0804.3346/ms.tex
1: \documentclass[useAMS,usenatbib]{mn2e}
2: \usepackage{epsfig}
3: \newcommand{\bm}[1]{\mbox{\boldmath $#1$}}
4: %%%%%%%%%%%%%%%%%
5: % (1)TITLE PAGE %
6: %%%%%%%%%%%%%%%%%
7: 
8: \title[The Impact of a Supernova Explosion in a Very Massive Binary]{The Impact of a Supernova Explosion in a Very Massive Binary}
9: 
10: \author[J. Sato, M. Umemura and K. Sawada]{Jun'ichi Sato$^{1}$\thanks{Current address: Service d'Astrophysique, DSM/IRFU, UMR AIM, CEA-CNRS-Univ. Paris VII, Saclay, France;junichi.sato@ces.fr (JS)}, 
11: Masayuki Umemura$^{1}$ and Keisuke Sawada$^{2}$\\
12: $^{1}$Center for Computational Sciences, University of Tsukuba, Tsukuba 305-8577, Japan\\
13: $^{2}$Department of Aeronautics and Space Engineering, Tohoku University, Sendai 980-8579, Japan}
14: 
15: \begin{document}
16: 
17: \date{Accepted ........ Received ........ ; in original form ........}
18: 
19: \pagerange{\pageref{firstpage}--\pageref{lastpage}} \pubyear{2008}
20: 
21: \maketitle
22: 
23: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
24: % (2)Abstract  & Subject Headings %
25: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
26: 
27: \begin{abstract}
28: We consider the effect of a supernova (SN) explosion 
29: in a very massive binary that is expected to form in a portion
30: of Population III stars with the mass higher than 100$M_\odot$. 
31: In a Population III binary system, a more massive star can
32: result in the formation of a BH and a surrounding accretion disc.
33: Such BH accretion could be a significant source of the cosmic
34: reionization in the early universe. 
35: However, a less massive companion star evolves belatedly
36: and eventually undergoes a SN explosion, so that the accretion disc
37: around a BH might be blown off in a lifetime of companion star. 
38: In this paper, we explore the dynamical impact of a SN explosion on 
39: an accretion disc around a massive BH, and elucidate
40: whether the BH accretion disc is totally demolished or not. 
41: For the purpose, we perform three-dimensional hydrodynamic 
42: simulations of a very massive binary system, where we assume
43: a BH of $10^3 M_{\odot}$ that results from a direct collapse of a very
44: massive star and a companion star of $100 M_{\odot}$ that 
45: undergoes a SN explosion. 
46: We calculate the remaining mass of a BH accretion disc as a function
47: of time. As a result, it is found that a significant portion of gas disc
48: can survive through three-dimensional geometrical effects 
49: even after the SN explosion of a companion star. 
50: Even if the SN explosion energy is higher by two orders of magnitude
51: than the binding energy of gas disc, about a half of disc can be left over. 
52: The results imply that the Population III BH accretion disc 
53: can be a long-lived luminous source, and therefore
54: could be an important ionizing source in the early universe. 
55: \end{abstract}
56: \begin{keywords} 
57: accretion discs --- PopIII stars --- early universe  
58: --- hydrodynamics --- binaries: general --- cosmology: theory 
59: \end{keywords}
60: 
61: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%
62: %(3)TEXT & Acknowledgements %
63: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%
64: 
65: \section{Introduction}
66: \label{Sec1}
67: The studies on the formation of Population III (Pop III) stars
68: have shown that first stars in the universe are likely to 
69: be as massive as $100-1000M_\odot$ 
70: \citep{BCL99, Abel00,Abel02} or to
71: form in a bimodal initial mass function (IMF) with peaks of several $100M_\odot$ 
72: and $\sim 1M_\odot$ \citep{NU01}. 
73: Recently, it has been revealed that 
74: a significant fraction of Pop III stars can be expected 
75: to form in binary systems \citep{Saigo+04}. 
76: \citet{HW02} have studied the nucleosynthetic evolution of Pop III
77: stars, and have shown that Pop III stars in the mass range of 
78: $m \ga 260 M_\odot$
79: result in direct black hole formation, while the mass range of
80: $25 M_\odot \la m \la 140 M_\odot$ leads to the supernova (SN) explosions.
81: Hence, if there is mass difference in a Pop III binary, 
82: there can form a binary system composed of a BH with $m \ga 260 M_\odot$
83: and a massive star with $25 M_\odot \la m \la 140 M_\odot$.
84: Also, recent general relativistic simulations on a supermassive star 
85: ($m \ga 10^3 M_\odot$)
86: have shown that if a star possesses 
87: large angular momentum, a massive gas disc is left over
88: around a forming black hole 
89: \citep{SS02, Shibata04, Shapiro04, SS04}. 
90: The mass fraction of gas disc is $10^{-4} - 10^{-1}$ 
91: of the initial stellar mass,
92: depending on the equation of state and the degree of rotation. 
93: This disc can fuel the BH in the later evolutionary stage.
94: Even if a disc left over is quite small, the Roche lobe overflow from
95: a companion star may also fuel the BH in a close binary system
96: \citep{Lawlor08}.
97: 
98: On the other hand, it is suggested that 
99: Pop III BH accretion may be an important clue 
100: for the reionization of the universe \citep{mad04,RO04b}. 
101: For a BH of $\approx 10^3 M_{\odot}$, 
102: the accretion disc can emit a blackbody radiation with 
103: an effective temperature of $10^{6-7} \mbox{ K}$,
104: if the mass accretion rate is close to the Eddigton rate 
105: (e.g. \citet{Kato+98}).
106: When the disc mass is several $10 M_{\odot}$,
107: the accretion timescale is several $10^{6} \mbox{ years}$. 
108: Therefore, the accreting BH can be an ultraviolet radiation source
109: in a longer timescale than the lifetime of a massive companion star. 
110: Then, the accreting BH can be a significant contributor 
111: for the early reionization\citep{HU07}. 
112: However, if the accretion disc is blown off by the SN explosion 
113: of a companion star, the accretion timescale may be too short 
114: to play a significant role for the reionization. 
115: 
116: In this paper, we explore the disruption and stripping of a gas disc
117: around a BH by a SN explosion of a massive companion star.
118: For the purpose, we perform three-dimensional hydrodynamic simulations
119: with a second-order scheme, AUSM-DV. In particular, to resolve the propagation
120: of a blast wave and the deformation of a gas disc with high accuracy,
121: three-dimensional generalized curvilinear coordinates are used. 
122: Such a three-dimensional simulation on the collision of SN blast wave 
123: with a BH accretion disc has been never performed so far. 
124: With the simulations, we study the stripping efficiency as a function of 
125: the kinetic energy of the SN blast wave, 
126: by changing the separation of binary and the mass of gas disc.
127: Then, the dependence on the ratio of the blast wave kinetic energy 
128: to the binding energy of gas disc is analysed in detail.
129: 
130: The paper is organized as follows. 
131: In Section 2, the details of present numerical model and 
132: method are presented. 
133: In Section 3, we show the numerical results for the deformation
134: and stripping of gas disc around a BH by the collision with
135: a SN blast wave. Also, the dependence on the model parameters
136: is studied, and the resultant remaining mass fraction of disc
137: is argued. Section 4 is devoted to conclusions.
138: 
139: 
140: \section{Numerical Model and Method}
141: \label{Sec2}
142: 
143: \subsection{Model}
144: \label{Sec2-1}
145: 
146: We consider a binary system composed of a black hole 
147: of $1000 M_{\odot}$ and a Pop III companion star of $100 M_{\odot}$. 
148: In the present analyses, the model parameters are the separation
149: between the BH and the companion star, 
150: the mass of gas disc surrounding the BH, and the density profile of gas disc. 
151: Based on the analysis by \citet{Saigo+04}, 
152: we set the binary separation, $a$, to be $700 \mbox{ AU}$ as a fiducial case, 
153: and a half or one-fifth of the fiducial separation is also examined. 
154: As for the mass of gas disc, considering the disc formation 
155: around a black hole \citep{SS02, Shibata04, Shapiro04, SS04}
156: and also the mass loss from a Pop III star before a SN explosion
157: \citep{HW02}, we set the disc mass to be $30 M_{\odot}$, 
158: $10 M_{\odot}$ or $3 M_{\odot}$.
159: Regarding the density distribution of gas disc, 
160: we assume a power-law type distribution like $\rho(r) \propto r^n$
161: on the equatorial plane. 
162: The power-law index $n$ is assumed to be $n=0$, $-1$, or $-2$. 
163: The temperature of gas disc is presumed to be $10^{4} \mbox{K}$, 
164: since the gas disc is irradiated by ultraviolet radiation from a companion
165: star before the explosion. 
166: The rotation of the disc is assumed to be Keplerian, because the self-gravity
167: is negligible in all the cases considered. 
168: In vertical directions of the disc, we determine the density distributions by
169: solving a hydrostatic balance. 
170: Also, taking the tidal effect into account, the gas disc is truncated at the 
171: tidal radius given by \cite{Paczynski77} and \cite{Boffin01}. 
172: The tidal radius is $\sim 0.7$ Roche lobe radius of the BH. 
173: For instance, the outer radius of the gas disc is set to be $280 \mbox{AU}$ 
174: for a fiducial case. 
175: The space out of the gas disc is filled with tenuous and high temperature gas,
176: since we cannot treat a vacuum in the present hydrodynamic scheme.
177: In order to minimize its dynamical effect, 
178: we set the density to be negligibly small.
179: Here, the density and temperature are assumed to be $2.5 \times 10^{-16} 
180: \mbox{g cm}^{-3}$ and $10^{7} \mbox{K}$, respectively.
181: The parameters adopted are summarized in Table \ref{tab1}. 
182: 
183: To simulate a SN explosion, we employ a method similar to
184: that taken by \citet{KY05}, and \citet{BYH03}, where the explosion energy
185: is inserted as the thermal energy of the gas ejected from a SN. 
186: The ejected gas has no initial velocity 
187: and is accelerated by the gas pressure. 
188: The temperature is assumed to be $1.0 \times 10^{8} \mbox{K}$ 
189: for all cases, and the density and extent of hot gas 
190: is adjusted so that the explosion energy is $10^{51} \mbox{erg}$ 
191: \citep{HW02}.
192: If the separation is shorter, 
193: the density of hot gas is higher and the extent is smaller.
194: For model A1, the density of the hot gas is set to be 
195: $\rho_{\rm SN}=1.0 \times 10^{-11} \mbox{g cm}^{-3}$ and
196: the extent is to be $50 \mbox{AU}$. 
197: We tested the cases of different density and extent 
198: and confirmed that the impact of a SN explosion on the gas disc 
199: does not have a strong dependence on the choices, but
200: is basically determined by the explosion energy.
201: Also, we regulate the duration of SN explosion 
202: so that the total mass ejected by SN explosion is
203: $50 M_{\odot}$ \citep{HW02}. 
204: 
205: \subsection{Numerical Method}
206: \label{Sec2-2}
207: We solve the three-dimensional Euler equation for inviscid gas.
208: Basic equations can be written in the conservative form as 
209: \begin{equation}
210: \frac{\partial \tilde{Q}}{\partial t}+
211: \frac{\partial \tilde{E}}{\partial x}+
212: \frac{\partial \tilde{F}}{\partial y}+
213: \frac{\partial \tilde{G}}{\partial z}+
214: \tilde{H}=0, 
215: \label{eq-1}
216: \end{equation}
217: where $\tilde{Q}$, $\tilde{E}$, $\tilde{F}$, $\tilde{G}$ and $\tilde{H}$ 
218: are respectively the following vectors, 
219: \begin{eqnarray}
220: &&
221: \tilde{Q}=
222: \left(
223: \begin{array}{c}
224: \rho \\
225: \rho u \\
226: \rho v \\
227: \rho w \\
228: e
229: \end{array}
230: \right)
231: ,\quad
232: \tilde{E}=
233: \left(
234: \begin{array}{c}
235: \rho u \\
236: \rho u^{2}+p \\
237: \rho uv \\
238: \rho uw \\
239: (e+p)u
240: \end{array}
241: \right)
242: ,\quad
243: \tilde{F}=
244: \left(
245: \begin{array}{c}
246: \rho v \\
247: \rho uv \\
248: \rho v^{2}+p \\
249: \rho vw \\
250: (e+p)v
251: \end{array}
252: \right)
253: ,
254: \nonumber \\
255: &&
256: \tilde{G}=
257: \left(
258: \begin{array}{c}
259: \rho w \\
260: \rho uw \\
261: \rho vw \\
262: \rho w^{2}+p \\
263: (e+p)w
264: \end{array}
265: \right)
266: ,\quad
267: \tilde{H}=
268: \left(
269: \begin{array}{c}
270: 0 \\
271: \rho f_{x} \\
272: \rho f_{y} \\
273: \rho f_{z} \\
274: \rho \left( uf_{x}+vf_{y}+wf_{z} \right)
275: \end{array}
276: \right),
277: \label{eq-2}
278: \end{eqnarray}
279: (e.g, \citealt{Vinokur74}), 
280: where $\rho$ is the density of gas, $u$, $v$ and $w$ are 
281: respectively the $x$-, $y$- and $z$- components of velocities, 
282: and $f_{x}$, $f_{y}$ and $f_{z}$ are 
283: the $x$-, $y$- and $z$- components of gravity force by a black hole. 
284: As the gravity force, we take into account only the effect of black hole,
285: since it is a major component. 
286: The pressure $p$ and the total energy per unit volume $e$ 
287: are related by the equation of state, 
288: \begin{equation}
289: p=\left(\gamma-1\right)
290: \left\{e-\frac{\rho}{2}\left(u^{2}+v^{2}+w^{2}\right)\right\},
291: \label{eq-3}
292: \end{equation}
293: where $\gamma$ is the ratio of specific heats of gas, 
294: which is $\gamma=5/3$ here. 
295: 
296: The Reynolds number based on the molecular viscosity is very
297: large for the present case. Therefore, we have neglected the
298: effect of viscosity (and only considered the effect of ram
299: pressure), i.e. solved the Euler equations. However, we
300: should note that possible turbulence would provide a tendency
301: to reduce the remaining disc mass or to extinguish the disc.
302: Solving such case accounting for turbulent eddy viscosity
303: will be our future work
304: \citep{Takeda+85}. 
305: 
306: To solve the governing equation (\ref{eq-1}) numerically,
307: AUSM-DV scheme \citep{WL94} is employed. 
308: To properly adjust the boundary and solve the flow with high accuracy,
309: we use the three-dimensional generalized curvilinear coordinates as
310: shown in Fig. \ref{Fig1}. 
311: Because we assume the system is in a plane-symmetric with respect 
312: to the equatorial plane, we perform simulations in the computational domain 
313: that is composed of a hemisphere and an adjacent protuberant region.
314: The BH and the SN explosion are placed on the equatorial plane,
315: where the BH is located at the center of hemisphere, and 
316: the SN explosion occurs at the center of protuberant region. 
317: In Fig. \ref{Fig1}, only a half of the computational domain is 
318: shown after the domain is cut by a plane perpendicular to 
319: the equatorial plane that connects the BH and the SN explosion. 
320: The number of grid points is $51 \times 101 \times 31$. 
321: We treat a black hole as a small central hole with absorbing boundary, 
322: whose radius is $20 \mbox{AU}$ for the case of $a=700 \mbox{AU}$. 
323: The gas disc is set up around the small hole, and
324: the inner radius of the disc is assumed to be $50 \mbox{ AU}$. 
325: The grids are generated with higher resolution near the black hole 
326: and the equatorial plane of the gas disc. 
327: Such coordinates allow us to treat accurately the flow of the gas disc. 
328: The computations are carried out 
329: only in the upper half domain relative to the equatorial plane, 
330: because we assume that the system is symmetric about this plane. 
331: The leftward meshes are built up to represent a SN explosion. 
332: A free boundary condition on outer and inner boundaries is adopted
333: in these meshes. 
334: 
335: \section{Results}
336: \label{Sec3}
337: 
338: When a SN explosion produces a supersonic flow with the mass of 
339: $50 M_{\odot}$ and the velocity of $4000 \mbox{ km}/\mbox{sec}$, 
340: the kinetic energy of the blast wave is $\sim 1.1 \times 10^{50} \mbox{ erg}$
341: for the solid angle of a disc viewed from the SN explosion.
342: In the fiducial case, the binding energy of gas disc is evaluated 
343: to be $\sim 3.1 \times 10^{48} \mbox{ erg}$.
344: Hence, the kinetic energy of a SN blast wave is much larger than the disc binding 
345: energy. Then, the collision of the blast wave with a disc 
346: is anticipated to be devastating. 
347: However, the disc is deformed by the collision, and also streamlines are bent 
348: significantly. Therefore, it is unclear whether the whole of disc gas is 
349: stripped out by the collision of the SN blast wave.
350: 
351: 
352: \subsection{Disc Deformation and Stripping}
353: \label{deformation}
354: 
355: Figure \ref{Fig2} shows the numerical results at early stages for model A1.
356: The density distributions and velocity fields are shown in the $x-y$ plane 
357: with $z=0$ ({\it lower panels}) and in $x-z$ plane with
358: $y=0$ ({\it upper panels}). 
359: Figure \ref{Fig2}(a) shows the initial state, where 
360: high-density hot gas is ejected at $700 \mbox{AU}$ from a BH. 
361: The hot gas expands almost radially (Fig. \ref{Fig2}(b))
362: and rushes toward the gas disc (Fig. \ref{Fig2}(c)).
363: At $\sim 0.47 \mbox{ years}$, as seen in Fig. \ref{Fig2}(d),
364: the bulk of flow is bent to go through the disc. 
365: In this stage, some of gas disc is stripped out, but
366: much of the blast wave energy can escape owing to the bending 
367: of gas flow. On the other hand, 
368: the flow directed to the disc center is strongly
369: decelerated by the collision with the disc, and begins to
370: deform the disc (Fig. \ref{Fig2}(e)(f)).
371: 
372: Figure \ref{Fig3} shows the deformation and stripping of the disc 
373: at later stages. The figure zooms in the disc regions. 
374: The rotation period at the outer edge of gas disc is 
375: $\sim 148 \mbox{ years}$. 
376: Fig. \ref{Fig3} (a) is the initial state again. 
377: As seen in Figures \ref{Fig3} (b) and (c),
378: the blast wave pushes the left-side edge of gas disc and produces a dent
379: within a few years, because the time-scale of the SN explosion is 
380: much shorter than the rotation period of the outer edge of gas disc. 
381: The rotation of disc transfers this dent, resulting in the overall deformation 
382: of the disc (Fig. \ref{Fig3} (d)). 
383: Through this deformation, the rotation balance breaks down,
384: and resultantly some of gas is blow out by centrifugal force,
385: as seen in Figure \ref{Fig3} (e).
386: Eventually, the gas disc is settled in a quasi-steady state 
387: within about $100 \mbox{ years}$ (Fig. \ref{Fig3} (f)).
388: As a result, roughly 70\% of an original disc around the BH is left over.
389: In Figure \ref{Fig4}, the deformation and stripping processes
390: are shown with three-dimensional volume rendering visualization.
391: The survival of the gas disc after the heaving and ruffling 
392: by the blast wave collision is clearly shown. 
393: 
394: Previously, \citet{Wheeler+75} argued that in a binary system 
395: the mass stripping and ablation from a star by the impinging balastwave 
396: is well expressed by a non-dimensional parameter $\Psi$, which is defined as 
397: \begin{equation}
398: \Psi \equiv \frac{1}{4}\frac{M_{\rm SN}}{M_{\rm c}}
399: \frac{R^{2}}{a^{2}}\left(
400: \frac{v_{\rm SN}}{v_{\rm es}}-1
401: \right), 
402: \end{equation}
403: where $M_{\rm SN}$ is the mass of gas expelled by a SN, 
404: $M_{\rm c}$ and $R$ are respectively the mass and radius 
405: of object affected by blast wave, 
406: $v_{\rm SN}$ is the typical speed of the blast wave,  
407: $v_{\rm es}$ is the escape velocity. 
408: They obtained the mass fraction ejected by the stripping and ablation
409: as a function of $\Psi$. 
410: We can evaluate $\Psi$ in the present system, where
411: $M_{\rm SN}=50 M_{\odot}$, $M_{\rm c}=M_{\rm disc}=30 M_{\odot}$, 
412: $R=280 \mbox{AU}$, $a=700 \mbox{AU}$ 
413: and $v_{\rm SN}=4000 \mbox{ km}/\mbox{sec}$. 
414: $v_{\rm es}$ is approximated to be
415: \begin{equation}
416: v_{\rm es}=\left(\frac{2 G M_{\rm BH}}{R}\right)^{1/2},
417: \end{equation}
418: where $M_{\rm BH}=1000 M_{\odot}$. 
419: Then, we find $\Psi \simeq 3.28$ in our model A1. 
420: According to the criterion for a polytropic star with $n=3$ given
421: by \citet{Wheeler+75}, it is predicted that about a half of mass is 
422: ejected in the case of $\Psi \simeq 3$. 
423: In the present simulation, about 30\% of mass is ejected eventually.
424: Therefore, we can conclude that $\Psi$ is a fairly good measure for
425: the mass ejection even in a disk system. However, in the disk system,
426: a little more mass remains than the prediction for a spherical star.
427: In the collision of blast wave with a gas disk, the streamline is
428: more strongly bent above and below the disc. Hence,
429: the momentum transferred to the gas in a disk is reduced, and
430: more mass can be left over. 
431: Therefore, the difference from the spherical prediction
432: can be reasonably understood in terms of the geometrical effect.
433: 
434: 
435: \subsection{Dependence on Binary Separation}
436: 
437: In order to investigate the dependence on the binary separation,
438: we perform additional two simulations with $a=350 \mbox{ AU}$ (model B1),
439: which is a half separation of model A1,
440: and with $a=140 \mbox{ AU}$ (model B2), which is a one-fifth 
441: separation of model A1. 
442: We assume the mass of the disc to be the same as model A1. 
443: Therefore, the disc radius becomes smaller
444: and the density is higher for these models. 
445: Also, the energy of the SN explosion, 
446: the temperature, and the total mass of ejected gas 
447: to be the same as model A1.
448: The rotation period of the outer edge of gas disc for model B1 and B2 is 
449: $\sim 52 \mbox{ years}$ and $\sim 13 \mbox{ years}$, respectively. 
450: Figure \ref{Fig5} shows the density distributions and the velocity fields
451: in the final quasi-steady state, which is reached 
452: at $t=38.19 \mbox{ years}$ and $11.95 \mbox{ years}$ 
453: for models B1 and B2, respectively. 
454: Hence, the time to reach the final state is shorter 
455: for the shorter separation. 
456: Interestingly, it is found that more mass is left over for
457: shorter binary separation. Roughly 80 \% of disc survives for
458: model B1, and nearly 85 \% of disc does for model B2.
459: This can be understood in terms of the binding energy $E_b$
460: of the disc relative to the kinetic energy $E_k$ of 
461: the SN blast wave. The ratio $E_b/E_k$ is $\sim 0.1$ and
462: $\sim 0.2$ for models B1 and B2, respectively, while
463: $E_b/E_k \sim 0.03$ for model A. Thus, $E_b/E_k$ is
464: thought to be a key physical quantity that determines
465: the remaining mass of disc. This point is argued again later.
466: 
467: 
468: \subsection{Dependence on Density Distribution}
469: 
470: To check whether the deformation and stripping of disc depends on 
471: the density distribution of gas disc, simulations for discs 
472: with different density distributions are performed.
473: One is a flat density distribution model with $n=0$ on the equatorial plane 
474: (model C1), and the other is a centrally concentrated model with
475: $n=-2$ (model C2). 
476: The mass of gas disc is the same as model A1. 
477: Figure \ref{Fig6} shows the results for a flat density distribution disc
478:  (model C1).
479: The snapshots of density distributions and velocity fields 
480: at $t=0.00, 10.32, \mbox{ and } 110.16 \mbox{ years}$ are presented. 
481: The results look similar those shown in Figure \ref{Fig3} at 
482: the same evolutionary stages. Actually, the remaining mass is not
483: changed significantly. 
484: Figure \ref{Fig7} show the results for a centrally-concentrated disc (model C2).
485: The snapshots of density distributions and velocity fields 
486: at $t=0.00, 10.37, \mbox{ and } 105.69 \mbox{ years}$ 
487: are presented. 
488: It is found that, for model C2, the influence by the blast wave is
489: slightly weaker in the central regions of the disc (Fig. \ref{Fig7} (b)), 
490: compared to model C1 (Fig. \ref{Fig6} (b)). 
491: But, the remaining mass increases by only several \%. 
492: Therefore, we can conclude that the density distributions of disc
493: do not influence the final mass of disc strongly. 
494: 
495: 
496: \subsection{Remaining Mass}
497: 
498: In Figure \ref{Fig8}, the time sequences of the disc mass 
499: are summarized by the ratio of the remaining disc mass 
500: to the disc mass without a supernova explosion of companion star. 
501: For model A1, C1 and C2, the disc is settled in quasi-steady states 
502: in $60$ years,
503: and $\sim 70 \%$ of disc gas remains, almost regardless of density 
504: distributions of disc. 
505: For model B1 ($a=350 {\rm AU}$), a quasi-steady state is attained 
506: in $20$ years and
507: $\sim 80 \%$ of disc gas remains.
508: For model B2 ($a=140 {\rm AU}$), the disc reaches 
509: a quasi-steady state in $10$ years 
510: and $\sim 85 \%$ of disc gas remains.
511: 
512: As argued in Section \ref{deformation}, the remaining amount of gas 
513: can be understood in terms of the ratio of $E_{b}$ to $E_{k}$, 
514: where $E_{b}$ is the binding energy of the gas disc around the BH and  
515: $E_{k}$ is the kinetic energy of the blast wave 
516: taking into account the solid angle of the disc viewed 
517: from the SN explosion point. 
518: $E_{b}$ is evaluated by integrating potential energy 
519: of the gas around the BH.
520: Figure \ref{Fig9} shows the remaining mass fraction against
521: the $E_{b}/E_{k}$ ratio. 
522: In this diagram, two more simulations are added with changing 
523: the disc mass: model D1 has the disc mass of $10 M_{\odot}$, 
524: which is one-third of that in model A, and model D2 has 
525: the disc mass of $3 M_{\odot}$, which is one-tenth of that in model A.
526: Obviously, the remaining mass fraction correlates positively with 
527: $E_{b}/E_{k}$. 
528: Thus, a key physical parameter that determines the remaining
529: mass fraction is the $E_{b}/E_{k}$ ratio.
530: Interestingly, even if the binding energy is less than 1\% of the 
531: kinetic energy of SN blast wave, a considerable amount of mass
532: remains. This is a three-dimensional effect of deformation and
533: stripping, which is demonstrated in Figure \ref{Fig4}.
534: 
535: 
536: \section{Conclusions}
537: \label{Sec4}
538: 
539: We have explored the dynamical impact of SN explosion in a very massive binary
540: system. We have simulated the deformation and stripping of a gas disc
541: around a BH, using three-dimensional generalized curvilinear coordinates.
542: Also, the dependence on the binary separation, the density distribution of disc,
543: and the mass of disc have been studied. 
544: As a result, we have found that a SN blast wave is not so devastating that
545: a disc around a BH is totally evaporated. 
546: It has turned out that the remaining mass fraction is basically determined
547: by the ratio of the disc binding energy to the SN kinetic energy. 
548: As a matter of importance, even if the binding energy of gas disc 
549: is much smaller than the kinetic energy of blast wave, 
550: a significant amount of mass can remain. For instance,
551: when the binding energy is about 1\% of the 
552: kinetic energy of SN blast wave, about a half of mass is left over.
553: We have found that the three dimensionality of the disc deformation is 
554: essential to make a disc around a BH survive. 
555: 
556: The present results imply that in a massive binary system like
557: a Pop III binary, the BH accretion activity can be long-lived 
558: even after a massive companion star explodes. 
559: Such BH accretion can emit ultraviolet radiation, and therefore
560: a Pop III BH accretion disc can be an important candidate 
561: as a reionization source in the early universe.
562: 
563: 
564: \section*{Acknowledgments}
565: We thank K. Hiroi for many useful comments. 
566: Numerical simulations have been performed with computational facilities 
567: at Center for Computational Sciences in University of Tsukuba. 
568: This work was supported in part by Grants-in-Aid, Specially
569: Promoted Research, 16002003, from MEXT in Japan.
570: 
571: 
572: %%%%%%%%%%%%%%%%%%
573: % (4) Appendices %
574: %%%%%%%%%%%%%%%%%%
575: %No Appendix
576: %%%%%%%%%%%%%%%%%
577: % (5)References %
578: %%%%%%%%%%%%%%%%%
579: \begin{thebibliography}{99}
580: \bibitem[\protect\citeauthoryear{Abel, Bryan, \& Norman}{2000}]{Abel00} 
581: Abel, T., Bryan, G. ~L., \& Norman, M.~L. 2000, ApJ, 540, 39
582: \bibitem[\protect\citeauthoryear{Abel, Bryan, \& Norman}{2002}]{Abel02} 
583: Abel, T., Bryan, G.~L., \& Norman, M.~L.\ 2002, Science, 295, 93 
584: \bibitem[\protect\citeauthoryear{Boffin}{2001}]{Boffin01} 
585: Boffin H. M. J., 2001, 
586: in Boffin H. M. J., Steeghs D., Cuypers J., eds, 
587: Lecture Notes on Physics 573, Astrotomography, Springer, Heidelberg, p. 69
588: \bibitem[\protect\citeauthoryear{Bromm, Coppi \& Larson}{1999}]{BCL99} 
589: Bromm, V., Coppi, P. S., \& Larson, R. B. 1999, ApJ, 527, L5
590: \bibitem[\protect\citeauthoryear{Bromm, Yoshida \& Hernquist}{2003}]{BYH03} 
591: Bromm, V., Yoshida, N., \& Hernquist, L. 2003, 596, L135
592: \bibitem[\protect\citeauthoryear{Heger \& Woosley}{2002}]{HW02}
593: Heger, A., \& Woosley, S. E., 2002, ApJ, 567, 532
594: \bibitem[\protect\citeauthoryear{Hirose \& Umemura}{2007}]{HU07} 
595: Hirose, Y. \& Umemura, M. 2007, in preparation.
596: \bibitem[\protect\citeauthoryear{Kato et al.}{1998}]{Kato+98}
597: Kato, S., Fukue, J., \& Mineshige, S., eds. 1998, Black Hole Accretion Discs, 
598: (Kyoto University Press, Kyoto)
599: \bibitem[\protect\citeauthoryear{Kitayama \& Yoshida}{2005}]{KY05} 
600: Kitayama, T. \& Yoshida, N. 2005, ApJ, 630, 675 
601: \bibitem[\protect\citeauthoryear{Lawlor et al.}{2008}]{Lawlor08} 
602: Lawlor T.~M., Young T.~R., Johnson T.~A., MacDonald J., 2008, MNRAS, 384, 
603: 1533 
604: \bibitem[\protect\citeauthoryear{Madau et al.}{2004}]{mad04} 
605: Madau, P., Rees, M.~J., 
606: Volonteri, M., Haardt, F., \& Oh, S.~P.\ 2004, ApJ, 604, 484 
607: %\bibitem[\protect\citeauthoryear{Nakahashi et al.}{1999}]{Nakahashi+99} 
608: %Nakahashi, K., Sharov, D., Kano, S., \& Kodera, M., 
609: %1999, Int. J. Numer. Meth. Fluids, 31, 97
610: \bibitem[\protect\citeauthoryear{Nakamura \& Umemura}{2001}]{NU01}
611: Nakamura, F., \& Umemura, M., 2001, ApJ, 548,19
612: \bibitem[\protect\citeauthoryear{Paczynski}{1977}]{Paczynski77}
613: Paczynski B., 1977, ApJ, 216, 822
614: \bibitem[\protect\citeauthoryear{Ricotti \& Ostriker}{2004a}]{RO04a} 
615: Ricotti, M. \& Ostriker, J.~P.\ 2004a, MNRAS, 350, 539 
616: \bibitem[\protect\citeauthoryear{Ricotti \& Ostriker}{2004b}]{RO04b} 
617: Ricotti, M. \& Ostriker, J.~P.\ 2004b, MNRAS, 352, 547
618: \bibitem[\protect\citeauthoryear{Saigo et al.}{2004}]{Saigo+04}
619: Saigo, K., Matsumoto, T., \& Umemuda, M. 2004, ApJ, 615, L65
620: \bibitem[\protect\citeauthoryear{Sekiguchi \& Shibata}{2004}]{SS04}
621: Sekiguchi, Y., \& Shibata, M. 2004, Phys. Rev. D, 70, 084005
622: \bibitem[\protect\citeauthoryear{Shapiro}{2004}]{Shapiro04}
623: Shapiro, S. L., 2004, ApJ. 610, 913
624: \bibitem[\protect\citeauthoryear{Shibata}{2004}]{Shibata04}
625: Shibata, M., 2004, ApJ. 605, 350
626: \bibitem[\protect\citeauthoryear{Shibata \& Shapiro}{2002}]{SS02}
627: Shibata, M., \& Shapiro, S. L. 2002, ApJ. 572, L39
628: \bibitem[\protect\citeauthoryear{Takeda et al.}{1985}]{Takeda+85} 
629: Takeda, H., Matsuda, T., Sawada K. \& Hayashi, C., 
630: 1985, Prog. Theor. Phys. 74, 272.
631: \bibitem[\protect\citeauthoryear{Vinokur}{1974}]{Vinokur74}
632: Vinokur, M. 1974, J. Comput. Phys., 14, 105.
633: \bibitem[\protect\citeauthoryear{Wada \& Liou}{1994}]{WL94} 
634: Wada, Y. \& Liou, M. S. 1994, AIAA Paper, 94-0083
635: \bibitem[\protect\citeauthoryear{Wheeler et al.}{1975}]{Wheeler+75} 
636: Wheeler, J. C., Lecar, M. \& McKee, C. F. 1975, 200, 145
637: \end{thebibliography}
638: %\end{document}
639: 
640: 
641: %%%%%%%%%%%%%%%%%%%%%%%
642: % (6) Figure Captions %
643: %%%%%%%%%%%%%%%%%%%%%%%
644: \setcounter{figure}{0}
645: \clearpage
646: %%%%%%%%%%%%%%%%%%%%%%%%% Table 1 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
647: \begin{table*}
648: \caption[]{Adopted parameters for model calculations. 
649: In this table, $a$ represents the binary separation, 
650: $M_{\rm disc}$ does the mass of gas disc, 
651: and $n$ does the power-law index of density distributions 
652: on the equatorial plane. 
653: }
654: \centering
655: \begin{tabular}{r|rrrl}
656: \hline
657: Model & $a$ [AU] & $M_{\rm disc}$ & $n$ & Remarks \\
658: \hline
659: A1 & $700$ & $30$ & $-1$ & Fiducial model \\
660: B1 & $350$ & $30$ & $-1$ & Half of the separation in model A1 \\
661: B2 & $140$ & $30$ & $-1$ & One-fifth of the separation in model A1 \\
662: C1 & $700$ & $30$ & $0$ & Flat density distribution disc \\
663: C2 & $700$ & $30$ & $-2$ & Centrally concentrated density distribution disc \\
664: D1 & $700$ & $10$ & $-1$ & One-third of the disc mass in model A1 \\
665: D2 & $700$ & $3$ & $-1$ & One-tenth of the disc mass in model A1 \\
666: \hline
667: \end{tabular}
668: \label{tab1}
669: \end{table*}
670: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
671: 
672: \begin{figure*}
673: \centerline{\psfig{figure=f1.eps,clip=ture,width=\textwidth}}
674: \caption[dummy]{
675: The three-dimensional generalized curvilinear coordinate system 
676: used in this paper. 
677: The bottom plane is the equatorial plane. 
678: The center of the gas disc, i.e. the BH, is placed at the center 
679: of a rightward hemisphere, and the SN explosion occurs at the center
680: of an adjacent protuberant region. A half of grids in the computational
681: domain are shown, after the whole domain is cut by a plane 
682: including the BH and the SN explosion point perpendicular 
683: to the equatorial plane. 
684: The number of grid points is $51 \times 101 \times 31$.
685: }
686: \label{Fig1}
687: \end{figure*}
688: 
689: 
690: \begin{figure*}
691: \centerline{\psfig{figure=f2.eps,clip=ture,width=\textwidth}}
692: \caption[dummy]{
693: Early stages of the interaction of a SN blast wave with a disc around a BH 
694: for model A1. 
695: Snapshots show density distributions and velocity fields in the $x-y$ plane 
696: with $z=0$ ({\it lower panels}), 
697: and in $x-z$ plane with $y=0$ ({\it upper panels}).
698: }\label{Fig2}
699: \end{figure*}
700: 
701: \begin{figure*}
702: \centerline{\psfig{figure=f3.eps,clip=ture,width=\textwidth}}
703: \caption[dummy]{
704: Later stages of the interaction of a SN blast wave with a disc around a BH 
705: for model A1, where
706: the deformation and stripping processes of the gas disc are shown.
707: The density distributions and the velocity fields are shown with
708: zooming in disc regions. 
709: }\label{Fig3}
710: \end{figure*}
711: 
712: \begin{figure*}
713: \centerline{\psfig{figure=f4.eps,clip=ture,width=\textwidth}}
714: \caption[dummy]{
715: 3D volume rendering visualization of density distributions 
716: with velocity fields for model A1. 
717: Figures (a), (b), (c), (d), (e), (f), (g), (h) and (i) 
718: represent the snapshots at $t=0.00, 0.63, 0.78, 2.90, 11.17, 
719: 35.72, 53.98, 72.98, \mbox{ and } 110.24$ yrs, respectively.
720: }\label{Fig4}
721: \end{figure*}
722: 
723: \begin{figure*}
724: \centerline{\psfig{figure=f5.eps,clip=ture,width=\textwidth}}
725: \caption[dummy]{
726: The density distributions and the velocity fields 
727: in the final state for shorter separation models; model B1 is a model
728: with $a=350 \mbox{ AU}$ ({\it left panel}) and model B2 is a model
729: with $a=140 \mbox{ AU}$ ({\it right panel}).  
730: The snapshots at $t=38.19 \mbox{ years}$ and $11.95 \mbox{ years}$ 
731: are presented for model B1 and B2, respectively. 
732: }\label{Fig5}
733: \end{figure*}
734: 
735: \begin{figure*}
736: \centerline{\psfig{figure=f6.eps,clip=ture,width=\textwidth}}
737: \caption[dummy]{
738: Same as Fig. \ref{Fig3} but for a flat density distribution model 
739: with $n=0$ (model C1). 
740: The snapshots at $t=0.00, 10.32, \mbox{ and } 110.16$ years are presented. 
741: }\label{Fig6}
742: \end{figure*}
743: 
744: \begin{figure*}
745: \centerline{\psfig{figure=f7.eps,clip=ture,width=\textwidth}}
746: \caption[dummy]{
747: Same as Fig. \ref{Fig3} but for a central-concentration 
748: density distribution model with $n=-2$ (model C2).
749: The snapshots at $t=0.00, 10.37, \mbox{ and } 105.69$ years are presented. 
750: }\label{Fig7}
751: \end{figure*}
752: 
753: \begin{figure*}
754: \centerline{\psfig{figure=f8.eps,clip=ture,width=\textwidth}}
755: \caption[dummy]{
756: The time variation of the remaining mass of a disc affected 
757: by SN explosion, where the ratio of remaining mass to the mass 
758: of an undisturbed disc is shown. 
759: The horizontal axis is time in units of years. 
760: A solid line, a dashed line, a dash-dotted line, a dotted line and 
761: a log-dashed line show model A1, B1, B2, C1 and C2, respectively. 
762: Note that the ratio for each case starts exactly from unity.
763: But, the strong inflow by supernova explosion causes the transient
764: rapid growth of ratio by about 10\% at the very early stage.
765: }
766: \label{Fig8}
767: \end{figure*}
768: 
769: \begin{figure*}
770: \centerline{\psfig{figure=f9.eps,clip=ture,width=\textwidth}}
771: \caption[dummy]{
772: The resultant remaining mass fraction is shown against 
773: the ratio of $E_{b}$ to $E_{k}$, 
774: where $E_{b}$ is the binding energy of the gas disc around a BH and  
775: $E_{k}$ is the kinetic energy of blast wave 
776: taking into account the solid angle of the disc viewed 
777: from the SN explosion point. 
778: The names near the symbols are model names.  
779: }
780: \label{Fig9}
781: \end{figure*}
782: 
783: \end{document}
784: