1: % iaus2esa.tex -- sample pages for Proceedings IAU Symposium document class

2: % (based on v1.0 cca2esam.tex)

3: % v1.04 released 17 May 2004 by TechBooks

4: %% small changes and additions made by KAvdH/IAU 4 June 2004

5: % Copyright (2004) International Astronomical Union

6: 

7: \NeedsTeXFormat{LaTeX2e}

8: 

9: 

10: \documentclass{iaus}

11: \usepackage{graphicx}

12: 

13: \newcommand{\apj}       {\emph{ApJ}}

14: 

15: 

16: \newcommand     \slN    {{\hbox{\slantsym\char'116}}}   %calligraphic N

17: 

18: \newcommand     \Angstrom       {\,{\rm \AA}}

19: \newcommand     \AU     {\,{\rm AU}}

20: \newcommand     \cm     {\,{\rm cm}}

21: \newcommand     \days   {\,{\rm d}}

22: \newcommand     \diss   {{\rm diss}}

23: \newcommand     \erg    {\,{\rm ergs}}

24: 

25: \newcommand     \eV     {\,{\rm eV}}

26: \newcommand     \fl     {{\rm fl.}}

27: \newcommand     \g              {\,{\rm g}}

28: \newcommand     \gtsim  {\gtrsim}                %apj version

29: \newcommand     \He     {{\rm He}}

30: \newcommand     \HH     {{\rm H}_2}

31: \newcommand     \Hz     {\,{\rm Hz}}

32: \newcommand     \Jy     {\,{\rm Jy}}

33: \newcommand     \keV    {\,{\rm keV}}

34: \newcommand     \K      {\,{\rm K}}

35: \newcommand     \kms    {\,{\rm km~s}^{-1}}

36: \newcommand     \kpc    {\,{\rm kpc}}

37: \newcommand     \Lsol   {L_{\odot}}

38: \newcommand     \ltsim  {\lesssim}               %apj version

39: \newcommand     \Mpc    {\,{\rm Mpc}}

40: \newcommand     \Msol   {M_{\odot}}

41: \newcommand     \NN     {\tilde{N}}             % Number of molecules in shell

42: \newcommand     \pc     {\,{\rm pc}}

43: \newcommand     \pd     {\gamma}

44: \newcommand     \rmH    {{\rm H}}

45: \newcommand     \s      {\,{\rm s}}

46: \newcommand     \sr     {\,{\rm sr}}

47: \newcommand     \UV     {{\rm uv}}

48: \newcommand     \yr     {\,{\rm yr}}

49: \newcommand{\btdnote}[1]{{\bf[#1]}}

50: 

51: \newcommand{\smyr}      {{ M_\odot\ \rm yr^{-1}}}

52: \newcommand{\sm}        {{ M_\odot}}

53: 

54: 

55: \newcommand{\beq}       {\begin{equation}}

56: \newcommand{\eeq}       {\end{equation}}

57: \newcommand{\beqa}      {\begin{eqnarray}}

58: \newcommand{\eeqa}      {\end{eqnarray}}

59: \newcommand{\e}         {$^{-1}$}

60: \newcommand{\ee}        {$^{-2}$}

61: \newcommand{\eee}       {$^{-3}$}

62: \newcommand{\caln}      {{\cal N}}

63: 

64: \newcommand{\fkep}      {f_{\rm Kep}}

65: \newcommand{\krho}      {{k_\rho}}

66: \newcommand{\mbe}       {M_{\rm BE}}

67: \newcommand{\mds}       {\dot m_*}

68: \newcommand{\msd}       {m_{*d}}

69: \newcommand{\msdo}       {m_{*d,\, 0}}

70: \newcommand{\mdsd}      {\dot m_{*d}}

71: \newcommand{\mdsdo}      {\dot m_{*d,\, 0}}

72: \newcommand{\tff}       {t_{\rm ff}}

73: \newcommand{\mcore}     {M_{\rm core}}

74: \newcommand{\msf}       {m_{*f}}

75: 

76: 

77: \newcommand{\htwo}      {H$_2$}

78: \newcommand{\nh}        {n_{\rm H}}

79: \newcommand{\mdir}      {M_{\rm dir}}

80: \newcommand{\muh}       {\mu_{\rm H}}

81: \newcommand{\scr}       {S_{\rm cr}}

82: \newcommand{\vp}        {\varpi}

83: \newcommand{\vpo}       {\varpi_0}

84: \newcommand{\atwo}      {\alpha^{(2)}}

85: 

86: 

87: \newcommand{\lal}       {Lyman-$\alpha$}

88: \newcommand{\btl}       {\bar\tau_L}

89: \newcommand{\calf}      {{\cal F}}

90: \newcommand{\crit}      {{\rm crit}}

91: \newcommand{\dnd}       {\Delta\nu_D}

92: \newcommand{\dvd}       {\Delta v_D}

93: \newcommand{\dvds}      {\Delta v_{D,\, 6}}

94: \newcommand{\iso}       {{\rm iso}}

95: \newcommand{\necr}      {n_{e,\,\rm cr}}

96: \newcommand{\rhk}       {\rho_{\rm HK}}

97: \newcommand{\taueff}    {\bar\tau_{\rm eff}}

98: \newcommand{\vecfrad}   {{\bf f}_{\rm rad}}

99: \newcommand{\prad}      {{P}_{\rm rad}}

100: \newcommand{\pradi}     {{P}_{\rm rad,\, iso}}

101: \newcommand{\pradrr}    {{P}_{{\rm rad},rr}}

102: \newcommand{\urad}      {{u}_{\rm rad}}

103: \newcommand{\uradi}      {{u}_{\rm rad,\, iso}}

104: \newcommand{\rvecprad}  {{\bf P}_{\rm rad}}

105: \newcommand{\vecF}      {{\bf F}}

106: \newcommand{\vecFhat}   {{\bf \hat F}}

107: \newcommand{\vecnabla}  {{\bf\nabla}}

108: \newcommand{\vecnhat}	{{\bf \hat n}}

109: 

110: 

111: 

112: \newcommand{\avg}[1]    {\langle #1 \rangle}

113: \newcommand{\pbyp}[2]   {\frac{\partial #1}{\partial #2}}

114: \newcommand{\cgc}       {c_{gc}}

115: \newcommand{\esd}       {\epsilon_{*d}}

116: \newcommand{\esdb}       {\bar\epsilon_{*d}}

117: \newcommand{\fsh}       {f_{\rm sh}}

118: \newcommand{\mdst}      {\dot m_{*,\,-3}}

119: \newcommand{\mdsdt}     {\dot m_{*d,\,-3}}

120: \newcommand{\mst}       {m_{*,\,2}}

121: \newcommand{\msdt}      {m_{*d,\,2}}

122: \newcommand{\mss}       {m_{*,\, s}}

123: \newcommand{\meq}       {M_{\rm eq}}

124: \newcommand{\ssc}       {\left(\frac{\Sigma}{\Sigma_c}\right)}

125: \newcommand{\kapr}      {\kappa_{\rm R}}

126: \newcommand{\kht}       {\kappa_{\rm T}}

127: \newcommand{\kph}       {\kappa_{\rm R,\, ph}}

128: \newcommand{\ledd}      {L_{\rm Edd}}

129: \newcommand{\ncr}       {n_{\rm cr}}

130: \newcommand{\rcr}       {r_{\rm cr}}

131: %\newcommand{\rph}      {r_{\rm ph}}

132: \newcommand{\rtwo}	{r_{\rm HII}}

133: \newcommand{\sph}       {\Sigma_{\rm ph}}

134: \newcommand{\teff}      {T_{\rm eff}}

135: \newcommand{\tcr}       {t_{\rm cr}}

136: \newcommand{\tyr}       {t_{\rm yr}}

137: \newcommand{\vff}       {v_{\rm ff}}

138: \newcommand{\vif}       {v_{\rm IF}}

139: \newcommand{\zph}       {z_{\rm ph}}

140: \newcommand{\zsr}       {z_{sr}}

141: \newcommand{\krs}       {\kappa_{{\rm R}, s}}

142: \newcommand{\phiedd} {\phi_{\rm Edd}}

143: 

144: \def\ion#1#2{#1$\;${\small\rm II}\relax}

145: \def\lesssim{\mathrel{\hbox{\rlap{\hbox{\lower4pt\hbox{$\sim$}}}\hbox{$<$}}}}

146: \def\gtrsim{\mathrel{\hbox{\rlap{\hbox{\lower4pt\hbox{$\sim$}}}\hbox{$>$}}}}

147: 

148: \title[Population III.1 stars] %% give here short title %%

149: {Population III.1 stars: formation, feedback and evolution of the IMF}

150: 

151: 

152: \author[Jonathan C. Tan]   %% give here short author list %%

153: {Jonathan C. Tan$^1$

154: %%  \thanks{Present address: Fluid Mech Inc., 24 The Street, Lagos, Nigeria.},

155: }

156: 

157: \affiliation{$^1$Dept. of Astronomy, University of Florida, Gainesville, FL 32611, USA\\ email: {\tt jt@astro.ufl.edu}}

158: 

159: 

160: \pubyear{2008}

161: \volume{255}  %% insert here IAU Symposium No.

162: \pagerange{119--126}

163: % \date{?? and in revised form ??}

164: \setcounter{page}{119}

165: \jname{Low-Metallicity Star Formation: From the First Stars to Dwarf Galaxies}

166: \editors{L.K. Hunt, S. Madden \& R. Schneider, eds.}

167: \begin{document}

168: 

169: \maketitle

170: 

171: \begin{abstract}

172: I discuss current theoretical expectations of how primordial, Pop

173: III.1 stars form. Lack of direct observational constraints makes

174: this a challenging task. In particular predicting the mass of

175: these stars requires solving a series of problems, which all affect,

176: perhaps drastically, the final outcome. While there is general

177: agreement on the initial conditions, $\rm H_2$-cooled gas at the

178: center of dark matter minihalos, the subsequent evolution is more

179: uncertain. In particular, I describe the potential effects of dark

180: matter annihilation heating, fragmentation within the minihalo,

181: magnetic field amplification, and protostellar ionizing

182: feedback. After these considerations, one expects that the first stars

183: are massive $\gtrsim 100\sm$, with dark matter annihilation heating

184: having the potential to raise this scale by large factors. Higher

185: accretion rates in later-forming minihalos may cause the

186: Pop III.1 initial mass function to evolve to higher masses.

187: %including the initial conditions, protostellar evolution,

188: %magnetic field growth, radiative and mechanical feedback, and possible

189: %effects due to dark matter annihilation. I discuss the predicted

190: %initial mass function and its evolution with redshift. Finally I

191: %discuss some implications for galaxy and supermassive black hole

192: %formation.

193: \keywords{stars: formation, galaxies: formation, dark matter, cosmology: theory}

194: %% add here a maximum of 10 keywords, to be taken form the file <Keywords.txt>

195: \end{abstract}

196: 

197: \firstsection % if your document starts with a section,

198:               % remove some space above using this command.

199: \section{Introduction: The Importance of Pop III.1 Stars and their IMF}

200: 

201: The first, essentially metal-free (i.e. Population III), stars are

202: expected to have played a crucial role in bringing the universe out of

203: the dark ages: initiating the reionization process, including the

204: local effects of their \ion{H}{2} regions in generating shocks and

205: promoting formation of molecular coolants in the relic phase;

206: photodissociating molecules; amplifying magnetic fields to possibly

207: dynamically important strengths; and generating the mechanical

208: feedback, heavy elements and possible neutron star or black hole

209: remnants associated with supernovae. In these ways Pop III stars laid

210: the foundations for galaxy formation, including supermassive black

211: holes and globular clusters. Many of these processes are theorized to

212: depend sensitively on the initial mass function (IMF) of Pop III

213: stars, thus motivating its study. The formation of the first Pop III

214: stars in a given region of the universe is expected to have been

215: unaffected by other astrophysical sources and these have been termed

216: Pop III.1, in contrast to Pop III.2 (McKee \& Tan 2008, hereafter

217: MT08). Pop III.1 are important for influencing the initial conditions

218: for future structure formation and for having their properties

219: determined solely by cosmology. There is also the possibility,

220: described in Sect.~\ref{S:DM}, that Pop III.1 star formation may be

221: sensitive to the properties of weakly interacting massive particle

222: (WIMP) dark matter.

223: 

224: Unfortunately, at the present time and in the near future we expect

225: only indirect observational constraints on the Pop III IMF. The epoch

226: of reionization can be constrained by CMB polarization (Page et

227: al. 2007) and future high redshift 21~cm HI observations (e.g. Morales

228: \& Hewitt 2004). Metals from individual Pop III supernovae may have

229: imprinted their abundance patterns in very low metallicity Galactic

230: halo stars (Beers \& Christlieb 2005) or in the Ly-$\alpha$ forest

231: (Schaye et al. 2003; Norman, O'Shea, \& Pascos 2004). Light from the

232: first stars may contribute to the observed NIR background intensity,

233: (e.g. Santos, Bromm, \& Kamionkowski 2002), and its fluctuations

234: (Kashlinsky et al. 2004; c.f. Thompson et al. 2007). If massive,

235: supernovae marking the deaths of the first stars may be observable by

236: JWST (Weinmann \& Lilly 2005). If these supernovae produce gamma-ray

237: bursts then these may already be making a contribution to the

238: population observed by SWIFT (Bromm \& Loeb 2002).

239: 

240: 

241: %The first stars thus set the stage for galaxy formation,

242: %for which observational constraints now exist at redshifts 8 to 10

243: %(Stark et al. 2007), and supermassive black hole formation and growth,

244: %with $\sim$billion solar mass black holes being seen out to redshift

245: %6.4 (Fan et al. 2003; Willott et al. 2003).

246: 

247: The lack of direct observations of Pop III star formation means

248: theoretical models lack constraints, which is a major problem for

249: treating such a complicated, nonlinear process. Numerical simulations

250: have been able to start with cosmological initial conditions and

251: advance to the point of protostar formation (see Yoshida et al., these

252: proceedings), but progressing further through the protostellar

253: accretion phase requires additional modeling of complicated

254: processes, including a possible need to include extra physics such as

255: WIMP annihilation and magnetic fields. Building up a prediction of the

256: final mass achieved by the protostar, i.e. the initial mass (function)

257: of the star (population), is akin to building a house of cards: the

258: reliability of the structure becomes more and more precarious.

259: 

260: In this article we summarize theoretical attempts to understand the

261: formation process and resulting IMF of Pop III.1 stars. We have

262: reviewed much of these topics previously (Tan \& McKee 2008), so here

263: we concentrate on a discussion of some of the more uncertain aspects

264: in these models, including the potential effects of WIMP

265: annihilation on Pop III.1 star formation, fragmentation during Pop

266: III.1 star formation, the generation of magnetic fields, the

267: uncertainties in predicting the IMF from feedback models, and the

268: evolution of the Pop III.1 IMF. Note, when discussing possible

269: fragmentation during the formation of a Pop III.1 star, we will

270: consider all stars that result from the same minihalo to be Pop III.1,

271: i.e. they are unaffected by astrophysical sources external to their own

272: minihalo.

273: 

274: 

275: 

276: %\cite[Anders \& Zinner (1993)]{AndersZinner93} and 

277: %\cite[Ott (1993)]{Ott93}.

278: 

279: 

280: %(Fig.\,\ref{fig1}).

281: 

282: 

283: 

284: 

285: \section{Initial Conditions and Possible Effects of WIMP Annihilation}

286: 

287: The initial conditions for the formation of the first stars are

288: thought to be relatively well understood: they are determined by the

289: growth of small-scale gravitational instabilities from cosmological

290: fluctuations in a cold dark matter universe. The first stars are

291: expected to form at redshifts $z\sim 10-50$ in dark matter

292: ``minihalos'' of mass $\sim 10^6 M_\odot$ (Tegmark et al. 1997).  In

293: the absence of any elements heavier than helium (other than trace

294: amounts of lithium) the chemistry and thermodynamics of the gas are

295: very simple. Once gas collects in the relatively shallow potential

296: wells of the minihalos, cooling is quite weak and is dominated by the

297: ro-vibrational transitions of trace amounts of $\rm H_2$ molecules

298: that cool the gas to $\sim 200$~K at densities $n_{\rm H} \sim

299: 10^4\:{\rm cm^{-3}}$ (Abel, Bryan, \& Norman 2002; Bromm, Coppi, \&

300: Larson 2002). Glover et al. (these proceedings) review the effects of

301: other potential coolants, finding they are small for Pop III.1 star

302: formation.

303: 

304: As the gas core contracts to greater densities, the $\rm H_2$ cooling

305: becomes relatively inefficient and the temperature rises to $\sim

306: 1000$~K. At densities $\sim 10^{10}\:{\rm cm^{-3}}$ rapid 3-body

307: formation of $\rm H_2$ occurs, creating a fully molecular region that

308: can cool much more efficiently. This region starts to collapse

309: supersonically until conditions become optically thick to the line and

310: continuum cooling radiation, which occurs at densities $\sim

311: 10^{17}\:{\rm cm^{-3}}$.  Recent 3D numerical simulations have

312: advanced to densities of order $10^{21}$~cm$^{-3}$ (see contribution

313: by Yoshida et al., these proceedings), but have trouble proceeding

314: further given the short timesteps required to resolve the dynamics of

315: the high density gas of the protostar. Further numerical progress can

316: be achieved by introducing sink particles (Bromm \& Loeb 2004) or with

317: 1D simulations (Omukai \& Nishi 1998; Ripamonti et al. 2002).

318: 

319: Alternatively, given the above initial conditions, the subsequent

320: accretion rate to the protostar can be calculated analytically (Tan \&

321: McKee 2004, hereafter TM04).  The accretion rate depends on the

322: density structure and infall velocity of the gas core at the point

323: when the star starts to form. Omukai \& Nishi (1998) and Ripamonti et

324: al. (2002) showed that the accreting gas is isentropic with an

325: adiabatic index $\gamma\simeq 1.1$ due to H$_2$ cooling; i.e., each

326: mass element satisfies the relation $P=K\rho^\gamma$ with the

327: ``entropy parameter'' $K=$~const.  In hydrostatic equilibrium---and

328: therefore in a subsonic contraction---such a gas has a density profile

329: $\rho \propto r^{-k_\rho}$ with $k_\rho\simeq 2.2$, as is seen in

330: simulations.

331: %corresponding to the hydrostatic equilibrium configuration of a

332: %polytrope with pressure $P(r)=K_p\rho(r)^{\gamma_p}$ and polytropic

333: %index $\gamma_p\simeq 1.1$.   

334: %This is confirmed by numerical simulations, which show

335: %that just before protostar formation the gas at the center of the

336: %minihalo has organized itself into an approximately singular

337: %polytropic sphere (with a modest degree of flattening about a rotation

338: %axis) with $k_\rho=2.2$. 

339: TM04 describe the normalization of the core density

340: structure via the ``entropy parameter''

341: \beq 

342: K'\equiv (P/\rho^\gamma)/1.88\times 10^{12}~{\rm cgs}=

343: (\teff'/300~{\rm K})(n_{\rm H}/10^4{\rm cm}^{-3})^{-0.1},

344: \label{eq:kp}

345: \eeq

346: where $\teff' \equiv T+\mu\sigma^2_{\rm turb}/k$ is an effective

347: temperature that includes the modest effect of subsonic turbulent motions

348: that are seen in numerical simulations (Abel et al. 2002).

349: 

350: %The normalization of the accretion rate also depends on the velocity

351: %of gas in the envelope at the time of protostar formation.

352: For the infall velocity at the time of protostar formation,

353: simulations show the gas is inflowing subsonically at about a third of

354: the sound speed (Abel et al 2002). Hunter's (1977) solution for mildly

355: subsonic inflow (Mach number =0.295) is the most relevant for this

356: case. It has a density that is 1.189 times greater than a singular

357: isothermal sphere (Shu 1977) at $t=0$, and an accretion rate that is

358: 2.6 times greater.

359: 

360: Feedback from the star, whether due to winds, photoionization, or

361: radiation pressure, can reduce the accretion rate of the star. TM04

362: and MT08 define a hypothetical star$+$disk mass, $\msdo$, and

363: accretion rate, $\mdsdo$, in the absence of feedback. In this case,

364: the star$+$disk mass equals the mass of the part of the core (out to

365: some radius, $r$, that has undergone inside-out collapse) from which

366: it was formed, $\msdo=M(r)$.  The instantaneous and mean star

367: formation efficiencies are $\esd\equiv \mdsd / \mdsdo$ and $\esdb

368: \equiv \msd / \msdo = \msd/M$, respectively.

369: 

370: Assuming the Hunter solution applies for a singular polytropic sphere

371: with $\gamma=1.1$, the accretion rate is then (TM04)

372: \beq

373: \label{eq:mdot}

374: \mdsd=0.026 \epsilon_{*d}K'^{15/7} (M/\sm)^{-3/7}~M_\odot~{\rm yr}^{-1},

375: \eeq

376: with the stellar mass smaller than the initial enclosed core mass via

377: $m_* \equiv \msd/(1+f_d) =\esdb M/(1+f_d)$. We choose a fiducial value

378: of $f_d=1/3$ appropriate for disk masses limited by enhanced viscosity

379: due to self-gravity.

380: 

381: \subsection{Possible Effects of Dark Matter Annihilation}\label{S:DM}

382: 

383: %\begin{figure}[b]

384: %% \vspace*{-2.0 cm}

385: %\begin{center}

386: % \includegraphics[width=5in]{f5.eps} 

387: %% \vspace*{-1.0 cm}

388: % \caption{Ratio of dark matter annihilation heating rate, $H(r)$, to

389: %baryonic cooling rate, $C(r)$, (from Natarajan et al. 2008) for dark

390: %matter density fits \#1, i.e. a power law with $\alpha_\chi\simeq 1.5$

391: %(top panels), and \#2, i.e. a steeper power law with

392: %$\alpha_\chi\simeq 2.0$ (bottom panels), for three simulated minihalos

393: %A (solid lines), B (long-dashed lines), C (dot-dashed lines) for

394: %fiducial $m_\chi=100$~GeV. Results for simulation A with $m_\chi =

395: %10$~GeV (dotted lines) and $m_\chi = 1$~TeV (dashed lines) are also

396: %shown.  }

397: %   \label{fig:dm}

398: %\end{center}

399: %\end{figure}

400: 

401: Pop III.1 stars form at the centers of dark matter (DM)

402: minihalos. While the mass density is dominated by baryons inside $\sim

403: 1$~pc, adiabatic contraction ensures that there will still be a

404: peak of DM density co-located with the baryonic

405: protostar. As discussed by Spolyar et al. (2008), if the dark matter

406: consists of a weakly interacting massive particle (WIMP) that self

407: annihilates, then this could lead to extra heating that can help

408: support the protostar against collapse. Spolyar et al. calculated

409: that, depending on the dark matter density profile, WIMP mass, and

410: annihilation cross section, the local heating rate due to dark matter

411: could exceed the baryonic cooling rate for densities $n_{\rm H}\gtrsim

412: 10^{14}\:{\rm cm^{-3}}$, corresponding to scales of about 20~AU from

413: the center of the halo/protostar.

414: 

415: Natarajan, Tan, \& O'Shea (2008) revisited this question by

416: considering several minihalos formed in numerical simulations. While

417: there was some evidence for adiabatic contraction leading to a

418: steepening of the dark matter density profiles in the centers of the

419: minihalos, this was not well resolved on the scales where heating may

420: become important. Thus various power law ($\rho_\chi\propto

421: r^{-\alpha_\chi}$) extrapolations were considered for the DM

422: density. A value of $\alpha_\chi\simeq1.5$ was derived based on the

423: numerically well-resolved regions at $r\sim1$~pc. A steeper value of

424: $\alpha_\chi\simeq 2.0$ was derived based on the inner regions of the

425: simulations. In the limit of very efficient adiabatic contraction, one

426: expects the dark matter density profile to approach that of the

427: baryons, which would yield $\alpha_\chi\simeq2.2$.

428: %In Figure~\ref{fig:dm} (from Natarajan et al. 2008) we show the ratio

429: %of the local dark matter annihilation heating rates to baryonic

430: %cooling rates in these models, as a function of radius and baryon

431: %density. These results illustrate the range of conditions under which

432: %dark matter heating can become important. 

433: For the density profiles with $\alpha_\chi\simeq 2.0$, Natarajan et

434: al. (2008) found that dark matter heating inevitably becomes

435: dominant. Natarajan et al. also considered the global

436: quasi-equilibrium structures for which the total luminosity generated

437: by WIMP annihilation that is trapped in the protostar, $L_{\chi,0}$, equals that

438: radiated away by the baryons, assuming both density distributions are

439: power laws truncated at some radius, $r_c$, with a constant density

440: core. This core radius was varied to obtain the equilibrium

441: luminosity. Typical results were $L_{\chi,0}\sim

442: 10^3\:L_\odot$ and $r_c\simeq$ to a few to a few tens of AU.

443: 

444: These scales at which equilibrium is established are important for

445: determining the subsequent evolution of the protostar, which will

446: continue to gain baryons and probably additional dark matter via

447: adiabatic contraction. Even in the limit where no further dark matter

448: becomes concentrated in the protostar, that which is initially present

449: can be enough to have a major influence on the subsequent protostellar

450: evolution. As the protostar gains baryonic mass it requires a greater

451: luminosity for its support. If there was no dark matter heating, the

452: protostar would begin to contract once it becomes older than its local

453: Kelvin-Helmholz time, i.e. on timescales much longer than the stellar

454: dynamical time. If dark matter is present, it will become concentrated

455: as the protostar contracts, and the resulting annihilation luminosity

456: will grow as $L_\chi \simeq L_{\chi,0} (r_*/r_{*,0})^{-3}$, assuming a

457: homologous density profile. For a starting luminosity of $L_{\chi,0} =

458: 1000\:L_\odot$ and radius of $r_{*,0}\simeq r_c = 10$~AU, this can

459: mean luminosities that are easily large enough to support

460: $\sim100\:M_\odot$ stars, i.e. $\sim10^6\:L_\odot$, at sizes of $\sim

461: 1$~AU, i.e. much greater than their main sequence radii, which would

462: be $\simeq 5 R_\odot =0.02$~AU. These estimates are of course very

463: sensitive to the initial size of the protostar.

464: 

465: Full treatment of the protostellar evolution (see Freese et al. 2008

466: for an initial model) requires a model for the evolution of the

467: stellar DM content, which grows by accumulation of surrounding

468: WIMPs, but also suffers depletion due to the annihilation process. The

469: mean depletion time in the star is $t_{\rm dep} =

470: (\rho_\chi/\dot{\rho}_\chi) \simeq m_\chi/(\rho_\chi <\sigma_a v>)

471: \rightarrow 105(m_\chi/100~{\rm GeV}) (\rho_\chi/10^{12}{\rm

472: GeV\:cm^{-3}})^{-1}\:{\rm Myr}$, where we have normalized to typical

473: values of $\rho_\chi$ in the initial DM core (Natarajan et

474: al. 2008). If the protostar contracts from an initial radius of 10~AU

475: to 1~AU then $t_{\rm dep}\simeq 10^5\:{\rm yr}$. This becomes

476: comparable to the growth time of the protostar (i.e. the time since

477: its formation, its age), $t_*=2.92 \times 10^4

478: K'^{-15/7}(m_*/100M_\odot)^{10/7} \: {\rm yr}$ (TM04).

479: We see that, if replenishment of WIMPs in the protostar is negligible,

480: then depletion can become important for AU scale protostars of $\sim

481: 100\:M_\odot$.

482: 

483: 

484: Protostars swollen by DM heating would have much cooler photospheres

485: and thus smaller ionizing feedback than if they had followed standard

486: protostellar evolution leading to contraction to the main sequence by

487: about 100~$M_\odot$. Ionizing feedback is thought to be important in

488: terminating accretion and thus setting the Pop III.1 IMF (MT08; see

489: \S\ref{S:feedback} below). The reduced ionizing feedback of DM-powered

490: protostars may allow them to continue to accrete to much higher masses

491: than would otherwise have been achieved.

492: 

493: 

494: 

495: \section{Protostellar Accretion and Disk Fragmentation}\label{S:accfrag}

496: 

497: Another process that may affect the IMF of the first stars is

498: fragmentation of the infalling gas after the first protostar has

499: formed. TM04 and MT08 considered the growth and evolution of the

500: protostar in the case of no fragmentation (and no DM heating): the

501: final mass achieved by the protostar is expected to be $\sim

502: 100-200\sm$ and set by a balance between its ionizing feedback and its

503: accretion rate through its disk (\S\ref{S:feedback}).

504: 

505: The accretion disk of the protostar 

506: %Once the central protostar has gained a certain amount of mass, the

507: %angular momentum of the newly infalling gas inevitably leads to

508: %formation of a rotationally supported disk. This 

509: does present an environment in which density fluctuations can grow,

510: since there will typically be many local dynamical timescales before

511: the gas is accreted to the star. TM04 calculated the expected disk size, $r_d(m_*)$,

512: assuming conservation of angular momentum inside the sonic point,

513: $r_{\rm sp}$, of the inflow, finding \beq r_d =

514: 1280\left(\frac{\fkep}{0.5}\right)^2\left(

515: \frac{\msdt}{\esdb}\right)^{9/7}K'^{-10/7}~{\rm AU} \rightarrow

516: 1850\left(\frac{\fkep}{0.5}\right)^2 \frac{\mst^{9/7}}{K'^{10/7}}~{\rm

517: AU}

518: %\nonumber\\

519: %        &\rightarrow& 1850\left(\frac{\fkep}{0.5}\right)^2

520: %        \frac{\mst^{9/7}}{K'^{10/7}}~{\rm AU},

521: \label{eq:rd}

522: \eeq where $m_{*d,2}=m_{*d}/100\sm$, $m_{*,2}=m_*/100\sm$, the

523: $\rightarrow$ is for the case with $f_d=1/3$ and $\fkep\equiv v_{\rm

524: rot}(r_{\rm sp}) / v_{\rm Kep}(r_{\rm sp})$, with a typical value of

525: 0.5 seen in numerical simulations.

526: 

527: The high accretion rates of primordial protostars make it likely that

528: the disk will build itself up to a mass that is significant compared

529: to the stellar mass. At this point the disk becomes susceptible to

530: global ($m=1$ mode) gravitational instabilities (Adams, Ruden, \& Shu

531: 1989; Shu et al. 1990), which are expected to be efficient at driving

532: inflow to the star, thus regulating the disk mass. Thus TM04 assumed a

533: fixed ratio of disk to stellar mass, $f_d=1/3$.

534: 

535: Accretion through the disk may also be driven by local instabilities,

536: the effects of which can be approximated by simple Shakura-Sunyaev

537: $\alpha_{\rm ss}$-disk models. Two dimensional simulations of clumpy,

538: self-gravitating disks show self-regulation with $\alpha_{\rm

539: ss}\simeq (\Omega t_{\rm th})^{-1}$ up to a maximum value $\alpha_{\rm

540: ss}\simeq 0.3$ (Gammie 2001), where $\Omega$ is the orbital angular

541: velocity, $t_{\rm th}\equiv \Sigma kT_{\rm c,d}/(\sigma T_{\rm

542: eff,d}^4)$ is the thermal timescale, $\Sigma$ is the surface density,

543: $T_{\rm c,d}$ is the disk's central (midplane) temperature, and

544: $T_{\rm eff,d}$ the effective photospheric temperature at the disk's

545: surface. 

546: 

547: Gammie (2001) found that fragmentation occurs when $\Omega t_{\rm

548: th}\lesssim 3$. This condition has the best chance of being satisfied

549: in the outermost parts of the disk that are still optically

550: thick. However, Tan \& Blackman (2004, hereafter TB04) considered the

551: gravitational stability of constant $\alpha_{\rm ss}=0.3$ disks fed at

552: accretion rates given by eq. \ref{eq:mdot} and found that the

553: optically thick parts of the disk remained Toomre stable ($Q>1$)

554: during all stages of the growth of the protostar. Note that the

555: cooling due to dissociation of $\rm H_2$ and ionization of H was

556: included in these disk models.

557: 

558: We therefore expect that during the early stages of typical Pop III.1

559: star formation, the accretion disk will grow in mass and mass surface

560: density to a point at which gravitational instabilities, both global

561: and local, act to mediate accretion to the star. The accretion rates

562: that can be maintained by these mechanisms are larger than the infall

563: rates of eq.~\ref{eq:mdot}, and so the disk does not fragment. 

564: 

565: We note that if fragmentation does occur and leads to formation of

566: relatively low-mass secondary protostars in the disk, then one

567: possible outcome is the migration of these objects in the disk until

568: they eventually merge with the primary protostar. The end result of

569: such a scenario would not be significantly different from the case of

570: no fragmentation. Another possibility is that a secondary fragment

571: grows preferentially from the circumbinary disk leading to the

572: formation of a massive twin binary system (Krumholz \& Thompson

573: 2007). If both stars are massive, this star formation scenario would

574: be qualitatively similar to the single star case in terms of the

575: effect of radiative feedback limiting accretion. A massive binary

576: system would mean that the accreting gas needs to lose less angular

577: momentum and binary-excited spiral density waves provide an

578: additional, efficient means to transfer angular momentum, compared to

579: the single star case. For close binaries, new stellar evolution

580: channels would be available involving mass transfer and merger, with

581: possible implications for the production of rapidly rotating

582: pre-supernova progenitors and thus perhaps gamma-ray bursts.

583: 

584: 

585: 

586: 

587: 

588: 

589: %Assuming dark matter annihilation does not have a significant effect on

590: %the main phase of protostellar accretion and noting that there may be

591: %a range of accretion rates depending on the formation redshift (O'Shea

592: %et al. 2007), we now consider what happens to the collapsing gas as it

593: %forms a protostar and accretion disk at rates that are initially

594: %$\gtrsim 10^{-2}\smyr$ (i.e. eq.~\ref{eq:mdot} with $K^\prime\sim1$).

595: 

596: %The accretion rate to the protostar is important for determining the

597: %size of the star, the structure of the accretion disk, and the

598: %accretion luminosity (Tan \& McKee 2004). The accretion rate, being

599: %related to the density of gas in the infalling envelope, will also

600: %determine how much feedback is required to halt the accretion process

601: %(McKee \& Tan 2008).

602: 

603: %The above models, and most others in the literature, have assumed a

604: %single protostar forms at the center of the minihalo. If fragmentation

605: %were to produce a cluster of many stars, then each would typically

606: %have a much smaller accretion rate, and a qualitatively different mode

607: %of star formation would result. If fragmentation leads to a binary

608: %system, then the star formation process, i.e. the properties of the

609: %protostar(s), would be qualitatively similar to the single star

610: %case. The major difference would be that the gas needs to lose less

611: %angular momentum and binary-excited spiral density waves provide an

612: %additional, efficient means to transfer angular momentum.

613: 

614: 

615: Fragmentation will only be significant for the IMF if it occurs

616: vigorously and leads to a cluster of lower mass stars instead of a

617: massive single or binary system. Clark, Glover, \& Klessen (2008)

618: claimed such an outcome from the results of their smooth particle

619: hydrodynamical simulation of the collapse of a primordial

620: minihalo. They allowed dense, gravitationally unstable gas to be

621: replaced by sink particles. They found a cluster of 20 or so

622: protostars formed.  As discussed by Clark et al. (2008) (see also

623: Glover et al., these proceedings), there are a number of caveats

624: associated with this result. The initial conditions (a sphere of

625: radius 0.17~pc with an uniform particle density of $5\times 10^5\:{\rm

626: cm^{-3}}$, and ratios of rotational and turbulent energy to

627: gravitational of 2\% and 10\%, respectively) were not derived from ab

628: initio simulations of cosmological structure formation. In particular,

629: cosmologically-formed minihalos evolve towards structures that have very

630: steep density gradients, centered about a single density peak. This is

631: likely to allow the first, central protostar to initiate its formation

632: long before other fluctuations have a chance to develop. The

633: development of a massive central object will create tidal forces in

634: the surrounding gas that will make it more difficult for gravitational

635: instabilities to develop. Furthermore, the surrounding gas is

636: infalling on about a local free fall time, so density perturbations

637: have few local dynamical timescales in which to grow. Another caveat

638: with the Clark et al. fragmentation results is the use of a simple

639: tabulated equation of state, in which gas can respond instantaneously

640: to impulses that induce cooling. This, and the form of the equation of

641: state used, lead to near isothermal conditions in the fragmenting

642: region.

643: 

644: 

645: 

646: %Tan \& McKee (2004) presented Shakura-Sunyaev $\alpha$-disk models for

647: %these accretion disks, calculating the structure assuming the above

648: %accretion rates for constant values of $\alpha=0.01,0.3$. The latter

649: %value was the largest seen in the 2D simulations of Gammie (2001)

650: %before gravitational instability led to fragmentation. 

651: 

652: \section{Magnetic Fields and Hydromagnetic Outflows}\label{S:magnetic}

653: 

654: TB04 considered the growth of magnetic fields in the accretion disk of

655: Pop III.1 protostars. They estimated minimum seed field strengths

656: $\sim 10^{-16}G$. Xu et al. (2008) have recently reported field

657: strengths of up to $10^{-9}G$ generated by the Biermann battery

658: mechanism in their simulations of minihalo formation. Such seed fields

659: are expected to be amplified by turbulence in the disk, attaining

660: equipartition strengths by the time the protostar has a mass of a few

661: solar masses or so. If the turbulence generates large scale helicity,

662: as in the model of Blackman \& Field (2002), then this can lead to the

663: creation of dynamically-strong fields that are ordered on scales large

664: compared to the disk. Such fields, coupled to the rotating accretion

665: disk, are expected to drive hydromagnetic outflows, such as disk winds

666: (Blandford \& Payne 1982).

667: 

668: TB04 then considered the effect of such outflows on the accretion of

669: gas from the minihalo, following the analysis of Matzner \& McKee

670: (2000). The force distribution of centrifugally-launched hydromagnetic

671: outflows is collimated along the rotation axes, but includes a

672: significant wider-angled component. Using the sector approximation,

673: TB04 found the angle from the rotation/outflow axis at which the

674: outflow had enough force to eject the infalling minihalo gas. This

675: angle increased as the protostellar evolution progressed, especially

676: as the star contracted to the main sequence, leading to a deeper

677: potential near the stellar surface and thus larger wind velocities.

678: The star formation efficiency due to protostellar outflow winds

679: remains near unity until $m_*\simeq 100\sm$, and then gradually

680: decreases to values of 0.3 to 0.7 by the time $m_*\simeq 300\sm$,

681: depending on the equatorial flattening of surrounding gas

682: distribution. Comparing these efficiencies to those from ionizing

683: feedback (\S\ref{S:feedback}), we conclude that the latter is more

684: important at determining the Pop III.1 IMF (see also Tan \& McKee 2008).

685: 

686: \section{How Accretion and Feedback Set the IMF}\label{S:feedback}

687: 

688: MT08 modeled the interaction of ionizing feedback on the accretion

689: flow to a Pop III.1 protostar. In the absence of WIMP annihilation

690: heating, the protostar contracts to the main sequence by the time

691: $m_*\simeq 100\sm$, and from there continues to accrete to higher

692: masses. At the same time, the ionizing luminosity increases, leading

693: to ionization of the infalling envelope above and below the plane of

694: the accretion disk. Once the \ion{H}{2} region has expanded beyond the

695: gravitational escape radius for ionized gas from the protostar,

696: pressure forces begin to act to reverse the infall. In the fiducial

697: case, by the stage when $m_*\simeq 100\sm$ we expect infall to have

698: been stopped from most directions in the minihalo. Only those regions

699: shadowed from direct ionizing flux from the protostar by the accretion

700: disk are expected to remain neutral and be able to accrete. 

701: 

702: In these circumstances the protostar starts to drive an ionized wind

703: from its disk (Hollenbach et al. 1994). Ionization

704: from the protostar creates an ionized atmosphere above the neutral

705: accretion disk, which then scatters some ionizing photons down on to

706: the shielded region of the outer disk, beyond $r_g$. An ionized

707: outflow is driven from these regions at a rate

708: \beq

709: \dot{m}_{\rm evap} 

710: \simeq 4.1\times 10^{-5} S_{\rm 49}^{1/2}T_{i,4}^{0.4}\msdt^{1/2}~~\smyr,

711: \eeq

712: where $S_{\rm 49}$ is the H-ionizing photon luminosity in units of

713: $10^{49}$ photons~$\rm s^{-1}$ and $T{i,_4}$ is the ionized gas temperature in units of $10^4$K. 

714: 

715: MT08 used the condition $\dot{m}_{\rm evap}>\dot{m}_*$ for determining

716: the final mass of the protostar. From numerical models they found it

717: is about $140 M_\odot$ in the fiducial case they considered, and

718: Table~\ref{tab:m} summarizes other cases. MT08 also made an analytic

719: estimate, assuming the H-ionizing photon luminosity is mostly due to the

720: main sequence luminosity of the star:

721: \beq

722: S\simeq 7.9\times 10^{49}\; \phi_S \mst^{1.5}~~~~{\rm ph\ s^{-1}},

723: \label{eq:ss}

724: \eeq which for $\phi_S=1$ is a fit to Schaerer's (2002) results,

725: accurate to within about 5\% for $60 M_\odot \lesssim m_*\lesssim 300

726: M_\odot$. Then the photoevaporation rate becomes 

727: \beq \dot{m}_{\rm evap} = 1.70\times 10^{-4}\phi_S^{1/2}(1+f_d)^{1/2} \left(\frac{T_{i,4}}{2.5}\right)^{0.4} \mst^{5/4}~\smyr.

728: \label{eq:evap2}

729: \eeq

730: The accretion rate onto the star-disk system is given by equation

731: (\ref{eq:mdot}). Equating this with equation (\ref{eq:evap2}), we find

732: that the resulting maximum stellar mass is

733: \beq

734: \label{eq:maxevap}

735: {\rm Max}\; m_{*f,2}=

736: 6.3\;\frac{\esd^{28/47}\esdb^{12/47}

737: K'^{60/47}}{\phi_S^{14/47} (1+f_d)^{26/47}} 

738:         \left(\frac{2.5}{T_{i,4}}\right)^{0.24} \rightarrow 1.45,

739: \eeq

740: where the $\rightarrow$ assumes fiducial values $\esd=0.2$,

741: $\esdb=0.25$, $K'=1$, $\phi_S=1$, $f_d=1/3$, and $T_{i,4}=2.5$ (see MT08

742: for details; note also here in eq.~\ref{eq:maxevap} we have corrected

743: a sign error in the index for $\phi_S$). This analytic estimate

744: therefore also suggests that for the fiducial case ($K'=1$) the mass

745: of a Pop III.1 star should be $\simeq 140 M_\odot$.

746: 

747: The uncertainties in these mass estimates include: (1) the assumption

748: that the gas distribution far from the star is approximately spherical

749: --- in reality it is likely to be flattened towards the equatorial

750: plane, thus increasing the fraction of gas that is shadowed by the

751: disk and raising the final protostellar mass; (2) uncertainties in the

752: disk photoevaporation mass loss rate due to corrections to the

753: Hollenbach et al. (1994) rate from the flow starting inside $r_g$ and

754: from radiation pressure corrections; (3) uncertainties in the

755: \ion{H}{2} region breakout mass due to hydrodynamic instabilities and

756: 3D geometry effects; (4) uncertainties in the accretion rate at late

757: times, where self-similarity may break down (Bromm \& Loeb 2004); (5)

758: the simplified condition, $\dot{m}_{\rm evap}>\mdsd$, used to mark the

759: end of accretion; (6) the possible effect of protostellar outflows

760: (discussed above); (7) the neglect of WIMP annihilation heating

761: (discussed above) and (8) the effect of rotation on protostellar

762: models, which will lead to cooler equatorial surface temperatures and

763: thus a reduced ionizing flux in the direction of the disk.

764: 

765: Here we discuss briefly the last of these effects. Using the results

766: of Ekstr\"om et al. (2008) and Georgi et al. (these proceedings), we

767: estimate that for a zero age main sequence protostar with

768: $\Omega/\Omega_{\rm crit}=0.99$ (i.e. rotating very close to

769: break-up), the surface temperature at an angle $80^\circ$ from the pole

770: (i.e. the direction relevant for the accretion disks modeled by MT08)

771: the surface temperature is reduced by a factor of 0.7. For

772: $m_*=140\sm$ this would cause $T_{\rm eff,*}$ to be reduced from

773: $1.0\times10^5\:{\rm K}$ to $7\times 10^4\:{\rm K}$ causing a

774: reduction in the ionizing flux (and thus also $\phi_S$) by a factor of

775: about 3. From eq.~\ref{eq:maxevap} we see that the mass of Pop III.1

776: star formation would be increased by about a factor of 1.4, to

777: $200\sm$ in the fiducial case.

778: 

779: 

780: 

781: 

782: \begin{table}

783: \begin{center}

784: \caption{Mass Scales of Population III.1 Protostellar Feedback}

785: \label{tab:m}

786: {\scriptsize

787: \begin{tabular}{cccccc}

788: \hline

789: $K^\prime$ & $f_{\rm Kep}$ & $T_{i,4}$ & $m_{\rm *,pb}$ ($\sm$)$^1$ & $m_{\rm *,eb}$ ($\sm$)$^2$ & $m_{\rm *,evap}$ ($\sm$)$^3$\\

790: \hline

791: 1 & 0.5 & 2.5 & 45.3 & 50.4 & 137$^4$\\

792: \hline

793: 1 & 0.75 & 2.5 & 37 & 41 & 137\\

794: 1 & 0.25 & 2.5 & 68 & 81 & 143\\

795: 1 & 0.125 & 2.5 & 106 & 170 & 173\\

796: 1 & 0.0626 & 2.5 & 182 & 330$^5$ & 256\\

797: \hline

798: 1 & 0.5 & 5.0 & 35 & 38 & 120\\

799: 1 & 0.25 & 5.0 & 53.0 & 61 & 125\\

800: \hline

801: 0.5 & 0.5 & 2.5 & 23.0 & 24.5 & 57\\

802: \hline

803: 2.0 & 0.5 & 2.5 & 85 & 87 & 321\\

804: \hline

805: \end{tabular}

806: }

807: \end{center}

808: \vspace{1mm}

809: \scriptsize{

810: {\it Notes:}\\

811: $^1$Mass scale of HII region polar breakout.\\

812: $^2$Mass scale of HII region near-equatorial breakout.\\

813: $^3$Mass scale of disk photoevaporation limited accretion.\\

814: $^4$Fiducial model.\\

815: $^5$This mass is greater than $m_{\rm *,evap}$ in this case because it is calculated without allowing for a reduction in $\dot{m}_*$ during the evolution due to polar HII region breakout (see MT08).

816: }

817: \end{table}

818: 

819: 

820: \section{Evolution of the Pop III.1 IMF}

821: 

822: %Population III.1 stars, i.e. those whose formation is unaffected by

823: %external astrophysical sources, are likely to be relatively rare

824: %objects. 

825: 

826: As the universe evolves and forms more and more structure, regions of

827: Pop III.1 star formation will become ever rarer. Indeed, because the

828: effects of radiation from previous stellar generations can propagate

829: relatively freely compared to the spreading and mixing of their metals

830: in supernovae, most metal-free star formation may be via Pop III.2

831: (Greif \& Bromm 2006). Nevertheless, understanding Pop III.1 star

832: formation is necessary as it establishes the initial conditions of

833: what follows.

834: 

835: O'Shea \& Norman (2007) studied the properties of Pop III.1

836: pre-stellar cores as a function of redshift. They found that cores at

837: higher redshift are hotter in their outer regions, have higher free

838: electron fractions and so form larger amounts of $\rm H_2$ (via $\rm

839: H^-$), although these are always small fractions of the total mass. As

840: the centers of the cores contract above the critical density of

841: $10^4\:{\rm cm^{-3}}$, those with higher $\rm H_2$ fractions are able

842: to cool more effectively and thus maintain lower temperatures to the

843: point of protostar formation. The protostar thus accretes from

844: lower-temperature gas and the accretion rates, proportional to $c_s^3

845: \propto T^{3/2}$, are smaller. Measuring infall rates at the time of

846: protostar formation at the scale of $M=100\sm$, O'Shea \& Norman find

847: accretion rates of $\sim 10^{-4}\smyr$ at $z=30$, rising to $\sim

848: 2\times 10^{-2}\smyr$ at $z=20$. If Hunter's (1977) solution applies,

849: the mass accretion rates to the protostar will be higher by a factor

850: of 3.7 by the time $\msd=100\sm$. These accretion rates then correspond to $K'$=0.37 ($z=30$) to

851: 4.3 ($z=20$). A naive application of eq.~\ref{eq:maxevap} would imply

852: a range of masses of $40\sm$ to $900\sm$. This suggests that the very

853: first Pop III.1 stars were relatively low-mass massive stars,

854: e.g. below the mass required for pair instability supernovae

855: ($140-260\;M_\odot$ in the models of Heger \& Woosley 2002). Such

856: stars would have had relatively little influence on their cosmological

857: surroundings, thus allowing Pop III.1 star formation to continue to

858: lower redshifts. It is not yet clear from simulations when Pop III.1

859: star formation was finally replaced by other types, since this depends

860: on the early IMFs of Pop III.1, III.2 and II stars. This transition

861: presumably occurred before reionization was complete.

862: 

863: 

864: \acknowledgements

865: We thank the organizers of IAU255 for a very stimulating meeting. The research of JCT is supported by NSF CAREER grant AST-0645412.

866: 

867: 

868: 

869: \begin{thebibliography}{}

870: 

871: \bibitem[]{abn2002}

872: Abel, T., Bryan, G. L., \& Norman, M. L. 2002, \emph{Science}, 295, 93

873: 

874: \bibitem[]{adams1989}

875: Adams, F. C., Ruden, S. P, \& Shu, F. H. 1989, \apj, 347, 959

876: 

877: \bibitem[]{beers2005}

878: Beers, T. C., \& Christlieb, N. 2005, \emph{ARA\&A}, 43, 531

879: 

880: \bibitem[]{blackman02}

881: Blackman, E. G., \& Field, G. B. 2002, Phys. Rev. Lett., 89, 265007

882: 

883: \bibitem[]{blandford82}

884: Blandford R. D., \& Payne D. G. 1982, MNRAS, 199, 883

885: 

886: \bibitem[]{bcl2002}

887: Bromm, V., Coppi, P. S., \& Larson, R. B. 2002, \emph{ApJ}, 564, 23

888: 

889: \bibitem[]{bromm2002}

890: Bromm, V., \& Loeb, A. 2002, \emph{ApJ}, 575, 111--116

891: 

892: \bibitem[]{bromm2004}

893: Bromm, V., \& Loeb, A. 2004, \emph{New Astron.}, 9, 353

894: 

895: \bibitem[]{clark2008}

896: Clark, P. C., Glover, S. C. O., \& Klessen, R. S. 2008, \emph{ApJ}, 672, 757

897: 

898: %\bibitem[]{fan2003}

899: %Fan, X., Strauss, M. A., Schneider, D. P. et al. 2003, \emph{AJ}, 125, 1649

900: 

901: \bibitem[]{ekstrom08}

902: Ekstr\"om, S., Meynet, G., Chiappini, C., Hirschi, R., \& Maeder, A. 2008, A\&A, in press (arXiv:0807.0573)

903: 

904: %\bibitem[]{fernandez2006}

905: %Fernandez, E. R., \& Komatsu, E. 2006, \emph{ApJ}, 646, 703

906: 

907: \bibitem[]{freese08}

908: Freese, K., Bodelheimer, P., Spolyar, D. \& Gondolo, P. 2008, arXiv: 0806.0617

909: 

910: \bibitem[]{gammie2001}

911: Gammie, C. F. 2001, \apj, 553, 174

912: 

913: \bibitem[]{greif2006}

914: Greif T. H., \& Bromm, V. 2006, \emph{MNRAS}, 373, 128

915: 

916: \bibitem[]{heger2002}

917: Heger, A., \& Woosley, S. E. 2002, \apj, 567, 532

918: 

919: \bibitem[]{hollenbach1994}

920: Hollenbach, D., Johnstone, D., Lizano, S., Shu, F. 1994, \apj, 428, 654

921: 

922: \bibitem[]{hunter1977}

923: Hunter, C. 1977, \emph{ApJ}, 218, 834

924: 

925: 

926: \bibitem[]{kashlinsky2004}

927: Kashlinsky, A. Arendt, R., Gardner, J. P. et al. 2004, \emph{ApJ}, 608, 1

928: 

929: \bibitem[]{krumholz07}

930: Krumholz, M. R., \& Thompson, T. A. 2007, \apj, 661, 1034

931: 

932: \bibitem[]{matzner2000}

933: Matzner, C. D. \& McKee, C. F. 2000, \apj, 545, 364

934: 

935: \bibitem[]{mckee2008}

936: McKee, C. F. \& Tan, J. C. 2008, \apj, 681, 771 (MT08)

937: 

938: \bibitem[]{morales2004}

939: Morales, M. F. \& Hewitt, J. 2004, \emph{ApJ}, 615, 7

940: 

941: \bibitem[]{}

942: Natarajan, A., Tan, J. C., \& O'Shea, B. W. 2008, \apj, submitted (arXiv:0807.3769)

943: 

944: \bibitem[]{norman2004}

945: Norman, M. L., O'Shea, B. W., \&  Paschos, P. 2004, \emph{ApJ}, 601, L115

946: 

947: \bibitem[]{omukai1998}

948: Omukai, K., \& Nishi, R. 1998, \emph{ApJ}, 508, 141

949: 

950: \bibitem[]{o'shea2007}

951: O'Shea, B. W., \& Norman, M. L. 2007, \apj, 654, 66

952: 

953: \bibitem[]{page2007}

954: Page, L., Hinshaw, G., Komatsu, E. et al. 2007, \emph{ApJS}, 170, 335

955: 

956: 

957: \bibitem[]{ripamonti2002}

958: Ripamonti, E., Haardt, F., Ferrara, A., \& Colpi, M. \emph{MNRAS}, 334, 401

959: 

960: \bibitem[]{santos2002}

961: Santos, M. R., Bromm, V., \& Kamionkowski, M. 2002, \emph{MNRAS}, 336, 1082

962: 

963: \bibitem[]{schaerer2002}

964: Schaerer, D. 2002, \emph{A\&A}, 382, 28

965: 

966: \bibitem[]{schaye2003}

967: Schaye, J., Aguirre, A., Kim, T-S. et al. 2003, \emph{ApJ}, 596, 768

968: 

969: 

970: \bibitem[]{shu1977}

971: Shu, F. H. 1977, \apj, 214, 488

972: 

973: \bibitem[]{shu1990}

974: Shu, F. H., Tremaine, S., Adams, F. C., \& Ruden, S. P. 1990, \apj, 358, 495

975: 

976: \bibitem[Spolyar et al. 2008]{sfg1}

977: Spolyar, D., Freese, K.,  \&  Gondolo, P., 2008, Physical Review Letters, 100, 051101

978: 

979: %\bibitem[]{stark2007}

980: %Stark, D. P., Ellis, R. S., Richard, J. et al. 2007, \emph{ApJ}, 663, 10

981: 

982: 

983: \bibitem[]{tan2004b}

984: Tan, J. C., \& Blackman, E. G. 2004, \apj, 603, 401 (TB04)

985: 

986: 

987: \bibitem[]{tan2004}

988: Tan, J. C., \& McKee, C. F. 2004, \emph{ApJ}, 603, 383 (TM04)

989: 

990: \bibitem[]{tan2008}

991: Tan, J. C., \& McKee, C. F. 2008, First Stars III, eds. O'Shea et al., AIP Conf. Proc., 990, p47

992: 

993: \bibitem[]{tegmark1997}

994: Tegmark, M., Silk, J., Rees, M.J., Blanchard, A., Abel, T., \& Palla, F. 1997, \apj, 474, 1

995: 

996: \bibitem[]{thompson2007}

997: Thompson, R. I., Eisenstein, D., Fan, X. et al. 2007, \emph{ApJ}, 657, 669

998: 

999: \bibitem[]{xu08}

1000: Xu, H., O'Shea, B. W., Collins, D. C., et al. 2008, \apj, submitted (arXiv:0807.2647)

1001: 

1002: \bibitem[]{weinmann2005}

1003: Weinmann, S. M., \&  Lilly, S. J. 2005, \emph{ApJ}, 624, 526

1004: 

1005: %\bibitem[]{willott2003}

1006: %Willott, C. J., McLure, R. J., \& Jarvis, M. J. 2003, \emph{ApJ}, 587, L15

1007: 

1008: 

1009: 

1010: 

1011: 

1012: 

1013: %\bibitem[]{giroux1996}

1014: %M. L. Giroux, \& P. R. Shapiro, \emph{ApJS}, \textbf{102}, 191--238 (1996).

1015: 

1016: %\bibitem[]{shapiro2004}

1017: %P. R. Shapiro, I. T. Iliev, \& A. C. Raga, \emph{MNRAS}, \textbf{348}, 753--782 (2004).

1018: 

1019: %\bibitem[]{omukai2005}

1020: %K. Omukai, T. Tsuribe, R. Schneider, and A. Ferrara, \emph{ApJ}, \textbf{626}, 627--643 (2005)

1021: 

1022: %\bibitem[]{bromm2003}

1023: %V. Bromm, and A. Loeb, \emph{Nature}, \textbf{425}, 812--814 (2003).

1024: 

1025: %\bibitem[]{uehara2000}

1026: %H. Uehara, and S-I. Inutsuka, \emph{ApJ}, \textbf{531}, L91--94 (2000).

1027: 

1028: 

1029: 

1030: 

1031: 

1032: 

1033: 

1034: 

1035: 

1036: 

1037: 

1038: 

1039: %\bibitem[]{yoshida2006}

1040: %N. Yoshida, K. Omukai, L. Hernquist, and T. Abel, \apj, \textbf{652}, 6--25 (2006).

1041: 

1042: 

1043: 

1044: %\bibitem[]{spolyar2007}

1045: %D. Spolyar, K. Freese, P. Gondolo, astro-ph/0705.0521 (2007).

1046: 

1047: %\bibitem[]{stahler1986}

1048: %S. W. Stahler, F. Palla, \& E. E. Salpeter, \apj, \textbf{302}, 590--605 (1986). 

1049: 

1050: %\bibitem[]{omukai2003}

1051: %K. Omukai, and F. Palla, \apj, \textbf{589}, 677--687 (2003).

1052: 

1053: 

1054: 

1055: 

1056: 

1057: 

1058: 

1059: %\bibitem[]{balbus1998}

1060: %S. A. Balbus, and J. F. Hawley, Rev. Mod. Phys., \textbf{70}, 1--53 (1998).

1061: 

1062: %\bibitem[]{kulsrud1997}

1063: %R. M. Kulsrud, R. Cen, J. P. Ostriker, \& D. Ryu, \apj, \textbf{480}, 481--491 (1997).

1064: 

1065: 

1066: 

1067: %\bibitem[]{shakura1973}

1068: %N. I. Shakura, and R. A. Sunyaev, A\&A, \textbf{24}, 337--355 (1973).

1069: 

1070: %\bibitem[]{frank1995}

1071: %J. Frank, A. King, \& D. Raine, Accretion Power in Astrophysics (Cambridge: Cambridge Univ. Press) (1995).

1072: 

1073: %\bibitem[]{artemova1996}

1074: %I. V. Artemova, G. S. Bisnovatyi-Kogan, G. Bjoernsson, \& I. D. Novikov, \apj, \textbf{456}, 119--123 (1996).

1075: 

1076: %\bibitem[]{shu1992}

1077: %F. H. Shu, The Physics of Astrophysics Vol II: Gas Dynamics (Mill Valley: University Science Books) (1992).

1078: 

1079: %\bibitem[]{iglesias1996}

1080: %C. A. Iglesias, and F. J. Rogers, \apj, \textbf{464}, 943--953 (1996).

1081: 

1082: 

1083: 

1084: 

1085: 

1086: 

1087: %\bibitem[]{matzner1999}

1088: %C. D. Matzner, and C. F. McKee, \apj, \textbf{526}, L109--L112 (1999).

1089: 

1090: %\bibitem[]{machida2006}

1091: %M. N. Machida, K. Omukai, T. Matsumoto, S-I. Inutsuka, \apj, \textbf{647}, L1--L4 (2006).

1092: 

1093: %\bibitem[]{tan2004c}

1094: %J. C. Tan, and C. F. McKee, in IAU Symp. 221, Star Formation at High Angular Resolution, ed. M. Burton, R. Jayawardhana, \& T. Bourke (San Francisco: ASP), astro-ph/0309139 (2004).

1095: 

1096: 

1097: %\bibitem[]{whalen2008}

1098: %D. Whalen, B. W. O'Shea, J. Smidt, M. L. Norman, \apj, submitted, astro-ph/0708.1603 (2008).

1099: 

1100: %\bibitem[]{glover2001}

1101: %S. C. O. Glover, and P. W. J. L. Brand, MNRAS, \textbf{321}, 385--397 (2001).

1102: 

1103: %\bibitem[]{susa2007}

1104: %H. Susa, \apj, \textbf{659}, 908--917 (2007).

1105: 

1106: %\bibitem[]{ahn2007}

1107: %K. Ahn, and P. R. Shapiro, MNRAS, \textbf{375}, 881--908 (2007).

1108: 

1109: %\bibitem[]{kudritzki2002}

1110: %R. Kudritzki, \apj, \textbf{577}, 389--408 (2002).

1111: 

1112: %\bibitem[]{meynet2006}

1113: %G. Meynet, S. Ekstr\"om, A. Maeder, A\&A \textbf{447}, 623--639 (2006).

1114: 

1115: 

1116: %\bibitem[]{tumlinson2004}

1117: %J. Tumlinson, A. Venkatesan, and J. M. Shull, \apj, \textbf{612}, 602--614 (2004).

1118: 

1119: %\bibitem[]{scannapieco2006}

1120: %E. Scannapieco, D. Kawata, C. B. Brook, R. Schneider, A. Ferrara,

1121: %B. K. Gibson, \apj, \textbf{653}, 285--299 (2006).

1122: 

1123: %\bibitem[]{stacy2007}

1124: %A. Stacy, and V. Bromm, MNRAS, in press, astro-ph/0705.3634, (2007).

1125: 

1126: %\bibitem[]{jappsen2007}

1127: %A. K. Jappsen, R.~S. Klessen, S. C. O. Glover, \& M.-M. Mac Low, astro-ph/0709.3530 (2007).

1128: 

1129: %\bibitem[]{mueller2002}

1130: %K. E. Mueller, Y. L. Shirley, N. J. Evans, \& H. R. Jacobson, ApJS, \textbf{143}, 469--497 (2002).

1131: 

1132: %\bibitem[]{figer2005}

1133: %D. F. Figer, Nature, \textbf{434}, 192--194 (2005).

1134: 

1135: %\bibitem[]{larson1971}

1136: %R. B. Larson, and S. Starrfield, A\&A, \textbf{13}, 190--197 (1971).

1137: 

1138: %\bibitem[]{yorke2002}

1139: %H. W. Yorke, and C. Sonnhalter, \apj, \textbf{569}, 846--862 (2002).

1140: 

1141: %\bibitem[]{krumholz2005a}

1142: %M. R. Krumholz, C. F. McKee, R. I. Klein, in {\it Massive Star Birth:

1143: %A Crossroads of Astrophysics}, ed. R. Cesaroni, M. Felli,

1144: %E. Churchwell, \& M. Walmsley (Cambridge: Cambridge University Press),

1145: %p. 231 (2005).

1146: 

1147: %\bibitem[]{krumholz2005b}

1148: %M. R. Krumholz, C. F. McKee, R. I. Klein, \apj, \textbf{618}, L33-36 (2005).

1149: 

1150: %\bibitem[]{mckee2003}

1151: %C. F. McKee, and J. C. Tan, \apj, \textbf{585}, 850--871 (2003).

1152: 

1153: 

1154: \end{thebibliography}

1155: 

1156: 

1157: 

1158: \end{document}

1159: