astro-ph0110263/ms.tex
1: \documentstyle[11pt,newpasp,twoside,epsf]{article}
2: \markboth{Izumi Hachisu}{APS Conf. Ser. Style}
3: \pagestyle{myheadings}
4: \nofiles
5: 
6: \def\edcomment#1{\iffalse\marginpar{\raggedright\sl#1\/}\else\relax\fi}
7: \marginparwidth 1.25in
8: \marginparsep .125in
9: \marginparpush .25in
10: \reversemarginpar
11: 
12: \begin{document}
13: \title{Evolution of binaries producing Type Ia supernovae,
14: luminous supersoft X-ray sources, and recurrent novae
15: } 
16: 
17: \author{Izumi Hachisu} 
18: 
19: \affil{College of Arts and Sciences, University of Tokyo, 
20: Meguro-ku, Tokyo 153-8902, Japan}
21: 
22: \begin{abstract}
23: We have been proposing two evolutionary paths to Type Ia supernovae
24: (SNe Ia), which are called the supersoft X-ray source (SSS) channel 
25: and the symbiotic channel, depending on the orbital separation 
26: just prior to an SN Ia explosion.
27: The essential difference of our treatment is inclusion of strong,
28: optically thick winds on mass-accreting white dwarfs (WDs) 
29: in the elementary processes of binary evolution when the mass 
30: accretion rate on to WDs exceeds a critical rate of 
31: $\dot M_{\rm cr} \sim 1 \times 10^{-6} M_\odot$ yr$^{-1}$.
32: Once optically thick winds begin to blow from the WD,
33: the binary can avoid forming a common envelope in some cases.  
34: We call this {\it accretion wind}.  So that the WDs are able 
35: to grow up to the Chandrasekhar mass and explode as an SN Ia, 
36: showing SSS or recurrent nova phenomena in the way to SNe Ia. 
37: Thus, the accretion wind process of WDs can open new channels 
38: to SNe Ia.  We have modeled the LMC supersoft source RX J0513.9-6951 
39: as an example of the systems in the accretion wind phase.
40: Further inclusions of two other elementary 
41: processes make the channels much wider; these are
42: the case BB mass transfer in the SSS channel 
43: and the strong orbital shrinkage during the superwind phase of the 
44: primary star in the symbiotic channel.  As a result, the estimated 
45: birth rate of SNe Ia via these two channels becomes compatible with 
46: the observation in our Galaxy.  Interestingly, the U Sco and T CrB 
47: subclasses of recurrent novae can be naturally understood as a part
48: of evolutionary stages in these two SSS and symbiotic channels 
49: to SNe Ia, respectively.   Thus we have a unified picture
50: of binary evolutions to SNe Ia, luminous SSS, and recurrent novae. 
51: \end{abstract}
52: 
53: \section{Introduction}
54:      Type Ia supernovae (SNe Ia) are one of the most luminous 
55: explosive events of stars.  Recently, SNe Ia have been used 
56: as good distance indicators that provide a promising tool 
57: for determining cosmological parameters (Perlmutter et al. 1999;
58: Riess et al. 1998) 
59: because of their almost uniform maximum luminosities.
60: However, one of the most important unresolved problems on SNe Ia is
61: that we do not know the exact progenitor systems of SNe Ia.
62: \par
63:     It is widely accepted that the exploding star itself is 
64: an accreting white dwarf (WD) in a binary
65: (Nomoto 1982; Nomoto,  Thielemann, \& Yokoi 1984).
66: However, the companion star (and thus the observed binary system)
67: is not known.  Several objects have ever been considered, 
68: which include merging double white dwarfs
69: (Iben \& Tutukov 1984; Webbink 1984),
70: recurrent novae (Starrfield, Sparks, \& Truran, 1985),
71: symbiotic stars (Munari \& Renzini 1992) etc 
72: (Livio 2000, for recent summary). 
73: \par
74:      Among these models, the systems of double white dwarfs 
75: (the double degenerate model) are not theoretically
76: supported; when two WDs merge, carbon ignites not at the center
77: but near the surface layer of the C+O WD; 
78: so that the C+O WD is converted into an O+Ne+Mg WD (Saio \& Nomoto 1998).  
79: If the total mass of the merger exceeds the Chandrasekhar mass,
80: the merged WD does not explode as an SN Ia but collapses to a neutron star 
81: (Nomoto \& Kondo 1991).  
82: 
83: \begin{figure}
84: %%%%%\plottwo{sn1afig.eps}{acc_map_z02.eps}
85: \plottwo{hachisu3_f1.eps}{hachisu3_f2.eps}
86: \caption{{\bf Left:} 
87: Optically thick winds blow from mass-accreting white dwarfs
88: when the mass transfer rate exceeds a critical 
89: rate, i.e., $\dot M_{\rm acc} > \dot M_{\rm cr}$.  Here, we assume 
90: that the white dwarf accretes mass from equatorial region and, 
91: at the same time, blows winds from polar regions.  {\bf Right:} 
92: Response of white dwarfs to mass accretion is illustrated
93: in the white dwarf mass and the mass accretion rate plane,
94: i.e., in the $M_{\rm WD}$-$\dot M_{\rm acc}$ plane.
95: Strong optically thick winds blow above the line of
96: $\dot M_{\rm acc} > \dot M_{\rm cr}$.  
97: There is no steady state burning below 
98: $\dot M_{\rm acc} < \dot M_{\rm std}$.  Instead, intermittent
99: shell flashes occur.  The envelope mass at which hydrogen ignites
100: to flash is also shown (taken from Fig. 9 of Nomoto 1982).
101: }
102: \end{figure}
103: 
104:      On the other hand, the single degenerate systems has become more
105: promising, in which a C+O WD gains mass from a non-degenerate 
106: companion star and grows its mass up to the Chandrasekhar mass.
107: Such promising progenitors include luminous supersoft X-ray sources
108: (SSS), recurrent novae (RNe), and symbiotic stars.   
109: However, the most important problem in these single degenerate systems 
110: is that we did not know the mechanism how to grow the mass of the WDs.
111: Instead, there are effective mass-losing mechanisms 
112: such as nova ejection 
113: and common envelope evolutions that get rid of the mass accumulated 
114: on the WD.  Here, I review and summarize our recent developments
115: of the mechanisms that enable to grow the mass of WDs in the 
116: single degenerate systems and give a unified picture of
117: evolutions of SSSs, recurrent novae, and SNe Ia. 
118: 
119: 
120: \section{Accretion wind evolution}
121:      First, we briefly explain how the mass of WDs in the single
122: degenerate systems grow toward the Chandrasekhar mass limit.
123: The reason that the standard binary evolution theory 
124: fails to explain the evolutionary path to SNe Ia and recurrent novae 
125: lies mainly in the theoretical prediction of a common envelope
126: formation and the ensuing spiral-in process, because these processes 
127: inhibit the growth of white dwarfs in binary systems.  As discussed
128: in many previous papers on binary evolution, 
129: it has been widely accepted that a hydrogen-rich envelope on 
130: a mass-accreting white dwarf expands to a red giant size 
131: when its mass accretion rate, 
132: $\dot M_{\rm acc}$, exceeds a critical limit, i.e.,
133: \begin{equation}
134: \dot M_{\rm cr} = 8.3 \times 10^{-7} 
135: ({{M_{\rm WD}} \over {M_\odot}} - 0.4) ~M_\odot 
136: \mbox{~yr}^{-1},
137: \end{equation}
138: (see, e.g., Fig. 9 of Nomoto 1982)
139: and easily forms a common envelope.
140: Once a common envelope is formed, two stars begin to spiral-in 
141: each other as a result of viscous drag, thus producing
142: a double degenerate system (Iben \& Tutukov 1984; Webbink 1984).
143: \par
144:      However, Hachisu, Kato, \& Nomoto (1996)
145: found that white dwarfs begin to blow 
146: strong winds ($v_{\rm wind} \sim 1000$ km s$^{-1}$ and
147: $\dot M_{\rm wind} > 1 \times 10^{-6} M_\odot$ yr$^{-1}$)
148: when $\dot M_{\rm acc} > \dot M_{\rm cr}$,
149: as illustrated in Fig. 1, thus 
150: preventing the binary from collapsing.
151: The binary does not spiral-in but keeps its separation. 
152: Hydrogen burns steadily, and therefore the helium layer of 
153: the white dwarf can grow at a rate of 
154: $\dot M_{\rm He} \approx \dot M_{\rm cr}$.
155: The other transferred matter is blown in the wind 
156: ($\dot M_{\rm wind} \approx \dot M_{\rm acc} - \dot M_{\rm cr}$),
157: where $\dot M_{\rm wind}$ is the wind mass loss rate. 
158: We call this {\it accretion wind} because it begins to blow 
159: when the accretion rate exceeds the critical rate.
160: \par
161:      Therefore, we have to revise Fig. 9 of Nomoto (1982),
162: which has been used widely in binary evolution scenarios to show the 
163: basic processes of mass-accreting white dwarfs.
164: The revision to Nomoto's Fig. 9 is shown in Fig. 1.
165: The most important difference of Fig. 1 from 
166: Nomoto's Fig. 9 is that white dwarfs increase their masses 
167: in a much wider parameter region. 
168: In our new diagram, the status of a mass-accreting white dwarf 
169: is specified by the following three phases:
170: (1) accretion wind phase ($\dot M_{\rm acc} > \dot M_{\rm cr}$); 
171: (2) steady shell-burning phase 
172: ($\dot M_{\rm std} < \dot M_{\rm acc} < \dot M_{\rm cr}$);
173: and (3) intermittent shell-flash phase
174: ($\dot M_{\rm acc} < \dot M_{\rm std}$), where
175: $\dot M_{\rm std}$ means the lowest limit of the mass accretion rate
176: for steady hydrogen shell-burning.  
177: The new growing region of white dwarfs,
178: $\dot M_{\rm acc} > \dot M_{\rm std}$, 
179: is much wider than old Nomoto's (1982) narrow strip,
180: $\dot M_{\rm std} < \dot M_{\rm acc} < \dot M_{\rm cr}$.
181: 
182: 
183: \begin{figure}
184: %%%%%\plottwo{ciaql_2.eps}{ciaql_1.eps}
185: \plottwo{hachisu3_f3.eps}{hachisu3_f4.eps}
186: \caption{{\bf Left:}
187: Geometrical configuration of our RX J0513.9-6951 model in the optical 
188: high state (wind phase).  The cool component is a slightly evolved 
189: MS companion filling up its inner critical Roche lobe.
190: The hot component, WD, blows a strong wind.
191: The accretion disk is blown off outward and has a larger size than
192: the no-wind phase.  {\bf Right:}
193: Geometrical configuration in the optical low state (no-wind phase).
194: }
195: \end{figure}
196: 
197: \begin{figure}
198: %%%%%\plotone{vmagfitall1.eps}
199: \plotone{hachisu3_f5.eps}
200: \caption{
201: Light curve is plotted against time for a $1.35 M_\odot$ white 
202: dwarf model together with the MACHO data (Alcock et al. 1996). 
203: We assume a larger size of the accretion disk 
204: (1.0 times the Roche lobe size) in the optical high state
205: (the wind phase) but a smaller size (0.8 times the Roche lobe size)
206: in the optical low state (the static, no-wind phase). 
207: The transition from the optical high to low state 
208: corresponds to the cease point of the strong wind 
209: (large open circles).  
210: }
211: \end{figure}
212: 
213: 
214: \subsection{RX J0513.9-6951 as an example of accretion wind phase}
215: As an example of the {\it accretion wind} phase, 
216: we have modeled long-term optical light curves of the LMC supersoft 
217: X-ray source RX J0513.9-6951 based on our wind model of 
218: mass-accreting white dwarfs.  The source shows a quasi-periodic 
219: variability, the period of which is $\sim 170$ d and 
220: the amplitude of which is $\sim 1$ mag (Alcock et al. 1996). 
221: The source has a relatively longer ($\sim 130$ d) optical high-state 
222: ($V \sim 16.2$) and a shorter ($\sim 40$ d) optical low-state 
223: ($V \sim 17.0$) with the transition time of several days.  
224: X-rays were observed only in the optical low-states.  
225: This indicates that a strong wind blows only in the optical high-state 
226: because X-rays are self-absorbed by the wind.  
227: To examine the idea of the strong wind, 
228: we have calculated envelope solutions of mass-accreting WDs 
229: and found that optically thick winds really occur when the mass 
230: accretion rate exceeds a critical rate of  $\dot M_{\rm cr}$
231: for $M_{\rm WD} > 0.7 M_\odot$ 
232: even in a low metallicity of $Z=0.004$ of LMC. 
233: \par
234:      We attribute the $100-150$ d bright 
235: phase to our strong wind phase with an expanded white dwarf 
236: photosphere, whereas the $\sim 40$ d 
237: less bright phase to our static (no-wind) phase with a shrinking 
238: envelope of the WD.  The duration and 
239: the brightness of the optical high state is explained by 
240: the model of a WD mass, $M_{\rm WD}= 1.35 ~M_\odot$, 
241: with an extending accretion disk edge over the Roche lobe size
242: during the wind phase as illustrated in Fig. 2. 
243: Irradiation of the accretion disk by the WD photosphere
244: plays an essential role to reproduce the optical light curve 
245: because a large part of the optical light comes from the 
246: accretion disk. 
247: \par
248:      The switching mechanism from the optical high to low state 
249: is understood as follows: 
250: Strong winds from the WD themselves make the separation increase, 
251: because the high velocity wind do not carry away large angular momentum.
252: The winds further strip off the very surface layer of the companion star.
253: As a result, the mass transfer rate decreases and it finally 
254: stops the strong winds.  After the wind stops, the surface of
255: the companion recovers in a thermal timescale of the envelope
256: and the mass transfer rate increases up and strong winds begin to blow
257: again.  Thus, the semi-periodic cycles repeat.
258: \par
259:      Light curve fittings indicate that the optical high state 
260: is initiated by a sudden increase in the mass transfer rate of
261: $\dot M_{\rm acc} \sim 1 \times 10^{-5} ~M_\odot$. 
262: Then, the envelope of the white dwarf expands to 
263: $R_{\rm ph} \sim 0.1 ~R_\odot$, not exceeding over the 
264: accretion disk size of $\sim 2 ~R_\odot$.  Our light curve
265: fittings are shown in Fig. 3.  In our models,
266: about 40\% of the transferred matter can be accumulated on
267: the white dwarf but the rest is blown off in the wind. 
268: This suggests that it takes $\sim 1 \times 10^5$ yr for the white dwarf 
269: of $1.35 ~M_\odot$ to grow to $1.38 ~M_\odot$ 
270: and to explode as a Type Ia supernova.  
271: 
272:  
273: \subsection{Two evolutionary paths to RNe and SNe Ia}
274:      Based on the new mechanism of the accretion wind evolution as 
275: shown in Fig. 1, Hachisu, Kato, \& Nomoto (1999a) and Hachisu et al. (1999b) 
276: have proposed two paths to SNe Ia (see also, Li \& van den Heuvel 1997).
277: In these two paths, WDs can accrete mass continuously from 
278: the companion and grow at a rate of $\dot M_{\rm cr}$ of eq. (1).
279: Therefore, candidates for SN Ia progenitors are systems at the final
280: stages on these paths; i.e., one is a supersoft
281: X-ray source (SSS) consisting of a WD and a lobe-filling,
282: more massive, main-sequence (MS) companion as shown in the late stage of
283: Fig. 4 (the WD+MS systems; see also Li \& van den Heuvel 1997)
284: and the other is a symbiotic star consisting of a WD
285: and a mass-losing red giant (RG) as shown in the late stage of Fig. 5 
286: (the WD+RG systems, see also Hachisu et al. 1996). 
287: Both the systems contribute to main parts of the SN Ia birth rate
288: (Hachisu et al. 1999a, 1999b).
289: \par
290:      Hachisu et al. (1999a, 1999b) have followed many binary evolutions
291: until SN Ia explosions.  The final outcome is summarized 
292: in Fig. 6 for the initial orbital period, $P_0$ (in units of days),
293: and the initial donor mass, $M_{d,0}$, which is representing 
294: $M_{\rm RG,0}$ (the initial mass of the red-giant component), 
295: or $M_{\rm MS,0}$ (the initial mass of the slightly 
296: evolved main-sequence component).  In this figure,
297: the initial mass of the white dwarf is assumed to be 
298: $M_{\rm WD,0}= 1.0 M_\odot$.  
299: The right-hand region, longer orbital periods of
300: $P_0 \sim 100$---1000 d, represents 
301: our results for the WD+RG system. 
302: In the left-hand region, shorter orbital periods of $P_0 \sim 0.3$---10 d, 
303: are shown the results for the WD+MS system.
304: \par
305:      We plot the initial parameter regions of binaries (thin solid 
306: curves in Fig. 6) that finally produce an SN Ia, which are corresponding to 
307: stage (F) in Fig. 4 or stage (D) in Fig.5.
308: Left diagram in Fig. 6 shows such two regions of the WD+MS systems 
309: (SSS channel) and the WD+RG systems (symbiotic channel).  We have added 
310: in this figure the final parameter regions (thick solid curves)
311: just before an SN Ia explosion occurs.  Starting from the initial region
312: encircled by the thin solid curves, binary systems evolve and 
313: can explode as an SN Ia in the regions enclosed 
314: by the thick solid curves.  Hachisu \& Kato (2001)
315: showed that some recurrent novae lie on the boarder of their 
316: parameter region that can produce an SN Ia 
317: (see also Kato 2001 in this volume).
318: Among six recurrent novae with known orbital periods,
319: five fall in the final regions of SN Ia progenitors 
320: just before the explosion.  Only T Pyx is far out of the regions 
321: of SNe Ia.
322: 
323: \begin{figure}
324: %%%%%\plottwo{hemtsn1a.eps}{windsn1a.eps}
325: \plottwo{hachisu3_f6.eps}{hachisu3_f7.eps}
326: \caption{ SSS channel.  {\bf Left:} 
327: Early evolutionary path through the common envelope evolution to
328: the helium (Case BB) mass transfer.  {\bf Right:}  
329: Late evolutionary path to an SN Ia explosion in our wind model.
330: }
331: \end{figure}
332: 
333: \begin{figure}
334: %%%%%\plottwo{channel1.eps}{channel2.eps}
335: \plottwo{hachisu3_f8.eps}{hachisu3_f9.eps}
336: \caption{ Symbiotic channel.
337: {\bf Left:} Early evolutionary path through the first common
338: envelope evolution to a symbiotic system. 
339: {\bf Right:} Late evolutionary path through 
340: the symbiotic channel to an SN Ia.
341: }
342: \end{figure}
343: 
344: 
345: \section{Case BB mass transfer}
346:       It has been shown that the ejecta of the recurrent nova
347: U Sco are helium-rich as much as He/H$\sim 2$ by number
348: (Williams et al. 1981).  The accretion disk of U Sco is also
349: suggested to be helium-rich (Hanes 1985).  Many of the SSSs 
350: have a stronger He {\scriptsize II} line than Balmer lines, 
351: thus also suggesting that the companion of the WD+MS systems 
352: (SSS channel) is helium-rich.  
353: In this section, therefore, we describe 
354: the reason that the cool component of the WD+MS systems has 
355: a helium-rich envelope.
356: \par
357:      Previous works consider only the initial systems 
358: consisting of a more massive AGB star with a C+O core and a less massive 
359: main-sequence star.  This system undergoes a common envelope evolution 
360: and it finally yields a binary system of a mass-accreting C+O WD and 
361: a lobe-filling MS/subgiant star.   Thus, they do not
362: include another important evolutionary path, in which
363: a more massive component fills up its inner 
364: critical Roche lobe when it develops a {\it helium core} 
365: of $\sim 0.8-2.0 M_\odot$ in its {\it red giant phase} 
366: as shown in the early stage of Fig. 4.
367: \par
368:      We consider, for example, a close binary with the primary mass of 
369: $M_{1,i} = 7 M_\odot$ and the secondary mass of $M_{2,i} = 2 M_\odot$
370: with the initial separation of $a_{i} \sim 80-600 ~R_\odot$.
371: When the primary has evolved to a red giant developing a helium 
372: core of $M_{\rm 1,He} \sim 1.0-1.6 M_\odot$, 
373: it fills up its inner critical Roche lobe.
374: Then, the binary undergoes a common envelope (CE) evolution, 
375: which yields a much more compact close binary 
376: of a naked helium star of $M_{1,{\rm He}} \sim 1.0-1.6 M_\odot$ and 
377: a main-sequence star of $M_2= 2 M_\odot$ 
378: with the final separation of $a_{\rm f, CE} \sim 4-50 ~R_\odot$. 
379: Here, we assume a relation of  
380: \begin{equation}
381: {{a_{\rm f,CE}} \over {a_i}} \sim \alpha_{\rm CE} 
382: \left({{M_{\rm 1,core}} \over {M_{1,i}}} \right) 
383: \left({{M_2} \over {M_{1,i}-M_{\rm 1,core}}} \right),
384: \end{equation}
385: with the efficiency factor of $\alpha_{\rm CE}=1.0$ for the
386: common envelope evolutions.
387: \par
388:      After the common envelope evolution, 
389: the radii of the inner critical Roche lobes of the primary and 
390: the secondary become $R_1^* \sim 0.36 a_{\rm f,CE}$ and 
391: $R_2^* \sim 0.4 a_{\rm f,CE}$, respectively.
392: Since the secondary radius should be smaller than its inner critical 
393: Roche lobe, i.e., $1.7 R_\odot \le R_2 < R_2^*$
394: ($1.7 R_\odot$ is the radius of a $2 M_\odot$ ZAMS star), 
395: the initial separation $a_{i}$ should exceeds $80 ~R_\odot$.  
396: The upper bound of the initial separation 
397: is obtained from the maximum radius of the $7 M_\odot$ star 
398: which has formed a helium core, i.e., 
399: $a_{i} \approx 2 R_{\rm 1,max} <~ 2 \times 300 ~R_\odot$. 
400: Thus, the allowable range of the initial separations is
401: $80 ~R_\odot <~ a_{i} <~ 600 ~R_\odot$ for a pair
402: of $M_{1,i}= 7 M_\odot$ and $M_{2,i}= 2 M_\odot$. 
403: Then, we have $R_1^* \sim 1.5-18 ~R_\odot$ and $R_2^* \sim 1.7-20 ~R_\odot$
404: after the first common envelope evolution.
405: \par 
406:      After the hydrogen-rich envelope is stripped away and 
407: hydrogen shell burning vanishes, the naked helium core 
408: contracts to ignite a central helium burning and 
409: becomes a helium main sequence star.  
410: When the helium star forms a C+O core, its helium envelope 
411: expands to fills its inner critical Roche lobe again.  
412: The helium is transferred stably to the secondary 
413: on an {\it evolutionary time scale} of $\tau_{\rm EV} \sim 10^5$ yr
414: because the mass ratio is smaller than $0.79$ ($q= M_1/M_2 < 0.79$).
415: The resultant mass transfer rate is 
416: $\dot M_1 \sim 10^{-5} M_\odot$ yr$^{-1}$. 
417: No common envelope is formed for such a low rate.
418: After the helium envelope is lost, the primary becomes a C+O WD 
419: of $M_{\rm C+O} \sim 0.9-1.1 M_\odot$ and the separation increases 
420: by 10\%$-$40\%, i.e., 
421: $a_{\rm f, He} \sim (1.1-1.4) a_{\rm f, CE} \sim (4-70) R_\odot$. 
422: Here, we assume the conservations of the total mass and angular momentum 
423: during the helium mass transfer
424: to obtain the separation after the helium mass transfer, 
425: $a_{\rm f, He}$.
426: The secondary receives almost pure helium matter of 
427: $\Delta M_{\rm He} \sim 0.1-0.5 M_\odot$ 
428: to form a helium-enriched envelope.
429: The secondary's hydrogen content decreases to 
430: $X \sim 0.6$ if helium is completely mixed into the star.  
431: For the $9+2.5 M_\odot$ case, much more helium is transferred,
432: but much less helium for the $6+2 M_\odot$ case.
433: 
434: 
435: 
436: \section{Shrink of orbital separation during superwind phase}
437:      Based on the population synthesis analysis, 
438: Yungelson \& Livio (1998) claimed that almost no realization frequency
439: is derived for the original WD+RG model (Hachisu et al. 1996). 
440: In this section, we point out that very wide binaries with initial 
441: separations of $a_i >~ 1500 ~R_\odot$, which were not included in their
442: analysis, are essentially important in our SN Ia modeling.
443: \par
444:      The more massive component (the initial mass of $M_{1,i}$) of 
445: a binary first evolves to a red giant (AGB stage) and fills its inner 
446: critical Roche lobe.  After a common envelope phase, 
447: the more massive component leaves a C+O WD, 
448: and the separation of the binary decreases by a factor of 
449: $a_f/a_i  \sim 1/40 - 1/50$ from eq.(2)
450: for $M_{\rm WD} \sim 1 ~M_\odot$ 
451: and $M_2 \sim 1 ~M_\odot$ because a $\sim 1 ~M_\odot$ WD descends 
452: from a main-sequence star of $M_{1,i} \sim 7-8 ~M_\odot$.
453: Yungelson \& Livio (1998) assumed that 
454: the separation of interacting binaries is {\it smaller than} 
455: $a_i <~ 1500 ~R_\odot$. 
456: Then, the widest binaries have separations of 
457: $a_f <~ 30-40 ~R_\odot$  after common envelope evolution.  
458: The orbital period is 
459: $P_0 <~ 20$ days for $M_{\rm WD,0} \sim 1 ~M_\odot$ 
460: and $M_{\rm RG,0} \sim 1 ~M_\odot$.  
461: There is no SN Ia region of the WD+RG systems 
462: for $P_0 <~ 20$ days as seen from Fig. 6. 
463: Thus, Yungelson \& Livio concluded that we cannot expect 
464: any SN Ia explosions from the right-hand SN Ia region 
465: (WD+RG system) in Fig. 6.
466: \par
467:      The reason that Yungelson \& Livio's modeling fails to reproduce
468: our SN Ia progenitors is their assumption of $a_i <~ 1500 ~R_\odot$. 
469: In what follows, we show that WD+RG binaries wide enough 
470: to have $P_0 \sim 100-1000$ d are born from initially very wide 
471: binaries with $a_i \sim 1500-40000 ~R_\odot$.
472: \par
473:      A star with a zero-age main-sequence mass of 
474: $M_{1,i} < 8 ~M_\odot$ ends its life by ejecting its envelope in a
475: wind of relatively slow velocities ($v \sim 10-40$ km s$^{-1}$) 
476: as shown in Fig. 5.  
477: These wind velocities are as low as the orbital velocities of 
478: binaries with separations of $a_i \sim 1500-40000 ~R_\odot$
479: for a pair of $M_{1,i} \sim 7 ~M_\odot$ and $M_{2,i} \sim 1 ~M_\odot$.
480: When the wind velocity is as low as the orbital velocity, 
481: we have the separation shrinking due to effective loss of
482: the angular momentum by the outflowing gas, i.e., 
483: $\dot a/a \approx 2 \dot M_1 / M_2 < 0$ (see Hachisu et al. 1999a).
484: \par
485:      Once the binary system begins to shrink, its evolution becomes 
486: similar to a common envelope evolution 
487: (see {\it stages B-D} in Fig. 5). 
488: Thus, the separation is reduced by a factor of $1/40-1/50$, i.e., 
489: $a_f \sim 30-1000 ~R_\odot$ 
490: for $M_{1,i} \sim 7 ~M_\odot$ and $M_{2,i} \sim 1 ~M_\odot$.
491: The orbital period becomes $P_0 \sim 15-3000$ days for a pair of
492: $M_{\rm WD,0} \sim 1 ~M_\odot$ and $M_2 \sim 1 ~M_\odot$
493: ({\it stage D} in Fig. 5).
494: These initial sets of the parameters are very consistent with 
495: the initial conditions of our WD+RG progenitor systems. 
496: 
497: 
498: \begin{figure}
499: %%%%%\plottwo{rn_region.eps}{phase_sp.eps}
500: \plottwo{hachisu3_f10.eps}{hachisu3_f11.eps}
501: \caption{
502: {\bf Left:} 
503: Regions producing SNe Ia are plotted  
504: in the $\log P - M_{\rm d}$ (orbital period --- donor mass) plane 
505: for the WD+MS system ({\it left}) and the WD+RG system ({\it right}).
506: Here we assume an initial white dwarf mass of $1.0 M_\odot$. 
507: Currently known positions of each recurrent nova are indicated 
508: by a star mark ($\star$) for U Sco (Hachisu et al. 2000),
509: a filled triangle for T CrB (Belczy\'nski \& Mikolajewska 1998),
510: an open rectangle for T Pyx (Patterson et al. 1998),
511: but by arrows for the other three recurrent novae, V394 CrA,
512: CI Aql, and RS Oph,
513: because the mass of the companion is not yet available explicitly.
514: {\bf Right:}
515: The accretion winds of WDs open new two channels to SNe Ia
516: while the two evolutionary processes of the case BB mass transfer
517: and the orbital shrinkage during the superwind phase enlarge 
518: the initial phase volume of SN Ia progenitors 
519: for the SSS channel and the symbiotic channel, respectively. 
520: }
521: \end{figure}
522: 
523: 
524: \section{Type Ia supernova rates}
525:      We estimate the SN Ia rate from the WD+RG and WD+MS systems 
526: in our Galaxy by 
527: \begin{equation}
528: \nu = 0.2 \int d q \int d~\log A 
529: \int {{d M} \over M^{2.5}} 
530: \quad {\rm yr}^{-1},
531: \end{equation}
532: where $q$, $A$, and $M$ are the mass ratio ($q=M_2/M_1 \le 1$),
533: the initial separation in units of solar radius, 
534: and the primary initial mass
535: in solar mass units (from eq. [1] of Iben \& Tutukov 1984).
536: The SN Ia rate is essentially the same as the volume of
537: the initial phase space consisting of $q$, $A$, and $M$ 
538: as shown in Fig. 6.  
539: \par
540:      The accretion winds of WDs open new two channels to SNe Ia.
541: The case BB mass transfer in the SSS channel extends 
542: the phase volume towards the low mass ratio region as shown
543: in Fig. 6.  On the other hand, the orbital shrinkage 
544: during the superwind phase in the symbiotic channel enlarges
545: the phase space towards the large initial separation
546: also shown in Fig. 6.
547: Therefore, the estimated rate of WD+RG/WD+MS systems becomes close 
548: to the observed rate in our Galaxy, i.e., $\nu \sim 0.003$ yr$^{-1}$,
549: which comes from the SSS channel of $\nu_{\rm MS} \sim 0.001$ yr$^{-1}$
550: and the symbiotic channel of $\nu_{\rm RG} \sim 0.002$ yr$^{-1}$.
551: 
552: 
553: \begin{references} 
554: \reference Alcock, C. et al. 1996, \mnras, 280, L49
555: \reference Belczy\'nski, K., \& Mikolajewska, J. 1998, \mnras, 296, 77
556: \reference Hachisu, I., \& Kato, M. 2001, \apj, 558, 323
557: \reference Hachisu, I., Kato, M., Kato, T, \& Matsumoto, K. 2000, \apj, 
558: 528, L97
559: \reference Hachisu, I., Kato, M., \& Nomoto, K. 1996, \apjl, 470, L97
560: \reference Hachisu, I., Kato, M., \& Nomoto, K. 1999a, \apj, 522, 487
561: \reference Hachisu, I., Kato, M., Nomoto, K., \& Umeda, H. 1999b, \apj, 519, 314 
562: \reference Hanes, D. A. 1985, \mnras, 213, 443
563: \reference Iben, I. Jr., \& Tutukov, A. V. 1984, \apjs, 54, 335
564: \reference Kato, M. 2001, in The Physics of Cataclysmic Variables and
565: Related Objects (this volume), in press
566: \reference Livio, M. 2000, in Type Ia Supernovae: Theory and Cosmology, 
567: ed. J. C. Niemeyer \& J. W. Truran (Cambridge: Cambridge Univ. Press), 33
568: \reference Li, X.-D., \& van den Heuvel, E. P. J. 1997, \aap, 322, L9
569: \reference Munari, U., \& Renzini, A. 1992, \apj, 397, L87
570: \reference Nomoto, K. 1982, \apj, 253, 798
571: \reference Nomoto,  K., \& Kondo, Y. 1991, \apjl, 367, L19
572: \reference Nomoto,  K., Thielemann, F., \& Yokoi, K. 1984, \apj, 286, 644
573: \reference Patterson, J. et al. 1998, \pasp, 110, 380
574: \reference Perlmutter, S. et al. 1999, \apj, 1999, 517, 565
575: \reference Riess, A. G. et al. 1998, \aj, 116, 1009
576: \reference Saio, H., \& Nomoto, K. 1998, \apj, 500, 388
577: \reference Starrfield, S., Sparks, W. M., \& Truran, J. W. 1985, \apj, 
578: 291, 136
579: \reference Webbink, R. F. 1984, \apj, 277, 355
580: \reference Williams, R. E., Sparks, W. M., Gallagher, J. S., Ney, E. P.,
581: Starrfield, S. G., \& Truran, J. W. 1981, \apj, 251, 221
582: \reference Yungelson, L., \& Livio, M. 1998, \apj, 497, 168
583: \end{references}
584: \end{document}
585: