1: % !iTeXMac(typeset): altpdflatex --keep-psfile ${iTMInput}

2: % !iTeXMac(compile): "./local Command"

3: %\documentclass[12pt,preprint]{aastex}

4: \documentclass{emulateapj}

5: \usepackage{natbib}

6: \usepackage{apjfonts}

7: \usepackage{epsfig}

8: 

9: \newcommand{\tq}{$t_{Q}$}

10: \newcommand{\tQ}{t_{Q}}

11: \newcommand{\etal}{et al.}

12: \newcommand{\NH}{$N_{H}$}

13: \newcommand{\Mdot}{\dot{M}}

14: \newcommand{\Lbol}{L_{\rm bol}}

15: \newcommand{\dEdt}{\epsilon\Mdot c^{2}}

16: %\newcommand{\LB}{\nu_{B} L_{\nu_{B}}}

17: \newcommand{\LB}{L_{B}}

18: \newcommand{\LBo}{L_{B,obs}}

19: \newcommand{\LBm}{L_{B,min}}

20: \newcommand{\Lcut}[1]{10^{#1}\,L_{\sun}}

21: \def\cm{{\rm\thinspace cm}}

22: \def\erg{{\rm\thinspace erg}}

23: \def\eV{{\rm\thinspace eV}}

24: \def\g{{\rm\thinspace g}}

25: \def\kg{{\rm\thinspace kg}}

26: \def\G{{\rm\thinspace G}}

27: \def\ga{{\rm\thinspace gauss}}

28: \def\Jy{{\rm\thinspace Jy}}

29: \def\K{{\rm\thinspace K}}

30: \def\keV{{\rm\thinspace keV}}

31: \def\MeV{{\rm\thinspace MeV}}

32: \def\km{{\rm\thinspace km}}

33: \def\kpc{{\rm\thinspace kpc}}

34: \def\Lsun{\hbox{$\rm\thinspace L_{\odot}$}}

35: \def\m{{\rm\thinspace m}}

36: \def\mJy{{\rm\thinspace mJy}}

37: \def\Mpc{{\rm\thinspace Mpc}}

38: \def\Msun{\hbox{$\rm\thinspace M_{\odot}$}}

39: \def\Rsun{\hbox{$\rm\thinspace R_{\odot}$}}

40: \def\pc{{\rm\thinspace pc}}

41: \def\ph{{\rm\thinspace ph}}

42: \def\s{{\rm\thinspace s}}

43: \def\yr{{\rm\thinspace yr}}

44: \def\Zsun{\hbox{$\rm\thinspace Z_{\odot}$}}

45: \def\Lstar{\hbox{$\rm\thinspace L^{\star}$}}

46: \def\sr{{\rm\thinspace sr}}

47: \def\Hz{{\rm\thinspace Hz}}

48: 

49: \def\ltsima{$\; \buildrel < \over \sim \;$}

50: \def\simlt{\lower.5ex\hbox{\ltsima}}

51: \def\gtsima{$\; \buildrel > \over \sim \;$}

52: \def\simgt{\lower.5ex\hbox{\gtsima}}

53: \def\Mdot{\hbox{$\dot M$}}

54: \def\mdot{\hbox{$\dot m$}}

55: \def\<{\thinspace}

56: \def\ss{\s\ }           %four

57: \def\Mpc{{\rm\thinspace Mpc}}

58: \def\km{{\rm\thinspace km}}

59: \def\msun{{\rm M_{\odot}}}

60: \def\s{{\rm\thinspace s}}

61: \def\pcm{\hbox{$\cm^{-3}\,$}}

62: \def\pcmK{\hbox{$\cm^{-3}\K$}}

63: \def\phpcmsqps{\hbox{$\ph\cm^{-2}\s^{-1}\,$}}

64: \def\pHz{\hbox{$\Hz^{-1}\,$}}

65: \def\ps{\hbox{$\s^{-1}\,$}}

66: \def\psqcm{\hbox{$\cm^{-2}\,$}}

67: \def\psr{\hbox{$\sr^{-1}\,$}}

68: \def\pyr{\hbox{$\yr^{-1}\,$}}

69: \def\kmps{\hbox{$\km\s^{-1}\,$}}

70: \def\kmpspMpc{\hbox{$\kmps\Mpc^{-1}$}}

71: \def\erg{{\rm\thinspace erg}}

72: \def\ergps{\hbox{$\erg\s^{-1}\,$}}

73: \def\hmpc{\;h^{-1}{\rm Mpc}}

74: \def\hkpc{h^{-1}{\rm kpc}}

75: \def\kms{{\rm \;km\;s^{-1}}}        

76: 

77: 

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

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

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

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

82: 

83: 

84: \shorttitle{Direct cosmological simulations of the growth of black

85: holes and galaxies} \shortauthors{Di Matteo \etal}

86: \slugcomment{Submitted to ApJ 05/14/07}

87: \begin{document}

88: 

89: \title{Direct cosmological simulations of the growth of black holes

90: and galaxies} \author{ Tiziana Di Matteo,\altaffilmark{1} J\"org

91: Colberg,\altaffilmark{1} Volker Springel,\altaffilmark{2} Lars

92: Hernquist\altaffilmark{3} \&\ Debora Sijacki\altaffilmark{2} }

93: \altaffiltext{1} {Physics Department, Carnegie Mellon University, 5000

94: Forbes Avenue, Pittsburgh, PA 15213}

95: \altaffiltext{2}{Max-Planck-Institut f\"{u}r Astrophysik,

96: Karl-Schwarzchild-Stra\ss e 1, 85740 Garching bei M\"{u}nchen,

97: Germany} \altaffiltext{3}{Harvard-Smithsonian Center for Astrophysics,

98: 60 Garden Street, Cambridge, MA 02138, USA}

99: 

100: 

101: \begin{abstract}

102:   We investigate the coupled formation and evolution of galaxies and their

103:   embedded supermassive black holes using state-of-the-art hydrodynamic

104:   simulations of cosmological structure formation. For the first time, we

105:   self-consistently follow the dark matter dynamics, radiative gas cooling,

106:   star formation, as well as black hole growth and associated energy feedback

107:   processes, starting directly from initial conditions appropriate for the

108:   $\Lambda$CDM cosmology.  Our modeling of the black hole physics is based on

109:   an approach we have recently developed and tested in simulations of isolated

110:   galaxy mergers. Here we apply the same model in cosmological simulations to

111:   examine: (i) the predicted global history of black hole mass assembly in

112:   galaxies, (ii) the evolution of the local black hole-host mass correlations

113:   and (iii) the conditions that allow rapid growth of the first quasars,

114:   as well

115:   as the properties of their hosts and descendants today. We find that our

116:   simulations produce a total black hole mass density $\rho_{\rm BH}\simeq 2

117:   \times 10^{5} \Msun\Mpc^{-3}$ by $z=0$, 

118:   in good agreement with observational

119:   estimates. The black hole accretion rate density, $\dot{\rho}_{\rm BH}$,

120:   peaks at lower redshift and evolves more strongly at high redshift than the

121:   star formation rate density, $\dot{\rho}_{*}$, with an approximate scaling

122:   as $\dot{\rho}_{\rm BH}/\dot{\rho}_{*} \propto (1+z)^{-4}$ at $z\ge3$. On

123:   the other hand, the ratio $\rho_{\rm BH}/\rho_{*} \sim (1+z)^{-0.6}$ of

124:   black hole to stellar mass densities shows only a moderate evolution at low

125:   redshifts $z \lesssim 3$. For the population of galaxies identified in the

126:   simulations at $z=1$ we find strong correlations between black hole mass and

127:   velocity dispersion or mass of the stellar systems. The predicted

128:   correlations agree well with the measured local $M_{\rm BH}-\sigma$ and

129:   $M_{\rm BH} -M_{*}$ relationships, but also suggest a weak evolution with

130:   redshift in the normalization, and in particular the slope. However, the

131:   magnitude of this effect is sensitive to the range of masses being probed.

132:   For

133:   stellar masses of $M_{*}\ge 3\times 10^{10}$, we predict a trend of

134:   increasing $M_{*}/ M_{\rm BH}$ with redshift, in agreement with recent

135:   direct estimates of the BH to host stellar mass ratio at high redshift and

136:   the conjecture that a more fundamental relation (a BH fundamental plane)

137:   should involve both $M_{*}$ and $\sigma$.  We find that our simulation

138:   models can also produce quite massive black holes at high redshift, as a

139:   result of extended periods of exponential growth in relatively isolated,

140:   rare regions that collapse early and exhibit strong gas

141:   inflows. Interestingly, when followed to their descendants, these first

142:   supermassive BH systems are not necessarily the most massive ones today,

143:   since they are often overtaken in growth by quasars that form later.

144: \end{abstract}

145: 

146: \keywords{quasars: general --- galaxies: formation --- galaxies: active --- 

147: galaxies: evolution --- cosmology: theory --- hydrodynamics}

148: 

149: \section{Introduction\label{sec:intro}}

150: Following the discovery of quasars \citep{Schmidt1963, Greenstein1963} it was

151: suggested that supermassive black holes ($10^6-10^9$~\Msun) lie at the centers

152: of galaxies, and that the quasar activity is fueled by the release of

153: gravitational energy from their accreted matter. The remnants of quasar phases

154: at early times are probably the supermassive black holes found at the centers

155: of galaxies in our local Universe. Interestingly, the properties of these

156: supermassive black holes are tightly coupled to the mass \citep{Magorrian1998}

157: and velocity dispersion of their host galaxies, as manifested in the $M_{\rm

158:   BH}-\sigma$ relation of spheroids \citep{Ferrarese2000, Gebhardt2000}. In

159: addition, the black hole mass is correlated with the concentration or Sersic

160: index \citep{Graham2006}.  Most recently, \citet{Hopkins2007a} have shown 

161: that these various correlations are not independent, and can be

162: understood as projections of a ``black hole fundamental plane'' (BHFP),

163: similar to that describing properties of elliptical galaxies.

164: 

165: The existence of highly luminous quasars also

166: constrains the formation and evolution of massive galaxies and the epoch of

167: reionization. Quasars with inferred black hole masses in excess of

168: $10^9$~\Msun\ have now been discovered out to $z\sim 6$ \citep{Fan2003},

169: indicating an early formation time for black holes and galaxy spheroids and

170: posing a significant challenge for theoretical models of high-redshift quasar

171: and galaxy formation.

172: 

173: This growing observational evidence, drawn from local galaxies to high

174: redshift quasars, argues for a close connection between the formation and

175: evolution of galaxies and of their central supermassive black holes. However,

176: the physical nature of this relationship has yet to be understood in detail.

177: Indeed, there are significant gaps in our observational and theoretical

178: knowledge of the history of black hole formation and evolution in galaxies.

179: 

180: For example, current velocity dispersion measurements are inconclusive about

181: the important question whether the tight scaling relations evolve with

182: redshift \citep{Woo2006,Shields2006}, or are essentially invariant

183: \citep{Shields2003} as a function of time. We note that some

184: evolution in the ratio of black hole to halo mass

185: is suggested by clustering constraints \citep[e.g][]{Adelberger2005,

186: Lidz2006}, but these measurements do not directly address

187: the relationship between black hole and properties of the

188: luminous host galaxy.  More relevant are 

189: comparisons of the black hole mass inferred from quasar observations

190: to the host stellar mass, both observationally

191: \citep{Merloni2004} and theoretically \citep{Hopkins2006a},

192: which indicate an evolution in e.g. the Magorrian relation, in

193: the sense that black holes are more massive relative to luminous

194: spheroids at high redshifts than at $z=0$.  

195: 

196: Theoretical studies of the co-evolution of black holes and galaxies have so

197: far mostly used so-called semi-analytical modeling

198: \citep[e.g.][]{Kauffmann2000, Cattaneo1999, Wyithe2003a, Volonteri2003,

199: DiMatteo2003, Granato2004, Springel2005c, Cattaneo2005, Croton2006,

200: DeLucia2006, Malbon2007} of galaxy formation, in which the growth of

201: galaxies and their embedded black holes is followed with simple physical

202: parameterizations on top of dark matter merging history trees. Many of these

203: models assume that quasar activity is triggered by major galaxy mergers,

204: motivated by hydrodynamical simulations that have shown that gravitational

205: tidal fields during major mergers of gas rich galaxies produce strong gas

206: inflows \citep{BarnesHernquist1991, BarnesHernquist1996}, which lead to a

207: burst of nuclear star formation \citep{Mihos1996} and are likely the

208: prerequisite for rapid black hole growth and quasar activity.  Nearby quasars

209: are indeed preferentially found in tidally disturbed objects

210: \citep[e.g][]{Jogee2004}, corroborating the importance of galaxy interactions

211: and mergers for major black hole growth.

212: 

213: Many theoretical explanations for the observed correlations between galaxy

214: properties and black hole mass rely on some form of self-regulated growth of

215: the BHs. For example, it has been suggested that the central black holes grow

216: until they release sufficient energy to unbind the gas that feeds them from

217: their host galaxy \citep{CiottiOstriker1997, Silk1998, Fabian1999,

218: Wyithe2003a}. We recently explored such a local energy feedback for the first

219: time with self-consistent, fully three-dimensional hydrodynamic simulations of

220: galaxies \citep{DiMatteo2005,Springel2005a} that include a treatment of

221: accretion on supermassive black holes and their associated energy

222: feedback. These simulations have demonstrated that the fundamental BH-host

223: correlation including the $M_{\rm BH}-\sigma$ relation can indeed be

224: reproduced in feedback-regulated models of BH growth \citep{DiMatteo2005}, in

225: accordance with theoretical conjectures. At the same time, the dynamical

226: coupling in the simulations of hydrodynamical gas inflow, star formation,

227: black hole growth and associated feedback processes gives them substantial

228: predictive power well beyond that of simplified analytical and semi-analytical

229: models. Besides the $M_{\rm BH}-\sigma$ or $M_{\rm BH}-M_{*}$

230: \citep{DiMatteo2005, Robertson2006a} relationships, the simulation models can

231: for example predict the detailed properties of the spheroidal galaxies forming

232: in major mergers and how they correlate with the BH masses. In fact, they

233: suggest the existence of a fundamental plane relation for BHs ($M_{BH} \propto

234: \sigma^{3.0} R_e^{0.5}$ \citep{Hopkins2007a}, provide an explanation for the

235: red colors of massive elliptical galaxies \citep{Springel2005b}, and

236: describe the fundamental plane of elliptical galaxies \citep{Robertson2006b}.

237: They also suggest luminosity-dependent quasar lifetimes, leading to a new

238: interpretation for the origin of the quasar luminosity function and its

239: evolution over cosmic history~\citep{Hopkins2005, Hopkins2006a}.

240: 

241: In the present paper, we extend these earlier studies by carrying out fully

242: cosmological hydrodynamic simulations of the $\Lambda$CDM model that jointly

243: follow the growth of galaxies and supermassive black holes, as well as their

244: associated feedback processes. Our approach is based on the same methodology

245: that we have developed and applied in the high-resolution simulations of

246: galaxy mergers, augmented with a suitable mechanism to seed emerging new dark

247: matter halos with a small black hole that can then grow by gas accretion later

248: on. While much more restricted in numerical resolution than simulations of

249: individual galaxy mergers, our modeling of star formation and black hole

250: physics in terms of a sub-resolution treatment provides quite accurate results

251: already at comparatively coarse resolution, an important prerequisite for

252: attempting to model these processes in cosmological simulations. Nevertheless,

253: numerical resolution is clearly an important limitation of our cosmological

254: results, an aspect that we will discuss in more detail where appropriate. With

255: this caveat in mind, we would like to stress however that the unambiguous

256: initial conditions of direct cosmological simulation make them in principle

257: the most powerful and accurate tool for studying the interplay of galaxy

258: formation and black hole growth.  Our aim is therefore to examine how well our

259: current model for treating BH physics in simulations does in present

260: state-of-the-art hydrodynamical calculations of cosmic structure formation,

261: and what we can learn from them to advance our theoretical understanding of

262: the co-evolution of galaxies and supermassive black holes. In this study we

263: shall focus on basic properties of the black hole population, like the

264: evolution of the cosmic BH mass density, and the correlations between BH

265: masses and host galaxy properties. In \citet{Sijacki2007} we also study an

266: extension of our feedback model with a `radio mode' that is active at low

267: accretion rates and is distinct from the normal quasar activity, allowing us

268: to study the formation of AGN in rich galaxy clusters at low redshifts, where

269: it is likely to be important.

270: 

271: This paper is structured as follows.  In \S2 we describe our simulation set

272: and the numerical modeling adopted for the ISM, star formation and gas

273: accretion onto black holes. In \S3 we present our results for the evolution of

274: the global black hole mass density and compare it to the cosmic history of

275: star formation and the evolution of the stellar density.  In \S4 we examine our

276: results for the fundamental BH/host correlations measured from the simulation

277: and for their evolution from high to low redshift. Finally, we summarize and

278: discuss our findings in \S5.

279: 

280: \section{Methodology\label{sec:method}}

281: 

282: \subsection{Numerical code}

283: In this study we focus on a $\Lambda$CDM cosmological model with parameters

284: chosen according to the first year results from the Wilkinson Microwave

285: Anisotropy Probe \citep[WMAP1;][] {Spergel2003}, $\Omega_0 = 0.3$,

286: $\Omega_{\Lambda} = 0.7$, Hubble constant $H_0 = 100 h \kmpspMpc$ with $h

287: =0.7$ and a scale invariant primordial power spectrum with index $n=1$, with a

288: normalization of the amplitude of fluctuations $\sigma_{8} = 0.9$.

289: \footnote{The largest simulation presented here had already been started by

290:   the time the updated third year constraints have become available

291:   \citep[WMAP3;][]{Spergel2006}. We comment on effects on the growth of the

292:   halo mass function owing to the lower amplitude of fluctuations, $\sigma_{8}$

293:   implied by WMAP3 in ~\citet{Li2007, Sijacki2007}}. We use a significantly

294:   extended version of the parallel cosmological TreePM-SPH code {\small

295:   GADGET2}~\citep{Springel2005d} to evolve a realization of $\Lambda$CDM

296:   initial conditions from high to low redshift. The combination of a

297:   high-resolution gravitational solver with individual and adaptive timesteps

298:   allows this code to bridge a large dynamic range both in length- and

299:   timescales.  Gas dynamics is followed with the Lagrangian smoothed particle

300:   hydrodynamics (SPH)~\citep[e.g] []{Monaghan1992} technique, which we employ

301:   in a formulation that manifestly conserves energy and entropy, despite the

302:   use of fully adaptive SPH smoothing lengths~\citep{Springel2002}.  Radiative

303:   cooling and heating processes are computed as in \citet{Katz1996}, with a

304:   spatially uniform photoionizing UV background that is imposed externally.

305: 

306: Within cosmological (or galaxy-sized) numerical simulations, it is presently

307: (and for some time to come) not feasible to follow the physics of star

308: formation and black hole accretion from first principles down to scales of

309: individual stars or black holes.  Any numerical model of galaxy formation

310: therefore needs to make substantial approximations for some of the relevant

311: physics on unresolved scales.

312: 

313: 

314: For modeling star formation and its associated supernova feedback we use the

315: sub-resolution multiphase model for the interstellar medium developed by

316: \citet{Springel2003a}. In this model, a thermal instability is assumed to

317: operate above a critical density threshold $\rho_{\rm th}$, producing a two

318: phase medium consisting of cold clouds embedded in a tenuous gas at pressure

319: equilibrium. Stars form from the cold clouds, and short-lived stars supply an

320: energy of $10^{51}\,{\rm ergs}$ to the surrounding gas as supernovae. This

321: energy heats the diffuse phase of the ISM and evaporates cold clouds, thereby

322: establishing a self-regulation cycle for star formation.  $\rho_{\rm th}$ is

323: determined self-consistently in the model by requiring that the equation of

324: state (EOS) is continuous at the onset of star formation. The cloud

325: evaporation process and the cooling function of the gas then determine the

326: temperatures and the mass fractions of the two hot and cold phases of the

327: ISM, such that the EOS of the model can be directly computed as a function of

328: density. The latter is encapsulating the self-regulated nature of star

329: formation owing to supernovae feedback in a simple model for a multiphase ISM.

330: As in the \citet{Springel2003a} model we have included a model

331: for supernova-driven galactic winds with an initial wind speed of $v \sim 480

332: \kmps$.

333: 

334: For the parameter settings adopted here, the model reproduces the observed

335: star formation rate surface densities in isolated spiral

336: galaxies~\citep{Kennicutt1989, Kennicutt1998}. Using a large number of nested

337: cosmological simulations, the approach we adopt (and the parameters we use)

338: has also been shown to lead to a numerically converged estimate

339: for the cosmic star formation history of the universe that agrees reasonably

340: well with low redshift observations \citep{Springel2003b,Hernquist2003}.

341: For the modeling

342: of BH accretion and feedback we adopt a similar strategy as for star

343: formation, which we discuss next.

344: 

345: \subsection{Accretion and Feedback from Supermassive black holes}

346: 

347: 

348: \begin{figure*}

349: \begin{center}

350: \resizebox{7.6cm}{!}{\includegraphics{f1a.ps}} %

351: \resizebox{7.6cm}{!}{\includegraphics{f1b.ps}}\\

352: \resizebox{7.6cm}{!}{\includegraphics{f1c.ps}} %

353: \resizebox{7.6cm}{!}{\includegraphics{f1d.ps}}\\

354: \resizebox{7.6cm}{!}{\includegraphics{f1e.ps}} %

355: \resizebox{7.6cm}{!}{\includegraphics{f1f.ps}}

356: \end{center}

357: \caption{Projected baryonic density field in slices of thickness

358: $5000\,h^{-1}\kpc$ through our high resolution simulation, color-coded

359: by temperature and with brightness proportional to the logarithm of

360: the gas density. Each panel shows the same region of space at

361: different redshifts, as labeled.  The circles mark the positions of

362: the black holes, with a size that encodes the BH mass, as indicated in

363: the top left panel.}

364: \label{fig:D6slices}

365: \end{figure*}

366: 

367: As for star formation, we adopt a sub-resolution model to capture the main

368: features of accretion and associated feedback on supermassive black holes

369: \citep[as introduced in][]{Springel2005a, DiMatteo2005}.  To this end, we

370: represent black holes by collisionless `sink' particles that can grow in mass

371: by accreting gas from their immediate environments, or by merging with other

372: black holes. We estimate the gas accretion rate onto a black hole using a

373: Bondi-Hoyle-Lyttleton parameterization \citep{Bondi1952, BondiHoyle1944,

374: Hoyle1939}. In this description, the accretion rate onto the black hole is

375: given by $\dot{M}_{\rm B} \, = \, {{4\pi \, \, [G^2 M_{\rm BH}^2 \, \rho}] /

376: {(c_s^2 + v^2)^{3/2}}} \, $ where $\rho$, $c_s$ are the density and sound

377: speed of the ISM gas, respectively, and $v$ is the velocity of the black hole

378: relative to the surrounding gas. In this study we also allow the black hole to

379: accrete at mildly super-Eddington values and impose a maximum allowed

380: accretion rate equal to $2x$ Eddington rate, $\dot{M}_{\rm Edd}$. As we shall

381: see such conditions are generally achieved at high redshift.  We note that the

382: detailed relativistic accretion flow onto the black hole is {\em unresolved}

383: in our simulations, but if the limitating factor for rapid growth of BHs lies

384: in the larger-scale gas distribution around the black hole, {\em which is

385: resolved}, then the Bondi prescription should capture the dependence of the

386: mean accretion rate onto the conditions of the gas in the region around the

387: black hole.

388: 

389: The radiated luminosity, $L_{\rm r}$ is related to the gas accretion rate,

390: $\dot {M}_{\rm BH}$, by the radiative efficiency $ \eta \, = \, {{L_{\rm

391: r}}/({\dot {M}_{\rm BH} \, c^2}}) \,$, which simply gives the mass to energy

392: conversion efficiency, set by the amount of energy that can be extracted from

393: the innermost stable orbit of an accretion disk around a black hole. We will

394: adopt a fixed value of $\epsilon_{\rm r} =0.1$, which is the mean value for a

395: radiatively efficient \citep{Shakura1973} accretion disk onto a Schwarzschild

396: black hole.

397: 

398: We assume that some small fraction $\epsilon_{\rm f}$ of the radiated

399: luminosity $L_{\rm r}$ can couple to the surrounding gas in the form of

400: feedback energy, viz. $ \dot{E}_{\rm feed} \, = \, \epsilon_{\rm f} \, L_{\rm

401:   r}$

402: % \,\, = \, \epsilon_{\rm f} \, 0.1 \, \dot{M}_{\rm BH} \,

403: %c^2 \, $. 

404: In accordance with our previous studies of galaxy merger simulations we take

405: $\epsilon_{\rm f} = 5 \% $.  This value governs the normalization of the

406: $M_{\rm BH} - \sigma$ relation, and $\epsilon_{\rm f} = 5 \% $ brings it into

407: agreement with current observations \citep{DiMatteo2005}. We note that

408: $\epsilon_{\rm f}$ is effectively the only free parameter in our black hole

409: model. After fixing it to reproduce the normalization of the observed $M_{\rm

410:   BH} - \sigma$ from the galaxy models we do not vary it in any of our

411: cosmological simulations.

412: 

413: For simplicity, we deposit the feedback energy isotropically in the region

414: around the black hole.  Lack of spatial resolution precludes us from

415: considering mechanical modes of releasing the energy, e.g.~in the form of a

416: jet. However, we note that is is plausible that such other forms of energy are

417: thermalized eventually as well, and that the final impact of the feedback

418: depends primarily on the total amount of energy released and less on the form

419: it is released in.  This is likely to be true generally, provided that

420: the energy (or momentum) is imparted to the surrounding gas on length

421: scales small and time scales short compared with those that characterize

422: the host galaxy.  In that event, the impact of black hole feedback will

423: be explosive in nature and, indeed, the blowout phase of evolution in

424: our simulations is well-described by a generalized Sedov-Taylor

425: blast-wave solution \citep{Hopkins2006b,HH2006}.

426: In any case, we emphasize that despite an isotropic release of the energy,

427: the response of the gas can still be decidedly anisotropic, e.g.~when a dense

428: gas disk is present that channels the gas response into a collimated outflow.

429: 

430: The idea that we follow with our feedback modeling here is the rapid

431: accretion phases of BHs at times close to their critical growth phases. Such

432: `quasar' phases are typically relatively short-lived and require galaxy

433: mergers to produce the strong gravitational tidal forcing necessary for

434: sufficient nuclear gas inflow rates. It is presently unclear whether the

435: accretion disks in such modes actually produce mechanical jets of release

436: their feedback in another way. It is however evident that at low redshift

437: (e.g.; $z < 1$ which we are not studying directly here) that when BHs that are

438: embedded in the hot gas atmospheres of large groups and clusters channel their

439: energy in mechanical feedback in the form of relativistic jets, generating

440: buoyantly rising radio bubble. In \citet{Sijacki2007} we consider this form

441: of feedback mode.

442: 

443: %This assumption is physically reasonable, and there is not likely

444: %to be a feasible alternative at present.

445: %Our understanding of the physics of accretion and how it couples to

446: %jets/outflows from supermassive black holes is sufficiently poor that it

447: %is not clear what range in spatial scales would be required to obtain

448: %a complete description of the impact of black hole growth and feedback

449: %on galactic scales. Even state-of-the-art 3-D numerical calculations

450: %of individual accretion disks with magneto-rotational instability and

451: %full-GR treatment \citep{reference} currently need to be carried out

452: %under quite restrictive assumptions (e.g., they are typically

453: %non-radiative).

454: 

455: An important remaining question in our model concerns the ultimate origin of

456: the BHs. Since our accretion rate estimate can only grow a BH that already

457: exists, we assume that a physical process that produces small seed BHs is

458: operating sufficiently efficiently that effectively all halos above a certain

459: threshold mass contain at least one such seed BH. Whether or not they can then

460: grow to larger masses by gas accretion will be determined by the local gas

461: conditions, as described in our model above. In order to achieve such a

462: seeding at a technical level in cosmological runs we use an on-the-fly

463: `friends-of-friends' group finder algorithm which is called at intervals

464: equally spaced in the logarithm of the scale-factor $a$, with $\Delta \log{a} =

465: \log{1.25}$. This provides the locations and mass of all halos in the

466: simulation. If a halo is more massive than our threshold and does not contain

467: any black hole yet, we endow it with one by converting its densest gas

468: particle into a sink particle with a seed black hole mass of $10^5\, h^{-1}

469: \Msun$.  The further growth of the black hole sink can then proceed by gas

470: accretion, at a rate that depends sensitively on the local conditions, or by

471: mergers with other black hole sink particles.  The total cumulative black hole

472: mass introduced in this way as seeds is negligible compared to the mass growth

473: by gas accretion. We note that being able to run a fast, parallel

474: `friends-of-friends' algorithm on the fly during simulations is an important

475: technical prerequisite of our technique.

476: 

477: Further motivation for this choice of seeding procedure is based on the

478: currently proposed scenarios for the seed black holes in galaxies. To grow a

479: supermassive black hole to a mass of $\sim 10^{9} \Msun$ in less than a

480: billion years, as required by presence of the $z=6$ SDSS quasars, may require

481: (1) the catastrophic collapse of a supermassive star that forms a large

482: initial black hole of mass $10^4-10^6\Msun$ \citep{Carr1984, Bromm2003,

483: Begelman2006}, or (2) alternatively, smaller black hole seeds ($M \sim 10^2

484: \Msun$) may form from the first PopIII stars at $z \sim 30$ and grow

485: exponentially from then on \citep{Abel2002, Bromm2004, Yoshida2006}.

486: In our simulations,

487: black hole seeds of mass $M=10^5 \,h^{-1}\Msun$ are introduced into galaxies

488: as they initially reach $M_{\rm halo} = 10^{10}\,h^{-1} \Msun$.  This choice

489: is a good approximation to what is expected for both of the hypotheses

490: outlined above.  For (1) this is roughly in the correct range; whereas for (2)

491: Eddington growth predicts that the black hole has grown to roughly these

492: values by the time of collapse of $M \sim 10^{10}h^{-1} \Msun$ perturbations,

493: which occurs at $z \sim 10$ in our standard cold dark matter scenario.

494: Additionally, although not required, this value of the initial black hole mass

495: to galaxy ratio fits the observed relations at low redshift

496: \citep{Magorrian1998, Ferrarese2000}. It is important to note that the

497: dominant growth of black holes always occurs in exponential Eddington phases

498: induced by gas accretion so that our results are rather insensitive to the

499: specific choice of the seed mass.

500: 

501: \begin{table}

502: \begin{center}

503: \caption{Numerical parameters of cosmological simulations with BHs}

504: \label{tab:simul}

505: \begin{tabular}{ccccccc}

506: \hline\hline

507: Run  &  Boxsize & $N_{p}$ & $m_{\rm DM}$ & $m_{\rm gas}$ & $\epsilon$& $z_{\rm en

508: d}$ \\

509:      &  $h^{-1}$Mpc &&  $h^{-1} \Msun$ &$h^{-1} \Msun$ &  $h^{-1}$ kpc &  \\

510: \hline

511:   D4 & 33.75 & $2\times 216^3$& $2.75 \times 10^{8}$ & $4.24 \times 10^{7}$  & 6.25

512: & 0.00 \\

513: D6 ({\it BHCosmo}) & 33.75 & $2\times 486^3$& $2.75\times 10^{7}$ &  $4.24\times 10^{6}$&

514: 2.73 & 1.00 \\

515: E6 & 50 & $2\times 486^3$& $7.85\times 10^{7}$ &  $1.21\times 10^{7}$&

516: 4.12 & 4.00 \\

517: \hline\\

518: \end{tabular}

519: \end{center}

520: \vspace{-1cm}

521: \end{table}

522: 

523: When galaxies and their surrounding dark halos merge to form a single dark

524: matter halo, their central black holes are also expected to merge eventually,

525: so hierarchical black hole mergers contribute to the growth of the central

526: black holes.  Whether the forming black hole binaries can really coalesce

527: efficiently is however a matter of debate.  In a stellar environment, it has

528: been argued that the binary hardens only very slowly \citep{Begelman1980,

529: Milosav2003}, while in gaseous environments binaries may coalesce rapidly

530: owing to strong dynamical friction with the gas \citep{Makino2004,

531: Escala2004}. In our galaxy-sized simulations, and even more so in the

532: cosmological boxes, it is not possible to treat in detail the problem of

533: binary hardening, nor to directly calculate the ejection of black holes by

534: gravitational recoil, or by three-body sling-shot ejection of black holes in

535: triple systems.  Because galaxies have typically large central concentrations

536: of gas we instead assume that two black hole particles merge quickly if they

537: come within the spatial resolution of the simulation and their relative speed

538: lies below the gas sound speed. In practice, this means that two sink

539: particles that fulfill these conditions are merged into a single BH particle,

540: with their masses combined.

541: 

542: \begin{figure*}

543: \begin{center}

544: \resizebox{16.0cm}{!}{\includegraphics{f2.eps}}

545: \end{center}

546: \caption{Three level zooms into the simulation region marked by the white

547: rectangle in the $z=3$ panel of Fig~\ref{fig:D6slices}.  The three panels show

548: the gas surface density, color-coded by temperature.  The panels show slices

549: of thickness 5000$ h^{-1} \kpc$ and of decreasing width (from A to C) as we

550: zoom into the region around a black hole of mass, $M_{BH} \sim 7\times

551: 10^{7}$. The yellow circles in the bottom right panel show the black holes in

552: the region, with a symbol size that is related to the BH mass as in

553: Fig.~\ref{fig:D6slices}.}

554: \label{fig:D6zoomsgas}

555: \end{figure*}

556: 

557: 

558: \begin{figure*}

559: \begin{center}

560: \resizebox{16.0cm}{!}{\includegraphics{f3.eps}}

561: \end{center}

562: %\resizebox{9.0cm}{!}{\includegraphics{zoom_slice_stars2027.ps}}}

563: %\resizebox{18.0cm}{!}{\includegraphics{Stars_Map_conv.ps}}}

564: %\resizebox{9.0cm}{!}{\includegraphics{zoom2_slice_stars027.ps}}}}

565: \caption{Stellar density in a three level zoom into the same regions

566: shown in Figure~\ref{fig:D6zoomsgas}, and which is marked by a white

567: rectangle in the $z=3$ panel of Fig~\ref{fig:D6slices}. In the most

568: zoomed-in panel C, a stellar disk and a small bulge component can be

569: seen for the central object.}

570: \label{fig:D6zoomsstars}

571: \end{figure*}

572: 

573: \subsection{Simulation runs}

574: Cosmological simulations which include quasar formation must model

575: sufficiently large volumes to sample a representative part of the universe,

576: but also have high enough resolution to model the full hydrodynamics. This is

577: a substantial challenge, given that the brightest quasars at $z\sim 6$ have a

578: low space density and are believed to reside in fairly massive dark matter

579: halos of mass $M\sim 10^{11}-10^{12} \Msun$, or even larger.  

580: At redshifts $z=2-3$, quasars

581: have a much larger space density, comparable to $L^{*}$ galaxies at $z=0$.

582: 

583: The strategy we choose here is model the universe with a periodic box of

584: moderate size $33.75\,h^{-1}{\rm Mpc}$ that is homogeneously sampled with

585: particles, and which we simulate with $2\times 486^3$ particles, one of the

586: {\it highest resolutions} so far achieved in a full cosmological

587: hydrodynamical calculation of galaxy formation. In this paper, we refer to

588: this largest simulation among our simulation set as the {\it BHCosmo} run.  We

589: will also compare it with two additional simulations which differ in mass and

590: spatial resolution, and/or box size, to test for resolution effects.

591: 

592: 

593: The fundamental numerical parameters of our simulation runs are listed in

594: Table~\ref{tab:simul}, where $N_p$ is the number of dark matter and gas

595: particles, $m_{\rm DM}$ and $m_{\rm gas}$ are their masses (initial mass in

596: the case of the gas). Finally, $\epsilon $ gives the comoving gravitational

597: softening length, and $z_{\rm end}$ the final redshift of the simulation.

598: 

599: 

600: For the physical problem at hand we prefer relatively high resolution in order

601: to capture the physics in high density regions appropriately, but we also need

602: a large volume to study the growth of deep gravitational potentials. Our

603: choice of $33.75\,h^{-1}{\rm Mpc}$ represents a compromise in this respect.

604: While this box-size is too small to be evolved to redshifts lower than

605: $z\simeq 1$ (otherwise the fundamental mode would become non-linear), it is

606: sufficiently large to provide a representative model for $L_\star$-objects at

607: higher redshift, even though very rare systems in the exponential tail of the

608: mass function will not be sampled well.  We also note that the choice of

609: box-size and particle number in the {\it BHCosmo} run is such that the

610: physical resolution at $z\sim 6$ is comparable to that in some of our previous

611: works on galaxy mergers, namely runs which used only $N\simeq 10000$ particles

612: for each galaxy. In this prior work, we have shown that despite the low

613: resolution the results for the black hole mass growth agreed well with those

614: obtained in runs with 128 times higher resolution, and can therefore be

615: considered converged with respect to this quantity. This overlap in

616: resolution between our cosmological runs and our previous work on isolated

617: galaxy mergers gives us confidence that the results of our cosmological runs

618: are not dominated by resolution effects, although it is clear that this needs

619: to be tested separately.  We remark that as part of our previous work on

620: mergers we have run a suite of several hundred galaxy merger simulations

621: \citep{Robertson2006a}, varying all the parameters describing star formation

622: and feedback from supernovae and black hole growth and accretion, besides

623: carrying out numerical resolution studies. The galaxy merger simulations are

624: clearly much better suited for investigating the full parameter space of our

625: model, while for the cosmological runs we have to restrict ourselves to our

626: default model owing to their much larger computational cost.

627: 

628: In addition to the above considerations, the choice of box-size in our new

629: simulations is also motivated by the set of simulations presented in

630: \citet{Springel2003b}. In fact, the {\it BHCosmo} run would be called `D6' in

631: their naming scheme. Being able to directly refer to their runs simplifies the

632: comparison of the physical properties of simulations with and without black

633: holes, e.g.~with respect to the star formation history.

634: 

635: %\subsection{Convergence}

636: 

637: %{\bf (still missing subsection)}

638: 

639: \section{Results}

640: 

641: \subsection{Visualization of the structure and black hole growth}

642: In Figure~\ref{fig:D6slices}, we show slices through the {\it BHCosmo}

643: simulation at a range of redshifts in order to visualize the evolution of the

644: baryonic density field and the growth of black holes.  The slice has a

645: thickness of $5\,h^{-1} {\rm Mpc}$ and shows the full box of size $33.75\,

646: h^{-1}{\rm Mpc}$ on a side. In each panel, the projected gas density field is

647: color-coded according to the gas temperature, with the brightness of each

648: pixel being proportional to the logarithm of the gas surface density. Circles

649: of different size are drawn to mark the locations of BHs of different mass, as

650: labelled.

651: 

652: The images show that black holes emerge in halos starting at high redshift (as

653: early as $z\sim 12$) and subsequently grow by gas accretion, driven by gas

654: inflows that accompany the hierarchical build-up of ever larger halos through

655: merging. As the simulation evolves, the number of black holes rapidly

656: increases and larger halos host increasingly more massive black holes.  We

657: note that the particular slice of the box shown in Figure~\ref{fig:D6slices}

658: does not contain the largest black hole in the simulation's volume, which turns

659: out to be located in the highest density region in the simulation.  We will

660: discuss this region separately in later sections.

661: 

662: By plotting the density field color-coded by temperature we can also see

663: traces of heating effects from strong gas outflows, which are caused by black

664: hole feedback and galactic winds from star formation. In our numerical mode,

665: quasar-driven outflows occur once a central BH gas grown so much that its

666: energy feedback in accretion phases is able to transfer sufficient energy to

667: the remaining gas to unbind and expel part of it from the galaxy potential as

668: a wind. This then terminates further strong growth of the BH, which is key

669: factor in establishing a self-regulated nature of the growth of black holes in

670: our model, as we discussed in detail in our previous work on galaxy merger

671: simulations.

672: 

673: To illustrate how well the mass resolution of the simulation captures details

674: of galaxy formation sites, we zoom in onto a region of $6\,h^{-1}{\rm Mpc}$ on

675: a side at $z=3$ (as indicated by the box drawn in Fig~\ref{fig:D6slices}),

676: showing the gas density in Figure~\ref{fig:D6zoomsgas}, and the stellar

677: density of the same region in Figure~\ref{fig:D6zoomsstars}. The middle

678: panels (labeled B) in both figures show a second zoom-level into a region of

679: $2\,h^{-1}{\rm Mpc}$ on a side (indicated by the square in the A panels).

680: Finally, the top panels show a further enlargement by a factor of 8.

681: 

682: In accordance with our seeding procedure, black holes are located in the

683: highest density regions. The more massive ones are found within the largest

684: halos, which have also undergone more prominent star formation. Such a

685: correspondence is expected in our model since large-scale gas inflows into the

686: centers of halos lead both to star formation (and starbursts) as well as to

687: nuclear black hole growth. In the highest level zoom of

688: Figures~\ref{fig:D6zoomsgas} and \ref{fig:D6zoomsstars}, the central galaxy,

689: with an extent of $\sim50h^{-1} {\rm kpc}$ (comoving), has a very rough 

690: disk-like morphology with a central stellar bulge.

691: Nevertheless, it is clear that our cosmological simulations in general still

692: have too low resolution for properly resolving galaxy morphologies. We also

693: note that producing disk galaxies with the right size and abundance in

694: cosmological hydrodynamical simulations is an essentially unsolved problem,

695: and the outlook for obtaining a solution to this long-standing challenge has

696: only slightly improved by recent works on disk galaxy formation

697: \citep{Robertson2004, Governato2007, Okamoto2007}.

698: %For our simulations, this implies that

699: %they cannot be expected to be able to accurately resolve large amounts of cold

700: %gas in centrifugal support in thin, gas-rich spirals. This probably leads to

701: %inaccuracies in the detailed growth history of our central BHs.  Instead of

702: %being `parked' at large radii until the final coalescence of two galaxies in a

703: %major merger, the gas will be fed prematurely to the nuclei in our runs,

704: %washing out some of the details of the time evolution of the merging systems.

705: %However, our resolution tests demonstrate that even under conditions of very

706: %poor resolution, quasar activity is triggered by mergers in a robust way and

707: %the BH grows to the right final mass expected for the velocity dispersion of

708: %the spheroid. 

709: 

710: \subsection{The evolution of the global black hole mass density}

711: The black hole mass density at the present epoch is estimated from direct

712: measurements of black hole masses in local galaxies (to establish, e.g., the

713: $M_{\rm BH}-M_\star$ relationship), combined with a suitable integration over

714: the galaxy luminosity function ~\citep{Fabian1999, YuTremaine2002,

715: Marconi2004}. The local value thus obtained matches that of the total relic BH

716: mass density estimated from a time integration of the luminosity output of

717: active galactic nuclei and quasars at X-ray and optical wavelengths

718: \citep{Soltan1982, Marconi2004, Shankar2004}, or in a less model-dependent

719: manner from an empirical determination of the bolometric quasar

720: luminosity function \citep{Hopkins2007b}.

721: 

722: \begin{figure}

723: \hspace*{-0.5cm}\vspace*{-1.0cm}\\

724:  \resizebox{9.0cm}{!}{\includegraphics{f4.ps}}

725: \caption{{\it Top panel}: The global black hole mass density evolution in the

726: {\it BHCosmo}/D5 run, shown by a solid line. The the star symbols and the

727: dashed line give the corresponding stellar mass density evolution, multiplied

728: by 0.0007 for easier comparison.  The grey dotted line shows the stellar mass

729: density in the D5 simulation which did not include black hole accretion and

730: associated quasar feedback ~\citep{Springel2003a}. The thin dashed line shows

731: the results from the lower resolution box, D4 described in

732: Table~\ref{tab:simul}. Different colors simply indicate the different

733: redshifts consistent with the scheme used in other figures.  The shaded grey

734: triangle indicates observational constraints taken from the literature

735: \citep{Marconi2004, Shankar2004}.  {\it Bottom Panel}: The global history of

736: the black hole accretion rate (solid line) and star formation rate (dot-dashed

737: line with stellar symbols) densities. The SFR is rescaled by $10^{-3}$ for

738: graphical clarity. In addition, we show the SFR history in the D5 simulations

739: without black holes (grey dotted line). Most of the black hole and stellar

740: mass is assembled by $z\sim 2-3$, but the peak in the BHAR density function is

741: far more pronounced than that of the SFR density. }

742:     \label{fig:D6acc_den}

743: \end{figure}

744: 

745: Figure~\ref{fig:D6acc_den} shows our simulation prediction for the global

746: density $\rho_{\rm BH}$ and its evolution with redshift ({\it thick black

747:   line}). We find that the normalization of the black hole mass density is in

748: agreement with the observational estimate of $\rho_{\rm BH} (0) =

749: 4.6^{+1.9}_{-1.4} \times 10^{5} h_{0.7}^2 \Msun \Mpc^{-3}$ of

750: \citet{Marconi2004} and its extrapolation to $z\sim 3$, derived by exploiting

751: hard X-ray and optically selected AGNs and quasars.  The grey area in

752: Figure~\ref{fig:D6acc_den} shows the region delimited by observational

753: constraints examined in detail in the literature \citep[e.g][]{Salucci1999,

754:   Marconi2004, Shankar2004}. The agreement we see here is very reassuring.  It

755: shows that our black hole accretion and feedback model in the {\it BHCosmo}

756: run is adequate and provides a realistic account for the dominant mode of

757: global black hole growth in our universe.

758: 

759: The black hole mass density evolves rapidly at high redshift, increasing by

760: four orders of magnitude between $z \sim 10$ to $z \sim 3$. While below this

761: redshift, there is some further growth, it only accounts for roughly a

762: doubling of the BH density down to $z= 1$. This is corroborated by the bottom

763: panel of Figure~\ref{fig:D6acc_den}, where we plot the history of the global

764: black hole accretion rate (BHAR) density in our {\it BHCosmo} run. It is

765: evident that the black hole mass growth occurs mainly by accretion above $z

766: \sim 3$.  Indeed, the BHAR density rises steadily at early times to a peak at

767: $z \sim 3$, where it reaches a value $ \sim 10^{-4} \Msun \yr^{-1} \Mpc^{-3}$.

768: Below this redshift, it drops rather rapidly, becoming more than an order of

769: magnitude lower by $z= 1$.  We note that the 'spike' seen in the BHAR density

770: at $z \simeq 6.5$ is caused by a single object in the box, namely the rapid

771: formation of the most massive black hole in the simulation at this epoch. We

772: will discuss the history of the corresponding halo and its embedded black hole

773: in more detail in \S 5.

774: 

775: \begin{figure}

776: \hspace{-1cm}

777:       \resizebox{10.cm}{!}{\includegraphics[width=4.in]{f5.ps}}

778:     \caption{The evolution of the black hole mass function in the {\it

779: BHCosmo} run. The different colors indicate different redshifts as

780: labeled in the figure. For comparison, the dark grey shaded region

781: shows the black hole mass function derived from local galaxies

782: \citep{Marconi2004}. The light gray area adds the constraint from the

783: integration of the hard X-ray luminosity function \citep{Shankar2004,

784: Merloni2004, Marconi2004}.  The largest uncertainties are at the high

785: and low mass end. The simulation results are in good agreement with

786: the observed mass function at the high mass end, and in reasonable

787: agreement at intermediate masses.}

788:     \label{fig:bhmf}

789: \end{figure}

790: 

791: 

792: \begin{figure}

793: \hspace{-1cm}

794:       \resizebox{10cm}{!}

795: {\includegraphics[width=4.in]{f6.ps}}

796: \caption{The time evolution of the accretion rate distribution as a

797:   function of the Eddington ratio, for the {\it BHCosmo} run. The

798:   different colors denote our measurements at different redshifts, as

799:   indicated in the legend. For the $z=1$ distribution function, we

800:   separately show three components corresponding to different regions

801:   of the black hole mass function.  In particular, the dotted,

802:   dot-dashed, and dot-dot-dashed lines give the separate contributions

803:   from the three different mass bins $M_{\rm BH} > 10^{8} \Msun$,

804:   $10^7 < M_{\rm BH}/\Msun \le 10^{8}$ and $10^6 \leq M_{\rm BH}/\Msun

805:   <10^{7}$, respectively. 

806: %\bf (I suppose the 'log' in the vertical

807: %  axis label applies to the whole expression? If so, it might be

808: %  better to make a bracket about everything.  Should we give the

809: %  numbers normalized to volume?)} 

810: }

811:     \label{fig:bhaf}

812: \end{figure}

813: 

814: \subsection{Black hole mass function and accretion rate function}

815: \label{sec:massfn}

816: In Figure~\ref{fig:bhmf}, we plot the black hole mass function at a

817: number of different redshifts. We find that the final black hole mass function

818: in our simulation (for $z=1$) is is quite good agreement with the one measured

819: locally, especially on the high-mass side. The $z=0$ constraint is indicated

820: by the {\em dark grey area} taken from the compilation of \citet{Marconi2004},

821: which is based on a combination of different observational data

822: \citep{kochanek2001, Nakamura2003, Bernardi2003, Sheth2003}. There is a small

823: deficit in our model at intermediate BH masses, but note that this will be

824: filled in at least partly by the expected residual growth from $z=1$ to $z=0$.

825: The {\em light grey area} adds an additional constraint for the contribution

826: of relic AGN, derived from an integration of the the hard X-ray luminosity

827: function \citep{Shankar2004}.  (Note that in this latter case the

828: normalization of the mass function depends on the value assumed for the

829: radiative efficiency when converting from the luminosity function to the mass

830: function.)

831:  

832: \begin{figure*}

833: \begin{center}

834: \resizebox{8.5cm}{!}{\includegraphics{f7.ps}} %

835: \resizebox{8.5cm}{!}{\includegraphics{f7a.ps}}

836: \end{center}

837:     \caption{An illustration of a large group in the {\it BHCosmo} run at

838: $z=1$.  {\it Left panel:} the gas density distribution. {\it Right

839: panel:} the stellar distribution. We have run a sub-group finder to

840: identify all the systems (galaxies) within each halo and analyze their

841: properties accordingly. The stellar system shown in yellow on the

842: right hand side is the main galaxy within this large group, while the

843: satellite galaxies are shown in grey. The images are 400 kpc on a side.

844:     \label{fig:subgroups}}

845: \end{figure*}

846: 

847: \begin{figure*}

848: \begin{center}

849: \resizebox{18.0cm}{!}{\includegraphics{f8.ps}}

850: \end{center}

851:     \caption{The evolution of the $M_{\rm BH} - \sigma$ relation in the {\it

852: BHCosmo} simulation. The masses of BHs and the projected stellar velocity

853: dispersions within the half mass radius ($R_e$) have been measured in our

854: simulated galaxies and are plotted at $z=1$, $2$, $3$, $4$, $5$, and $6.5$. We

855: compared our measurements at all redshifts with the best-fit to the local

856: $M_{\rm BH}- \sigma$ relation of \citet[][hereafter T02]{Tremaine2002}, which

857: is shown as a thick gray line. Linear regression fits to our simulated BHs are

858: shown by solid lines at each redshift, with $1-\sigma$ errors indicated by the

859: hatched regions. For ease of comparison, the dotted lines in each panel show

860: the best fit relations at all redshifts. The points are color-coded

861: according to their accretion rates in units of Eddington, as indicated in the

862: color bar at the top right hand corner of the figure. 

863: %{\bf (I think the

864: %circles are too large in this figure, producing too much overlap among

865: %them. Also, the total number of symbols appears somewhat small for the volume

866: %- have you imposed a selection criterion for the galaxies of some kind? I

867: %would also suggest to draw the solid black lines above the thick grey one, and

868: %not the other way around, then one can see it better. The tick marks on the

869: %horizontal axes are a bit on the short side. Finally, I think the axes-labels

870: %and panel captions would look much nicer if you would use postscript fonts for

871: %them.)}

872: }

873: \label{fig:mbhsigma}

874: \end{figure*}

875: 

876: 

877: \begin{figure*}

878: \begin{center}

879:   \resizebox{18.0cm}{!}{\includegraphics{f9.ps}}

880: \end{center}

881:     \caption{The evolution of the $M_{\rm BH} - M_{*}$ relation in the

882: {\it BHCosmo} simulation. The masses of BHs and the corresponding

883: stellar mass have been measured in our simulated galaxies and are

884: plotted at $z=1$, $2$, $3$, $4$, $5$, and $6.5$. They are compared

885: with the best-fit power law to the local $M_{\rm BH}- M_{*}$ relation

886: by \citet{Haring2004}, shown by the thick gray line. Our fits to

887: the simulation results at each redshift are shown as solid black

888: lines. The relation is tight with small scatter ($1-\sigma$

889: uncertainty ranges are plotted as hatched regions but are hardly

890: visible at low redshift). As in Figure~\ref{fig:mbhsigma}, the points

891: are color coded by accretion rate. The dotted lines show in each panel

892: the best fit relation at the other redshifts. 

893: %{\bf (as in the previous

894: %figure, the description of the dotted lines is a bit confusing, and

895: %I'd suggest to omit them. Like for the previous figure, I'd suggest to

896: %change the size of the circles, make the x-tickmarks longer and switch

897: %to postscript fonts)}}

898:     \label{fig:mbhmstar}}

899: \end{figure*}

900: 

901: As expected in a hierarchical formation scenario, the high mass end of the

902: mass function shifts to larger masses with redshift. However, it is

903: interesting that the mass function for high masses grows rather rapidly at

904: early times relative to the low mass end.  On the other hand, for redshifts $z

905: \lesssim 2$, the high mass end is virtually fully assembled while the BH mass

906: function continues to grow for low to intermediate masses, leading effectively

907: to a steepening of the mass function for $z \simlt 3$.  This appears

908: consistent with the emerging observational picture of the evolution of the

909: supermassive black hole population according to which massive BHs ($M_{\rm BH}

910: \simgt 10^9 \Msun$) are assembled early and are then likely undergoing

911: comparatively passive evolution in the centers of large spheroids

912: \citep{Shields2003, Vestergaard2004, Adelberger2005}. More generally, this

913: phenomenon is described by the idea of an `anti-hierarchical' black hole

914: growth, or equivalently a `downsizing' of black hole activity

915: \citep{Merloni2004, Marconi2004}, which is derived from constraints on the

916: accretion history from X-ray luminosity functions.

917: 

918: %Intermediate mass BHs ($7.0 \lesssim \log

919: %(M_{\rm BH} / M_{\sun}) \lesssim 8.3$) in the mass function undergoing their

920: %most significant evolution below $z\sim 2$. The growth of the mass function in

921: %this regime, contributes also to the slow climb of the black hole mass density

922: %(Fig.~\ref{fig:D6acc_den} below $z \sim 2$. The formation of the most massive

923: %black holes (and spheroids) provides the dominant contribution to the global

924: %black hole mass density.

925: 

926: Finally, in Figure~\ref{fig:bhaf}, we show the evolution of the corresponding

927: accretion rate distribution function for the black holes in our simulation,

928: expressed in units of the Eddington rate. This function is strongly peaked at

929: a few percent below the critical Eddington value, with most black holes

930: accreting at $10^{-2} \lesssim \mdot/\mdot_{\rm Edd} \lesssim 1 $ at redshifts

931: $z\simgt 6$. The distribution becomes wider and develops a small amplitude

932: tail at low Eddington rates for $ 6 < z < 3$.  Below $z \sim 3$ the peak of

933: the accretion rate function shifts to $\mdot/\mdot_{\rm Edd} \sim 10^{-2} -

934: 10^{-3}$ and an increasingly large population of sources is present accreting

935: down to $10^{-6}$ Eddington. At $z\sim 1$ the distribution function is

936: extremely wide and dominated by sub-Eddington (to severely sub-Eddington)

937: sources, in terms of number. While hence the number of sources accreting at

938: sub-Eddington rates increases sharply with decreasing redshift, the

939: quasar-like population, i.e.~the sources accreting close to Eddington,

940: decreases with decreasing redshift, with a maximum at $z\sim 3$.  For $z\sim

941: 1$, we also include in Figure~\ref{fig:bhaf} measurements of the separate

942: contributions to the accretion rate function from three different black hole

943: mass ranges.  The dotted line gives the Eddington ratio distribution for

944: $M_{\rm BH} > 10^{8} \Msun$, and the dash-dotted and dot-dot-dashed lines

945: are for masses $10^7 < M_{\rm BH}/\Msun \le 10^{8}$ and $10^6 \leq M_{\rm

946: BH}/\Msun <10^{7}$, respectively.

947: 

948: Taken together, the evolution of the black hole mass and accretion rate

949: functions implies that most massive black holes assemble early and do so at

950: close to their critical rate. At low redshift, progressively lower

951: luminosities and lower mass systems start dominating the black hole activity.

952: This is in accord with recent results from studies of the hard X-ray

953: luminosity function of quasars and AGN, e.g.~typical Seyfert-like objects with

954: $M_{\rm BH} \sim 10^7 - 10^8 \Msun$ accreting at a few percent of the

955: Eddington rate~\citep[e.g.][]{Ueda2003, Hasinger2005}.

956: 

957: \subsection{Comparison of the histories of black hole growth and star formation}In Figure~\ref{fig:D6acc_den}, we have already shown and discussed the

958: evolution of the black hole mass and accretion rate densities. In the

959: same Figure, we show corresponding results for the evolution of the

960: stellar mass density, $\rho_{*}$, and the star formation rate (SFR)

961: density, with their normalizations rescaled by factors of $7\times

962: 10^{-3}$ and $10^{-3}$, respectively, for plotting purposes. For

963: comparison, we also include results for the D5 simulation of

964: \citet{Springel2003b} which did not include black holes and any

965: associated accretion or feedback processes (dotted grey lines).

966: 

967: We can see that $\rho_{*}$ far exceeds $\rho_{\rm BH}$ at all redshifts, with

968: $\rho_{*}$ evolving less strongly with redshift than $\rho_{\rm BH}$ for $z

969: \simgt 3$. At early times, $\rho_{\rm BH}$ rises more rapidly than the star

970: formation density, while it tracks its evolution below this redshift. If we

971: parameterize the ratio of $\rho_{\rm BH}/ \rho_{*}$ by an evolutionary factor

972: $(1+z)^{\alpha}$ we find that its evolution is approximately given by

973: \begin{equation}

974: \rho_{\rm BH}/ \rho_{*} \sim \left\{ \begin{array}{ll} \phi_{3} (1+z)^{-3} & \mbox{if $z \ge

975: 3.5$} \nonumber \\ & \\ \phi_1 [(1+z)]^{-0.6} & \mbox{if $z <

976: 3.5$}

977: \end{array} \right. , \label{eq:rho}

978: \end{equation}

979: with $\phi_{3} = 5 \times 10^{-2}$ and

980: $\phi_{1}=2.2\times10^{-3}$. Accordingly, up to $z\sim 3$, the evolution of

981: the star formation rate density is considerably shallower than that of the

982: black hole accretion rate density. Below this redshift, the BHAR and SFR

983: densities closely track each other.  As a result, the BHAR density has much

984: more pronounced peak, which we find to lie at slightly lower redshift than the

985: peak of the SFR. Parameterizing the evolution of the BHAR and SFR density

986: ratio with a power law in $(1+z)$ leads to

987: \begin{equation}

988: \dot{\rho_{BH}} / \dot{\rho}_{*} \sim\left\{ \begin{array}{ll} \dot{\rho}_{3} (1+z)^{-4} & \mbox{if $z \ge

989: 3$} \nonumber \\ & \\ \dot{\rho}_1 [(1+z)]^{-0.1} & \mbox{if $z <

990: 3$}

991: \end{array} \right. , \label{eq:drho}

992: \end{equation}

993: with $\dot{\phi}_{3} = 0.2$ and $\dot{\phi}_{1} = 10^{-3}$. 

994: 

995: Note that Equations~(\ref{eq:rho}) and (\ref{eq:drho}) are only meant to

996: provide approximate scalings for our results from the simulations. The

997: important point we want to emphasize is that our results imply a different and

998: much stronger evolution of the black hole mass and accretion rate densities at

999: high redshift relative to the stellar density and star formation rate density.

1000: The black hole mass density tends to be assembled later than the stellar mass,

1001: despite the growth of (a small number of) very massive BHs already at high

1002: redshift. However, for $z \sim 3$ and below, our models predict that the black

1003: hole mass and stellar mass densities, as well as the BHAR and SFR densities,

1004: closely track each other allowing only for small amounts of relative evolution

1005: between the two. 

1006: 

1007: Another important point from Figure~\ref{fig:D6acc_den} is that the feedback

1008: we associate with black hole accretion does not significantly affect the

1009: global assembly of stellar mass. The peak of the SFR density is unaffected by

1010: the inclusion of BH feedback, but the drop in SFR density ($z < 3$) is

1011: slightly more abrupt in the simulations with black holes. This effect becomes

1012: more pronounced at $z\sim 1$, the final redshift for our simulation.  At still

1013: lower redshift, we expect that BH feedback will become important in regulating

1014: the cooling and star formation in very massive halos. This is then ascribed

1015: not to quasar growth but rather to a `radio mode'. We explore this different

1016: feedback channel in a companion paper by \citet{Sijacki2007}.

1017: 

1018: \section{The $M_{\rm BH} - \sigma$ and $M_{\rm BH} - M_{*}$ relations}

1019: \label{msigmarel}

1020: 

1021: \subsection{Identification of groups and subgroups}

1022: As a prerequisite for being able to study correlations between black

1023: hole and host galaxy properties in our simulation we first need to

1024: apply a suitable group finding algorithm that reliably identifies the

1025: stellar mass associated with the different galaxies. Note that

1026: especially the more massive halos identified by our basic

1027: friends-of-friends grouping algorithm used for finding virialized

1028: objects often contain a number of galaxies. This is illustrated in

1029: Figure~\ref{fig:subgroups}, where we show a large cluster-sized group

1030: selected from the $z=1$ output of the {\it BHCosmo} simulation. The

1031: panel on the left shows the gas density distribution in this large

1032: group, while the panel on the right hand side displays the stellar

1033: distribution. It is evident that the group contains several,

1034: gravitationally bound galaxies.

1035: 

1036: Our sub-group finder identifies all the galaxies within each group.  Our

1037: method to identify galaxies within a given group is based on a variant of the

1038: {\small SUBFIND} algorithm \citep{Springel2001}.  We first compute an

1039: adaptively smoothed baryonic density field for all stars and gas particles,

1040: allowing us to robustly identify centers of individual galaxies.  We find the

1041: extent of these galaxies by processing the gas and star particles in the order

1042: of declining density, adding particles one by one to the galaxies attached to

1043: the galaxy to which its nearest denser neighbor already belongs~\citep[see

1044: also][]{Nagamine2004}. Note we are interested in the stellar and gas content

1045: of galaxies and associating the gaseous component to galaxies, as they

1046: typically they contain very dense star-forming gas, makes the method very

1047: robust in finding galaxies. This allows us to compute physical properties

1048: such as stellar mass, star formation rate, stellar velocity dispersion,

1049: metallicity, black hole mass and BH accretion rate separately for each

1050: galaxy. As an illustration, Figure~\ref{fig:subgroups} shows the stellar

1051: distribution associated with the largest galaxy in the group (which also hosts

1052: the most massive BH in the group) in yellow.

1053: 

1054: For each galaxy/subgroup that contains stars and a black hole, we calculate the

1055: projected (spherically averaged) half-mass effective radius, $R_e$, and the

1056: mass-weighted stellar velocity dispersion $\sigma$ within $R_e$.  Note however

1057: that in order to make the measurements of $\sigma$ and $R_e$ somewhat accurate

1058: we only consider those objects that contain more than 100 stars particles

1059: within $R_e$.  This determination of $\sigma$ closely resembles the procedure

1060: for measuring the velocity dispersion from observational data

1061: \citep{Gebhardt2000}, allowing for a direct comparison.

1062: 

1063: \subsection{The predicted $M_{\rm BH} - \sigma$ and $M_{\rm BH} -M_{*}$

1064:   relationships and their evolution} Figures~\ref{fig:mbhsigma} and

1065: \ref{fig:mbhmstar} plot the $M_{\rm BH} - \sigma$ and $M_{\rm BH} -M_{*}$

1066: relations, respectively, for our simulated galaxies at redshifts $z=1$, $2$,

1067: $3$, $4$, $5$ and $6.5$ (from top left to bottom right). Each measurement is

1068: color-coded according to the accretion rate of the corresponding black

1069: hole. We find a strong power-law correlation between both the velocity

1070: dispersion and the stellar mass with the black hole mass. Furthermore, at

1071: $z=1$, these correlations agree very well with the ones observed at the

1072: present epoch~\citep{Ferrarese2000, Gebhardt2000, Tremaine2002, Haring2004} 

1073: over a large dynamic range. This is a remarkable confirmation of the basic

1074: merger-driven scenario for self-regulated BH growth that we previously

1075: explored in isolated high-resolution galaxy merger simulations. Based on the

1076: same model, and without any fine-tuning or change of the model parameters, we

1077: here obtain a good match to the observed $M_{\rm BH} - \sigma$ relation

1078: directly from simulations that start from cosmological initial conditions and

1079: self-consistently account for the hierarchical build up galaxies in a

1080: $\Lambda$CDM universe.

1081: 

1082: \subsubsection{The evolution of $M_{\rm BH} -\sigma$}

1083: At higher redshifts, and in particular for $z > 3$, our results for the

1084: $M_{\rm BH} - \sigma$ relation shows some degree of evolution relative to the

1085: local relations. To compare with the local  data as a function of time,

1086:  we

1087: fit the $M_{\rm BH}-\sigma$ relation at each redshift 

1088: with a simple power-law of the form \beq \log \left(\frac{M_{\rm

1089:       BH}}{\Msun}\right) = a\, \log \left(\frac{\sigma}{200 \kmps}\right) + b.

1090: \label{eqn:msigmafit}

1091: \eeq The best-fit relations thus obtained are shown in

1092: Figure~\ref{fig:mbhsigma} as solid black lines, with the dispersion indicated

1093: by hatched regions. The dotted lines are the best-fit relations for all

1094: redshifts combined. We compare with the observed relation as determined by

1095: T02, which is described by a slope $a=4.02$, a normalization $b=8.2$ and a

1096: dispersion $\Delta \sim 0.25-0.30$ (grey thick line in

1097: Fig.~\ref{fig:mbhsigma}).

1098: 

1099: \begin{table}

1100: \begin{center}

1101: \caption{Parameters of best-fit $M_{\rm BH} -\sigma$ relations}

1102: \label{tab:msigma}

1103: \begin{tabular}{ccccc}

1104: \hline\hline\\

1105: z  &  slope (a) & normalization, (b) & scatter $\Delta$ & slope (a$_{s}$)\\

1106:  \hline\\

1107: 1 ....... &  $3.95\pm 0.10$ & $8.29\pm0.03$ & 0.10 & 3.9  \\

1108: 2 ....... &  $4.01\pm 0.15$ & $8.42\pm0.04$ & 0.16 & 4.1  \\

1109: 3 ....... &  $4.21\pm 0.22$ & $8.32\pm0.07$ & 0.23 & 5.2 \\

1110: 4 ....... &  $4.54\pm 0.35$ & $8.17\pm0.10$ & 0.36 & 6.7 \\

1111: 5 ....... &  $3.37\pm 0.45$ & $7.70\pm0.13$ & 0.47 &  --   \\

1112: 6.5 ..... &  $3.26\pm 0.85$ & $7.50\pm0.25$ & 0.89 &  --  \\

1113: \hline\\

1114: \end{tabular}

1115: \end{center}

1116: \vspace{-1cm}

1117: \end{table}

1118: 

1119: The constants $a$ and $b$ for our best-fit relations and their dispersions are

1120: tabulated in Table~\ref{tab:msigma} for different redshifts.  Note that we

1121: here do not attempt to assess the statistical significance of the correlations

1122: in detail as the sources of systematic errors in the numerical measurements of

1123: $\sigma$ and $M_{\rm BH}$ cannot be easily quantified in the simulations.  In

1124: particular, the cosmological simulations cannot determine the morphological

1125: properties of galaxies and therefore do not provide a direct measure

1126: of spheroid masses or their velocity dispersions.  Our

1127: fitting procedure is merely intended to provide a first characterization of

1128: the overall evolution of the slope and normalization of the relations in the

1129: simulation model.  As Figure~\ref{fig:mbhsigma} and Table~\ref{tab:msigma}

1130: indicate, the $M_{\rm BH}-\sigma$ relation predicted from our simulation is

1131: consistent with a slope $\sim 4$ at low redshift, as observed. At $z=3-4$, the

1132: slope appears to be slightly steeper and at $z=5-6$ slightly shallower, but the

1133: small number of systems at $z \sim 6$ makes the latter trend uncertain.

1134: 

1135: Inspection of Figure~\ref{fig:mbhsigma} shows a qualitative trend whereby the

1136: larger systems with high $\sigma \simgt 150 \kmps$ (which are also those that

1137: are best resolved in our simulation) appear to populate the high mass end of

1138: the $M_{\rm BH}-\sigma$ relation already at $z > 2-3$.  The lower mass end of

1139: the $M_{\rm BH}-\sigma$ relation is then increasingly filled in towards $z\sim

1140: 1$. To illustrate this trend more explicitly, we perform a fit of the relation

1141: using only those systems with $\sigma \ge 150\kmps$; this is shown by the

1142: dashed line in Figure~\ref{fig:mbhsigma}. The slope $a_{s}$ obtained for these

1143: `high' $\sigma$ black holes is typically steeper than for the full relation

1144: (our measured values for $a_{s}$ are listed in Table~\ref{tab:msigma}). At a

1145: fixed and relatively high $\sigma$, the mean $M_{\rm BH}$ is larger at $z\sim

1146: 3-4$ than at $z\sim 1$.  This trend suggests that black hole growth predates

1147: the final growth of the spheroid potential at scales $\sigma \sim 170\,{\rm

1148: km\,s^{-1}}$, which is consistent with the recent measurements of the $M_{\rm

1149: BH}-\sigma$ relation in a sample of Seyferts at $z\sim 0.35$ by

1150: \citet{Woo2006}. This relatively more efficient black hole growth at a fixed

1151: $\sigma$ at high redshift may thus be responsible for a mild change in the

1152: high mass end of the $M_{\rm BH}-\sigma$ relation. In the remainder of this

1153: section we will discuss possible additional dependences in the $M_{\rm

1154: BH}-\sigma$ \citep{Hopkins2007a} relation which may produces a small change in

1155: the slope of the relation with time.

1156: 

1157: Note also that at $z \ge 4$ black holes are more likely to accrete close to

1158: their critical Eddington rates, as we showed earlier. In our models, the

1159: $M_{\rm BH}-\sigma$ relation is a natural consequence of the self-regulated

1160: growth of black holes \citep{DiMatteo2005}, where a black hole grows until its

1161: released energy is sufficient to expel the gas in its surroundings in a quasar

1162: driven wind, which then terminates nuclear accretion. For this reason, we

1163: expect the relation to show more scatter at times when most systems are still

1164: actively growing and accreting close to their Eddington values (see

1165: Table~\ref{tab:msigma}). This is expected as the primary path for assembly BH

1166: mass is via accretion during major mergers so that the relations converge and

1167: get increasingly tighter as galaxies undergo major mergers and continue to

1168: merge.

1169: 

1170: \begin{figure}

1171: \hspace*{-0.4cm}

1172: \resizebox{8.5cm}{!}{\includegraphics{f10.ps}}

1173: \caption{The stellar velocity dispersion $\sigma$ versus stellar

1174: $M_{*}$ at $z=1$, $2$, $3$, $4$, $5$, $6.5$ (indicated by different

1175: colors from blue to pink, the same ones as used in

1176: Figs.~\ref{fig:D6acc_den}-\ref{fig:bhaf}). The best-fit power-law to

1177: the trend is shown with a dotted line at each redshift, and with a

1178: solid line at $z=1$. The dispersion $\sigma$ at fixed $M_{*}$

1179: increases with increasing redshift, which can be interpreted as a weak

1180: evolution in the Faber \& Jackson relation.

1181:    \label{fig:sigma_mstar}}

1182: \end{figure}

1183: 

1184: 

1185: \begin{figure}

1186:   \resizebox{9cm}{10cm}{\includegraphics{f11.ps}}

1187:     \caption{The redshift evolution of the projected half mass radius

1188: $R_{e}$ (top panel), cold gas fraction $f_{\rm gas}$ (middle panel)

1189: and characteristic black hole to stellar mass ratio $\left <M_{\rm

1190: BH}/M_{*}\right>$ (bottom panel), for systems with $M_{*} > 3\times

1191: 10^{10} \Msun$ (dashed lines) or $M_{*} > 6 \times 10^{10} \Msun$

1192: (solid) in the {\em BHCosmo} run. These threshold values were chosen

1193: to compare with the observed evolution determined by

1194: \citet{Trujllo06}. The increase in cold gas content in high redshift

1195: progenitor hosts, and the trend to more compact systems with an

1196: increasing ratio of $M_{\rm BH}/M_{*}$ at a fixed stellar mass is

1197: consistent with the recent results of \citet{Trujllo06} and

1198: \citet{Peng2006}, as well as the BHFP \citep{Hopkins2007a}.

1199:     \label{fig:fgasre}}

1200: \end{figure}

1201: 

1202: \begin{figure*}

1203: \begin{center}

1204: \resizebox{9.5cm}{!}{\includegraphics{f12a.ps}}%

1205: \hspace*{-1cm}\resizebox{9.5cm}{!}{\includegraphics{f12b.ps}}

1206: \end{center}

1207:  \caption{Simulation prediction for two projections of the `Black Hole

1208: Fundamental Plane' (BHFP) relation at $z=1$, in terms of $\sigma$ and

1209: $M_{*}$ (left panel), and $\sigma$ and $R_e$ (right panel).  We

1210: compare with the best-fit relations from \citet{Hopkins2007a}, shown

1211: as dotted lines. The simulation agrees well with the conjecture of a

1212: BHFP, which confirms the overall trends we have found in the $M_{\rm

1213: BH}-\sigma$ and $M_{\rm BH}-M_{*}$ relations. This likely owes to

1214: the scatter in the measurements of both $\sigma$ and in particular

1215: $R_e$ which are noisy measurements in the simulations. The colors

1216: indicate accretion rate values (as in previous figures and indicated 

1217: in the color bar).}

1218:     \label{fig:fplane}

1219: \end{figure*}

1220: 

1221: The overall normalization of the $M_{\rm BH}-\sigma$ relation evolves

1222: weakly with redshift when compared to the TO2 result for the present

1223: epoch. Below $z \sim 3$, there is virtually no evolution (bearing in

1224: mind the weak trend discussed above for $\sigma > 150 \kmps$), while

1225: we find a weak drift of the normalization beyond $z\sim 3$.  A rough

1226: parameterization of this evolution in the normalization of the $M_{\rm

1227: BH}-\sigma$ relation is given by $ M_{\rm BH}/\sigma^4 \propto

1228: (1+z)^{\alpha}$ with $\alpha =-0.2$.

1229: 

1230: These results for an overall modest evolution in the normalization are in

1231: qualitative agreement with those from an analysis of isolated galaxy merger

1232: simulations by \citet{Robertson2006a}.  We note that in the latter study a

1233: redshift scaling of galaxy properties and a specific set of structural

1234: properties of the merging galaxies had to be assumed, which introduced an

1235: important systematic uncertainty. While having lower numerical resolution per

1236: merger, the simulation we analyze here eliminates this caveat by providing a

1237: fully self-consistent cosmological history for the formation and evolution of

1238: galaxies and their black holes (albeit at a lower spatial resolution). This

1239: provides an important confirmation of the analysis of \citet{Robertson2006a},

1240: who however did not find evidence for an evolution of the slope at the high

1241: mass end of the $M_{\rm BH}-\sigma$ relation.

1242: 

1243: %Note that, as shown in the isolated merger studies, the normalization of the

1244: %$M_{BH}-\sigma$ relationship is also set by adjusting the feedback

1245: %parameter, $\epsilon_{\rm f}$. As we have kept this fixed at $5\%$ consistent

1246: %with these latter studies, the slope of the relations is not adjustable or

1247: %tunable but a natural consequence of the self-regulated black hole growth.

1248: 

1249: %here

1250: 

1251: \subsubsection{The evolution of $M_{\rm BH} -M_{*}$}

1252: In Figure~\ref{fig:mbhmstar}, we show the $M_{\rm BH} -M_{*}$ relation

1253: from the simulations at $z=1$, $2$, $3$, $4$, $5$ and $6.5$, alongside

1254: the local observational relation determined by \citet[][thick grey

1255: lines]{Haring2004}. Our best-fit relation at each redshift is shown by

1256: a solid line while the dotted lines in each panel show results for the

1257: other redshifts. Table~\ref{tab:mbhmstar} gives the slope $c$,

1258: normalization $d$, and dispersion $\Delta_m$ for our best-fit

1259: relations of the form

1260: \beq \log \left(\frac{M_{\rm BH}}{\Msun}\right) = c \log

1261: \left(\frac{M_{*}}{10^{11} \Msun}\right) + d.

1262: \label{eqn:mbhmstarfit}

1263: \eeq As before, our fitted values for $c$ and $d$ are intended to

1264: indicate general trends in the evolution rather than to be used as

1265: statistically rigorous measurements.  The observed relationship

1266: \citep{Haring2004} has a slope $c=1.12$ and normalization $d=8.2$.

1267: 

1268: Overall there appears to be only limited evolution in the $M_{\rm BH} -M_{*}$

1269: relation, but there is a slight steepening at $z=2-4$.  To highlight this

1270: trend, we restrict our fits to the high mass end with $M_{*}\ge 5\times

1271: 10^{10}\Msun$ (dashed line in Fig.~\ref{fig:mbhmstar}). In this range, the

1272: relation is significantly steeper, implying slopes $c_s \sim 1.9$ at $z=3-4$

1273: and $\sim 1.5$ at $z\sim 2$.  This is more significant than the evolution

1274: found in the slope of $M_{\rm BH}-\sigma$, and implies that there is some

1275: evolution in the ratio of black hole mass to stellar mass relative to the

1276: local observations. More precisely, systems with $M_{*} \simgt 10^{10} \Msun$

1277: have larger black hole masses at fixed $M_{*}$ than at $z=1$, where the ratio

1278: is in good agreement with the relation observed at the present epoch.  This

1279: trend of an increasing $M_{\rm BH}/M_{*}$ ratio as a function of redshift

1280: appears consistent with the recent measurements of high redshift (up to $z

1281: \sim 3.5$) BH masses and host luminosities by \citet{Peng2006},

1282: as well as the BHFP \citep{Hopkins2007a}. We will

1283: further analyze this effect in \S~\ref{sec:fgas}.

1284: 

1285: \begin{table}

1286: \begin{center}

1287: \caption{Parameters of best-fit $M_{\rm BH} -M_{*}$ relations}

1288: \label{tab:mbhmstar}

1289: \begin{tabular}{ccccc}

1290: \hline\hline\\

1291: $z$  &  slope $c$ & normalization $d$ & scatter $\Delta$ & c$_{s}$\\

1292:  \hline\\

1293: 1 ....... &  $1.18\pm0.02$ & $8.10\pm0.03$ & 0.03 &1.2 \\

1294: 2 ....... &  $1.23\pm0.03$ & $8.09\pm0.03$ & 0.04 & 1.5 \\

1295: 3 ....... &  $1.25\pm 0.04$ & $8.04\pm0.04$ & 0.06 & 1.9 \\

1296: 4 ....... &  $1.30\pm 0.05$ & $8.04\pm0.05$ & 0.07 &1.9 \\

1297: 5 ....... &  $1.17\pm 0.10$ & $7.90\pm0.10$ & 0.14 & 2.0 \\

1298: 6.5 ..... &  $1.01\pm 0.22$ & $7.78\pm0.25$ & 0.34 & -- \\

1299: \hline\\

1300: \end{tabular}

1301: \end{center}

1302: \vspace{-1cm}

1303: \end{table}

1304: 

1305: \subsection{Evolution of $\sigma$, gas fraction, and $R_e$,

1306:  at fixed stellar host mass}

1307: \label{sec:fgas}

1308: We now analyze some of the physical properties of the host galaxies

1309: and their evolution with redshift to investigate the physical origin

1310: for the trends we have found in the $M_{\rm BH}-\sigma$ and $M_{\rm

1311: BH} -M_{*}$ relations.  Figure~\ref{fig:sigma_mstar} shows the stellar

1312: velocity dispersion $\sigma$ versus the stellar mass $M_{*}$ for each

1313: galaxy as a function of redshift. The dotted lines and the solid line

1314: (for $z=1$) show our best-fit relations. At a fixed $M_{*}$, the

1315: velocity dispersion $\sigma$ increases with increasing redshift. This

1316: is consistent with the results from the merger remnants in

1317: \citet{Robertson2006a} and reflects changes in the structural

1318: properties of stellar spheroids, which become smaller towards higher

1319: redshift.

1320: 

1321: In Figure~\ref{fig:fgasre}, we show the projected half mass radius $R_e$, the

1322: cold gas fraction $f_{\rm gas}$ (a proxy for the disk gas fraction), and the

1323: $\left<M_{\rm BH}/M_{*}\right>$ ratio as a function of redshift. We show

1324: results for systems with $M_{*} \ge 3 \times 10^{10} \Msun$ (dashed lines) and

1325: $M_{*} \ge 6 \times 10^{10} \Msun$ (solid lines). We find higher cold gas

1326: fractions in higher redshift systems with $f_{gas} \propto

1327: (1+z)^{1.5}$. The increased gas content, dissipation rate in high redshift

1328: progenitor is consistent with them being more centrally concentrated, as

1329: shown by the trend of decreasing $R_{e}$ and increasing $\sigma$ with

1330: increasing redshift, leading to larger ratios $M_{\rm BH} /M_{*}$ at high

1331: redshifts, with $M_{\rm BH} /M_{*} \sim (1+z)^{0.5}$, at least over these

1332: ranges of $M_{*}$. Larger values of $\sigma$ at fixed $M_{*}$ at high redshift

1333: are also consistent with an inverse overall trend (e.g.; $(1+z)^{-0.2}$) in

1334: the evolution of the $M_{\rm BH}-\sigma$ relation.

1335: 

1336: The above results are consistent with the general picture for a merger-driven

1337: quasar growth we have developed based on our galaxy merger

1338: simulations~\citep{DiMatteo2005, Springel2005a, Robertson2006a, Hopkins2005,

1339: Hopkins2006a, Hopkins2007a} , which also implies that the $M_{\rm BH}-\sigma$

1340: and $M_{\rm BH}-M_{*}$ relations are connected to each other. This connection

1341: is most clearly demonstrated by the `black hole fundamental plane' (BHFP)

1342: relation discussed by \citet{Hopkins2007a}, which arises from the joint

1343: formation process of spheroids and massive black holes in mergers. Our

1344: cosmological simulations also should agree with the BHFP if most of the BH

1345: growth is associated with mergers.  In Figure~\ref{fig:fplane}, we show our

1346: simulation measurements for the BHFP and compare it to the the best-fit from

1347: \citet{Hopkins2007a}, shown as a solid line.  We find very good agreement with

1348: the predictions for the BHFP by \citet{Hopkins2007a}, in both of its

1349: formulations, i.e.~for $R_e$ and $\sigma$ or $M_{*}$ and $\sigma$, although

1350: the former shows a larger scatter than the latter, owing to the noisy

1351: measurements of $R_e$ in the cosmological simulation.

1352: 

1353: 

1354: \section{The first and the most massive black holes}

1355: \begin{figure*}

1356: \begin{center}

1357: \resizebox{17.0cm}{!}{\includegraphics{f13.ps}}\vspace*{-1.8cm}

1358: \end{center}

1359:  \caption{Individual mass assembly and accretion rate histories for

1360: the six most massive black holes and two intermediate mass BHs (chosen

1361: randomly) in the {\it BHCosmo} run. The top panel shows the black hole

1362: mass as a function of redshift while the bottom four panels give the

1363: detailed accretion rate history for the black holes with the

1364: corresponding colors in the top panel. The first supermassive black

1365: hole forming in the {\it BHCosmo} run, and the most massive one at the

1366: end of the run, are shown with thicker lines in pink and blue,

1367: respectively. From the bottom panels we see that phases of high

1368: Eddington accretion occur at different times of the history of

1369: different black holes.

1370: \label{fig:accrhistD6}}

1371: \end{figure*}

1372: In the previous sections we have discussed the predictions from our

1373: simulation for the global history of black hole mass assembly in

1374: galaxies from the high redshift Universe to today. We have compared

1375: the predictions for the evolution of the black hole mass and accretion

1376: rate density to the history of the star formation rate density and

1377: discussed the growth of the black hole mass function.  We have also

1378: discussed the $M_{\rm BH} - \sigma$ and $M_{\rm BH} -M_{*}$

1379: relationships as a function of time, and examined which physical

1380: properties drive their cosmological evolution.

1381: 

1382: However, our simulation methodology not only allows us to make statistical

1383: statements about the black hole population. Rather, it can also be used to

1384: study the detailed growth history of individual black holes, from the moment

1385: they are seeded to today, which provides a particularly powerful way to follow

1386: the evolution of black holes over cosmic time. The gas which fuels black hole

1387: activity ultimately has its origins in the intergalactic medium, draining

1388: along filaments into forming galaxies. Because of this, BH radiative histories

1389: are directly linked to the formation of large-scale structure in the Universe,

1390: from supercluster scales down to the immediate environment of host galaxies.

1391: Being able to follow these large-scale processes and their impact on

1392: individual black holes self-consistently is the key to a qualitatively better

1393: level of understanding. For example, we have the tools to examine what turns a

1394: small black hole into a supermassive one at $z=6$, or whether some BHs grow

1395: hardly at all after this initial phase (as we shall see below in some

1396: examples). We can ask how quasar lifetimes are related to their clustering,

1397: how important BH mergers versus gas accretion are in the cosmological growth

1398: of BH mass, what BH light curves look like in detail over a Hubble time,

1399: and how specific outbursts correlate with galaxy and cluster merger and

1400: accretion events.

1401: 

1402: In future work we will return to address such questions in more detail; for

1403: example in \citet{Colberg07} we present an extensive study of BH environments

1404: as the universe evolves.  For now we will show that there are some striking

1405: inferences to be drawn from following the detailed histories of even a few

1406: black holes.  As an example we choose to focus on the formation and fate of

1407: the first large black holes that form in our simulation, as well as the

1408: properties of their hosts and their descendants. This allows us to identify

1409: the prerequisite conditions for the growth of supermassive black holes already

1410: in the early universe.

1411: 

1412: 

1413: \begin{figure*}

1414: \begin{center}

1415: \resizebox{17.0cm}{!}{\includegraphics[angle=270]{f14.ps}}

1416: \end{center}

1417:  \caption{Time evolution of the environment around the host galaxy of

1418: the first supermassive massive black hole at $z=6$ (bottom panels),

1419: and of the most massive BH at $z=1$ (top panels) in the {\it BHcosmo}

1420: run. The location and masses of the supermassive BHs are marked by

1421: arrows of different size, as labeled. While the bottom system hosts

1422: the most massive black hole at high redshift, it does not end up

1423: hosting also the most massive black hole at the center of the largest

1424: galaxy at low redshift. Instead, the system shown in the top panels

1425: overtakes it in growth at intermediate redshifts, when it is formed in

1426: the highest density region in the simulation, which is a protocluster

1427: region.

1428:     \label{fig:2bigbhs}}

1429: \end{figure*}

1430: 

1431: Observations of luminous SDSS quasars at redshifts as high as $z \sim

1432: 6$ \citep{Fan2003} present a number of challenges for models of high

1433: redshift quasar and galaxy formation. Their low space density suggests

1434: that they reside in the rarest dark matter density peaks at this early

1435: epoch, yet the apparent lack of companion galaxies in the field has

1436: been used to argue that these quasars reside in far more common halos

1437: \citep{Carilli2004, Willott2005}. Extensive follow-up observations of

1438: the highest redshift ($z=6.42$) quasar, SDSS J11148+5251, are starting

1439: to constrain the properties of its host galaxy \citep{Walter2004}.  CO

1440: observations indicate a relatively small stellar spheroid, however,

1441: the existence of CO, iron and carbon emission indicates a heavily

1442: enriched ISM and vigorous star formation \citep{Maiolino2005}. In any

1443: case, rapid growth is clearly required to produce large supermassive

1444: black holes and spheroids in the $\sim 800$ million years available to

1445: $z\sim 6$.  Unsurprisingly, the type of object associated with the

1446: 'first quasar' is therefore still a matter of debate.

1447: 

1448: %Because of the rarity - models assume 10^13 mass halos - millenium run

1449: %pick the halo here we do not pick --- this is the major difference...

1450: % WRITE SOME OF THIS STUFF>>..

1451: 

1452: The relatively small volume of our simulation prevents us from directly

1453: addressing the problem of the `first quasars', simply because their space

1454: density is so low. Properly identifying one of the rare host systems requires

1455: at the very least box-sizes of $500\,h^{-1}{\rm Mpc}$

1456: \citep{SpringelMillennium2005}, or more adequately $1\,h^{-1}{\rm Gpc}$

1457: \citep{Li2007}.  Nevertheless, our simulation allows us to examine the nature

1458: of the environments and the hosts where exponential growth of BHs can first

1459: occur at early times. This has some relevance to the problem of where the

1460: first quasars are likely to form.

1461: 

1462: In order to retrieve the required information for all the black holes in the

1463: simulation we have constructed full merger trees for each of them. This allows

1464: us to track the growth and accretion history of individual black holes of

1465: different masses, make detailed lightcurves from their accretion histories,

1466: and study the properties of their host galaxies.

1467: 

1468: 

1469: \subsection{Individual BH mass and accretion rate as function of $z$}

1470: In Figure~\ref{fig:accrhistD6}, we show the accretion histories of a

1471: number of black holes in the {\it BHCosmo} run (top panel) and some

1472: examples of their corresponding accretion rate histories in the bottom

1473: four panels. The sample in the top panel consists of the six most

1474: massive black holes in the simulation and two randomly chosen ones

1475: with an intermediate mass at $z=1$. The four bottom panels show the

1476: accretion rate in units of Eddington for four black holes with

1477: corresponding line colors, including the most massive black hole at

1478: $z=6$, the most massive and second most massive at $z=1$, and finally

1479: one of the intermediate mass ones.  Note that this is a small

1480: sub-sample of the several thousand black holes in the simulation. We

1481: focus here on the evolution of the most massive ones, and in

1482: particular on their early formation and the fate of their descendants.

1483: In \citet{Colberg07}, we study a complete sample of individual BH

1484: histories as a function of environments for the entire simulation.

1485: 

1486: The first interesting result shown in Figure~\ref{fig:accrhistD6} is

1487: that even in this small volume conditions exist that are conducive for

1488: exponential black hole growth, as required by the presence of quasars

1489: at $z\sim 6$ with the large observed masses. The evolution of the mass

1490: of the largest black hole in our simulation at early times is shown by

1491: the thick pink line.  The corresponding accretion rate history (in the

1492: same color) shows a rapid succession of numerous phases of high

1493: accretion at the critical Eddington rate, between $ 5 \simlt z \simlt

1494: 7.5$. Below $z\sim5$, the black hole is drastically more quiescent,

1495: except for a couple of sporadic Eddington phases at around $z\sim 2$.

1496: Interestingly, this first most massive black hole at $z \sim 6$ is not

1497: the most massive one at $z=1$. Instead, it is `overtaken' in growth by

1498: another black hole shown with the blue thick line, which is the most

1499: massive black hole at the end of the simulation.  The Eddington growth

1500: phases for this objects start at $z < 6$ and extend all to way to

1501: $z\sim 3.5$. By $z \sim 4.5$, the black hole mass in this system

1502: catches up with that of the first massive black hole, and then keeps

1503: growing exponentially to $z\sim4$.

1504: 

1505: In Figure~\ref{fig:2bigbhs}, we show projected gas density maps of the

1506: host galaxies and the surrounding structures (at redshifts $z=7.5$,

1507: $4$, and $1.6$) for two of these black holes, the one that is most

1508: massive at $z=1$, and the one that is most massive at $z=6$.  This

1509: provides good clues for the origin of the evolutionary difference

1510: between these two systems. Although the hosts of both black holes at

1511: $z=7.5$ are halos of similar mass, $\sim 10^{10} \Msun$, one of them

1512: lies at the intersection of three small filaments, leading to

1513: efficient gas cooling and a high star formation of $\sim 100

1514: \Msun$yr$^{-1}$. The host galaxy grows rapidly until $z=6$ but then

1515: its gas supply dwindles, and at $z\sim4$ it is overtaken by the other

1516: black hole.  This latter one lies on a very massive filament, which

1517: grows even more vigorously at this later time than its host galaxy

1518: (which has a SFR of $\sim 1000 \Msun$yr$^{-1}$ at $z\sim 2$). In this

1519: way a large stellar spheroid is produced around the BH, which

1520: eventually will end up at the center of a rich cluster of galaxies.

1521: Our results therefore show that a massive black hole that is found in

1522: the cD of large galaxy cluster at late times was not necessarily the

1523: most massive one at $z=6$, as it has been often assumed in the

1524: literature \citep[e.g.][]{SpringelMillennium2005}. Since the growth

1525: history of black holes is intertwined with the non-linear processes of

1526: structure formation, individual growth histories of BHs can be complex

1527: and need not preserve the rank order in a group of BHs that start out

1528: with similar masses.

1529: 

1530: 

1531: 

1532: \begin{figure}

1533: \hspace{-0.5cm}

1534: \resizebox{8.5cm}{!}{\includegraphics{f15.ps}}

1535:  \caption{Comparison of individual black hole mass histories in the

1536: {\it BHCosmo} run (blue lines, D6) and in the larger simulation volume

1537: of the E6 run (red lines). The growth of the first supermassive black

1538: holes at $z\sim 6-7$ is more widespread in the larger volume of the

1539: E6. The 'catch-up' of larger black holes that form later and in higher

1540: density regions (see text) can also be seen, and is comparable to the

1541: case shown in Fig.~\ref{fig:accrhistD6}.

1542:     \label{fig:accrhistD6E6}}

1543: \end{figure}

1544: 

1545: \begin{figure*}

1546: \begin{center}

1547: \resizebox{17.0cm}{!}{\includegraphics{f16.ps}}%

1548: \vspace*{-0.8cm}

1549: \end{center}

1550:  \caption{Black hole merger history trees shown as a function of

1551: redshift ($y$-axis) and position (along the $x$-axis), for the most

1552: massive black hole at $z=1$ in the {\it BHCosmo} run.  The full tree

1553: is shown in the top left panel, while the other three panels split up

1554: the tree according to accretion rate in units of Eddington, as

1555: indicated by the colors. The black hole mass is given by the size of

1556: the symbols, as labeled.

1557: \label{fig:tree1}}

1558: \end{figure*}

1559: 

1560: 

1561: \begin{figure*}

1562: \begin{center}

1563: \resizebox{17.0cm}{!}{\includegraphics{f17.ps}}

1564: \vspace*{-0.8cm}

1565: \end{center} 

1566: \caption{Black hole merger history tree shown as a function of redshift

1567: ($y$-axis) and position ($x$-axis) for the first high-redshift massive black

1568: hole that forms in the simulation. The full tree is shown in the top left

1569: panel, while the other three panels split up the tree according to accretion

1570: rate in units of Eddington, as indicated by the colors. The black hole mass is

1571: given by the size of the symbols, as labeled. The evolution of this system is

1572: qualitatively very different from the one shown in Fig.~\ref{fig:tree1} Note

1573: that discreteness in the outputs for black holes that remain inactive cause

1574: the gap in the rightmost branch of this tree.

1575:     \label{fig:tree6}}

1576: \end{figure*}

1577: 

1578: Further confirmation of this result from the {\it BHCosmo} simulation

1579: come from our E6 run (see Table~\ref{tab:simul}), which includes the

1580: same physics and model parameters as the {\it BHCosmo} simulation but

1581: samples rare halos better, thanks to its larger box-size of $50

1582: \,h^{-1}{\rm Mpc}$ on a side.  In this larger volume, there are a

1583: handful of more examples of early exponential black hole growth

1584: between $8\simlt z \simlt 6$, as shown by the red black hole mass

1585: growth curves in Figure~\ref{fig:accrhistD6E6}.  Consistent with our

1586: previous finding, there are again black holes that 'catch-up' in

1587: growth to the most massive black hole at some earlier time. We note

1588: that we have run this larger volume simulation with the specific

1589: purpose of checking the {\it BHCosmo} results for the first

1590: supermassive black holes by obtaining better statistics for large

1591: halos at $z\simgt 6$. For this reason, this simulations was not

1592: evolved beyond $z=4$.

1593: 

1594: 

1595: 

1596: \subsection{BH merger trees for the first and the most massive black hole}

1597: Figures~\ref{fig:tree1} and \ref{fig:tree6} show two example black

1598: hole merger trees, one for the most massive black hole at $z=1$, and

1599: one for the $z=1$ descendant of the first supermassive black hole at $z

1600: \sim 6$. All progenitor black holes that merge together to build up

1601: the final BH are included in the trees.  The black hole masses along

1602: the tree are represented by different sizes of the circles, while the

1603: color encodes their accretion rate in units of the Eddington rate.  In

1604: both figures, the top-left panel shows the full BH merger tree,

1605: whereas the top-right and bottom two panels split the tree according

1606: to accretion rate of the BHs.

1607: 

1608: Inspection of these two BH trees immediately reveals a complex and

1609: rich merger history for the most massive black holes at $z=1$

1610: (Fig.~\ref{fig:tree1}), as opposed to the comparatively isolated

1611: evolution of the most massive object at $z=6$, which only features

1612: three major branches for its tree (Fig.~\ref{fig:tree6}). The large

1613: majority of the mass in the latter case is built up by early phases of

1614: critical gas accretion at $z \simlt 6$. On the other hand, in

1615: Figure~\ref{fig:tree1}, Eddington accretion phases are spread out

1616: along many branches of the tree and cover a much larger range of

1617: redshifts.  At redshifts $z < 3$, the progenitors are preferentially

1618: accreting at low Eddington rates, so that the residual growth of the final

1619: black hole mass increasingly occurs from ``dry'' mergers with other

1620: black holes, consistent with our earlier results for the global black

1621: hole growth.

1622: 

1623: 

1624: \section{Summary and discussion\label{sec:conclusions}}

1625: In this paper we have investigated the coupled formation and evolution

1626: of black holes and galaxies using state-of-the-art cosmological

1627: hydrodynamic simulations of the $\Lambda$CDM model. For the first

1628: time, we have incorporated black hole growth and associated feedback

1629: from quasar activity self-consistently in cosmological hydrodynamic

1630: simulations. Our approach has been based on the methodology recently

1631: developed and tested in simulations of galaxy mergers

1632: \citep{DiMatteo2005, Springel2005a}. In this initial paper we have

1633: focused on investigating the model predictions for (i) the global

1634: history of black hole mass assembly and its relation to the history of

1635: star formation, from high redshift to the present, for (ii) the

1636: evolution of the BH mass function and accretion rate function and its

1637: connection to the observational ``downsizing'' phenomenon, for (iii)

1638: the correlations between black hole mass, velocity dispersion, and

1639: stellar mass of host galaxies, and finally, for (iv) the formation and

1640: fate of the first quasars and the properties of their hosts.

1641: 

1642: An important and highly encouraging first result has has been that the

1643: cosmological black hole mass density $\rho_{\rm BH}$ predicted by our

1644: high-resolution simulation reproduces the locally measured value and

1645: its extrapolation to higher redshift ($z< 2.5$), inferred from

1646: integrating the X-ray luminosity functions \citep{YuTremaine2002, Marconi2004,

1647: Shankar2004}, or, more directly, the bolometric quasar

1648: luminosity function \citep{Hopkins2007b}.

1649: We predict a steep evolution of $\rho_{\rm BH}$ as a

1650: function of redshift, with a rise that is more rapid than that of the

1651: star formation rate. At $z > 3$, $\rho_{\rm BH}/\rho_{*} \propto

1652: (1+z)^{-3}$ whereas there is at most weak evolution in the ratio

1653: $\rho_{\rm BH}/\rho_{*}$ at $z \lesssim 3$. Similarly, whilst the star

1654: formation rate density $\dot\rho_{*}$ broadly tracks the BHAR density

1655: $\dot\rho_{\rm BH}$ below $z\sim 2.5-3$, their ratio evolves steeply

1656: at higher redshifts, as $\dot\rho_{\rm BH}/ \dot\rho_{*} \propto

1657: (1+z)^{-4}$.  The SFR density peaks earlier than the BHAR density and

1658: exhibits a more gradual evolution with redshift compared to the BHAR.

1659: Only at redshifts below the peak of the BHAR, the BHAR and SFR rate

1660: densities start tracking each other.

1661: 

1662: Our results for the evolution of the BHAR density are broadly

1663: consistent with constraints obtained by \citet{Hopkins2007b} from a

1664: comprehensive analysis of a large sample of observational data sets,

1665: which allowed them to synthesize the evolution of the bolometric

1666: quasar luminosity density with redshift and to show that the

1667: luminosity density indeed peaks at $z\sim 2-3$, with a sharp drop

1668: towards higher redshifts.

1669: 

1670: 

1671: We have shown that the growth of the black hole mass function shows

1672: signs of `anti-hierarchical' behavior, or `downsizing'.  The high mass

1673: end of the BH mass function is established at comparatively high

1674: redshift and then significantly slows down in evolution below $z \sim

1675: 2$, where only the abundance of intermediate mass BHs is still growing

1676: appreciably.  We have found more direct evidence for black hole

1677: ``downsizing'' by inspecting the distribution of accretion rates, in

1678: unis of the critical Eddington rate. The accretion rate function

1679: shifts from a narrow distribution dominated by high Eddington rates at

1680: high redshift to a broad distribution with a small fraction of

1681: Eddington accretion for low redshifts ($z < 3$). High mass black holes,

1682: forming in the high density peaks at high redshift are built up

1683: rapidly by vigorous accretion. However, the effects of gas depletion

1684: and AGN feedback drive a strong decline in the accretion rate of

1685: massive black holes towards late times, while during these later times

1686: the peak of the accretion activity shifts to progressively smaller

1687: mass scales.

1688: 

1689: %Gas depletion in large halos and suppression of further cooling by feedback

1690: %processes are likely to play a role in these trends.

1691: 

1692: The BH masses of our simulated galaxies are strongly correlated with the

1693: stellar velocity dispersions and stellar masses of their host galaxies, and

1694: the correlations agree remarkably well with the local $M_{\rm BH} - \sigma$

1695: and $M_{\rm BH}-M_{*}$ relationships, over a large dynamic range. We

1696: previously showed with simulations of isolated galaxy mergers

1697: \citep{DiMatteo2005, Robertson2006a, Hopkins2007a} that our AGN feedback

1698: prescription leads to a self-regulated BH growth that can explain the 

1699: $M_{\rm BH} - \sigma$ relation. 

1700: It is highly reassuring that the same simulation

1701: model, with an unmodified feedback efficiency $\epsilon_f$ reproduces the

1702: observed $M_{\rm BH} - \sigma$ in full cosmological simulations as well. We

1703: emphasize that the free parameter $\epsilon_f$ sets the normalization of the

1704: obtained relationship, but the slope and scatter of the relation obtained from

1705: the simulations are not adjustable and a non-trivial consequence of the

1706: self-regulated BH growth.

1707: 

1708: Our simulation also suggests a weak evolution with redshift of the

1709: normalization of the $M_{\rm BH}-\sigma$ relation, as $M_{\rm BH}

1710: /\sigma \sim (1+z)^{-0.2}$. However, we find that this evolutionary

1711: trend is sensitive to range of masses being probed. When we focus on

1712: the better resolved more massive systems, there appears to be some

1713: mild evolution in the slope of the BH scaling relations at high

1714: $\sigma$ and for high $M_{*}$. In particular, we find a trend of

1715: increasing $M_{*}/ M_{\rm BH}$ with redshift, in agreement with recent

1716: direct estimates of the BH to host stellar mass ratio at high redshift

1717: by \citet{Peng2006}. This trend is accompanied by an increase in the

1718: cold gas fraction in the host galaxy, and a smaller $R{_e}$ at a fixed

1719: stellar mass.  These results are consistent with the suggestion that

1720: the $M_{\rm BH} - \sigma$ and $M_{\rm BH}-M_{*}$ relations are

1721: projections of a more basic correlation,

1722: the black hole fundamental plane \citep{Hopkins2007a}, 

1723: $M_{\rm BH} \propto \sigma^3

1724: R_{e}^{0.5}$, and the interpretation that gas-rich systems at high

1725: redshift produce more dissipative mergers that lead to more

1726: concentrated BH hosts.

1727: 

1728: 

1729: Our findings for the BH scaling relations appear fully consistent with a

1730: scenario where quasar activity is driven by galaxy mergers, as suggested by

1731: our simulations of isolated mergers \citep{DiMatteo2005, Springel2005a,

1732: Robertson2006a, Hopkins2006a, Hopkins2007a}. While gas is available for star

1733: formation over a large range of mass scales, it can only gets to central

1734: regions of galaxies in large amounts as a result of the inflows that accompany

1735: major mergers. At the same time, the mergers produce spheroids, which together

1736: with the growth-limiting quasar feedback establishes the $M_{\rm BH} - \sigma$

1737: relationship.  The detailed time evolution of the quasar activity in an

1738: individual galaxy depends on the angular momentum of the gas and the disk size

1739: and morphology. It is clear that the limited spatial resolution of our

1740: cosmological simulation is a severe limitation on how faithfully this can be

1741: tracked in our calculation, but because the final BH masses in merger remnants

1742: are robustly predicted by our simulation methodology even at coarse resolution

1743: this does not seriously affect the quantities we study here.

1744: 

1745: Our simulated volumes are too small to directly check if a population

1746: of supermassive black holes as massive the $z\sim 6$ Sloan quasars can

1747: form at high redshift. However, we do find that our simulation model

1748: produces black holes at early times that spend a large fraction of

1749: their time at high accretion rates close to their Eddington rates. We

1750: therefore predict the presence of massive black holes at high

1751: redshift. Using high-resolution simulations of multiple mergers that

1752: were constructed to represent a high redshift merging tree of one of

1753: the most massive protocluster regions in a $(1\,h^{-1}{\rm Gpc})^3$

1754: volume, \citet{Li2007} have recently made the match with the Sloan

1755: quasar much more explicit. They were able to show that the gas-rich

1756: mergers expected in this rare overdensity can indeed grow a BH of mass

1757: $\sim 10^{9}\,{\rm M}_\odot$ early enough.

1758: 

1759: We have also found that the first supermassive black hole in our simulation

1760: forms in relative isolation, as a result of strong gas inflows and merging at

1761: the intersection of large-scale filaments. However, by following this system

1762: to lower redshift it turned out that it did not evolve into the most massive

1763: black hole today. Instead, it was overtaken in growth by a black hole that

1764: acquired most of its mass in major mergers in a high density region at lower

1765: redshift. Such a change in the relative rank of BHs as a function of time

1766: appears quite generic, as we explicitly checked with a simulation of larger

1767: volume. This clearly complicates attempts to directly link high redshift

1768: progenitor systems to low redshift descendants. As we have demonstrated, the

1769: cosmological simulation methodology we introduced here provides however an

1770: excellent tool for studying the evolution of the cosmic BH evolution. In the

1771: present study, we focused only on the radiatively efficient accretion mode of

1772: AGN, which is associated with quasar activity. For following AGN activity also

1773: in clusters of galaxies, and hence down to $z=0$ in large volumes, we also

1774: need to account radio activity, which becomes important in very massive halos

1775: at low redshift. In \citet{Sijacki2007} we present an extension of our

1776: simulation methodology that accounts for this physics, and we discuss results

1777: obtained with this unified model for AGN feedback both for clusters of

1778: galaxies and cosmological boxes.

1779:  

1780: Our cosmological approach to black hole formation enables us to follow the

1781: fuel for black hole activity from its ultimate source, the early intergalactic

1782: medium, and so link the large-scale structure and environments of black holes

1783: directly with their growth. As a result, countless different avenues of

1784: research are opened up and we plan to explore many in future work. For

1785: example, in \citet{Colberg07} we use merger trees constructed for every black

1786: hole in the simulation to carry out a systematic study of their environments

1787: and histories and to determine the amount of black hole growth owing to

1788: mergers versus gas accretion; the latter should be testable with upcoming

1789: gravitational wave experiments. Future issues which will be addressed include

1790: the nature of AGN clustering and its evolution and its relation to the

1791: underlying dark matter clustering, predictions for quasar lifetimes and

1792: luminosity functions, and how feedback from AGN may manifest itself on

1793: large-scales and can be detected through the Sunyaev-Zeldovich effect. With

1794: our new approach we can address the question of how the first black hole

1795: formed and grew we will be able to make predictions for the ionizing

1796: background owing to the first miniquasars, an important but currently uncertain

1797: ingredient in current models of reionization.

1798: 

1799: 

1800: \acknowledgements

1801: We thank Rupert Croft for many discussions and reading the manuscript.

1802: The simulations were performed at Carnegie Mellon University and the

1803: Pittsburgh Supercomputer Center (PSC). 

1804: This work has been supported in part through NSF AST-0607819.

1805: 

1806: %We thank Karen Tozzi, for design...

1807: 

1808: \bibliographystyle{apj}

1809: \bibliography{ms}

1810: 

1811: 

1812: \end{document}

1813: 

1814: 

1815: