0806.4122/ms.tex
1: %version: 25 June 2008
2: %\documentclass[12pt,preprint]{aastex}
3: \documentclass{aastex}
4: \usepackage[onecolumn,numberedappendix]{emulateapj5}
5: %\usepackage{epsf}
6: 
7: %\renewcommand{\baselinestretch}{2}
8: %\slugcomment{preparation for submitting to ApJ}
9: 
10: \usepackage{epsfig}
11: 
12: \shortauthors{T. Hosokawa \& K. Omukai}
13: \shorttitle{Evolution of Massive Protostars}
14: 
15: \begin{document}
16: 
17: \title{Evolution of Massive Protostars with High Accretion Rates
18: }
19: \author{Takashi Hosokawa and Kazuyuki Omukai}
20: 
21: \affil{Division of Theoretical Astronomy, 
22: National Astronomical Observatory of Japan, Mitaka, Tokyo 181-8588, Japan}
23: \email{hosokawa@th.nao.ac.jp; omukai@th.nao.ac.jp}
24: 
25: \begin{abstract}
26: Formation of massive stars by accretion requires a 
27: high accretion rate of $\dot{M}_\ast > 10^{-4}~M_{\odot}/{\rm yr}$
28: to overcome the radiation pressure barrier of the forming stars.
29: Here, we study evolution of protostars accreting at 
30: such high rates, by solving the structure of the central 
31: star and the inner accreting envelope simultaneously.
32: The protostellar evolution is followed starting from small initial cores 
33: until their arrival at the stage of the Zero-Age Main Sequence (ZAMS) stars. 
34: An emphasis is put on evolutionary features different 
35: from those with a low accretion rate of 
36: $\dot{M}_\ast \sim 10^{-5}~M_{\odot}/{\rm yr}$, 
37: which is presumed in the standard scenario for low-mass star formation. 
38: With the high accretion rate of 
39: $\dot{M}_\ast \sim 10^{-3}~M_{\odot}/{\rm yr}$, 
40: the protostellar radius becomes very large and exceeds 
41: $100~R_{\odot}$.
42: Unlike the cases of low accretion rates, 
43: deuterium burning hardly affects the evolution, and
44: the protostar remains radiative even after its ignition. 
45: It is not until the stellar
46: mass reaches $\simeq 40~M_{\odot}$ that hydrogen burning begins and
47: the protostar reaches the ZAMS phase, and this ZAMS arrival mass
48: increases with the accretion rate.
49: These features are similar to those of the first star formation 
50: in the universe, where high accretion rates are also expected, 
51: rather than to the present-day low-mass star formation. 
52: At a very high accretion rate of $> 3 \times 10^{-3}~M_{\odot}/{\rm yr}$,
53: the total luminosity of the protostar becomes so high 
54: that the resultant radiation pressure inhibits the 
55: growth of the protostars under steady accretion
56: before reaching the ZAMS stage. 
57: Therefore, the evolution under the critical accretion rate
58: $3 \times 10^{-3}~M_{\odot}/{\rm yr}$ gives the upper mass
59: limit of possible pre-main-sequence stars at $\simeq 60~M_{\odot}$.
60: The upper mass limit of MS stars is also set by the radiation
61: pressure onto the dusty envelope under the same 
62: accretion rate at $\simeq 250~M_{\odot}$.
63: We also propose that the central source enshrouded 
64: in the Orion KL/BN nebula has effective temperature and luminosity
65: consistent with our model, and is a possible candidate for such 
66: protostars growing under the high accretion rate. 
67: \end{abstract}
68: 
69: 
70: \keywords{accretion -- stars: early-type -- stars: evolution 
71: -- stars: formation -- stars: pre-main-sequence}
72: 
73: %%%%%%%%%%%%%%%%%%%%%%%%
74: \section{Introduction}
75: \label{sec:intro}
76: %%%%%%%%%%%%%%%%%%%%%%%%
77: 
78: Massive ($ > 8~M_{\odot}$) stars make significant impacts on 
79: the interstellar medium via various feedback processes:, e.g., 
80: UV radiation, stellar winds, and supernova explosions.
81: Such feedback processes sometimes trigger or regulate nearby 
82: star-formation activity.
83: Their thermal and kinetic effects are important factors
84: in the phase cycle of the interstellar medium.  
85: Strong coherent feedback by many massive stars
86: causes galactic-scale dynamical phenomena, such 
87: as galactic winds. Furthermore, massive stars
88: dominate the light from distant galaxies. 
89: The cosmic star formation rate is mostly estimated with
90: the light observed from massive stars \citep[e.g.,][]{Md96, HB06}. 
91: Despite such importance, the question of the formation process of 
92: massive stars still remains open.  
93: As for lower-mass stars, there is a widely-accepted formation scenario, 
94: where gravitational collapse of molecular cloud cores leads to subsequent 
95: mass accretion onto tiny protostars 
96: \citep[e.g.,][]{SAL87}.
97: If one directly applies this scenario to massive star
98: formation, however, some difficulties arise in the main 
99: accretion phase \citep[e.g., see][for recent reviews]{ZY07, MO07}.
100: The main difficulty in the formation of massive stars is the very
101: strong radiation pressure acting on a dusty envelope.
102: The radiative repulsive effect becomes quite strong at
103: the dust destruction front, where the accretion flow  
104: gets much of the outward momentum of radiation.  
105: Therefore, a necessary condition for massive star formation
106: is to overcome this barrier at the dust destruction front.
107: However, some theoretical work has shown that the accretion rate of 
108: $\dot{M}_\ast \sim 10^{-5}~M_{\odot}/{\rm yr}$ expected
109: of lower-mass protostars is too low to overcome the barrier
110: \citep[e.g.,][hereafter WC87]{WC87}.
111: Because of this difficulty, various scenarios different
112: from the standard accretion paradigm have been
113: proposed such as stellar mergers \citep[e.g.,][]{SPH00}
114: and competitive accretion \citep[e.g.,][]{BBZ98}.
115: 
116: An origin of this disputed situation is uncertainty concerning 
117: the initial condition for massive star formation. 
118:  This uncertainty partly originates from the scarcity of
119: massive stars; most of the massive star forming regions
120: are distant. 
121: In addition, the initial condition is easily disrupted
122: by feedback from newly formed massive stars.
123: Since the accretion rate reflects
124: the thermal state of the original molecular core, 
125: it is still uncertain which rates should be favored 
126: for growing high-mass protostars.
127: In fact, some observations of young high-mass sources
128: suggest high accretion rates.
129: Signatures of infall motion are detected with various 
130: line observations toward high-mass protostellar objects 
131: (HMPOs) and hyper-/ultra-compact H~II regions,
132: and the derived accretion rates are  
133: $10^{-4} - 10^{-3}~M_{\odot}/{\rm yr}$ 
134: \citep[e.g.,][]{Fl05, Bl06, KW06}.
135: Similar high accretion rates are also inferred by SED
136: fitting of hot cores \citep{OLD99} and HMPOs \citep{Fz07, KG08}.
137: Molecular outflows are also ubiquitous in high-mass 
138: star-forming regions, and estimated high mass outflow rates 
139: also suggest high accretion rates \citep[e.g.,][]{B02, Zh05}. 
140:  Such high accretion rates have the advantage of 
141: overcoming the radiation pressure barrier \citep{LS71, KH74}.
142: Theoretically, \citet{Nk00} have suggested a protostar growing 
143: at the very high accretion rate of 
144: $\sim 10^{-2}~M_{\odot}/{\rm yr}$ to explain the low radiation temperature
145: of scattered light from Orion KL region \citep{Mn98}.
146: \citet{MT02, MT03} have predicted that massive molecular
147: cloud cores, from which a few massive stars form, should be
148: dominated by supersonic turbulence, envisioning that 
149: such cores, if they exist, are embedded in a high-pressure environment.
150: \citet{KKM07} have simulated collapse of the turbulent cores
151: performing radiation-hydrodynamical calculations, and demonstrated
152: that the accretion rates attain more than $10^{-4}~M_{\odot}/{\rm yr}$
153: in their simulations.
154: Some recent observations are getting a close look at the initial
155: condition of massive star formation, and have found
156: strong candidates for high-mass pre-stellar
157: cores \citep[e.g.,][]{RSJ07, Mt07}. 
158: Further detailed observations will verify the scenario.
159: 
160: The main targets of this paper are protostars growing at the
161: high accretion rates of $\dot{M}_\ast \geq 10^{-4}~M_{\odot}/{\rm yr}$.
162: Detailed modeling of such protostars should be useful 
163: for future high-resolution observations and simulations of
164: massive star formation.
165: However, previous studies on such protostars are fairly limited.
166: The protostellar evolution has been well studied for low-mass 
167: ($< 1~M_{\odot}$) and intermediate-mass ($< 8~M_{\odot}$) protostars
168: by detailed numerical calculations solving the stellar structure,
169: but these studies have focused on evolution at the low 
170: accretion rate of $\sim 10^{-5}~M_{\odot}/{\rm yr}$
171: (e.g., Stahler, Shu \& Taam 1980a,b, hereafter SST80a,b;
172: Palla \& Stahler 1990, 1991, hereafter PS91, 1992; 
173: Beech \& Mitalas 1994)
174: Maeder and coworkers have calculated the protostellar evolution
175: under accretion rates growing with the stellar mass, which
176: finally exceeds $10^{-4}~M_{\odot}/{\rm yr}$ 
177: \citep{BM96, BM01, NM00}, 
178: but it is still uncertain how the evolution changes with accretion 
179: rates, and which physical mechanisms cause differences.
180: Some authors have used polytropic one-zone models 
181: originally invented for low accretion rates
182: even at such a high rate \citep[e.g.,][]{Nk00, MT03, KT07}, but 
183: its validity remains obscure owing to the lack of detailed calculations.
184: By coincidence, the similar high accretion rates are expected
185: for protostars forming in the early universe \citep[e.g.,][]{ON98, Ys06}. 
186: This simply reflects high temperatures in primordial
187: pre-stellar clouds. Evolution of such primordial protostars has
188: been studied by \citet[][hereafter SPS86]{SPS86} and 
189: \citet{OP01, OP03}.  Therefore, our
190: other motivation is to investigate how the protostellar evolution 
191: changes with metallicities.
192: 
193: To summarize, our goal in this paper is to answer the following questions;
194: \begin{itemize}
195: \item What are the properties of massive protostars 
196:  (e.g., radius, luminosity, and effective temperature)
197:  growing at high accretion rates?
198:   Can we observe any signatures of their characteristic properties?
199: \item
200:  How different are high-mass protostars from low- and 
201:  intermediate-mass ones, for which lower accretion
202:  rates are presumed?
203:  How about the difference from the protostellar evolution of 
204:  primordial stars?
205:  Furthermore, if the protostellar evolution significantly varies with 
206:  accretion rates and metallicities, what causes such differences? 
207: \item 
208:   What are the consequences of protostellar evolution with
209:   high accretion rates for the formation and feedback processes
210:   of massive stars?
211:   How massive can a star get before its arrival at the 
212:   Zero-Age Main Sequence (ZAMS)?
213:   What is the maximum mass of stars that can be formed? 
214: \end{itemize}
215: In order to answer these questions, we solve the 
216: structure of accreting protostars with various accretion rates and 
217: metallicities with detailed numerical calculations
218: \citep[also see][for a similar effort]{YB08}.
219: 
220: The organization of this paper is as follows:
221: In \S~\ref{sec:num}, we briefly review the basic procedure to construct 
222: numerical models of accreting protostars and their surrounding envelopes.
223: The subsequent \S~\ref{sec:result} is the main part of this
224: paper, where our numerical results are presented.
225: First, we investigate the protostellar evolution of two fiducial 
226: cases with the accretion rates of $10^{-3}$ and 
227: $10^{-5}~M_{\odot}/{\rm yr}$ in \S~\ref{ssec:md_1em3} 
228: and \S~\ref{ssec:md_1em5}. After that, we show more general 
229: variations of protostellar evolution,
230: e.g., mass accretion rates in \S~\ref{ssec:md_dep},
231: and metallicities in \S~\ref{ssec:metal}.
232: In \S~\ref{ssec:edd}, we present the protostellar evolution
233: at very high accretion rates exceeding 
234: $10^{-3}~M_{\odot}/{\rm yr}$, and show that the steady accretion
235: is limited by a radiation pressure barrier acting on a gas
236: envelope.
237: We further discuss some implications based on our numerical results.
238: In \S~\ref{sec:stopacc}, we examine which feedback effects can
239: affect the growth of protostars for a given accretion rate.
240: We also explore some observational possibilities of detecting signatures 
241: of the supposed high accretion rates in \S~\ref{sec:obs}.
242: Finally, \S~\ref{sec:sum} is assigned to summary and conclusions.
243: 
244: %\clearpage
245: 
246: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
247: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
248: \section{Numerical Modeling}
249: \label{sec:num}
250: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
251: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
252: 
253: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
254: \subsection{Outline}
255: \label{ssec:outline}
256: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
257: 
258: %-----------------------------------------------------------------%
259: \begin{figure}[t]
260:   \begin{center}
261: \epsfig{ file=f1.eps,
262:          angle=0,
263:          width=4in}
264: \end{center}
265: %prost_shcem.eps   
266: \caption{
267: A schematic figure of a protostar and surrounding accretion 
268: flow.
269: The accretion shock forms at the stellar surface.
270: When the flow becomes optically thick before hitting on the shock, 
271: the photosphere forms outside the stellar surface.
272: The optically thick part of the accreting envelope is called 
273: the radiative precursor.
274: A dust cocoon surrounds the protostar at larger radius.
275: The inner boundary of the dust cocoon is a dust destruction front.
276: Most of the light from the protostar is absorbed at the 
277: dust destruction front once, and reemitted as infrared light.
278: In our calculations, we solve the detailed structure only of
279: the protostar and accretion flow far inside the dust destruction 
280: front, which are enclosed by the thick dashed line.
281: We discuss possible feedback effects on the outer dust cocoon
282: in \S~\ref{sec:stopacc}.
283: }
284: \label{fig:prost_schem}
285: \end{figure}
286: 
287: %-------------------------------------------------------------------%
288: 
289: We employ the method of calculation developed by SPS86 and PS91. 
290: In this section, only the outline of modeling is overviewed.
291: A detailed description of the method is presented 
292: in Appendix~\ref{sec:method}.
293: 
294: As shown in Figure \ref{fig:prost_schem},
295: the whole system can be divided into two parts of different nature, 
296: i.e., an accreting envelope and a hydrostatic core.
297: This hydrostatic core is also called a protostar since it 
298: eventually grows into a star by accretion.
299: The outer part of the accreting envelope contains dust grains, 
300: while the warm ($\gtrsim 2000$K) part near the protostar
301: is dust-free as a result of their evaporation.
302: Despite the absence of the dust, the innermost dense part 
303: of the envelope becomes optically thick to gas continuum opacity 
304: and the photosphere appears outside the accretion shock front
305: in the case of a high accretion rate.
306: Such an inner envelope is called the radiative precursor. 
307: We here consider only the protostar and the radiative precursor
308: under a given constant accretion rate (see Fig.~\ref{fig:prost_schem}).
309: The dusty outer envelope is not included in our formulation,
310: although stellar feedback exerted on the envelope may be important in 
311: the last stage of accretion: the mass accretion can be terminated finally by 
312: radiation pressure or other protostellar feedback 
313: processes (e.g., stellar wind) exerted on the dusty envelope.
314: For the present, we presume that the protostar continues 
315: to grow at a given mass accretion rate and look for solutions 
316: of protostars with steady-state accretion. 
317: Discussion on feedback that halts the accretion is 
318: deferred to \S~\ref{sec:stopacc}.
319: 
320: We calculate protostellar evolution by constructing a
321: time sequence of quasi-steady structures of the protostar and 
322: accreting envelope. 
323: We solve the stellar structure equations for the protostar. 
324: For the accreting envelope, we adopt different treatments 
325: depending on whether or not it is opaque to the gas opacity. 
326: If the flow remains optically thin and no radiative precursor exists, 
327: the free-fall is assumed to be outside the star. 
328: For the opaque flow, on the other hand,
329: we solve the structure of the radiative precursor 
330: by using the equations for a steady-state flow.
331: At the outer boundary, which is taken to be at the photosphere, 
332: the flow is assumed to be in free fall.
333: After solving the protostar and accreting envelope individually, 
334: these solutions are connected at the accretion shock front 
335: by the radiative shock condition (e.g., SST80a).
336: The shooting method is adopted for solving radial structure.
337: This procedure is repeated until the required boundary
338: conditions are satisfied. 
339: 
340: We start calculation from a very small protostar, typically 
341: $M_{\ast, 0}= 0.01M_{\sun}$.
342: The initial models are constructed following SST80b 
343: (see Appendix~\ref{ssec:initial} for detail). 
344: Although this choice is rather arbitrary, the star converges
345: immediately to a certain structure appropriate for accretion 
346: at a given rate. 
347: Specific conditions of the initial models do not affect 
348: the evolution thereafter.   
349: The evolution is followed by increasing
350: the stellar mass owing to accretion step by step.
351: This procedure is repeated until the star reaches the ZAMS phase
352: after the onset of hydrogen burning.
353: In a few runs with very high accretion rates, however,
354: steady accretion becomes impossible before arrival at 
355: the ZAMS and the calculation is terminated at this moment.
356: 
357: 
358: %------------------------------------------------------------------------
359: \begin{table}[t]
360: %\caption{Model Parameters}
361: \label{tb:md}
362: \begin{center}
363: Table 1. Calculated Runs and Input Parameters \\[3mm]
364: %{\scriptsize
365: \begin{tabular}{l|ccccccc}
366: \hline
367: Run    & $\dot{M}_\ast~(M_{\odot}/{\rm yr})^a$ & 
368:          ${\rm [D/H]}~(10^{-5})^b$ &
369:          $Z^c$ & $M_{*,0}~(M_{\odot})^d$ & $R_{*,0}~(R_{\odot})^e$ & reference$^f$ \\
370: \hline
371: \hline
372: MD6x3     & $6 \times 10^{-3}$  & 2.5   & 0.02  & 0.5 & 43.9
373:                                 & \S~\ref{ssec:edd}    \\
374: MD4x3     & $4 \times 10^{-3}$  & 2.5   & 0.02  & 0.3 & 33.0
375:                                 & \S~\ref{ssec:edd}    \\
376: MD3x3     & $3 \times 10^{-3}$  & 2.5   & 0.02  & 0.2 & 27.8
377:                                 & \S~\ref{ssec:edd}    \\
378: MD3x3-z0  & $3 \times 10^{-3}$  & 2.5   & 0.0   & 0.2 & 26.4
379:                                 & \S~\ref{ssec:edd}    \\
380: MD3       & $10^{-3}$  & 2.5    & 0.02  & 0.05  & 15.5
381:                        &  \S~\ref{ssec:md_1em3}, \ref{ssec:md_dep}
382:                                                , \ref{ssec:metal}, \ref{ssec:edd}  \\
383: MD3-noD   & $10^{-3}$  & 0.0   & 0.02   & 0.05 & 15.5
384:                        &  \S~\ref{ssec:md_dep}   \\
385: MD3-z0    & $10^{-3}$  & 2.5   & 0.0   &  0.05 & 13.0
386:                        &  \S~\ref{ssec:metal}  \\
387: MD4       & $10^{-4}$  & 2.5   & 0.02  &  0.01 & 7.9
388:                        &  \S~\ref{ssec:md_dep}   \\
389: MD4-noD   & $10^{-4}$  & 0.0   & 0.02  &  0.01 & 7.9
390:                        &  \S~\ref{ssec:md_dep}   \\
391: MD5       & $10^{-5}$  & 2.5   & 0.02  &  0.01  & 3.7
392:                        &  \S~\ref{ssec:md_1em5}, \ref{ssec:md_dep} 
393:                         , \ref{ssec:metal} \\
394: MD5-noD   & $10^{-5}$  & 0.0   & 0.02  &  0.01  & 3.7
395:                        &  \S~\ref{ssec:md_1em5}, \ref{ssec:md_dep}   \\
396: MD5-dh1   & $10^{-5}$  & 1.0   & 0.02  &  0.01  & 3.7
397:                        &  Appendix \ref{ssec:dab}  \\
398: MD5-dh3   & $10^{-5}$  & 3.0   & 0.02  &  0.01  & 3.7
399:                        &  Appendix \ref{ssec:dab}  \\
400: MD5-z0    & $10^{-5}$  & 2.5   & 0.0   &  0.01  & 2.9
401:                        &  \S~\ref{ssec:metal}  \\
402: MD5-ps91$^g$  & $10^{-5}$  & 2.5   & 0.02  &  1.0 & 4.2
403:                        &  Appendix \ref{ssec:psini}  \\
404: MD6       & $10^{-6}$  & 2.5   & 0.02  &  0.01  &  1.6 
405:                        &  \S~\ref{ssec:md_dep}   \\
406: MD6-noD   & $10^{-6}$  & 0.0   & 0.02  &  0.01  & 1.6
407:                        &  \S~\ref{ssec:md_dep}   \\
408: \hline                           
409: \end{tabular}
410: %}
411: \noindent
412: \end{center} 
413: $a$ : mass accretion rate, $b$ : initial number fractional abundance of
414: deuterium, $c$ : metallicity, $d$ : mass of initial core model,
415: $e$ : radius of initial core model,  
416: $f$ : subsections where numerical results of each run are presented, 
417: $g$ : initial model is the same as PS91
418: 
419: \end{table}
420: %--------------------------------------------------------------------------
421: 
422: 
423: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
424: \subsection{Calculated Runs}
425: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
426: 
427: The calculated runs and their input parameters are listed in Table.1. 
428: Cases with a wide range of accretion rates are studied,
429: starting from $10^{-6}~M_{\odot}/{\rm yr}$ up to 
430: $6 \times 10^{-3}~M_{\odot}/{\rm yr}$. 
431: The lowest values of $10^{-6}- 10^{-5}~M_{\odot}/{\rm yr}$
432: are typical for low-mass star formation, 
433: while high rates $>10^{-4}~M_{\odot}/{\rm yr}$
434: are those envisaged in the accretion scenario of 
435: massive star formation. 
436: The initial deuterium abundance of [D/H] = 
437: $2.5 \times 10^{-5}$ is adopted as a fiducial value,  
438: following previous work (e.g., SST80a, PS91) for comparison. 
439: To assess the role of deuterium burning, 
440: we also calculate ``noD'' runs, 
441: where the deuterium is absent.
442: Most of the runs are for the solar metallicity $Z_{\sun}$(=0.02).
443: Two runs with $10^{-3}$ and $10^{-5} M_{\odot}/{\rm yr}$ are 
444: calculated for the metal-free gas (runs MD3-z0 and MD5-z0)
445: to see the effects of different metallicities.
446: For $10^{-5}~M_{\odot}/{\rm yr}$, variations in deuterium abundances 
447: (runs MD5-dh1 and MD5-dh3)
448: and in initial models (MD5-ps91) are studied and 
449: presented in Appendix \ref{ap:prev} for comparison with 
450: previous calculations.
451: 
452: Initial stellar mass in each run is taken to be sufficiently small, 
453: typically $M_{*,0} = 0.01~M_{\odot}$.
454: For high accretion rates $\dot{M}_\ast \geq 10^{-3}~M_{\odot}/{\rm yr}$,
455: somewhat more massive initial stars (see Table.1) are used 
456: since convergence of calculation was not achieved for lower mass ones.
457: The radii of initial models are also listed in Table.1.
458: 
459: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
460: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
461: \section{Protostellar Evolution with Different Accretion Rates}
462: \label{sec:result}
463: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
464: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
465: 
466: 
467: There are two important timescales for evolution of 
468: accreting protostars. The first one is the accretion timescale, 
469: \begin{equation}
470: t_{\rm acc} \equiv \frac{M_{\ast}}{\dot{M}_\ast} ,
471: \end{equation}
472: over which the protostar grows by mass accretion.
473: This is an evolutionary timescale of our calculations.
474: The second one is the Kelvin-Helmholtz (KH) timescale, 
475: \begin{equation}
476: t_{\rm KH} \equiv \frac{G M_{\ast}^2}{R_{\ast} L_{\ast}} ,
477: \end{equation}
478: over which the protostar loses energy by radiation.
479: The balance among these timescales is crucial to
480: the protostellar evolution.
481: When $t_{\rm KH} > t_{\rm acc}$, the radiative energy loss
482: hardly affects the protostellar evolution. 
483: The evolution is controlled only by the effect of mass accretion.
484: When $t_{\rm KH} < t_{\rm acc}$, on the other hand,
485: the evolution is mainly controlled by the radiative energy loss.
486: 
487: Luminosity from accreting protostars has two sources:
488: one is from the stellar interior, the other from the accretion
489: shock front.
490: In this paper, we call the former the interior luminosity $L_{\ast}$, 
491: and the latter the accretion luminosity,
492: \begin{equation}
493: L_{\rm acc} \equiv \frac{G \dot{M}_\ast M_{\ast}}{R_{\ast}} .
494: \label{eq:lacc}
495: \end{equation}
496: Total luminosity from the star is the sum of the interior 
497: and accretion luminosity: $L_{\rm tot} = L_{\ast} + L_{\rm acc}$.
498: Note that the balance between $t_{\rm KH}$ and $t_{\rm acc}$ also 
499: determines the dominant component of luminosity from accreting 
500: protostars, i.e., 
501: the ratio $L_{\rm acc}/L_{\ast}$ is equivalent to
502: $t_{\rm KH}/t_{\rm acc}$.
503: 
504: During the growth of the protostar, the KH timescale 
505: significantly decreases.  
506: As a result, the ratio $t_{\rm KH}/t_{\rm acc}$, which is
507: initially very large, falls below unity in the protostellar evolution. 
508: Below, we show that the decrease of $t_{\rm KH}/t_{\rm acc}$
509: indeed leads to a variety of evolutionary phases of accreting
510: protostars.
511: 
512: 
513: 
514: 
515: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
516: \subsection{Case with High Accretion Rate
517: $\dot{M}_\ast = 10^{-3}~M_{\odot}/{\rm yr}$}
518: \label{ssec:md_1em3}
519: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
520: 
521: First, we see a protostar growing at a 
522: high accretion rate $\dot{M}_\ast = 10^{-3}~M_{\odot}/{\rm yr}$ (run MD3), 
523: which is relevant to massive star formation.
524: The upper panel of Figure \ref{fig:str_fdtmax_1em3} shows
525: evolution of the protostellar radius (thick solid) and 
526: the interior structure as a function of the protostellar mass. 
527: The mass increases monotonically with time $t$, 
528: $M_{\ast}=M_{\ast,0}+\dot{M}_\ast t$, and can be regarded as 
529: a time coordinate. 
530: The entire evolution can be divided into the following 
531: four phases according to their characteristic features;
532: (I) the adiabatic accretion (for $M_{\ast} \lesssim 6~M_{\odot}$), 
533: (II) swelling ($6~M_{\odot} \lesssim M_{\ast} \lesssim 10~M_{\odot}$),
534: (III) KH contraction 
535: ($10~M_{\odot} \lesssim M_{\ast} \lesssim 30~M_{\odot}$), 
536: and (IV) main-sequence accretion ($M_{\ast} \gtrsim 30~M_{\odot}$) phases.
537: Below, we see the protostellar evolution in each phase.
538: 
539: %----------------------------------------------------------------------%
540: \begin{figure}[t]
541:   \begin{center}
542: \epsfig{ file=f2.eps,
543:          angle=0,
544:          width=4in}
545: %str_fdtmax_1em3.eps
546: \caption{ 
547: Evolution of a protostar with the accretion rate 
548: $\dot{M}_\ast = 10^{-3}~M_{\odot}/{\rm yr}$ (run MD3).
549: {\it Upper panel} : The interior structure of the protostar.
550: The thick solid curve represents the protostellar radius ($R_{\ast}$), 
551: which is the position of the accretion shock front. 
552: Convective layers are shown by gray-shaded area.
553: The hatched areas indicate layers of active nuclear burning,
554: where the energy production rate exceeds the steady rate
555: $L_{\rm D,st}/M_{\ast}$ for the deuterium burning, 
556: and $L_{\ast}/M_{\ast}$ for the hydrogen burning. 
557: The thin dotted curves represent the loci of 
558: mass coordinates ; $M = 0.1$, 0.3, 1, 3, 10, and 30~$M_{\odot}$.
559: {\it Lower panel} :
560: Evolution of the mass-averaged deuterium concentration,
561: $f_{\rm d, av}$ (solid line) and the maximum temperature within 
562: the star $T_{\rm max}$ (dot-dashed line).
563: In both panels, the shaded background shows the 
564: four evolutionary phases ;
565: (I) adiabatic accretion, (II) swelling, (III) Kelvin-Helmholtz
566: contraction, and (IV) main-sequence accretion phases.
567: }
568: \label{fig:str_fdtmax_1em3}
569:   \end{center}
570: \end{figure}
571: %---------------------------------------------------------------------%
572: %------------------------------------------------------------------------%
573: \begin{figure}[t]
574:   \begin{center}
575: \epsfig{ file=f3.eps,
576:          angle=0,
577:          width=6in}
578: %slpf_1em3.eps
579: \caption{Radial profiles of the specific entropy and
580: the luminosity at different epochs for the protostar with 
581: the accretion rate $\dot{M}_\ast = 10^{-3}~M_{\odot}/{\rm yr}$ (run MD3)
582: shown as functions of the mass coordinate $M$.
583: For each snapshot, total stellar mass ($M_{\ast}$) at its moment is labeled.
584: The four panels correspond to the four evolutionary phases, 
585: (I) adiabatic accretion (upper left), (II) swelling 
586: (upper right), (III) Kelvin-Helmholtz contraction (lower left), 
587: and (IV) main-sequence accretion (lower right) phases.
588: In the entropy profiles of the upper left panel, 
589: the filled circles denote the post-shock values.
590: In the upper right panel, 
591: the open circles indicate the bottom edges of convective layers.
592: }
593: \label{fig:slpf_1em3}
594:   \end{center}
595: \end{figure}
596: %------------------------------------------------------------------------%
597: %------------------------------------------------------------------------%
598: \begin{figure}[t]
599: \begin{center}
600: \epsfig{ file=f4.eps,
601:          angle=0,
602:          width=4in}
603: %t_comp_1em3.eps
604: \caption{
605: Comparison between the local accretion and cooling timescales
606: in the surface settling layer, defined by equations 
607: (\ref{eq:taccs}) and (\ref{eq:tcools}).
608: These timescales are shown as functions of 
609: the mass coordinate measured from the accretion shock front
610: $m \equiv M_{\ast} - M$.
611: The accretion timescale is presented by the solid line. 
612: The cooling timescales at $M_{\ast} = 1~M_{\odot}$, 
613: 3~$M_{\odot}$, and 5~$M_{\odot}$ are presented
614: by the dot-dashed, dashed, and dotted lines, 
615: respectively. }
616: \label{fig:t_comp_1em3}
617:   \end{center}
618: \end{figure}
619: %----------------------------------------------------------------------%
620: %---------------------------------------------------------------------%
621: \begin{figure}[t]
622:   \begin{center}
623: \epsfig{ file=f5.eps,
624:          angle=0,
625:          width=4in}
626: %lum_enuc_tsc_1em3.eps
627: \caption{ 
628: {\it Top panel} : Evolution of the interior luminosity
629: $L_{\rm *}$ (dotted), accretion luminosity 
630: $L_{\rm acc}$ (dashed), and total luminosity 
631: $L_{\rm tot} \equiv L_{\rm *} + L_{\rm acc}$
632: (solid) of an accreting protostar with 
633: $\dot{M}_\ast = 10^{-3}~M_{\odot}/{\rm yr}$ (run MD3).
634: The total luminosity calculated with radius and luminosity
635: of the ZAMS stars is also presented by the thin solid line.
636: {\it Middle panel} : Evolution of various contributions
637: to the interior luminosity of the protostar.
638: The thick dotted line represents the stellar luminosity,
639: $L_{\rm *}$ (also shown in the top panel).
640: The total burning rate of each nuclear reaction is
641: shown for the deuterium burning ($L_{\rm d}$, coarse dotted), 
642: pp-chain ($L_{\rm pp}$, dot-dot-dashed), and CN-cycle 
643: ($L_{\rm CN}$, dot-dashed line). The horizontal coarse dotted line indicates 
644: the steady deuterium burning rate $L_{\rm d,st}$.
645: The maximum luminosity within the star is plotted with the coarse dashed
646: line.
647: {\it Bottom panel} : Evolution of the accretion timescale $t_{\rm acc}$
648: (dashed), and Kelvin-Helmholtz timescale $t_{\rm KH}$ 
649: (dotted) for the same protostar. 
650: In all panels, the shaded background shows the four evolutionary 
651: phases, as in Fig.~\ref{fig:str_fdtmax_1em3}. 
652: }
653: \label{fig:lum_enuc_tsc_1em3}
654:   \end{center}
655: \end{figure}
656: %------------------------------------------------------------------------%
657: 
658: \paragraph{Adiabatic Accretion Phase}
659: 
660: Upper panels of Figure \ref{fig:slpf_1em3} present the evolution 
661: of entropy profile within the protostar. 
662: This shows that the entropy profiles at different epochs completely 
663: overlap: at a fixed mass element, the specific entropy 
664: remains constant. The entropy is originally generated at the accretion 
665: shock front, and embedded into the stellar interior.
666: Thus, during this earliest phase,  
667: the entropy profile inside the star just traces the history of 
668: entropy at the post-shock point. Owing to a high value of opacity, 
669: radiative heat transport is inefficient in the interior of the
670: protostar. This makes the interior luminosity small, then
671: $t_{\rm KH} \gg t_{\rm acc}$ in this phase 
672: (see Fig.~\ref{fig:lum_enuc_tsc_1em3} bottom panel, below).
673: Once the accreted matter settles into the opaque interior, the
674: specific entropy is just conserved.
675: 
676: The snapshots of the luminosity profile are also presented in 
677: lower panels of Figure \ref{fig:slpf_1em3}.
678: Near the surface is a spiky structure where the gradient
679: $\partial L/\partial M$ is very large owing to a sharp decrease in opacity there.
680: Although a significant entropy loss occurs in this luminosity spike as 
681: indicated by the relation 
682: $(\partial s/\partial t)_M \simeq - 1/T~(\partial L/\partial M)_t$ 
683: (from eq. (\ref{eq:ene})), 
684: this layer is extremely thin and the time for a mass element 
685: to pass though it is so short that the entropy hardly changes.
686: This can be confirmed by comparing the two timescales;
687: the local accretion time 
688: \begin{equation}
689: t_{\rm acc,s} (m) = \frac{m}{\dot{M}_\ast},
690: \label{eq:taccs}
691: \end{equation}
692: and local cooling time
693: \begin{equation}
694: t_{\rm cool,s} (m) = 
695: \frac{m~\delta s}{\int_0^m \frac1T \frac{\partial L}{\partial m} dm'},
696: \label{eq:tcools}
697: \end{equation}
698: where $m = M_{\ast} - M$ is the depth (in mass coordinate) of a mass shell 
699: from the accretion shock, 
700: and $\delta s$ denotes arbitrary small entropy change.
701: The former, $t_{\rm acc, s}$, is the time for 
702: a thin shell of mass $m$ at the surface to be replaced 
703: by the newly accreted material, while
704: the latter, $t_{\rm cool, s}$, is the time for
705: the same shell to lose the entropy $m~\delta s$ by outward
706: radiation. 
707: Comparison between these timescales in the settling layer is presented
708: in Figure \ref{fig:t_comp_1em3}.
709: where $\delta s = 0.1~k_{\rm B}/m_{\rm H}$ 
710: is adopted for numerical evaluation.
711: This indicates that $t_{\rm acc, s}$ is always shorter than $t_{\rm cool, s}$:
712: the accreted material is swiftly embedded in the interior before 
713: losing the entropy $\delta s$ by radiation.
714: Therefore, the adiabaticity remains valid 
715: in spite of the spike in luminosity profile.
716: Note that the short $t_{\rm acc,s}$ is a result of the high accretion
717: rate. 
718: We will see that the situation changes for much lower 
719: accretion rates in \S~\ref{ssec:md_1em5} and \ref{ssec:md_dep} below.
720: 
721: With increasing protostellar mass $M_{\ast}$, the accretion shock 
722: strengthens and the post-shock entropy increases.
723: Then, the interior entropy increases as well.
724: This causes a gradual expansion of the stellar radius in the adiabatic 
725: accretion phase (Fig.~\ref{fig:str_fdtmax_1em3}, upper panel).
726: Using the typical density and pressure within a star 
727: of mass $M_{\ast}$ and radius $R_{\ast}$ 
728: \citep[e.g.,][]{CG68};
729: \begin{equation}
730: \rho \sim \frac{M_{\ast}}{R_{\ast}^3}, 
731: \qquad P \sim G \frac{M_{\ast}^2}{R_{\ast}^4}.
732: \label{eq:typ_rhop}
733: \end{equation}
734: and the expression for specific entropy of ideal monatomic gas 
735: \begin{equation}
736: s = \frac{3 \cal{R}}{2 \mu} \ln \left( \frac{P}{\rho^{5/3}} \right) + const.
737: \label{eq:s_gene}
738: \end{equation}
739: where ${\cal R}$ is the gas constant and $\mu$ is the mean
740: molecular weight, 
741: the stellar radius and entropy is related as \citep{St88};
742: \begin{equation}
743: R_{\ast} \propto M_{\ast}^{-1/3} \exp\left[    
744:                             \frac{2 \mu}{3 {\cal R}} s
745:                            \right],
746: \label{eq:r_srel}
747: \end{equation} 
748: i.e., the stellar radius is larger for the higher entropy 
749: within the star. 
750: SPS86 have derived the mass-radius relation for 
751: protostars in the adiabatic accretion phase;
752: \begin{equation}
753: R_{\ast} \simeq  26~R_{\odot} \left( \frac{M_{\ast}}{M_{\odot}} \right)^{0.27} 
754:                      \left( \frac{\dot{M}_\ast}{10^{-3}~M_{\odot}/{\rm yr}} \right)^{0.41},
755: \label{eq:r_sps86}
756: \end{equation}
757: under the condition that opacity in the radiative precursor 
758: is dominated by H$^{-}$ bound-free absorption.
759: As this condition holds in our case,  
760: our calculated mass-radius relation is in a good agreement with
761: equation (\ref{eq:r_sps86}).
762: Equation (\ref{eq:r_sps86}) suggests the large radius
763: ($> 10~R_\odot$) of the rapidly-accreting protostar.
764: 
765: The interior of the protostar remains radiative throughout this phase 
766: (Fig.~\ref{fig:str_fdtmax_1em3}, upper panel):
767: all the energy transport is via radiation.
768: This is in high contrast with low $\dot{M}_\ast$ 
769: ($\sim 10^{-5}~M_{\odot}/{\rm yr}$) cases,
770: where most of the interior becomes convective owing to 
771: deuterium burning for $M_{\ast} \ga 0.4M_{\sun}$ 
772: (see Fig.~\ref{fig:str_fdtmax_1em5} below).
773: This can be attributed to a difference in the interior temperature.
774: From equation (\ref{eq:typ_rhop}), typical temperature within the star is  
775: \begin{equation}
776: T = \frac{\mu}{\cal{R}} \frac{P}{\rho} 
777:   \sim \frac{G}{\cal{R}} \frac{\mu M_{\ast}}{R_{\ast}}.
778: \label{eq:t_typ}
779: \end{equation}
780: Therefore, the large stellar radius leads to the
781: low temperature in the stellar interior.
782: In fact, the maximum temperature in this phase does not exceed 
783: the threshold for deuterium burning
784: (lower panel of Fig.~\ref{fig:str_fdtmax_1em3}).
785: 
786: Once the accreted matter has passed through the outermost layer
787: with the luminosity spike, the luminosity gradient becomes much milder.
788: The luminosity has a maximum at some radius:
789: outside the luminosity maximum, the gradient $\partial L / \partial M < 0$, 
790: while $\partial L / \partial M > 0$ inside.
791: This means that heat is removed from the deep interior
792: and absorbed in the outer layer.
793: This entropy transfer, however, remains small and does not
794: modify the entropy distribution significantly during this phase.
795: Efficiency of the outward entropy transfer is related to the
796: value of opacity.
797: In most of the stellar interior, major sources of opacity are
798: bound-bound and bound-free transitions of heavy elements, which 
799: approximately obey Kramers' law, 
800: $\kappa \propto \rho T^{-3.5}$ \citep{HHS62, Cl68}.
801: With the increase in stellar mass and then 
802: the interior temperature, the opacity decreases.
803: This results in steady increase of the outward heat flux 
804: with evolution.
805: In fact, as shown in the lower panels of Figure \ref{fig:slpf_1em3}, 
806: the amplitude of the luminosity increases with the growth of the 
807: protostar.
808: Also the maximum luminosity within the star $L_{\rm max}$ 
809: increases as a power-law function of $M_{\ast}$ 
810: as seen in the middle panel of Figure \ref{fig:lum_enuc_tsc_1em3}.
811: This dependence can be understood as follows:
812: for a star with radiative stratification and Kramers' opacity, 
813: the luminosity scales as
814: $L_{\rm rad} \propto M_{\ast}^{11/2} R_{\ast}^{-1/2}$ 
815: (e.g., Cox \& Giuli 1968).
816: We have confirmed that $L_{\rm max}$ roughly obeys,
817: \begin{equation}
818: L_{\rm max} \simeq 0.2~L_{\odot} 
819:                 \left( \frac{M_{\ast}}{M_{\odot}}  \right)^{11/2}
820:                 \left( \frac{R_{\ast}}{R_{\odot}}  \right)^{-1/2} 
821: \label{eq:l_max_a}
822: \end{equation}
823: in our calculations. 
824: Using equations (\ref{eq:r_sps86}) and (\ref{eq:l_max_a}), 
825: we obtain $L_{\rm max} \propto M_{\ast}^{5.4}$ for a constant
826: accretion rate.
827: When opacity becomes sufficiently small, the radiative heat transport 
828: becomes significant even in the deep interior.
829: Consequently, the entropy distribution and thus stellar structure are
830: drastically altered, which leads to the subsequent swelling phase.
831: 
832: 
833: 
834: \paragraph{Swelling Phase}
835: 
836: At the mass of about 6~$M_{\sun}$, 
837: the protostar begins to swell up dramatically.
838: The stellar radius suddenly increases by a factor of about three 
839: and eventually exceeds $100~R_{\odot}$. 
840: This swelling is caused by redistribution of entropy within the star.
841: As explained above, the outward heat flow continues to increase,
842: and ceases to be negligible.
843: Figure \ref{fig:slpf_1em3} shows that the entropy distribution changes 
844: significantly with time, i.e., with $M_{\ast}$: 
845: the entropy decreases in the deep interior, 
846: and increases in the outer layer. 
847: The extent of the inner entropy-losing region 
848: becomes larger and larger, and approaches the surface, 
849: as more of the interior becomes less opaque.
850: 
851: As seen in the bottom panels of Figure~\ref{fig:slpf_1em3}, 
852: the peak of the luminosity, which is the boundary between 
853: the entropy-losing interior and the outer absorbing layer, 
854: gradually moves toward the surface with increasing height.
855: This outward propagation of the luminosity peak is called the 
856: luminosity wave (SPS86). 
857: Only a thin outlying layer absorbs the 
858: entropy transported from the deep interior.
859: This rapid increase of the entropy near the surface 
860: causes swelling of the radius. 
861: Actually, only a small fraction of mass near the 
862: surface contributes to this swelling.
863: For example, at the maximum expansion of the radius 
864: $R_{\ast} \simeq 140~R_{\odot}$ at $M_{\ast} \simeq 10~M_{\odot}$, 
865: only 0.03\% of the total mass is contained 
866: in the outer layers of $R_{\ast} > 70~R_{\sun}$.
867: 
868: Note that the swelling is not due to the shell burning 
869: of deuterium.
870: \citet{PS90} have presented that a similar swelling occuring
871: in intermediate-mass protostars at the low accretion rate 
872: of $\dot{M}_\ast = 10^{-5}~M_{\odot}/{\rm yr}$, and concluded that
873: this is caused by the shell burning of deuterium. 
874: In our case, however, deuterium burning is not important 
875: for the swelling.
876: Even in the run without deuterium burning (MD3-noD),
877: the swelling is caused by redistribution of entropy in the same fashion
878: (also see \S~\ref{ssec:md_dep} below). 
879: During this phase, the deuterium burning occurs in the inner region
880: (Fig.~\ref{fig:str_fdtmax_1em3}, upper panel).
881: Despite the active deuterium burning, 
882: most of the stellar interior remains radiative.
883: Because of the higher entropy and thus lower density in the star, 
884: the opacity is lower in our case when the active deuterium burning
885: begins. 
886: This allows efficient radiative transport of the generated entropy 
887: \citep{St88}. 
888: The middle panel of Figure \ref{fig:lum_enuc_tsc_1em3} shows that
889: the maximum radiative luminosity $L_{\rm max}$ always exceeds
890: that by the deuterium burning $L_{\rm D}$.
891: This indicates a large capacity of radiative heat transport.
892: Also a thin convective layer near the surface 
893: (Fig.~\ref{fig:str_fdtmax_1em3}) is not brought about by the deuterium burning.
894: Although this layer receives the transported entropy from inside, 
895: the higher opacity due to ionization prevents the outward heat flow. 
896: This causes a negative entropy gradient and then convection near the surface.
897: 
898: \paragraph{Kelvin-Helmholtz Contraction}
899: 
900: With the approach of the luminosity wave to the surface, 
901: an increasing amount of energy flux escapes from the star 
902: without being absorbed inside.
903: This results in a significant increase in the interior luminosity 
904: $L_{\ast}$ for $M_{\ast} > 5 M_{\odot}$ (Fig.~\ref{fig:lum_enuc_tsc_1em3}) 
905: and thus the shorter KH timescale $t_{\rm KH}$ ($\propto 1/L_{\ast}$) 
906: of the star.
907: The arrival of the luminosity wave at the stellar surface marks
908: the moment that all parts of the star begin to lose heat.
909: Almost simultaneously, $t_{\rm KH}$ becomes as short as the accretion
910: timescale $t_{\rm acc}$ (Fig.~\ref{fig:lum_enuc_tsc_1em3}). 
911: While maintaining the virial equilibrium against the radiative energy loss,
912: the protostar turns to contraction.
913: This is the so-called KH contraction phase. 
914: The epoch of the radial turn-around is roughly estimated 
915: by equating $t_{\rm acc}$ and 
916: $t_{\rm KH,lmax} \equiv G M_{\ast}^2 / R_{\ast} L_{\rm max}$, 
917: where $L_{\rm max}$
918: is given by equation (\ref{eq:l_max_a}),
919: \begin{equation}
920: M_{\rm *,rmax} = 10.8~M_{\odot} \left(  
921:                            \frac{\dot{M}_\ast}{10^{-3}~M_{\odot}/{\rm yr}}
922:                           \right)^{2/9}.
923: \label{eq:m_rmax}
924: \end{equation}
925: In the above, we have omitted a weak dependence on $R_{\ast}$ and just 
926: substituted $R_{\ast} \sim 10~R_{\odot}$ for simplicity. 
927: During the contraction of the protostar, the KH timescale remains slightly
928: shorter than the accretion timescale. 
929: Since the ratio $t_{\rm KH}/t_{\rm acc}$, or 
930: equivalently $L_{\ast}/L_{\rm max}$,
931: is less than unity, the interior luminosity exceeds the accretion 
932: luminosity and becomes the dominant source of the total luminosity 
933: of the protostar (upper panel of Fig.~\ref{fig:lum_enuc_tsc_1em3}).
934: 
935: % D-burning
936: Since the stellar interior remains radiative in spite of the 
937: deuterium burning, the accreted deuterium is not transported
938: to the deep interior by convection.
939: Deuterium in the deep interior is thus soon consumed. 
940: Thereafter the deuterium burning continues in a shell-like outer layer
941: (Fig.~\ref{fig:str_fdtmax_1em3}, upper panel).
942: This surface burning proceeds at a rate slightly exceeding
943: the steady-burning one \citep[e.g.,][]{PS90}, 
944: \begin{equation}
945: L_{\rm D, st} \equiv \dot{M}_\ast \delta_{\rm D} = 
946:                 1500~L_{\odot} \left( 
947:                                \frac{\dot{M}_\ast}{10^{-3}~M_{\odot}/{\rm yr}} 
948:                              \right)
949:                              \left(
950:                                \frac{[ {\rm D/H} ]}{2.5 \times 10^{-5}} 
951:                              \right), 
952: \label{eq:l_dst}
953: \end{equation}
954: where $\delta_{\rm D}$ is the energy available from
955: the deuterium burning per unit gas mass.
956: Consequently, the mass-averaged deuterium concentration $f_{\rm d,av}$ 
957: significantly decreases during the KH contraction
958: (Fig.~\ref{fig:str_fdtmax_1em3}, lower panel).
959: 
960: 
961: \paragraph{Arrival at a Main-sequence Star Phase}
962: 
963: With the KH contraction, the interior temperature increases gradually.
964: At $M_{\ast} \simeq 20~M_{\odot}$ the maximum temperature reaches $10^7$K
965: and the nuclear fusion of hydrogen begins (Fig.~\ref{fig:str_fdtmax_1em3}).
966: Although the hydrogen burning is initially dominated by the 
967: pp-chain reactions, energy generation by the CN-cycle reactions 
968: immediately overcome this. 
969: The energy production by the CN-cycle compensates 
970: radiative loss from the surface at $M_{\ast} \simeq 30~M_{\odot}$ 
971: (Fig.~\ref{fig:lum_enuc_tsc_1em3}, middle), where the 
972: KH contraction terminates.
973: A convective core emerges owing to the rapid entropy generation 
974: near the center.
975: The stellar radius increase after that, 
976: obeying the mass-radius relation of main-sequence stars.
977: 
978: %-------------------------------------------------------------------------%
979: \begin{figure}[t]
980: \begin{center}
981: \epsfig{ file=f6.eps,
982:          angle=0,
983:          width=4in}
984: %prec_1em3.eps
985: \caption{ 
986: {\it Upper panel} :
987: Positions of the accretion shock front (solid line), 
988: and photosphere (dotted line) of a growing protostar 
989: with $\dot{M}_\ast = 10^{-3}~M_\odot/{\rm yr}$
990: (run MD3).
991: The layer between the accretion shock front and photosphere
992: corresponds to the radiative precursor.
993: {\it Lower panel} : 
994: the optical depth within the radiative precursor $\tau_{\rm rec}$, 
995: and effective temperatures at the photosphere $T_{\rm ph}$
996: and accretion shock front 
997: $T_{\rm sh} \equiv (L_{\ast}/4 \pi R_{\ast}^2 \sigma)^{1/4}$. 
998: The temperatures are normalized as $T_4 = (T/10^4 {\rm K})$.
999: The shaded background shows the four evolutionary phases,
1000: as in Fig.~\ref{fig:str_fdtmax_1em3}. 
1001: }
1002: \label{fig:prec_1em3}
1003: \end{center}
1004: \end{figure}
1005: %-------------------------------------------------------------------------%
1006: 
1007: 
1008: \paragraph{Evolution of the Radiative Precursor}
1009: 
1010: Figure \ref{fig:prec_1em3} shows the evolution of 
1011: protostellar ($R_{\ast}$) and photospheric ($R_{\rm ph}$) radii.
1012: Except in a short duration, the photosphere is formed outside 
1013: the stellar surface: the radiative precursor persists 
1014: throughout most of the evolution.
1015: In the adiabatic accretion phase, the photospheric radius is
1016: slightly outside the protostellar radius and increases gradually with 
1017: the relation $R_{\rm ph} \simeq 1.4 R_{\ast}$, which is derived 
1018: analytically by SPS86.
1019: Although the precursor temporarily disappears in the swelling phase
1020: around $M_{\ast} \simeq 10 M_{\odot}$, 
1021: it emerges again in the subsequent KH contraction phase.
1022: In the KH contraction phase, the precursor extends spatially and
1023: becomes more optically thick.
1024: Due to the extreme temperature sensitivity of H$^-$ bound-free 
1025: absorption opacity, the region in the accreting envelope with 
1026: temperature $\ga 6000$K becomes optically thick.
1027: Thus, the photospheric temperature $T_{\rm ph}$ remains at $6000-7000$K 
1028: throughout the evolution (Fig.~\ref{fig:prec_1em3}, lower panel).
1029: Since the photospheric radius is related to the total luminosity by 
1030: \begin{equation}
1031: R_{\rm ph}=\left(L_{\rm tot}/4\pi \sigma T_{\rm ph}^4 \right)^{1/2},
1032: \end{equation}
1033: for the constant $T_{\rm ph}$,
1034: the rapid increase in the luminosity in the contraction phase 
1035: causes the expansion of the photosphere.
1036: 
1037: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1038: \subsection{Case with Low Accretion Rate
1039:             $\dot{M}_\ast = 10^{-5}~M_{\odot}/{\rm yr}$}
1040: \label{ssec:md_1em5}
1041: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1042: 
1043: Next, we revisit the evolution of an accreting protostar
1044: under the lower accretion rate of $\dot{M}_\ast = 10^{-5}~M_{\odot}/{\rm yr}$
1045: (run MD5). 
1046: Although this case has been extensively studied by previous 
1047: authors, a brief presentation should be helpful
1048: to underline the effects of different accretion rates on 
1049: protostellar evolution. 
1050: Different properties of the protostar 
1051: even at the same accretion rate owing to updates 
1052: from the previous works,   
1053: e.g, initial models and opacity tables, 
1054: will also be described.
1055: For a thorough comparison between our and some previous calculations,
1056: see appendix~\ref{ap:prev}.
1057: 
1058: The lower accretion rate means the longer accretion timescale
1059: $t_{\rm acc}\propto 1/\dot{M}_\ast$:
1060: even at the same protostellar mass $M_{\ast}$, the evolutionary 
1061: timescale is longer in the case of low $\dot{M}_\ast$.
1062: Therefore, the protostar has ample time to lose heat before 
1063: gaining more mass.
1064: This results in the lower entropy and thus a smaller radius
1065: at the same protostellar mass in the low $\dot{M}_\ast$ case 
1066: as shown in the upper panel of Figure \ref{fig:str_fdtmax_1em5}.
1067: Although with the smaller value, the overall evolutionary 
1068: features of the high and low $\dot{M}_\ast$ radii are
1069: similar.
1070: 
1071: Again, the protostellar evolution can be divided 
1072: into four characteristic stages, i.e., 
1073: (I) convection ($M_{\ast} \lesssim 3~M_{\odot}$), 
1074: (II) swelling ($3~M_{\odot} \lesssim M_{\ast} \lesssim 4~M_{\odot}$),
1075: (III) KH contraction ($4~M_{\odot} \lesssim M_{\ast} \lesssim 7~M_{\odot}$),
1076: and (IV) main-sequence accretion ($M_{\ast} \gtrsim 7~M_{\odot}$) phases.
1077: Note that the first phase is now the convection phase instead of 
1078: the adiabatic accretion. 
1079: In the low $\dot{M}_\ast$ case, the deuterium burning 
1080: drastically affects the protostellar evolution in early phases 
1081: \citep[e.g.,][]{St88}.
1082: Since the evolution in the phases (III) and (IV) is similar to 
1083: the previous high-$\dot{M}_\ast$ case (\S~\ref{ssec:md_1em3}), 
1084: we emphasize the two early phases (I) and (II) in the following. 
1085: 
1086: %--------------------------------------------------------------------%
1087: \begin{figure}[t]
1088: \begin{center}
1089: \epsfig{ file=f7.eps,
1090:          angle=0,
1091:          width=4in}
1092: %str_fdtmax_1em5.eps
1093: \caption{ Same as Fig.~\ref{fig:str_fdtmax_1em3} but for the lower
1094: accretion rate of $\dot{M}_\ast = 10^{-5}~M_{\odot}/{\rm yr}$ (run MD5). 
1095: In the lower panel, deuterium concentration in the convective layer
1096: $f_{\rm d,cv}$ is also presented.
1097: In both upper and lower panels, the shaded background shows the 
1098: four evolutionary phases ;
1099: (I) convection, (II) swelling, (III) Kelvin-Helmholtz
1100: contraction, and (IV) main sequence accretion phases}
1101: \label{fig:str_fdtmax_1em5}
1102: \end{center}
1103: \end{figure}
1104: %---------------------------------------------------------------------%
1105: %-----------------------------------------------------------------------%
1106: \begin{figure}[t]
1107:   \begin{center}
1108: \epsfig{ file=f8.eps,
1109:          angle=0,
1110:          width=6in}
1111: %slpf_1em5.eps
1112: \caption{Same as Fig.~\ref{fig:slpf_1em5} but for the lower accretion
1113: rate of $\dot{M}_\ast = 10^{-5}~M_{\odot}/{\rm yr}$ (run MD5).
1114:  }
1115: \label{fig:slpf_1em5}
1116:   \end{center}
1117: \end{figure}
1118: %------------------------------------------------------------------------%
1119: %------------------------------------------------------------------------%
1120: \begin{figure}[t]
1121: \begin{center}
1122: \epsfig{ file=f9.eps,
1123:          angle=0,
1124:          width=6in}
1125: %slpf_1em5_nod.eps
1126: \caption{Same as Fig.~\ref{fig:slpf_1em5} but for switching off
1127: deuterium burning in the calculation (run MD5-noD). 
1128: Only profiles in the early adiabatic accretion phase 
1129: and swelling phase are presented.
1130: }
1131: \label{fig:slpf_1em5_nod}
1132: \end{center}
1133: \end{figure}
1134: %---------------------------------------------------------------------%     
1135: %---------------------------------------------------------------------%
1136: \begin{figure}
1137:   \begin{center}
1138: \epsfig{ file=f10.eps,
1139:          angle=0,
1140:          width=4in}
1141: %t_comp_1em5.eps
1142: \caption{Same as Fig.~\ref{fig:t_comp_1em3} but for the case with
1143: the lower accretion rate of $\dot{M}_\ast = 10^{-5}~M_{\odot}/{\rm yr}$ 
1144: and without deuterium burning (run MD5-noD). 
1145: The local cooling timescales given by equation (\ref{eq:tcools}) 
1146: at $M_{\ast} = 1.0$ (dot-dashed), 
1147: 1.5 (dashed), and 2.0~$M_{\odot}$ (dotted line) are presented.
1148: }
1149: \label{fig:t_comp_1em5}
1150:   \end{center}
1151: \end{figure}
1152: %--------------------------------------------------------------------%   
1153: %---------------------------------------------------------------------%
1154: \begin{figure}[t]
1155: \begin{center}
1156: \epsfig{ file=f11.eps,
1157:          angle=0,
1158:          width=4in}
1159: %str_1em5_nod.eps
1160: \caption{ Same as the upper panel of Fig.~\ref{fig:str_fdtmax_1em5}
1161: but for the case without the deuterium burning
1162: ([D/H] = 0, run MD5-noD).
1163: The dot-dashed line represents the mass-radius relation 
1164: for the case with the fiducial deuterium abundance, 
1165: [D/H] = $2.5 \times 10^{-5}$ 
1166: (run MD5, also see Fig.~\ref{fig:str_fdtmax_1em5}).
1167: }
1168: \label{fig:str_1em5_nod}
1169: \end{center}
1170: \end{figure}
1171: %----------------------------------------------------------------------%
1172: %----------------------------------------------------------------------%
1173: \begin{figure}[t]
1174:   \begin{center}
1175: \epsfig{ file=f12.eps,
1176:          angle=0,
1177:          width=4in}
1178: %lum_enuc_tsc_1em5.eps
1179: \caption{ 
1180: Same as Fig.~\ref{fig:lum_enuc_tsc_1em3} but for 
1181: the lower accretion rate of 
1182: $\dot{M}_\ast = 10^{-5}~M_{\odot}/{\rm yr}$ (run MD5). 
1183: In each panel, the stellar luminosity 
1184: $L_{\rm *}$ ({\it Top panel}, thin dotted line), 
1185: maximum luminosity within the star $L_{\rm max}$ 
1186: ({\it Middle panel}, thin coarse dashed line),
1187: and Kelvin-Helmholtz time $t_{\rm KH}$ 
1188: ({\it Bottom panel}, thin dotted line) for the case without 
1189: deuterium burning (run MD5-noD) are presented for comparison.
1190: In all panels, the shaded background shows the four characteristic 
1191: phases, as in Fig.~\ref{fig:str_fdtmax_1em5}. 
1192: }
1193: \label{fig:lum_enuc_tsc_1em5}
1194:   \end{center}
1195: \end{figure}
1196: %-----------------------------------------------------------------------%
1197: 
1198: 
1199: 
1200: \paragraph{Convection Phase}
1201: 
1202: In the low $\dot{M}_\ast$ case, the D burning begins in an earlier phase 
1203: and has a more profound effect on the evolution
1204: than in the high-$\dot{M}_\ast$ case (run MD3).
1205: As a result of the smaller radius, the temperature, as well as density, 
1206: is higher in the low $\dot{M}_\ast$ case 
1207: (cf. lower panels of Fig.~\ref{fig:str_fdtmax_1em3} and 
1208: Fig.~\ref{fig:str_fdtmax_1em5}).
1209: The maximum temperature reaches $10^6$~K as early as 
1210: $M_{\ast} \simeq 0.2~M_{\odot}$ 
1211: and active deuterium burning begins subsequently
1212: (Fig.~\ref{fig:str_fdtmax_1em5}), while in the high-$\dot{M}_\ast$ case 
1213: the D ignition does not occur until $M_{\ast} \simeq 6~M_{\odot}$ 
1214: (Fig.~\ref{fig:str_fdtmax_1em3}).
1215: Unlike in the high-$\dot{M}_\ast$ case,
1216: the deuterium burning makes most of the interior convective.
1217: This is a result of high density and thus high opacity 
1218: at the ignition of deuterium burning, occurring 
1219: at a fixed temperature of $\simeq 10^{6}$ K.
1220: Since high opacity hinders the radiative heat transport,
1221: entropy generated by deuterium burning cannot be transported radiatively.
1222: This necessarily gives rise to convection.
1223: 
1224: Figure \ref{fig:str_fdtmax_1em5} shows that
1225: the extent of the convective layers experiences a complicated history.
1226: At $M_{\ast} \simeq 0.3~M_{\odot}$, a convective layer appears 
1227: for the first time slightly off-center. 
1228: Inside the convective layer, entropy is homogenized and its value 
1229: increases owing to the D burning (Fig.~\ref{fig:slpf_1em5}, I).
1230: With this increased entropy, the outer layer is gradually incorporated 
1231: into the convective region. 
1232: Fresh deuterium is supplied from the newly incorporated layer  
1233: and spreads over the convective region, and finally is consumed 
1234: by nuclear burning at the bottom of the region. 
1235: However, the expansion of the convective region cannot be maintained 
1236: if the supply of deuterium becomes insufficient.
1237: Without the fuel, the deuterium concentration in the convective region 
1238: $f_{\rm d,cv}$ drops and the nuclear burning ceases: 
1239: the convective region disappears.   
1240: This occurs at $M_{\ast} \simeq 0.6~M_{\odot}$ 
1241: (Fig.~\ref{fig:str_fdtmax_1em5}, upper).
1242: At $M_{\ast} \simeq 0.7~M_{\odot}$, another episode of deuterium burning 
1243: begins just outside the first convective region, where fresh deuterium 
1244: still remains.
1245: The second convective region develops there and immediately reaches 
1246: the stellar surface.
1247: Although this complicated evolution of the convective regions is 
1248: different from that reported in SST80b, this difference can be 
1249: attributable to high sensitivity of the evolution on adopted 
1250: initial deuterium abundance.
1251: We discuss this in greater details in Appendix~\ref{ssec:dab}.
1252: 
1253: Since the energy generation by D burning is very sensitive 
1254: to the temperature, temperature remains constant at 
1255: $\simeq 1.5 \times 10^{6}$K during the active deuterium burning 
1256: (Fig.~\ref{fig:str_fdtmax_1em5}, lower panel).
1257: This is the so-called thermostat effect (e.g., Stahler 1988).
1258: If this works,
1259: $M_{\ast}/R_{\ast} \propto T \simeq const.$ from equation 
1260: (\ref{eq:t_typ}):
1261: the protostellar radius $R_{\ast}$ increases as $\propto M_{\ast}$.
1262: This linear increase in $R_{\ast}$ with $M_{\ast}$ can be observed 
1263: in the range $0.7M_{\odot} \lesssim M_{\ast} \lesssim 1M_{\odot}$
1264: (Fig.~\ref{fig:str_fdtmax_1em5}, upper panel).
1265: 
1266: When the deuterium concentration in the convective region
1267: drops to a few percent, the thermostat effect ceases to work.
1268: The maximum temperature starts increasing again and 
1269: the radial expansion slows down after that.
1270: Owing to the high temperature, opacity decreases 
1271: in the deep interior with increasing stellar mass.
1272: This allows efficient radiative energy transport and 
1273: renders a large portion of the interior radiative.
1274: The radiative core gradually extends to the outer layer,  
1275: and the convective layer moves outward.
1276: Since fresh deuterium is no longer supplied to the radiative interior 
1277: by the mixing, deuterium is soon exhausted there. 
1278: The expansion of the radiative interior occurs in a later phase 
1279: in the calculation of PS91.
1280: This difference can be attributed to different initial models: 
1281: whereas PS91 assumed a fully convective star of $1~M_{\odot}$
1282: as an initial model, we started calculation 
1283: from the initial mass $M_{\ast,0} = 0.01~M_{\odot}$.
1284: At $M_{\ast} = 1~M_{\odot}$, the protostar is not fully convective in our case.
1285: The outer layers of $M > 0.25~M_{\odot}$ are convective, 
1286: where the D burning occurs at the bottom, while 
1287: the remaining inner core is radiative at this moment.
1288: We confirmed that, starting from the same initial model, 
1289: the evolution becomes similar to that of PS91 
1290: as discussed in Appendix \ref{ssec:psini}.
1291: 
1292: Because of the early ignition of deuterium burning, 
1293: entropy and luminosity profiles in this phase look rather different from
1294: those in the high $\dot{M}_\ast$ case 
1295: (see Fig.~\ref{fig:slpf_1em3} and \ref{fig:slpf_1em5}).
1296: For easier comparison, let us see the case without deuterium burning, 
1297: where the similar profiles are in fact reproduced 
1298: (Fig.~\ref{fig:slpf_1em5_nod}).
1299: However, one important difference exists, the spiky structure in the 
1300: entropy profiles near the surface.
1301: This shows significant entropy loss near the surface 
1302: in the low $\dot{M}_\ast$ case, which leads to the low entropy within the star.
1303: The reason can be seen clearly in Figure \ref{fig:t_comp_1em5}, where 
1304: the comparison between the local accretion timescale $t_{\rm acc,s}$ and 
1305: cooling timescale $t_{\rm cool,s}$ (eq.s \ref{eq:taccs} and \ref{eq:tcools})
1306: is presented.
1307: In contrast to the the case of high $\dot{M}_\ast$, 
1308: in most of the settling layer, $t_{\rm acc,s}$ is shorter
1309: than $t_{\rm cool,s}$: accreted matter loses entropy 
1310: radiatively near the surface before settling into the adiabatic interior.
1311: 
1312: 
1313: \paragraph{Swelling Phase}
1314: 
1315: In the short period of $3~M_{\odot} \lesssim M_{\ast} \lesssim 4~M_{\odot}$, 
1316: the stellar radius jumps up by a factor of two. 
1317: At the same time, the whole interior becomes radiative 
1318: (Fig.~\ref{fig:str_fdtmax_1em5}, upper). 
1319: Although deuterium burning continues near the surface, 
1320: it does not drive convection any more.
1321: 
1322: The cause of this swelling has been attributed to the shell burning 
1323: of deuterium, which also occurs in our calculation (Palla \& Stahler 1990).
1324: However, we conclude that the main driver of the swelling is 
1325: not the D shell burning, but the outward entropy transportation 
1326: as in the case of the high accretion rate.
1327: The reasons are as follows.
1328: First, evolutionary features of entropy and luminosity profiles 
1329: in this phase are similar to those in the swelling
1330: phase in the high $\dot{M}_\ast$ case 
1331: (run MD3; cf. Figs.\ref{fig:slpf_1em3}, \ref{fig:slpf_1em5}):
1332: in both cases, entropy decreases in the deep interior 
1333: and increases significantly near the stellar surface; 
1334: simultaneously, the peak of the luminosity distribution moves to the surface
1335: as the luminosity wave.
1336: Second, even in the run without deuterium burning (run MD5-noD),
1337: this swelling occurs as shown in Figure \ref{fig:str_1em5_nod}: 
1338: the protostar swells up at $M_{\ast} \simeq 3~M_{\odot}$.
1339: 
1340: The interior luminosity $L_{\ast}$ jumps up at 
1341: $M_{\ast} \simeq 0.7~M_{\odot}$ with the arrival 
1342: of the convective region at the surface 
1343: as a result of high efficiency of convective heat transfer 
1344: (Fig.~\ref{fig:lum_enuc_tsc_1em5}, upper panel).
1345: Although this rise in $L_{\ast}$ decreases $t_{\rm KH}$,  
1346: $t_{\rm KH}$ remains much longer than $t_{\rm acc}$ 
1347: (Fig.~\ref{fig:lum_enuc_tsc_1em5}, lower panel).
1348: It is not until another sudden rise in $L_{\ast}$ at 
1349: $\simeq 3-4 M_{\odot}$ due to the arrival of the luminosity wave 
1350: that $t_{\rm KH}$ becomes as short as $t_{\rm acc}$. 
1351: The maximum radius is attained at this epoch since the swelling is 
1352: also caused by the luminosity wave, as seen above.
1353: Then, the analytical argument leading to equation (\ref{eq:m_rmax}) 
1354: also holds in this case:
1355: for $\dot{M}_\ast = 10^{-5}~M_{\odot}/{\rm yr}$, this leads to 
1356: the maximum radius at $M_{\rm *, rmax} \simeq 3.9~M_{\odot}$, 
1357: which agrees with our numerical result.
1358: 
1359: Simultaneous returning to the radiative structure 
1360: with the swelling has been explained by \citet{PS90}.
1361: The luminosity from the steady deuterium burning
1362: is comparable to the accretion luminosity by chance:
1363: \begin{equation}
1364: L_{\rm D, st} = \dot{M}_\ast \delta_{\rm D} \sim
1365:                \dot{M}_\ast \frac{G M_{\ast}}{R_{\ast}} = L_{\rm acc} .
1366: \label{eq:ldst_lacc}
1367: \end{equation}
1368: The epoch of the radial turn-around obeys equation
1369: (\ref{eq:m_rmax}), which is derived by equating
1370: $t_{\rm acc}$ and $t_{\rm KH,lmax} = G M_*^2/R_* L_{\rm max}$.
1371: This means that $L_{\rm max}$ becomes comparable to $L_{\rm acc}$ 
1372: and thus to $L_{\rm D,st}$ in the swelling phase.
1373: Hence, the radiative luminosity now has a capacity to transport
1374: the energy generated by deuterium burning: 
1375: $L_{\rm max} \sim L_{\rm D,st}$.  
1376: The deuterium burning no longer gives rise to any convective layer 
1377: after the swelling. 
1378: 
1379: 
1380: 
1381: \paragraph{Later Phases}
1382: 
1383: Subsequently, the protostar enters the KH contraction phase.
1384: The evolution thereafter is similar to the previous high-$\dot{M}_\ast$ case. 
1385: The maximum temperature within the star increases with the contraction, and
1386: hydrogen burning begins at $M_{\ast} \simeq 7~M_{\odot}$.
1387: 
1388: %-----------------------------------------------------------------------%
1389: \begin{figure}[t]
1390:   \begin{center}
1391: \epsfig{ file=f13.eps,
1392:          angle=0,
1393:          width=4in}
1394: %prec_1em5.eps
1395: \caption{Same as Fig.~\ref{fig:prec_1em3} but for the lower
1396: accretion rate of $\dot{M}_\ast = 10^{-5}~M_{\odot}/{\rm yr}$ (run MD5). 
1397: The shaded background shows the four characteristic phases,
1398: as in Fig.~\ref{fig:str_fdtmax_1em5}. 
1399: }
1400: \label{fig:prec_1em5}
1401:   \end{center}
1402: \end{figure}
1403: %------------------------------------------------------------------------%
1404: 
1405: 
1406: \paragraph{Evolution of the Radiative Precursor}
1407: 
1408: Figure \ref{fig:prec_1em5} shows
1409: that the accretion flow is optically thin except 
1410: in the early phase of $M_{\ast} < 1~M_{\odot}$
1411: at the low accretion rate of $10^{-5}~M_{\odot}/{\rm yr}$.
1412: With the expansion at the active deuterium burning, 
1413: the stellar surface emerges above 
1414: the photosphere at $M_{\ast} \simeq 1~M_{\odot}$ 
1415: and the radiative precursor disappears thereafter. 
1416: Without the precursor, 
1417: the stellar surface is directly visible, 
1418: with effective temperature rising remarkably 
1419: in the KH contraction phase, i.e., for 
1420: $M_{\ast} > 4 M_{\odot}$.
1421: This is in stark contrast to the case of 
1422: high accretion rate $10^{-3}~M_{\odot}/{\rm yr}$, 
1423: where the effective temperature remains several thousand K
1424: throughout the evolution (Fig.~\ref{fig:prec_1em3}).
1425: 
1426: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1427: \subsection{Dependence on Mass Accretion Rates}
1428: \label{ssec:md_dep}
1429: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1430: 
1431: %-------------------------------------------------------------------%
1432: \begin{figure}[t]
1433:   \begin{center}
1434: \epsfig{ file=f14.eps,
1435:          angle=0,
1436:          width=4.5in}
1437: %m_rtmax.eps
1438: \caption{
1439: Evolution of the protostellar radii (upper panel) 
1440: and the maximum temperatures within the stars (lower panel) 
1441: for different mass accretion rates. 
1442: The cases of $\dot{M}_\ast=10^{-6}$ (run MD6; solid), 
1443: $10^{-5}$ (MD5; dashed), $10^{-4}$ (MD4; dot-dashed), 
1444: and $10^{-3} M_{\odot}/{\rm yr}$ (MD3; dotted line) are depicted.
1445: Also shown by the thin lines are the runs without 
1446: deuterium burning (``noD'' runs) for the same accretion rates.
1447: In both panels, all the curves finally converge to single lines, 
1448: which is the relations for the main-sequence stars.
1449: }
1450: \label{fig:m_rtmax}
1451:   \end{center}
1452: \end{figure}
1453: %--------------------------------------------------------------------%
1454: %--------------------------------------------------------------------%
1455: \begin{figure}[t]
1456: \begin{center}
1457: \epsfig{ file=f15.eps,
1458:          angle=0,
1459:          width=4in}
1460: %s_1msun_nod.eps
1461: \caption{
1462: Distributions of the specific entropy for the case without deuterium
1463: (``no-D'' runs in Table 1) at the epoch of $M_{\ast} = 1~M_{\odot}$ 
1464: for different accretion rates. 
1465: Histories of the post-shock values are also plotted by the dotted lines
1466: for the cases with $\dot{M}_\ast = 10^{-6}$ and 
1467: $10^{-5}~M_{\odot}/{\rm yr}$.
1468: For the cases with $\dot{M}_\ast = 10^{-4}$ and $10^{-3}~M_{\odot}/{\rm yr}$, 
1469: those curves completely overlap with the entropy distributions
1470: at $M_{\ast} = 1~M_{\odot}$.
1471: }
1472: \label{fig:s_1msun_nod}
1473: \end{center}
1474: \end{figure}
1475: %--------------------------------------------------------------------%
1476: %--------------------------------------------------------------------%
1477: \begin{figure}[h]
1478:   \begin{center}
1479: \epsfig{ file=f16.eps,
1480:          angle=0,
1481:          width=4in}
1482: %tcl_tacc_min.eps
1483: \caption{ 
1484: Evolution of the minimum value 
1485: of the ratio $t_{\rm cool,s}(m)/t_{\rm acc,s}(m)$ 
1486: in the surface settling layer.
1487: The no-D cases are shown for 
1488: $\dot{M}_\ast=10^{-6}$ (solid), $10^{-5}$ (dashed), 
1489: $10^{-4}$ (dot-dashed), and $10^{-3}~M_{\odot}/{\rm yr}$ (dotted).
1490: For the ratio $t_{\rm cool,s}(m)/t_{\rm acc,s}(m)$ exceeding unity, 
1491: the accreted material settles into the interior without losing 
1492: its post-shock entropy radiatively. 
1493: }
1494: \label{fig:r_tcl_tacc}
1495:   \end{center}
1496: \end{figure}
1497: %--------------------------------------------------------------------%
1498: 
1499: Here, we summarize the effects of the different accretion rates 
1500: on the protostellar evolution.
1501: The mass-radius relations of protostars with 
1502: accretion rates $\dot{M}_\ast=10^{-6}$, 
1503: $10^{-5}$, $10^{-4}$, and $10^{-3} M_{\odot}/{\rm yr}$
1504: are shown in the upper panel of Figure \ref{fig:m_rtmax}.
1505: 
1506: First, the protostellar radius is larger for the higher accretion
1507: rate at the same protostellar mass, 
1508: as discussed in \S~\ref{ssec:md_1em3}.  
1509: For example, at $M_{\ast} = 1 M_{\odot}$, the radius
1510: $R_{\ast} \simeq 25 R_{\odot}$ 
1511: for $10^{-3} M_{\odot}/{\rm yr}$, while it is only 
1512: $2.5 R_{\odot}$ for $10^{-6} M_{\odot}/{\rm yr}$.
1513: For easier comparison, the mass-radius relations in the no-deuterium 
1514: cases are also shown by thin lines in Figure \ref{fig:m_rtmax}.
1515: The deuterium burning enhances the stellar radius,  
1516: in particular, in low $\dot{M}_\ast$ cases
1517: (see in this section below).
1518: Nevertheless, the trend of larger $R_{\ast}$ for higher $\dot{M}_\ast$ 
1519: remains valid even without the deuterium burning.
1520: For simplicity, we do not include the effect of 
1521: deuterium burning in the following argument.
1522: The origin of this trend is higher specific entropy for 
1523: higher accretion rate.
1524: Recall that the stellar radius is larger with the high
1525: entropy in the stellar interior as shown by equation (\ref{eq:r_srel}).
1526: The entropy distributions at $M_{\ast} = 1 M_{\odot}$ 
1527: for the no-D runs are shown in Figure \ref{fig:s_1msun_nod}, 
1528: which indeed demonstrates that 
1529: the entropy is higher for the higher $\dot{M}_\ast$ cases.
1530: The interior entropy is set in the post-shock settling layer 
1531: where radiative cooling can reduce the entropy from 
1532: the initial post-shock value 
1533: as observed in spiky structures near the surface 
1534: (e.g., the case of $\dot{M}_\ast \lesssim 10^{-5}~M_{\odot}/{\rm yr}$ 
1535: in Fig.~\ref{fig:s_1msun_nod}). 
1536: For efficient radiative cooling,
1537: the local cooling time $t_{\rm cool, s}$ in the settling layer
1538: must be shorter than the local accretion time $t_{\rm acc,s}$,
1539: as explained in \S~\ref{ssec:md_1em3} and \ref{ssec:md_1em5}.
1540: Figure \ref{fig:r_tcl_tacc} shows that the ratio 
1541: $t_{\rm cool,s}/t_{\rm acc,s}$ 
1542: is smaller for lower $\dot{M}_\ast$ and becomes less than unity
1543: for $\lesssim 10^{-5}~M_{\odot}/{\rm yr}$.
1544: Thus the accreted gas cools radiatively in the settling layer 
1545: in low $\dot{M}_\ast$ ($\lesssim 10^{-5} M_{\odot}/{\rm yr}$) cases, 
1546: while in the higher $\dot{M}_\ast$ cases, the gas is embedded into the stellar 
1547: interior without losing the post-shock entropy.
1548: The difference in entropy among high 
1549: $\dot{M}_\ast$ ($\gtrsim 10^{-4} M_{\odot}/{\rm yr}$) cases comes from 
1550: the higher post-shock entropy for higher $\dot{M}_\ast$.
1551: In those cases, the protostar is enshrouded with the 
1552: radiative precursor, whose optical depth is 
1553: larger for the higher accretion rate 
1554: (c.f.,Figs.\ref{fig:prec_1em3} and \ref{fig:prec_1em5}, bottom).
1555: Consequently, more heat is trapped in the precursor, 
1556: which leads to the higher post-shock entropy.
1557: 
1558: Figure \ref{fig:m_rtmax} shows not only that 
1559: not only the swelling of the stellar radius occurs even in the ``no-D'' runs,
1560: but also that their timings are almost the same as in the cases with D.
1561: This supports our view that the cause of the swelling is 
1562: outward entropy transport by the luminosity wave
1563: rather than the deuterium shell-burning 
1564: (\S~\ref{ssec:md_1em3} and \ref{ssec:md_1em5}).
1565: Also, in all cases, the epoch of the swelling obeys equation 
1566: (\ref{eq:m_rmax}), which is derived by equating
1567: the accretion time $t_{\rm acc}$ and the KH time $t_{\rm KH,lmax}$.
1568: 
1569: Another clear tendency is that
1570: the deuterium burning begins at higher stellar mass for 
1571: higher accretion rate.
1572: Similarly, the onset of hydrogen burning and subsequent arrival at the ZAMS 
1573: are postponed until higher protostellar mass for the higher accretion rate.
1574: For example, the ZAMS arrival is at $M_{\ast} \simeq 4 M_{\odot}$ 
1575: (40$M_{\odot}$) for $10^{-6}$ ($10^{-3}$, respectively) 
1576: $M_{\odot}/{\rm yr}$.
1577: These delayed nuclear ignitions reflect the lower temperature 
1578: within the protostar for the higher $\dot{M}_\ast$ at the same $M_{\ast}$ 
1579: (Fig.~\ref{fig:m_rtmax}, lower panel).
1580: Since $R_{\ast}$ is larger with higher $\dot{M}_\ast$ at a given $M_{\ast}$,
1581: typical temperature is lower within the star.
1582: Substituting numerical values in equation (\ref{eq:t_typ}), we obtain
1583: \begin{equation}
1584: T_{\rm max} \sim 10^7~{\rm K} 
1585:                  \left( \frac{M_{\ast}}{M_{\odot}}   \right)
1586:                  \left( \frac{R_{\ast}}{R_{\odot}}   \right)^{-1} ,
1587: \label{eq:t_max_aly}
1588: \end{equation}
1589: which is a good approximation for our numerical results.
1590: 
1591: Finally, we consider the significance of deuterium burning 
1592: for protostellar evolution. Comparing  with ``no-D'' runs,
1593: we find that the effect of deuterium burning not only appears
1594: later, but also becomes weaker with increasing accretion rates.
1595: At the low accretion rates
1596: $\dot{M}_\ast \leq 10^{-5}~M_{\odot}/{\rm yr}$, 
1597: convection spreads into a wide portion of the star owing to the vigorous 
1598: deuterium burning. 
1599: In addition, the thermostat effect works and the stellar radius increases 
1600: linearly during this period. At the higher accretion rates 
1601: $\dot{M}_\ast \geq 10^{-4}~M_{\odot}/{\rm yr}$, on the other hand, 
1602: deuterium burning affects the protostellar evolution only slightly. 
1603: Convective regions do not appear even after the beginning 
1604: of deuterium burning for $\dot{M}_\ast = 10^{-3} M_{\odot}/{\rm yr}$ 
1605: (\S~\ref{ssec:md_1em3}).
1606: 
1607: A key quantity for understanding this variation is the
1608: maximum radiative luminosity at the ignition of deuterium burning.
1609: The maximum radiative luminosity within the star $L_{\rm max}$
1610: at deuterium ignition can be evaluated by 
1611: eliminating $R_{\ast}$ and $M_{\ast}$ in equation (\ref{eq:l_max_a})
1612: with (\ref{eq:r_sps86}), (\ref{eq:t_max_aly}), and setting
1613: $T_{\rm max} \sim 10^6$~K: 
1614: \begin{equation}
1615: L_{\rm max} \sim 10^3~L_{\odot} 
1616:                     \left( \frac{\dot{M}_\ast}{10^{-3}~M_{\odot}/{\rm yr}} 
1617:                     \right)^{2.8} .  
1618: \end{equation}
1619: On the other hand, the luminosity from deuterium burning is 
1620: (eq. \ref{eq:l_dst})
1621: $L_{\rm D,st} \sim 10^3~L_{\odot}~(\dot{M}_\ast/10^{-3}~M_{\odot}/{\rm yr})$. 
1622: For accretion rate $\dot{M}_\ast \geq 10^{-3}~M_{\odot}/{\rm yr}$, 
1623: $L_{\rm max}> L_{\rm max, D}$:
1624: radiation is able to carry away all the entropy generated by 
1625: the deuterium burning. 
1626: Therefore, the protostellar interior remains radiative 
1627: in the high $\dot{M}_\ast$ cases (e.g., run MD3, see \S~\ref{ssec:md_1em3}), 
1628: while convection is excited for the lower $\dot{M}_\ast$ 
1629: (e.g., run MD5, see \S~\ref{ssec:md_1em5}).
1630: 
1631: 
1632: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1633: \subsection{Comparison with Primordial Star Formation}
1634: \label{ssec:metal}
1635: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1636: 
1637: %-----------------------------------------------------------------%
1638: \begin{figure}[t]
1639:   \begin{center}
1640: \epsfig{ file=f17.eps,
1641:          angle=0,
1642:          width=4in}
1643: %m_r_z.eps
1644: \caption{The protostellar mass-radius relations for different
1645: metallicities.
1646: The cases of the solar (dotted; runs MD5 and MD3) 
1647: and zero (solid; MD5-z0 and MD3-z0) metallicities 
1648: are presented 
1649: for the accretion rates $\dot{M}_\ast=10^{-5}$ and $10^{-3} 
1650: M_{\odot}/{\rm yr}$.
1651: }
1652: \label{fig:m_r_z}
1653:   \end{center}
1654: \end{figure}
1655: %-----------------------------------------------------------------%
1656: 
1657: Besides the massive star formation in the local universe,  
1658: high accretion rates have been invoked in the primordial 
1659: star formation in the early universe.
1660: Without important metal coolants, the temperature 
1661: remains at several hundred K in the primordial gas 
1662: \citep[e.g.,][]{SZ67, MST67, PSS83}.
1663: The accretion rate, which is given by  
1664: \begin{equation}
1665: \dot{M}_\ast \simeq \frac{c_{\rm s}^3}{G} 
1666: = 10^{-3} M_{\sun}/{\rm yr}
1667: \left( \frac{T}{500{\rm K}} \right)^{3/2},
1668: \end{equation}
1669: where $c_{\rm s}$ and $T$ are the sound speed and temperature, respectively, 
1670: in star-forming clouds, 
1671: becomes as high as $10^{-3}M_{\odot}/{\rm yr}$ even without turbulence 
1672: (SPS86).
1673: As a result of similar accretion rates, 
1674: protostellar evolution of the primordial stars well resembles  
1675: with those of the high $\dot{M}_\ast$ cases studied in this paper  
1676: (SPS86 ; Omukai \& Palla 2001,2003).
1677: The mass-radius relations for the primordial ($Z=0$) protostars are 
1678: presented in Figure \ref{fig:m_r_z}, along with their solar-metallicity 
1679: ($Z=0.02$) counterparts for comparison.
1680: This indicates that basic evolutionary features are 
1681: similar despite differences in metallicity:
1682: as in the present-day protostars, 
1683: primordial protostars pass through the same four characteristic
1684: phases (see \S~\ref{ssec:md_1em3} and \ref{ssec:md_1em5}).
1685: The following two differences, however, still exist.
1686: First, the swelling of the protostar and subsequent 
1687: transition to the KH contraction
1688: occur earlier in the $Z=0$ cases.
1689: For example, for $\dot{M}_\ast=10^{-3}M_{\sun}/{\rm yr}$, 
1690: the maximum radius is attained and the KH contraction 
1691: begins at $10M_{\sun}$ in the $Z=0.02$ case, while it is $7M_{\sun}$
1692: in the $Z=0$ case.
1693: Recall that the transition to those phases is caused by 
1694: the decrease in opacity with increasing $M_{\ast}$ (\S~\ref{ssec:md_1em3}).
1695: Owing to the lack of heavy elements, the opacity is lower 
1696: in the primordial gas at the same density and temperature.
1697: Efficient radiative heat transport begins earlier, which 
1698: results in the earlier transitions in the evolutionary phases.
1699: Second, the stellar radius in the main-sequence phase 
1700: is smaller as a result of higher interior temperature 
1701: ($\simeq 10^{8}$K) for the primordial stars.
1702: Due to the lack of C and N, sufficient energy production 
1703: by the CN cycle is not available at a temperature similar to that 
1704: of the solar metallicity case ($\simeq 10^{7}$K).
1705: The KH contraction is not halted until 
1706: the temperature reaches $\simeq 10^{8}$K, where 
1707: a small amount of carbon produced by the He burning
1708: enables the operation of the CN cycle
1709: \citep{EC71, OP03}.
1710: Since more contraction is needed, the arrival at the ZAMS can be later 
1711: in the $Z=0$ case, despite the earlier beginning of the KH contraction:
1712: for example, it is $30M_{\odot}$ ($50M_{\odot}$) in the $Z_{\odot}$ 
1713: ($Z=0$, respectively) case.
1714: 
1715: 
1716: 
1717: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1718: \subsection{Radiation Pressure Barrier on the Accreting Gas Envelope}
1719: \label{ssec:edd}
1720: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1721: 
1722: %------------------------------------------------------------------------%
1723: \begin{figure}[t]
1724:   \begin{center}
1725: \epsfig{ file=f18.eps,
1726:          angle=0,
1727:          width=4in}
1728: %high_mdot.eps
1729: \caption{ 
1730: {\it Upper panel} :
1731: Evolution of the protostellar radii for cases
1732: at the very high accretion rates
1733: of $10^{-3}$ (solid, run MD3), $3 \times 10^{-3}$ 
1734: (dashed, MD3x3), $4 \times 10^{-3}$ 
1735: (dot-dashed, MD4x3), and $6 \times 10^{-3}~M_{\odot}/{\rm yr}$ 
1736: (dotted, MD6x3).
1737: For comparison, we also plot the case of a
1738: primordial ($Z = 0$) protostar with $3 \times 10^{-3}~M_{\odot}/{\rm yr}$ 
1739: (thin dashed, MD3x3-z0).
1740: {\it Lower panel} : Evolution of a ratio between the total luminosity
1741: of protostars and Eddington luminosity for the same cases. 
1742: }
1743: \label{fig:m_r_blowup}
1744:   \end{center}
1745: \end{figure}         
1746: %--------------------------------------------------------------------%
1747: 
1748: The higher the mass accretion rate, the more massive the 
1749: protostar grows before the arrival at the ZAMS phase
1750: (\S~\ref{ssec:md_1em3} and \ref{ssec:md_dep}).
1751: However, this trend does not continue unlimitedly: 
1752: for a rate exceeding a threshold value, stars have an upper mass 
1753: limit set by the radiation pressure onto the radiative 
1754: precursor. 
1755: 
1756: The evolution of the protostellar radius is shown in 
1757: Figure \ref{fig:m_r_blowup} (upper panel) for cases with 
1758: very high accretion rates $>10^{-3}~M_{\odot}/{\rm yr}$.
1759: Up to a certain moment in the KH contraction phase, 
1760: the evolution in all cases proceeds in a qualitatively 
1761: similar way 
1762: although the radius is larger and the onset of swelling 
1763: and the KH contraction are later for higher $\dot{M}_\ast$ cases.
1764: However, for accretion rates 
1765: $\dot{M}_\ast > 3 \times 10^{-3} M_{\odot}/{\rm yr}$, 
1766: the contraction is halted at $\simeq 40M_{\sun}$, followed by
1767: abrupt expansion at $\simeq 50M_{\sun}$.
1768: In such cases, further steady accretion onto the star is not possible
1769: owing to strong radiative pressure exerted onto the radiative precursor.
1770: This phenomenon has been found for the primordial star formation 
1771: by \citet{OP01, OP03}.
1772: They found that abrupt expansion occurs when the total
1773: luminosity from a protostar,
1774: $L_{\rm tot} = L_{\ast} + L_{\rm acc}
1775: =L_{\ast} + G M_{\ast} \dot{M}_{\ast}/R_{\ast}$, 
1776: becomes close to the Eddington luminosity 
1777: $L_{\rm Edd}$.
1778: The strong radiation pressure decelerates the accretion flow and then 
1779: reduces the ram pressure onto the stellar surface. 
1780: With reduced inward ram pressure,  
1781: an outward thrust by the interior radiation 
1782: blows a thin surface layer away.
1783: 
1784: The critical accretion rate $\dot{M}_{\ast}$ is defined as the rate above 
1785: which the abrupt stellar expansion occurs in the KH contraction phase.
1786: By using the condition that $L_{\rm tot}$ reaches $L_{\rm Edd}$ before 
1787: the protostar reaches the ZAMS, \citep{OP03} derived the 
1788: critical rate as follows:
1789: \begin{equation}
1790: L_{\rm ZAMS} + \frac{G M_{\rm ZAMS} \dot{M}_{\rm \ast, cr}}{R_{\rm ZAMS}} 
1791: = L_{\rm Edd},
1792: \label{eq:lcr_pm}
1793: \end{equation}
1794: that is,
1795: \begin{equation}
1796: \dot{M}_{\rm \ast, cr} = \frac{4 \pi c R_{\rm ZAMS}}{\kappa_{\rm esc}}
1797:                    \left( 1 - \frac{L_{\rm ZAMS}}{L_{\rm Edd}}  \right)
1798:                    ,
1799: \label{eq:mcr_pm}
1800: \end{equation}
1801: where $\kappa_{\rm esc}$ is the electron-scattering opacity, 
1802: and the suffix ``ZAMS'' means quantities of the ZAMS stars. 
1803: Although the critical rate $\dot{M}_{\rm \ast, cr}$ 
1804: by equation (\ref{eq:mcr_pm}) depends on 
1805: the stellar mass, the dependence is very weak and 
1806: $\dot{M}_{\rm \ast, cr} \simeq 4 \times 10^{-3} M_{\odot}/{\rm yr}$ 
1807: for the primordial stars.
1808: Since the ZAMS radius increases with metallicity 
1809: (see \S~\ref{ssec:metal}), 
1810: $\dot{M}_{\rm \ast, cr}$ is expected to increase with metallicity.
1811: For example, for $Z=10^{-4}$, 
1812: $\dot{M}_{\rm \ast, cr}=9 \times 10^{-3} M_{\odot}/{\rm yr}$ by 
1813: equation (\ref{eq:mcr_pm}), and the evolutionary calculation 
1814: demonstrated that abrupt stellar expansion occurs in fact 
1815: for higher accretion rates \citep{OP03}.
1816: Our calculations, however, indicate that the critical accretion rate in 
1817: the solar metallicity case is $\simeq 3 \times 10^{-3} 
1818: M_{\odot}/{\rm yr}$, slightly smaller than that of primordial protostars.
1819: This apparent contradiction comes from the fact that, 
1820: in the solar metallicity case, opacity in the accreting envelope is 
1821: higher than that of the electron scattering one $\kappa_{\rm es}$ 
1822: postulated in the above derivation for the critical accretion rate.
1823: In fact, the violent stellar expansion is found to take place
1824: with a total luminosity of about a half of the Eddington luminosity, 
1825: as shown in the lower panel of Figure \ref{fig:m_r_blowup}.
1826: A marginal case is Run MD3x3, where the contraction is 
1827: about to be reversed, but the star barely reaches 
1828: the ZAMS after some loitering. 
1829: In this case, the Eddington ratio 
1830: $f_{\rm Edd}=L_{\rm tot}/L_{\rm Edd}$ slightly exceeds 
1831: 0.5 at $M_{\ast} \simeq 45~M_{\odot}$.
1832: For comparison, 
1833: the case of the primordial star is shown for the same accretion rate, 
1834: $\dot{M}_\ast = 3 \times 10^{-3}~M_{\odot}/{\rm yr}$, 
1835: which is slightly lower than the critical accretion rate (run MD3x3-z0).
1836: Because of the earlier transition to the KH contraction phase, 
1837: the Eddington ratio also begins to increase
1838: earlier than the present-day counterpart (run MD3x3).
1839: Despite high luminosity almost reaching the Eddington limit, 
1840: the KH contraction is not reversed and the protostar can grow 
1841: further.
1842: Taking into account the higher opacity in the envelope, 
1843: we add the Eddington factor of $0.5$ to 
1844: equation (\ref{eq:lcr_pm}):
1845: \begin{equation}
1846: L_{\rm ZAMS} + \frac{G M_{\rm ZAMS} \dot{M}_{\rm \ast, cr}}{R_{\rm ZAMS}}
1847: \simeq 0.5 L_{\rm Edd}
1848: \label{eq:lcr_present}
1849: \end{equation}
1850: for present-day protostars. 
1851: The critical accretion rate becomes 
1852: $\dot{M}_{\rm \ast, cr} \simeq 3 \times 10^{-3} M_{\odot}/{\rm yr}$,
1853: which agrees well with our numerical results. 
1854: Although this critical rate is a function of the stellar mass, 
1855: its dependence is weak as in the primordial case.
1856: 
1857: Our results suggest an upper
1858: mass limit of PMS stars at $\simeq 60~M_\odot$, 
1859: which corresponds to the mass where the protostar reaches 
1860: the ZAMS at the critical accretion rate of 
1861: $M_{\rm cr} \sim 3 \times 10^{-3}~M_\odot/{\rm yr}$.
1862: More massive PMS stars cannot be formed by steady
1863: mass accretion. With higher accretion rates, the steady 
1864: accretion is halted by the abrupt expansion at
1865: $M_\ast \simeq 50~M_\odot$. 
1866: What happens after the abrupt expansion is speculated
1867: upon in the next section.
1868: 
1869: 
1870: 
1871: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1872: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1873: \section{Upper Limit on the Stellar Mass by Radiation Feedback}
1874: \label{sec:stopacc}
1875: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1876: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1877: 
1878: %-----------------------------------------------------------------%
1879: \begin{figure}[t]
1880:   \begin{center}
1881: \epsfig{ file=f19.eps,
1882:          angle=0,
1883:          width=5in}
1884: %wc.eps
1885: \caption{ 
1886: Limits on the stellar mass for each accretion rate by various
1887: feedback effects. 
1888: The three horizontal arrows represent 
1889: protostellar evolution at the accretion rate of (i)
1890: $\dot{M}_\ast = 10^{-5}$ (run MD5), (ii) $10^{-3}$ (run MD3), 
1891: and (iii) $6 \times 10^{-3}~M_{\odot}/{\rm yr}$ (run MD6x3).
1892: Also shown are the stellar masses where the protostellar radii
1893: reach the maximum (dashed line) and where the protostars arrive
1894: at the ZAMS (dotted line).
1895: The region shaded by light gray represents where the
1896: total luminosity of the protostar, 
1897: $L_{\rm tot} = L_{\ast} + L_{\rm acc}$ exceeds the dust Eddington
1898: luminosity $L_{\rm Edd,d}$, given by equation (\ref{eq:eddd}).
1899: The lower-right hatched region represents where the ram
1900: pressure of the free-fall flow is insufficient to overcome 
1901: the radiation pressure barrier at the dust destruction front.
1902: The upper-right hatched region denotes the similar
1903: prohibited region but for the radiation pressure acting on
1904: a gas envelopes just around the protostar. 
1905: The extra arrow (iv) indicates a possible path along which
1906: the protostar will evolve after the gas envelope is disrupted
1907: by the radiation pressure (see text).
1908: }
1909: \label{fig:wc}
1910:   \end{center}
1911: \end{figure}         
1912: %--------------------------------------------------------------------%
1913: 
1914: As long as the central star is not so massive and
1915: its feedback on the accreting envelope is weak, 
1916: the accretion proceeds at a rate set by the condition of 
1917: the parent star-forming core. 
1918: Our assumption concerning the accretion at a pre-determined rate 
1919: remains valid.
1920: However, when the central star grows massive enough and
1921: its feedback, e.g. radiative pressure on 
1922: the dusty outer envelope, thermal pressure in the HII region, etc., 
1923: can no longer be neglected,
1924: the accretion rate is now regulated by the stellar feedback. 
1925: The mass accretion is eventually terminated  
1926: and the final stellar mass is set at this moment.
1927: 
1928: Among possible types of feedback, the radiation pressure 
1929: on the outer dusty envelope is considered to be the most important 
1930: in terminating the protostellar accretion (e.g., WC87).
1931: Since no such outer envelope has been included in our calculation, 
1932: we evaluate here the epoch of accretion termination  
1933: by stellar feedback with the assumption that the accretion rate 
1934: is roughly constant up to this moment.
1935: 
1936: Most of the stellar radiation, emitted in the optical or UV 
1937: range, is absorbed in a thin innermost layer 
1938: of the dusty envelope, i.e., the dust destruction front, 
1939: and the outward momentum of photons 
1940: is transfered to the flow (see Fig.~\ref{fig:prost_schem}).
1941: Overcoming this radiation pressure barrier is a necessary 
1942: condition for continuing accretion.
1943: The condition that the outward radiation pressure falls below
1944: the inward ram pressure of the accretion flow is expressed as
1945: \begin{equation}
1946: \frac{L_{\rm tot}}{4 \pi R_{\rm d}^2 c} < \rho u^{2},
1947: \end{equation}
1948: where $R_{\rm d}$ is the radius of the dust destruction front, 
1949: and $\rho$ and $u$ are the flow density and velocity there.
1950: The limit on the stellar mass 
1951: by this condition is shown in Figure \ref{fig:wc}
1952: as a function of the mass accretion rate (i.e, ``$P_{\rm ram}$ limit'').
1953: Here, we have assumed that the flow is in the free-fall.
1954: In Figure \ref{fig:wc}, the epochs of the maximum swelling and the arrival 
1955: at the ZAMS are also presented from our results. 
1956: Except in cases with very high accretion rates 
1957: ($> 3 \times 10^{-3}~M_{\odot}/{\rm yr}$; \S~\ref{ssec:edd}),
1958: the radiation pressure limit is attained 
1959: after the stars reach the ZAMS.
1960: This verifies previous authors' treatment in assuming 
1961: that the central star is already in the ZAMS phase when 
1962: the feedback becomes important (e.g., WC87). 
1963: 
1964: In reality, the accretion flow is decelerated
1965: before reaching the dust destruction front. 
1966: The radiation absorbed at the dust destruction
1967: front is re-emitted in infrared wavelengths and 
1968: this re-emitted radiation imparts pressure onto 
1969: the outer flow.
1970: This radiation pressure exceeds the gravitational pull of the star 
1971: if the luminosity exceeds (McKee \& Ostriker 2007):
1972: \begin{equation}
1973: L_{\rm Edd,d} = \frac{4 \pi c G M_{\ast}}{\kappa_{\rm d, IR}}
1974: \simeq  1600 L_{\odot} \left( \frac{M_{\ast}}{M_{\odot}}  \right), 
1975: \label{eq:eddd}
1976: \end{equation}
1977: which is the ``Eddington luminosity'' for dusty gas, 
1978: calculated by using the dust opacity for the far-infrared light
1979: $\kappa_{\rm d, IR} \simeq 8~{\rm cm^2 / g}$  
1980: instead of the electron-scattering opacity.
1981: The adopted value $8~{\rm cm^2 / g}$ is the Rosseland mean opacity 
1982: at $T \simeq 600$~K, where the opacity takes the maximum value
1983: \citep{Pol94}. 
1984: In the outermost region of the dust cocoon, 
1985: where $T < 600$~K and $\kappa_{\rm d, IR} < 8~{\rm cm^2 / g}$,
1986: the deceleration effect is somewhat milder.
1987: This effect becomes significant when the flow reaches
1988: the inner warm region.
1989: As shown by the light-shaded region in Figure \ref{fig:wc},
1990: the deceleration of dusty flow occurs ($L_{\rm tot} > L_{\rm Edd,d}$) 
1991: in a wide range of $M_{\ast}$ and $\dot{M}_\ast$.
1992: In this case, the upper mass limit by the radiation pressure 
1993: (``$P_{\rm ram}$ limit'' in Fig.~\ref{fig:wc}), which is derived 
1994: by assuming the free-fall flow, becomes lower: the boundary of the 
1995: ``$P_{\rm ram}$ limit'' region shifts to the left to some extent.
1996: To evaluate this effect more quantitatively and then
1997: to identify how mass accretion is finally terminated,
1998: further studies will be needed.
1999: It has been shown that the deceleration
2000: effect of infrared light is diminished by reduced dust opacity (WC87),
2001: non-spherical accretion \citep[e.g.,][]{Nk89, YS02},
2002: and other various processes 
2003: (e.g., see McKee \& Ostriker 2007 for a recent review).
2004: Note that even with the reduced dust opacity, the ``$P_{\rm ram}$
2005: limit'' still remains as long as the dusty envelope is thick enough
2006: for optical/UV (not infrared) light from the star. 
2007: 
2008: As seen in \S~\ref{ssec:edd}, for a very high accretion rate of  
2009: $\dot{M}_\ast > 3 \times 10^{-3}~M_{\odot}/{\rm yr}$, 
2010: the radiation pressure acting on the gas 
2011: envelope terminates the steady mass accretion. 
2012: Figure \ref{fig:wc} shows that 
2013: this limit (``quasi-$L_{\rm Edd}$ limit'') is more severe 
2014: than the radiation pressure limit 
2015: for $\dot{M}_\ast > 3 \times 10^{-3}~M_{\odot}/{\rm yr}$ 
2016: although the deceleration of the dusty flow is already important.
2017: Note that WC87's result on this quasi-Eddington limit 
2018: differs from ours (see their Fig.~5).
2019: This discrepancy comes from their assumption that the forming star is 
2020: already in the ZAMS, which is not valid in that case.
2021: On the other hand, we have consistently calculated the evolution  
2022: until this limit is reached and the star begins abrupt expansion.
2023: For such a high accretion rate, 
2024: what happens when the steady accretion
2025: becomes impossible is not clear in realistic situations.
2026: The accreting envelope may be completely 
2027: blown away and growth of the protostar may be terminated 
2028: at this point.
2029: Without accretion, the protostar quickly relaxes 
2030: to a ZAMS star. 
2031: Another possibility is that since the strong radiation 
2032: force originates in the accretion luminosity itself, 
2033: the accretion may continue in a self-regulated way 
2034: by reducing its rate, for example, 
2035: by an outflow.
2036: This process may proceed in a sporadic way:
2037: the protostar continues to grow alternately repeating 
2038: eruption and accretion.
2039: If this is the case, the protostar evolves along the trajectory (iv)
2040: in Figure \ref{fig:wc}, which runs along the boundary of the 
2041: prohibited region and the maximum stellar mass is 
2042: about $250~M_{\odot}$ set by the radiation pressure feedback 
2043: at the critical accretion rate 
2044: $\dot{M}_{\rm \ast, cr} \simeq 3 \times 10^{-3}~M_{\odot}/{\rm yr}$ 
2045: (Fig.~\ref{fig:wc}).
2046: This is a somewhat higher value than the likely upper mass 
2047: cut-off of the Galactic initial mass function 
2048: $\sim 150~M_{\odot}$ \citep[e.g.,][]{WK04, Fg05}.
2049: This disagreement might be due to the deceleration of the outer 
2050: dusty envelope, which is not included in this analysis.
2051: 
2052: 
2053: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2054: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2055: \section{Observational Signatures of High Accretion Rates}
2056: \label{sec:obs}
2057: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2058: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2059: 
2060: %---------------------------------------------------------------------%
2061: \begin{figure}[t]
2062:   \begin{center}
2063: \epsfig{ file=f20.eps,
2064:          angle=0,
2065:          width=5in}
2066: %hr.eps
2067: \caption{
2068: The stellar birthlines for different accretion rates. 
2069: The cases for accretion rates $10^{-6}~M_{\odot}/{\rm yr}$ 
2070: (solid; run MD6), $10^{-5}$ (dashed; MD5), $10^{-4}$ (dot-dashed; MD4), 
2071: and $10^{-3}~M_{\odot}/{\rm yr}$ (dotted; MD3) are presented.
2072: Each track shows the evolution from the initial model, 
2073: and filled circles on the track represent points of the stellar
2074: mass $M_{\ast}=$ 1, 3, 5, 9, and 20$M_{\odot}$ 
2075: from the lower right in this order. 
2076: The thick broken line represents the positions for 
2077: zero-age main-sequence from \citet{Sch92}.
2078: The open squares along the line denote the positions for the 
2079: same masses as above.
2080: The dotted straight lines indicate the loci for the 
2081: constant stellar radius of $1 R_{\odot}$,
2082: $10 R_{\odot}$, and $100 R_{\odot}$.  }
2083: \label{fig:hr}
2084:   \end{center}
2085: \end{figure}
2086: %--------------------------------------------------------------------%
2087: %----------------------------------------------------------------%
2088: \begin{figure}[t]
2089:   \begin{center}
2090: \epsfig{ file=f21.eps,
2091:          angle=0,
2092:          width=4in}
2093: %orionKL.eps
2094: \caption{
2095: Evolution of the effective temperatures and bolometric luminosities 
2096: of protostars with accretion rate $\dot{M}_\ast=10^{-3}$ (solid; run MD3),
2097: $3 \times 10^{-3}$ (dashed; MD3x3), and 
2098: $6 \times 10^{-3} M_{\odot}/{\rm yr}$ (dot-dashed; MD6x3).
2099: The condition required for the protostar in Orion KL region \citep{Mn98}
2100: is indicated by the horizontal lines with arrows;
2101: $L_{\rm tot} \geq 4 \times 10^4~L_{\odot}$ and $T_{\rm eff} < 5500$K.
2102: The mass range satisfying this condition is denoted by the shade 
2103: for the case of $\dot{M}_\ast=6 \times 10^{-3} M_{\odot}/{\rm yr}$. } 
2104: \label{fig:KL}
2105:   \end{center}
2106: \end{figure}
2107: %-------------------------------------------------------------------------%
2108: 
2109: 
2110: Massive protostars with high accretion rates 
2111: have some characteristic observational features.
2112: In this section, we explore possibilities of identifying them
2113: observationally and of verifying the postulated high accretion rates.
2114: Although protostars still in the main accretion phase are hard to 
2115: observe directly, pre-main sequence (PMS) stars just 
2116: after this phase are optically visible.
2117: The distribution of such PMS stars in the Hertzprung-Russell (HR) 
2118: diagram retains a footprint of the previous accretion phase.
2119: The birthline is the initial loci of the PMS stars
2120: on the HR diagram as a function of the stellar mass. 
2121: The PMS stars move down-left toward the ZAMS line 
2122: on the diagram thereafter.
2123: It has been shown that the birthline for 
2124: $\dot{M}_\ast \sim 10^{-5}~M_{\odot}/{\rm yr}$ successfully traces 
2125: the upper envelope of observed distributions of low-mass and 
2126: intermediate-mass PMS stars in the HR diagram 
2127: \citep{PS92, NM00}.
2128: In Figure \ref{fig:hr}, we show the birthlines for 
2129: various accretion rates. 
2130: These tracks are calculated from the interior luminosity $L_{\ast}$ 
2131: and the stellar radius $R_{\ast}$ at the end of the accretion, i.e., 
2132: when the protostellar mass reaches $M_{\ast}$.
2133: Without the accretion, contribution to the total luminosity 
2134: is now only by the interior one $L_{\ast}$ and the effective 
2135: temperature is given by $T_{\rm *, eff} \equiv 
2136: (L_{\ast}/4 \pi \sigma R_{\ast}^2)^{1/4}$.
2137: With higher $\dot{M}_\ast$, protostars have larger radii and 
2138: their birthline shifts toward the upper-right. 
2139: This dependence on $\dot{M}_\ast$ is consistent with previous
2140: calculations \citep{PS92, NM00}. 
2141: 
2142: If a high accretion rate of $\dot{M}_\ast > 10^{-4}~M_{\odot}/{\rm yr}$ 
2143: is indeed achieved in massive star formation, 
2144: stars as massive as $10~M_{\odot}$ have not yet arrived at the 
2145: ZAMS at the end of the accretion phase.
2146: Thus, the presence of such massive PMS stars is peculiar to the high 
2147: $\dot{M}_\ast$ scenario. 
2148: These PMS stars are very luminous with $L_{\ast} > 10^4 L_{\odot}$, 
2149: and their effective temperature is much lower than that of the ZAMS stars. 
2150: Although some candidates have been reported very close to the ZAMS line,
2151: no such object has yet been firmly detected \citep[e.g.,][]{HHC97}. 
2152: One explanation for the lack of detection
2153: is the very short KH timescale $t_{\rm KH}$ of such PMS stars.
2154: For example, in the case with $\dot{M}_\ast = 10^{-3} M_{\odot}/{\rm yr}$
2155: (run MD3), $t_{\rm KH}$ falls below $10^4$yr for $M_{\ast} > 10 M_{\odot}$ 
2156: (Fig.~\ref{fig:lum_enuc_tsc_1em3}, bottom). 
2157: On the other hand, from the case with 
2158: $\dot{M}_\ast = 10^{-5}~M_{\odot}/{\rm yr}$ (run MD5)
2159: we see that $t_{\rm KH} > 10^6$~yr ($10^5$~yr) for low- 
2160: (intermediate-, respectively) mass protostars 
2161: (Fig.~\ref{fig:lum_enuc_tsc_1em5}, bottom).
2162: In addition to the small number of massive stars, such a short 
2163: duration in the PMS phase severely limits the possibility 
2164: of detection.
2165: 
2166: As referred to in \S~\ref{sec:intro}, some information of 
2167: dust-enshrouded protostars has been gleaned from the light 
2168: re-emitted in the envelope.
2169: On the other hand, \citet{Mn98} reported 
2170: a unique indirect observation of an embedded massive protostar.
2171: They observed the infrared light from the Orion BN/KL nebula,
2172: which is scattered by dust grains far away from the exciting
2173: protostar. Direct light from the protostar is blocked by a 
2174: torus-like structure on the line of sight toward us, but 
2175: likely leaks in the polar direction and is scattered by the nebula. 
2176: The color temperature of the scattered light is evaluated as 
2177: $T_{\rm eff} \simeq 3000-5500$K from features of the absorption lines. 
2178: The bolometric luminosity of the protostar is 
2179: $L_{\rm bol} > 4 \times 10^4 L_{\odot}$. 
2180: If the scattered light is from the photosphere of the star, 
2181: this low $T_{\rm eff}$ cannot be explained by a MS star:
2182: $T_{\rm eff}$ of a ZAMS star with the same luminosity 
2183: is about 35000K, much higher than the above estimation.
2184: \citet{Mn98} concluded that a very large stellar radius
2185: of $> 300 R_{\odot}$ is needed to explain the 
2186: low effective temperature. 
2187: Stimulated by this result, \citet{Nk00} studied the 
2188: evolution of protostars with very high accretion rates 
2189: $\sim 10^{-2} M_{\sun}/{\rm yr}$ using a one-zone model.
2190: However, they concluded that stellar 
2191: radius does not exceed $30~R_{\odot}$ and then
2192: failed to reproduce the observed low effective temperature.
2193: In contrast, using more detailed modeling, 
2194: we have shown that the stellar radius in fact reaches 
2195: as large as $\simeq 100M_{\sun}$ in the cases of high accretion rate .
2196: In Figure \ref{fig:KL}, we present evolution of 
2197: photospheric temperature and bolometric luminosity 
2198: for three cases with $\dot{M}_\ast \geq 10^{-3}~M_{\odot}/{\rm yr}$. 
2199: This figure shows that the required condition is 
2200: satisfied, for example, with the case 
2201: of $\dot{M}_\ast \geq 6 \times 10^{-3}~M_{\odot}/{\rm yr}$ for 
2202: $M_{\ast} \simeq 20 - 25~M_{\odot}$. 
2203: As a conclusion, the observation can be explained  
2204: if the accretion rate is higher than $4 \times 10^{-3}~M_{\odot}/{\rm yr}$.
2205: The allowed mass range corresponds to the duration of the swelling of 
2206: the protostar by the luminosity wave (see \S~\ref{ssec:md_1em3}).
2207: Recall that this swelling is caused by very inhomogeneous entropy 
2208: distribution: the matter near the surface receives a large amount of 
2209: entropy temporarily.
2210: Since this is not included in Nakano et al. (2000)'s one-zone model, 
2211: their model gave a smaller radius than ours. 
2212: In Appendix~\ref{sec:one_zone}, we present calibration of 
2213: parameters for a polytropic one-zone model to include this effect 
2214: approximately.
2215: 
2216: 
2217: 
2218: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2219: \section{Summary and Conclusions}
2220: \label{sec:sum}
2221: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2222: 
2223: We have studied evolution of accreting protostars 
2224: by solving the structure of the central growing stars 
2225: and the surrounding accreting envelopes simultaneously.
2226: Particular attention is paid to cases with high accretion rates of 
2227: $\dot{M}_\ast \geq 10^{-4}~M_{\odot}/{\rm yr}$, which are
2228: envisaged in some current scenarios of massive star formation.
2229: The protostellar evolution at a high mass accretion rate of 
2230: $\dot{M}_\ast = 10^{-3}~M_{\odot}/{\rm yr}$ (run MD3 in Table 1) can 
2231: be summarized as follows:
2232: 
2233: The entire evolution is divided into four characteristic
2234: phases;
2235: (I) adiabatic accretion ($M_{\ast} \lesssim 6~M_{\odot}$), 
2236: (II) swelling ($6~M_{\odot} \lesssim M_{\ast} \lesssim 10~M_{\odot}$),
2237: (III) KH contraction 
2238: ($10~M_{\odot} \lesssim M_{\ast} \lesssim 30~M_{\odot}$), 
2239: and (IV) main-sequence accretion 
2240: ($M_{\ast} \gtrsim 30~M_{\odot}$) phases.
2241: 
2242: A main driver of transition between the evolutionary phases 
2243: above is a decrease in opacity in the stellar interior with increasing 
2244: temperature, and thus protostellar mass.
2245: Due to the fast accretion, the matter accreted onto the stellar surface 
2246: is embedded into the interior before radiatively losing  
2247: the entropy produced at the accretion shock.
2248: Early in the evolution, the opacity in the stellar interior, which 
2249: is mainly by free-free absorption, is very high owing to low temperature.
2250: As a result of the high opacity, the star keeps a large amount of 
2251: entropy imported by the accreted matter.
2252: This is the adiabatic accretion phase (I).
2253: With the decrease in opacity, 
2254: the entropy is transported outward gradually.
2255: When the matter near the surface temporarily 
2256: has a large amount of entropy, 
2257: the star expands as large as $\simeq 200~R_{\odot}$.
2258: This is the swelling phase (II).
2259: After that, all parts of the protostar lose energy and 
2260: the protostar contracts. 
2261: This is the KH contraction phase (III).
2262: With contraction, the temperature increases and the hydrogen 
2263: burning eventually begins.
2264: When the nuclear burning provides sufficient 
2265: energy to balance with that lost from the star radiatively, 
2266: the star stops the KH contraction and reaches the zero-age main-sequence 
2267: (ZAMS) phase at $\simeq 30~M_{\sun}$ (IV). 
2268:  
2269: Deuterium burning plays little role in the evolution.
2270: Even if the deuterium burning is turned off by hand, 
2271: the result hardly changes. 
2272: The swelling (II) is caused by outward transfer of entropy within 
2273: the star, which is observed as a propagating luminosity wave.
2274: 
2275: In general, at the higher accretion rate, the 
2276: protostellar radius is larger and then 
2277: the maximum temperature is lower
2278: at the same stellar mass, 
2279: owing to the higher entropy within the protostar.
2280: As a result of the lower temperature, the onset of the 
2281: nuclear burning is postponed to the higher protostellar mass.
2282: 
2283: For very high accretion rate 
2284: $\dot{M}_\ast > 3 \times 10^{-3}~M_{\odot}/{\rm yr}$, 
2285: in the course of the KH contraction,
2286: the radiation pressure onto the inner accreting envelope 
2287: becomes so strong that the flow is retarded before hitting 
2288: the stellar surface.
2289: The reduced ram pressure onto the stellar surface causes
2290: abrupt expansion of the star.
2291: The steady-state accretion is not possible thereafter.
2292: At this moment, the accretion might be halted, or 
2293: might continue at a reduced rate or in a sporadic way.
2294: In either case, the evolution at the critical accretion
2295: rate of $M_{\rm cr} = 3 \times 10^{-3}~M_{\odot}/{\rm yr}$ gives 
2296: the upper mass limit of PMS stars at $\simeq 60~M_\odot$.
2297: Further growth of the star should be finally limited by the
2298: radiation pressure onto the outer dusty envelope. 
2299: We have found this limit with $M_{\rm cr}$ 
2300: gives the maximum mass of MS stars at $\simeq 250~M_\odot$. 
2301: 
2302: Such a high accretion rate has also been expected and studied for 
2303: the first star formation in the universe.
2304: The evolutions of primordial and solar-metallicity 
2305: protostars are very similar at the same accretion rate.
2306: However, the lower opacity of the primordial gas results in the 
2307: earlier transitions between the evolutionary phases above.
2308: For example, with $\dot{M}_\ast = 10^{-3}~M_{\odot}/{\rm yr}$,
2309: the accreting star enters the KH contraction phase at $10~M_{\sun}$ 
2310: in the solar metallicity case, while it enters 
2311: at $7M_{\sun}$ in the primordial case.
2312: 
2313: Distinguishing features of those massive protostars
2314: accreting at high rates include the large stellar radius 
2315: sometimes exceeding $100~R_{\odot}$, and 
2316: then low color temperature $\simeq 6000$K.
2317: A massive protostar in Orion BN/KL nebula indeed has  
2318: such features and is a possible candidate of such objects. 
2319: 
2320: {\acknowledgements 
2321: The authors thank Francesco Palla, Nanda Kumar, Jonathan Tan, 
2322: Mark Krumholz, and Shu-ichiro Inutsuka for helpful comments and discussions. 
2323: This study is supported in part by Research Fellowships of the Japan
2324: Society for the Promotion of Science for Young Scientists (TH) and 
2325: by the Grants-in-Aid by the Ministry of Education, Science and 
2326: Culture of Japan (18740117, 18026008, 19047004: KO). }
2327: 
2328: 
2329: \begin{thebibliography}{}
2330: 
2331: \bibitem[Alexander \& Ferguson(1994)]{AF94}
2332: Alexander, D.R. \& Ferguson, J. W., 1994, \apj, 437, 879
2333: 
2334: \bibitem[Asplund, Grevesse \& Sauval(2005)]{AGS05}
2335: Asplund, M., Grevesse, N. \& Sauval, A.J., 2005, ASP Conf. Ser., 336, 25
2336: 
2337: \bibitem[Beech \& Mitalas(1994)]{BM94}
2338: Beech, M. \& Mitalas, R., 1994, \apjs, 95, 517
2339: 
2340: \bibitem[Behrend \& Maeder(2001)]{BM01}
2341: Behrend, R. \& Maeder, A., 2001, \aap, 373, 190
2342: 
2343: \bibitem[Beltr\'an et al.(2006)]{Bl06}
2344: Beltr\'an, M.T., Cesaroni, R., Codella, C., Testi, L., Furuya, R.S. \&
2345: 				     Olmi, L. 2006, Nature, 443, 427
2346: 
2347: \bibitem[Beuther et al.(2002)]{B02}
2348: Beuther, H., Schilke, P., Sridharan, T.K., Menten, K.M., Walmsley,
2349: 				C.M. \& Wyrowski, F. 2002, \aap, 383, 892
2350: 
2351: \bibitem[Bernasconi \& Maeder(1996)]{BM96}
2352: Bernasconi, P.A. \& Maeder, A. 1996, \aap, 307, 829
2353: 
2354: \bibitem[Bonnell, Bate \& Zinnecker(1998)]{BBZ98}
2355: Bonnell, I.A., Bate, M.R. \& Zinnecker, H., 1998, \mnras, 298, 93
2356: 
2357: \bibitem[Clayton(1968)]{Cl68}
2358: Clayton D., 1968, {\it Principles of Stellar Evolution and
2359: 				Nucleosynthesis}. 
2360:             McGraw-Hill, New York
2361: 
2362: \bibitem[Cox \& Giuli(1968)]{CG68}
2363: Cox, J.P. \& Giuli, R.T., 1968, {\it Principles of Stellar Structure.}, 
2364: Gordon \& Breach, New York
2365: 
2366: \bibitem[Cunha, Hubeny \& Lanz(2006)]{CHL06}
2367: Cunha, K., Hubeny,I. \& Lanz,T. 2006, \apj, 647, L143
2368: 
2369: \bibitem[Ezer \& Cameron(1971)]{EC71}
2370: Ezer, D. \& Cameron, A.G.W. 1971, Ap\&SS, 14, 399
2371: 
2372: \bibitem[Fazal et al.(2007)]{Fz07}
2373: Fazal, F.M., Sradharan, T.K., Qiu, K., Whitney, B., Robiraille, T. \&
2374: 				Zhang, Q., 2007, astro-ph/0711.2261
2375: 
2376: \bibitem[Figer(2005)]{Fg05}
2377: Figer, D.F., 2005, Nature, 434, 192
2378: 
2379: \bibitem[Fuller, Williams \& Sridharan(2005)]{Fl05}
2380: Fuller,G.A., Williams,S.J. \& Sridharan,T.K., \aap, 442, 949
2381: 
2382: \bibitem[Grevesse \& Noels(1993)]{GN93}
2383: Grevesse, N. \& Noels, A., 1993, Phys. Scr., T47, 133
2384: 
2385: \bibitem[Hanson, Howarth \& Conti(1997)]{HHC97}
2386: Hanson, M.M., Howarth, I.D. \& Conti, P.S., 1997, \apj, 489, 698
2387: 
2388: \bibitem[Hayashi, Hoshi \& Sugimoto(1962)]{HHS62}
2389: Hayashi, C., Hoshi, R. \& Sugimoto, D., 1962, Prog. Theor. Phys. Suppl., 22, 1
2390: 
2391: \bibitem[Hopkins \& Beacom(2006)]{HB06}
2392: Hopkins, A.M. \& Beacom, J.F., 2006, \apj, 651, 142
2393: 
2394: \bibitem[Igresias \& Rogers(1996)]{IR96}
2395: Igresias, C.A. \& Rogers, F.J., 1996, \apj, 464, 943
2396: 
2397: 
2398: \bibitem[Kahn(1974)]{KH74}
2399: Kahn, F.D., 1974, \aap, 37, 149
2400: 
2401: \bibitem[Keto \& Wood(2006)]{KW06}
2402: Keto, E. \& Wood, K., 2006, \apj, 637, 850
2403: 
2404: \bibitem[Krumholz \& Thompson(2007)]{KT07}
2405: Krumholz, M.R. \& Thompson, T.A., 2007, \apj, 661, 1034
2406: 
2407: \bibitem[Krumholz, Klein \& McKee(2007)]{KKM07}
2408: Krumholz, M.R., Klein, R.I. \& McKee, C.F., 2007, \apj, 656, 959
2409: 
2410: \bibitem[Kumar \& Grave(2008)]{KG08}
2411: Kumar, M.S.N. \& Grave, J.M.C., 2008, astro-ph/0802.2073
2412: 
2413: \bibitem[Larson \& Starrfield(1971)]{LS71}
2414: Larson, R.B. \& Starrfield, S. 1971, \aap, 13, 190
2415: 
2416: \bibitem[Madau et al.(1996)]{Md96}
2417: Madau,P., Ferguson,H.C., Dickinson,M.E., Giavalisco,M., Steidel,C.C,
2418: 				     Fruchter,A. \mnras, 283, 1388
2419: 
2420: \bibitem[Matsuda, Sato \& Takeda(1969)]{MST67}
2421: Matsuda, T., Sato, H., \& Takeda, H. 1969, Prog. Theor. Phys., 42, 219
2422: 
2423: \bibitem[McKee \& Tan(2002)]{MT02}
2424: McKee, C.F. \& Tan, J.C., 2002, Nature, 416, 59
2425: 
2426: \bibitem[McKee \& Tan(2003)]{MT03}
2427: McKee, C.F. \& Tan, J.C., 2003, \apj, 585. 850
2428: 
2429: \bibitem[McKee \& Ostriker(2007)]{MO07}
2430: McKee, C.F. \& Ostriker, E.C., 2007, \araa, 45, 565
2431: 
2432: \bibitem[Morino et al.(1998)]{Mn98}
2433: Morino, J., Yamashita, T., Hasegawa, T. \& Nakano, T., 1998, Nature,
2434: 				393, 340
2435: 
2436: \bibitem[Motte et al.(2007)]{Mt07}
2437: Motte,F., Bontemps,S., Schilke,P., Schneider,N., Menten,K.M. \&
2438: 				     Broguiere,D., 2007, \aap, 476, 1243
2439: 
2440: \bibitem[Nakano(1989)]{Nk89}
2441: Nakano, T. 1989, \apj, 345, 464
2442: 
2443: \bibitem[Nakano, Hasegawa \& Norman(1995)]{NHN95}
2444: Nakano, T., Hasegawa, T., \& Norman, C., 1995, \apj, 450, 183
2445: 
2446: \bibitem[Nakano et al.(2000)]{Nk00}
2447: Nakano, T., Hasegawa, T., Morino, J. \& Yamashita, T., 2000, \apj, 534, 976
2448: 
2449: \bibitem[Norberg \& Maeder(2000)]{NM00}
2450: Norberg, P. \& Maeder, A., 2000, \aap, 359, 1025
2451: 
2452: \bibitem[Omukai \& Nishi(1998)]{ON98}
2453: Omukai, K. \& Nishi, R., 1998, \apj, 508, 1410
2454: 
2455: \bibitem[Omukai \& Palla(2001)]{OP01}
2456: Omukai, K. \& Palla, F., 2001, \apj, 561, 550L
2457: 
2458: \bibitem[Omukai \& Palla(2003)]{OP03}
2459: Omukai, K. \& Palla, F., 2003, \apj, 589, 677
2460: 
2461: \bibitem[Osorio, Lizano \& D'Alessio(1999)]{OLD99}
2462: Osorio, M., Lizano, S. \& D'Alessio, P., 1999, \apj, 525, 808
2463: 
2464: \bibitem[Palla, Salpeter \& Stahler(1983)]{PSS83}
2465: Palla, F., Salpeer, E.E. \& Stahler, S.W., 1983, \apj, 271, 632
2466: 
2467: \bibitem[Palla \& Stahler(1990)]{PS90}
2468: Palla, F. \& Stahler, S.W., 1990, \apj, 360, 47L
2469: 
2470: \bibitem[Palla \& Stahler(1991)]{PS91}
2471: Palla, F. \& Stahler, S.W., 1991, \apj, 375, 288 (PS91)
2472: 
2473: \bibitem[Palla \& Stahler(1992)]{PS92}
2474: Palla, F. \& Stahler, S.W., 1992, \apj, 392, 667
2475: 
2476: \bibitem[Palla \& Stahler(1993)]{PS93}
2477: Palla, F. \& Stahler, S. W., 1993, \apj, 418, 414
2478: 
2479: \bibitem[Pollack et al.(1994)]{Pol94}
2480: Pollack, J.B., Hollenbach, D., Beckwith, S., Simonelli, D.P., Roush, T.,
2481: 				     Fong, W. 1994, 421, 615
2482: 
2483: \bibitem[Rathborne, Simon \& Jackson(2007)]{RSJ07}
2484: Rathborne, J. M., Simon, R. \& Jackson, J. M. 2007, \apj, 662, 1082
2485: 
2486: \bibitem[Saslaw \& Zipoy(1967)]{SZ67}
2487: Saslaw, W. C., \& Zipoy, D. 1967, \nat, 216, 976 
2488: 
2489: \bibitem[Schaller et al.(1992)]{Sch92}
2490: Schaller, G., Schaerer, D., Meynet, G. \& Maeder, A., 1992, A\&AS, 96, 269
2491: 
2492: \bibitem[Shu, Adams \& Lizano(1987)]{SAL87}
2493: Shu, F.H., Adams, F.C. \& Lizano,S., 1987, \araa, 25 23
2494: 
2495: \bibitem[Stahler(1983)]{St83}
2496: Stahler, S.W., 1983, \apj, 274, 822
2497: 
2498: \bibitem[Stahler(1988)]{St88}
2499: Stahler, S.W., 1988, \apj, 332, 804
2500: 
2501: \bibitem[Stahler, Palla \& Salpeter(1986)]{SPS86}
2502: Stahler, S.W., Palla, F. \& Salpeter, E.E., 1986, \apj, 302, 590 (SPS86)
2503: 
2504: \bibitem[Stahler, Shu \& Taam(1980a)]{SST80a}
2505: Stahler, S.W., Shu, F.H. \& Taam, R.E., 1980, \apj, 241, 637 (SST80a)
2506: 
2507: \bibitem[Stahler, Shu \& Taam(1980b)]{SST80b}
2508: Stahler, S.W., Shu, F.H. \& Taam, R.E., 1980, \apj, 242, 226 (SST80b)
2509: 
2510: \bibitem[Stahler, Palla \& Ho(2000)]{SPH00}
2511: Stahler, S.W., Palla, F. \& Ho, P.T.P., 2000, Protostars and Planets IV, 327 
2512: 
2513: \bibitem[Tan \& McKee(2004)]{TM04}
2514: Tan, J.C. \& McKee, C.F., 2004, \apj, 603, 383
2515: 
2516: \bibitem[Weidner \& Kroupa(2004)]{WK04}
2517: Weidner, C. \& Kroupa, P., 2004, \mnras, 348, 187
2518: 
2519: \bibitem[Wolfire \& Cassinelli(1987)]{WC87}
2520: Wolfire, M.G. \& Cassinelli, J.P., 1987, \apj, 319, 850 (WC87)
2521: 
2522: \bibitem[Yorke \& Sonnhalter(2002)]{YS02}
2523: Yorke, H.W. \& Sonnhalter, C., 2002, \apj, 569, 864
2524: 
2525: \bibitem[Yorke \& Bodenheimer(2008)]{YB08}
2526: Yorke, H.W. \& Bodenheimer,P., 2008, in Astronomical Society of the
2527: 				     Pacific Conference Series,
2528: 				     Vol. 387, ed. H.Beuther, H.Linz \&
2529: 				     T.Henning, p189
2530: 
2531: \bibitem[Yoshida et al.(2006)]{Ys06}
2532: Yoshida, N., Omukai, K., Hernquist, L. \& Abel, T., 2006, \apj, 652, 6
2533: 
2534: \bibitem[Zhang et al.(2005)]{Zh05}
2535: Zhang, Q., Hunter, T.R., Brand, J., Sridharan, T.K., Cesaroni, R.,
2536: Molinari, S., Wang, J. \& Kramer, M., 2005, \apj, 625, 864
2537: 
2538: \bibitem[Zinnecker \& Yorke(2007)]{ZY07}
2539: Zinnecker, H. \& Yorke, H.W., 2007, \araa, 45, 481
2540: 
2541: 
2542: \end{thebibliography}
2543: 
2544: 
2545: \clearpage
2546: 
2547: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2548: %%%%%%%%%%%%%%%%%%%%%%%%%% Appendix %%%%%%%%%%%%%%%%%%%%%%%%%%%
2549: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2550: 
2551: \appendix
2552: 
2553: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2554: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2555: \section{Method of Calculations}
2556: \label{sec:method}
2557: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2558: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2559: 
2560: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2561: \subsection{Evolutionary Calculations}
2562: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2563: 
2564: %------------------------%
2565: \subsubsection{Protostar}
2566: \label{sssec:prost}
2567: %------------------------%
2568: 
2569: For protostellar structure, the following four stellar 
2570: structure equations are solved:
2571: \begin{equation}
2572: \left( \frac{\partial r}{\partial M} \right)_t = \frac{1}{4 \pi \rho r^2},
2573: \label{eq:con} 
2574: \end{equation}
2575: \begin{equation}
2576: \left( \frac{\partial P}{\partial M}  \right)_t = - \frac{GM}{4 \pi r^4}, 
2577: \label{eq:mom}
2578: \end{equation}
2579: \begin{equation}
2580: \left( \frac{\partial L}{\partial M} \right)_t 
2581: = \epsilon - T \left( \frac{\partial s}{\partial t}  \right)_M ,
2582: \label{eq:ene}
2583: \end{equation}
2584: \begin{equation}
2585: \left( \frac{\partial s}{\partial M} \right)_t
2586: = \frac{G M}{4 \pi r^4} \left( \frac{\partial s}{\partial p} \right)_T
2587:   \left( \frac{L}{L_s} - 1  \right) C ,
2588: \label{eq:heat}
2589: \end{equation}
2590: with a spatial variable of Lagrangian mass coordinate $M$.
2591: In the above, $\epsilon$ is the energy production rate 
2592: by nuclear fusion, $s$ is the entropy per unit mass,  
2593: $L_s$ is the radiative luminosity with adiabatic temperature
2594: gradient, and other quantities have ordinary meanings.
2595: The coefficient, $C$ in equation (\ref{eq:heat}) is unity if 
2596: $L < L_s$ (i.e., in radiative layers), and given by the 
2597: mixing-length theory if $L > L_s$ (i.e., in convective layers). 
2598: The momentum equation (\ref{eq:mom}) simply states the
2599: hydrostatic equilibrium.
2600: This is valid because the stellar mechanical timescale is generally 
2601: much shorter than the accretion timescale.
2602: In energy equation (\ref{eq:ene}), on the other hand, the 
2603: non-equilibrium term $-T (\partial s/\partial t)_M$ should be included.
2604: This is because the KH timescale can be longer than or
2605: comparable to the accretion timescale in our calculations.
2606: Thus, the thermal equilibrium is not generally satisfied in 
2607: accreting protostars.
2608: Nuclear reactions included in $\epsilon$ are 
2609: deuterium burning and hydrogen burning via the pp-chain and CN-cycle.
2610: The elements are homogeneously distributed
2611: in convective layers by the convective mixing.
2612: 
2613: There are two alternative schemes to integrate equations 
2614: (\ref{eq:ene}) and (\ref{eq:heat}), which are the explicit 
2615: and implicit schemes:
2616: 
2617: 
2618: \paragraph{Explicit Scheme}
2619: In the explicit scheme, equation (\ref{eq:ene})
2620: is considered as a time development equation of the entropy.
2621: In order to update the entropy at each spatial grid
2622: from a time step $t - \Delta t$ to $t$, 
2623: we use $(\partial L/\partial M)_t$
2624: evaluated at the previous time step $t - \Delta t$ and 
2625: calculate $(\partial s/\partial t)_M$ by equation (\ref{eq:ene}).
2626: Equation (\ref{eq:heat}) is transformed as,
2627: \begin{equation}
2628: L = L_s \left[
2629:          1 + \frac{4 \pi r^4}{G M} 
2630:              \left( \frac{\partial P}{\partial s}  \right)_T
2631:              \left( \frac{\partial s}{\partial M} \right)_t
2632:         \right] ,
2633: \label{eq:lpfile}
2634: \end{equation}
2635: with $C=1$ (radiative).
2636: We substitute the updated entropy gradient $(\partial s/\partial M)_t$
2637: to equation (\ref{eq:lpfile}) and calculate the luminosity profile 
2638: $L(M,t)$.
2639: This scheme enables the precise integration where the heat transport among
2640: mass cells is negligible, i.e., $(\partial L/\partial M)_t$
2641: is much smaller than other terms in equation (\ref{eq:ene}).
2642: For example, this quasi-adiabatic situation is created in the 
2643: very opaque interior of the protostar.
2644: 
2645: \paragraph{Implicit Scheme}
2646: In the implicit scheme, on the other hand, time-derivative 
2647: term $- T (\partial s/\partial t)_M$ in equation (\ref{eq:ene}) 
2648: is regarded as a source term to calculate the luminosity profile
2649: by spatial integration. We simultaneously integrate equations 
2650: (\ref{eq:ene}) and (\ref{eq:heat}),
2651: evaluating $(\partial s/\partial t)_M$ at every increment of spatial grids.
2652: This scheme works effectively where the heat transport
2653: among mass cells works to redistribute entropy.
2654: For example, this is the case near the stellar surface. 
2655: Devote special attention to discretizing the time-derivative 
2656: $( \partial s / \partial t)_M$ in equation (\ref{eq:ene}).
2657: For example, consider the gas element $\dot{M}_\ast \Delta t$ 
2658: measured from the surface in a stellar model.
2659: This element accretes to the star from a time step 
2660: $t - \Delta t$ to $t$, and is still in the accreting envelope 
2661: at the previous time step $t - \Delta t$.
2662: The accreting matter plunges to the star through the accretion shock
2663: front, which is treated as a mathematical discontinuity in
2664: our formulation. Describing the derivative across the discontinuity
2665: is a problem. 
2666: We avoid this difficulty adopting a coordinate conversion (SST80b),
2667: \begin{equation}
2668:  \left( \frac{\partial s}{\partial t}  \right)_M = 
2669:   \left( \frac{\partial s}{\partial t} \right)_m 
2670: + \dot{M}_\ast  \left( \frac{\partial s}{\partial m}  \right)_t ,
2671: \label{eq:relm}
2672: \end{equation}
2673: where $m$ is the inverse mass coordinate $m \equiv M_{\ast} - M$.
2674: The term $(\partial s/\partial t)_m$ can be obtained only with
2675: the entropy distribution within the shock front.
2676: This treatment is fairly valid, because the structure
2677: of the settling layer hardly changes over
2678: a few time steps, observing from the shock front.
2679: 
2680: 
2681: In our calculations, we properly adopt different schemes
2682: in different evolutionary stages.
2683: In the early phase of the evolution, we use both the explicit
2684: and implicit schemes to construct one stellar model (e.g., SST80b); 
2685: the explicit scheme is used in the deep interior
2686: and is switched to the implicit scheme near the stellar surface.
2687: First, we solve the interior with the explicit scheme;
2688: we integrate equations (\ref{eq:con}) and (\ref{eq:mom}) outward 
2689: with a guess of central pressure $P_c$. 
2690: When the explicitly solved region is radiative, 
2691: we calculate $L(M,t)$ with equation (\ref{eq:ene}). 
2692: Once active deuterium burning begins, however, 
2693: the explicit update of entropy sometimes makes negative 
2694: entropy gradients $(\partial s/\partial M)_t < 0$.
2695: Such a layer should be regarded as a convective layer,
2696: and equation (\ref{eq:lpfile}) is not valid any more.
2697: We temporarily change the integration method in such a layer.
2698: We correct the entropy distribution as $(\partial s/\partial M)_t = 0$
2699: there after the explicit update, following the 
2700: ``equal-area rule'' developed by SST80b.
2701: Once the entropy profile is corrected, we calculate 
2702: $(\partial s/\partial t)_M$ and obtain the luminosity distribution 
2703: by integrating equation (\ref{eq:ene}).
2704: As the integration approaches the surface, the term 
2705: $(\partial L/\partial M)_t$ in equation (\ref{eq:ene}) gradually grows. 
2706: We switch to the implicit method where $(\partial L/\partial M)_t$ 
2707: attains $0.5 \dot{M}_\ast T (\partial s/\partial m)_t$.
2708: We integrate all equations (\ref{eq:con}) -- (\ref{eq:heat}) with 
2709: a guess of the entropy at the switching point $s_{\rm sw}$.
2710: When the integration reaches the stellar surface, we check
2711: the differences from the required boundary conditions
2712: (see \S~\ref{sssec:accflow} below).
2713: We improve the guessed values of $P_c$ and $s_{\rm sw}$, and 
2714: iterate this procedure until the model satisfies the outer
2715: boundary conditions.
2716: 
2717: The above method successfully works in the early phase of evolution.
2718: However, only in the later stage of evolution, where the entropy 
2719: redistribution among mass cells works even in the deep stellar interior,
2720: we need to use another method with the implicit scheme.
2721: This occurs when a widespread convective layer appears during 
2722: active deuterium burning, or when radiative heat transport becomes 
2723: efficient due to the opacity decrease in the deep interior
2724: (see \S~\ref{sec:result} for detail).
2725: In the fully implicit method, we integrate the equations
2726: (\ref{eq:con}) -- (\ref{eq:heat}) outward from the center with 
2727: a guess of the central pressure $P_c$ and entropy $s_c$. 
2728: If the integration reaches the stellar surface, we advance to 
2729: the same fitting process at the surface as in the above method.
2730: If the integration diverges and fails before reaching the
2731: surface, we adopt another shooting method to a halfway fitting point.
2732: We additionally guess the radius $R_{\ast}$ and luminosity $L_{\ast}$ 
2733: at the stellar surface, and integrate
2734: backward to the fitting point. 
2735: We adjust boundary values to match the physical quantities at 
2736: the fitting point, which are obtained by the forward and backward 
2737: integrations.
2738: 
2739: %- shooting splitting 
2740: %- maximum e_D in radiative layer
2741: 
2742: 
2743: 
2744: %----------------------------------------------------------------%
2745: \subsubsection{Accreting Gas Envelope and Accretion Shock Front}
2746: \label{sssec:accflow}
2747: %----------------------------------------------------------------%
2748: 
2749: The outer boundary condition of the protostar is given 
2750: by constructing the structure of an accreting gas envelope.
2751: The boundary layer connecting the accreting envelope and
2752: protostar actually has detailed structure.
2753: The acreted gas initially hits a standing shock front and
2754: is heated up by compression there. 
2755: The shocked gas next enters the relaxation layer behind 
2756: the shock front and is cooled down by radiative loss.
2757: In our formulation, this detailed structure is not solved,
2758: but treated as a mathematical discontinuity (SST80).
2759: We define the following two points across the discontinuity:
2760: a post-shock point, where gas and radiation temperature
2761: first become equal behind the relaxation layer, and a post-shock
2762: point ahead of the shock front.
2763: In our calculations, the former is the outer boundary 
2764: of the protostar, and the latter is the inner boundary of the
2765: accreting envelope.
2766: Physical states at each point are related by a radiative 
2767: shock jump condition. 
2768: 
2769: Here, we focus on cases where the post-shock quantities
2770: are given in advance by the trial outward 
2771: integration of the protostar (see \S~\ref{sssec:prost}). 
2772: When the outward integration fails and the shooting to a halfway
2773: fitting point is needed, we only slightly modify the same
2774: procedure (e.g., see PS91). 
2775: First, we judge if the accreting envelope is optically thin or thick.
2776: Temporarily assuming that the the accretion flow is optically 
2777: thin and in the free-fall, velocity and mass density at 
2778: the pre-shock point are,
2779: \begin{equation}
2780: u_{\rm pre} = - \sqrt{\frac{2 G M_{\ast}}{R_{\ast}} } ,
2781: \label{eq:u_pre}
2782: \end{equation}
2783: \begin{equation}
2784: \rho_{\rm pre} = \frac{\dot{M}_\ast}{4 \pi R_{\ast}^2 u_{\rm pre}} .
2785: \label{eq:rho_pre}
2786: \end{equation}
2787: Ahead of the shock front, outward radiation comes from both
2788: the stellar interior and relaxation layer behind the shock front.
2789: Therefore, radiation temperature at the pre-shock point is written
2790: as,
2791: \begin{equation}
2792: T_{\rm rad, pre} = 
2793: \left( 
2794: \frac{L_{\ast} + L_{\rm acc}}{4 \pi R_{\ast}^2 \sigma} 
2795: \right)^{1/4} , 
2796: \label{eq:t_pre}
2797: \end{equation}
2798: where $\sigma$ is the Stefan-Boltzmann constant, and $L_{\rm acc}$ is the
2799: accretion luminosity defined by equation (\ref{eq:lacc})
2800:   Using equations (\ref{eq:u_pre}) -- (\ref{eq:t_pre}), optical thickness
2801: of the accreting envelope is estimated as, 
2802: \begin{equation}
2803: \tau_{\rm pre} = R_{\ast} \rho_{\rm pre} \kappa(\rho_{\rm pre}, T_{\rm pre}),
2804: \label{eq:tau_pre}
2805: \end{equation}
2806: where $\kappa$ is the Rosseland mean opacity.
2807: If $\tau_{\rm pre} < 1$, the accreting envelope is optically thin
2808: and initial adoption of the thin envelope is verified.
2809: The jump condition at the accretion shock front is provided by
2810: considering the flow structure across the shock front (SST80b). 
2811: In the optically-thin case, the post-shock temperature is related
2812: to the stellar radius and luminosity (PS91),
2813: \begin{equation}
2814: T_{\rm post} = \left[ \frac{1}{4 \pi R_{\ast}^2 \sigma}   
2815:                 \left( \frac12 L_{\ast} + \frac34 L_{\rm acc} \right)
2816:                \right]^{1/4} . 
2817: \label{eq:tpt_thin}
2818: \end{equation}
2819: This is the outer boundary condition of the protostar.
2820: The protostellar models are constructed for
2821: the post-shock quantities to satisfy equation (\ref{eq:tpt_thin}).
2822: 
2823: If $\tau_{\rm pre} > 1$, on the other hand, the accreting envelope
2824: should actually be optically thick. In this case, we solve 
2825: structure of the flow inside the photosphere, i.e., radiative precursor.
2826: First, we determine the locus of the photosphere. 
2827: Assuming the free-fall flow outside the photosphere, 
2828: optical depth to the photosphere $\tau_{\rm ph}$ 
2829: can be written following equations 
2830: (\ref{eq:u_pre}) -- (\ref{eq:tau_pre}) by substituting physical
2831: quantities at the photosphere. The only unknown quantity is the 
2832: luminosity at the photosphere $L_{\rm ph}$,
2833: which is needed to calculate temperature in equations (\ref{eq:t_pre}).
2834: We calculate this photospheric luminosity using the conservation
2835: law of the net energy outflow rate $L_e$ through the precursor,
2836: \begin{equation}
2837: L_e = L_{\rm ph} - \dot{M}_\ast
2838:       \left( h_{\rm ph} + \frac12 u_{\rm ph}^2 - 
2839:       \frac{G M_{\ast}}{R_{\rm ph}}  \right) ,
2840: \label{eq:le1}
2841: \end{equation}
2842: where $h_{\rm ph} = h(\rho_{\rm ph}, T_{\rm ph})$ is the enthalpy 
2843: of both gas and radiation at the photosphere.
2844: Using the known post-shock quantities, we can give the numerical value of
2845:  $L_e$ at the bottom of the accretion flow, 
2846: \begin{equation}
2847: L_e = L_{\ast} - \dot{M}_\ast
2848:       \left( 
2849:       h_{\rm post} + \frac12 u_{\rm pre}^2 - \frac{G M_{\ast}}{R_{\ast}}  
2850:       \right) .
2851: \label{eq:le2}
2852: \end{equation}
2853: Assigning this value to $L_e$ in equation (\ref{eq:le1}),
2854: the photospheric luminosity is written with
2855: the infall velocity and thermal state at $r = R_{\rm ph}$. 
2856: Consequently, $T_{\rm ph}$ and $\tau_{\rm ph}$ are given as 
2857: implicit functions of $R_{\rm ph}$.
2858: Therefore, we can fix $R_{\rm ph}$ in an iterative fashion 
2859: by requiring $\tau_{\rm ph} = 1$.
2860: Once the photospheric radius is fixed, we solve the structure of the
2861: precursor by integrating the following equations inward from 
2862: the photosphere, 
2863: \begin{equation}
2864: \rho = - \frac{\dot{M}_\ast}{4 \pi r^2 u} ,
2865: \label{eq:cont_prec}
2866: \end{equation}
2867: \begin{equation}
2868: u \frac{\partial u}{\partial r} = - \frac{G M_{\ast}}{r^2} 
2869:                         - \frac{1}{\rho} \frac{\partial P}{\partial r}
2870:                         + \frac{\kappa}{c} F ,
2871: \end{equation}
2872: \begin{equation}
2873: \frac{\partial T}{\partial r} = - \frac{3 \rho \kappa}{16 \sigma T^3} F , 
2874: \end{equation}
2875: where $F$ is the radiative flux.
2876: Using equations (\ref{eq:le1}) and (\ref{eq:cont_prec}), $F$
2877: is written as a function of $\rho$, $u$, and $T$,
2878: \begin{equation}
2879: F = - \rho u \left( 
2880:              \frac{L_e}{\dot{M}_\ast} + h + \frac12 u^2 - \frac{G M_{\ast}}{r}
2881:              \right) .
2882: \end{equation}
2883: When the integration reaches the pre-shock point, a jump condition 
2884: bridges the the pre-shock and post-shock points. 
2885: An isothermal jump condition is applied for the optically-thick 
2886: radiative shock front, 
2887: \begin{equation}
2888: T_{\rm post} = T_{\rm pre}.
2889: \label{eq:tpt_thick}
2890: \end{equation}
2891: This is another outer boundary condition in addition to 
2892: equation (\ref{eq:tpt_thin}).
2893: 
2894: 
2895: 
2896: 
2897: %%%%%%%%%%%%%%%%%%%%%%%%%%%%
2898: \subsection{Initial Models}
2899: \label{ssec:initial}
2900: %%%%%%%%%%%%%%%%%%%%%%%%%%%%
2901: 
2902: We construct initial models by solving their structure.
2903: The method of calculation is based on that of evolutionary
2904: calculations, but with slight modifications.
2905: First, we assume the entropy distribution in the stellar interior 
2906: as an increasing function with $M$, 
2907: \begin{equation}
2908: s(M) = s_{c,0} 
2909: + \beta \frac{k_{\rm B}}{m_{\rm H}} \frac{M}{M_{*,0}} ,
2910: \label{eq:s0}
2911: \end{equation}
2912: where $s_{c,0}$ is specific entropy at the stellar center, 
2913: $k_{\rm B}$ is Boltzmann constant, $m_{\rm H}$ is atomic mass 
2914: unit, and $\beta~(>0)$ is a free parameter.
2915: In the core interior, we integrate equations (\ref{eq:con}) 
2916: and (\ref{eq:mom}) outward beginning with guessed $s_{c,0}$ 
2917: and central pressure, $p_c$. 
2918: According to equation (\ref{eq:lpfile}), the luminosity profile there 
2919: is calculated using an entropy gradient obtained from (\ref{eq:s0}).
2920: Whereas this procedure is valid in the adiabatic interior, 
2921: entropy and luminosity profiles should be consistently calculated
2922: in place of equation (\ref{eq:s0}) near the surface
2923: (also see \S~\ref{sssec:prost}).
2924: The switching point is defined as follows.
2925: Using equations (\ref{eq:ene}) and (\ref{eq:relm}), the energy
2926: equation can be written as,
2927: \begin{equation}
2928: \left( \frac{\partial L}{\partial M} \right)_t = 
2929:  \dot{M}_\ast T \left( \frac{\partial s}{\partial M}  \right)_t ,
2930: \label{eq:ene_ini}
2931: \end{equation}
2932: omitting the nuclear energy production and time-derivative terms.
2933: Once the luminosity gradient $(\partial L/\partial M)_t$ becomes 
2934: comparable to $\dot{M}_\ast T (\partial s/\partial M)_t$ calculated with 
2935: equation (\ref{eq:s0}), we solve the full equations including
2936: equation (\ref{eq:ene_ini}).
2937: The guessed $s_{c,0}$ and $p_c$ are adjusted so that the core satisfies
2938: outer boundary conditions, as explained in \S~\ref{sssec:accflow}.
2939: We have confirmed that some variations of initial models only affect
2940: subsequent protostellar evolution in a very early phase.
2941: In this paper, we adopt $\beta = 1$ as the entropy slope in 
2942: equation (\ref{eq:s0}).
2943: 
2944: 
2945: %%%%%%%%%%%%%%%%%%%%%%%%%%%%
2946: \subsection{Opacity Tables}
2947: \label{ssec:op}
2948: %%%%%%%%%%%%%%%%%%%%%%%%%%%%
2949: 
2950: For the opacity, we adopt OPAL tables \citep[e.g.,][]{IR96} 
2951: with the composition given by \citet{GN93}
2952: for temperature $T > 7000$~K. 
2953: For $T < 7000$~K, we use other opacity tables based on calculations
2954: by \citet{AF94}. 
2955: Contribution from grains is excluded in 
2956: the adopted composition \citep{AGS05, CHL06}.
2957: In all runs, we adopt $X = 0.7$, $Y = 1 - X- Z$, and
2958: relative abundances of heavy elements following \citet{GN93}
2959: composition.
2960: 
2961: 
2962: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2963: \section{Comparison with Previous Work \\
2964:          -- Evolution with the Low Accretion Rate
2965:          of $10^{-5}~M_{\odot}/{\rm yr}$ --}
2966: \label{ap:prev}
2967: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2968: 
2969: In \S~\ref{ssec:md_1em5}, we have presented the calculated 
2970: protostellar evolution at the low accretion rate of
2971: $10^{-5}~M_{\odot}/{\rm yr}$ (run MD5). Whereas our numerical results
2972: reproduce basic features obtained by previous work 
2973: (e.g., SST80a, PS91), there are still some differences.
2974: In this appendix, we examine causes of these differences by 
2975: exploring some cases with different deuterium
2976: abundances and initial models.
2977: 
2978: 
2979: %---------------------------------------------%
2980: \subsection{Dependence on Deuterium Abundance}
2981: \label{ssec:dab}
2982: %---------------------------------------------%
2983: The effect of different deuterium abundances 
2984: on the stellar interior structure in early phases
2985: is shown in Figure \ref{fig:str_ddep} for cases with 
2986: ${\rm [D/H]}=3\times 10^{-5}, 2.5\times 10^{-5}$, and $1\times 10^{-5}$.
2987: The middle panel is the same as Figure \ref{fig:str_fdtmax_1em5} (upper), 
2988: but reproduced for comparison.
2989: As described in \S~\ref{ssec:md_1em5}, our run MD5
2990: exhibits somewhat complex evolution of convective layers in the early
2991: phase. 
2992: A convective layer first appears at $M_{\ast} \simeq 0.3~M_{\odot}$,
2993: just after the deuterium burning begins. 
2994: This convective layer disappears at
2995: $M_{\ast} \simeq 0.6~M_{\odot}$. The second convective layer emerges
2996: just outside the former outer boundary of the first one 
2997: at $M_{\ast} \simeq 0.7~M_{\odot}$, 
2998: and quickly reaches the surface. 
2999: SST80a calculated the same model, but did not observe 
3000: such a complex evolution. In their calculation, 
3001: the first convective layer survives and eventually incorporates 
3002: most of the stellar interior.
3003: 
3004: This difference comes from the fact that the evolution of 
3005: convective layers sensitively depends on the initial 
3006: deuterium abundance. 
3007: For example, in the case with the deuterium abundance 
3008: [D/H] = $3 \times 10^{-5}$ 
3009: (Figure \ref{fig:str_ddep} upper; run MD5-dh3), slightly higher than 
3010: the fiducial value, 
3011: the convective layer first appears at the same epoch as the fiducial case 
3012: with [D/H] = $2.5 \times 10^{-5}$, and extends outward.
3013: In this case, another convective layer soon emerges at the
3014: surface and begins to extend inward.
3015: These two convective layers finally merge at $M_{\ast} \simeq 0.6~M_{\odot}$,
3016: and render most of the interior convective.
3017: At $M_{\ast} \simeq 1~M_{\odot}$, the convective region contains about
3018: 97\% of the total mass. These features are very similar to
3019: those found by SST80a, who terminated their calculation 
3020: at $M_{\ast} = 1~M_{\odot}$.
3021: Our calculation shows that the subsequent evolution in this case 
3022: is almost the same as in our fiducial run MD5. 
3023: The central radiative core gradually
3024: expands as the stellar mass increases, and finally occupies 
3025: most of the interior at $M_{\ast} \simeq 4~M_{\odot}$. 
3026: 
3027: This sensitive dependence on the deuterium abundance is explained
3028: as follows. 
3029: The expansion of convective layers is caused by 
3030: increase of the specific entropy within it, which 
3031: is generated by the nuclear burning at the bottom of the layer.
3032: For continuing expansion of the convective layer, 
3033: deuterium needs to be supplied continuously.
3034: Fresh deuterium becomes available when the convective layer newly 
3035: incorporates the radiative region.
3036: This deuterium immediately spreads over the
3037: convective layer, and is consumed for further nuclear fusion.
3038: For sufficiently high deuterium abundance,
3039: acquired deuterium is high enough to maintain the active nuclear burning. 
3040: In this case, the convective layer continuously grows 
3041: as in the case with ${\rm [D/H]}= 3\times 10^{-5}$ (run MD5-dh3). 
3042: On the other hand, for low deuterium abundances,
3043: the nuclear burning becomes too weak to maintain development 
3044: of the convective layer.
3045: The convective layer ceases to grow and disappears as in 
3046: the case with ${\rm [D/H]}= 2.5 \times 10^{-5}$ (run MD5).
3047: For even smaller deuterium abundance [D/H] = $1 \times 10^{-5}$ 
3048: (Fig.~\ref{fig:str_ddep}, lower; run MD5-dh1), 
3049: no convective layer appears even after ignition 
3050: of the deuterium burning until at $M_{\ast} \sim 0.8~M_{\odot}$, 
3051: when the surface convective layer finally emerges.
3052: In this case, the subsequent mass-radius relation is 
3053: almost same as in the case without deuterium burning (run MD5-noD)
3054: although the surface convective layer pesists 
3055: until $M_{\ast} \sim 3~M_{\odot}$. 
3056: Our fiducial run MD5 is just intermediate between 
3057: the deuterium-rich (run MD5-dh3) and -poor (MD5-dh1) cases 
3058: presented in Figure \ref{fig:str_ddep}.
3059: Therefore, a slight difference in input physics affects
3060: the internal structure significantly.
3061: 
3062: %-----------------------------------------------------------------%
3063: \begin{figure}[b]
3064:   \begin{center}
3065: \epsfig{ file=f22.eps,
3066:          angle=0,
3067:          width=4in}
3068: %str_ddep.eps
3069: \caption{The interior evolution in early phases 
3070: for different deuterium abundances with the accretion
3071: rate $\dot{M}_\ast = 10^{-5}~M_{\odot}/{\rm yr}$.
3072: The deuterium abundances are [D/H] = $3 \times 10^{-5}$ 
3073: (top; run MD5-dh3), $2.5 \times 10^{-5}$ (middle; MD5), 
3074: and $1 \times 10^{-5}$ (bottom; MD5-dh1).
3075: Properties of the interior structure are presented in the same
3076: manner as in Fig.~\ref{fig:str_fdtmax_1em3}.
3077: In each panel, the thick solid line represents the position
3078: of the accretion shock front. The convective regions are shown 
3079: by the gray-shaded area. 
3080: In the top (bottom) panel, the position of the accretion shock front
3081: for [D/H] = $2.5 \times 10^{-5}$ ([D/H]=0, run MD5-noD) is 
3082: plotted with the dot-dashed curve for comparison. 
3083: }
3084: \label{fig:str_ddep}
3085:   \end{center}
3086: \end{figure}
3087: %-------------------------------------------------------------------%
3088: 
3089: %-----------------------------------------------%
3090: \subsection{Effect of Difference in Initial Models}
3091: \label{ssec:psini}
3092: %-----------------------------------------------%
3093: 
3094: %-------------------------------------------------------------------%
3095: \begin{figure}[t]
3096:   \begin{center}
3097: \epsfig{ file=f23.eps,
3098:          angle=0,
3099:          width=4in}
3100: %str_fdtmax_1em5_ps.eps
3101: \caption{
3102: Same as Fig.~\ref{fig:str_fdtmax_1em5} but for the different
3103: initial model calculated following PS91 
3104: (run MD5-ps91).
3105: In the upper panel, the filled circles denote the protostellar radius 
3106: by PS91 with the same parameter as our fiducial case (MD5), 
3107: which is shown by the dashed curve.
3108: PS91 started calculation at $M_{*,0} = 1~M_{\odot}$, while the 
3109: initial mass in our calculation is much smaller 
3110: $M_{*,0} = 0.01~M_{\odot}$. 
3111: The dot-dashed curve also presents the protostellar radius
3112: in run MD5-ps91, where the deuterium 
3113: burning is turned off at $M_{\ast} = 2.7~M_{\odot}$, 
3114: when a radiative barrier emerges within the star (see text).
3115:  }
3116: \label{fig:str_fdtmax_1em5_ps}
3117:   \end{center}
3118: \end{figure}
3119: %-------------------------------------------------------------------%
3120: 
3121: Evolution of accreting intermediate-mass 
3122: ($1~M_{\odot} < M_{\ast} < 8~M_{\odot}$)
3123: protostars have been studied in some previous work (e.g., PS91). 
3124: Our results agree well with the previous work, 
3125: but still show some differences even with the same accretion rate, 
3126: which can be ascribed to different initial models. 
3127: Whereas our initial model is a tiny radiative protostar with 
3128: $\ll 1M_{\sun}$ (see \S~\ref{ssec:initial}), the initial model of PS91, for
3129: example, is a $1~M_{\odot}$ fully-convective protostar. 
3130: We confirmed that our code reproduces almost the same evolution 
3131: as that of PS91, if we begin the calculation with the same initial model.
3132: 
3133: The initial model of PS91 is constructed in a manner similar to
3134: that described in \S~\ref{ssec:initial}.
3135: The homogeneous entropy distribution is adopted in place of 
3136: equation (\ref{eq:s0}), and its value is taken from the
3137: semi-analytic model of \citet{St88}. 
3138: We have reconstructed the model of \citet{St88}, and obtained 
3139: $s_{c,0} = -4.12~k_{\rm B}/m_{\rm H}$ for a $1~M_{\odot}$ star 
3140: with our adopted opacity tables.
3141: As in \S~\ref{ssec:initial}, we separately solve the core interior
3142: and the surface super-adiabatic layer. 
3143: First, we integrate the core interior outward from the center with a 
3144: guessed value of the central pressure.  
3145: We also guess the luminosity at the bottom of
3146: the surface super-adiabatic layer in advance, and record the entropy 
3147: gradient $\partial s/\partial M$ using equation (\ref{eq:heat}). 
3148: Once the recorded $\partial s/\partial M$ attains a small finite value, 
3149: we switch the scheme and integrate all equations 
3150: (\ref{eq:con}) -- (\ref{eq:heat}).
3151: The guessed quantities are adjusted for the core to satisfy the  
3152: outer boundary conditions.
3153: The constructed model is a fully convective star, 
3154: whose mass and radius are $1~M_{\odot}$ and $4.2~R_{\odot}$. 
3155: 
3156: We calculate the
3157: subsequent evolution beginning with this initial model (run MD5-ps91).
3158: Figure \ref{fig:str_fdtmax_1em5_ps} shows the evolution of such a 
3159: protostar. Our calculation reproduces 
3160: the basic features of a calculation done by PS91, which adopted
3161: the same accretion rate and accretion-shock outer boundary.
3162: In an early phase, the star remains fully convective and 
3163: deuterium burning occurs around the stellar center. 
3164: Deuterium concentration significantly falls in this
3165: phase. This is because the central temperature continuously increases,
3166: and total burning rate $L_{\rm D}$ slightly exceeds
3167: the steady burning rate $L_{\rm D,st}$ \citep[e.g.,][]{PS93}.
3168: These features are somewhat different from our fiducial run MD5, 
3169: where a radiative core persists throughout the evolution and gradually 
3170: expands with the increase in $M_{\ast}$ by accretion 
3171: (see Fig.~\ref{fig:str_fdtmax_1em5}). 
3172: These differences come from different initial models.
3173: Figure \ref{fig:str_fdtmax_1em5} also shows that mass-averaged deuterium 
3174: concentration $f_{\rm d,av}$ does not fall much below 0.01 in our run MD5.
3175: This is because the deuterium burning layer gradually moves to 
3176: the outer layer, where temperature is always around $10^6$~K
3177: and $L_{\rm D} \sim L_{\rm D,st}$. 
3178: At $M_{\ast} = 2~M_{\odot}$, for example, $f_{\rm d,av} \sim 0.04$ in
3179: run MD5, but only $f_{\rm d,av} \sim 10^{-5}$ in run MD5-ps91.
3180: Therefore, our calculation predicts that intermediate-mass 
3181: protostars have a larger amount of deuterium than assumed by PS91. 
3182: Figure \ref{fig:str_fdtmax_1em5_ps} shows that the fully convective 
3183: phase ends at $M_{\ast} \simeq 2.6~M_{\odot}$, when a thin
3184: radiative layer, i.e., radiative barrier, suddenly appears 
3185: within the star. 
3186: The radiative barrier blocks inward convective transport of 
3187: the accreted deuterium.
3188: Consequently, the inner region, where deuterium has run out, 
3189: immediately returns to being radiative.
3190: After that, the convective layer remains only above 
3191: the radiative core, and deuterium burning occurs 
3192: at the bottom of the convective layer.
3193: As in our fiducial run MD5, the protostar swells up at
3194: $M_{\ast} \simeq 3.4~M_{\odot}$. 
3195: PS91 have concluded that this swelling
3196: is caused by the shell-burning of deuterium.  
3197: To see the role of the deuterium shell-burning in the swelling, 
3198: we also calculate evolution with cessation of the deuterium burning 
3199: after the appearance of the radiative barrier. 
3200: We find that the swelling occurs similarly, although it is 
3201: slightly delayed, even without deuterium burning. 
3202: This is caused by the outward transport of embedded entropy,
3203: which is observed as propagation of a luminosity wave 
3204: (see \S~\ref{ssec:md_1em3} and \ref{ssec:md_1em5}). 
3205: Although PS91 have attributed the swelling only to the shell-burning of
3206: deuterium, we conclude that the main driver of the swelling is always 
3207: the propagation of the luminosity wave.
3208: After the turn-around of the radius, the star enters the subsequent
3209: KH contraction phase. The calculated mass-radius relation thereafter 
3210: gradually converges with that in our fiducial run MD5.
3211: The epoch of the Hydrogen burning, $M_{\ast} \simeq 7~M_{\odot}$, 
3212: also agrees with that in the fuducial run.
3213: 
3214: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
3215: \section{Calibration of One-zone Model}
3216: \label{sec:one_zone}
3217: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
3218: 
3219: %------------------------------------------------------------------%
3220: \begin{figure}[t]
3221:   \begin{center}
3222: \epsfig{ file=f24.eps,
3223:          angle=0,
3224:          width=4in}
3225: %1zone_fit.eps
3226: \caption{Comparison between the protostellar radii calculated by 
3227: the one-zone models based on \citet{MT03} but with parameters 
3228: calibrated as in Sec.\ref{sec:one_zone} (solid) and by 
3229: our numerical models (dotted). 
3230: The cases with the accretion rates from $10^{-6}$ to 
3231: $10^{-3}~M_{\odot}/{\rm yr}$ are shown.
3232:  }
3233: \label{fig:1zone}
3234:   \end{center}
3235: \end{figure}
3236: %-------------------------------------------------------------------%
3237: 
3238: While detailed numerical calculations as in this paper 
3239: are a direct method to study the structure of accreting protostars,
3240: one-zone modeling with the polytropic equation of state 
3241: $P = K \rho^{1 + 1/N}$ is also a useful way. 
3242: In the one-zone models of protostellar evolution, 
3243: which were originally developed by Nakano, Hasegawa, \& Norman (1995), 
3244: an energy budget of the protostar is considered.
3245: The total energy of the star $E$ and its time-derivative $dE/dt$
3246: are written as functions of the stellar mass $M_\ast$ and radius $R_\ast$
3247: respectively. 
3248: The mass-radius relation is obtained by substituting the
3249: equation $E$ into $dE/dt$.
3250: \citet{Nk00} applied such a model to study protostellar 
3251: evolution at the very high accretion rate of 
3252: $\dot{M}_\ast \sim 10^{-2}~M_{\odot}/{\rm yr}$.
3253: \citet{MT03} improved the model by calibrating it
3254: with numerical calculations at the accretion
3255: rates of $\leq 10^{-4}~M_{\odot}/{\rm yr}$ (e.g., PS91).
3256: \citet{TM04} also showed that the one-zone models can reproduce 
3257: the mass-radius relations of primordial protostars
3258: at the high accretion rates $\sim 10^{-3}~M_{\odot}/{\rm yr}$
3259: \citep[e.g.,][]{OP03}.
3260: 
3261: Here, we present one-zone models reproducing our numerical 
3262: results (Fig.~\ref{fig:1zone}).
3263: These models are constructed following \citet{MT03} with some
3264: improvements. 
3265: Our numerical calculations have shown that 
3266: protostars are initially radiative for any accretion rate.
3267: With an accretion rate less than $10^{-4}~M_{\odot}/{\rm yr}$,
3268: the convective layers gradually extend as deuterium burning 
3269: becomes significant.
3270: Since the polytropic models with index $N = 1.5$ and $N = 3$ 
3271: approximate a fully convective and radiative star respectively
3272: \citep[e.g.,][]{CG68}, we adopt intermediate values: 
3273: $N = 2.5$ for $\dot{M}_\ast \le 10^{-5}~M_{\odot}/{\rm yr}$ 
3274: and slightly larger values for higher accretion rates, 
3275: $N = 2.5 + 0.25~\log (\dot{M}_\ast/10^{-5}~M_{\odot}/{\rm yr})$.
3276: Initial masses and radii are arbitrarily taken as 
3277: $M_{*,0} = 0.1~M_{\odot}$ and 
3278: $R_{*,0} = 2.5~R_{\odot} (\dot{M}_\ast/10^{-5}~M_{\odot}/{\rm yr})^{1/3}$.
3279: The steady deuterium burning is assumed to begin when the 
3280: central temperature reaches $1.5 \times 10^6$~K. 
3281: After that, we adopt $N = 1.75$, which is slightly larger 
3282: than the fully convective value $N = 1.5$, 
3283: to resemble our numerical result, where a radiative core 
3284: gradually expands with the increase of the stellar mass
3285: (see \S~\ref{ssec:md_1em5}).
3286: Subsequently, the protostar swells up with the propagation 
3287: of the luminosity wave. 
3288: We include this effect into the model 
3289: by artificially increasing the stellar radius 
3290: by a factor of 3 to fit the calculated mass-radius relations. 
3291: We assume that this occurs when the ratio $t_{\rm KH}/t_{\rm acc}$ 
3292: reaches 1.75. 
3293: Other adjustments of models are the same as those of 
3294: \citet{MT03}. The presented one-zone models fit our 
3295: mass-radius relations within 30\% deviation except in 
3296: very early phases for high accretion rates.
3297: 
3298: 
3299: \end{document}
3300: 
3301: 
3302: 
3303: 
3304: 
3305: 
3306: 
3307: 
3308: 
3309: 
3310: 
3311: 
3312: 
3313: 
3314: 
3315: 
3316: 
3317: 
3318: