astro-ph0503451/ms.tex
1: %\documentclass[12pt,preprint]{aastex}
2: \documentclass[apj]{emulateapj}
3: \usepackage{apjfonts}
4: 
5: %%\received{}
6: %%\accepted{}
7: %%\journalid{}{}
8: %%\articleid{}{}
9: 
10: \slugcomment{Accepted to {\it{The Astrophysical Journal Letters}}}
11: \shortauthors{Gonzalez et al.}
12: \shorttitle{Galaxy Cluster Assembly at $z=0.37$}
13: 
14: 
15: 
16: \citestyle{aa}
17: \begin{document}
18: 
19: \newcommand{\kms}{~km~s$^{-1}$~}
20: \newcommand{\logh}{$+5\log h_{70}$}
21: \newcommand{\chandra}{{\it{Chandra~}}}
22: \newcommand{\ho}{~}
23: \newcommand{\msun}{M$_\odot$~}
24: \newcommand{\sg}{SG1120-1202}
25: \title{Galaxy Cluster Assembly at $z=0.37$$^{1}$}
26: 
27: \author{Anthony H. Gonzalez$^{2,3}$, Kim-Vy H. Tran$^4$, Michelle N. Conbere$^3$, 
28: \& Dennis Zaritsky$^5$ }
29: 
30: \footnotetext[1]{Based on observations with the Chandra X-ray
31:   Observatory, the VLT (program 072.A-0367), the Magellan Baade telescope, 
32: and the MMT Observatory, a joint venture of the Smithsonian Astrophysical 
33: Observatory and the University of Arizona.}
34: \footnotetext[2]{NSF Astronomy and Astrophysics Postdoctoral Fellow}
35: \footnotetext[3]{Department of Astronomy, University of Florida,
36:   Gainesville, FL 32611}
37: \footnotetext[4]{Institute for Astronomy, ETH Z\"urich, CH-8093
38:   Z\"urich, Switzerland}
39: \footnotetext[5]{Steward Observatory, University of Arizona, 933 
40: North Cherry Avenue, Tucson, AZ 85721}
41: 
42: 
43: \begin{abstract}
44: 
45:   We present X-ray and spectroscopic confirmation of a cluster assembling from
46:   multiple, distinct galaxy groups at $z=0.371$.  Initially detected in the
47:   Las Campanas Distant Cluster Survey, the structure contains at least four
48:   X-ray detected groups that lie within a maximum projected separation of 4
49:   Mpc and within $\Delta \mathrm{v}=550$\kms of one another.  Using \chandra
50:   imaging and wide-field optical spectroscopy, we show that the individual
51:   groups lie on the local $\sigma-T$ relation, and derive a total mass of
52:   $M\ge 5\times 10^{14}$ \msun for the entire structure.  We demonstrate that
53:   the groups are gravitationally bound to one another and will merge into a
54:   single cluster with $\gtrsim\frac{1}{3}$ the mass of Coma.  We also find
55:   that although the cluster is in the process of forming, the individual
56:   groups already have a higher fraction of passive members than the field.
57:   This result indicates that galaxy evolution on group scales is key to
58:   developing the early-type galaxies that dominate the cluster population by
59:   $z\sim0$.
60:   
61: \end{abstract}
62: 
63: \keywords{galaxies: evolution --- galaxies: clusters: general ---  X-rays: galaxies: clusters}
64: 
65: \section{Introduction}
66: 
67: A fundamental prediction of hierarchical structure formation is that galaxy
68: clusters assemble at late times from the merging and accreting of smaller
69: structures \citep{peebles:70}.  While many examples of the accretion of groups
70: and galaxies onto massive clusters exist
71: \citep[e.g.][]{abraham:96,hughes:98,tran:05}, we lack clear early-stage
72: examples of clusters being assembled from an ensemble of galaxy groups.  The
73: identification of such protoclusters is challenging due to the extensive
74: mapping required to trace the substructure that will eventually form the
75: cluster.
76: 
77: Nonetheless, detection of these rare systems is well worth the effort.  They
78: provide a detailed snapshot of the process of filamentary collapse and
79: assembly in the quasi-linear regime, and enable a {\it direct} study of a
80: class of progenitors that does not yet include a massive cluster.  The latter
81: is particularly critical for determining if ``pre-processing'' in groups is
82: responsible for the bulk of the observed differences in morphology and stellar
83: populations between cluster and field galaxies \citep{zabludoff:98,kodama:01}
84: 
85: We present confirmation of a protocluster identified optically in the Las
86: Campanas Distant Cluster Survey \citep[LCDCS;][]{gonzalez:01}.  Designed to
87: provide a catalog of clusters at $z\approx0.35-0.9$, the LCDCS extends down to
88: group masses at the lowest redshifts and hence can be used to identify
89: filamentary structures at $z\sim0.4$.  One such candidate structure includes
90: four LCDCS candidates within a 3.5$\arcmin$ radius region with estimated
91: redshifts of $z_{est}=0.35-0.50$.  We demonstrate that we are witnessing the
92: assembly of a cluster from what can be described in analogy to superclusters
93: as a supergroup. Throughout this paper, we use a standard cosmology
94: ($\Omega_M=0.3$, $\Omega_\Lambda=0.7$, $H_0=70$\kms).
95: 
96: \vskip 0.5cm
97: \section{X-ray Observations with Chandra}
98: 
99: The target field was observed with \chandra\ ACIS-I in Very Faint (VF) mode on
100: March 24, 2002 (Obs ID 3235) for 70.13 ks, with the detector oriented such
101: that one LCDCS candidate was imaged in each ACIS-I chip.  Data reduction was
102: performed in the standard fashion using CIAO 3.0.  The level 1 event file was
103: reprocessed to correct for charge transfer inefficiency and to reduce the VF
104: mode particle background,\footnote{See
105: \url{http://cxc.harvard.edu/cal/Acis/Cal\_prods/vfbkgrnd/} } and bad pixels
106: and events with bad grades were removed in the standard fashion.  Strong
107: flares were removed using the $\mathrm{lc\_clean}$ routine written by Maxim
108: Markevitch, yielding a net exposure time of 64 ks. The high energy background
109: in the remaining time interval is consistent with the quiescent rate from the
110: "blank-sky" background file to within 1\%.\footnote{See
111: \url{http://cxc.harvard.edu/ciao/threads/acisbackground/}}
112: 
113: 
114: Image analysis is performed in the 0.8-3 keV band. Using wavelet detection, we
115: identify six distinct extended sources at $>5\sigma$ significance (Figure
116: \ref{fig:xray}). Four of these sources are coincident to within $10\arcsec$
117: with LCDCS candidates, while the other two extended sources have no LCDCS
118: counterparts. One of these latter two (source 5) could not have been detected
119: by the LCDCS due to its proximity to a bright star; the other is the faintest
120: of the six sources and appears to be a very poor group.
121: 
122: 
123: Spectral analysis is performed in the 0.7-8 keV band.  Temperatures are
124: measured within $400$\ho kpc circular apertures to maximize signal-to-noise
125: (S/N), excluding point sources, and using the appropriately normalized ACIS
126: "blank-sky" file for the background.  We fit the spectra using a MEKAL model,
127: with the absorption fixed to the galactic value $n_H=5.13\times10^{20}$
128: cm$^{-2}$ and the metal abundance fixed at $Z=0.3$ \citep{mushotzky:1997}.
129: For four of the extended sources, we fix the redshift to the median velocity
130: of the associated galaxies (see \S \ref{sec:spec}), while for LCDCS 0259
131: (source 2), which lies outside the spectroscopic survey region, we use the
132: single redshift obtained for the brightest associated galaxy.  The best-fit
133: temperatures (Table \ref{tab:tab1}, with 68\% uncertainties) range from
134: $T=1.7-3$ keV and are consistent with the temperatures found for groups or
135: poor clusters similar to Virgo \citep{shibata:01}.  Only for the faintest
136: source do we lack sufficient S/N to obtain a temperature.
137: 
138: \begin{figure}
139: \epsscale{1.0}
140: \plotone{f1}
141: \caption{Smoothed X-ray map ($13.2'\times 13.2'$) of the structure with
142:   point sources excised; north is up and east is to the left. The
143:   crosses denote the locations from the LCDCS catalogs and the numeric
144:   labels correspond to the designations in Table \ref{tab:tab1}.
145:   Redshifts have been obtained for sources 1-5.
146: \label{fig:xray}}
147: \end{figure}
148: 
149: 
150: 
151: \section{Optical Spectroscopy}
152: \label{sec:spec}
153: 
154: \subsection{MMT, Magellan, and VLT Observations}
155: 
156: To confirm that the X-ray sources are part of a larger structure, long-slit
157: spectra of the candidate brightest group galaxies (BGGs) associated with each
158: of the six X-ray peaks were obtained with Magellan and the MMT in February and
159: April 2003, respectively.  Redshifts obtained for the BGG's in groups 1-5
160: confirm that at least four of the X-ray regions are associated with a single
161: physical complex at $z=0.37$.  Only the BGG candidate for group 1 was found to
162: lie at a different redshift ($z=0.48$).  These initial results motivated a
163: more extensive spectroscopic program with the VLT.
164: 
165: \begin{figure}
166: \epsscale{1.0}
167: \plotone{f2}
168: \caption{{\it Top left} --
169:   Redshift distribution of the 101 confirmed members (defined by $0.36<z<0.38$)
170:   in \sg~($z=0.3710\pm0.0004$).  Although the members trace a very
171:   large structure (projected spatial distance $\sim4$\ho Mpc) and
172:   include galaxies from 3 of the groups at $z=0.37$, they have a
173:   remarkably narrow redshift distribution and a 
174:   velocity dispersion of only $\sigma=616\pm50$\kms. 
175:   {\it Top right and lower panels} -- Redshift distributions for the individual groups; the
176: arrow denotes the average redshift of the members.  To isolate
177:   members of each individual group, here we consider only members
178:   within 500\ho kpc of the X-ray centers.  Even with our
179:   conservative selection criterion, the maximum redshift difference
180:   between the groups is only $\Delta z=0.0025$. 
181: \label{fig:zhist1}}
182: \end{figure}
183: 
184: 
185: 
186: Using VIMOS \citep{lefevre:03}, we measure redshifts and determine membership
187: for galaxies near groups 1, 3, 4, and 5; group 2 was excluded due to an
188: interchip gap.  From a $14'\times16'$ $R$-band image (1820s) taken with VIMOS,
189: we generated a catalog of $\sim2100$ objects with $R\leq22.5$.  Our
190: observations used three multi-slit masks with $1''-$wide slits and targeted
191: 443 of these objects; each mask had a total integration time of 2400 seconds.
192: While ours is a magnitude limited survey and targets were not selected by
193: morphology, preference was given to objects in visually overdense regions.
194: The intermediate resolution (MR) grism on VIMOS gave us a spectral range of
195: $0.5-1.0\mu$m and spectral resolution of 12.2\AA.
196: 
197: To reduce and extract spectra, we use a combination of IRAF\footnote{IRAF is
198: distributed by the National Optical Astronomy Observatories, which are
199: operated by the Association of Universities for Research in Astronomy, Inc.,
200: under cooperative agreement with the National Science Foundation.} routines
201: and custom software provided by D. Kelson \citep{kelson:98}; a more detailed
202: explanation of the reduction pipeline can be found in \citet{tran:05}.  We use
203: a spectrophotometric standard to correct for the telluric A and B bands and
204: flux calibrate the spectra.
205: 
206: We determine redshifts with the IRAF cross-correlation routine XCSAO
207: \citep{kurtz:92}.  Our final redshift catalog has 413 objects, corresponding
208: to a target success rate of 93\%.  The average redshift uncertainty estimated
209: by XCSAO is $\sim30$\kms; while XCSAO tends to underestimate the errors, this
210: problem does not affect our results.
211: 
212: \subsection{Redshift Distribution}
213: \label{sec:redshifts}
214: 
215: From spectroscopy of 4 of the 6 regions with extended X-ray emission, we
216: identify a peak corresponding to LCDCS 0258 ($z=0.48$, source 1), and another
217: corresponding to three X-ray groups at $z=0.37$, hereafter referred to as
218: \sg~(SG standing for ``supergroup''). The mean redshift of the 101 galaxies
219: belonging to \sg~ is $z=0.3710\pm0.0004$.  Although the members lie in
220: multiple, distinct X-ray bright regions and trace a large structure with a
221: projected spatial distance of $\sim4$\ho Mpc, their redshift distribution is
222: remarkably narrow (Fig.~\ref{fig:zhist1} and Table~\ref{tab:tab1}) and their
223: velocity dispersion surprisingly small ($\sigma=616\pm50$\kms).  The mean
224: redshift, velocity dispersion, and associated errors are determined using the
225: biweight and jacknife methods \citep{beers:90}.  To compare the dynamics of
226: the three X-ray groups to each other, members of each group are selected as
227: galaxies that lie within 500\ho kpc of their respective X-ray peak.  The
228: redshift distributions of the three groups are shown in Figure
229: \ref{fig:zhist1} and their velocity dispersions are listed in Table
230: \ref{tab:tab1}.
231: 
232: \subsection{Spectral Populations in \sg}
233: 
234: To quantify the recent star formation histories of galaxies in \sg, we
235: separate the members by [OII] equivalent width into absorption
236: ([OII]~$\lambda3727<5$\AA; "passive") and emission
237: ([OII]~$\lambda3727\geq5$\AA; "active") line galaxies.  We use the same
238: bandpasses as in \citet{fisher:98} to measure the [OII] doublet, and our
239: wavelength coverage includes the [OII] doublet for 95 of the 101 confirmed
240: members.
241: 
242: The fraction of passive galaxies in \sg~ is $61\pm8$\%.  This fraction is less
243: than in CL~1358+62 ($81\pm6$\%), a more massive cluster at comparable redshift
244: \citep[$z=0.33$, $\sigma=1027$\kms;][]{fisher:98}, but twice as high as the
245: field value \citep[$27\pm4$\% at $0.2<z<0.5$;][]{tran:04b}. The passive galaxy
246: fraction in \sg~ most closely resembles that found in X-ray luminous groups in
247: the nearby universe \citep[$69\pm7$\%;][]{tran:01}.  Direct comparison to
248: these other samples is valid because all four are magnitude-selected and use
249: the same [OII] selection criteria.
250: 
251: \begin{deluxetable}{llllllll}
252: \tabletypesize{\scriptsize}
253: \tighten
254: \tablecaption{Group Sample}
255: \tablewidth{0pt}
256: \tablehead{
257: \colhead{ID} &
258: \colhead{$\alpha$ } &
259: \colhead{$\delta$ } &
260: \colhead{LCDCS} &
261: \colhead{$z$} &
262: \colhead{$T$} &
263: \colhead{$\sigma$ } &
264: \colhead{$N_{g}$\tablenotemark{a}}\\
265: \colhead{} &
266: \colhead{(J2000)} &
267: \colhead{(J2000)} &
268: \colhead{ID} &
269: \colhead{} &
270: \colhead{(keV)} &
271: \colhead{\kms} &
272: \colhead {} 
273: }
274: \startdata
275: 1 & 11:19:55.2  &$-$12:02:28 & 0258 & 0.4794 &    $2.3^{+0.4}_{-0.3}$ & $820\pm101$ & 17  \\
276: 2 & 11:20:07.6  &$-$12:05:13 & 0259 & 0.3707&  $2.2^{+0.7}_{-0.4}$ & --- & 1   \\
277: 3 & 11:20:13.2  &$-$11:58:44 & 0260 & 0.3704 &    $1.7^{+0.5}_{-0.3}$ & $369\pm76$ & 17 \\
278: 4 & 11:20:22.3  &$-$12:01:45 & 0264 & 0.3713 &    $1.8^{+1.2}_{-0.5}$ & $446\pm83$ & 22 \\
279: 5& 11:20:10.0  &$-$12:08:50 & ---  & 0.3688 &    $3.0^{+1.2}_{-1.0}$ & $557\pm93$ & 11  \\
280: 6 & 11:20:15.6  &$-$11:56:03 & --- &  ---  &    ---                & ---  & --- \\
281: \enddata
282: \tablenotetext{a}{Number of galaxies used to calculate the mean redshift and velocity dispersion.}
283: \label{tab:tab1}
284: \end{deluxetable}
285: 
286: 
287: 
288: 
289: 
290: 
291: \section{Cluster Assembly}
292: 
293: \sg's extended, roughly linear geometry is seen in both its X-ray emission and
294: the spatial distribution of its galaxies.  Coupled with the small dispersion
295: in the mean group redshifts ($\Delta z=0.0025$) and visible signs of
296: interaction in the X-ray map (groups 2+5 and 3+6), this geometry argues that
297: we are witnessing the initial stages of filamentary collapse and the formation
298: of a more massive cluster.  To test this hypothesis, we estimate \sg's total
299: mass, assess the probability that the system is bound, and estimate the
300: dynamical timescale for the system to merge.  We include groups 2-5 in this
301: analysis; we are confident that group 2 is part of the complex despite having
302: only one redshift, both because the redshift is for the BGG (a 3 $L_*$
303: elliptical within 3$\arcsec$ of the X-ray peak) and because asymmetry in the
304: X-ray map indicates that groups 2 and 5 are interacting.
305: 
306: \subsection{Total Mass}
307: 
308: Given the consistency of the groups with the local $\sigma-T$ relation (Figure
309: \ref{fig:sigt}), we assume hydrostatic equilibrium and use the local
310: mass-temperature ($M-T$) relation in \citet[][Equation 16]{shimizu:03} to
311: estimate the masses of the individual groups.\footnote{Evolution in the $M-T$
312: relation should be negligible over this redshift interval \citep{ettori:04}.}
313: The four spectroscopically confirmed groups at $z=0.37$ have $T=1.7-3$ keV,
314: yielding a combined mass of $M=5.3^{+1.6}_{-0.9}\times10^{14}$ \msun.  This
315: mass is a lower limit on the total mass of the system because it does not
316: include group 6, groups that may lie outside the \chandra~field, or material
317: in filaments between groups.  It is equivalent to that of a single cluster
318: with $T\sim5$ keV, and at least a third the mass of the Coma cluster
319: \citep{hughes:1989,honda:1996}.
320: 
321: \begin{figure}
322: \epsscale{1.0}
323: \plotone{f3}
324: \caption{ Comparison of the groups in this study with the local $\sigma-T$
325:   relation \citep{xue:00}. The solid symbols with error bars correspond to the
326:   groups in our sample, circles denoting members of the supergroup ($z=0.37$)
327:   and the triangle LCDCS 0258 ($z=0.48$, source 1). The other points are the local data
328:   from \citet{xue:00} and the solid line is their best fit for this combined
329:   group and cluster sample. Our sample is consistent with this local relation.
330: \label{fig:sigt}}
331: \end{figure}
332: 
333: 
334: 
335: \subsection{Dynamical State}
336: 
337: That the individual groups in \sg~lie on the local $\sigma-T$ relation argues
338: that each group is approximately virialized and that, if the larger structure
339: is bound, the groups are infalling for the first time.  We perform a simple
340: dynamical analysis to assess the likelihood that the \sg~complex is indeed
341: gravitationally bound.
342: 
343: The individual groups will be bound to the larger system if 
344: \begin{equation}
345: \frac{GM}{R_{p} (\sin i)^{-1}} \geq \frac{\mathrm{v}_{pec}^2}{2}
346: \end{equation}
347: where $M$ is the total mass, $\mathrm{v}_{pec}$ is the pairwise peculiar
348: velocity for groups, $R_{p}$ is the projected distance between the groups, and
349: $i$ is the opening angle.  We assume the local value of
350: $<\mathrm{v}_{pec}>=325\pm175$ \kms from \citet{padilla:01} and conservatively
351: take $R_{p}= 3$\ho Mpc; this value corresponds to the largest separation
352: between the X-ray groups.
353: 
354: 
355: Using our X-ray estimate of the total mass, the above inequality holds for
356: $i\ga4^\circ$. Geometrically, the probability is 99.5\% that the opening angle
357: is at least this large. Even assuming $\mathrm{v}_{pec}=500$\kms ($1\sigma$
358: above the mean value from \citealt{padilla:01}), the inequality holds for
359: $i\ga10^\circ$ and the probability that the groups are bound remains high
360: (97\%).
361: 
362: The above analysis indicates that an unbound system is unlikely, but does not
363: conclusively eliminate the possibility.  We can strengthen the constraint by
364: exploring the implications of an unbound system.  For $i \la 10^\circ$, the
365: groups are separated by $\ga20$ Mpc, and the corresponding Hubble flow
366: redshift offset is $(\Delta z)_H\ga0.0075$ (2250 \kms). In this case, $(\Delta
367: z)_H$ is a factor of three greater than the observed maximum redshift
368: separation between the groups.  Thus relative peculiar velocities of $\ga1500$
369: \kms would be required to counterbalance the Hubble flow and explain the small
370: observed redshift separation; this is an implausibly large value if the groups
371: are truly separated by $\ga20$ Mpc and not infalling.
372: 
373: We conclude that the groups in \sg~are bound to each other.  The dynamical
374: time for this system is
375: \begin{equation}
376: t_{dyn}\approx \sqrt{\frac{R^3}{GM}} \approx 1.2 (\sin i)^{-3/2} \;\;\mathrm{Gyrs},
377: \end{equation}
378: for $R_{p}=1.5$\ho Mpc and $M=5\times 10^{14}$ \msun.  For $i\ga26^{\circ}$,
379: $t_{dyn}$ is less than the lookback time.
380: 
381: 
382: An alternative treatment is to assume that the groups are bound and on radial
383: orbits that have not yet crossed the center of the potential, and use the
384: timing argument \citep{kahn:1959} to determine their line-of-sight distance by
385: fitting the projected separation and radial velocity in the center of mass
386: frame. For the four groups we find solutions with line-of-sight distances from
387: the center-of-mass of between 7.5 and 10 Mpc.  These solutions have $11^\circ
388: < i < 27^\circ$, and hence suggest that the complete collapse of the system
389: will occur after the current time.  The main conclusion from this analysis is
390: that a bound solution exists with the measured masses that can reproduce the
391: positions and radial velocities of the groups. Detailed descriptions of the
392: orbits and future merger history require more information.
393: 
394: \section{Conclusions}
395: 
396: Our observations confirm that \sg~is a protocluster being built up by the
397: assembly of multiple galaxy groups.  From deep \chandra imaging and optical
398: spectroscopy, we find that \sg~ contains a minimum of four X-ray luminous
399: groups at $z=0.371$ within a projected 4\ho Mpc of one another.  The groups
400: have X-ray temperatures of $T=1.7-3$ keV and their mean redshifts span a
401: narrow range of $\Delta z=0.0025$ (550\kms).  The group velocity dispersions
402: and X-ray temperatures fall on the local $\sigma-T$ relation, indicating that
403: they are each virialized systems with a combined mass $M\ge5\times 10^{14}$
404: \msun.  Using dynamical arguments, we demonstrate that the X-ray groups are
405: gravitationally bound to one another and should merge into a single cluster;
406: the resulting cluster will have $\gtrsim\frac{1}{3}$ the mass of Coma.
407: 
408: 
409: While \sg~represents only one of a variety of accretion histories that can
410: yield a moderate mass cluster at $z=0$, this assembly path -- the late merging
411: of multiple roughly equal mass subhalos -- is of particular interest.
412: \sg~provides a unique laboratory for assessing the impact of local density
413: upon the evolution of member galaxies, and for quantifying the degree to which
414: the group environment acts to ``pre-process'' galaxies prior to cluster
415: assembly.  Already we see that the individual groups have twice as many
416: passive galaxies as the field, indicating that galaxy evolution on group
417: scales is key to developing the early-type galaxies that dominate the cluster
418: population by $z\sim0$.
419: 
420: 
421: \acknowledgements
422: 
423: We are indebted to A. Vikhlinin and M. Markevitch for their assistance
424: and for providing us with codes employed in the \chandra analysis.  Support
425: for this work was provided by NASA through Chandra Award GO2-3183X3 issued
426: by the CXO Center, which is operated by SAO for and on behalf of NASA under
427: contract NAS8-03060. AHG is funded by an NSF Astronomy \& Astrophysics
428: Postdoctoral Fellowship under award AST=0407085, KT acknowledges support from
429: the Swiss National Science Foundation, and DZ acknowledges fellowships from
430: the David and Lucile Packard Foundation and Alfred P. Sloan Foundation.
431: 
432:  
433: %Facilities: 
434: %   \facility{CXO(ACIS)}, \facility{VLT(VIMOS)}, 
435: %   \facility{MMT}, \facility{Magellan}
436: 
437: \bibliographystyle{apj}
438: \bibliography{ms}
439: 
440: 
441: \end{document}
442: