1: \documentclass[usenatbib]{mn2e}

2: \usepackage{graphicx}

3: 

4: \begin{document}

5: \topmargin-1cm

6: 

7: %Make my life significantly easier

8: \newcommand\approxgt{\mbox{$^{>}\hspace{-0.24cm}_{\sim}$}}

9: \newcommand\approxlt{\mbox{$^{<}\hspace{-0.24cm}_{\sim}$}}

10: \newcommand{\be}{\begin{equation}}

11: \newcommand{\ee}{\end{equation}}

12: \newcommand{\bea}{\begin{eqnarray}}

13: \newcommand{\eea}{\end{eqnarray}}

14: \newcommand{\lexp}{\mathop{\langle}}

15: \newcommand{\rexp}{\mathop{\rangle}}

16: \newcommand{\rexpc}{\mathop{\rangle_c}}

17: \newcommand{\nbar}{\bar{n}}

18: \newcommand{\zmax}{Z_{\rm max}}

19: \newcommand{\amass}{10^{13} \,M_\odot}

20: \newcommand{\hmass}{10^{14} \,M_\odot}

21: \newcommand{\mmin}{M_{\rm{min}}}

22: 

23: %for some crazy reason mnras style files don't 

24: %have these key'd in somewhere

25: \newcommand{\apj}{ApJ}

26: \newcommand{\apjl}{ApJL}

27: \newcommand{\apjs}{ApJS}

28: \newcommand{\mnras}{MNRAS}

29: \newcommand{\aap}{AAP}

30: \newcommand{\prd}{PRD}

31: \newcommand{\aj}{AJ}

32: 

33: 

34: \title{Constraints on the Merging Timescale of Luminous Red Galaxies,

35:   Or, Where Do All the Halos Go?}

36: 

37: \author[Conroy et al.]

38: {Charlie Conroy${}^1$\thanks{cconroy@astro.princeton.edu},

39: Shirley Ho${}^1$ and Martin White$^{2}$\\

40: $^{1}$ Department of Astrophysical Sciences, Peyton Hall,

41: Princeton University, Princeton, NJ 08544, USA.\\

42: $^{2}$ Departments of Physics and Astronomy, 601 Campbell Hall,

43: University of California Berkeley, CA 94720, USA.}

44: \date{\today}

45: \maketitle

46: 

47: \begin{abstract}

48:   In the $\Lambda$CDM cosmology dark matter halos grow primarily

49:   through the accretion of smaller halos.  Much of the mass in a halo

50:   of $\hmass$ comes in through accretion of $\sim\amass$ halos.  If

51:   each such halo hosted one luminous red galaxy (LRG) then the

52:   accretion of so many halos is at odds with the observed number of

53:   LRGs in clusters unless these accreted LRGs merge or disrupt on

54:   relatively short timescales ($\sim2$ Gyr).  These timescales are

55:   consistent with classical dynamical friction arguments, and imply

56:   that $2-3$ LRGs have merged or disrupted within each halo more

57:   massive than $\hmass$ by $z=0$.  The total amount of stellar mass

58:   brought into these massive halos by $z=0$ is consistent with

59:   observations once the intracluster light (ICL) is included.  If

60:   disrupted LRGs build up the ICL, then the hierarchical growth of

61:   massive halos implies that a substantial amount of ICL should also

62:   surround satellite LRGs, as suggested by recent observations of the

63:   Virgo cluster.  Finally, we point out that these results are

64:   entirely consistent with a non-evolving clustering strength and halo

65:   occupation distribution, and note that observations of the latter in

66:   fact support the hypothesis that merging/disruption of massive

67:   galaxies does indeed take place at late times.

68: 

69: \end{abstract}

70: 

71: %\keywords{large-scale structure of universe}

72: 

73: \section{Introduction} \label{sec:introduction}

74: 

75: The formation and evolution of massive red galaxies provide a critical

76: testing ground for modern theories of galaxy formation based on

77: hierarchical merging of dark matter halos.  Ongoing growth of massive

78: halos via mergers is a generic feature of hierarchical models, such as

79: cold dark matter (CDM).  However evidence for the ongoing assembly of

80: massive galaxies is at best inconclusive.  Evolution in the galaxy

81: stellar mass and luminosity functions at the massive/luminous end

82: appears quite modest since $z=1$ \citep[e.g.][]{Drory04, Bundy05,

83:   Borch06, Fontana06,Faber06, Willmer06, Brown07, Caputi06, Wake06}

84: though estimates of the merger rate of massive galaxies present a less

85: consistent picture \citep{vanDokkum05,Bell06,Masjedi06, MWhite07}.  If,

86: as theory predicts, massive halos are constantly accreting halos that

87: are themselves hosts of massive galaxies, what is the fate of these

88: accreted galaxies?

89: 

90: Two physical effects can cause satellite galaxies to `disappear' from

91: an observational sample.  Tidal forces acting on a satellite as it

92: orbits in the host halo potential can cause it to \emph{disrupt}.  At

93: the same time, dynamical friction (DF) causes a satellite to lose

94: energy to the background dark matter halo and eventually causes the

95: satellite to sink toward the center and \emph{merge} with the central

96: galaxy of the host halo.  While such notions, and their relevance to

97: the evolution of galaxies within clusters, have been known for decades

98: \citep[e.g.][]{Chandrasekhar43, Ostriker75, Ostriker77, Merritt84},

99: accurate merger times of satellite galaxies have been historically

100: hard to calculate, and are poorly constrained observationally.

101: Unfortunately, the problem cannot at present by circumvented by

102: brute-force simulations due both to severe resolution requirements and

103: the uncertain effects of baryon condensation on the survival of

104: satellite halos \citep[see e.g.][]{Moore99, Klypin99, Diemand04,

105:   Gao04a, Gao04b, Reed05}.  In addition, while DF is usually

106: considered in a collisionless medium (such as dark matter), DF acting

107: in a collisional medium (such as intracluster gas) is stronger

108: (weaker) than in the collisionless case for satellites traveling at

109: supersonic (subsonic) speeds \citep{Ostriker99}.  Observational

110: constraints on the merging timescale of satellites would hence provide

111: valuable insight into this complex dynamical process.

112: 

113: This paper explores observational constraints on the average merging

114: timescale of luminous red galaxies (LRGs).  We assign LRGs to dark

115: matter halos that have grown more massive than $M\sim\amass$ and use

116: an $N$-body simulation to follow their accretion onto larger dark

117: matter halos with $z=0$ mass comparable to observed rich groups and

118: clusters $(M>\hmass)$.  Comparison with the observed multiplicity

119: function of LRGs at $z\sim0.3$ implies that accreted LRGs must merge

120: on timescales comparable to those predicted by Chandrasekhar's formula

121: ($\sim2$ Gyr).  While this may not be surprising, the relative flood

122: of massive halos onto more massive halos implies that a substantial

123: number of LRGs have disrupted over the history of the Universe.

124: Though these numbers may at first glance appear large (on average

125: $2-3$ disrupted LRGs per $z=0$ halo with $M>\hmass$) we show that the

126: total stellar mass brought in by these accreted LRGs is consistent

127: with the observed stellar mass in clusters so long as one counts both

128: observed massive galaxies and the observed intracluster light.

129: 

130: The following sections describe in more detail the salient accretion

131: properties of massive dark matter halos ($\S$\ref{s:accretion}), the

132: inferred merging timescale of LRGs, if LRGs correspond to massive

133: halos ($\S$\ref{s:merge}), and the implied total stellar mass brought

134: into massive $z\sim0$ dark matter halos by these accreted LRGs

135: ($\S$\ref{s:stellar}).  We conclude in $\S$\ref{s:conc}.  Throughout

136: we assume a flat $\Lambda$CDM cosmology with $(\Omega_m,

137: \Omega_{\Lambda}, h,\sigma_8) = (0.25,0.75,0.72,0.8)$, and use a

138: virial mass definition, $M_{200}$, corresponding to the mass contained

139: within a region that has mean density equal to $200\times$ the

140: critical density \citep[see e.g.][]{Evrard07}.

141: 

142: \section{The Accretion History of Massive Dark Matter

143:   Halos}\label{s:accretion}

144: 

145: \begin{figure}

146: \begin{center}

147: \resizebox{3.5in}{!}{\includegraphics{f1.eps}}

148: \end{center}

149: \caption{Fraction of halos at $z=0$ with $M>\hmass$, $f$, that have

150:   absorbed $N_{\rm sat}$ halos with mass $>\amass$.  The thick solid line

151:   is for halos accreted at all epochs, while the dashed (dotted) line

152:   indicates only those halos accreted after (before) $z=0.8$,

153:   i.e. when the Universe was half its present age for our assumed

154:   cosmology.  No distinction is made here between halos that have

155:   dissolved and halos that remain as bound satellites.  On average,

156:   $3.2$ halos with mass $>\amass$ have accreted onto these more

157:   massive halos by $z=0$.}

158: \label{fig:mergers}

159: \end{figure}

160: 

161: A robust expectation of a Universe dominated by cold dark matter is

162: the hierarchical growth of structure, and in particular the growth of

163: dark matter halos via the accumulation of smaller halos.  An

164: illustrative example of the accretion history of dark matter halos is

165: shown in Figure \ref{fig:mergers}.  There we plot the multiplicity

166: function for halos more massive than $10^{14}M_\odot$, i.e.~the

167: distribution of the number of halos with $M>\amass$ that have accreted

168: onto halos with $z=0$ mass greater than $10^{14} M_\odot$ (see the

169: appendix for details regarding the simulation used to compile this

170: information).  We refer to these more massive halos as ``hosts''

171: throughout.  There are 2339 such hosts in our simulation at $z=0$,

172: corresponding to a number density of $\sim 2 \times 10^{-5}$

173: Mpc$^{-3}$.  Note that these distributions are not symmetric.  On

174: average, halos more massive than $\hmass$ have been bombarded by $3.2$

175: halos with mass $>\amass$ over a Hubble time.  Here we do not

176: distinguish between halos that were accreted directly onto the host

177: halo and those that were accreted onto an intermediate halo that later

178: accreted onto the host, although such a distinction will be utilized

179: in the following sections.  The accretion of such massive halos is

180: roughly equally important both at low and high redshift: on average

181: two such halos have been accreted at $z<0.8$ (the Universe was about

182: half its present age at $z=0.8$ for the cosmology assumed herein).

183: 

184: Halos more massive than $\sim\amass$ are expected to contain at least

185: one massive galaxy at their center \citep{Zehavi05} even at moderate

186: redshifts \citep[e.g.][]{Yan03,Yan04,Coil06b}.  From Figure

187: \ref{fig:mergers} we are lead to the conclusion that, \emph{in the

188:   absence of mergers,} observed clusters with $M>\hmass$ should

189: contain on average $3.2$ massive galaxies (and certainly more if

190: accreted halos of lower masses also contain massive galaxies), with a

191: significant tail toward much larger numbers.  However, reproducing the

192: observed clustering of massive galaxies at $z\sim 0$ \citep{Zehavi05b}

193: would require closer to $1.2$ galaxies in such halos in our

194: simulation, in agreement with other work \citep{Masjedi06,

195:   Kulkarni07}.  While these statements are only qualitative, they will

196: be confirmed in the more quantitative discussion that follows.  In

197: order to reconcile the accretion properties of halos with

198: observations, we are thus lead to consider the fate of these massive

199: halos and the galaxies within them.

200: 

201: \section{The Merging Timescale of LRGs}\label{s:merge}

202: 

203: LRGs are massive galaxies with very little ongoing star-formation;

204: they thus constitute the tip of the red sequence.  They have uniform

205: spectral energy distributions marked by numerous features and hence

206: their redshifts are relatively straightforward to estimate

207: photometrically \citep[][redshift uncertainties are $\delta

208: z\sim0.03$]{Padmanabhan05}.  Modeling of their spectral energy

209: distributions has lead to the conclusion that these galaxies formed

210: the bulk of their stars at $z>2$ \citep[e.g.][]{Trager00, Jimenez06,

211:   Thomas05}, and hence are expected to evolve largely

212: dissipationlessly at $z<1$.  Their clustering strength is large,

213: suggesting that they live in massive dark matter halos $M>\amass$

214: \citep{Zehavi05b}.

215: 

216: Recently, \citet{Ho07} has measured the multiplicity function of LRGs

217: extracted from the Sloan Digital Sky Survey \citep[SDSS;][]{DR4} for

218: 43 clusters over the redshift range $0.2<z<0.5$.\footnote{The primary

219:   spectral feature used to measure photometric redshifts of LRGs is

220:   the $4000\AA$ break; at $z<0.2$ this feature moves out of the SDSS

221:   bandpass filters.  Hence our sample is restricted to $z>0.2$.

222:   Although Ho et al.'s sample extends to $z\sim0.6$, for our purposes

223:   we truncate it at $z=0.5$ to limit the amount of possible evolution

224:   within the sample.} Cluster virial masses were derived from

225: \emph{ROSAT} $X$-ray data and range from

226: $10^{14.1}<M_{200}/M_\odot<10^{14.9}$.  The average stellar mass of

227: the LRGs in this sample is $M_{\rm star}=10^{11.6} M_\odot$, as

228: determined from a color-based stellar mass estimator \citep[for a

229: Chabrier IMF;][]{Bell03}.  These clusters contain on average $2.5$

230: LRGs.  The reader is referred to \citet{Ho07} for further details

231: regarding these observations.

232: 

233: In this sample there are approximately five clusters that contain no

234: LRGs\footnote{We say approximately because Ho et al. statistically

235:   remove interlopers based on photometric redshift uncertainties and

236:   hence clusters contain a non-integer number of LRGs.}.  When

237: plotting the observed multiplicity function below we include both the

238: one reported in Ho et al. and one where clusters with $N<1$ are

239: artificially assigned $N=1$.  This is done to afford a more robust

240: comparison to our simple model (see below) where we \emph{assume} that

241: each cluster halo contains at least one LRG at the center.  As

242: discussed below, our conclusions are insenstive to this distinction.

243: 

244: \begin{figure}

245: \begin{center}

246: \resizebox{3.5in}{!}{\includegraphics{f2.eps}}

247: \end{center}

248: \caption{The multiplicity function of observed LRGs (\emph{dashed line

249:     joining points}) and accreted dark matter halos for two accreted

250:   mass thresholds: $M >\amass$ (\emph{solid line}) and $M >2\times

251:   10^{13} M_\odot$ (\emph{dotted line}).  The observed curves diverge

252:   at $N<1$ due to whether or not we artifically assign an LRG at the

253:   center of observed clusters with $N<1$ (see text for details).

254:   \emph{Top Four Panels:} Accreted halos are assumed to have disrupted

255:   after a time $\tau$, shown in the upper right corner of each panel.

256:   It is apparent that if LRGs can be identified with halos of mass

257:   $>\amass$ then they must on average merge within $\sim2$ Gyr.

258:   \emph{Bottom Panel:} Accreted halos are assumed to have merged after

259:   a dynamical friction timescale (see Equation \ref{eqn:df}).  This

260:   panel also includes halos more massive than $M>5\times10^{12}

261:   M_\odot$ (\emph{dot-dashed line}) for comparison.}

262: \label{fig:mf}

263: \end{figure}

264: 

265: 

266: The dark matter halo accretion history of massive halos (e.g. Figure

267: \ref{fig:mergers}) is closely related to the LRG multiplicity

268: function.  The former can be converted into the latter if one knows

269: both the minimum halo mass (measured at the epoch of accretion)

270: associated with accreted LRGs, $\mmin$, and the average time it takes

271: for LRGs to merge and/or disrupt\footnote{Throughout we use the words

272:   ``merge'' and ``disrupt'' interchangeably since our analysis does

273:   not distinguish between these two possibilities.} once accreted.

274: Below we argue for reasonable values of $\mmin$ and then attempt to

275: directly constrain the average LRG merging timescale.  We parameterize

276: the probability that an LRG will have merged by a time $t_{\rm{acc}}$

277: since accretion onto the host halo via:

278: \be

279: P_{\rm{merge}} = 1 - e^{-t_{\rm{acc}}/\tau}

280: \ee

281: where $\tau$ is the merging timescale.  The number of LRGs predicted

282: by this simple model is then 

283: \be

284: N_{\rm LRG} = 1 + \sum_i e^{-t_{\rm{acc,i}}/\tau}

285: \ee

286: where the first term counts one LRG at the center of the host halo and

287: the second term counts those satellites with accretion epoch mass

288: $>\mmin$ that have not merged.  For the purposes of generating a

289: multiplicity function we round the second term to the nearest integer.

290: Note that in generating a multiplicity function we do not have to make

291: a distinction between accretion events that did or did not occur

292: within the main host halo.  This distinction will only become relevant

293: when discussing the merger rates of LRGs.

294: 

295: \begin{figure}

296: \begin{center}

297: \resizebox{3.5in}{!}{\includegraphics{f3.eps}}

298: \end{center}

299: \caption{Number of accreted halos with $M>\amass$ that have merged

300:   with the host halo by redshift $z$, per $z=0$ host halo with

301:   $M>\hmass$.  The merging time is computed in two different ways: a

302:   constant timescale of either 0 or 2 Gyr (\emph{dashed} and

303:   \emph{solid lines}) and a timescale determined by the Chandrasekhar

304:   dynamical friction formula (\emph{dotted line}).}

305: \label{fig:mtime}

306: \end{figure}

307: 

308: 

309: The minimum LRG stellar mass in the observed sample is $10^{11.3}

310: M_\odot$.  A minimum halo mass associated with LRGs can be estimated

311: by assuming the universal baryon fraction $f_b=0.17$ and an efficiency

312: factor $\eta$ of converting baryons into stars.  While this factor is

313: only poorly constrained even at low redshift, values of order

314: $\sim0.1$ are likely reasonable for these massive galaxies

315: \citep{Hoekstra05,Mandelbaum06} and imply a minimum LRG hosting halo

316: mass of $\mmin\sim\amass$.  Note that the minimum halo mass for

317: hosting LRGs is here used to identify those accreted halos that are

318: likely to host an LRG, with a significant fraction of the accretion

319: occurring at $z\sim1$ or higher.  Thus this minimum mass will likely

320: not directly correspond to the minimum halo mass hosting LRGs at

321: $z\sim 0$, since the LRGs that have survived to the present epoch will

322: have accreted much more dark matter, resulting in a larger value for

323: $\mmin$ at the present epoch.  In our simulations halos today are on

324: average five times more massive than they were when they first crossed

325: $\mmin$, making our estimate consistent with that inferred from $z\sim

326: 0$ clustering \citep{Zehavi05b,Kulkarni07}.  As we discuss in

327: \S\ref{s:stellar}, $\mmin$ much larger than $\amass$ would require

328: unreasonably long dynamical friction times and $\mmin$ either much

329: larger or much smaller would be in conflict with stellar mass

330: estimates in clusters.

331: 

332: Figure \ref{fig:mf} plots the resulting LRG multiplicity function both

333: for LRGs in accreted halos more massive than $\amass$ (solid lines)

334: and observations (dashed lines).  We also include predictions for LRGs

335: associated with halos twice as massive as our fiducial minimum mass

336: (dotted lines) in order to illustrate the sensitivity to our assumed

337: LRG halo mass threshold.  Each panel in Figure \ref{fig:mf} is the

338: multiplicity function for a different merger timescale.  The top four

339: panels assume that the merger timescale, $\tau$, is constant. It is

340: apparent from these panels that if $\mmin\sim\amass$ then LRGs must

341: merge on a characteristic timescale of $\sim2$ Gyr.  This timescale

342: implies an average number of LRGs per cluster of 2.5, satisfyingly

343: close to the observed value of 2.6.

344: 

345: Note that in order to compare to the observations we have weighted

346: host halos at $z=0.3$ (of which there are 460 in our simulation within

347: the observed mass range) in such a way as to reproduce the mass

348: distribution of the observed clusters.  It is the combination of these

349: two effects (higher redshift and different mass distribution) that

350: does not allow a direct comparison between Figures \ref{fig:mergers}

351: and \ref{fig:mf}.

352: 

353: If LRGs never merged, there would be on average $5.8$ LRGs per $z=0.3$

354: host halo (averaged over the observed distribution of halo masses).

355: Comparing this number to the observed $2.5$ LRGs per cluster

356: highlights the importance and prevalence of LRG mergers.  

357: 

358: The bottom panel in Figure \ref{fig:mf} assumes that the merger

359: timescale is equal to the dynamical friction timescale \citep{Binney87}:

360: \be

361: \label{eqn:df}

362: t_{\rm DF} = 0.1\,t_H\, \,\frac{M_h/M_s}{\ln(1+M_h/M_s)}

363: \ee

364: where $M_h$ and $M_s$ are the host and satellite masses, $z$ is the

365: redshift, $t_H$ is the Hubble time and all quantities are measured at

366: the epoch of accretion.  The average mass ratio at accretion is $\sim

367: 6$ for our sample.  The pre-factor, $0.1\,t_H$, is the characteristic

368: time for a halo with mean density ${\mathcal O}(10^2)$ times the

369: critical density.  It is important to note that $t_{\rm DF}$ gets

370: shorter both at higher redshift and for merger mass ratios closer to

371: unity.  In this lower panel we have additionally included results for

372: halos of mass $M>5\times 10^{12} M_\odot$ (dot-dashed line) for

373: comparison.

374: 

375: This comparison with simple dynamical friction estimates provides a

376: satisfying cross check to the results in the upper panels.  In

377: particular, for our fiducial minimum LRG halo mass of $\amass$, the

378: dynamical friction timescale averaged over all the accreted halos is

379: $2.4\,$Gyr (median time is $1.8\,$Gyr --- the distribution is highly

380: asymmetric), which is quite similar to the constant merger timescale

381: that best matches the observed multiplicity function ($\sim2$ Gyr).

382: From Figure \ref{fig:mf} it is apparent that the simple DF timescale

383: would not have reproduced the observed LRG multiplicity function if

384: the minimum halos mass capable of hosting LRGs were substantially more

385: or less massive than $\amass$.  The implication here is clear: if

386: $\mmin$ is in fact considerably larger or smaller than $\amass$ then

387: simple DF arguments do not apply to the LRG population.

388: 

389: In fact, it is not at all clear that Equation \ref{eqn:df} should

390: apply here or in general to the dynamical evolution of satellite

391: galaxies, as it is strictly valid for a point mass moving in an

392: infinite, uniform background density field.  Indeed, much work has

393: gone into both testing the validity of Equation \ref{eqn:df} with

394: simulations \citep[e.g.][]{SWhite83,vdb99, Velazquez99,Read06} and

395: developing extensions to it \citep[e.g.][]{Tremain84,Colpi99},

396: including numerically following the evolution of the satellite orbit,

397: including the mass loss due to tidal forces \citep{Benson02,Taffoni03,

398:   Taylor04,Zentner05}.  This body of work has shown that Equation

399: \ref{eqn:df} is at best a crude approximation to the realistic,

400: time-dependent problem.  For these and other reasons it is quite

401: surprising, if not entirely coincidental, that the classical

402: Chandrasekhar DF timescale adequately captures the merging timescale

403: of LRGs.

404: 

405: As can be seen from the upper four panels in Figure \ref{fig:mf},

406: there is a degeneracy between the merger timescale and the minimum

407: halo mass associated with LRGs in the sense that a larger $\mmin$

408: coupled to a larger timescale can produce roughly the same

409: multiplicity function.  Thus, if one thought that LRGs lived in more

410: massive halos than what we have assumed here, then one would infer a

411: longer merging timescale for LRGs.  However, this is exactly opposite

412: to what one would infer from dynamical friction arguments since

413: $t_{\rm DF}\propto M_s^{-1}$.  Furthermore, increasing $\mmin$ to

414: $5\times\amass$ would result in far too few LRGs in massive halos

415: compared to observations, even if $\tau=\infty$; in this case the

416: average number of LRGs per halo would be $1.6$.  As we describe in

417: \S\ref{s:stellar}, $\mmin$ is further constrained by observations of

418: the stellar light in massive halos.

419: 

420: These merger timescales can easily be cast into a discussion of LRG

421: merger rates.  For this discussion we consider the full host halo

422: population at $z=0$, rather than the population at $z=0.3$ meant to

423: coincide with the data from Ho et al., in order to draw more general

424: conclusions about LRG mergers.  Figure \ref{fig:mtime} plots the

425: cumulative distribution of \emph{merged} LRGs as a function of

426: redshift, per host halo.  The figure includes constant merger

427: timescales of $0$ and $2\,$Gyr and a timescale set by dynamical

428: friction.  The $0\,$Gyr case can equivalently be thought of as the

429: distribution of \emph{accreted} LRGs, since in this case the accretion

430: and merging epochs are coincident.  In this figure we only count LRGs

431: that merge within the main progenitor of the $z=0$ host halo.  This

432: figure is thus not directly comparable to Figure \ref{fig:mergers}.

433: In other words, if a halo merges within a halo that itself later

434: mergers with the host halo then it is not counted here.  This plot is

435: thus meant to capture the number of mergers actually occurring

436: \emph{within the main progenitor} of the host halo.

437: 

438: The dynamical friction timescale is shorter than $2\,$Gyr at high

439: redshift and longer than $2\,$Gyr at low redshift; this results in a

440: more gradual increase in the merger rate per unit redshift compared to

441: a constant merger time of $2\,$Gyr.  Since the constant and dynamical

442: friction timescales are different at redshifts both greater and less

443: than $0.3$, comparisons to the multiplicity function at different

444: epochs can in principle rule out either (or both) of these timescales.

445: Both the constant $2\,$Gyr timescale and that determined by dynamical

446: friction imply that $2-2.5$ LRGs have merged with the host halo by

447: $z=0$.  Moreover, the figure indicates that a substantial number of

448: LRGs are merging/disrupting at $z<1$.  In the next section we set this

449: in the context of recent observational results of the stellar mass

450: budget in groups and clusters.

451: 

452: \begin{figure}

453: \begin{center}

454: \resizebox{3.5in}{!}{\includegraphics{f4.eps}}

455: \end{center}

456: \caption{The total stellar-to-virial mass ratio as a function of $z=0$

457:   halo virial mass.  Only galaxies more massive than $M_{\rm

458:     star}=10^{11.3} M_\odot$ and the ICL are included in the stellar

459:   mass budget.  Observations from \citet[][\emph{solid

460:     circles}]{Lin04a} have been converted to total stellar masses, and

461:   are compared to total stellar masses estimated by assuming that

462:   every accreted halo with mass $>\amass$ hosts an LRG with $M_{\rm

463:     star}=10^{11.6} M_\odot$ (\emph{diamonds}).}

464: \label{fig:mstarmvir}

465: \end{figure}

466: 

467: \section{The Total Cluster Stellar Mass} \label{s:stellar}

468: 

469: The merger rate of LRGs found in the previous section indicates that a

470: significant number of LRGs must have disrupted over the history of

471: massive halos.  In particular, a merger timescale of $\tau=2\,$Gyr

472: implies that on average $2-3$ LRGs have disrupted within each host

473: halo more massive than $\hmass$ by $z=0$ (and approximately $2.1$

474: since $z=1$).  In this section we discuss how reasonable such a

475: disruption rate is in light of the total stellar mass observed in

476: clusters at $z\sim0$.

477: 

478: The amount of stellar mass brought into massive halos by LRGs can be

479: estimated in the following way.  We consider halos that have a mass at

480: $z=0$ greater than $\hmass$ and assign a stellar mass of $10^{11.6}

481: M_\odot$, the mean stellar mass of the observed LRG sample, to the

482: center of each.  Then, each halo that was accreted onto these $z=0$

483: halos is assigned the same amount of stellar mass if the halo mass at

484: accretion is $>\amass$.  This exercise is thus meant to count the

485: total amount of stellar mass that was at some point associated with

486: LRGs.  We make no distinction between disrupted and non-disrupted LRGs

487: except in one case: if, according to our best-guess LRG merger

488: timescale (2 Gyr), an LRG halo would have disrupted not in the main

489: progenitor of the $z=0$ halo but rather in some smaller halo that

490: would itself later accrete onto the main progenitor but does not

491: merge, then we do not count this LRG in the final stellar mass budget.

492: In this case the disrupted LRG contributes to the satellite's ICL

493: (i.e. ICL that surrounds the satellite and is distinct from the

494: central ICL).  As it turns out, only $20$\% of all accreted halos more

495: massive than $\amass$ fall into this category and including these

496: halos in the stellar mass budget does not appreciably change our

497: conclusions.

498: 

499: We compare to data presented in \citet{Lin04a} who have compiled

500: information on $93$ clusters at $z<0.1$, including $X$-ray

501: observations used to derive cluster virial masses and luminosities of

502: cluster members derived from 2MASS photometry.  From these data Lin et

503: al. have estimated the luminosity function (LF) of each cluster,

504: assuming that the faint-end slope is fixed at $\alpha=-1.1$.  Using

505: their LFs we are able to estimate the total luminosity in galaxies

506: brighter than $L_K=2.8\times10^{11} L_\odot$ which corresponds to the

507: minimum luminosity of the Ho et al. LRG sample.  This total luminosity

508: is converted into stellar mass by assuming a mass-to-light ratio of

509: $M_{\rm star}/L_K=0.72$ which is appropriate for red galaxies with a

510: Chabrier IMF \citep{Bell03}.  The LFs reported in \citet{Lin04a} do

511: not include the brightest cluster galaxy (BCG); we thus add these in

512: separately.  Finally, we have assumed that each cluster contains

513: intracluster light (ICL) with a stellar mass equal to the mass of the

514: BCG identified by Lin et al.~(i.e. $L_{\rm ICL} = L_{\rm BCG}$).  This

515: amount of light associated with the ICL is consistent with recent

516: observations \citep{Zibetti05, Gonzalez05}.  In this ``total'' cluster

517: luminosity we do not include any possible ICL associated with

518: satellites; it is for this reason that we did not include LRGs that

519: disrupted in halos which themselves later merged with the host.

520: 

521: Figure \ref{fig:mstarmvir} presents a comparison between the data from

522: \citet{Lin04a} and the stellar mass associated with accreted halos.

523: The agreement is encouraging.  Note that varying any one of our

524: assumptions can change the results from both the data and our model;

525: the important point to take away from this comparison is that the

526: influx of massive galaxies embedded within accreted halos appears to

527: roughly agree with the total stellar mass within observed clusters at

528: $z\sim0$.  This provides further support to our identification of

529: halos with mass $>\amass$ as being host to LRGs and suggests that

530: disrupted LRGs deposit their stars into a combination of the central

531: galaxy and ICL.  Increasing or decreasing $\mmin$ by a factor of two

532: would result in substantial disagreement with the observations shown

533: in Figure \ref{fig:mstarmvir}.  This is due to the fact that the

534: number of accreted halos does not scale linearly with the accreted

535: halo mass, and provides further support for our choice of

536: $\mmin=\amass$.

537: 

538: The simple model presented here also provides a straightforward means

539: for understanding the observed trend of decreased scatter in $M_{\rm

540:   star}/M_{\rm host}$ with increasing $M_{\rm host}$.  This arises

541: because the number of accreted halos with mass $>\amass$ is a weak

542: function of $M_{\rm host}$.  This is in contrast to the observed

543: number of satellites, which appears to be closer to linear in $M_{\rm

544:   host}$ \citep{Lin04a, Popesso07}.  The difference implies that

545: fractionally more halos/LRGs are merging in lower mass halos compared

546: to higher mass halos.

547: 

548: Interestingly, we find that it is not uncommon for $z=0$ host halos to

549: contain disrupted LRGs that did not disrupt within the host halo (and

550: hence were not counted in the above figure) but rather disrupted in a

551: smaller halo that later accreted onto the host (and remained as a

552: satellite to $z=0$).  If these disrupted LRGs are depositing some

553: fraction of their stars into ICL, then this suggests that there could

554: be a significant amount of ICL that is not centered on the central

555: galaxy but is instead centered on cluster satellites.  Such a scenario

556: is corroborated by recent observations of the Virgo cluster that show

557: significant amounts of ICL surrounding several of the most massive

558: satellites \citep{Mihos05}.

559: 

560: 

561: \section{Discussion}\label{s:conc}

562: 

563: The results of the previous sections suggest the following picture.

564: If LRGs are associated with halos more massive than $\amass$ at the

565: time when they are accreted onto more massive host halos, then the

566: observed multiplicity function of LRGs at $z\sim0.3$ implies that LRGs

567: must merge and/or disrupt on timescales of $\sim2$ Gyr.  Such a merger

568: rate implies that $2-3$ such LRGs have disrupted in halos more massive

569: than $\hmass$ by $z=0$.  This merger timescale is consistent with

570: classical dynamical friction arguments and suggests that a rather

571: simplistic dynamical prescription for the evolution of LRGs is

572: applicable when considering ensemble averages.

573: 

574: Moreover, the amount of total stellar mass in clusters that was at one

575: point associated with these infalling LRGs (ignoring for the moment

576: whether or not this stellar mass is locked up in satellite galaxies)

577: is consistent with observations when the observed amount of stars in

578: the intracluster light (ICL) is accounted for.  This in turn suggests

579: that the disrupting LRGs are depositing their stars into a combination

580: of the ICL and central galaxy, which is consistent with previous

581: modeling \citep{Monaco06,Murante07,Purcell07, Conroy07b}. Finally,

582: there appears to be a significant number of LRGs that have disrupted

583: within halos that only later accreted onto (but did not merge with)

584: what would become the $z=0$ host halo.  This suggests that there could

585: be a significant amount of ICL surrounding cluster satellites, in

586: addition to what is known to be associated with the central galaxy.

587: 

588: It has been historically challenging to constrain the merger rate of

589: galaxies.  Previous studies have relied on either morphological

590: disturbances \citep[e.g.][]{Conselice03, vanDokkum05, Bell06} or close

591: pair counts \citep[e.g.][]{Masjedi06} as probes of the merger rate of

592: massive galaxies.  Unfortunately, both methods are rather indirect

593: since the connection between either morphological disturbances or

594: close pair counts and merger rates is uncertain.  The most recent

595: inferred LRG-LRG merger rate is from \citet{Masjedi06} who find a rate

596: of $0.6\times 10^4$ Gyr$^{-1}$ Gpc$^{-3}$.  Averaging over all halos

597: between $z=0.5$ and $z=0.2$, the model presented herein implies an

598: LRG-LRG merger rate of $(1.0-1.3)\times 10^4$ Gyr$^{-1}$ Gpc$^{-3}$,

599: depending on whether the constant 2 Gyr or dynamical friction

600: timescale is used.  The agreement with \citet{Masjedi06} is

601: encouraging, especially given the (different) uncertainties in both

602: approaches.  These rates are also consistent with current predictions

603: from cosmological hydrodynamic simulations \citep{Maller06}.

604: 

605: Many studies have attempted to constrain the stellar mass growth of

606: massive galaxies from their inferred merger rates.  However, as argued

607: in \citet{Conroy07b} and herein, the merging of massive galaxies will

608: often not correspond to significant growth of the resulting galaxy

609: because a substantial amount of stars can be transfered to the ICL.

610: We hence caution against using merger rates to constrain the stellar

611: mass growth of galaxies.\footnote{This issue is intimately related to

612:   the way in which one counts galaxy light.  Of course the

613:   \emph{combined} light of both the central galaxy and its ICL will

614:   increase after a merger event.}  In fact, significant growth of the

615: ICL via merging at late times provides a means for reconciling two

616: apparently contradictory facts: one the one hand, observations at

617: $z<1$ indicate that central massive red galaxies grow little in mass

618: \citep[e.g.][]{Brown07, Fontana06, Bundy06, Wake06}, while on the

619: other hand, merging/disruption of galaxies within groups and clusters

620: at late times appears relatively common \citep[e.g.][]{MWhite07}.

621: 

622: \citet{MWhite07} outlined an approach for measuring the merging rate

623: of massive galaxies similar to the one presented herein.  Using the

624: observed evolution in the clustering of massive galaxies, these

625: authors concluded that $\sim1/3$ of massive satellites merge/disrupt

626: between $z\simeq0.9$ and $z\simeq0.5$.  In the present work we find

627: roughly $50$\% of massive satellites have disrupted over similar

628: epochs.  Our fraction is slightly higher because we have focused on

629: more massive galaxies than in \citet{MWhite07}.  The more general

630: conclusion from these two studies is, however, robust --- the

631: population of massive galaxies experiences significant amounts of

632: merging/disruption, even at $z<1$.

633: 

634: There is an important implication of considering the evolution of

635: galaxies within the context of the hierarchical growth of halos.  At

636: first glance, the lack of evolution in the observed correlation

637: function of massive galaxies and their halo occupation distribution at

638: $z<1$ suggests that massive galaxies do not disrupt or merge over this

639: epoch.  However, these galaxies are embedded within dark matter halos

640: that are continually merging and accreting new galaxies, which instead

641: suggests that massive galaxies \emph{must} merge in order that these

642: observed quantities not evolve appreciably at late times.  This

643: statement is further corroborated by dissipationless simulations which

644: show explicitly that the average number of subhalos within host halos

645: does not evolve appreciably at $z<1$ because the accretion and

646: disruption rate of subhalos are approximately equal \citep{Reed05,

647:   Conroy06a}.  If satellite galaxies reside within these subhalos then

648: the observed non-evolution of the clustering and halo occupation of

649: massive galaxies at $z<1$ is in fact consistent with significant

650: amounts of merging at late times.

651: 

652: Our results highlight the power of using purely dissipationless

653: simulations coupled to simple relations between galaxies and dark

654: matter to infer the evolution of galaxies and their relation to the

655: underlying dark matter with time.  The approach outlined herein can

656: easily be extended to other datasets to provide additional constraints

657: on the merger rate of galaxies.

658: 

659: \section*{Acknowledgments}

660: 

661: We thank David Hogg, Mike Boylan-Kolchin, Morad Masjedi, Eliot

662: Quataert, and Jerry Ostriker for comments on an earlier draft.

663: S.H. thanks David Spergel, Jim Gunn, Jerry Ostriker, Chris Hirata and

664: Nikhil Padmanabhan for insightful discussions and comments.

665: C.C. thanks nature for being complicated, but not too complicated.

666: M.W. is supported by NASA. The simulations were performed on the

667: supercomputers at the National Energy Research Scientific Computing

668: center.

669: 

670: 

671: \input{ms.bbl}

672: %\bibliographystyle{mn2e}

673: %\bibliography{master_refs}

674: 

675: 

676: \begin{appendix}

677:   

678:   \section{The Simulation, Halo Catalog, and Merger

679:     Trees} \label{sec:simulation}

680: 

681: We use a high resolution simulation of a $\Lambda$CDM cosmology

682: ($\Omega_M=0.25=1-\Omega_\Lambda$, $\Omega_B=0.043$, $h=0.72$,

683: $n_s=0.97$ and $\sigma_8=0.8$).  The linear theory power spectrum is

684: computed by evolution of the coupled Einstein, fluid and Boltzmann

685: equations using the code described in \citet{MWhite96}.  This code agrees

686: well with {\sl CMBfast\/} \citep{Seljak96}, see e.g.~\citet{Seljak03}.  The

687: simulation employs $1024^3$ particles of mass $8\times

688: 10^{9}\,h^{-1}M_\odot$ in a periodic cube of side $500\,h^{-1}$Mpc

689: using a {\sl TreePM\/} code \citep{MWhite02}.  The Plummer equivalent

690: softening is $18\,h^{-1}$kpc (comoving).

691: 

692: The phase space data for the particles exists at 50 outputs, spaced

693: equally in conformal time between $z\simeq 3$ and $z=0$.  For each

694: output we generate a catalog of halos using the Friends-of-Friends

695: (FoF) algorithm \citep{Davis85} with a linking length of $0.168\times$

696: the mean inter-particle spacing.  This procedure partitions the

697: particles into equivalence classes, by linking together all particle

698: pairs separated by less than a distance $b$.  The halos correspond

699: roughly to particles with $\rho>3/(2\pi b^3)\simeq 100$ times the

700: background density.  For each halo we compute a number of properties,

701: including the mass $M_{200}$ interior to $r_{200}$ within which the

702: mean density is $200\times$ the critical density.  $M_{200}$ is

703: computed from a fit of an NFW profile \citep{NFW97} to the particles

704: in the FoF group.

705: 

706: Merger trees are computed from the set of halo catalogs by identifying

707: for each halo a ``child'' at a later time.  The child is defined as

708: that halo which contains, at the later time step, more than half of

709: the particles in the parent halo at the earlier time step (weighting

710: each particle equally).  For the purposes of tracking halos this

711: simple linkage between outputs suffices (note that we do not attempt

712: to track subhalos within larger halos, which generally requires

713: greater sophistication).  From the merger trees it is straightforward

714: to compute the time when a halo `falls in' to a larger halo, the

715: number and masses of the progenitors etc.

716: 

717: \end{appendix}

718: 

719: 

720: \end{document}

721: