M_{\rm a}$

1226:   equal a few tens of percent, confirming the original suggestion by

1227:   \citet{hartmannstauffer89} that stellar winds may be capable of

1228:   removing accreted angular momentum.}

1229: 

1230: \item{For most cases in spin equilibrium, the disk truncation radius

1231:   was very close to the corotation radius, though this is not a

1232:   general requirement of the APSW model (\S \ref{sec_state}).}

1233: 

1234: \item{Accretion power is generally sufficient to power a stellar wind

1235:   that is capable of solving the angular momentum problem (\S

1236:   \ref{sec_power}), as suggested in Paper~I.  The energy requirements

1237:   for most of the cases considered here is relatively large, and more

1238:   work is needed to further constrain the energy coupling.}

1239: 

1240: \item{Under the assumption that the stellar wind is accretion powered,

1241:   the cases we examined suggested a hard upper limit of $\dot M_{\rm

1242:     w} / \dot M_{\rm a} \la 0.6$.}

1243: 

1244: \item{Finally, in section \ref{sec_semianalytic} we used the results

1245:   from Paper~II to derive a semi-analytic formulation for the

1246:   equilibrium spin rate predicted by the APSW model.  We found the

1247:   that spin rate, expressed as a fraction of breakup speed, generally

1248:   depends only on the two dimensionless parameters $\dot M_{\rm w} /

1249:   \dot M_{\rm a}$ and $\Psi \equiv B_*^2 R_*^2 (\dot M_{\rm a} v_{\rm

1250:     esc})^{-1}$.}

1251: 

1252: \end{enumerate}

1253: 

1254: 

1255: 

1256: The APSW model incorporates several previous ideas.  As in all other

1257: models that emphasize the role of stellar magnetic fields, the

1258: interaction of the magnetized star with the disk leads to the

1259: truncation of the disk and accretion of material along field lines

1260: onto the star.  The APSW model adopts the finding of the Ghosh \&

1261: Lamb-type models \citep[e.g.,][]{ghoshlamb78, konigl91,

1262:   armitageclarke96}, which is also supported by numerical simulations

1263: \citep[][]{romanovaea02, long3ea05}, that the angular momentum of

1264: accreting material is transferred to the star.  However, in contrast

1265: to the Ghosh \& Lamb-type models, we found that for slow rotators, any

1266: spin-down torque arising from the star-disk interaction is negligible

1267: (item 1 above). Instead, the stellar spin-up torque from accretion is

1268: counteracted by an accretion-driven stellar wind, which carries a

1269: comparable amount of angular momentum out of the system.

1270: 

1271: Compared to all existing models, APSW is distinct in that it

1272: conceptually links the driving of the stellar wind to the energy

1273: released by the accretion process (via $\epsilon$).  In other ways,

1274: the general picture of APSW is similar to other angular momentum

1275: models that utilize winds.  In particular, the X-wind \citep{shuea94,

1276:   ostrikershu95}, the Reconnection X-wind \citep{ferreira3ea00}, and

1277: other works considering stellar winds \citep{hartmannstauffer89,

1278:   paatzcamenzind96} all find that, in order to carry away significant

1279: angular momentum, the mass outflow rate needs to be of the order of

1280: 10\% of the accretion rate (item 2 above).  Except for the X-wind, in

1281: all of the above scenarios, the outflow is magnetically connected to

1282: the star, and thus extracts angular momentum directly from the star.

1283: By contrast, the X-wind outflow is magnetically connected to the disk.

1284: Furthermore, the X-wind is unique in that it assumes that the

1285: accretion of material does not deposit angular momentum onto the star.

1286: 

1287: As mentioned in item 3 above, for most of the specific cases of our

1288: simulated winds, we found that in the spin equilibrium state, the

1289: truncation radius was very close to the corotation radius.  This is

1290: similar to the prediction of the disk locking models, which includes

1291: both the Ghosh \& Lamb-type models and the X-wind.  However, in

1292: contrast with the disk locking models, it is not a requirement of APSW

1293: that $R_{\rm t}$ be close to $R_{\rm co}$.  Measurements of the

1294: location of the inner edge of the gas disk \citep{najita3ea03, carr07}

1295: suggest that $R_{\rm t}$/$R_{\rm co}$ is typically $\sim 70$\%.  We

1296: leave a more detailed comparison between models for future work.

1297: 

1298: 

1299: 

1300: %The APSW model is a synthesis of many previous ideas.  As in the

1301: %Ghosh \& Lamb-type models \citep[e.g.,][]{ghoshlamb78, konigl91,

1302: %armitageclarke96}, the stellar magnetic field truncates the disk, and

1303: %the accretion of material from the disk adds angular momentum to the

1304: %star.  This contrasts with the X-wind model \citep{shuea94}, which

1305: %proposes that disk material arrives at the star entirely stripped of

1306: %its angular momentum.  Like the X-wind, and unlike Ghosh \& Lamb, the

1307: %APSW model requires a wind to remove angular momentum from the

1308: %system.  But, in contrast with the X-wind, where the outflow is

1309: %magnetically connected to the disk, the APSW is magnetically

1310: %connected to the star. We leave a more detailed comparison between

1311: %APSW and other models for future work.

1312: 

1313: There is observational evidence that the outflow rates from accreting

1314: young stellar systems are of the order of 10\% of the accretion rates

1315: \citep[e.g.,][]{hartigan3ea95, calvet97} and are therefore accretion

1316: powered \citep{cabritea90}.  However, it appears that a large fraction

1317: of this flow (which is usually probed by forbidden emission coming

1318: from large spatial scales) originates in the disk, rather than the

1319: star \citep{ferreira3ea06}.  It is not yet clear how much of the total

1320: observed flow may originate in a stellar wind.  There is some evidence

1321: specifically for stellar winds from CTTSs \citep{beristain3ea01,

1322: edwardsea03, dupreeea05, edwardsea06, kwan3ea07}, as distinct from

1323: disk winds, and that these are accretion powered

1324: \citep[e.g.,][]{edwardsea03, edwardsea06}, but the mass outflow rates

1325: are not yet well constrained \citep{dupreeea05}.  Additional work

1326: constraining the value of $\dot M_{\rm w} / \dot M_{\rm a}$, the

1327: stellar wind driving mechanism, and the stellar magnetic field

1328: strength and geometry will help to provide stringent and quantitative

1329: tests for the APSW model.

1330: 

1331: The predictions of the spin equilibrium state can also be checked

1332: observationally \citep{johnskrullgafford02}.  This will likely require

1333: large samples of stars, due to large uncertainties in measured

1334: parameters, and since intrinsic variability in real systems

1335: \citep[e.g.,]{hartmann97} may only allow a spin equilibrium state to

1336: be achieved in a time-averaged sense.  The Ghosh \& Lamb, X-wind, and

1337: APSW models all predict an equilibrium spin rate that depends on

1338: $\Psi$, but of these three, only the APSW model contains an additional

1339: dependence on the stellar wind mass outflow rate.  For the power law

1340: fits to the simulations of Paper~II, the APSW spin equilibrium

1341: predicts a slightly different power-law of spin vs.\ $\Psi$ than the

1342: other models (\S \ref{sec_semianalytic})---though the exact dependence

1343: of the stellar wind torque has not been determined for all parameters.

1344: 

1345: In this series of papers, we have focused on the global problem of

1346: calculating the magnitude of stellar wind torques and comparing them

1347: with other torques acting on accreting stars.  In order to refine the

1348: APSW model further and make the predictions more precise, more work is

1349: required.  In particular, it is not yet clear how the presence of an

1350: accretion disk will influence the stellar wind torque, and conversely,

1351: how a stellar wind may influence the accretion process.  Also,

1352: although it is clear that there is enough accretion energy to power

1353: the stellar wind, it is still not known what actually drives the

1354: stellar wind and how the accretion power may transfer to it.  We

1355: suspect that a strong flux of hydromagnetic waves can be excited near

1356: the base of the accretion shock and can tap the energy released there,

1357: which may provide an efficient driver for the APSW.  We defer a

1358: rigourous investigation of the APSW driving mechanism to future work.

1359: 

1360: 

1361: 

1362: \acknowledgements

1363: 

1364: 

1365: 

1366: We wish to thank many people for discussions regarding this work,

1367: including: Gibor Basri, Sylvie Cabrit, Andrea Dupree, Suzan Edwards,

1368: Will Fischer, Shu-ichiro Inutsuka, Chris Johns-Krull, Marina Romanova,

1369: Frank Shu, Keivan Stassun, Jeff Valenti, and Sydney Wolff.  We also

1370: thank the referee, Jonathan Ferreira, for his useful suggestions for

1371: improving the paper and KITP for hosting us while finishing the

1372: manuscript.  This research was supported in part by the National

1373: Science Foundation under Grant No. PHY05-51164.  SM is supported by

1374: the University of Virginia through a Levinson/VITA Fellowship

1375: partially funded by The Frank Levinson Family Foundation through the

1376: Peninsula Community Foundation.  REP is supported by a grant from

1377: NSERC.

1378: 

1379: 

1380: \appendix

1381: \section{Determination of the Disk Truncation Radius}

1382: 

1383: We follow MP05b to calculate the location of the disk truncation

1384: radius, $R_{\rm t}$, and the reader will find details in that paper.

1385: For convenience, we list the relevant equations here.  As in section

1386: \ref{sec_tdsd}, we adopt $\gamma_{\rm c} = 1$.

1387: 

1388: To determine $R_{\rm t}$, we first find the magnetic connectivity

1389: state using the criterion

1390: \begin{eqnarray}

1391: \label{eqn_trans}

1392: f < (1 - \beta) (2^{3/2} \Psi)^{-3/7},

1393: \end{eqnarray}

1394: where $\Psi$ is defined by equation (\ref{eqn_psi}).  If condition

1395: (\ref{eqn_trans}) is satisfied, the system is in ``state 1."

1396: Otherwise, it is in ``state 2."  In state 1, we determine the

1397: truncation radius using

1398: \begin{eqnarray}

1399: \label{eqn_rt1}

1400: R_{\rm t} = (2^{3/2} \Psi)^{2/7} R_*.

1401: % = (2)^{2/7} (G M_*)^{-{1/7}} (\dot M_{\rm a})^{-{2/7}} B_*^{4/7} R_*^{12/7}.

1402: \end{eqnarray}

1403: In state 2, we determine the truncation radius by solving

1404: \begin{eqnarray}

1405: \label{eqn_rt2}

1406: \left({R_{\rm t} \over R_{\rm co}}\right)^{-7/2}

1407:   \left[{1 - \left({R_{\rm t} \over R_{\rm co}}\right)^{3/2}}\right] =

1408:   {\beta \over 2^{3/2} \Psi} f^{-{7/3}}.

1409: \end{eqnarray}

1410: 

1411: 

1412: 

1413: % ---------- BIBLIOGRAPHY -------------

1414: 

1415: %\bibliography{../../references}

1416: %\bibliographystyle{../../bibstyles/apj}

1417: 

1418: \begin{thebibliography}{56}

1419: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi

1420: 

1421: \bibitem[{{Armitage} \& {Clarke}(1996)}]{armitageclarke96}

1422: {Armitage}, P.~J. \& {Clarke}, C.~J. 1996, \mnras, 280, 458

1423: 

1424: \bibitem[{{Beristain} {et~al.}(2001){Beristain}, {Edwards}, \&

1425:   {Kwan}}]{beristain3ea01}

1426: {Beristain}, G., {Edwards}, S., \& {Kwan}, J. 2001, \apj, 551, 1037

1427: 

1428: \bibitem[{{Bertout} {et~al.}(1988){Bertout}, {Basri}, \&

1429:   {Bouvier}}]{bertout3ea88}

1430: {Bertout}, C., {Basri}, G., \& {Bouvier}, J. 1988, \apj, 330, 350

1431: 

1432: \bibitem[{{Bouvier} {et~al.}(1997){Bouvier}, {Forestini}, \&

1433:   {Allain}}]{bouvier3ea97}

1434: {Bouvier}, J., {Forestini}, M., \& {Allain}, S. 1997, \aap, 326, 1023

1435: 

1436: \bibitem[{{Cabrit} {et~al.}(1990){Cabrit}, {Edwards}, {Strom}, \&

1437:   {Strom}}]{cabritea90}

1438: {Cabrit}, S., {Edwards}, S., {Strom}, S.~E., \& {Strom}, K.~M. 1990, \apj, 354,

1439:   687

1440: 

1441: \bibitem[{{Calvet}(1997)}]{calvet97}

1442: {Calvet}, N. 1997, in IAU Symposium, Vol. 182, Herbig-Haro Flows and the Birth

1443:   of Stars, ed. B.~{Reipurth} \& C.~{Bertout}, 417--432

1444: 

1445: \bibitem[{{Calvet} \& {Gullbring}(1998)}]{calvetgullbring98}

1446: {Calvet}, N. \& {Gullbring}, E. 1998, \apj, 509, 802

1447: 

1448: \bibitem[{{Camenzind}(1990)}]{camenzind90}

1449: {Camenzind}, M. 1990, in Reviews in Modern Astronomy, ed. G.~{Klare}, 234--265

1450: 

1451: \bibitem[{{Cameron} \& {Campbell}(1993)}]{cameroncampbell93}

1452: {Cameron}, A.~C. \& {Campbell}, C.~G. 1993, \aap, 274, 309

1453: 

1454: \bibitem[{{Carr}(2007)}]{carr07}

1455: {Carr}, J.~S. 2007, to appear in proceedings of IAU Symposium No.\ 243,

1456:   Star-Disk Interaction in Young Stars

1457: 

1458: \bibitem[{{Davidson} \& {Ostriker}(1973)}]{davidsonostriker73}

1459: {Davidson}, K. \& {Ostriker}, J.~P. 1973, \apj, 179, 585

1460: 

1461: \bibitem[{{Decampli}(1981)}]{decampli81}

1462: {Decampli}, W.~M. 1981, \apj, 244, 124

1463: 

1464: \bibitem[{{Dupree} {et~al.}(2005){Dupree}, {Brickhouse}, {Smith}, \&

1465:   {Strader}}]{dupreeea05}

1466: {Dupree}, A.~K., {Brickhouse}, N.~S., {Smith}, G.~H., \& {Strader}, J. 2005,

1467:   \apjl, 625, L131

1468: 

1469: \bibitem[{{Edwards} {et~al.}(2006){Edwards}, {Fischer}, {Hillenbrand}, \&

1470:   {Kwan}}]{edwardsea06}

1471: {Edwards}, S., {Fischer}, W., {Hillenbrand}, L., \& {Kwan}, J. 2006, \apj, 646,

1472:   319

1473: 

1474: \bibitem[{{Edwards} {et~al.}(2003){Edwards}, {Fischer}, {Kwan}, {Hillenbrand},

1475:   \& {Dupree}}]{edwardsea03}

1476: {Edwards}, S., {Fischer}, W., {Kwan}, J., {Hillenbrand}, L., \& {Dupree}, A.~K.

1477:   2003, \apjl, 599, L41

1478: 

1479: \bibitem[{{Ferreira}(1997)}]{ferreira97}

1480: {Ferreira}, J. 1997, \aap, 319, 340

1481: 

1482: \bibitem[{{Ferreira} {et~al.}(2006){Ferreira}, {Dougados}, \&

1483:   {Cabrit}}]{ferreira3ea06}

1484: {Ferreira}, J., {Dougados}, C., \& {Cabrit}, S. 2006, \aap, 453, 785

1485: 

1486: \bibitem[{{Ferreira} {et~al.}(2000){Ferreira}, {Pelletier}, \&

1487:   {Appl}}]{ferreira3ea00}

1488: {Ferreira}, J., {Pelletier}, G., \& {Appl}, S. 2000, \mnras, 312, 387

1489: 

1490: \bibitem[{{Ghosh} \& {Lamb}(1978)}]{ghoshlamb78}

1491: {Ghosh}, P. \& {Lamb}, F.~K. 1978, \apjl, 223, L83

1492: 

1493: \bibitem[{{Hartigan} {et~al.}(1995){Hartigan}, {Edwards}, \&

1494:   {Ghandour}}]{hartigan3ea95}

1495: {Hartigan}, P., {Edwards}, S., \& {Ghandour}, L. 1995, \apj, 452, 736

1496: 

1497: \bibitem[{{Hartmann}(1997)}]{hartmann97}

1498: {Hartmann}, L. 1997, in IAU Symposium, Vol. 182, Herbig-Haro Flows and the

1499:   Birth of Stars, ed. B.~{Reipurth} \& C.~{Bertout}, 391--405

1500: 

1501: \bibitem[{{Hartmann} \& {Stauffer}(1989)}]{hartmannstauffer89}

1502: {Hartmann}, L. \& {Stauffer}, J.~R. 1989, \aj, 97, 873

1503: 

1504: \bibitem[{{Herbst} {et~al.}(2007){Herbst}, {Eisl{\"o}ffel}, {Mundt}, \&

1505:   {Scholz}}]{herbstea07}

1506: {Herbst}, W., {Eisl{\"o}ffel}, J., {Mundt}, R., \& {Scholz}, A. 2007, in

1507:   Protostars and Planets V, ed. B.~{Reipurth}, D.~{Jewitt}, \& K.~{Keil},

1508:   297--311

1509: 

1510: \bibitem[{{Jayawardhana} {et~al.}(2006){Jayawardhana}, {Coffey}, {Scholz},

1511:   {Brandeker}, \& {van Kerkwijk}}]{jayawardhanaea06}

1512: {Jayawardhana}, R., {Coffey}, J., {Scholz}, A., {Brandeker}, A., \& {van

1513:   Kerkwijk}, M.~H. 2006, \apj, 648, 1206

1514: 

1515: \bibitem[{{Johns-Krull} \& {Gafford}(2002)}]{johnskrullgafford02}

1516: {Johns-Krull}, C.~M. \& {Gafford}, A.~D. 2002, \apj, 573, 685

1517: 

1518: \bibitem[{{Joy}(1945)}]{joy45}

1519: {Joy}, A.~H. 1945, \apj, 102, 168

1520: 

1521: \bibitem[{{Keppens} \& {Goedbloed}(2000)}]{keppensgoedbloed00}

1522: {Keppens}, R. \& {Goedbloed}, J.~P. 2000, \apj, 530, 1036

1523: 

1524: \bibitem[{{K\"onigl}(1991)}]{konigl91}

1525: {K\"onigl}, A. 1991, \apjl, 370, L39

1526: 

1527: \bibitem[{{Kwan} {et~al.}(2007){Kwan}, {Edwards}, \& {Fischer}}]{kwan3ea07}

1528: {Kwan}, J., {Edwards}, S., \& {Fischer}, W. 2007, \apj, 657, 897

1529: 

1530: \bibitem[{{Lamb} {et~al.}(1973){Lamb}, {Pethick}, \& {Pines}}]{lamb3ea73}

1531: {Lamb}, F.~K., {Pethick}, C.~J., \& {Pines}, D. 1973, \apj, 184, 271

1532: 

1533: \bibitem[{{Long} {et~al.}(2005){Long}, {Romanova}, \& {Lovelace}}]{long3ea05}

1534: {Long}, M., {Romanova}, M.~M., \& {Lovelace}, R.~V.~E. 2005, \apj, 634, 1214

1535: 

1536: \bibitem[{{Lynden-Bell} \& {Pringle}(1974)}]{lyndenbellpringle74}

1537: {Lynden-Bell}, D. \& {Pringle}, J.~E. 1974, \mnras, 168, 603

1538: 

1539: \bibitem[{{Lyo} \& {Lawson}(2005)}]{lyolawson05}

1540: {Lyo}, A.-R. \& {Lawson}, W.~A. 2005, Journal of Korean Astronomical Society,

1541:   38, 241

1542: 

1543: \bibitem[{{Matt} \& {Pudritz}(2005{\natexlab{a}})}]{mattpudritz05l}

1544: {Matt}, S. \& {Pudritz}, R.~E. 2005{\natexlab{a}}, \apjl, 632, L135 (Paper~I)

1545: 

1546: \bibitem[{{Matt} \& {Pudritz}(2005{\natexlab{b}})}]{mattpudritz05}

1547: ---. 2005{\natexlab{b}}, \mnras, 356, 167 (MP05b)

1548: 

1549: \bibitem[{{Matt} \& {Pudritz}(2007)}]{mattpudritz07coolstars}

1550: ---. 2007, to appear in proceedings of the 14th Cambridge Workshop on Cool

1551:   Stars, Stellar Systems, and the Sun, astro-ph/0701648

1552: 

1553: \bibitem[{{Matt} \& {Pudritz}(2008)}]{mattpudritz08II}

1554: ---. 2008, \apj, in press (arXiv:0801.0436) (Paper~II)

1555: 

1556: \bibitem[{{Muzerolle} {et~al.}(2001){Muzerolle}, {Calvet}, \&

1557:   {Hartmann}}]{muzerolle3ea01}

1558: {Muzerolle}, J., {Calvet}, N., \& {Hartmann}, L. 2001, \apj, 550, 944

1559: 

1560: \bibitem[{{Najita} {et~al.}(2003){Najita}, {Carr}, \& {Mathieu}}]{najita3ea03}

1561: {Najita}, J., {Carr}, J.~S., \& {Mathieu}, R.~D. 2003, \apj, 589, 931

1562: 

1563: \bibitem[{{Ostriker} \& {Shu}(1995)}]{ostrikershu95}

1564: {Ostriker}, E.~C. \& {Shu}, F.~H. 1995, \apj, 447, 813

1565: 

1566: \bibitem[{{Paatz} \& {Camenzind}(1996)}]{paatzcamenzind96}

1567: {Paatz}, G. \& {Camenzind}, M. 1996, \aap, 308, 77

1568: 

1569: \bibitem[{{Pelletier} \& {Pudritz}(1992)}]{pelletierpudritz92}

1570: {Pelletier}, G. \& {Pudritz}, R.~E. 1992, \apj, 394, 117

1571: 

1572: \bibitem[{{Rebull} {et~al.}(2004){Rebull}, {Wolff}, \& {Strom}}]{rebull3ea04}

1573: {Rebull}, L.~M., {Wolff}, S.~C., \& {Strom}, S.~E. 2004, \aj, 127, 1029

1574: 

1575: \bibitem[{{Rebull} {et~al.}(2002){Rebull}, {Wolff}, {Strom}, \&

1576:   {Makidon}}]{rebullea02}

1577: {Rebull}, L.~M., {Wolff}, S.~C., {Strom}, S.~E., \& {Makidon}, R.~B. 2002, \aj,

1578:   124, 546

1579: 

1580: \bibitem[{{Romanova} {et~al.}(2002){Romanova}, {Ustyugova}, {Koldoba}, \&

1581:   {Lovelace}}]{romanovaea02}

1582: {Romanova}, M.~M., {Ustyugova}, G.~V., {Koldoba}, A.~V., \& {Lovelace},

1583:   R.~V.~E. 2002, \apj, 578, 420

1584: 

1585: \bibitem[{{Safier}(1998)}]{safier98}

1586: {Safier}, P.~N. 1998, \apj, 494, 336

1587: 

1588: \bibitem[{{Shakura} \& {Sunyaev}(1973)}]{shakurasunyaev73}

1589: {Shakura}, N.~I. \& {Sunyaev}, R.~A. 1973, \aap, 24, 337

1590: 

1591: \bibitem[{{Shu} {et~al.}(1994){Shu}, {Najita}, {Ostriker}, {Wilkin}, {Ruden},

1592:   \& {Lizano}}]{shuea94}

1593: {Shu}, F., {Najita}, J., {Ostriker}, E., {Wilkin}, F., {Ruden}, S., \&

1594:   {Lizano}, S. 1994, \apj, 429, 781

1595: 

1596: \bibitem[{{Tout} \& {Pringle}(1992)}]{toutpringle92}

1597: {Tout}, C.~A. \& {Pringle}, J.~E. 1992, \mnras, 256, 269

1598: 

1599: \bibitem[{{Ustyugova} {et~al.}(1999){Ustyugova}, {Koldoba}, {Romanova},

1600:   {Chechetkin}, \& {Lovelace}}]{ustyugovaea99}

1601: {Ustyugova}, G.~V., {Koldoba}, A.~V., {Romanova}, M.~M., {Chechetkin}, V.~M.,

1602:   \& {Lovelace}, R.~V.~E. 1999, \apj, 516, 221

1603: 

1604: \bibitem[{{Uzdensky}(2004)}]{uzdensky04}

1605: {Uzdensky}, D.~A. 2004, \apss, 292, 573

1606: 

1607: \bibitem[{{Uzdensky} {et~al.}(2002){Uzdensky}, {K{\" o}nigl}, \&

1608:   {Litwin}}]{uzdensky3ea02}

1609: {Uzdensky}, D.~A., {K{\" o}nigl}, A., \& {Litwin}, C. 2002, \apj, 565, 1191

1610: 

1611: \bibitem[{{Vogel} \& {Kuhi}(1981)}]{vogelkuhi81}

1612: {Vogel}, S.~N. \& {Kuhi}, L.~V. 1981, \apj, 245, 960

1613: 

1614: \bibitem[{{Wang}(1995)}]{wang95}

1615: {Wang}, Y.-M. 1995, \apjl, 449, L153

1616: 

1617: \bibitem[{{Washimi} \& {Shibata}(1993)}]{washimishibata93}

1618: {Washimi}, H. \& {Shibata}, S. 1993, \mnras, 262, 936

1619: 

1620: \bibitem[{{Yi}(1995)}]{yi95}

1621: {Yi}, I. 1995, \apj, 442, 768

1622: 

1623: \end{thebibliography}

1624: 

1625: 

1626: 

1627: 

1628: %------------- TABLES -------------

1629: 

1630: %\newpage

1631: 

1632: 

1633: 

1634: 

1635: %------------- FIGURES -------------

1636: 

1637: 

1638: 

1639: 

1640: 

1641: \label{lastpage}

1642: \end{document}

1643: 

1644: