1: \documentclass[12pt,preprint]{aastex}
2: \usepackage[monochrome]{color}
3: %About bibun
4: \newcommand{\del}[2]%
5: {\frac{\mathrm{d}{#2}}{\mathrm{d}{#1}}}
6: \newcommand{\Del}[2]%
7: {\frac{\mathrm{D}{#2}}{\mathrm{D}{#1}}}
8: \newcommand{\ddel}[2]%
9: {\frac{\mathrm{d}^2{#2}}{\mathrm{d}{#1}^2}}
10:
11: \newcommand{\pdel}[2]%
12: {\frac{\partial{#2}}{\partial{#1}}}
13: \newcommand{\pddel}[2]%
14: {\frac{\partial^2{#2}}{\partial{#1}^2}}
15: \newcommand{\udel}[1]{\partial_{#1}}
16: \newcommand{\tdel}[1]{\partial^{#1}}
17:
18: \newcommand{\grad}{\nabla}
19: \renewcommand{\div}{\nabla\cdot}
20: \newcommand{\rot}{\nabla\times}
21: \newcommand{\laplace}{\bigtriangleup}
22: \renewcommand{\vec}[1]{\mathbf{#1}}
23:
24: \newcommand{\Ms}{M_{\odot}}
25: \newcommand{\km}{\mathrm{km}}
26: \newcommand{\cm}{\mathrm{cm}}
27: \newcommand{\ms}{\ \mathrm{ms}}
28: \newcommand{\cmps}{\mathrm{cm/s}}
29: \newcommand{\MRI}{\mathrm{MRI}}
30: \newcommand{\WRAP}{\mathrm{WRAP}}
31: \newcommand{\gauss}{\mathrm{G}}
32: \newcommand{\erg}{\mathrm{erg}}
33: \newcommand{\psec}{\mathrm{s}^{-1}}
34: \newcommand{\Nabog}{N_{\mathrm{A}}}
35: \newcommand{\gpcmc}{\mathrm{g/cm^3}}
36: \newcommand{\MeV}{\ \mathrm{MeV}}
37: \newcommand{\ergps}{\ \mathrm{erg/s}}
38: \newcommand{\ergpcmc}{\ \mathrm{erg/cm^3}}
39:
40: \def\Alfven{{Alfv\'{e}n}~} % ??
41: %\newcommand{\section}[1]{{\Large #1}\\}
42: %teisu
43: %\def\i{\mathrm{i}}
44: %\def\e{\mathrm{e}}
45: \def\d{\rm{d}}
46:
47: \newcommand{\myemail}{takiwaki@utap.phys.s.u-tokyo.ac.jp}
48:
49: \slugcomment{Accepted to the ApJ}
50: \shorttitle{SRMHD Simulations of Magnetically-dominated Jets}
51: \shortauthors{Takiwaki et al.}
52:
53: \begin{document}
54:
55: \title{Special Relativistic Simulations of Magnetically-dominated Jets
56: in Collapsing Massive Stars}
57:
58: \author{Tomoya Takiwaki\altaffilmark{1},
59: Kei Kotake\altaffilmark{2,3},
60: and Katsuhiko Sato\altaffilmark{1,4,5}}
61:
62: \affil{\altaffilmark{1}Department of Physics,
63: School of Science, the University of Tokyo, 7-3-1 Hongo,
64: Bunkyo-ku,Tokyo 113-0033, Japan}
65: \email{takiwaki@utap.phys.s.u-tokyo.ac.jp}
66: \affil{$^2$Division of Theoretical Astronomy, National Astronomical Observatory of Japan, 2-21-1, Osawa, Mitaka, Tokyo, 181-8588, Japan}
67: \affil{$^3$Max-Planck-Institut f\"{u}r Astrophysik, Karl-Schwarzshild
68: -Str. 1, D-85741, Garching, Germany}
69: \email{kkotake@th.nao.ac.jp,kotake@MPA-Garching.MPG.DE}
70: \affil{\altaffilmark{4}
71: Research Center for the Early Universe,
72: School of Science, the University of Tokyo,7-3-1 Hongo,
73: Bunkyo-ku, Tokyo 113-0033, Japan}
74: \affil{\altaffilmark{5}
75: The Institute for the Physics and Mathematics of the Universe, The
76: University of Tokyo,
77: Kashiwa, Chiba, 277-8568 Japan}
78: \begin{abstract}
79: We perform a series of two-dimensional magnetohydrodynamic
80: core-collapse simulations of rapidly rotating and strongly magnetized massive stars.
81: %the rotational core-collapse of magnetized massive stars.
82: %We systematically investigate how the strong magnetic
83: %fields and the rapid rotation affect properties of the magnetic explosions.
84: To study the properties of magnetic explosions for a longer time stretch
85: of postbounce evolution,
86: we develop a new code under the framework of special relativity including
87: a realistic equation of state with a multiflavor neutrino leakage scheme.
88: %achieve a longer stretch of the hydrodynamic evolution.
89: %As for the microphysics, we employ a realistic equation of state
90: %and approximate the neutrino
91: %transport by the multiflavor leakage scheme,
92: %We found two types of jets.
93: %Depending on the initial rotation rates and field strength,
94: Our results show the generation of the
95: magnetically-dominated jets in the two ways.
96: One is launched just after the core-bounce in a prompt way and another
97: is launched at $ \sim 100 $ ms after the stall of the prompt shock.
98: %This difference on the duration reflects how fast the magnetic field is
99: % amplified.
100: We find that the shock-revival occurs when the
101: magnetic pressure becomes strong, due to the field wrapping, enough
102: to overwhelm the ram pressure of the accreting matter. The critical
103: toroidal magnetic fields for the magnetic shock-revival
104: are found to be universal of $\sim 10^{15}\mathrm{G}$ behind the jets.
105: We point out that the time difference before the
106: % magnetic
107: shock-revival has a strong correlation with
108: the explosions energies.
109: %Since the jet propagations are suppressed by the accreting matter from the outer core,
110: %the required magnetic fields for the magnetic explosions are
111: %determined by the ram pressure of the accreting matter.
112: %Reflecting similar accretion rate among the models, toroidal magnetic field
113: % near the pole have similar value of $10^{15}\mathrm{G}$.
114: %Therefore proto-magnetars are generated in the center.
115: %Rapidly rotating protoneutron stars endowed with the high fields
116: Our results suggest that the magnetically dominated jets are accompanied by the
117: formation of the magnetars.
118: %We speculate that rapidly rotating protoneutron
119: %stars with the high magnetic fields may be the newly-born magnetars.
120: %We speculate that the mildly relativistic jets might be the origin of
121: %the observed X-ray flashes in gamma-ray bursts.
122: Since the jets are mildly relativistic, we speculate that they might be the origin of
123: some observed X-ray flashes.
124: %We speculate that the collimated
125: % jets are associated with the formation of the magnetars.
126: %The magnetically dominated jets might be related to the observed X-ray flashes
127: % in gamma-ray bursts.
128: %Our simulations suggest that the collimated magnetically dominated
129: % jets are associated with the formation of the magnetars.
130: % This study will be the foundation for understanding the relation between the formation of the magnetars and the X-Ray flashes observed in gamma-ray bursts.
131: \end{abstract}
132:
133: \keywords{supernovae: collapse, rotation ---
134: magnetars: pulsars, magnetic field ---
135: methods: numerical ---
136: MHD --- special relativity ---
137: gamma rays: bursts}
138:
139: \section{Introduction}
140: There have been growing evidences shedding lights on the relations
141: between the high energy astrophysical phenomena and their origins.
142: A number of host galaxies of long-duration gamma-ray bursts (GRBs) are
143: recently identified as metal poor galaxies whose metalicities are
144: lower than that of average massive star-forming galaxies \citep[]
145: [and reference therein]{sava06,stan06}.
146: The preponderance of short-lived massive star formation in such
147: young galaxies, as well as the identification of SN Ib/c light curves
148: peaking after the bursts in a few cases, has provided strong support for
149: a massive stellar collapse origin of the long GRBs \citep{pacz98,gala98,stan03}.
150: The duration of the long GRBs
151: %($2\mathrm{s}<t_{\gamma}< 10^3\mathrm{s}$)
152: may correspond to the accretion of debris falling into the central black
153: hole (BH)\citep{piro98}, which suggests the observational consequence
154: of the BH formation likewise the supernova of neutron star formation.
155: There is also a growing observational evidence of supermagnetized
156: neutron stars with the magnetic fields of $\sim 10^{14}-10^{15}$ G,
157: the so-called magnetars (\citet{dunc92}, see \citet{latt07} for
158: a recent review).
159: The magnetic fields are determined by the measured period and derivative
160: of period under the assumption that the spin-down is caused due to the usual
161: magnetic dipole radiation \citep{zhan00,hard06}.
162: Tentative detections of spectral features during the burst phase also
163: indicate $B\sim 10^{15}\mathrm{G}$ when interpreted as proton cyclotron
164: lines \citep{gavr02,ibra03,rea03}.
165: Recently X-ray flash (XRF), which is a low energy analogue of the GRB, is
166: receiving great attentions as a possible
167: relevance to the magnetar formations \citep{maza06,toma07}.
168: A large amount of neutron rich Ni ejected by
169: SN2006aj associated with XRF060218 is interpreted to be the formation of such
170: objects, not the black hole after the explosion \citep{maed07a}.
171:
172: So far a number of numerical simulations have been done towards the
173: understanding of the formation mechanisms of these compact
174: objects such as neutron stars, magnetars, and the black holes in
175: combination with their possible consequences like GRBs and XRFs.
176: The leading model for the long-duration GRBs is the
177: collapsar model \citep{macf99}. In the model, the core of massive
178: stars with significant angular momentum collapses into a black hole (BH).
179: The neutrinos emitted from the rotation-supported accretion disk
180: around the BH heat the matter of the funnel region of the disk, to
181: launch the GRB outflows. The relativistic flows are expected to ultimately form a
182: fireball, which is
183: good for the explanation of the observed afterglow (e.g., \citet{pira99}).
184: In addition, it is suggested that the strong magnetic fields in the
185: cores of order of $10^{15} \gauss$ play also an active role both for driving
186: the magneto-driven jets and for extracting a significant amount of
187: energy from the central engine (e.g.,
188: \cite{usov92,whee00,thom04,uzde07a} and see references therein).
189:
190: In order to understand such scenarios,
191: the ultimate necessity of the stellar core-collapse simulations
192: is to perform the simulations tracing all the phases
193: in a consistent manner starting from the stellar core-collapse,
194: core-bounce, shock-stall, stellar explosion (phase 1) or BH formation
195: and the formation of accretion disk (phase 2), energy
196: deposition to the funnel region by neutrinos and/or magnetic fields
197: (phase 3), to the launching of the fireballs (phase 4).
198: Here for convenience we call each stage as phase 1, 2, etc.
199: The requirement for the numerical modeling to this end is highly
200: computationally expensive, namely the multidimensional MHD
201: simulations not only with general relativity for handling the BH
202: formation, but also with the multi-angle neutrino transfer for
203: treating highly anisotropic neutrino radiation from the disks.
204: So various approximative approaches to each phase
205: have been undertaken.
206: As we mention below, these studies are complimentary in the sense that
207: the different epochs are focused on, with the different initial
208: conditions for the numerical modeling being taken.
209: %One of the difficulties on these studies is that the GRBs and XRFs are
210: %generally considered to be jet-like to circumvent the baryon loading
211: %problem \citep{piran_rev}.
212: %To clarify the mystery on the central activity,
213: %we must focus on the aspherical effect of the gravitational collapse,
214: %such as rotation and magnetic fields.
215: %There are various approaches to investigate the central engine of these
216: %phenomena
217:
218: In addition to the elaborate studies in the conventional supernova context
219: (see recent reviews for \citet{kota06,jank07}), much attention has
220: been paid recently to the roles of rapid rotation and magnetic fields for studying
221: the formation of magnetars and its possible application to
222: the collapsars {\color{red}
223: \citep{yama04,taki04,kota04a,sawa05,matt06,mois06,ober06a,nish06,burr07,cerd07,scei07,komi07b}.
224: }% komi07b is added for the referr report
225: % In the simulations of this phase 1, much emphasis is mainly placed on the
226: %sophistication of the microphysics such as a realistic equation of
227: %nuclear matters and the neutrino transport.
228: After the failed or weak explosion,
229: % \citep{heger},
230: the accretion to the central objects may lead to the formation of a BH (phase 2).
231: Several general relativistic studies are on the line for the understanding
232: of the hydrodynamics at the epoch of the BH formation, in which
233: more massive progenitors ($> \sim 25 M_{\odot}$) than those of the study
234: in the phase 1 are generally employed \citep{shib06,seki07}.
235: Treating the BH as an absorbing boundary or using the fixed metric
236: approaches, the numerical studies of
237: the phase 3 are concerned with the initiation of the outflows from the funnel
238: region of the disk to the acceleration of the jets as a result of the
239: neutrino heating and/or MHD processes till the jets become mildly relativistic{\color{red}
240: \citep{koid98,macf99,prog03,nisi05,devi05,krol05,hawl06,mizu06,fuji06,uzde06,mcki07b,komi07a,naga07,suwa07a,suwa07b,bark08}.% uzde06 \& bark08 are added for the referr report
241: }Numerical studies of the phase 4 are
242: mainly concerned with the dynamics later on, namely, the jet
243: propagation to the breakout from the star, when the acceleration of
244: the jets to the high Lorentz factor is expected{\color{red}
245: \citep{ston00,aloy00,zhan03,leis05,mizt06,mcki06,mizu07}
246: }.% mcki06 is added for the referr report
247:
248: Our previous study was devoted to the phase 1, in which
249: we performed a series of 2D core-collapse
250: simulations of rotating and magnetized massive stars under the framework of
251: the Newtonian magnetohydrodynamics \citep{taki04}. We found that the magneto-driven
252: jet-like shocks were launched from the protoneutron
253: stars just after core-bounce. However at the moment, we were unable to
254: follow the dynamics much later on until when the collimated jets
255: reach further out from the center.
256: The Alfv\'{e}n velocity of the jet
257: propagating into the outer layer of the iron core can be estimated by
258: the following simple order-of-magnitude estimation,
259: \begin{equation}
260: v_{A} = \frac{B}{\sqrt{4 \pi \rho}} \sim 10^{10}\mathrm{cm/s}\frac{B/{10^{13}\mathrm{G}}}{\sqrt{\rho/\left(10^5\mathrm{g/cm^3}\right)}},
261: \end{equation}
262: with $\rho$ and $B$ being the typical density and magnetic field
263: there. It can be readily inferred that the Alfv\'{e}n velocity
264: can exceed the speed of light unphysically in the Newtonian simulation.
265: To avoid this problem we construct a new code under the framework of
266: special relativistic magnetohydrodynamics.
267: %In this paper we challenge similar initial setup again
268: %and investigate how the initial strengths of the rotation and
269: %the magnetic fields affect the explosion.
270: We take a wider parametric range for the strength of the rotation than
271: that of our previous work. By so doing, we hope to study
272: more systematically than before how the strong magnetic
273: fields and the rapid rotation affect the properties of the magnetic
274: explosions.
275:
276: We summarize the numerical methods in section \ref{sec:NM}.
277: Section 3 is devoted to the initial models.
278: In section 4, we show the numerical results.
279: In section 5, we summarize our study and discuss the implications of
280: our model for the magnetars and the X-ray flashes.
281: Details of the numerical scheme and the code tests are given in the appendix.
282:
283: \section{Numerical Methods}\label{sec:NM}
284: The results presented in this paper are calculated by the newly
285: developed special relativistic magnetohydrodynamic (SRMHD) code.
286: The novel point of this code is that the detailed microphysical
287: processes relevant for the stellar-core-collapse simulations
288: are also coupled to the magneto-hydrodynamics (MHD).
289: We briefly summarize the numerical methods in the following.
290:
291: The MHD part of the code is based on the formalism of \citet{devi03}.
292: %The definition of the primary variables is as follows.
293: Before going to the basic equations, we write down the definition of the
294: primary code variables.
295: The state of the relativistic fluid element at each point in the space time
296: is described by its density, $\rho$; specific energy, $e$; velocity,
297: $v^i$; and pressure, $p$.
298: And the magnetic field in the rest {\color{red}frame} of the fluid is described by
299: the 4-vector $\sqrt{4\pi}b^{\mu}={^*F}^{\mu\nu}U_{\nu}$, where $^*F^{\mu\nu}$ is the
300: dual of the electro-magnetic field strength tensor and $U_{\nu}$ is the
301: 4-velocity.
302:
303: After some mathematical procedures presented in Appendix \ref{app:DeBE},
304: the basic equations of SRMHD are described as follows:
305: \begin{eqnarray}
306: \pdel{t}{D}
307: +\frac{1}{\sqrt{\gamma}}\udel{i}{\sqrt{\gamma}Dv^i} &=&0 \label{eq:mass_consv}\\
308: \pdel{t}{E}
309: +\frac{1}{\sqrt{\gamma}}\udel{i}{\sqrt{\gamma}Ev^i}
310: &=&-p\pdel{t}{W}
311: -\frac{p}{\sqrt{\gamma}}\udel{i}{\sqrt{\gamma}W v^i} - L_{\nu}\label{eq:ene_consv}\\
312: \pdel{t}{S_i-b^tb_i}
313: +\frac{1}{\sqrt{\gamma}}
314: \udel{j}{\sqrt{\gamma}\left(S_i v^j-b_ib^j\right)}
315: &=&
316: -\frac{1}{2}
317: \left(
318: \rho h \left(Wv_k\right)^2
319: - \left(b_k\right)^2
320: \right)\udel{i}{\gamma^{kk}}\nonumber\\
321: & &
322: -\left(\rho h W^2- {b^t}^2\right) \udel{i}{\Phi}\nonumber\\
323: & &
324: -\udel{i}{\left(p+\frac{\|b\|^2}{2}\right)}\label{eq:mom_consv}\\
325: \pdel{t}{B^i}
326: +\udel{j}{\left(Wv^jb^i-Wv^ib^j\right)}
327: &=&0\label{eq:induction}\\
328: \tdel{k}{{\udel{k}{\Phi}}}&=&\rho hW^2
329: -\left(p+\frac{\left|b\right|^2}{2}\right)
330: -\left(b^t\right)^2\label{eq:poisson}
331: \end{eqnarray}
332: where $W=\frac{1}{\sqrt{1-v^kv_k}}$, $D=\rho W$, $E=e W$ and
333: $S_i=\rho hW^2v_i$ are
334: the Lorentz boost factor, auxiliary variables correspond to density,
335: energy, and momentum, respectively.
336: Eq. (\ref{eq:mass_consv}), Eq. (\ref{eq:ene_consv}) and Eq. (\ref{eq:mom_consv})
337: represents the mass, energy, and momentum conservations.
338: $L_{\nu}$ in the right hand side of Eq. (\ref{eq:ene_consv})
339: is a total neutrino cooling rate determined by microphysical
340: processes which will be later explained.
341: In Eq. (\ref{eq:mom_consv}) it is noted that the relativistic enthalpy,
342: $h=(1+e/\rho+p/\rho+\left|b\right|^2/\rho)$
343: includes magnetic energy.
344: Eq. (\ref{eq:induction}) is the induction equation for the magnetic
345: fields. $B^i$ are related to that in the rest frame of fluid
346: as $B^i=Wb^i -Wb_t b^i$.
347: Here $b_t$ is a time component of the 4-vector, $b_{\mu}$.
348: Eq. (\ref{eq:poisson}) is the Poisson equation for the gravitational
349: potential, $\Phi$.
350:
351: This newly developed code is an Eulerian code based on the finite-difference
352: method.
353: The numerical approach for solving the basic equations of
354: (\ref{eq:mass_consv}), (\ref{eq:ene_consv}), and
355: (\ref{eq:mom_consv}), consists of the two steps, namely,
356: the transport and the source step.
357: These procedures are essentially the same as those of ZEUS-2D \citep{ston92}.
358: At the transport step, the second order upwind scheme of Van Leer is implemented \citep{vlee77}.
359: To handle the numerical oscillations, we employ an artificial viscosity.
360: In the special relativistic treatments, many forms for the compression
361: heating are possible \citep{hawl84b}.
362: In our code, we employ the form of
363: $\frac{\rho h}{\sqrt{\gamma}}\udel{i}{\sqrt{\gamma}Wv^{i}}$ as the
364: compression heating, which becomes the well-known artificial viscosity of
365: von Neumann and Richtmyer under the Newtonian approximation.
366: While not explicitly included in the above expression for the enthalpy, the
367: contribution from the compression heating on the inertia is included in our calculations.
368: The detailed status on the shock capturing using this term is shown at
369: Appendix \ref{TP}.
370:
371: The time evolution of the magnetic fields is
372: solved by induction equation, Eq. (\ref{eq:induction}).
373: In so doing, the code utilizes the so-called
374: constrained transport method, which ensures
375: the divergence free ($\nabla\cdot\vec{B}=0$)
376: of the numerically evolved magnetic fields at all times.
377: Furthermore, the method of characteristics (MOC)
378: is implemented to propagate accurately all modes of MHD waves.
379: The detailed explanation and the numerical tests are delivered in the
380: appendix \ref{ap:AWP}.
381: The self-gravity is managed by solving the Poisson equation, Eq. (\ref{eq:poisson})
382: with the incomplete Cholesky decomposition conjugate gradient method.
383:
384: Together with these hydrodynamic procedures,
385: the following microphysical processes are implemented in this code.
386: We approximate the neutrino transport by a multiflavor leakage
387: scheme \citep{epst81,ross03}, in which three neutrino flavors:
388: electron neutrino, $\nu_{e}$; electron antineutrino, $\bar{\nu}_{e}$; and
389: the heavy-lepton neutrinos, $\nu_{\mu}$, $\bar{\nu}_{\mu}$,
390: $\nu_{\tau}$, $\bar{\nu}_{\tau}$ (collectively referred to
391: as $\nu_{X}$), are taken into account. The
392: neutrino reactions included are electron capture on
393: proton and free nuclei; positron capture on neutron; photo-, pair, plasma
394: processes \citep{full85,taka78,itoh89,itoh90}.
395: We added a transport equation for the lepton fraction $Y_l (= Y_e -
396: Y_{e^{+}} + Y_{\nu_e} - Y_{\bar{\nu}_e})$,
397: \begin{equation}
398: \pdel{t}{Y_l}
399: +\frac{1}{\sqrt{\gamma}}\udel{i}{\sqrt{\gamma}Y_l} = - \gamma_l
400: \end{equation}
401: to treat their
402: change due to the relevant charged current reactions, whose reaction
403: rates are collectively represented by $\gamma_l$ here, with $Y_e,
404: Y_{e^{+}},Y_{\nu_e},Y_{\bar{\nu}_e}$, $\gamma_l$ being electron, positron,
405: electron neutrino, anti-electron neutrino fraction, respectively
406: (see \citet{epst81,ross03,kota03a} for details of the estimation of $\gamma_l$).
407: $L_{\nu}$ in Eq. (\ref{eq:ene_consv}) represents the total
408: neutrino cooling rate which is also estimated by the scheme.
409: As for the equation of state (EOS), we employ a realistic one based on
410: the relativistic mean field theory \citep{shen98}.
411: Since the pressure is not represented as the analytic function of
412: density and internal energy like in the case of polytropic EOS,
413: an iterative algorithm are employed to update the fundamental variables
414: (see Appendix \ref{CVFV} for detail).
415:
416: %Together with these hydrodynamic procedures,
417: %the following microphysical processes are implemented in this code.
418: %we employ a realistic equation of state based on
419: %relativistic mean field theory \citep{shen98} (see \citealt{kota03} for
420: % the implementation) and treat the neutrino
421: %cooling by a multiflavor leakage scheme, in which 6 species of neutrinos
422: % with pair, photo, and plasma processes by \cite{itoh89} in addition
423: %to the standard charged current neutrino cooling reactions
424: % are included.
425: %First we added an transport equation for the lepton fractions to
426: %treat the standard charged current neutrino cooling reactions.
427: %fraction and neutrinos fractions,
428: %to treat neutrino reaction and neutrino transport by the
429: %so-called leakage scheme \citep{epst81}.
430: %We take into account three neutrino flavors:
431: %electron neutrino, $\nu_{e}$; electron antineutrino, $\bar{\nu}_{e}$; and
432: %the heavy-lepton neutrinos, $\nu_{\mu}$, $\bar{\nu}_{\mu}$,
433: %$\nu_{\tau}$, $\bar{\nu}_{\tau}$, which are collectively referred to as $\nu_{X}$.
434: %The cooling rate, $L_{\nu}$, is also estimated by the scheme.
435: %The neutrino reactions included in this code are electron capture on
436: %proton and free nuclei; positron capture on neutron; photo-, pair, plasma
437: %processes \citep{full85,taka78,itoh89,itoh90}.
438: %Second we have incorporated the tabulated equation of state (EOS)
439: %based on relativistic mean field theory \citep{shen98}.
440: %Since the pressure is not represented as the analytic function of
441: %density and internal energy,
442: %an iterative algorithm are employed to update the fundamental variables.
443: %Please see Appendix \ref{CVFV} for detail.
444:
445: In our two dimensional simulations, the spherical coordinate is
446: used with 300($r$) $\times$ 60($\theta$) grid points to cover the
447: computational domain. Axial symmetry and reflection symmetry across
448: the equatorial plane are assumed.
449: The radial grid is nonuniform,
450: extending from $0$ to $4.0\times 10^8 \cm$ with finer
451: grid near the center. The finest grid is set to $10^5 \cm$.
452: The polar grid uniformly covers from $\theta=0$
453: to $\theta=\frac{\pi}{2}$. {\color{red} This choice of the grid numbers is
454: sufficient for the aim of this paper as will be discussed in section 5.}
455:
456:
457: Finally we summarize the difference on the numerical approach from our
458: previous work \citep{taki04}.
459: Most major development is the fully special
460: relativistic treatment on magneto-hydrodynamics.
461: And for the microphysical parts
462: the cooling terms by neutrino contains contributions from not only
463: $\nu_{\mathrm{e}}$ but also $\bar{\nu_{\mathrm{e}}}$ and
464: $\nu_{\mathrm{X}}$.
465: These advances provide more reliable results on the
466: magneto-rotational core-collapse.
467:
468: \section{Initial Models}\label{sec:NGIM}
469:
470: We make precollapse models by taking the profiles of density,
471: internal energy, and electron fraction distribution from
472: a rotating presupernova model of E25 by \citet{hege00}.
473: This model has mass of $25M_{\odot}$ at the zero age main
474: sequence, however loses the hydrogen envelope and becomes a Wolf Rayet
475: star of 5.45 $M_\odot$ before core-collapse.
476: Our computational domain involves the whole iron-core of $1.69 M_{\odot}$.
477: It is noted that this model seems to be a good candidate as a progenitor of the GRB since
478: the lack of the line spectra of the ejected envelopes are reconciled
479: with the observations of the supernovae associated with GRBs (e.g., \citet{mesz06}).
480: %the line spectra of the ejected envelopes is not found in many
481: %supernovae that accompany GRBs (e.g., \citet{mesz06}).
482:
483:
484: Since little is known about the spatial distributions of the rotation and
485: the magnetic fields in the evolved massive stars (see, however,
486: \citet{spru02}), we add the following
487: rotation and magnetic field profiles
488: in a parametric manner to the non-rotating core mentioned above.
489: For the rotation profile, we assume a cylindrical rotation of
490: \begin{equation}
491: \Omega(X,Z)=\Omega_{0}\frac{X_{0}^2}{X^2+X_{0}^2}\frac{Z_{0}^4}{Z^4+Z_{0}^4}\label{eq:CR},
492: \end{equation}
493: where $\Omega$ is the angular velocity and X and Z denote distance
494: from the rotational axis and the equatorial plane.
495: We adopt values of the parameters, $X_0$ and $Z_0$, as $10^7 \cm,10^8
496: \cm$, respectively.
497: The parameter, $X_{0}$ represents the degree of differential rotation.
498: %The value of $X_0$ taken here corresponds $1/20$ of the radius
499: We assume the strong differential rotation as
500: in our previous study \citep{taki04}.
501:
502: As for the initial configuration of the magnetic fields,
503: we assume that
504: the field is nearly uniform and parallel to the rotational axis in the core
505: and dipolar outside.
506: For the purpose, we consider the following effective vector potential,
507: \begin{equation}
508: A_r=A_\theta=0,
509: \end{equation}
510: \begin{equation}
511: A_\phi=\frac{B_0}{2}\frac{r_0^3}{r^3+r_0^3}r\sin\theta,\label{vec_phi}
512: \end{equation}
513: where $A_{r,\theta,\phi}$ is the vector potential in the $r,\theta,\phi$
514: direction, respectively, $r$ is the radius, $r_0$ is the radius of the core, and
515: $B_0$ is the model constant.
516: In this study, we adopt the value of $r_0$ as $2\times10^8$ cm which
517: is approximately the size of the iron core at a precollapse stage.
518: This vector potential can produce the uniform magnetic fields when $r$
519: is small compared with $r_0$, and the dipole magnetic fields for vice versa.
520: Since the outer boundary is superposed at $r=4\times10^8~{\rm cm}$, the magnetic
521: fields are almost uniform in the computational domain
522: as the previous work \citep{taki04}.
523: It is noted that this is a far better way than the loop current method
524: for constructing the dipole magnetic fields \citep{symb84},
525: because our method produces no divergence of the magnetic fields
526: near the loop current.
527: We set the outflow boundary conditions for the magnetic fields at the
528: outer boundary of the calculated regions.
529:
530: %For the magnetic fields we assume vector potentials of this form:
531: %\begin{eqnarray}
532: % A_{r}&=&A_{\theta}=0. \label{eq:MF1}\\
533: % A_{\phi}&=&\frac{B_{0}}{2}\frac{r_{0}^3}{r^3+r_{0}^3}r \sin\theta, \label{eq:MF2}
534: %\end{eqnarray}
535: % where $r_{0}=2\times 10^8 \cm$ that is approximately the radius of the iron core.
536: %The magnetic fields are given by $\vec{B}=\rot{\vec{A}}$,
537: %automatically satisfying the condition of $\div{\vec{B}}=0$.
538: %This vector potentials reproduce the uniform magnetic fields parallel to
539: %the rotational axis near the center
540: %and the dipole magnetic fields far outside the iron core surface.
541:
542:
543: %We prefer parametric approach to investigate various possibility.
544: We compute 9 models changing
545: the total angular momentum and the strength of magnetic fields
546: by varying the value of $\Omega_0$ and $B_{0}$.
547: The model parameters are shown in Table \ref{tab:model}.
548: The models are named
549: after this combination,
550: with the first letters, B12, B11, B10,
551: representing strength of the initial magnetic field,
552: the following letter, TW4.0, TW1.0, TW0.25
553: representing the initial $T/|W|$, respectively.
554: Here $T/|W|$ indicates the ratio of
555: the rotational energy to the absolute value of the gravitational energy.
556: %The total rotational energy, $T$, is given by the integration of the local
557: %rotational energy determined by the matter configuration and Eq. (\ref{eq:CR}).
558: The corresponding values of $\Omega_{0}$ are
559: $151 \mathrm{rad/s}$, $76 \mathrm{rad/s}$, $38 \mathrm{rad/s}$
560: for TW4.0, TW1.0, TW0.25, respectively.
561: %These values of the rotational energy here are relatively large compared to
562: It is noted that the value of $T/|W|$ is $0.15\%$ of the progenitor by
563: \citet{hege00} and also that the specific angular momenta ranges from
564: $0.5$ to $1.5$ $j_{16}$ for TW0.25 to TW4.0 models
565: with $j_{16}\equiv 10^{16}~\rm{cm}^2~{\rm s}^{-1}$,
566: which are in good agreement with the requirement of the collapsar model \citep{macf99}.
567: % By exploring to the models with more rapid rotation,
568: %we here hope to study the magnetorotational effects on the dynamics systematically.
569: %Therefore this models correspond to rapidly rotating progenitors.
570: %These values of the rotational energy here seem relatively large compared to
571: %that of the original data of Heger whose
572: %$T/|W|$ is $0.15\%$ \citep{hege00}.
573: %Therefore this models correspond to rapidly rotating progenitors.
574: {\color{red}
575: Current stellar evolution calculations predict that the rapidly rotating
576: massive stars with smaller metalicity experiences the so-called chemically homogeneous
577: cores during its evolution \citep{yoon}.
578: Such stars are considered to satisfy the requirements of the collapsar model,
579: namely rapid rotation of the core \citep{woos06}.
580: According to a GRB progenitor model of 35OB in \citet{woos06},
581: the magnetic field strength of the core reaches to $\sim 10^{12}$ G and
582: the specific angular momentum is the order of $j_{16} \sim 1$), with which our choices
583: for the initial magnetic field and the initial rotation rate are reconciled.}
584:
585: \begin{table}[h]
586: \begin{center}
587: \caption{Models and Parameters }\label{tab:model}
588: \begin{tabular}{cc|ccc}
589: \tableline\tableline
590: & & & {$T/|W|(\%)$}& \\
591: & & 0.25\% & 1.0\% &4.0\%\\
592: \tableline
593: &$10^{10}\gauss$ & B10TW0.25 & B10TW1.0 & B10TW4.0 \\
594: $B_{0}(\mathrm{Gauss})$ & $10^{11}\gauss$ & B11TW0.25 & B10TW1.0 & B11TW4.0 \\
595: &$10^{12}\gauss$ & B12TW0.25 & B10TW1.0 & B12TW4.0 \\
596: \tableline
597: \end{tabular}
598: \tablecomments{
599: Model names are labeled by the initial strength of magnetic fields and rotation.
600: $T/|W|$ represents the ratio of the rotational energy to the absolute value of
601: the gravitational energy.
602: The corresponding values of $\Omega_{0}$ in Eq. (\ref{eq:CR}) are
603: $151 \mathrm{rad/s}$, $76 \mathrm{rad/s}$, $38 \mathrm{rad/s}$
604: for TW4.0, TW1.0, TW0.25, respectively.
605: $B_{0}$ represents the strength of the poloidal magnetic fields (see Eq.(\ref{vec_phi})).
606: The corresponding values of $E_{m}/|W|$ is $2.5\times 10^{-8}$,
607: $2.5\times 10^{-6}$ and $2.5\times 10^{-4}$ for $10^{10}\gauss$,
608: $10^{11}\gauss$ and $10^{12}\gauss$, respectively with $E_{m}$ being
609: the magnetic energy.
610: }
611: \end{center}
612: \end{table}
613:
614: \clearpage
615:
616: \section{Results}
617: %In this section we present numerical results calculated by the newly
618: %developed SRMHD code.
619: %As stated in section \ref{sec:NGIM}
620: %we performed 9 computations by varying the strengths of the magnetic field and the
621: %rotation.
622: \subsection{Hydrodynamics before Core-Bounce}
623: First of all, we briefly mention the dynamics before core bounce,
624: when the gross features are rather similar among the computed models.
625: The characteristic properties are summarized in Table \ref{tab:PROCPS}.
626:
627: The story before core-bounce is almost the same as the canonical
628: core-collapse supernovae with rapid rotation (see, e.g.,
629: \citet{kota03a}). The core begins to collapse
630: due to electron captures and the photodissociation of the iron nuclei,
631: and eventually experiences the bounce at the subnuclear density
632: by the additional support of the centrifugal forces.
633: In fact, the central densities at bounce becomes smaller and the epoch
634: till bounce is delayed as the initial
635: rotation rates become larger (see $\rho_{\rm bnc}$ and $t_{\rm bnc}$
636: in Table \ref{tab:PROCPS}).
637:
638: As the compression proceeds,
639: the rotational energy increases and reaches a few $10^{52}
640: \mathrm{erg}$ at the moment of the bounce
641: (seen from $T/|W|_{\rm bnc} \times W_{\rm bnc}$ in Table
642: \ref{tab:PROCPS}).
643: Given the same initial rotation rates, the values of $T/|W|_{\rm bnc}$
644: do not depend on the initial field strength so much. This means that the
645: angular momentum transfer is negligible before bounce, which is also
646: the case of the Newtonian hydrodynamics \citep{yama04}.
647: At bounce, the unshocked core becomes more flattened
648: as the initial rotation rate becomes larger (compare panels in Figure \ref{fig:ds_bnc}).
649: \textcolor{red}{The central protoneutron stars rotate very rapidly reaching to $\sim
650: 3000$ rad/s with the typical surface magnetic fields of $\sim 10^{13}$
651: G to $\sim 10^{15}$ G for B10 and B12 models, respectively.}
652: From the table, it is also seen that the amplification rates of the magnetic
653: fields ($A_{\mathrm{amp}}$)
654: are mainly determined by the initial rotational rates.
655: One exception is the model B12TW4.0. Due to very rapid rotation with the highest
656: magnetic fields initially imposed, the model bounces predominantly due
657: to
658: the magnetic force. As a result, the core bounce occurs
659: earlier with the lower central density with less gravitational
660: energy of the inner core than the models with the same initial rotation rate
661: (see Table \ref{tab:PROCPS}). This earlier magnetically-supported bounce
662: leads to the suppression of the amplification rate, which is exceptionally
663: observed for this model.
664: %the magnetic field is not so important
665: %for the field amplification.
666: %%Only the rotation affects the dynamics and field amplification.
667: %In fact, table \ref{tab:PROCPS} shows that
668: %the amplification rates of magnetic energy, $A_{\mathrm{amp}}$, are not
669: %so different among the models with the same initial rotation rates.
670:
671: In this way, the hydrodynamic properties before bounce are
672: mainly governed by the differences of the initial rotation rates.
673: On the other hand, the differences of the magnetic field strength begin to play an important role on
674: the dynamics later on.
675: We will mention them in detail from the next sections.
676:
677: \begin{table}[h]
678: \caption{MHD properties till core bounce}\label{tab:PROCPS}
679: \begin{tabular}{cccccccccc}
680: \tableline\tableline
681: %\backslashbox
682: Model&
683: $T/|W|_{{\mathrm{bnc}}}$ &
684: $\rho_{\mathrm{bnc}}$ & $|W_{\mathrm{bnc}}|$ &
685: $E_m/|W|_{\mathrm{ini}}$ &
686: $E_m/|W|_{\mathrm{bnc}}$ &
687: $\frac{E_{p}}{E_m}$ &
688: $A_{\mathrm{amp}}$ & $t_{\mathrm{bnc}}$ \\
689: Names & & {\small $[10^{14}\mathrm{g/cm^3}]$ } & $[10^{53}\mathrm{erg}]$ &
690: & & & & $[\mathrm{ms}]$ \\
691: \tableline
692: B12TW0.25 &
693: $0.10$ &
694: $2.1$ & $1.1$ &
695: $2.5 \times 10^{-4}$ & $1.0 \times 10^{-3}$ &
696: $0.3$ &
697: $100$ & $245$ \\
698: B11TW0.25 &
699: $0.10$ &
700: $2.1$ & $1.1$ &
701: $2.5 \times 10^{-6}$ & $1.0 \times 10^{-5}$ &
702: $0.3$ &
703: $100$ & $245$ \\
704: B10TW0.25 &
705: $0.10$ &
706: $2.1$ & $1.1$ &
707: $2.5 \times 10^{-8}$ & $1.0 \times 10^{-7}$ &
708: $0.3$ &
709: $100$ & $245$ \\
710: \tableline
711: B12TW1.0 &
712: $0.18$ &
713: $1.3$ & $1.1$ &
714: $2.5 \times 10^{-4}$ & $9.0 \times 10^{-3}$ &
715: $0.07$ &
716: $720$ & $295$ \\
717: B11TW1.0 &
718: $0.18$ &
719: $1.3$ & $1.1$ &
720: $2.5 \times 10^{-6}$ & $7.0 \times 10^{-5}$ &
721: $0.07$ &
722: $610$ & $295$ \\
723: B10TW1.0 &
724: $0.18$ &
725: $1.3$ & $1.1$ &
726: $2.5 \times 10^{-8}$ & $7.0 \times 10^{-7}$ &
727: $0.07$ &
728: $610$ & $295$ \\
729: \tableline
730: B12TW4.0 &
731: $0.20$ &
732: $0.095$ & $0.68$ &
733: $2.5 \times 10^{-4}$ & $20 \times 10^{-3}$ &
734: $0.3$ &
735: $800$ & $477$ \\
736: B11TW4.0 &
737: $0.19$ &
738: $0.11$ & $0.74$ &
739: $2.5 \times 10^{-6}$ & $29 \times 10^{-5}$ &
740: $0.1$ &
741: $4400$ & $484$ \\
742: B10TW4.0 &
743: $0.19$ &
744: $0.11$ & $0.74$ &
745: $2.5 \times 10^{-8}$ & $31 \times 10^{-7}$ &
746: $0.1$ &
747: $4400$ & $484$ \\
748: \tableline
749: \end{tabular}
750: \tablecomments{
751: Characteristic properties before core bounce.
752: %The meaning of the column from left to right is as follows.
753: $T/|W|_{\mathrm{bnc}}$ is the rotational energy per gravitational energy
754: at bounce.
755: $\rho_{\mathrm{bnc}}$ is the maximum density at bounce.
756: % $W_{\mathrm{bnc}}$
757: % is the absolute value of the gravitational energy at bounce.
758: $E_m/|W||_{\mathrm{ini}}$ and $E_m/|W||_{\mathrm{bnc}}$
759: is the magnetic energy per the gravitational energy initially and at
760: bounce, respectively.
761: % is $5\times 10^{51}\mathrm{erg}$ in all models.
762: % is the magnetic energy per gravitational energy
763: % when the simulations start and is that at bounce.
764: $\frac{E_{p}}{E_m}$ is the ratio of the poloidal magnetic energy to the total
765: magnetic energy at bounce.
766: $A_{\mathrm{amp}}$ represents the amplification rate of magnetic
767: energy until core bounce, which is defined as $A_{\mathrm{amp}}
768: \stackrel{\mathrm{def}}{\equiv} (E_m|_{\mathrm{bnc}})/(E_m|_{\mathrm{ini}})$.
769: $t_{\mathrm{bnc}}$ represents the time till bounce.
770: %In the rapidly rotating models the core bounce occurs mildly, therefore
771: % the definition of that have some ambiguity.
772: }
773:
774: \end{table}
775:
776: \begin{figure}[ht]
777: \begin{center}
778: \includegraphics[width=.99\linewidth]{fig1.eps}
779: \caption{Snapshots for models B10TW0.25 (left) and
780: B10TW4.0 (right), showing the degree of the rotational flattening at core bounce.
781: In each panel, contour of density [$\mathrm{g/cm^3}$] (left) and entropy per
782: baryon [$k_B$] (right) is shown. Flattening of the unshocked core
783: (the central low entropy region seen in the right side of each panel) is
784: remarkable for the right panel.
785: Note that the unit of the horizontal and the vertical axis is in cm.
786: %R Left: Snapshot at the core bounce for Model B10TW0.25 ($245$ ms after the beginning of the simulation).
787: %Right: Snapshot at the core bounce for Model B10TW4.0 ($477$ ms after
788: % the beginning of the simulation).
789: %Rotational flattening of the unshocked inner core is remarkable seen for the rapidly
790: %rotating model of B10TW4.0 (right panel).
791: %The central density and the compactness of the core are different for
792: % the initial rotation rate.
793: }\label{fig:ds_bnc}
794: \end{center}
795: \end{figure}
796: \clearpage
797:
798: \subsection{Prompt vs. Delayed MHD Exploding Model}\label{ssec:pvd}
799: After bounce,
800: we can categorize the computed models into two groups, by the criterion whether
801: the shock generated at bounce promptly reach the surface of the
802: iron core or not.
803: For later convenience, we call the former and the latter models as
804: prompt and delayed MHD exploding model, respectively throughout the
805: paper.
806: The models and the corresponding groups are shown in Figure \ref{fig:two_groups}.
807: %The models colored by red and by green correspond to the ``prompt
808: %MHD exploding models'' and ``delayed MHD exploding models'', respectively.
809: To begin with, we choose typical model from the two groups and mention
810: their properties in detail.
811:
812: \begin{figure}[ht]
813: \begin{center}
814: \includegraphics[width=.66\linewidth]{fig2.eps}
815: \caption{Classification of the computed models into the prompt (red
816: blocks) or delayed (green blocks) MHD exploding model by the difference
817: of $t_{1000\mathrm{km}}$ shown in this table, which is the
818: shock-arrival time to the radius of $1000~{\rm km}$ after bounce.
819: %This value is the duration required for the jet to reach 1000km after
820: % bounce.
821: %The red blocks are for the ``prompt MHD exploding models'' and the green ones
822: % for the ``delayed MHD exploding models''.
823: }\label{fig:two_groups}
824: \end{center}
825: \end{figure}
826:
827: \paragraph{Prompt MHD Exploding Model}
828: The models classified into this group have strong magnetic
829: fields and rapid rotation initially.
830: %We show the typical dynamics of this group taking B12TW1.0 as an example.
831: %In this model, the $E_{m}/|W|$ reaches $1\%$ at bounce.
832: %Since the initial $E_m/|W|$ is $2.5\times 10^{-4}$, therefore
833: %$36$ times amplified by rotation during the infall epoch
834: %(see table \ref{tab:PROCPS}).
835: Figure \ref{fig:b12tw100seq} shows the dynamics near core bounce. As
836: seen from the bottom left panel, the shock at core bounce stalls
837: in the direction to the equatorial plane
838: at $\sim 1.4 \times 10^{7} \cm$ promptly ($\sim 3$ ms) after bounce. However the shock in the direction of the
839: rotational axis does not stall and becomes a collimated jet (see top
840: right and bottom right). The wound-up magnetic fields are an important
841: agent to explain these properties.
842:
843: The magnetic fields for the promptly MHD models are {\color{red}strong enough} to power
844: the jet already at the epoch of bounce.
845: That is clearly shown in the top left panel, showing that the ``plasma
846: $\beta$'' $\stackrel{\mathrm{def}}{\equiv} p/\frac{B^2}{8\pi}$, being the ratio of {\color{red}the matter to
847: the magnetic pressure}, outside the unshocked core near
848: the poles becomes very low (typically $10^{-2}$). From the right side
849: of the bottom left panel, the toroidal magnetic field strength there
850: reaches over $10^{15}\mathrm{G}$.
851: The dynamics around the poles are strongly affected by these strong magnetic fields.
852: %It should be noted that the bounce itself is not
853: %magneto-driven (i.e. matter-pressure-driven) because
854: %the plasma $\beta$ inside the unshocked core is still high.
855:
856: The three dimensional plots of Figure \ref{fig:3djet} are
857: useful to see how the field wrapping occurs.
858: From the top left panel, it is seen that the field lines are strongly
859: wound around the rotational axis.
860: The white lines in the top right shows
861: the streamlines of the matter.
862: A fallback of the matter just outside of the head of the jet downwards
863: to the equator (like a
864: cocoon) is seen. In this jet with a cocoon-like structure,
865: the magnetic pressure is always {\color{red} dominant over} the matter pressure (see
866: the region where plasma $\beta$ less than 1 in the right side of the
867: top right panel of Figure \ref{fig:b12tw100seq}).
868: \textcolor{red}{
869: This magneto-driven jet does not stall and penetrate to the surface of
870: the iron core, which is essentially the reproduction of the pioneering
871: results in the MHD supernova simulations by \citet{leblanc} and its analysis by
872: \citet{meier}.}
873: %This magneto-driven jet does not stall and penetrate to the surface of
874: %the iron core.
875: The speed of the head of the jet is mildly relativistic of $\sim 0.3 c$,
876: with $c$ being the speed of light (the right side of bottom right panel of Figure \ref{fig:b12tw100seq}).
877: At $20 \ms$ after bounce, the jet finally reaches the surface of the
878: iron core of $\sim 10^8 \cm$.
879: At this moment, the explosion energy, which will be a useful quantity for comparing
880: the strength of the explosion among the models later,
881: reaches $1.4 \times 10^{50}\ \mathrm{erg}$.
882: %Since the explosion energy increases with time,
883: %this value does not simply represent the observed explosion energy.
884: %This value is used for comparing the strength of explosion among the
885: %models later.
886:
887: \begin{figure}[h]
888: \begin{center}
889: \includegraphics[width=.99\linewidth]{fig3.eps}
890: \caption{Time evolution of various quantities characterizing
891: the dynamics near after bounce
892: for a prompt MHD exploding model
893: (see text in section 4.2 of Prompt MHD exploding models).
894: This is for model B12TW1.0.
895: In each panel, the left side
896: represents the logarithm of density [$\mathrm{g/cm}^3$]. Time in each
897: panel is measured from the epoch of bounce. At the top panels, the right
898: side is {\color{red} the logarithm of} the ``plasma $\beta$'' $\stackrel{\mathrm{def}}{\equiv} p/\frac{B^2}{8\pi}$, indicated by ``Beta''. At the bottom left
899: panel, the right side is the logarithm of toroidal component of the magnetic fields
900: [$\mathrm{G}$], indicated by ``$B_{\rm phi}$''.
901: At the bottom right panel, the right side is the radial
902: velocity in unit of the speed of light $:c$.
903: Note that the unit of the horizontal and the vertical axis of all
904: panels are in cm.
905: }\label{fig:b12tw100seq}
906: \end{center}
907: \end{figure}
908:
909:
910: \begin{figure}[h]
911: \begin{center}
912: \includegraphics[width=.99\linewidth]{fig4.eps}
913: \caption{
914: Three dimensional plots of entropy with the magnetic field lines
915: (left) and the streamlines of the matter (right) during the jet
916: propagation for models of B12TW1.0 (top) and B10TW1.0 (bottom), at
917: $20 \ms$ and $94 \ms$ after bounce, respectively.
918: The outer edge of the sphere colored by blue represents the radius of
919: $7.5 \times 10^7 \cm$.
920: Note that the model of the top and bottom panel belongs to
921: the prompt and delayed MHD exploding model, respectively.
922: These panels highlight not only the wound up magnetic field around the
923: rotational axis (left), but also the fallback of the matter from the head of the
924: jet downwards to the equator, making a cocoon-like structure behind
925: the jet (right).
926: }\label{fig:3djet}
927: \end{center}
928: \end{figure}
929: \clearpage
930:
931: \paragraph{Delayed MHD Exploding Model}
932: The models with weaker initial magnetic fields belong to the delayed
933: MHD exploding model (see Figure \ref{fig:two_groups}).
934: %Figure \ref{fig:two_groups} shows more detailed informations on the models.
935: In the following, we explain their properties taking model B10TW1.0 as an example.
936: It is noted again that this model has the same initial
937: rotation rate with model B12TW1.0 of the previous section,
938: but with the two orders-of-magnitudes weaker initial magnetic fields.
939:
940: In the case of model B10TW1.0, the shock wave at bounce stalls
941: in all directions at $\sim 1.5 \times 10^{7} \cm$.
942: %The magnetic energy per
943: %the gravitational energy is very weak ($5\times 10^{-7}$) at the core bounce.
944: %The initial $E_m/|W|$ is $2.5\times 10^{-8}$, therefore
945: %$28$ times amplified by the rotation.
946: As shown in the top left panel of Figure
947: \ref{fig:b10tw100seq}, the plasma $\beta$ is so high that the magnetic
948: fields play no important role before bounce.
949: After the shock stalls, the stalled shock begins to oscillate.
950: %shape of shock front becomes oblate
951: %and begins to oscillate.
952: The middle left and the bottom left panel shows the prolate and
953: oblate phase during the oscillations, respectively.
954: Until $\sim 70 \ms$ after bounce, the oscillation of the shock front
955: continues diminishing its amplitude. Approximately the number of the
956: oscillations is about $5$ times this time. Without the magnetic
957: fields, the oscillation should cease settling into the equilibrium
958: state with the constant accretion through the stalled-shock to the center. However during this
959: oscillation, the magnetic fields behind the stalled shock
960: gradually grow due to the field wrapping and the plasma $\beta$ around the
961: polar regions becomes low as seen from the right side of the top right panel.
962: Soon after the toroidal magnetic fields become as high as $\sim 10^{15}\mathrm{G}$
963: behind the stalled shock (see the middle right panel),
964: the stalled shock near the pole suddenly begins to propagate along the
965: rotational axis and
966: turns to be a collimated jet (see the bottom right panel).
967: This revived jet does not stall in the iron-core. This is the reason
968: why we call this model as the delayed MHD exploding model.
969: %The radial velocity in unit of the speed of light is shown in the bottom right
970: %panel of Figure \ref{fig:b10tw100seq}.
971: The speed of the jet reaches about $5.5\times 10^9 \mathrm{cm/s}$ (see
972: the bottom right panel).
973: Also in this jet, the toroidal component of the magnetic fields is
974: dominant over the poloidal one and a fallback of the matter
975: is found in the outer region of the jet (cocoon) as in the case of the
976: promptly MHD exploding model (see the bottom two panels of Figure \ref{fig:3djet}).
977: %The bottom two panels
978: %of Figure \ref{fig:3djet} illustrate that three dimensional snap shots of this jet.
979: At $\sim 96 \ms$ after bounce, the jet reaches $\sim 10^8 \cm$.
980: The explosion energy at that time reaches
981: $0.094 \times 10^{50} \mathrm{ergs}$.
982:
983: As mentioned, the dynamical behaviors between the prompt and delayed MHD
984: exploding models after bounce seem apparently different. However
985: there are some important similarities between them, which we discuss
986: from the next section.
987: %However there are some similarities among them, which we discuss in the next section.
988:
989: \begin{figure}[h]
990: \begin{center}
991: \includegraphics[width=.8\linewidth]{fig5.eps}
992: \caption{Same as Figure 3 but for the quantities showing the dynamics near
993: after bounce for a delayed exploding model (see text in section 4.2:Delayed MHD
994: exploding model). This is for model B10TW1.0. During the oscillations
995: of the stalled shock after bounce (from middle left to
996: bottom left), the magnetic fields behind the stalled shocks become
997: large enough, due to the field wrapping (top and middle right), leading to the
998: shock-revival for the formation of the magnetically-dominated jet (bottom right).
999: Note that the unit of the horizontal and the vertical axis of all
1000: panels is in cm.
1001: %In each panel, the left side represents the logarithm of density
1002: %[$\mathrm{g/cm}^3$]. Time in each panel is measured from the epoch of bounce.
1003: %Top Left Panel: Logarithm of density [$\mathrm{g/cm}^3$] (left) and plasma $\beta$ (right) is shown at the core bounce.
1004: %Middle Left Panel: Logarithm of density [$\mathrm{g/cm}^3$] (left) and
1005: % entropy per baryon [$k_{\mathrm{B}}$](right) is shown at $13 \ms$
1006: % after bounce.
1007: %Bottom Left Panel: Logarithm of density [$\mathrm{g/cm}^3$] (left) and
1008: % entropy per baryon [$k_{\mathrm{B}}$](right) is shown at $25 \ms$
1009: % after bounce.
1010: %Top Right Panel: Logarithm of density [$\mathrm{g/cm}^3$] (left) and
1011: % logarithm of plasma $\beta$ (right)is shown at $70 \ms$ after bounce.
1012: %Middle Right Panel: Logarithm of density [$\mathrm{g/cm}^3$] (left) and
1013: % logarithm of toroidal magnetic field [$\mathrm{G}$] (right) is
1014: % shown at $74 \ms$ after bounce.
1015: %Bottom Right Panel: Logarithm of density [$\mathrm{g/cm}^3$] (left) and
1016: % radial velocity in the unit of light speed (right) is
1017: % shown at $77 \ms$ after bounce.
1018: }\label{fig:b10tw100seq}
1019: \end{center}
1020: \end{figure}
1021: \clearpage
1022: \subsection{Similarities of Prompt and Delayed MHD Exploding Model}
1023:
1024: In this section we focus on the similarities between the prompt and
1025: delayed MHD exploding models.
1026:
1027: From Figure \ref{fig:vrB}, it can be seen that
1028: the radial velocities and the magnetic fields of the jets are quite
1029: similar among the models regardless of the prompt or delayed exploding
1030: models.
1031: Typical values of the toroidal magnetic fields are $10^{14}-10^{16}\mathrm{G}$
1032: and typical velocities are $10-30\%$ of the speed of light.
1033: The opening angles of the jets are also similar.
1034: The width of this jet is about $8 \times 10^6 \cm$ when the jet
1035: reaches $7.5\times 10^7 \cm$, which means that the half opening angle
1036: of the jets is about $6^{\circ}$ at this time.
1037: These characteristic values of the jets are summarized in Table \ref{tab:PROJET}.
1038:
1039: Detailed properties of the jets {\color{red} in the vicinity along the rotational axis}
1040: are shown in Figures \ref{fig:JPTW100}
1041: and \ref{fig:JPB11} to see the origin of these similarities.
1042: We fix the initial rotation
1043: rate and the initial field strength in Figure \ref{fig:JPTW100} and
1044: \ref{fig:JPB11}, respectively, to see their effects separately.
1045: In Figure \ref{fig:JPTW100}, the initial rotation rate is
1046: $T/|W|=1.0\%$ and the different lines correspond to the
1047: difference between the initial magnetic fields from $10^{12}$ (B12) to
1048: $10^{10}$ G (B10). In Figure \ref{fig:JPB11}, the initial magnetic
1049: field is $10^{11}\mathrm{G}$ and the different lines corresponds the
1050: difference in the initial rotation rates.
1051:
1052: From the top and middle panels of Figures \ref{fig:JPTW100} and
1053: \ref{fig:JPB11}, we find that the radial profiles of the
1054: toroidal magnetic field,
1055: the plasma $\beta$ ($0.1-0.01$), the density,
1056: and the velocity, are rather similar behind the shock whose position
1057: can be seen from the discontinuity at $\sim 700$ km.
1058: Above all, it is surprising to see the
1059: remarkable similarity in the profiles of the toroidal magnetic fields
1060: behind the shock among the models (top left in Figures \ref{fig:JPTW100} and
1061: \ref{fig:JPB11}). The typical strength behind the shock
1062: is seen to be $\sim 10^{15}\mathrm{G}$. This critical strength of the toroidal
1063: magnetic field for the shock-revival is estimated as follows.
1064: The matter behind the stalled-shock is pushed inwards by the ram pressure of
1065: the accreting matter. This ram pressure is estimated as,
1066: \begin{equation}
1067: P=4 \times
1068: 10^{28}\left(\frac{\rho}{10^{10}\mathrm{g/cm^3}}\right)\left(\frac{\Delta
1069: v}{2 \times 10^9\mathrm{cm/s}}\right)^2\mathrm{erg/cm^3},
1070: \end{equation}
1071: where the typical density and the radial velocity are taken
1072: from Figure \ref{fig:ds_bnc} and the bottom right panel of
1073: Figure \ref{fig:b10tw100seq}, respectively.
1074: When the toroidal magnetic fields are amplified as large as $\sim
1075: 10^{15}\mathrm{G}$ due to the field wrapping behind the shock,
1076: the resulting magnetic pressure, $\frac{B^2}{8\pi}$, can overwhelm the ram
1077: pressure, leading to the magnetic shock-revival.
1078: The origin of the similarity of the jets seen in
1079: Figure \ref{fig:b12tw100seq} comes from this mechanism.
1080: We find that this process works in all the computed models.
1081: {\color{red}It is noted that the importance of the magnetic-shock revival
1082: was noticed also in the analytic models by \citet{uzde07a,uzde07b}.
1083: In addition to their expectations, our simulations show that the
1084: explosion energy becomes smaller than their estimations because
1085: the magnetic tower cannot be wider as they assumed.}
1086:
1087: From the bottom panels of Figure \ref{fig:JPB11}, it can be seen
1088: that the poloidal fields behind the shock front
1089: do not depend on the initial rotation rate so much given the same
1090: initial field strength,
1091: while the difference of the poloidal magnetic fields behind the shock
1092: in the bottom panels of Figure \ref{fig:JPTW100} simply comes from the
1093: difference in the initial field strength.
1094: This feature is regardless of the prompt or delayed models.
1095: %On the other hand,
1096: %it is found from the bottom panels of Figure
1097: %\ref{fig:JPB11} that the poloidal fields do not depend on
1098: %the initial rotation rate so much, given the same initial field
1099: %strength.
1100: %surprising that they reflects the
1101: %difference of the initial magnetic fields given the same initial
1102: %rotational rate (the bottom left panel of Figure \ref{fig:JPTW100}).
1103: %Interestingly it is found from the bottom panels of Figure
1104: %\ref{fig:JPB11} that the poloidal fields in the jet does
1105: %not depend on the initial rotation rate .
1106: % It is natural that the poloidal magnetic fields reflects the
1107: %difference of the initial magnetic fields given the same initial
1108: %rotational rate (the bottom left panel of Figure \ref{fig:JPTW100}).
1109:
1110: %It is surprising to see the
1111: %remarkable accordance in the profiles of the toroidal magnetic fields
1112: %behind the shock among the models.
1113: %The typical strength is seen to be $\sim 10^{15}\mathrm{G}$.
1114:
1115: %The origin of the similarity of the jets seen in
1116: %Figure \ref{fig:b12tw100seq} comes from this accordance.
1117: % The critical strength of the toroidal
1118: %magnetic field for the shock-revival is estimated as follows.
1119:
1120: % The matter behind the stalled-shock is pushed inwards by the ram pressure of
1121: % the accreting matter. This ram pressure is estimated as,
1122: %\begin{equation}
1123: % P=4 \times
1124: % 10^{28}\left(\frac{\rho}{10^{10}\mathrm{g/cm^3}}\right)\left(\frac{\Delta
1125: % v}{2 \times 10^9\mathrm{cm/s}}\right)^2\mathrm{erg/cm^3},
1126: %\end{equation}
1127: %where the typical density and the radial velocity are taken
1128: %from Figure \ref{fig:ds_bnc} and the bottom right panel of
1129: %Figure \ref{fig:b10tw100seq}, respectively.
1130: %When the toroidal magnetic fields are amplified as large as $\sim
1131: %10^{15}\mathrm{G}$ due to the field wrapping behind the shock,
1132: %the resulting magnetic pressure, $\frac{B^2}{8\pi}$, can overwhelm the ram
1133: %pressure, leading to the magnetic shock-revival.
1134: %In this time the jet could break out a ``Wall'' produced by the accretion.
1135:
1136:
1137: \begin{figure}[ht]
1138: \begin{center}
1139: \includegraphics[width=.90\linewidth]{fig6.eps}
1140: \caption{Contour of radial velocity (left side) and toroidal magnetic
1141: fields (right side in each panel) for various models, showing the similarity of the jets between the
1142: prompt and delayed MHD exploding models.
1143: The right and left panels correspond to the rapidly rotating ($T/|W|=4.0\%$)
1144: and the slowly rotating $T/|W|=0.25\%$ models, respectively.
1145: From top to bottom panels, the initial strength of the
1146: magnetic fields changes from strong (B12 models) to weak (B10 models).
1147: The time from bounce is shown in the top right part of each panel,
1148: indicating the difference between the prompt and the delayed models.
1149: Note that the unit of the horizontal and the vertical axis of all
1150: panels is in cm.}\label{fig:vrB}
1151: \end{center}
1152: \end{figure}
1153:
1154: \begin{figure}[ht]
1155: \begin{center}
1156: \includegraphics[width=.99\linewidth]{fig7.eps}
1157: \caption{
1158: Various profiles of jets along the rotational axis, for models
1159: B12TW1.0 at $20 \ms$ (solid), B11TW1.0 at $24 \ms$ (short-dashed)
1160: and B10TW1.0 (dotted) at $94 \ms$ after bounce, respectively, showing
1161: the effects of the difference of the initial magnetic fields (fixing
1162: the initial rotation rate of $T/|W|=1.0\%$).
1163: Top left and right is the toroidal magnetic field ($B_{\phi}$) and
1164: plasma beta ($\beta = P/\frac{B^2}{8\pi}$).
1165: Middle left and right is the radial velocity [cm/s] and the density [$\mathrm{g/cm^3}$].
1166: Bottom left and right is the poloidal magnetic field ($B_{p}$) and the ratio
1167: of toroidal to poloidal magnetic field ($B_{\phi}/B_{p}$).
1168: Note that the shock position is approximately $700$ km as seen from
1169: the discontinuity of these profiles.
1170: {\color{red}
1171: It should be noted that the toroidal fields here are not
1172: just at $\theta=0^{\circ}$ (the fields are zero there)
1173: but at the closest mesh to the axis of
1174: $\theta=1.5^{\circ}$.}
1175: }\label{fig:JPTW100}
1176: \end{center}
1177: \end{figure}
1178:
1179: \begin{figure}[ht]
1180: \begin{center}
1181: \includegraphics[width=.99\linewidth]{fig8.eps}
1182: \caption{Same as Figure \ref{fig:JPTW100}, but for models
1183: B11TW4.0 at $28 \ms$ (solid), B11TW1.0 at $24 \ms$ (short-dashed),
1184: and B11TW0.25 (dotted) at $64 \ms$ after bounce, respectively,
1185: showing the effects of the difference of the initial rotation rates
1186: (fixing the initial field strength of $B11$).
1187: }\label{fig:JPB11}
1188: \end{center}
1189: \end{figure}
1190:
1191: \clearpage
1192: \begin{table}
1193: \begin{center}
1194: \caption{Characteristic properties of jets }\label{tab:PROJET}
1195: \begin{tabular}{cccccc}
1196: \tableline\tableline
1197: %\backslashbox
1198: Name&
1199: $t_{\mathrm{exp}}$ & $t_{\mathrm{1000km}}$ &
1200: $v_{\mathrm{jet}}$ & ${E_{\mathrm{exp}}}_{\mathrm{1000km}}$ \\
1201: & $[\mathrm{ms}]$ & $[\mathrm{ms}]$ &
1202: $[\mathrm{cm/s}]$ & $[10^{50}\mathrm{ergs}]$ \\
1203: \tableline
1204: B12TW0.25 &
1205: $ 0$ & $ 32 $ & $ 6.0 \times 10^{9}$ & $ 1.3$ \\
1206: B11TW0.25 &
1207: $48$ & $ 72 $ & $ 3.7 \times 10^{9}$ & $0.05$ \\
1208: B10TW0.25 &
1209: $92$ & $ 122$ & $ 3.8 \times 10^{9}$ & $0.02$ \\
1210: \tableline
1211: B12TW1.0 &
1212: $ 0$ & $ 20 $ & $ 8.0 \times 10^{9}$ & $1.4$ \\
1213: B11TW1.0 &
1214: $10$ & $ 27 $ & $ 8.0 \times 10^{9}$ & $0.23$ \\
1215: B10TW1.0 &
1216: $70$ & $ 96 $ & $ 5.5 \times 10^{9}$ & $0.094$ \\
1217: \tableline
1218: B12TW4.0 &
1219: $ 0$ & $ 25$ & $ 6.0 \times 10^{9}$ & $1.0$ \\
1220: B11TW4.0 &
1221: $16$ & $ 32$ & $ 6.0 \times 10^{9}$ & $0.10$ \\
1222: B10TW4.0 &
1223: $65$ & $104$ & $ 5.5 \times 10^{9}$ & $0.006$ \\
1224: \tableline
1225: \end{tabular}
1226: \tablecomments{
1227: Properties of the jets.
1228: $t_{\mathrm{exp}}$ is the duration from bounce to the revival of the
1229: stalled shock due to the field wrapping.
1230: $t_{\mathrm{1000km}}$ represents the duration required for the jet to
1231: reach $\sim \mathrm{1000km}$ after bounce.
1232: Explosion energy, ${E_{\mathrm{exp}}}_{\mathrm{1000km}}$, and the jet
1233: velocity, $ $ $v_{\mathrm{jet}}$ is estimated at the moment.
1234: ${E_{\mathrm{exp}}}_{\mathrm{1000km}}$ are normalized as $10^{50}$erg.
1235: For the definition of the explosion energy, see Eq. (\ref{eq:expene}).
1236: }
1237: \end{center}
1238: \end{table}
1239: \clearpage
1240:
1241:
1242: \subsection{Dependence of Jet Arrival Times and Explosion Energies on
1243: Initial Rotation Rates and Magnetic Field Strengths}\label{sec:dependence}
1244: In the previous section, we discuss the similarity among the computed models.
1245: From this section, we move on to discuss the differences among them.
1246:
1247: \paragraph{Jet Arrival Time}
1248: First we discuss the ``jet arrival time'' shown in Table
1249: \ref{fig:two_groups}, which is the timescale when the jet reaches the outer edge of
1250: the iron core of $\sim 1000\mathrm{km}$.
1251: % and
1252: %reflects the duration required for launching the jet.
1253: As discussed in the previous section, this timescale
1254: is mainly determined how long it takes for the magnetic fields behind the
1255: shock to become as large as the critical toroidal magnetic fields
1256: ($\sim 10^{15}\mathrm{G}$) as a result of the field wrapping.
1257: %As we classified into two groups in Table \ref{fig:two_groups},
1258: %the jet arrival time is different for each model.
1259: %This value depends on how the critical magnetic fields are
1260: %generated near the pole behind the shock wave generated by the core-bounce.
1261: %As we discussed in the previous section,
1262: %the critical value of is determined by the
1263: %accretion rate.
1264:
1265: From the top left panel of Figure \ref{fig:exptime},
1266: it is seen that the strong initial magnetic fields shorten the jet arrival time.
1267: %the initial strength of the magnetic fields affects
1268: %the jet arrival time.
1269: %The shows the magnetic dependence
1270: %on the duration.
1271: %It is clear that strong initial magnetic fields shorten the jet arrival time.
1272: This tendency is seen in all the computed models regardless of the
1273: prompt or delayed exploding models.
1274: When the initial magnetic fields are strong enough ($\sim 10^{-4}$ of
1275: the gravitational energy), the jet arrival times between the different
1276: initial rotational models become almost the same.
1277: In this case, the critical magnetic fields for
1278: the shock-revival are already generated by the
1279: compression before core bounce. So the strong magneto-driven jets
1280: can produce the prompt MHD explosions in a similar way.
1281: For the rapidly rotating
1282: models (the sequence of TW1.0 and TW4.0),
1283: it is seen that the decrease in the rate of the jet arrival time
1284: as a function of the initial $E_m/|W|$ becomes smaller
1285: when the initial $E_m/|W|$ is larger than $\sim 10^{-6}$ (see the kink
1286: in the panel). This is because too strong
1287: magnetic fields transport the angular momentum of the protoneutron
1288: star outwards, leading to the suppression of the efficiency of the
1289: field wrapping after bounce.
1290:
1291: In the top right panel of Figure \ref{fig:exptime}, the
1292: dependence of the jet arrival time on the initial rotation rate is shown.
1293: By intuition, the jet arrival time may become shorter as the initial
1294: rotation rates become larger since the field-wrapping should become
1295: more efficient. The panel shows that this is true for moderately
1296: rotating models of the initial $T/|W|$ less than $0.01$,
1297: but not true for the more rapidly rotating models. This can be
1298: explained as follows.
1299: %, i.e., rapid rotation does not
1300: %necessarily shorten the jet arrival time.
1301: Too rapid rotation of the core hinders the central core from collapsing
1302: due to the stronger centrifugal forces.
1303: This feature is clearly shown in the middle left panel of Figure
1304: \ref{fig:exptime} showing the density profiles.
1305: The density near the center is $\sim 100$ times lower than that for the slowly
1306: rotating models.
1307: {{\color{red}
1308: Since the angular momentum is well conserved before bounce (see section 4.1),
1309: the inner core ($\lesssim 20$ km) gains smaller angular velocities
1310: for rapidly rotating models by the weakened compression
1311: as seen in the middle right panel of Figure \ref{fig:exptime}.}
1312: %Consequently, the inner core cannot gain much angular
1313: %momentum by compression before the core-bounce as
1314: %seen in the middle panels of Figure \ref{fig:exptime}.
1315: %It is noted here again that the angular
1316: %momentum is well conserved before the core-bounce and distributed following
1317: %the density (see section 4.1).
1318: %The density and the angular velocity in rapidly rotating models are lower
1319: %than that of the slow rotating model.
1320: %These features are indeed
1321: Reflecting these aspects, the amplification rate
1322: of the magnetic fields ($\del{t}{E_m}/E_m$) near core-bounce becomes smaller
1323: for the most rapidly rotating model (TW4.0) as seen from the bottom panel of Figure 9.
1324: %, although the magnetic energy itself becomes larger
1325: %for the rapidly rotating model (bottom left panel).
1326: This suppression makes the jet arrival time almost constant or longer as the initial
1327: $T/|W|$ becomes larger than $\sim 0.01$ as in the top right panel of
1328: Figure \ref{fig:exptime}.
1329:
1330: \paragraph{Explosion Energies}
1331: In addition to a wide variety of the jet arrival times,
1332: we find a large difference in the strengths of the magnetic
1333: explosions.
1334:
1335: As a measure of the strength, we define the explosion energy as,
1336: \begin{equation}
1337: {E_{\mathrm{exp}}}_{\mathrm{1000km}}
1338: =
1339: \int_{\mathrm{D}}\d
1340: V~e_{\mathrm{local}}
1341: =
1342: \int_{\mathrm{D}}\d
1343: V\left(e_{\mathrm{kin}}+e_{\mathrm{int}}+e_{\mathrm{mag}}+e_{\mathrm{grav}}\right),\label{eq:expene}
1344: \end{equation}
1345: here $e_{\mathrm{local}}$ is the sum of $e_{\mathrm{kin}}$, $e_{\mathrm{int}}$, $e_{\mathrm{mag}}$ and
1346: $e_{\mathrm{grav}}$, with being the kinetic, internal, magnetic, and gravitational energy, respectively (see Appendix \ref{app:enedef} for
1347: their definitions in special relativity) and $\mathrm{D}$ represents
1348: the domain where the local energy is positive, indicating that the matter
1349: is not bound by the gravity. The explosion energy is evaluated when
1350: the jet arrives at the radius of $1000\km$ at the polar direction.
1351: The value of the explosion energy is summarized in Figure
1352: \ref{fig:expenergy}.
1353: Generally speaking, it is found that the explosion energies becomes
1354: larger for the prompt MHD exploding models (red) than the delayed MHD
1355: exploding models (green).
1356: %the models whose jet arrival times are short, have
1357: %larger explosion energies. In contrast, the
1358: %Even for the delayed exploding models, we found the magneto-driven jets,
1359: %however the explosion energies of these models become smaller.
1360:
1361: What makes the difference on the explosion energies
1362: {\color{red}
1363: at the shock breakout from the iron cores?}
1364: {\color{red} Firstly, the initial strength of the magnetic field
1365: is the primary agent to affect the explosion energies. The explosion energies are
1366: larger for models with the larger initial fields as seen in Figure 11.
1367: Secondly, the geometry
1368: of the jets has also effects on the explosion energies.}
1369: Figure \ref{fig:ble} shows the toroidal magnetic fields (left side)
1370: and the local energy (right side) in the jets from the
1371: stronger to the weak magnetic fields models (from top to bottom panels)
1372: {\color{red} at the shock breakout}. {\color{red}
1373: In each right panel, it is noted
1374: that the regions with the positive local energies
1375: ($e_{\mathrm{local}} > 0$ in Eq. (12)) are drawn with color scales and
1376: the regions with black are for the regions with the negative local energies.}
1377: It is seen that the regions where
1378: the local energy is positive mostly coincide with the regions where
1379: the strong toroidal magnetic fields are generated.
1380: As the initial field strength becomes larger, the regions where the
1381: local energy becomes positive, becomes larger ({\color{red} i.e. the jets become wider}), leading to the larger explosion energies.
1382: In the case of the delayed
1383: exploding model (the right panel in Figure \ref{fig:ble}),
1384: it is found that the width of the jets becomes
1385: narrower, which results in the smaller explosion energies.
1386: Although the properties of the jets just on the rotational axis are
1387: similar among the models seen from Figures 7 and 8,
1388: the lateral structures of the
1389: jets is found to have influence over the explosion energies.
1390:
1391:
1392: %there are large difference on the regions of the jets.
1393: %This difference makes wide divergence on the explosion energy.
1394:
1395: \begin{figure}[h]
1396: \begin{center}
1397: \includegraphics[width=.99\linewidth]{fig9.eps}
1398: \caption{Dependence of the jet arrival time on
1399: the initial magnetic field strength (top left) and the initial rotation rate
1400: (top right). Here
1401: $E_m/|W|$ and $T/|W|$ represent the ratio of initial magnetic and
1402: rotational energies to the gravitational energy, respectively. The jet arrival time
1403: is the duration for the jet to reach the outer edge of iron core of
1404: $\sim 1000\mathrm{km}$. Middle left and right panel shows the distribution
1405: of the density and the angular velocity for models B11TW4.0, B11TW1.0 and
1406: B11TW0.25 at core bounce as a function of the equatorial radius, respectively.
1407: Bottom panel shows the temporal amplification rate ($\del{t}{E_m}/E_m$) near
1408: core bounce.
1409: }\label{fig:exptime}
1410: \end{center}
1411: \end{figure}
1412: \clearpage
1413:
1414:
1415: \begin{figure}[ht]
1416: \begin{center}
1417: \includegraphics[width=.66\linewidth]{fig10.eps}
1418: \caption{
1419: The explosion energy for all the models.
1420: The energies are normalized to $10^{50} \mathrm{erg}$.
1421: Note again that the red and green blocks indicate the prompt and
1422: delayed MHD exploding models, respectively.
1423: See Eq. (\ref{eq:expene}) for the definition of the explosion energy.
1424: }\label{fig:expenergy}
1425: \end{center}
1426: \end{figure}
1427: \clearpage
1428:
1429: \begin{figure}[h]
1430: \begin{center}
1431: \includegraphics[width=.99\linewidth]{fig11.eps}
1432: \caption{Contour of toroidal magnetic field and the
1433: local energy for models B12TW1.0, B11TW1.0 and B10TW1.0 near the shock breakout
1434: from the iron core.
1435: In each panels, the logarithm of toroidal magnetic field[G] (left
1436: side) and the logarithm of local energy [$\mathrm{ergs/cm}^3$] (right
1437: side) is shown, respectively. It is noted for each right panel
1438: that the regions with the positive local energies ($e_{\mathrm{local}} > 0$ in Eq. (12))
1439: are drawn with color scales and
1440: the regions with black are for the regions with the negative local energies.
1441: The timescales given is measured from the core-bounce indicate the prompt or delayed MHD
1442: exploding models. The unit of the horizontal and the vertical axis of all
1443: panels are in cm.
1444: %Top: B12TW1.0 at $16 \ms$ after bounce.
1445: %Middle: B11TW1.0 at $21 \ms$ after bounce.
1446: %Bottom: B10TW1.0 at $77 \ms$ after bounce.
1447: }\label{fig:ble}
1448: \end{center}
1449: \end{figure}
1450: \clearpage
1451:
1452:
1453: \section{Summary and Discussion}
1454: We performed a series of two-dimensional MHD
1455: simulations of rotational core-collapse of magnetized massive stars.
1456: The main motivation was to clarify how the strong magnetic fields and
1457: the rapid rotation of the core affect the magnetic explosions.
1458: To handle the very strong magnetic fields,
1459: we developed a new code under the framework of special relativity.
1460: A novel point is that the microphysics such as the realistic equation
1461: of state and the neutrino cooling are implemented to the special relativistic MHD code.
1462: Due to these advantages, our computation can achieve
1463: a longer time-stretch of the evolutions compared to previous studies.
1464: %With these computations,
1465: The obtained results can be summarized as follows.
1466: \begin{itemize}
1467: \item
1468: Magnetically powered jets are commonly found in all the
1469: computed models. In the jets, the magnetic fields are dominated by the
1470: toroidal components as a result of the field wrapping. For the
1471: profiles and strengths of the toroidal fields behind the jets, we find a remarkable
1472: similarity. We find that
1473: the jet-like explosions occur when the
1474: magnetic pressure behind the shock becomes strong, due to the field
1475: wrapping, enough to overwhelm the ram pressure of the accreting matter. The required
1476: toroidal magnetic fields are similar of $\sim 10^{15}\mathrm{G}$,
1477: which can be also understood by a simple order-of-magnetite estimation.
1478: Reflecting the similarity in the mechanism of producing jets,
1479: global properties of the jets such as the velocities ($\sim 20\%$ of
1480: the speed of light) and the half opening
1481: angle of the jets ($\sim 6^{\circ}$) are also found to be similar among the computed
1482: models.
1483: %There are the two types of the generation of the jets.
1484: %One is launched just after the core-bounce and another is launched at
1485: %$\sim 100$ ms after bounce. The jet-like explosions occur when the
1486: %magnetic pressure behind the shock becomes strong, due to the field wrapping, enough
1487: %to overwhelm the ram pressure of the accreting matter. The required
1488: %toroidal magnetic fields are similar of $\sim 10^{15}\mathrm{G}$,
1489: %which can be also understood by a simple order-of-magnetite estimation.
1490: %Reflecting the similarity in the mechanism of producing jets,
1491: %global properties of the jets such as the velocities and the opening
1492: %angles are also found to be similar among the computed
1493:
1494: %dominant, which are generated by the rotational field wrapping.
1495: % For the strength of the toroidal field behind the shock, we find a remarkable
1496: % similarity.
1497: % The required strength for launching the jet is
1498: % $\sim 10^{15}\mathrm{G}$ that is determined by the ram pressure of
1499: % accreting matter.
1500: % Reflecting this, the velocities and the opening angles of the jets also
1501: % have similarities among the various initial conditions.
1502: % The velocities of the jet head are $\sim 20\%$ of the speed of
1503: % light.
1504: % And the half opening angles become $\sim 6^{\circ}$ commonly.
1505:
1506: \item The timescale before the onset of the magnetic shock-revival are
1507: quite different depending on the initial strengths of rotation and
1508: magnetic fields. When the initial strengths of rotation and
1509: magnetic field are larger, the jet can be launched just after the
1510: core-bounce, which we called as the promptly MHD exploding models.
1511: We furthermore find that even for the model with
1512: the weaker initial field and slow rotation, the jet-like explosions
1513: can occur after sufficient field wrapping to reach the critical field
1514: strength, which we called as the delayed MHD exploding models. In this
1515: case, the explosion can be delayed about $\sim 100\mathrm{ms}$ after
1516: bounce. The explosion energy also strongly depends on
1517: the time difference before the shock-revival.
1518: The stronger initial magnetic fields make wider
1519: exploding regions, leading to the larger explosion energy. The largest
1520: MHD-driven explosion energy obtained is $\sim 10^{50}$ erg.
1521: \end{itemize}
1522:
1523: %{\color{red} A role of the shock-revival as a result of the field wrapping,
1524: %was also noticed in the analytic studies by \citet{uzde07a,uzde07b}.
1525: %Our numerical simulations are in favor of their expectations.
1526: %In addition, our simulations show that the explosion energy becomes smaller than
1527: %the estimated ones in their studies because the magnetic tower cannot
1528: %be wider as they assumed. Thus the highly relativistic outflows of the GRBs seem
1529: %to be hardly explained by the jets obtained in our computations. As
1530: %discussed in the final part of this section,
1531: %we think that the jet might be good for the explanation
1532: %of the X-ray flashes.}
1533:
1534: \textcolor{red}{In addition to the magnetic shock-revival,
1535: the neutrino-driven shock revival, namely neutrino heating
1536: from the newly born protoneutron star may energetize the jets
1537: as suggested by \citet{thom07}.
1538: Although we treated only the neutrino coolings in our computations,
1539: we try to estimate the effect of the neutrino heating in the following way
1540: to see which one could be more important to produce the jets.
1541: We compare the energy gained by the neutrino heating
1542: to the magnetic energy.
1543: The specific neutrino heating rate due to neutrino absorptions
1544: ($ \nu_e + n \rightarrow e^{-}+ p$ and ${\bar\nu}_e + p \rightarrow e^{+}+ n$),
1545: which are the dominant heating processes at certain radius in the postbounce phase,
1546: can be estimated using Eq. (10) of \citet{qian96}.
1547: %$ \dot{q}_{\nu N} \sim 9.65 \Nabog \left[(1-Y_{e}) L_{\nu_{e},51}
1548: % \epsilon_{\nu_{e},\MeV}^2
1549: %\right](1-\sqrt{1-R_{\nu}^2/r^2})/R_{\nu,6}$,
1550: %where $\Nabog$ is the Avogadro's number, $L_{\nu,~51}$ is the neutrino luminosity
1551: %in $10^{51}\ergps$, $R_{\nu,6}$ is the radius of the neutrino sphere in $10^6\cm$,
1552: %and $\epsilon_{\nu_{e},\MeV}$ is the neutrino energy on the neutrino sphere in MeV
1553: %\citep{qian96}.
1554: For model B11TW1.0 at $22\ms$ after bounce,
1555: the jet reaches $\sim 5\times 10^{7} \cm$ and the density at the head of the jet,
1556: $\rho_{\rm jet}$, is $\sim 1.5 \times 10^8 \gpcmc$.
1557: At that time, the neutrino sphere locates at $\sim 7 \times 10^{6} \cm$.
1558: The average energies and the luminosity for the electron and anti-electron neutrinos
1559: there are about $10$ and $14 \MeV$, and $50$ and $\sim 7 \times 10^{51} \ergps$.
1560: In this setup, the heating rate due to the neutrino absorptions, $\dot{q}_{\nu N}$,
1561: reads $\sim 7.7 \times 10^{22} {\rm MeV}/{\rm s}/{\rm g}$.
1562: In the same way, the heating rate due to the neutrino pair-annihilation
1563: ($\nu + {\bar \nu} \rightarrow e^{-}+ e^{+}$) can be estimated as,
1564: $\sim 1.0 \times 10^{16} {\rm MeV}/{\rm s}/{\rm g}$, which is negligible
1565: compared to the neutrino
1566: absorptions, albeit with the general relativistic corrections
1567: \citep{salm99,asano1,asano2}, for far outside the neutrino
1568: spheres where the magnetic shock revival occurs.
1569: If the neutrino luminosity maintains during the characteristic time
1570: scale for the delayed MHD explosion, namely $\Delta t_{\rm delayed} \sim 100$ ms,
1571: the fluid element may gain $\dot{q}_{\nu N}(\rho_{\rm jet}
1572: /1.5 \times 10^8 \gpcmc)(\Delta t_{\rm delayed}/100{\rm ms}) \sim 1.6\times 10^{25}
1573: \ergpcmc$ from the neutrino heatings.
1574: This estimation shows that the energy deposition from neutrinos is quite smaller
1575: than the magnetic energy behind the jet,
1576: $\sim \frac{(B/10^{15}{\rm G})^2}{8\pi}\sim 5.4\times 10^{28}\ergpcmc$.
1577: Here, we omitted the neutrino cooling to maximize the effect of the neutrino heating.
1578: This simple estimation shows that for the models computed in this paper, the
1579: neutrino heatings could be minor compared to the magnetic effects. However for
1580: more weakly magnetized and slowly rotating cores, the neutrino heating may
1581: overwhelm the magnetic shock-revival. To seek the criteria which mechanisms can
1582: be dominant, is important, however, beyond scope of this study.}
1583:
1584: {\color{red}
1585: With respect to our numerical computations, we have to give some discussions.
1586: First of all, we discuss the convergence of our numerical results.
1587: As mentioned in section \ref{sec:NM}, we have taken the grid numbers
1588: of $N_r = 300$ and $N_{\theta} = 60$
1589: as a fiducial value with $N_r$ and $N_{\theta}$ being the grid numbers in radial and
1590: polar direction, respectively. For the convergence tests, we vary $N_{r}$ from $200$,
1591: $400$, to $1000$ fixing $N_{\theta} =60$ and $N_{\theta}$ from $30$, $90$, to $120$,
1592: while fixing $N_{r} =300$. Taking model B11TW1.0,
1593: we follow the dynamics for models with the different numerical resolutions
1594: till $\sim 30$ ms after bounce when the magneto-driven jets
1595: come out of the central iron cores.
1596: We pay attention to the (spatially-integrated) magnetic energies
1597: among the models, because they are a good indicator to see the degree of
1598: the amplification of the magnetic fields.
1599: Varying the coarser to the finer grid resolutions for the radial and polar direction,
1600: the relative change in the magnetic energies from the fiducial model is found to
1601: be in the range of $+0.15\%-0.35\%$ and $\pm 3.0\%$, respectively.
1602: These resolution tests seem to support the convergence of the numerical values
1603: obtained in this paper. However we shall note that
1604: better angular resolutions are needed to resolve the collimation of the jets
1605: further than $\sim 4\times 10^7$ cm from the center.
1606: This is because the intervals of the radial mesh, $\Delta r$, become logarithmically large in the spherical coordinates,
1607: which may lead to the overestimation of the collimation.
1608: Since the jets propagate rather cylindrically around the polar axis later on,
1609: we think it a good way to switch from
1610: the spherical coordinates to the cylindrical coordinates at certain radius
1611: to sustain better angular resolutions, however, beyond scope of this study.
1612: Next we estimate the numerical dissipation of the magnetic fields.
1613: During the infall epoch, the toroidal magnetic flux is a good quantity to
1614: see the flux conservation. The difference between the initial toroidal magnetic flux,
1615: $\Phi_{\mathrm{ini}}$, and the flux at the bounce, $\Phi_{\mathrm{bnc}}$, is
1616: estimated to be $\Delta\Phi = \frac{\Phi_{\mathrm{ini}}-\Phi_{\mathrm{bnc}}}{\Phi_{\mathrm{ini}}} \sim 1 \%$.
1617: This means that the numerical magnetic dissipation is
1618: treated to be small for our numerical code.}
1619:
1620: Then we move on to address a few imperfections in our simulations.
1621: Magneto-rotational Instability (MRI) has been pointed out to be active
1622: in the outer layer of protoneutron star
1623: which rotates strongly differentially \citep{akiy03,kota04b,masa06}.
1624: The wavelength of maximum growth rate of the linear instability,
1625: $\lambda=\frac{2\pi v_{a}}{\Omega}$, becomes
1626: $
1627: \sim 5
1628: \left(\frac{300 \mathrm{s}^{-1}}{\Omega}\right)
1629: \left(\frac{B}{10^{12}\mathrm{G}}\right)
1630: \left(\frac{10^{10}\mathrm{g/cm^3}}{\rho}\right)^{1/2}\mathrm{km}
1631: $
1632: \citep{balb98}, where our numerical grid is $\sim 3\mathrm{km}$
1633: there. Since $10-100$ times finer mesh than the wavelength is
1634: required for resolving the MRI \citep{shib06}, our simulations are
1635: insufficient to take into account the field amplification due to the MRI.
1636: This is a very important task remained to be investigated,
1637: although the computational costs are still expensive.
1638: %Our simulation resolution of $\sim
1639: %3\mathrm{km}$ there could not satisfy the required conditions.
1640: %Therefore the real amplification rate should becomes larger than ours.
1641: % It is an very important task remained to be investigated,
1642: %Non-axisymmetric MRI instabilities should be very
1643: %interesting to be investigated \citep{masa06}, however beyond our scope here.
1644: As for the microphysics, the neutrino heating is not included as mentioned above
1645: (see, however, \citet{burr07}). Since the timescale for the magnetic explosions in
1646: the polar direction are much earlier than the neutrino heating or
1647: the g-mode excited explosions, we think that both of them basically
1648: should play a minor role here.
1649: On the other hand, the neutrino
1650: heating could be helpful for producing the explosion in the direction
1651: of the equatorial plane where the field-wrapping induced
1652: magnetic explosions are unlikely to occur.
1653: Finally the general relativistic (GR) effects are
1654: not incorporated in our simulations.
1655: During our simulation time,
1656: the central protoneutron stars do not collapse to the black
1657: holes as inferred from a simple estimation of the Schwarzshild radius.
1658: Thus we think that the GR effects may not drastically change our
1659: results qualitatively.
1660: \textcolor{red}{
1661: After the jet break-out from the star leaving behind the
1662: narrow funnel, the mass accretion in the direction of the equatorial plane
1663: may lead to the black hole formation, which provides us with
1664: the initial condition of the collapsar. This would be another interesting
1665: subject as an application of this study.}
1666:
1667: Bearing these caveats in mind,
1668: we state some speculations based on our results.
1669: The protoneutron stars obtained here are with the poloidal magnetic fields of order $10^{15}[\rm{G}]$ and with its
1670: rotation period of an order of milliseconds, which could be the
1671: origin of the magnetar according to a hypothesis by
1672: \citet{dunc92}.
1673: %The protoneutron star rotates with
1674: %The strength of the poloidal component of the magnetic fields
1675: %in the protoneutron star is about $10^{14}-10^{15}\gauss$ in our models.
1676: %Thus such model might be related to the formation of so-called
1677: %magnetars.
1678: If so, it means that the magnetically-driven jets could be
1679: associated with the formation of the magnetars.
1680: Our results suggest that the toroidal component of the magnetic field
1681: is dominant in the young magnetars.
1682: The large magnetic energy of the toroidal component stored in deep crusts
1683: and cores of the magnetars would be transported outside via \Alfven waves
1684: and be released as giant-flares of the SGRs \citep{dunc92,thom01,thom02}.
1685: Some observational evidences support this picture.
1686: For example, quasi periodic oscillations discovered in a X-ray light
1687: curve of the giant-flare from SGR1806-20 would originate from the \Alfven
1688: wave from the interior of the star \citep{rea06,isra05}.
1689: There are several studies indicating that the magnetar formation yields
1690: the XRF \citep{maza06,maed07a,maed07b}.
1691: While the ordinary GRBs require the highly relativistic ejecta,
1692: the mildly relativistic ejecta is favorable for XRFs
1693: \citep{toma07,sode06,ghis07}, which may be the case here because
1694: the magneto-driven jets become only mildly relativistic due
1695: to the high baryon loading of the matter along the rotational axis.
1696: %As shown here,
1697: %due to the high baryon loading of the matter along the rotational axis,
1698: %from the magnetar tend not to reach high Lorentz
1699: %factor due to its high baryon load
1700: %Further investigations are necessary to clarify the connection between
1701: %the magnetar formation and the XRFs.
1702:
1703: There remain more rooms to be investigated applying our simulations.
1704: In this study, we employed one progenitor model. Since
1705: the accretion rate of the matter to the protoneutron stars should
1706: depend on the progenitor mass \citep{hege05,woos06}, we think it very
1707: important to investigate how the criterion of
1708: the magnetic shock-revival changes with the progenitor models.
1709: Moreover the initial configurations of the magnetic fields, which are
1710: still highly uncertain, could be
1711: changed in a systematic manner like in \citet{sawa05,sawa07}, to see
1712: their effects on dynamics.
1713: \textcolor{red}{While this study focused on the shock-propagation in the iron cores,
1714: in which the jets become only mildly relativistic,
1715: we plan to continue to follow the dynamics later on till the jets break out of the stars
1716: (phase 4 in the introduction), in which the jets are expected to be relativistic.}
1717: Very recent studies by \citet{komi07b,bucc07} are on this line.
1718: Our simulation can be more consistent than their studies in the sense
1719: that we start the simulations from the onset of the core-collapse, and
1720: that the protoneutron stars are not excised like their models.
1721: By continuing the simulations of the jet propagations till the
1722: shock-breakout, we plan to study the possible connection between the
1723: magnetically-driven jets obtained here and the origins of the XRFs
1724: in the forthcoming work (Takiwaki et al. in preparation).
1725: \acknowledgments
1726:
1727: We are grateful to S. Yamada for fruitful discussions.
1728: KK thanks to E. M\"{uller} for useful discussions.
1729: TT is grateful to S. Horiuchi for proofreading the manuscript.
1730: Numerical computations were in part carried on VPP5000
1731: and general common use computer system at the center for
1732: Computational Astrophysics, CfCA, the National
1733: Astronomical Observatory of Japan.
1734: This study was partially supported
1735: by Grants-in-Aid for the Scientific Research from the Ministry of Education, Science and
1736: Culture of Japan through No.S 19104006, No. 14079202 and No. 16740134.
1737: This work was supported by World Premier International Research Center Initiative (WPI Initiative), MEXT, Japan.
1738: %PRD
1739: %\bibliographystyle{h-physrev}
1740: %APJ
1741: \bibliographystyle{apj}
1742: \bibliography{sr}
1743:
1744:
1745: \appendix
1746:
1747: \section{Derivation of the Basic Equations for SRMHD}\label{app:DeBE}
1748: In this Appendix, we summarize formalisms on the basic equations and
1749: the numerical tests for our newly developed SRMHD code.
1750: For the formalisms,
1751: we follow the derivation of \citet{devi03,hawl84a,hawl84b}.
1752: For convenience, we proceed
1753: the derivation keeping the metric general forms, i.e.,
1754: \textcolor{red}{
1755: $\d s^2 = -\alpha^2 \d t^2 + \gamma_{ij}\left(dx^i + \beta^i dt \right)\left(dx^j + \beta^j dt \right)$} .
1756: where $\alpha$
1757: is the lapse function, $\beta$ is the
1758: shift vector and $\gamma_{ij}$ is the spatial $3$-metric.
1759: And we take the Minkowski metric later.
1760:
1761: There are four fundamental magnetohydrodynamic equations.
1762: The conservation of baryon number is
1763: \begin{eqnarray}
1764: \udel{\mu}{\rho U^\mu}&=&0 \label{barcons}
1765: \end{eqnarray}
1766: where $\rho$,$U^{\mu}(\mu=0,1,2,3)$
1767: are baryon mass density and 4-velocity at each point.
1768: The conservation of the stress-energy is
1769: \begin{eqnarray}
1770: \udel{\mu}{T^{\mu\nu}}&=&0 \label{tmncons}
1771: \end{eqnarray}
1772: where $T^{\mu\nu}$ is stress-energy tensor
1773: and Maxwell's equations
1774: \begin{eqnarray}
1775: \udel{\mu}{F^{\mu\nu}}&=&4\pi J^{\nu},\\
1776: \udel{\mu}{^{*}F^{\mu\nu}}&=&0\label{ind_eq}.
1777: \end{eqnarray}
1778: where $F^{\mu\nu}$ is the antisymmetric
1779: electro-magnetic tensor and
1780: \begin{eqnarray}
1781: {^{*}F}^{\mu\nu}=\frac{1}{2}\epsilon^{\mu\nu\delta\sigma}F_{\delta\sigma}\label{fs_def}
1782: \end{eqnarray}
1783: is dual of $F^{\mu\nu}$.
1784: Maxwell's equations are supplemented
1785: by the equation of the charge conservation $\udel{\mu}J^{\mu} = 0$.\\
1786: The energy momentum tensor consists of perfect fluid parts
1787: and electromagnetic parts, i.e.
1788: \begin{eqnarray}
1789: T^{\mu\nu}&=&\rho h^{*} U^{\mu}U^{\nu}+pg^{\mu\nu}
1790: +\frac{1}{4\pi}\left({F^{\mu}}_{\alpha}F^{\nu\alpha}
1791: -\frac{1}{4}F_{\alpha\beta}F^{\alpha\beta}g^{\mu\nu}
1792: \right)
1793: \end{eqnarray}
1794: where $h^{*}=(1+e/\rho+p/\rho)$ is the relativistic enthalpy with $e$
1795: and $p$ being the internal energy and the pressure, respectively.
1796:
1797: For later convenience,
1798: we define magnetic induction in
1799: the rest frame of the fluid,
1800: \begin{eqnarray}
1801: b^{\mu}&=&\frac{1}{\sqrt{4\pi}}{^{*}F^{\mu\nu}U_{\nu}}\label{b_def}.
1802: \end{eqnarray}
1803:
1804: We adopt the ideal MHD limit and assume infinite
1805: conductivity (the flux- freezing condition),
1806: where in the electric field in the fluid rest frame is zero, i.e., $F_{\mu\nu}U^{\nu}=0$.
1807:
1808: Combining Eq. (\ref{fs_def}) with (Eq. \ref{b_def})
1809: and conditions for infinite conductivity, we obtain
1810: \begin{equation}
1811: F_{\mu\nu} =\epsilon_{\alpha\beta\mu\nu}\sqrt{4\pi}b^{\alpha}U^{\beta}.
1812: \end{equation}
1813: The orthogonality condition
1814: \begin{equation}
1815: b^{\mu}U_{\mu}=0\label{b_on}
1816: \end{equation}
1817: follows directly from Eq. (\ref{b_def}).
1818:
1819: The induction Eq. (\ref{ind_eq}) can also be
1820: rewritten by substituting the definitions,
1821: \begin{equation}
1822: \udel{\alpha}{}\left(U^\alpha\,b^\beta- b^\alpha\,U^\beta\right) =0 .
1823: \end{equation}
1824: By expanding this equation using the product rule and applying the
1825: orthogonality condition Eq. (\ref{b_on}), we obtain the identity
1826: \begin{equation}
1827: U_\nu\,b^\mu{\nabla}_{\mu}\,U^\nu =0 .\label{bident}
1828: \end{equation}
1829:
1830: It is useful to rewrite the energy momentum tensor as
1831: \begin{eqnarray}
1832: T^{\mu\nu}&=&
1833: \left(\rho h^{*}+\left|b\right|^2 \right)U^{\mu}U^{\nu}
1834: +(p+\frac{\left|b\right|^2}{2})g^{\mu\nu}
1835: -b^{\mu}b^{\nu}.
1836: \end{eqnarray}
1837:
1838: We have to expand basic equations in terms of the code variable,
1839: and transform the equation for auxiliary density, energy and momentum functions
1840: $D=\rho W,E=eW,S_{i}=\rho h W^2 v_{i}$.
1841: Finally the set of variables $D,E,S_{i},B_{i}$
1842: will be evolved through the basic equations transformed here.
1843:
1844: The equation of baryon conservation (\ref{barcons})
1845: can be expanded in terms of the code
1846: variables easily,
1847: \begin{equation}\label{masscons}
1848: \partial_t\,D + {1 \over
1849: \sqrt{\gamma}}\,\partial_j\,(D\,\sqrt{\gamma}\,V^j) = 0.
1850: \end{equation}
1851:
1852: The equation of energy conservation is derived by
1853: contracting (\ref{tmncons}) with ${U}_{\nu}$,
1854: \begin{equation}\label{eq.1a}
1855: {U}_{\nu}\,{\nabla}_{\mu}{T}^{\mu\,\nu} =
1856: {U}_{\nu}\,{\nabla}_{\mu}\left\{
1857: \left(\rho\,h^{*}+{\|b\|}^2\right)\,U^{\mu}\,U^{\nu}+
1858: \left(P+{{\|b\|}^2 \over 2}\right){g}^{\mu\,\nu}-
1859: b^\mu\,b^\nu\right\} = 0 .
1860: \end{equation}
1861: By using the identity (\ref{bident}) and
1862: (\ref{barcons}), we obtain the local energy conservation
1863: \begin{equation}\label{econs}
1864: {\nabla}_{\mu} \left(\rho\,\epsilon\,U^{\mu}\right)+
1865: P\,{\nabla}_{\mu} U^{\mu}= 0 ,
1866: \end{equation}
1867: Applying the definition for the auxiliary energy function $E$,
1868: the energy equation is rewritten as follows:
1869: \begin{equation} \label{enfinal}
1870: \partial_{t}\left(E\right)+{1 \over \sqrt{\gamma}}\,
1871: \partial_{i}\left(\sqrt{\gamma}\,E\,V^i\right)
1872: + P\,\partial_{t}\left(W\right) +
1873: {P\over\sqrt{\gamma}}\,\partial_{i}\left(\sqrt{\gamma}\,W\,V^i\right)
1874: = 0.
1875: \end{equation}
1876:
1877: The momentum conservation equations follow from
1878: \begin{equation}
1879: \nabla_\mu\,{T^\mu}_\nu = {\nabla}_{\mu}\left\{
1880: \left(\rho\,h^{*}+{\|b\|}^2\right)\,U^{\mu}\,U_{\nu}+
1881: \left(P+{{\|b\|}^2 \over 2}\right){\delta^\mu}_\nu-
1882: b^\mu\,b_\nu\right\} = 0 .\label{mom.1}
1883: \end{equation}
1884: This equation can be rewritten as
1885: \begin{eqnarray}
1886: & &{1 \over \alpha\,\sqrt{\gamma}}\,
1887: \partial_\mu\,\sqrt{\gamma}\,S_\nu\,V^\mu +{1 \over 2\,\alpha}\,
1888: { S_\alpha\,S_\beta \over S^t }\,\partial_\nu\,g^{\alpha \beta}\nonumber\\
1889: &+&
1890: \partial_\nu\,\left(P+{{\|b\|}^2 \over 2}\right)-
1891: {1 \over \alpha\,\sqrt{\gamma}}\,
1892: \partial_\mu\,\alpha\,\sqrt{\gamma}\,b^\mu\,b_\nu-{1 \over 2}\,
1893: b_\alpha\,b_\beta\,\partial_\nu\,g^{\alpha \beta}
1894: = 0 .\label{mom.2}
1895: \end{eqnarray}
1896: To obtain the final form of the equations, multiply (\ref{mom.2})
1897: by the lapse $\alpha$, split the $\mu$ index into
1898: its space ($i$) and time ($t$) components, and restrict $\nu$ to the
1899: spatial indices ($j$) only:
1900: \begin{eqnarray}
1901: & &\partial_t\left(S_j-\alpha\,b_j\,b^t\right)+
1902: {1 \over \sqrt{\gamma}}\,
1903: \partial_i\,\sqrt{\gamma}\,\left(S_j\,V^i-\alpha\,b_j\,b^i\right)\nonumber\\
1904: &=&
1905: - {1 \over 2}\,\left({S_\epsilon\,S_\mu \over S^t}+
1906: \alpha\,b_\mu\,b_\epsilon\right)\,
1907: \partial_j\,g^{\mu\,\epsilon}-
1908: \alpha\,\partial_j\left(P+{{\|b\|}^2 \over 2}\right) .\label{mom.3}
1909: \end{eqnarray}
1910: The $\nu$ index can be restricted to the spatial indices because the
1911: equation that arises from $\nu=t$ for the time components of momentum
1912: and magnetic fields is redundant, corresponding to the total energy
1913: conservation equation.
1914: In our formalism, we solve the Eq.
1915: (\ref{enfinal}) separately for the internal energy.
1916: %Basic equations for GRMHD are derived.
1917: %We can easily make the basic equations for SRMHD taking metric as
1918: Taking the following metric,
1919: \begin{eqnarray}
1920: g_{\mu\nu}&=&
1921: \left(
1922: \begin{array}{cc}
1923: -1& 0 \\
1924: 0 & \gamma_{ij} \\
1925: \end{array}\right).
1926: \end{eqnarray}
1927: where $\gamma_{ij}$ is the spatial metric whose
1928: concrete description depends on coordinate system.
1929: Finally we describe our treatment of the gravity.
1930: Under the weak field limit, time-time component of the metric, $g_{tt}$, takes
1931: the form of $-\left(1-2\Phi\right)$ where
1932: $\Phi$ is Newtonian gravitational potential \citep[e.g.][]{shap83}.
1933: % it is possible to
1934: %take into account the general relativistic effects.
1935: %As a first step to include GR effects, we assume the following metric
1936: %form in case of the weak field,
1937: %\begin{eqnarray}
1938: % \d s^2 =-\left(1-2\Phi\right)\d t^2+\left(1+2\Phi\right)\d r^2+r^2\d \Omega.
1939: %\end{eqnarray}
1940: %In this case polar coordinate is taken for mathematical simplification.
1941: The third term of the momentum equation then becomes,
1942: \begin{eqnarray}
1943: -\frac{1}{2}\frac{S^{\alpha}S^{\beta}}{S^{t}}
1944: \udel{j}g_{\alpha\beta}
1945: \approx \rho h W^2 \udel{j}\Phi.
1946: \end{eqnarray}
1947: Under this limit,
1948: Einstein equation becomes the Poisson equation for the
1949: gravitational potential (see Eq. (\ref{eq:poisson})).
1950: Since the origin of the source term is
1951: the $tt$ component of the energy momentum tensor,
1952: we replace $\rho$ in the ordinary Newtonian limit with $T_{tt}$.
1953: as we discussed in the appendix.
1954:
1955: {{\color{red} The validity of using the Newtonian potential in the core-collapse
1956: simulations here may be discussed by seeing the
1957: value of compactness parameter, $\frac{2GM(r)}{rc^2}$, where $r$ is the radius and
1958: $M(r)$ is the enclosed mass
1959: within $r$. In the vicinity of the protoneutron star of $\sim 1.2 M_{\odot}$ with
1960: the typical size of $\sim 20 {\rm km}$, the parameter is $\sim 0.18$. So
1961: the error caused by neglecting higher order metric perturbations
1962: is estimated to below $\sim 3\%$. The qualitative features
1963: found in this paper may be unchanged due to the incursion of the GR. }
1964:
1965: \subsection{Energy Descriptions}\label{app:enedef}
1966: We need to modify the description of energy from the Newtonian one to the special
1967: relativistic one.
1968: The total local energy, $e_{\mathrm{local}}$, is defined by sum of
1969: the various energy:
1970: \begin{eqnarray}
1971: e_{\mathrm{local}}
1972: &=&
1973: e_{\mathrm{kin}}
1974: +e_{\mathrm{int}}
1975: +e_{\mathrm{mag}}
1976: +e_{\mathrm{grav}}
1977: \end{eqnarray}
1978: where
1979: $e_{\mathrm{kin}},e_{\mathrm{int}},e_{\mathrm{grav}}$ and $e_{\mathrm{mag}}$
1980: is kinetic energy, internal energy, gravitational energy
1981: and magnetic energy, respectively.
1982: Their specific description is as follows:
1983: \begin{eqnarray}
1984: e_{\mathrm{kin}}&=&\rho W \left(W-1\right),\\
1985: e_{\mathrm{int}}&=& e W^2+p\left(W^2-1\right),\\
1986: e_{\mathrm{grav}}&=&-\rho h W^2 \Phi,\\
1987: e_{\mathrm{mag}}&=& \vec{B}^2\left(1-\frac{1}{2W^2}\right)
1988: -\frac{{b^0}^2}{2W^2}.
1989: \end{eqnarray}
1990: These descriptions are used for the calculations of the explosion
1991: energy in Subsection \ref{sec:dependence}.
1992:
1993: \section{Special Relativistic MOC}
1994: The method of characteristics (MOC) is popularly used in the
1995: magneto-hydrodynamical simulations.
1996: In this algorithm the magnetic fields are evolved along the
1997: characteristic lines of
1998: the \Alfven waves.
1999: Detailed procedure for this algorithm for the Newtonian case is given
2000: in \citet{ston92}.
2001: For the special relativistic (SR) computations, we derive the solutions of
2002: the SR \Alfven waves in an analytic form,
2003: %the modification the velocity of \Alfven waves is however different from the Newtonian o%ne.
2004: %
2005: %Here we derive analytic solutions of the special relativistic
2006: %\Alfven wave.
2007: \begin{eqnarray}
2008: \left.\Del{t}{Wv_i+b_i/\sqrt{\rho h}}\right|_{-}&=&0,\\
2009: \left.\Del{t}{Wv_i-b_i/\sqrt{\rho h}}\right|_{+}&=&0,\\
2010: \left.\Del{t}{}\right|_{-}
2011: &\stackrel{\mathrm{def}}{\equiv}&\pdel{t}{}+\frac{v_j-\frac{b_j}{\sqrt{\rho h}W}}{(1-b^t/\sqrt{\rho h}W)}\pdel{x_j}{},\\
2012: \left.\Del{t}{}\right|_{+}
2013: &\stackrel{\mathrm{def}}{\equiv}&\pdel{t}{}+\frac{v_j+\frac{b_j}{\sqrt{\rho
2014: h}W}}{(1+b^t/\sqrt{\rho h}W)}\pdel{x_j}{},
2015: \end{eqnarray}
2016: where $W$, $\rho$, and $h$ is the Lorentz factor, density and enthalpy respectively.
2017: $v_j$ and $b_j$ is the perpendicular component of the velocity and the magnetic
2018: field to the $x_{i}$ directions.
2019:
2020: %In this time we develop special relativistic treatment on the MOC.
2021: In the subroutine for solving SR MOC in the code, the velocity and the magnetic fields
2022: are updated at half-time step along the characteristics using the above
2023: equations. By giving the analytic forms, it is readily seen that the speed
2024: of the propagation is guaranteed to be below the speed of light even
2025: for the regions where the density becomes low and the magnetic fields become strong,
2026: which is quite important for keeping the stable numerical
2027: calculations in good accuracy.
2028: %the Newtonian estimation for the \Alfven velocity, $\frac{b_j}{\sqrt{\rho}}$ exceed the speed of light.
2029:
2030: \paragraph{\Alfven Wave Propagation}\label{ap:AWP}
2031: The propagation of a liner \Alfven wave is a basic test problem
2032: of MHD simulation.
2033: We consider a constant background magnetic field, $B_x$, and
2034: fluid velocity, $v_{x}$.
2035: And we add small transverse perturbations with velocity,
2036: $v_z$ ($v_y$), and magnetic field, $B_z$ ($B_y$).
2037: In this situation
2038: $b^t=\sum v^k b_k\approx v_j b_j$,
2039: therefore the analytic solution for the \Alfven wave becomes
2040: \begin{eqnarray}
2041: \left.\Del{t}{Wv_z+b_z/\sqrt{\rho h}}\right|_{-}&=&0,\\
2042: \left.\Del{t}{Wv_z-b_z/\sqrt{\rho h}}\right|_{+}&=&0,\\
2043: \left.\Del{t}{}\right|_{-}
2044: &\approx&\pdel{t}{}
2045: +\frac{v_x-\frac{b_x}{\sqrt{Dh}W}}{1-v_x\frac{b_x}{\sqrt{Dh}W}}\pdel{x}{},\\
2046: \left.\Del{t}{}\right|_{+}
2047: &\approx&\pdel{t}{}
2048: +\frac{v_x+\frac{b_x}{\sqrt{Dh}W}}{1+v_x\frac{b_x}{\sqrt{Dh}W}}\pdel{x}{}.
2049: \end{eqnarray}
2050: If we take
2051: \begin{eqnarray}
2052: Wv_z+b_z/\sqrt{\rho h}=0, \label{nmmode}
2053: \end{eqnarray}
2054: the minus mode does not propagate.
2055: we assume $B_x=0.09$, $v_x=0.08$
2056: for the Newtonian \Alfven wave
2057: and $B_x=0.9$, $v_x=0.8$
2058: for the relativistic \Alfven wave.
2059: We take $v_z$ as $10^{-7}v_x\sin(2\pi x)$
2060: and $B_z$ is determined from $b_z$ in Eq. (\ref{nmmode}).
2061: The result is shown in Figure \ref{relalz}.
2062: In both Newtonian and relativistic cases, the form of the wave is not changed.
2063: It indicates that the computations are successfully performed in our code.
2064: The propagated waveforms are very smooth and no oscillations are found
2065: like the ones in the previous study \citep{devi03}.
2066:
2067:
2068: \begin{figure}[ht]
2069: \begin{center}
2070: \includegraphics[width=.99\linewidth]{fig12.eps}
2071: \caption[\Alfven wave propagation in $z$ direction]
2072: {Top panels:
2073: Newtonian \Alfven wave propagation in $z$ direction.
2074: The left figure show initial and final profile of $v_z$ and
2075: the right one does that of $B_z$.
2076: Bottom panels:
2077: Relativistic \Alfven wave propagation in $z$ direction.
2078: The left figure show initial and final profile of $v_z$ and
2079: the right one does that of $B_z$.}\label{relalz}
2080: %We omit same figures for the $y$ directions.}\label{relalz}
2081: \end{center}
2082: \end{figure}
2083: \clearpage
2084:
2085: \section{Conservative Variable and Fundamental Variables}\label{CVFV}
2086:
2087: In our numerical code, variables such as {$hWv^i,D,E,B^i$} are evolved.
2088: These variables are called conservative variables.
2089: It is necessary to compute
2090: fundamental variables such as $v^i,\rho ,e,B^i$ from these conservative variables.
2091: In our computations, pressure is
2092: not described as an analytic function of energy $e$,
2093: therefore algorithms used in other GRMHD simulations \citep{devi03} is not
2094: available here. If the value of the enthalpy, $h$, is found,
2095: the Lorentz factor, $W$, is obtained from the values of $hWv^{i}$
2096: and then all the fundamental variables are obtained.
2097: To determine the value of $h$, we search the root of the equation below:
2098: \begin{eqnarray}
2099: f(h)\stackrel{\mathrm{def}}{\equiv}(1+e/\rho+
2100: p/\rho+\left|b\right|^2/\rho)-h=0
2101: \label{eq:ite}
2102: \end{eqnarray}
2103: where $\left|b\right|^2=(B^2+b_0^2)/W^2$.
2104: The fundamental variables in the equation are obtained when $h$ is
2105: assumed,
2106: as stated above.
2107: We use simple bisection method to search the root of Eq. \ref{eq:ite}
2108: and make the error of the equation, $\Delta h/h$, below $10^{-4}$.
2109:
2110: \section{Test Problems}\label{TP}
2111: At last we show some results of the numerical tests on our code.
2112: We have done three typical problems
2113: in both non-relativistic case and relativistic case.
2114: Three problems are the shock tube problem, reflection shock problem
2115: and magnetic shock tube problem.
2116: We present the results one by one.
2117:
2118: \paragraph{Shock Tubes}
2119: We consider the shock tube
2120: used by Sod \citep{sod78} in his comparison of finite difference scheme.
2121: First we perform weak shock problem and strong shock problem in
2122: Newtonian case \citep{hawl84b}.
2123: For the weak one,
2124: the initial conditions of this problem
2125: gas with $\Gamma=1.4$ with pressure and density
2126: $P_{l}=1.0,\rho_{l}=10^5$ to the left of $x=0.5$
2127: and
2128: $P_{r}=0.1,\rho_{r}=0.125\times 10^5$ to the right.
2129: For the strong one,
2130: the initial conditions of this problem are gas with $\Gamma=1.4$ with pressure and density
2131: $P_{l}=0.67,\rho_{l}=100$ to the left of $x=0.5$
2132: and
2133: $P_{r}=0.67\times 10^{-7},\rho_{r}=1.0$ to the right.
2134: The numerical values greatly correspond to the analytic value.
2135:
2136: Next we perform relativistic shock tube problems.
2137: We fix the hydrodynamical variables
2138: except for the pressure of the left side,
2139: $\rho_{l}=10,\rho_{r}=1,p_{r}=10^{-6}$\citep{anni03}.
2140: We set three types of pressure, i.e.
2141: $p_{l}=1.33$($W=1.08$) for lowly relativistic case,
2142: $p_{l}=6.67$($W=1.28$) fir mildly relativistic case
2143: and $p_{l}=666.7$($W=3.28$) for highly relativistic case.
2144: The results are shown in the left panels of Figure \ref{fig:rst}.
2145: For lowly and mildly relativistic case,
2146: the numerical value greatly correspond to the analytic value.
2147: For highly relativistic case,
2148: the velocity of the numerical comparison
2149: doesn't reach that of the analytic solution.
2150: It is due to the artificial viscosity which
2151: converted kinetic energy to internal energy.
2152:
2153: \paragraph{Wall Reflections}
2154: A second test presented here
2155: is the wall shock problem involving
2156: the shock heating of cold fluid hitting a wall at the
2157: left boundary. When the fluid hits the wall
2158: a shock forms and travels to the right, separating the pre-shocked
2159: state composed of the initial data and the post-shocked state
2160: with solution in the wall frame
2161: \begin{eqnarray}
2162: V_S & = & \frac{\rho_{1} W_{1} V_{1}}{\rho_{2} - \rho_{1}W_{1}} , \\
2163: P_{2} & = & \rho_{2} (\Gamma - 1)(W_{1} - 1) , \\
2164: \rho_{2} & = & \rho_{1} \left[ \frac{\Gamma + 1}{\Gamma - 1} +
2165: \frac{\Gamma}{\Gamma - 1}(W_{1} - 1) \right] ,
2166: \end{eqnarray}
2167: where $V_S$ is the velocity of the shock front, and the pre-shocked
2168: energy and post-shocked velocity were both assumed
2169: negligible ($e = V_2 = 0$).
2170:
2171: The initial data are set up to be
2172: uniform across the grid with adiabatic index $\Gamma=4/3$,
2173: pre-shocked density $\rho_{1} = 1$,
2174: and pre-shocked pressure $P_{1} = 10^{-6}$.
2175: And we change the velocity of the unshocked region parametrically.
2176: The result is shown in the right panels of Figure \ref{fig:rst}.
2177: For all computations, the differences
2178: between the numerical solution and the analytic one are small,
2179: however in case of relativistic one,
2180: pressure of the numerical solution is bigger than
2181: the analytic one.
2182: It is also due to the artificial viscosity assumed here.
2183:
2184: \begin{figure}[ht]
2185: \begin{center}
2186: \includegraphics[width=.99\linewidth]{fig13.eps}
2187: \caption[Shock Tubes and Wall Reflection]{
2188: Left and right panel show the shock tube and the wall reflection tests,
2189: respectively.
2190: From top to bottom, the Lorentz factor becomes larger
2191: (for right panels, top: $v=0.1c$ ($W=1.01$),
2192: middle: $v=0.5c$($W=1.33$), and bottom:$v=0.9c$($W=5.26$)).
2193: Note that the pressure and density is normalized by $100$.
2194: %The top left panels is the case of $v=0.5c$($W=1.33$).
2195: %The top left panels is the case of $v=0.9c$($W=5.26$).
2196: %The top panel shows the lowly relativistic shock tube,
2197: % the middle does mildly relativistic one and
2198: %the bottom does highly relativistic one.
2199: %Right Panels: Wall Reflections.
2200: %In this figure the results of wall reflection problem
2201: %are presented.
2202: %The density is normalized by $100$ in all panels.
2203: %The top left panels is the case of $v=0.1c$($W=1.01$).
2204: %Its pressure is normalized by $100$.
2205: %The top left panels is the case of $v=0.5c$($W=1.33$).
2206: %The top left panels is the case of $v=0.9c$($W=5.26$).
2207: }\label{fig:rst}
2208: \end{center}
2209: \end{figure}
2210: \clearpage
2211:
2212: \paragraph{Magnetic Shock Tubes}
2213: At last we present magnetic shock tube problems \citep{brio88}.
2214: We show the initial condition and the results of the computations in
2215: Table \ref{tab:bw}.
2216: And we present mildly relativistic case in Figure \ref{fig:mrmst}, showing
2217: that our code can handle the various magneto-sonic waves as good as
2218: the code by \citet{devi03}.
2219: %The obtained results show qualitatively good agreements with
2220: %those of \citet{devi03}.
2221:
2222: \begin{figure}[ht]
2223: \begin{center}
2224: \includegraphics[width=.99\linewidth]{fig14.eps}
2225: \caption{Mildly relativistic magnetic shock tubes}\label{fig:mrmst}
2226: \end{center}
2227: \end{figure}
2228:
2229:
2230: \begin{table}[ht]
2231: \caption[Initial and intermediate states for
2232: shock tube tests.]{Initial and intermediate states for
2233: shock tube tests. This table list the measured values
2234: in each state of shock tube: ``Left'' is the initial left
2235: state for the given variables,
2236: ``FR'' is the value of the variable
2237: at the foot of leftmost fast rarefaction fan,
2238: ``SC'' is the value of the peak
2239: of the slow compound wave,
2240: ``$\mathrm{CD_l}$'' is the value of left
2241: of the constant discontinuity,
2242: ``$\mathrm{CD_r}$'' is the value of right
2243: of the constant discontinuity,
2244: ``FR'' is the value of the variable
2245: at the foot of the second fast rarefaction fan and
2246: ``Right'' is the initial right
2247: state for the given variables,
2248: }\label{tab:bw}
2249:
2250: \begin{tabular}{llrrrrrrr}
2251: & & & & & & & & \\
2252: \hline
2253: Test & Variable & Left & FR & SC & $\mathrm{CD_l}$ & $\mathrm{CD_r}$ & FR & Right \\
2254: \hline
2255: \hline
2256: Newtonian & $\rho$& 1.00 & 0.66 & 0.84 & 0.70 & 0.25 & 0.12 & 0.13 \\
2257: $B^x(\times 10^{-2})$ & $P (\times {10}^{-4})$
2258: & 1.00 & 0.44 & 0.73 & 0.50 & 0.50 & 0.09 & 0.10 \\
2259: $ =0.75$ & $v^x (\times {10}^{-3})$
2260: & 0.00 & 6.67 & 4.6 & 5.99 & 6.02 & -2.79 & 0.00 \\
2261: & $v^y (\times {10}^{-2})$
2262: & 0.00 &-0.25 &-1.10 &-1.58 &-1.58 &-0.20 & 0.00 \\
2263: & $\frac{B^y}{\sqrt{4\pi}} (\times {10}^{-2})$
2264: & 1.00 &0.6 &-0.5 &-0.03 &-0.54 &-0.9 &-1.00 \\
2265: \hline
2266: Mildly & $\rho$& 1.00 & 0.65 & 0.84 & 0.70 & 0.24 & 0.11 & 0.13 \\
2267: Relativistic & $P (\times {10}^{-2})$
2268: & 1.00 & 0.42 & 0.71 & 0.49 & 0.49 & 0.08 & 0.10 \\
2269: $B^x(\times 10^{-1})$ & $v^x(\times {10}^{-1})$ & 0.00 & 0.67 & 0.46 & 0.58 & 0.58 &-0.26 & 0.00 \\
2270: $=0.75$ & $v^y(\times {10}^{-1})$
2271: & 0.00 &-0.24 &-0.94 &-1.5 &-1.5 &-1.9 & 0.00 \\
2272: & $\frac{B^y}{\sqrt{4\pi}}(\times {10}^{-1})$& 1.00 &0.56 &0.3 &-0.52 &-0.52 &-0.88 &-1.00 \\
2273: \hline
2274: Relativistic & $\rho$& 1.00 & 0.59 & 0.70 & 0.65 & 0.31 & 0.11 & 0.13 \\
2275: $B^x$ & $P$ & 1.00 & 0.51 & 0.60 & 0.47 & 0.47 & 0.08 & 0.10 \\
2276: $=0.75$ & $v^x$ & 0.00 & 0.41 & 0.27 & 0.28 & 0.28 &-0.12 & 0.00 \\
2277: & $v^y$ & 0.00 &-0.07 &-0.62 &-0.58 &-0.11 &-0.11 & 0.00 \\
2278: & $\frac{B^y}{\sqrt{4\pi}}$& 1.00 &0.61 &0.19 &-0.24 &-0.24 &-0.81 &-1.00 \\
2279: \hline
2280: \hline
2281: \end{tabular}
2282: \end{table}
2283:
2284:
2285: \end{document}
2286: