1: %

2: %% Beginning of file 'main.tex'

3: %%

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

5: 

6: %\usepackage{bm}

7: 

8: \newcommand{\vdag}{(v)^\dagger}

9: %\newcommand{\myemail}{skywalker@galaxy.far.far.away}

10: 

11: %\slugcomment{Not to appear in Nonlearned J., 45.}

12: 

13: 

14: \shorttitle{Equatorially-Asymmetric Magnetized Supernovae}

15: \shortauthors{Sawai, Kotake, \& Yamada}

16: 

17: \begin{document}

18: 

19: \title{Numerical Simulations of Equatorially-Asymmetric Magnetized

20: Supernovae: Formation of Magnetars and Their Kicks}

21: 

22: \author{Hidetomo Sawai\altaffilmark{1}, Kei Kotake\altaffilmark{2,3} and

23: Shoichi Yamada\altaffilmark{1,4}}

24: \altaffiltext{1}{\textit{Science \& Engineering, Waseda University,

25: 3-4-1 Okubo,

26: Shinjuku, Tokyo 169-8555, Japan}}

27: \email{hsawai@heap.phys.waseda.ac.jp}

28: 

29: \altaffiltext{2}{\textit{Division of Theoretical Astronomy, National

30:     Astronomical Observatory of Japan, 2-21-1 Osawa, Mitaka, Tokyo 181-8588,

31:     Japan}}
 

32: 

33: \altaffiltext{3}{\textit{Max-Planck-Institute f\"{u}r Astrophysik,

34: Karl-Schwarzshild-Str. 1, D-85741, Garching, Germany}}

35: 

36: \altaffiltext{4}{\textit{Advanced Research Institute for Science \&

37: Engineering, Waseda University, 3-4-1 Okubo,

38: Shinjuku, Tokyo 169-8555, Japan}}

39: 

40: \begin{abstract}

41: A series of numerical simulations on magnetorotational core-collapse

42:  supernovae are carried out. Dipole-like configurations which are offset

43:  northward are assumed

44:  for the initially strong magnetic fields together with rapid

45:  differential rotations. Aims of our study are to

46:  investigate effects of the offset magnetic field on magnetar kicks

47:  and on supernova dynamics. Note that we study a regime where

48:  the proto-neutron star formed after collapse has a large magnetic field

49:  strength approaching that of a ``magnetar'', a highly

50:  magnetized slowly rotating neutron star. As a result, equatorially-asymmetric

51:  explosions occur with a formation of the bipolar jets. We find that the

52:  jets are fast and light in the north and a slow and heavy in the south

53:  for in rapid cases while they are fast and heavy in the north and slow

54:  and light in the south for slow rotation case.

55:  Resultant magnetar's kick velocities are $\sim 300-1000$~km~s$^{-1}$.  

56:  We find that the acceleration is mainly due to the magnetic pressure

57:  while the somewhat weaker magnetic tension works toward the opposite

58:  direction, which is due to stronger magnetic field in the northern

59:  hemisphere.

60:  Noted that observations of magnetar's proper motions are very

61:  scarce, our results supply a prediction for future observations. Namely,

62:  magnetars possibly have large kick velocities, several hundred~km~s$^{-1}$, as

63:  ordinary neutron stars do, and in an extreme case they could have those up

64:  to 1000~km~s$^{-1}$. In each model, the formed proto-magnetar is slow

65:  rotator with rotational period of more than 10~ms. It is also found that, in

66:  rapid rotation 

67:  models, the final configuration of the magnetic field in the

68:  proto-magnetar is a collimated dipole-like

69:  field pinched by the torus of toroidal field lines whereas the

70:  proto-magnetar produced in the slow 

71:  rotation model is totally poloidal-field dominant. 

72: \end{abstract}

73: 

74: 

75: \keywords{supernovae: general --- supernovae: general ---

76: pulsars: general --- stars: magnetic fields --- MHD --- methods: numerical  }

77: 

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

79:  Soft-gamma repeaters (SGRs) and anomalous

80:  X-ray pulsars (AXPs) are candidates of so-called ``magnetars'', highly

81:  magnetized slowly rotating neutron stars. This is based on

82:  the ``magnetar model'', i.e. their burst-like activities, persistent

83:  X-ray emissions, and large spin-down rates will originate from high 

84:  magnetic fields, $B \gtrsim 10^{14}-10^{15}$~G

85:  \citep{dun92,pac92,tho95,tho96}.

86:  It is notable that such a large magnetic field also may power supernova

87:  explosions (magnetorotational supernova). Thus, study of magnetorotational 

88:  supernovae may be important in the context of magnetar formations.  

89:    

90:  The first paper concerning magnetorotational supernovae written by

91:  \citet{leb70} appeared about 30 years before the idea of ``magnetar

92:  model'' has

93:  arisen. They found that strong magnetic field combined with rotation

94:  produced an asymmetric supernova explosion. 

95:  Only a few studies followed after this \citep{bis75, mul79, ohn83,

96:  sym84}, reflecting a feeling that

97:  magnetic fields large enough to drive supernovae were thought to be

98:  unrealistic. Now that such high magnetic fields are observationally

99:  supported, the magnetorotational supernova is attracting more

100:  attention than before. In the last several years some papers have been

101:  published on the magnetorotational supernovae \citep[e.g.][]{whe02,

102:  aki03, yam04, kot04, tak04, ard05, kot05, saw05, obe06, 

103:  nis06, moi06, shi06}. Although, with these studies we have developed

104:  our understanding of magnetorotational supernovae, many issues still

105:  remain to be investigated. For example, neutron star's kicks in the

106:  context of magnetorotational supernovae (magnetar's kicks) have not been studied fully numerically. (see, however, \citet{kot05}: they computed the

107:  parity-violating effects in strongly magnetized supernova cores and

108:  expected that neutron star's kicks would result.) In fact, all numerical

109:  simulations of magnetorotational core-collapse so far assumed equatorial

110:  symmetry and computed only a

111:  quarter of a meridional plane of the star.

112: 

113:  Observed neutron stars are known to have generally larger proper

114:  velocities than that of their progenitor stars \citep{gun79}. Their

115:  three-dimensional velocities are typically several

116:  hundred~km~s$^{-1}$ and some of them have a speed higher than

117:  1000~km~s$^{-1}$ \citep[e.g.][]{cor98,hob05}. The origin of the proper

118:  velocities is still controversial. Although many researchers

119:  agree that an asymmetric supernova explosion is a more promising origin

120:  rather than a binary-system

121:  disruption by a supernova explosion, several different mechanisms have been

122:  proposed, which can be categorized mainly into three

123:  classes \citep{lai03}. 

124: 

125:  The first mechanism is a hydrodynamically-driven

126:  kick. \citet{sch06} numerically studied hydrodynamic instabilities

127:  in neutrino-driven supernova explosions and showed that neutron stars

128:  were accelerated up to about 500 - 1000~km~s$^{-1}$ due to the dominance of $l

129:  = 1$ mode. In their numerical simulations  neutron stars are substituted

130:  with gravitating inner boundaries. \citet{fry04} performed

131:  three-dimensional SPH simulations and asserted that neutrino asymmetries

132:  decelerated a kicked neutron star significantly, which \citet{sch06}

133:  argued were negligible. The second mechanism is

134:  a neutrino-magnetic-field driven kick. \citet{lai98} studied a

135:  possibility that an asymmetric field topology could produce a large kick

136:  when combined with neutrino emission. They pointed out problems in

137:  former similar works and improved on how to deal with microphysics. The

138:  conclusion which they derived was that at least $\sim 10^{16}$~G difference of

139:  magnetic field strength between the north and south poles of a

140:  proto-neutron star is 

141:  necessary to get 300~km~s$^{-1}$ kick velocity. \citet{arr99}

142:  investigated parity violation effects on neutron star's kick with again

143:  improved microphysics. They argued that for the generation of the kick

144:  velocity of a few

145:  hundred~km~s$^{-1}$, at least $10^{15}-10^{16}$ G

146:  of dipole magnetic field is necessary in a proto-neutron star.  

147:  The third mechanism is an electromagnetically-driven kick, which was

148:  suggested by \citet{har75}. According to their

149:  analysis neutron star's kick velocity of several hundreds~km~s$^{-1}$ can

150:  be generated when the initial rotation period of pulsar is $\sim 1$ ms

151:  though with the initial rotation period of $\sim 10$ ms the result will

152:  be two orders of magnitude smaller than the former velocity. This mechanism

153:  does not depend on the strength of magnetic fields. The main difference

154:  from the other two mechanisms is a time scale of kick generation. With

155:  a surface field of $\sim 10^{12}$~G and an initial rotation period of

156:  $\sim 1$ ms, the acceleration time scale of a pulsar is approximately five

157:  years.

158: 

159: In this paper we consider an another class mechanism,

160: ``\textit{magnetohydrodynamically-driven kick}''. Consider the

161: situation that the rapidly rotating core has the dipole-like magnetic

162: field which is somewhat offset from the center of the core   

163: prior to collapse. Then, it is expected that a supernova with

164: equatorially-asymmetric bipolar jet will occur. This

165: will lead to a kick of the neutron star. Magnetar-class strong magnetic field

166: is required for this mechanism to work. If this were to a kick mechanism 

167: of ordinary neutron stars, the magnetic field must decay from 

168: $\sim10^{15}-10^{16}$~G to $\sim 10^{12}$~G. (This is

169: also the case for the second mechanism above.) \citet{gol92} estimated

170: the decay time scales of the magnetic field

171: in a neutron star due to Ohmic dissipation or ambipolar

172: diffusion. Their results imply that the magnetar-class magnetic fields

173: confined in 

174: neutron stars require at least $\sim 10^6 - 3 \times 10^7$ years to

175: decay to the strength of typical neutron stars ($\sim 10^{12}$ G). This

176: is much longer than the ages of not small number of ordinary

177: pulsars. Hence, we ignore the kicks of ordinary pulsars here, but are

178: concerned only with the kicks of magnetars.  

179: 

180:   According to the ``\textit{magnetohydrodynamically-driven kick}'' mechanism,

181:  we numerically study the core-collapse of massive star with off-centered 

182:  strong diploe-like magnetic fields and rapid rotation.

183: 

184:  The initial magnetic 

185:  field strength chosen here corresponds to the magnetar-class or larger 

186:  ($\sim10^{15}-10^{16}$ G) when the core contracts to a typical neutron-star 

187:  radius. This means that we adopt a so-called fossil origin hypothesis of 

188:  magnetic fields in magnetars. Although the origin of magnetic fields

189:  in magnetars are still a mystery, \citet{fer05} asserted that the magnetic

190:  flux of $\theta$ Orionis C, which is an O star, corresponds to that of

191:  magnetars and that the magnetism of magnetars could be explained as the

192:  fossil of progenitors as in the case of white dwarfs
 \citep{wic05}. In

193:  fact, the magnetic flux of $\theta$ Orionis C is

194:  $\sim~1\times10^{27}$~G~cm$^2$, which corresponds to $\sim~10^{15}$~G of

195:  the surface field if a neutron star is formed with the frozen magnetic

196:  field. There are several other OB stars suggested 

197:  observationally to possess magnetar-class magnetic fluxes. For example, 

198:  the estimated magnetic flux of HD191612 \citep{don02} corresponds to the

199:  surface field strength of $\sim~5\times10^{15}$~G, whereas several

200:  $\sim~10^{14}$~G of surface magnetic fields are expected for $\zeta$

201:  Cassiopeiae \citep{nei03a}, $\omega$ Orionis \citep{nei03b}, $\xi$ CMa

202:  \citep{hub06} and V2052 Ophiuchi \citep{nei03c} if their magnetic fields are

203:  compressed in the formation of a neutron star. Among them, $\theta$ Orionis C

204:  and HD191612 are O stars while the others are B stars. On the other hand,

205:  \citet{tho93} proposed that strong magnetic fields of magnetars originate

206:  from the convective dynamo process, which requires a rotation period shorter

207:  than $\sim 30$ ms. However, whether such a process really operates in neutron

208:  stars or not, is still uncertain. In this paper, we take a scenario that some

209:  OB stars have large magnetic fluxes and end up with a magnetar (fossil origin

210:  hypothesis). Although the initial magnetic fields assumed in this study are

211:  still a little larger than those implied by above observations, we are

212:  exploring the extreme limit. 

213: 

214: The offset of

215: the magnetic field in our computation may be supported by numerical 

216: simulations of fossil magnetic field in A stars and white dwarfs carried

217: by \citet{bra04}. They found that initially random magnetic

218: fields evolved into a coherent dipole-like field configuration stabilized by

219: a torus of twisted magnetic field lines. This magnetic field diffuse

220: within $\sim 10^9$~years which is much longer than life time of massive

221: stars. We would like to stress that the resultant magnetic field was

222: usually somewhat offset from the center of the star. Note that they assumed

223: fossil origin hypothesis of magnetic field which we also adopt here.   

224: 

225: The rotation period assumed in our rapid rotation models (see

226: \S~\ref{model}) is an order of magnitude shorter than the result of

227: \citet{heg05} who calculated the evolution of rotating massive

228: stars. In spite of the importance of their study, it is not still the

229: conclusion. The angular momentum distributions at the pre-collapse stage

230: should be studied further in the future. Our position is that they are

231: not well-known yet and just consider the extreme case.    

232: 

233: Since observations of magnetar kick velocities are very scarce so far,

234: our standpoint in this paper is predicting future 

235: observations of magnetar's proper velocities. Proper motions 

236: have been measured for only two magnetar candidates (both are AXPs), in

237: which just upper limits were put; $\lesssim 1400(D/10$ kpc$)$~km~s$^{-1}$ for

238: 4U0142+61  \citep{hul00,woo06} and $\lesssim 2500(D/3$ kpc$)$~km~s$^{-1}$

239: for 1E2259+586 \citep{oge05}. \citet{gae01} examined the reliability of

240: AXP/SGR association with supernova remnants (SNRs). The associations

241: which they considered probable enable us to estimate velocities of

242: AXPs and SGRs from their positions in the SNRs: $<$ 500~km~s$^{-1}$ for AXP

243: 1E 1841-045, $<$ 400~km~s$^{-1}$ for AXP 1E

244: 2259+586, and $<$ 500~km~s$^{-1}$ for AX J1845-0258 (AXP candidate)

245: \citep[][and reference their in]{gae01}. \citet{cli82} evaluated the

246: velocity of SGR 0526-66 assuming that it is associated

247: with SNR N49 and derived

248: $v_{kick} \sim 400-1400$ km~s$^{-1}$. \citet{cor99} derived $\sim

249: 800-1400$~km~s$^{-1}$ velocity of SGR 1627-41 from the association with

250: SNR G337.0-0.1. However, \citet{gae01} insisted

251: that these two associations between the SGRs and the SNRs are

252: less convincing. 

253: 

254: Recently, we reported the results of

255: numerical study that configurations of magnetic fields affect supernova

256: dynamics significantly \citep{saw05}. The effects of the offset dipole

257: field on the supernova explosion is another subject which we are

258: interested in this paper. We particularly focus on how dynamics and the

259: explosion energy are altered and that north-south difference of the

260: magnetic and rotational energies. 

261: 

262: The rest of this paper organized as follows. We describe numerical

263: models such as basic equations, a equation of state, magnetic field and

264: rotation, and the definition for proto-magnetar in $\S$~2. In $\S$~3,

265: the some numerical results including dynamics and kick velocities are

266: shown. In $\S$~4, we give some discussions and conclude the paper. 

267: 

268: \section{Models}\label{model}

269: We numerically simulate the core-collapse of magnetized massive stars

270: with the numerical code ZEUS-2D developed by \citet{sto92}.

271: A $1.5M_\odot$ core of $15 M_\odot$ star

272: provided by \citet{woo95} is chosen as a pre-collapse model, which gives

273: spherically

274: symmetric profiles of the density and specific internal energy. 

275: Magnetic field and rotation are just added to the core by hand (see \S

276: \ref{magrot}). 

277: 

278: \subsection{Basic Equations}

279: The following ideal MHD equations are solved:

280: \begin{eqnarray}

281:   \frac{D\rho}{Dt}+\rho\nabla\cdot\mbox{\boldmath $v$}=0\label{mac},\\

282:   \nonumber\\

283:   \rho\frac{D\mbox{\boldmath $v$}}{Dt}=-\nabla p-\rho\nabla\Phi+\frac{1}{4\pi}

284:    (\nabla\times\mbox{\boldmath $B$})\times\mbox{\boldmath $B$}\label{moc},\\

285:   \nonumber\\

286:   \rho\frac{D}{Dt}(\frac{e}{\rho})=-p\nabla\cdot

287:    \mbox{\boldmath $v$}\label{enc},\\

288:   \nonumber\\ 

289:   \frac{\partial\mbox{\boldmath $B$}}{\partial t}=\nabla\times

290:    (\mbox{\boldmath $v$}\times\mbox{\boldmath $B$}),\label{far}

291: \end{eqnarray}

292: where $\rho$, \mbox{\boldmath $v$}, $p$, $e$, $\Phi$, \mbox{\boldmath

293: $B$} are the density,

294: velocity, internal energy density, gravitational potential,

295: and magnetic field, respectively. We denote

296: the Lagrangian derivative as \mbox{$\frac{D}{Dt}$}.

297: 

298: A computational region is a half of the meridional plane with axisymmetric

299: assumption. $200(r) \times 60(\theta)$ grid points extending to 2000~km

300: are used initially. After central density reaches $10^{12}$ g

301: cm$^{-3}$, we change the radial mesh resolution and a radius of outer

302: boundary into $300(r)$ and 1500~km, respectively. The radial grid spacing are

303: not uniform with a finer resolution for smaller radii, while the angular

304: grid points are put uniformly. 

305: 

306: \subsection{Equation of State}\label{eos}   

307: As in the previous papers by \citet{yam04} and \citet{saw05}, we use a

308: parametric EOS

309: which was first introduced by \citet{tak84}. Although we

310: drastically simplify complicated microphysics such as the

311: neutrino transport, it is not a serious concern since our computation is

312: done only on the time scale of prompt explosion, during which the neutrino

313: transport is less important.

314: 

315: The parametric EOS we employed in this paper is as follows;

316: \begin{eqnarray}         

317:  p_{tot}=p_c(\rho)+p_t(\rho,e_t),\label{eosm}\\

318:  p_c(\rho)=K\rho^\Gamma,\label{eosc}\\ 

319:  p_t(\rho,e_t)=(\gamma_t-1)\rho\epsilon_t\label{eost}.

320: \end{eqnarray}

321: The pressure consists of two parts, the cold part ($p_c$) and the thermal

322: part ($p_t$). The thermal part is a function of the density and the

323: specific thermal energy, $\epsilon_t$, in which $\gamma_t$ is the

324: parameter called the thermal stiffness. The cold part is a function

325: of the density alone where the two constants, $K$ and $\Gamma$, take

326: account of the effect of the degeneracy of

327: leptons and the nuclear force. We choose the same values for $\gamma_t$,

328: $K$ and $\Gamma$ as in \citet{saw05}, so that a numerical simulation without

329: either magnetic filed or rotation fails to explode as in recent realistic

330: simulations. Readers are referred to the paper by \citet{saw05} for more details about this EOS.

331: 

332: \subsection{Magnetic field and Rotation}\label{magrot}

333: The initial magnetic field is assumed to have a dipole-like structure produced

334: by the sum of the line currents

335: which are uniformly distributed in the spherical region within a 1500~km

336: radius. The equatorially-asymmetric dipole magnetic field is obtained by

337: displacing equatorially-symmetric field along the 

338: rotation axis. Then the initial magnetic filed in

339: the cylindrical coordinate is given by

340: \begin{eqnarray}

341: B_{\varpi}(\varpi, z) = \frac{2J}{c} \sum_{\varpi_J} \sum_{z_J}\frac{z^{'}}{\varpi \sqrt{(\varpi_J + \varpi)^2 + z^{'}}}\Bigg[-K + \frac{\varpi_J^2 + \varpi^2 + z^{'2}}{(\varpi_J - \varpi)^2 + z^{'2}} E \Bigg],\label{emag1}\\

342: B_z (\varpi, z)= \frac{2J}{c} \sum_{\varpi_J} \sum_{z_J}\frac{1}{\sqrt{(\varpi_J + \varpi)^2 + z^{'}}}\Bigg[ K + \frac{\varpi_J^2 - \varpi^2 - z^{'2}}{(\varpi_J - \varpi)^2 + z^{'2}} E \Bigg],\label{emag2}\\

343: z^{'} = z - z_J - z_{off},\label{emag3}

344: \end{eqnarray}

345: where $R$ is a distance from the center, and $(\varpi_J, z_J)$ is the

346: mesh point where the line currents

347: exist. $J$, $K$, $E$, $c$, and $z_{off}$ are the current, complete

348: elliptical integral of the first kind, complete elliptical integral of

349: the second kind, the speed of light, and the degree of offset,

350: respectively. 

351: We compute with changing the magnetic field strength and

352: degree of offset (see Table~\ref{tmodels}). Note that, if the result of

353: \citet{bra04} is adopted to the iron core of a 2000~km radius, the

354: degree of displacement is several hundred~km. Fig.~\ref{fbfield} shows

355: the initial magnetic field configuration of models MR3 and SR10 (for the

356: name of models, see below).

357:  

358: The initial angular velocity in our simulations is assumed to have a

359: differentially rotating distribution,

360: \begin{equation}

361: \Omega(r)=\Omega_0\frac{r_0^2}{r_0^2+r^2},

362: \label{dif}

363: \end{equation} 

364: where $\Omega_0$ and $R_0$ are constants. With this distribution,

365: rotation is faster at small radii. In all models,

366: $R_0 = 1000 $~km is chosen, with which the initial differential rotation is

367: rather mild. We compute the models with rapid

368: rotation and slow rotation cases (see Table~\ref{tmodels}). The rapid

369: and slow rotation correspond to a neutron star rotating with the period

370: of $\sim 1$~ms and $\sim 10$~ms, respectively, if the angular

371: momentum conservation is assumed.

372: 

373: In Table \ref{tmodels}, six models computed are summarized. The

374: name of each

375: model consists of three parts, two characters and a number. The first

376: character denotes a strength of magnetic field; ``Moderate'' and

377: ``Strong''. The second character represents a rotation period, ``Slow''

378: and ``Rapid''. The attached number stands for a degree of

379: magnetic-field displacement.

380: For example, model MR3 has a moderate magnetic field, rapid rotation and

381: the displacement of 300~km. 

382: 

383: \subsection{Definitions of a Proto-Magnetar and its

384:   Velocity}\label{nsdef} 

385: We define a proto-magnetar (or a proto-neutron star) in our analysis

386: as a region which has a mass of $1.2 M_\odot$. The mass of each fluid

387: element is summed up descending order of the density. This is almost the 

388: maximal mass of fluid which remains through the simulations in the numerically

389: active region and does not go out of the outer boundary.  

390: Neutron stars in binary systems typically have the mass of around

391: $1.4M_{\odot}$ \citep[e.g.][]{tay82,tho99}. However, the mass of

392: magnetar candidates is 

393: unknown at present. \citet{sch06} discussed that

394: resultant mass of neutron star decreases as explosion time scale becomes

395: short. If explosion occurs in a prompt manner, a $1.2 M_\odot$ magnetar

396: may not be so bad. 

397: 

398: The velocity of the proto-magnetar is defined as follows,

399: \begin{eqnarray}         

400: v_{NS}(t) = \int_{t_0}^{t} \frac{F(t')}{M_{NS}} \mathrm{d}t', \label{evnsdef1}\\

401: F(t) = \oint_{S_{NS}} ( - P \mbox{\boldmath $n$} - \frac{B^2}{8 \pi} \mbox{\boldmath $n$} + \frac{B_z}{4 \pi} \mbox{\boldmath $B$} ) \cdot \mbox{\boldmath $n_z$} \mathrm{d} S - \int_{V_{NS}} \rho \nabla \Phi_{outer} \mathrm{d}V, \label{evnsdef2}

402: \end{eqnarray}

403: where, \mbox{\boldmath $n$}, \mbox{\boldmath $n_z$}, and

404: $\Phi_{outer}$ represent a local unit

405: vector normal to each surface element on a proto-magnetar, a unit

406: vector parallel to the rotation axis, and the gravitational potential

407: attributed 

408: to mass distribution in the outer layer of the proto-magnetar,

409: respectively. $t_0$ is the time when the central density reaches 

410: $1.0 \times10^{12}$~g cm$^{-3}$, before which the acceleration of the

411: proto-magnetar is negligible.

412: 

413: The velocity of a magnetar also could be defined in principle as the

414:  center of mass velocity. However we do not adopt this definition here

415:  because it leads to enormous numerical errors in calculating the volume

416:  integral of momenta over the region of the proto-magnetar. 

417:  The velocity at each grid point is obtained by the time-integration of the 

418:  volume forces (pressure gradient, Lorentz force and gravity) exerted there

419:  and the advection term. Hence, the volume integral of the momentum over 

420:  the proto-magnetar is equal to the time integration of the volume integral of

421:  these forces and the advection term over the proto-magnetar. 

422:  The difficulty in the numerical estimation of the proto-magnetar's kick

423:  velocity lies in the volume integral of the forces and the advection term. In

424:  fact, the error in calculating the proto-magnetar's kick velocity in this way

425:  comes not only from the surface of the proto-magnetar but also from its

426:  entire volume, and the problem is that the values of the integrand at the

427:  center are greater by a few orders of magnitude than those at the surface of

428:  the proto-magnetar and, as a result, errors of a few percent around the

429:  central region of the proto-magnetar correspond to errors of more than 100

430:  percent around the surface. On the other hand, if we calculate the

431:  proto-magnetar's kick velocity by time-integrating the surface forces, the

432:  resultant numerical errors originate from the local errors in the forces on

433:  the proto-magnetar surface and their magnitude will be comparable to the

434:  typical error, namely a few percent. This corresponds to an error of a few

435:  10 km/s in the proto-magnetar's kick velocity in model MR3, for example.

436:  It should be noted that the employment of a conservative scheme does not

437:  guarantte the correct spatial distribution of momentum. Moreover, since the

438:  self-gravity cannot be written in a conservative form, the strict

439:  conservation of the total momentum cannot be expected after its inclusion. 

440: 

441: 

442: \section{Results}\label{result}

443: We show the numerical results of our computation in this section.

444: The explosion dynamics, explosion energies, final states of

445: magnetic field and rotation, and magnetar kicks are given separately.

446: 

447: \subsection{Dynamics}\label{dyn}

448: We first describe the dynamics of each model. The important parameters

449: for all models are summarized in Table 2 and~\ref{texp}. 

450: 

451: Model MR0 has the same

452: initial parameters as model C10 of \citet{saw05} except for the magnetic

453: field configuration. The magnetic field of C10 was parallel

454: to the rotation axis and its strength was stronger near the rotation

455: axis while the magnetic field of model MR0 is dipole-like with larger

456: strength near the center. Namely, models MR0 is 

457: a more ``centrally'' concentrated counterpart of model C10. The

458: dynamical evolution of model MR10 is described as follows.

459: Model MR0 first bounces at 142.7 ms after the onset of collapse when the

460: central density reaches $3.75 \times 10^{14}$~g~cm$^{-3}$. However the

461: first shock wave is generated not by the nuclear force but by the magnetic

462: force  0.2 ms prior to the nuclear bounce around a radius of 30~km near the

463: rotation axis, which was also the case for model C10 in \citet{saw05}. 

464: Indeed the magnetic field around the region is strong enough by the time

465: of bounce 

466: due to the compression and field wrapping that the magnetic pressure becomes

467: much larger than the matter pressure. Then, the magnetic-force-dominant regions

468: gradually spread outward and shock waves propagate through the core

469: accompanied by the formation of a bipolar jet. This occurs by

470: transferring the magnetic energy to the kinetic

471: energy. As shown in left panel of Fig.~\ref{fengy} and its inset, the

472: toroidal energy is produced by the 

473: rotational energy especially around the time of the first bounce ($\sim

474: 143$~ms) and the subsequent second bounce ($\sim 147$~ms). After the

475: amplifications, the toroidal magnetic energy clearly 

476: decrease as the kinetic energy of matters whose radial velocity is

477: positive (the outward kinetic energy) increases. The toroidal energy is

478: not only consumed to launch the jets but also continues to be produced by

479: the rotational energy during this phase. This is why the rate of the outward

480: kinetic energy increase is 

481: larger than that of the toroidal energy decrease. The poloidal field

482: energy also will be transferred into the kinetic energy although the

483: total poloidal magnetic energy itself increases with time as the matter

484: falls. The outer most shock front reaches a 

485: radius of 1500~km, 20~ms after the bounce (see top left panel of

486: Fig.~\ref{fvbeta}). The 

487: explosion energy at that time is  $4.4\times 10^{51}$~erg, which is

488: almost twice as large as the most explosive model C10 in

489: \citet{saw05}. This implies that centrally concentrated magnetic fields

490: tends to explode supernovae more powerfully. 

491: 

492: Offset magnetic field models with rapid rotation, MR3, SR3 and SR10,

493: produce equatorially-asymmetric explosions with the faster 

494: jet northward, in which the degree of asymmetry is the most 

495: remarkable for model SR10. (see Fig.~\ref{fvbeta}).  

496: In these models, shock waves are

497: generated also prior to the nuclear bounce due to the magnetic force,

498: which occurs first in the northern hemisphere 

499: and then in the southern hemisphere with $\sim 0.1-0.6$~ms delay. Then

500: the shocks propagate through the core with larger speed northward.

501: The magnetic energy is more dominant in the

502: northern hemisphere than in the southern hemisphere almost always until

503: the shock reaches the surface of the core although the toroidal part is

504: larger in the southern hemisphere (Fig.~\ref{fengmag}). This is why

505: the jet and shock propagation are faster in the north.

506: 

507: We show the reason why the toroidal magnetic energy is larger in the

508: southern hemisphere.

509: The time evolutions of the rotational energies in Fig.~\ref{fengmag}

510: show that matter collapses more deeply in the northern 

511: hemisphere around the first bounce for models MR3 and SR3\footnote{The

512: two bumps in the rotational energy evolution in each panel of

513: Fig.~\ref{fengmag} roughly show how deeply the core collapses,

514: reflecting the gravitational energy evolution.}. This is because

515: the stronger magnetic fields in the northern hemisphere extract the

516: rotational energy more efficiently, which results in that the

517: centrifugal force reduces and the matter falls favorably. On the other

518: hand, the same interpretation is not valid for model SR10, in which the

519: collapse is deeper in the southern hemisphere in turn around the first

520: bounce. It is considered that magnetic field not only

521: encourages the collapse by 

522: extracting the rotational energy, but also discourages it. The latter

523: effect occurs simply because the

524: magnetic force preclude the matter from falling where it is dominant,

525: i.e. around the rotation axis with a radius more than $\sim 100$~km for

526: the rapid rotation models.

527: Note that, for every model computed here, there always locally exist the region

528: where the magnetic pressure dominates the matter

529: pressure during collapse, and so this effect is significant in spite of the

530: fact that the total magnetic energy is much smaller than the rotational

531: energy. This effect seems to work well in model SR10. It is

532: likely that too strong magnetic field is disadvantageous for the infall

533: of the matter. For the

534: second bounce in each model, the magnetic field rather

535: works to decrease the infall rate which leads to deeper collapse in the

536: southern hemisphere since the magnetic field grows larger than that at the

537: first bounce. As a result,

538: the rotational energy becomes larger in the

539: northern hemisphere after the second bounce, and thus the toroidal

540: magnetic energy also becomes larger. 

541:  

542: Table~\ref{texp} shows that the mass ejection in the

543: southern hemisphere is larger than that in the northern

544: hemisphere for model MR3\footnote{The ejected mass is defined as the sum

545: of the mass of fluid elements with the positive total energy. With this

546: definition, a part of mass in the proto-magnetar is also summed.}. 

547: This is explained by the effect mentioned above, i.e. the magnetic force

548: prevents the matter infall. An amount of ejected mass is determined by

549: how much matter is accumulated around the outside of the boundary between

550: the inner and outer cores when the shock waves are generated. In fact,

551: it is the outer core where the braking of infall by the magnetic

552: forces works. Owing to stronger magnetic field, the mass infall in the

553: outer core is more effectively hampered

554: in the northern hemisphere although the

555: larger gravitational energy is extracted in the northern hemisphere

556: around the first bounce for models MR3 and SR3 and so at least the inner

557: core collapses well in this half. As a

558: result, mass eruption is smaller in the northern hemisphere. 

559: According to \citet{yam94}, the amount of available gravitational energy

560: is sensitive to the ejected mass for the prompt explosion with pure

561: rotation. However it will not be said for magnetorotational case. 

562: 

563: In summary, we find that the jets are right and fast in the northern

564: hemisphere and heavy and slow in the southern hemisphere for the rapid

565: rotation cases.

566: 

567: The slow rotation models, MS3 and SS3, produce an order of magnitude

568: weaker explosions than rapid rotation models. The expelled masses are

569: also smaller by an order of magnitude (see Table~\ref{texp}).   

570: In each of the two models, the magnetic field plays a dynamical role

571: only just behind the shock fronts as they expands

572: outward. However, unlike in the rapid rotation cases, the toroidal

573: magnetic field plays almost no role to power the shock but probably the

574: poloidal field single-handedly does. In fact the slow rotation can not yield

575: strong toroidal fields whose energy is enough to help explosion (see

576: Fig.~\ref{fengy}). The farthest shock front reaches a radius of 1500~km

577: about 70 ms after the bounce, which is 3.5 times longer than for model MR3. 

578: Contrary to the rapid rotation models, the slow rotation models expel

579: more mass in the northern hemisphere (see Table~\ref{texp}). This is

580: because the magnetic 

581: pressure during bounce is not strong enough to hamper the collapse (see

582: right panel of Fig.~\ref{fengy}) and the magnetic field just acts to extract

583: the rotational energy. Thus for the slow rotation models, the

584: jets are heavy and fast in the northern hemisphere and right and slow in

585: the southern hemisphere.

586: 

587: \subsection{The Explosion Energy}\label{seng}

588: The explosion energies for all model is given in Table~\ref{texp}.

589: They are estimated by the sum of the energies over the

590: fluid elements with the positive total energy when the shock front

591: reaches a radius of 1500~km. It can be seen that for the rapid rotation

592: models the offset dipole field

593: weakens the explosion in comparison of MR0 with MR3 and of SR3 with

594: SR10. For model MR3, the explosion energy in the northern hemisphere is

595: somewhat smaller than that of model MR0 while the explosion energy in

596: the southern hemisphere is just slightly larger than that of model MR0,

597: which makes model MR3 weaker exploder. These features are just reflected

598: by the amounts of the ejected mass, i.e. the larger mass ejected, the

599: stronger the explosion becomes (see Table~\ref{texp}). Namely, in this case, the

600: explosion energy is also controlled by the degree of infall in the outer core as

601: the amount of the ejected mass. For the comparison between models SR3

602: and SR10, the situation is different. Although the expelled mass is

603: larger for model SR10, the explosion energy is larger for model SR3.

604: This is because the stored magnetic energy after the bounce for model

605: SR3 is much larger than that for model SR10 both in the northern and southern

606: hemispheres (see Fig.~\ref{fengmag}). For model SR10, the magnetic field

607: is very sparse in the southern hemisphere from the beginning. 

608: Meanwhile, in the northern hemisphere, in spite of initially very strong

609: magnetic field it can not be amplified so much since the increase of

610: rotational energy due to collapse is very small (see \S~\ref{dyn} ). 

611: Not same as the rapid rotation models, the explosion energy is larger in

612: the northern hemisphere for the slow rotation models. This is simply

613: because mass ejection is larger in the north (see \S~\ref{dyn}).    

614: After all, what we find here is that the magnetic fields which induce the

615: explosions also work against the energetic explosions.    

616: 

617: \subsection{Formation of A Proto-Magnetar}

618: The proto-magnetars which we call in this paper were defined in

619: \S~\ref{nsdef}. They have much different shapes

620: between the rapid rotation models and slow rotation models. In upper two panels

621: of Fig.~\ref{fmagrot}, the final shape of the proto-magnetars for 

622: model MR3 and MS3 can be seen. For model MR3, this corresponds to the

623: region whose density is more than $2.0 \times 10^{7}$~g~cm$^{-3}$ and the

624: proto-magnetar has a butterfly-like shape in the meridional plane,

625: occupying large fraction in the computational domain. On

626: the other hand, the proto-magnetar of model MS3 whose critical density

627: is $7.8 \times 10^{9}$~g~cm$^{-3}$, has only a radius of $\sim 100$~km

628: and its shape is a oblate ellipsoid.

629: 

630: Final rotation-period profiles for models MR3 and MS3 are also displayed

631: in upper two panels of Fig.~\ref{fmagrot}. For model MR3 the constant

632: $\Omega$ surface in the proto-magnetar is rather cylindrical than

633: spherical because the matter is expelled strongly toward both poles.

634: There exist a

635: slowly-rotating torus with a $\sim 100$~km radius in the inner region of

636: the proto-magnetar. Starting from the surface of the torus, narrow zones

637: where the angular velocities are also small run along the inner line of

638: the wings of the ``butterfly''.

639: In the proto-magnetar, there is almost no region whose rotation period

640: is smaller than 10~ms. In model MS3, the

641: angular momentum distribution is disarray especially at small radii

642: though at large distances from the rotation axis it is marginally

643: cylindrical. All fluid

644: elements rotate with a rotation period of about a few seconds in the

645: proto-magnetar.  

646: 

647: The magnetic field distribution at the end of

648: simulations is shown in lower panels of Fig.~\ref{fmagrot}. In model MR3,

649: a region where the poloidal field is strong spindles along the rotation

650: axis to the surface of the core due to the formation of the strong bipolar

651: jet. The poloidal field has a collimated dipole-like shape around the

652: rotation axis and no

653: longer much off-centered as before while the toroidal field is somewhat

654: equatorially-asymmetric. Around the foot of the jets and the inner

655: part of the proto-magnetar, the toroidal fields are very weak since their

656: energy is transferred into the kinetic energy of jets. On the other

657: hand, in the outer part of the proto-magnetar (the wings of the

658: ``butterfly), the toroidal field is

659: dominant in most area. The global magnetic field structure is roughly such that

660: the dipole magnetic field pinched by the toroidal magnetic field lines,

661: the region with more than $\sim 10^{13}$ G regarded.

662: How this magnetic fields

663: will evolve during cooling of the proto-magnetar to a magnetar of $\sim

664: 10$~km radius

665: is of another interest which should be investigated elsewhere in the

666: future. In model MS3, areas where the toroidal field is comparable to

667: the poloidal field are very small. There is a cylinder at the center of the

668: core with

669: $\sim 1400$~km length and $\sim 400$~km radius (a blue-colored cavity in

670: Fig.~\ref{fmagrot}) where is totally poloidal-field

671: dominant. The proto-magnetar is in the most inner part of this cylinder,

672: in which the toroidal 

673: field is especially weak. The poloidal field is also non-offset

674: dipole-like but is not as collimated as in the case of model MR3 owing

675: to weakly ejected mass.  

676: 

677: \subsection{Kick Velocity of a Nascent Magnetar}

678: We show, in Fig.~\ref{fvns}, the time evolution of velocities of the

679: proto-magnetars for all models. Models MR3, MS3, SR3 and SS3 produce kick

680: velocities of around $350-500$~km s$^{-1}$ while model SR10 yields a

681: velocity of more than 1000~km s$^{-1}$. Note that in the

682: equatorially-symmetric model MR0, the velocity of the proto-magnetar is less

683: than 70~km s$^{-1}$, which are due to un-removable numerical noises. 

684: The final velocities are estimated at 240 ms

685: from the beginning of computations. 

686: 

687: In model MR3, the proto-magnetar is

688: kicked southward and finally reaches a kick velocity of

689: 512~km~s$^{-1}$. As shown in  

690: Fig~\ref{fans}, the driving force is the magnetic pressure. Since the

691: magnetic pressure generally pushes the matter and the magnetic energy is

692: superior in the northern hemisphere (see

693: Fig.~\ref{fengmag}\footnote{This figure 

694: does not shows the magnetic energies just around the proto-magnetar but

695: those integrated over the northern or southern hemisphere. Hence, it

696: gives just a rough indication here.}), the

697: proto-magnetar is pushed southward. Meanwhile, the magnetic

698: tension which generally pulls the matter works northward again with the help

699: of stronger magnetic field in the northern hemisphere although this is

700: somewhat weaker than the magnetic 

701: pressure. Now we discuss what determines which of these two stresses

702: becomes dominant. The z-component of the force owing to magnetic

703: pressure and tension of the poloidal field in cylindrical coordinates are  

704: \begin{eqnarray}         

705: F_{mp,z} = -\frac{1}{8\pi}\int_{S}B^2 dS_z = -\frac{1}{8\pi}\int_{S}B^2 \cos\theta_2 dS,\label{efmp}\\

706: F_{mt,z} = \frac{1}{4\pi}\int_{S}B_z \mbox{\boldmath $B$}\cdot d\mbox{\boldmath $S$} = \frac{1}{4\pi}\int_{S} B^2 \cos\theta_1 \cos(\theta_1 - \theta_2) dS\label{efmt},

707: \end{eqnarray}

708: where, $B^2=B_z^2+B_R^2$, and $\theta_1$ and $\theta_2$ is the angles of

709: the magnetic field and 

710: surface element vector measured from the rotation axis, respectively. We

711: consider here in the range $\theta_2-90^\circ \le \theta_1 \le \theta_2 +

712: 90^\circ$ and $-90^\circ \le \theta_2 \le 90^\circ$, i.e. the magnetic

713: field is outward on the surface in the northern hemisphere. 

714: A contribution by the toroidal field is omitted for simplicity for the

715: moment. Where the magnetic tension is locally dominant, the following

716: relation works out from Eqs.~(\ref{efmp}) and ~(\ref{efmt}), 

717: \begin{eqnarray}         

718: 2B^2 \cos\theta_1 \cos(\theta_1 - \theta_2) - B^2 \cos \theta_2 = \cos(2\theta_1 - \theta_2) \ge 0,

719: \end{eqnarray}

720: and then $\theta_1$ satisfies

721: \begin{eqnarray}         

722: \frac{\theta_2}{2}-45^\circ \le \theta_1 \le \frac{\theta_2}{2}+45^\circ,

723: \end{eqnarray}

724: (see Fig.~\ref{fmagtp}). Regarding the northern hemisphere of model MR3

725: at 240 ms, Fig.~\ref{fmagrot} indicates that $\theta_2$ is often less than

726: -~60$^\circ$, where the magnetic force strongly works. Hence, the magnetic 

727: tension is dominant when $-75^\circ \lesssim \theta_1 \lesssim

728: 15^\circ$. If the effect of the toroidal field is taken into account, the

729: upper critical angle become somewhat larger because the toroidal field

730: is considered to be large in the southern hemisphere (see

731: \S~\ref{dyn}). On the other hand, the left-bottom panel of Fig.~\ref{fmagtp}

732: shows that $\theta_1$ is adequately more than $15^\circ$ almost

733: everywhere. Therefore the magnetic pressure is dominant

734: there and it becomes a leading force to accelerate the proto-magnetar. The same discussion can be done on the southern hemisphere. In fact in

735: model MR3 (and also in models SR3 and SR10), such a situation is common for

736: almost all time in the computation.  

737: 

738: Around $150-160$~ms in model MR3, the magnetic acceleration

739: decreases, which causes kicking-back of the proto-magnetar

740: northward. This decrease is likely to be yielded by the predominant

741: amplification of the toroidal field in the southern hemisphere around that time

742: (see Fig.~\ref{fengmag}), which makes the magnetic pressure work

743: northward. After kicking-back, the magnetic stress comes

744: back to life with even greater strength than before. 

745: This is because the difference of the magnetic stress

746: between the north and south surfaces becomes larger although the stresses

747: themselves are weaker than before. Then the

748: acceleration gradually decreases as the north-south difference in magnetic

749: stress diminishes. At the end of the computation, the acceleration is

750: quite small, and so the proto-magnetar no longer will be accelerated

751: significantly. Compared with the magnetic stress, the

752: pressure and gravity are less important for the acceleration of the

753: proto-magnetar except for the time of kicking-back. 

754: After the northward-shock

755: front passes the north surface of the proto-magnetar in first around

756: 150~ms, the pressure is larger in the vicinity of the north

757: surface. However, this is until the southward shock front passes the south

758: surface of the proto-magnetar with a few~ms delay. After this, the pressure

759: in the vicinity of the south surface is always larger because there is

760: more matter or the heat energy. The gravity pulls the

761: proto-magnetar to the direction where there is more matter. However,

762: after 162~ms, a part of matter is ejected out of the numerical domain, but we

763: neglect the gravity from the expelled matter. Hence, the value of the

764: gravitational acceleration in Fig.~\ref{fans} is not valid after this

765: time. Nevertheless,

766: keeping in mind that the gravitational acceleration is comparable to the

767: acceleration owing to the pressure, the effect of the gravity on the

768: kick velocity is no significance either way. The same can be said for

769: models SR3 and SR10. For models MS3 and SS3, since the ejected mass is

770: $\sim 0.03 M_\odot$, the gravitational

771: acceleration changes at most $\sim 10$ \% even if they are taken into account.

772: 

773: 

774: In models SR3 and SR10, the time evolution of magnetar velocities and

775: the acceleration mechanisms

776: are similar to that of model MR3. In model SR3, however, the magnetic

777: stress which has accelerated the proto-magnetar southward turns to work

778: oppositely around

779: 210 ms and the proto-magnetar starts to be decelerated. This is probably

780: because 

781: the magnetic field becomes locally stronger around the south surface of

782: the proto-magnetar. At the end of the

783: simulation, the magnetic stress still has a positive value. Thus, although

784: the proto-magnetar gets the final velocity of 353~km s$^{-1}$, this

785: will be decreased later in some degree. Model

786: SR10 produces the highest velocity of proto-magnetar among all

787: models. In spite of 

788: the smaller stored magnetic energy than that in model SR3, a considerable

789: north-south difference of them can produce a larger

790: accelerations. In this model, kicking-back occurs due not to reduction of

791: magnetic stress but to the pressure and gravity. At the end of

792: simulation the acceleration is still 

793: substantial. It will be negligible around 300 ms if linearly

794: extrapolated. The velocity of the proto-magnetar around that time will

795: be $\sim 1500-1600$~km~s$^{-1}$. 

796: 

797: The slow rotation models, MS3 and SS3, show more complex time evolution of

798: the proto-magnetar acceleration (see Fig~\ref{fans}). In each of these cases,

799: the magnetic pressure alone is not always the leading driving force, but

800: the magnetic 

801: tension, matter pressure or gravity also play important roles. This is

802: due to a low ratio of the magnetic pressure to

803: matter pressure in the vicinity of the proto-magnetar. Although these

804: models explode weakly, the final  

805: magnetar velocity are not differ significantly from those for the more

806: strong explosion models. This is because the kinetic energy of a

807: proto-magnetar whose

808: velocity is $\sim$ a few 100~km~s$^{-1}$ is just $\sim 10^{48}$~erg which is

809: even smaller than the explosion energies of model MS3 and SS3.

810: The time evolution of acceleration still varies 

811: violently even around the end of the computation. In order to get

812: reliable final velocities, longer numerical simulations are necessary.

813: 

814: \subsubsection{Uncertainties from the definition of A Proto-Magnetar}

815:  As noted in \S~\ref{nsdef}, we assume a proto-magnetar as region containing

816:  1.2~M$_\odot$, which are calculated by summing up the mass of each fluid

817:  element in descending order of the density. However, it is admittedly

818:  highly uncertain at 240~ms which part of the gas is finally settled to be a

819:  proto-magnetar. In fact, this will not be known for sure until computations

820:  are done up to the time, by which a fall back of matter has been finished.

821:  According to Zhang et al. (2007),

822:  this time is estimated to be about $10^6$~s, which is far longer

823:  than our computational time, 240~ms, and is, unfortunately, too long to

824:  dynamically follow with numerical simulations like ours. What we can do and

825:  we should do instead is to make clear the uncertainty of the kick

826:  velocities associated with the definition of a proto-magnetar. For this

827:  purpose, we have computed another model, which covers a greater portion of

828:  the star, and we have also calculated the kick velocities with several

829:  alternative definitions of a proto-magnetar, taking model MR3 as an example.

830: 

831:    We first investigate the uncertainties coming from the proto-magnetar

832:  mass. Five different masses of proto-magnetars, 1.2, 1.23, 1.26, 1.3,

833:  1.35~M$_\odot$, are tried. We take here the maximal mass to be

834:  1.35~M$_\odot$ anticipating that a proto-magnetar produced by a magnetic

835:  prompt explosion will be less massive than 1.4~M$_\odot$, the canonical value

836:  for the ordinary radio pulsars. We then introduce another criterion for a

837:  proto-magnetar, by which fluid elements are summed up in the descending order

838:  of "pressure" instead of density. Several masses of proto-magnetar are

839:  tried again with this "pressure criterion". In the study of proto-magnetars

840:  more massive than 1.2~M$_\odot$, a larger portion of the progenitor star

841:  needs to be computed. We thus run another numerical simulation, which

842:  computes the evolution of the inner 2.0~M$_\odot$ of the 15~M$_\odot$

843:  progenitor by \citet{woo95} for the same parameter values of $|E_m/W|$,

844: $|T/W|$, and z$_{off}$ as in model MR3, and obtain the kick velocities

845:  of 10 different proto-magnetars. We give each model the name as

846:  MR3env[D or P] [proto-magnetar's mass], where ``D'' and ``P'' denote the

847:  density- and pressure-criterion, respectively. For example, a model for

848:  1.23~M$_\odot$ proto-magnetar determined by the pressure-criterion is

849:  referred to as MR3envP1.23.

850: 

851:    The results are shown in Fig. 9, which illustrates the time evolution of

852:  the velocity for each proto-magnetar. We find that the difference in

853:  the criterion does not produce a significant variation in the kick

854:  velocities. On the other hand, it is found that the difference in the

855:  mass of proto-magnetar results in a noticeable variation in the kick

856:  velocities. In both series, MR3envD and MR3envP, the velocity tends to

857:  decrease with increasing mass. This is because a proto-magnetar with a higher

858:  mass has a larger radius and the magnetic field on the surface becomes

859:  weaker, which then leads to a smaller north-south difference in the magnetic

860:  forces. Note, however, that the most massive proto-magnetar models in our

861:  analysis, MR3envD1.35 and MR3envP1.35, still have kick velocities of

862:  $\sim$~200~km~s$^{-1}$.

863: 

864:    The kick velocities obtained with several different definitions of a 

865:  proto-magnetar cover a range of $\sim~200-500$~km~s$^{-1}$. We thus consider

866:  that the kick velocities we have procure in this study have uncertainty of

867:  this degree. We can thus claim that a proto-magnetar is possibly accelerated

868:  up to several hundred~km~s$^{-1}$ by the magnetohydrodynamic process

869:  considered in this paper.

870: 

871: 

872: \section{Discussion and Conclusion}

873: We have done series of numerical simulations of core-collapse supernovae

874: with strong offset-dipole magnetic fields (offset northward) and rapid

875: rotations. Main 

876: purposes of our study was investigating how dynamics and magnetar

877: kicks behave with such off-centered magnetic fields. We finally found that:

878: 

879: 1. Equatorially-asymmetric explosions occur for all models with a

880:  formation of bipolar jets. Jets are fast-light in the north and

881:  slow-heavy in the south for the rapid rotation models while they are

882:  fast-heavy in the north and slow-light in the south for the slow

883:  rotation models. 

884: 

885: 2. Off-centered magnetic field weaken the explosion for the rapid

886:  rotation case. In this case, the explosion energy is larger in

887:  the south since less matter is ejected or less magnetic energy is

888:  stored in the north. On the other hand, the explosion energy is larger

889:  in the north for the slow rotation models.

890: 

891: 3. The formed proto-magnetars are slow rotator with the rotational periods of

892:  more than 10~ms. The final magnetic field around the proto-magnetar has

893:  a collimated dipole-like configuration pinched by the toroidal field

894:  lines for the rapid rotation models whereas the proto-magnetar formed

895:  in the slow rotation models is totally poloidal-field dominant. 

896: 

897: 4. If the initial magnetic field is stronger in the northern hemisphere,

898:  proto-magnetars are kicked southward with the velocities of several

899:  hundred~km~s$^{-1}$, or in extreme case $\sim$~1000~km~s$^{-1}$.   

900:  In most cases, they are accelerated mainly by the magnetic pressure

901:  while the somewhat weaker magnetic tension works the opposite

902:  direction, which is due to stronger magnetic field in the northern

903:  hemisphere. 

904: 

905: Form the results of computations, we predict that magnetars also

906: possibly have large velocities as ordinary pulsars do and in some

907: extreme cases they could have $\sim 1000$~km~s$^{-1}$ velocities.

908: Current observations show that at least

909: some magnetar candidates have the velocities of less than

910: 500~km~s$^{-1}$ (see $\S$~\ref{intro}) whereas \citet{dun92} claimed that

911: magnetars will have large

912: kick velocities up to $\sim 1000$~km~s$^{-1}$. It is notable that our

913: results also show that the some moderate initial conditions lead to the

914: velocities less than $\sim~500$~km~s$^{-1}$ which is not inconsistent

915: with the observations. On the other hand, if the association between SGR

916: 0526-66 and N49 or SGR 

917: 1627-41 and G337.0-0.1 is true, these magnetar candidates should have

918: an extremely large kick velocities up to $\sim 1000$~km~s$^{-1}$. In the

919: context of our numerical simulations, these extraordinarily high

920: velocities are interpreted as the products of the extreme initial

921: condition, i.e. very strong and highly asymmetric

922: magnetic fields as in model SR10. 

923: 

924: Some observations of recent date imply

925: that some magnetars are originate from massive progenitor 

926: ($\gtrsim 20-50M_{\odot}$) \citep[e.g.][]{gae05,fig05,mun06}. We use a $15

927: M_\odot$ stellar model although our computation 

928: involves a formation of magnetars. This is because a $15 M_\odot$ star

929: is a canonical progenitor of core-collapse computations and it has not

930: known yet whether a massive-star-origin magnetar is common. However, it

931: is worth investigating in the future whether the similar results to us are

932: yielded by the numerical simulation with more massive progenitors. When

933: the massive progenitors are chosen, it is more natural for us to assume rapid

934: rotation since \citet{heg05} argued from their calculation

935: of stellar evolution that more massive stars tend to rotate more

936: rapidly. According to them, $35 M_\odot$ star will reach the rotation

937: period of the neutron star of $4.4$ ms in there standard model, though

938: the magnetic flux of the star is far weaker than that of magnetars and

939: again this result is not conclusive.

940: 

941: The other kick mechanisms introduced in \S~\ref{intro} does not work in

942: the situation considered here. For the hydrodynamically-driven and

943: neutrino-magnetic-field driven mechanism, this is because the kicks are

944: accompanied with delayed explosion while, in our simulation, prompt

945: explosion occurs with the help of strong magnetic field and rotation. 

946: In an electromagnetically-driven mechanism, the acceleration time

947: scale is $\sim $ 4 s if the field strength on the surface of the

948: (proto-) magnetar is $\sim 10^{16}$~G, which is much longer than that of

949: ``\textit{magnetohydrodynamically-driven kick}''. Our computation shows

950: that the initially rapid rotation becomes quite slow (more than 10~ms

951: rotation period) in $\lesssim~100$~ms after bounce (see

952: \S~\ref{magrot}). Noted that this mechanism requires the rotation period

953: of $\sim 1$~ms, the proto-magnetar cannot be substantially accelerated.  

954: 

955: The above discussion is valid only when the initial conditions which we

956: assume, strong magnetic fields, rapid rotation, and offset dipole fields, are

957: achieved. Although validity of these assumptions is discussed in

958: $\S$~\ref{magrot}, they are far from solid. What will happen

959: if our assumptions are partially fulfilled? With strong magnetic fields

960: but a slow rotation, the magnetorotational explosion considered here can

961: not occur. In this case, the engine of supernova might be the heating by

962: emitted neutrinos, and the second kick mechanism (neutrino-magnetic

963: field driven) introduced in $\S$~\ref{intro} might work. We

964: have no idea at the moment whether the hydrodynamical-instability-driven

965: kick (the first mechanism) could operate or not in the presence of

966: strong magnetic fields. This should be investigated in the future

967: works. If the initial magnetic field is weak and rotation is rapid,

968:  the magnetorotational explosion may ensue with a help of

969:  magnetorotational instability (MRI) or convective dynamo action

970:  \citep[e.g.][]{aki03, ard05, dun92}. However, it needs more detailed

971:  studies. Note that in order to reliably compute the growth of MRI from weak

972:  magnetic fields of $\sim~10^8-10^9$~G in the supernova core, the spatial grid

973:  size should be at least as small as $\sim~10^2$~cm, which no numerical

974:  simulation has achieved yet so far. Meanwhile, even in the present

975:  computations with $\sim~10^{12}-10^{13}$~G initially, the grid resolution is

976:  not fine enough by a factor of ten in some important regions to resolve MRI. 

977:  In this case, however, the compression and wrapping-up of the initial field

978:  amplify the field strength sufficiently before MRI sets in and we expect MRI

979:  will not change the dynamics very much even if it occurs.

980:  But, we do not intend to claim that MRI is not important. On the contrary,

981:  we are interested in the MRI-related issues such as the linear growth,

982:  nonlinear saturation, implications for explosion, pulsar kick and spin and so

983:  forth. We think they should be addressed in a separate paper. In so doing,

984:  three-dimensional computations are important, which we are now undertaking

985:  (Sawai et al. 2008 in preparation).  

986: 

987: We find that the centrally concentrated magnetic field has an advantage in

988: energetic explosion. Recently, \citet{pia06} reported that SN 2006aj

989: associated with X-ray flash 060218 is a factor 2-3 more luminous than

990: other normal type Ic supernovae although it is dimmer than so-called

991: hypernovae associated with gamma-ray bursts. \citet{maz06} claimed that

992: the remnant of SN 2006aj is not a black hole but a neutron star,

993: possibly a magnetar. If this is true, a highly

994: energetic explosion induced by centrally-concentrated magnetic fields can

995: be one of the origins of this middle-class supernova although the

996: mechanism to produce the X-ray flash is beyond the scope of this

997: paper. Note that not every supernova accompanied by a magnetar

998: should produce such a energetic explosion. It is reported that

999: for some supernovae, which are considered to be associated with 

1000: magnetar candidates, the explosion

1001: energies are just as large as those of ordinary supernovae

1002: \citep{vin05,sas04}.

1003: 

1004: Finally, how large a fraction do magnetars occupy among the whole population of

1005: isolated neutron stars? At present, the confirmed candidates of magnetars are

1006: counted up to eleven which is only $\sim 1$ \% of the observed

1007: ordinary pulsars. However it is too early to conclude that magnetars

1008: are such a minor population. Spin-down ages are

1009: measured in three of four SGRs, which are younger than 1900 years. This means

1010: a magnetar is born every $\sim 600$ years, corresponding to $\sim 15$ \% of

1011: ordinary neutron stars. If another magnetar, SGR 1627-41, has a similar

1012: spindown age, the number of magnetars are estimated to be $\sim

1013: 20$ \% of that of the ordinary pulsars. Here we omit AXPs since there

1014: may be a large number of AXPs which are too dim to detect

1015: \citep{woo06}. \citet{woo06} insisted that there

1016: may be a lot of magnetars which are too old to be detected

1017: and that the number of magnetars could be comparable to that of ordinary

1018: pulsars. If this is the case, the study of the supernova theory in the

1019: strong-magnetic-field regime will have a more significant meaning.

1020:  

1021: \acknowledgments

1022: H. S. thanks to E. M\"{u}ller and H.-Th. Janka for useful discussion during his

1023: stay at Max-Planck-Institute f\"{u}r Astrophysik. H. S. also thanks

1024: D. Lai for helpful discussion.     

1025: Some of the numerical simulations were done on the

1026: supercomputer VPP700E/128 at RIKEN and VPP500/80 at KEK (KEK

1027: Supercomputer Projects No. 108). This work was partially supported by

1028: the Japan Society for Promotion of Science (JSPS) Research Fellowships

1029: (H.S.),

1030: the Grants-in-Aid for the Scientific 

1031: Research (14079202, 17540267) from Ministry of Education, Science and

1032: Culture of Japan, and by the Grants-in-Aid for the 21th century COE

1033: program ``Holistic Research and Education Center for Physics of

1034: Self-organizing Systems''.

1035: 

1036: 

1037: 

1038: \begin{thebibliography}{}

1039: 

1040: 

1041: \bibitem[Akiyama et al.(2003)]{aki03} Akiyama, S., Wheeler, J. C.,

1042: 			       Meier, D. L., \& Lichtenstadt, I. 2003,

1043: 			       \apj, 584, 954

1044: 

1045: %\bibitem[Ardeljan et al.(1998)]{ard98} Ardeljan, N. V.,

1046: %			       Bisnovatyi-Kogan, G. S., \& Moiseenko

1047: %			       S. G. 1998, LPN, 506, 145A

1048: 

1049: %\bibitem[Ardeljan et al.(2000)]{ard00} Ardeljan, N. V.,

1050: %			       Bisnovatyi-Kogan, G. S., \& Moiseenko

1051: %			       S. G. 2000, Astron. Astrophys. , 355, 1181

1052: 

1053: \bibitem[Ardeljan et al.(2005)]{ard05} Ardeljan, N. V.,

1054: 			       Bisnovatyi-Kogan, G. S., \& Moiseenko

1055: 			       S. G. 2005, Mon. Not. R. Astron. Soc.,

1056: 			      359, 333

1057: 

1058: \bibitem[Arras \& Lai(1999)]{arr99} Arras, Ph. \& Lai, D. 1999,

1059: 			      Phys. Rev. D, 60, 043001

1060: 

1061: %\bibitem[Balbus \& Hawley(1991)]{bal91} Balbus, S. A., \& Hawley, J. F. 1991,

1062: %			       \apj, 376, 214

1063: 

1064: \bibitem[Bisnovatyi-Kogan et al.(1975)]{bis75} Bisnovatyi-Kogan, G. S.,

1065: 			       Popov, YU. P., \& Samochin, A. A. 1975,

1066: 			       Ap\&SS, 41, 287B

1067: 

1068: \bibitem[Braithwaite \& Spruit (2004)]{bra04} Braithwaite, J., \& Spruit

1069: 				      H. C., 2004, Nature, 431, 819

1070: 

1071: %\bibitem[Blondin et al.(2003)]{blo03} Blondin, J. M., Mezzacappa, A., \&

1072: %			      DeMarino, C. 2003, \apj, 584, 971

1073: 			      

1074: %\bibitem[Buras et al.(2003)]{bur03} Buras, R., Janka, H. -T., \&

1075: %			       Kifonidis, K. 2003,

1076: %			       Phys. Rev. Lett., 90, 241101

1077: 

1078: \bibitem[Cline et al.(1982)]{cli82} Cline, T. L. et al. 1982, \apj, 255, L45

1079: 

1080: \bibitem[Corbel et al.(1999)]{cor99} Corbel, S., Chapuis, C., Dame,

1081: 			      T. M., \& Durouchoux, P. 1999, \apj, 526, L29

1082: 

1083: \bibitem[Cordes \& Chernoff(1998)]{cor98} Cordes, J. M., \& Chernoff,

1084: 			      D. F. 1998, \apj, 505, 315

1085: 

1086: \bibitem[Donati et al.(2002)]{don02} Donati, J.-F., Babel, J., Harries, T. J., 

1087:                                      Howarth, I. D., Petit, P., \& Semel,

1088:                                      M. 2002, Mon. Not. R. Astron. Soc., 333,

1089:                                      55 

1090: 

1091: \bibitem[Donati et al.(2006)]{don06} Donati, J.-F., Howarth, I. D., Bouret,

1092: 				  J.-C., Petit, P., Catala, C., \& Landstreet,

1093: 				  J. 2006, Mon. Not. R. Astron. Soc., 365, L6 

1094: 

1095: \bibitem[Duncan \& Thompson(1992)]{dun92} Duncan, R. C., \& Thompson,

1096: 				  C. 1992, \apj, 392, L9

1097: 

1098: \bibitem[Ferrario \& Wickramasinghe(2005)]{fer05} Ferrario, L. \&

1099: 			      Wickramasinghe, D. T. 2005,

1100: 			      Mon. Not. Astron. Soc., 356, 615 

1101: 

1102: %\bibitem[Ferrario \& Wickramasinghe(2006)]{fer06} Ferrario, L. \&

1103: %			      Wickramasinghe, D. T. 2006,

1104: %			      Mon. Not. Astron. Soc., 367, 1323

1105: 

1106: \bibitem[Figaer et al.(2005)]{fig05} Figer, D. F., Najarro, F., Geballe,

1107: 			      T. R., Blum, R. D., \& Kudritzki,

1108: 			      R. P. 2005, \apj, 622, L49

1109: 

1110: %\bibitem[Foglizzo \& Tagger(2000)]{fog00} Foglizzo, T., \& Tagger,

1111: %			      M. 2000, Astron. Astrophys., 363, 174

1112: 

1113: \bibitem[Fryer \& Warren(2004)]{fry04} Fryer, C. L., \& Warren,

1114: 			      M. S. 2004, \apjl, 601, 391

1115: 

1116: \bibitem[Gaensler et al.(2005)]{gae05} Gaensler, B. M.,

1117: 			      McClure-Griffiths, N. M., Oey, M. S.,

1118: 			      Haverkorn, M., Dickey, J. M., \& Green,

1119: 			      A. J. 2005, \apj, 620, L95

1120: 

1121: \bibitem[Gaensler et al.(2001)]{gae01} Gaensler, B. M., Slane, P. O.,

1122: 			      Gotthelf, E. V., \& Vasisht, G. 2001,

1123: 			      \apj, 559, 963

1124: 

1125: \bibitem[Gunn \& Ostriker(1979)]{gun79} Gunn, J. E., \& Ostriker,

1126: 			      J. P. 1979, \apj, 160, 979

1127: 

1128: \bibitem[Goldreich \& Reisenegger(1992)]{gol92} Goldreich, P. \&

1129: 			      Reisenegger, A. 1992, \apj, 395, 250

1130: 

1131: \bibitem[Harrison \& Tademaru(1975)]{har75} Harrison, E. R. \& Tademaru,

1132: 			      E. 1975, \apj, 201, 447

1133: 

1134: %\bibitem[Heger et al.(2003)]{heg03} Heger, A., Woosley, S. E., Langer,

1135: %			       N., \& Spruit, H. C. 2003, Proc. IAU 215

1136: %			       ``Stellar Rotation''

1137: 

1138: \bibitem[Heger et al.(2005)]{heg05} Heger, A., Woosley, S. E.,

1139: 			       \& Spruit, H. C. 2005, \apj, 626, 350

1140: 

1141: \bibitem[Hobbs et al.(2005)]{hob05} Hobbs, G., Lorimer, D. R., Lyne,

1142: 			      A. G., \& Kramer, M. 2005,

1143: 			      Mon. Not. R. Astron. Soc., 360, 974

1144: 

1145: \bibitem[Hubrig et al.(2006)]{hub06} Hubrig, S., Briquet, M., Sch\"{o}ller,

1146: 			      M., De Cat, P., Mathys, G., \& Aerts, C.

1147: 			      2006, Mon. Not. Astron. Soc. 369, L61

1148: 

1149: \bibitem[Hulleman et al.(2000)]{hul00} Hulleman, F., van Kerkwijk,

1150: 			      M. H., \& Kulkarni, R. 2000, Nature, 408, 689

1151: 

1152: %\bibitem[Janka et al.(2005)]{jan05} Janka, H. -T., Buras, R., Kitaura

1153: %			      Joyanes, F. S., Marek, A., Rampp, M., \&

1154: %			      Scheck, L. 2005, Proc. ``Nuclei in the

1155: %			      cosmos 8'', to appear in Nucl. Phys. A,

1156: %			      astro-ph/0411347

1157: 

1158: %\bibitem[Kotake et al.(2003)]{kot03} Kotake, K., Yamada, S.,

1159: %			       \& Sato, K. 2003, \apj, 595, 304

1160: 

1161: 

1162: \bibitem[Kotake et al.(2004)]{kot04} Kotake, K., Sawai, H., Yamada, S.,

1163: 			       \& Sato, K. 2004, \apj, 608, 391

1164: 

1165: \bibitem[Kotake et al.(2005)]{kot05} Kotake, Yamada, S.,

1166: 			       \& Sato, K. 2005, \apj, 618, 474

1167: 

1168: \bibitem[Lai(2003)]{lai03} Lai, D. 2003, Proc. ``Cosmic explosions in

1169: 			      three dimensions: asymmetries in

1170: 			      supernovae and gamma-ray bursts'', p276 

1171: 

1172: \bibitem[Lai \& Qian(1998)]{lai98} Lai, D. \& Qian, Y. 1998, \apj, 505, 844

1173: 

1174: \bibitem[LeBlanc \& Wilson(1970)]{leb70} LeBlanc, J. M., \& Wilson,

1175: 				  J. R. 1970, \apj, 161, 541

1176: 

1177: %\bibitem[Leonard et al.(2000)]{leo00} Leonard, D. C., Filippenko, A. V.,

1178: %			      Barth, A. J., \& Matheson, T. 2000, \apj,

1179: %			      536, 239

1180: 

1181: %\bibitem[Liebend\"{o}rfer et al.(2001)]{lie01} Liebend\"{o}rfer, M.,

1182: %			       Mezzacappa, A., Thielemann, F. -K.,

1183: %			       Messer, O. E. B., Hix, W. R., \& Bruenn,

1184: %			       S. W. 2001, Phys. Rev. D, 63, 103004

1185: 

1186: %\bibitem[Liebend\"{o}rfer et al.(2004)]{lie04} Liebend\"{o}rfer, M.,

1187: %			       Rampp, M., Janka, H.-T., \& Mezzacappa,

1188: %			       A. 2004, \apj in press, astro-ph/0310662

1189: 

1190: \bibitem[Mazzali et al.(2006)]{maz06} Mazzali, P. A. et al. 2006,

1191: 			      Nature, 442, 1018

1192: 

1193: %\bibitem[Meier et al.(1976)]{mei76} Meier, D. L., Epstein, R. I.,

1194: %			      Arnett, W. D., \& Schramm, D. N. 1976,

1195: %			      \apj, 204, 869

1196: 

1197: \bibitem[Moiseenko et al.(2006)]{moi06} Moiseenko, S. G.,

1198: 			      Bisnovatyi-Kogan, G. S., \& Ardeljan,

1199: 			      N. V. 2006, Mon. Not. R. Astron. Soc.,

1200: 	        	      310, 501

1201: 

1202: \bibitem[M\"{u}ller \& Hillebrandt(1979)]{mul79} M\"{u}ller, E., \&

1203: 			      Hillebrandt, W. 1979, Astron. Astrophys.,

1204: 			      80, 147

1205: 

1206: \bibitem[Muno et al.(2006)]{mun06} Muno, M. P. et al. 2006, \apj, 636, L41

1207: 

1208: \bibitem[Neiner et al.(2003a)]{nei03a} Neiner, C., Geers, V. C., Henrichs,

1209:   H. F., Floquet, M., Fr\'{e}mat, Y., Hubert, A. -M., Preuss, O., \& Wiersema,

1210:   K. 2003a, A\&A, 406, 1019

1211: 

1212: \bibitem[Neiner et al.(2003b)]{nei03b} Neiner, C., Hubert, A. -M., Fr\'{e}mat,

1213:   Y., Floquet, M., Jankov, S., Preuss, O., Henrichs, H. F., \& Zorec, J. 2003b,

1214:   A\&A, 409, 275 

1215: 

1216: \bibitem[Neiner et al.(2003c)]{nei03c} Neiner, C., Henrichs, H. F., Floquet,

1217:   M., Fr\'{e}mat, Y., Preuss, O., Hubert, A. -M., Geers, V. C., Tijani, A. H.,

1218:   Nichols, J. S., \& Jankov, S. 2003c, A\&A, 411, 565

1219: 

1220: \bibitem[Nishimura et al.(2006)]{nis06} Nishimura, S., Kotake, K.,

1221: 			      Hashimoto, M., Yamada, S., Nishimura, N.,

1222: 			      Fujimoto, S., \& Sato, K. 2006, \apj, 642,

1223: 			      410

1224: 

1225: \bibitem[Obergaulinger et al.(2006)]{obe06} Obergaulinger, M., Aloy,

1226: 			      M. A., \& M\"{u}ller, E. 2006, A\&A, 450, 1107

1227: 

1228: \bibitem[\"{O}gelman and Tepedelenlio\v{g}lu(2005)]{oge05} \"{O}gelman,

1229: 			      H. \& Tepedelenlio\v{g}lu 2005,

1230: 			      astro-ph/0503215, submitted to \apjl

1231: 

1232: \bibitem[Ohnishi(1983)]{ohn83} Ohnishi, T. 1983,

1233: 				  Tech. Rep. Inst. At. En. Kyoto

1234: 				  Univ. No.198

1235: \bibitem[Paczy\'{n}ski(1992)]{pac92} Paczy\'{n}ski, B. 1992, Acta

1236: 				       Astronomica, 42, 145

1237: 

1238: \bibitem[Pian et al.(2006)]{pia06} Pian, E. et al. 2006, Nature, 442, 1101

1239: 

1240: %\bibitem[Rampp \& Janka(2000)]{ram00} Rampp, M., \& Janka, H. -T. 2000,

1241: 			%       \apj, 539, L33 

1242: 

1243: \bibitem[Sasaki et al.(2004)]{sas04} Sasaki, M., Plucinsky, P. P.,

1244: 			      Gaetz, T. J., Smith, R. K., Edgar, R. J.,

1245: 			      \& Slane, P. O. 2004, \apj, 617, 322

1246: 

1247: 

1248: \bibitem[Sawai et al.(2005)]{saw05} Sawai, H., Kotake, K. \& Yamada, S. 2005,

1249: 			      \apj, 631, 446

1250: 

1251: %\bibitem[Sawai et al.(2006)]{saw06} Sawai, H., Kotake, K. \& Yamada, S. 2006,

1252: %                              J. Phs.: Conf. Ser., 31, 151

1253: 

1254: \bibitem[Scheck et al.(2006)]{sch06} Scheck, L., Kifonidis, K., Janka,

1255: 			      H.-Th., \& M\"{u}ller, E. 2006, A\&A, 457, 963

1256: 

1257: \bibitem[Shibata et al.(2006)]{shi06} Shibata, M., Liu, Y. T., Shapiro,

1258: 			      S. L., \& Stephens, B. C. 2006, accepted

1259: 			      for publication in Phys. Rev. D, 

1260: 			      astro-ph/0610840

1261: 

1262: %\bibitem[Shimizu et al.(2001)]{shi01} Shimizu, T. M., Ebisuzaki, T.,

1263: 			  %    Sato, K., \& Yamada, S. 2001, \apj, 552, 756 

1264: 

1265: 

1266: \bibitem[Stone \& Norman(1992)]{sto92} Stone, J. M., \& Norman

1267: 				  M. L. 1992, \apjs, 80, 791 

1268: 

1269: %\bibitem[Sumiyoshi et al.(2004)]{sum04} Sumiyoshi, K., Suzuki, H.,

1270: %			      Yamada, S., \& Toki, H. 2004,

1271: %			      Nucl. Phys., A730, 227 

1272: 

1273: \bibitem[Symbalisty(1984)]{sym84} Symbalisty, E. M. D. 1984, \apj, 285,

1274: 				  729

1275: 

1276: \bibitem[Takahara \& Sato(1984)]{tak84} Takahara, M. \& Sato,

1277: 				  K. 1984, Prog. Theor. Phys., 71, 524

1278: 

1279: \bibitem[Takiwaki et al.(2004)]{tak04} Takiwaki, T., Kotake, K.,

1280: 			       Nagataki, S., \& Sato, K. 2004, \apj,

1281: 			      616, 1086

1282: 

1283: \bibitem[Taylor \& Weisberg(1982)]{tay82} Taylor, J. H. \& Weisberg,

1284: 			      J. M. 1982, \apj, 253, 908

1285: 	

1286: 

1287: \bibitem[Thompson \& Duncan(1993)]{tho93} Thompson, C., \& Duncan,

1288: 				  R. C. 1992, \apj, 408, 194

1289: 

1290: 

1291: \bibitem[Thompson \& Duncan(1995)]{tho95} Thompson, C., \& Duncan,

1292: 				  R. C. 1995, Mon. Not. R. Astron. Soc.,

1293: 				       275, 255 

1294: 

1295: 

1296: \bibitem[Thompson \& Duncan(1996)]{tho96} Thompson, C., \& Duncan,

1297: 				  R. C. 1996, \apj, 473, 322 

1298: 

1299: %\bibitem[Thompson et al.(2003)]{tho03} Thompson, T. A., Burrows, A., \&

1300: 			%       Pinto, P. A. 2003, \apj, 592, 434

1301: 

1302: \bibitem[Thorsett \& Chakrabarty(1999)]{tho99} Thorsett, S. E. \&

1303: 			      Chakrabarty, D. 1999, \apj, 512, 288

1304: 		

1305: 

1306: \bibitem[Vink \& Kuiper(2005)]{vin05} Vink, J. \& Kuiper, L. 2005,

1307: 			      Mon. Not. R. Astron. Soc., 370, L14 

1308: 

1309: %\bibitem[Walder et al.(2005)]{wal05} Walder, R., Burrows, A., Ott,

1310: 			 %     C. D., Livne, E., Jarrah, M. 2005,

1311: 			  %    submitted to \apj, astro-ph/0412187 

1312: 

1313: %\bibitem[Wang et al.(1996)]{wan96} Wang, L., Wheeler, J. C., Li, Z., \&

1314: 			   %   Clocchiatti, A. 1996, \apj, 467, 435 

1315: 

1316: %\bibitem[Wheeler et al.(2000)]{whe00} Wheeler, J. C., Yi, I.,

1317: 			    %  H\"{o}flich, P., \& Wang, L. 2000, \apj, 537, 810

1318: \bibitem[Wheeler et al.(2002)]{whe02} Wheeler, J. C., Meier, D. L., \&

1319: 			       Wilson, J. R. 2002, \apj, 568, 807

1320:  

1321: \bibitem[Wickramasinghe \& Ferrario(2005)]{wic05} Wickramasinghe,

1322: 			      D. T. \& Ferrario, L. 2005,

1323: 			      Mon. Not. Astron. Soc., 356, 1576

1324: 

1325: \bibitem[Woods \& Thompson(2006)]{woo06} Woods, P. M. \& Thompson,

1326: 			      C. 2006, Compact Steller X-ray Sources

1327: 			      (Cambridge Astrophysics Series), 547

1328: 

1329: \bibitem[Woosley(1995)]{woo95} Woosley, S. E. 1995, private

1330: 				  communication

1331: 

1332: \bibitem[Yamada \& Sato(1994)]{yam94} Yamada, S., \& Sato, K. 1994,

1333: 			       \apj, 434, 268 

1334:  

1335: \bibitem[Yamada \& Sawai(2004)]{yam04} Yamada, S., \& Sawai, H. 2004,

1336: 				  \apj, 608, 907

1337: 

1338: %\bibitem[Yamasaki \& Yamada(2005)]{yam05} Yamasaki, T., \& Yamada, S. 2005,

1339: 		%		  submitted to \apj, astro-ph/0412625

1340: 

1341: \bibitem[Zhang et al.(2007)]{zha07} Zhang, W., Woosley, S. E., \& Heger,

1342: 				  A. 2007, submitted to \apj, astro-ph/0701083

1343: 

1344: 

1345: \end{thebibliography}

1346: 

1347: \begin{figure}

1348: %\includegraphics[width=5cm]{bfield.eps}

1349: %\epsscale{0.1}

1350: \plotone{f1.eps}

1351:   \caption{The initial magnetic magnetic fields of models MR3 (left

1352:  panel) and SR10 (right panel). Gray colors show the strengths of

1353:  the poloidal magnetic fields and the vectors represent the directions of

1354:  the magnetic fields. The initial radius of the 1.5

1355:  M$_{\odot}$ core is 2000~km.} 

1356:  \label{fbfield}

1357: \end{figure}

1358: 

1359: \begin{figure}

1360: \epsscale{1}

1361: \plotone{f2.eps}

1362:   \caption{The time evolutions of the rotational, kinetic, poloidal magnetic,

1363:  and toroidal magnetic energy until the farthest shock reaches a

1364:  radius 1500~km. the left and right panels correspond to model MR0 and

1365:  MS3, respectively. ``Outward kinetic energy'' and ``inward kinetic

1366:  energy'' are defined as the kinetic energy of matters whose radial

1367:  velocity is positive and negative, respectively. } 

1368:  \label{fengy}

1369: \end{figure}

1370: 

1371: \begin{figure}

1372: \plotone{f3.eps}

1373:   \caption{The velocity fields on top of the density color contours (the

1374:  right half of each panel), and the contours of the ratio of the

1375:  magnetic pressure to matter pressure (the left half of each

1376:  panel) when the shock front

1377:  reaches a radius of about 1500~km. From left to right

1378:  and top to bottom, panels of model MR0, MR3, SR3,

1379:  SR10, MS3, and SS3 are displayed in sequence. In each panel, colors in

1380:  the region out of a radius of 1500~km are extrapolated.}  

1381:  \label{fvbeta}

1382: \end{figure}

1383: 

1384: %\begin{figure}

1385: %\epsscale{1}

1386: %\plotone{massej.eps}

1387: %  \caption{The time evolution of the mass ejection in the northern hemisphere

1388: % (solid) and the southern hemisphere (dashed) for model MR3 and that in the

1389: % northern hemisphere (or the southern hemisphere) for model MR0.} 

1390: % \label{fmass}

1391: %\end{figure}

1392: 

1393: \begin{figure}

1394: \epsscale{1}

1395: \plotone{f4.eps}

1396:   \caption{The time evolution of the total, poloidal, and toroidal magnetic

1397:  energy and the rotational energy. The energies of the northern and

1398:  southern hemisphere are drawn separately. The insets show the time evolution of the

1399:  central density. Figures of models MR3, SR3, and SR10 are put in order

1400:  from left to right and top to bottom.} 

1401:  \label{fengmag}

1402: \end{figure}

1403: 

1404: \begin{figure}

1405: \plotone{f5.eps}

1406:   \caption{Upper two panels: The density contours (the right half of

1407:  each panel), and the contours of the rotation period in millisecond

1408:  (the left half of each panel) at 240 ms from the beginning for models

1409:  MR3 (the left panel) and MS3 (the right panel). In the left half of

1410:  each panel, the white-colored and black-colored regions have rotation period

1411:  of more than 10 s and less than 10 ms, respectively. The vectors in the

1412:  right-half of each panel represent the velocity field. Lower two panels:

1413:  The contours of the poloidal field strength (the right half of

1414:  each panel), and the contours for the ratio of the toroidal to poloidal field

1415:  strength (the left half of each panel) at 240 ms from the beginning for models

1416:  MR3 (the left panel) and MS3 (the right panel). The vectors in the

1417:  right-half of each panel represent the directions of the poloidal

1418:  magnetic field.} 

1419:  \label{fmagrot}

1420: \end{figure}

1421: 

1422: \begin{figure}

1423: \plotone{f6.eps}

1424:   \caption{Time evolutions of the proto-magnetar velocities. The northward

1425:  velocity is taken to be positive.} 

1426:  \label{fvns}

1427: \end{figure}

1428: 

1429: 

1430: \begin{figure}

1431: \begin{center}

1432: \plotone{f7.eps}

1433: %\resizebox{14cm}{!}{\includegraphics{ans.eps}}

1434:   \caption{Time evolutions of the accelerations acting on the proto-magnetar

1435:  star. The magnetic acceleration divided into two parts, the

1436:  tension part and pressure part. From top

1437:  to bottom and left to right, panels of model MR3, SR3,

1438:  SR10, MS3, and SS3 are displayed in sequence. The northward

1439:  acceleration is taken to be positive.} 

1440:  \label{fans}

1441: \end{center}

1442: \end{figure}

1443: 

1444: \begin{figure}

1445: \begin{center}

1446: \plotone{f8.eps}

1447:   \caption{A schematic picture which shows the magnetic tension-dominant

1448:  area (strong gray colored). \mbox{\boldmath $B$} and $d$\mbox{\boldmath $S$}

1449:  is the magnetic field vector and surface vector, respectively. $\theta_1$ and

1450:  $theta_2$ are the angles from the

1451:  rotation axis to the magnetic field vector and surface vector, respectively,

1452:  taking clockwise rotation positive.}

1453:  \label{fmagtp}

1454: \end{center}

1455: \end{figure}

1456: 

1457: \begin{figure}

1458: \plotone{f9.eps}

1459:   \caption{Time evolutions of the proto-magnetar velocities for several

1460:  definitions of a proto-magnetar. Left and right panel corresponds to the

1461:  results 

1462:  of model series of MR3evnD and MR3envP, respectively. In each panel, solid,

1463:  dotted, short-dashed, long-dashed, and dotted-short-dashed, 

1464:  lines correspond to model of 1.2, 1.23, 1.26, 1.3, and 1.35~M$_\odot$

1465:  proto-magnetars, whereas dotted-long-dashed line is drawn for model MR3 for

1466:  comparison. The northward

1467:  velocity is taken to be positive.} 

1468:  \label{fvnsmr3env}

1469: \end{figure}

1470: 

1471: 

1472: \input{tab1.tex}

1473: \input{tab2.tex}

1474: \input{tab3.tex}

1475: 

1476: 

1477: \end{document}

1478: % LocalWords:  envP envD ller

1479: