astro-ph0608444/ms.tex
1: \newif\ifemulate
2: \emulatetrue
3: 
4: \ifemulate
5:   \documentclass{emulateapj}
6:   \usepackage{apjfonts}
7: \else
8:   \documentclass[12pt,preprint]{aastex}
9: \fi
10: 
11: \newcommand{\msun}{$M_\sun$}
12: 
13: % put all figures and tables in defines so can
14: % switch easily from preprint to emulateapj format
15: 
16: %\vspace*{-1.15cm}
17: \def\figa{
18: \begin{figure*}
19: \begin{center}
20: \leavevmode
21: \includegraphics[angle=0,width=16.75cm]{f1a.ps} \vspace{-1cm}
22: \includegraphics[angle=0,width=16.75cm]{f1b.ps} \vspace{-1cm}
23: \includegraphics[angle=0,width=16.75cm]{f1c.ps}
24: \includegraphics[angle=0,width=16.75cm]{f1d.ps}
25: \caption{Images of $z_{850}$-dropouts in the HUDF from 
26: HST/ACS ($i_{775},z_{850}$), NICMOS ($J_{110},H_{160}$), and
27: {\it Spitzer}/IRAC ($3.6\mu$m,$4.5\mu$m). All sources are 
28: undetected $(<2\sigma)$ at $i_{775}$ and bluer wavelengths, 
29: but we note that $z_{850}$-dropout galaxies can have some flux 
30: at $i_{775}$ due to incomplete absorption between
31: Lyman$-\alpha$ at rest-frame 1216\AA\ and the Lyman limit at 
32: 912\AA. The top two candidates are clearly detected in the IRAC $3.6$ and 
33: $4.5\mu$m images, while two others are 
34: marginally detected. Flux contribution from nearby sources was subtracted 
35: in the Spitzer images. Each panel is $4\farcs1\times4\farcs1$
36: in size, or $\approx$21~kpc at $z=7$.}
37: \end{center}
38: \leavevmode
39: \vspace{-0.9cm}
40: \end{figure*}
41: }
42: 
43: %\vspace*{-0.1cm}
44: \def\figb{
45: \begin{figure}
46: \includegraphics[angle=0,width=8.0cm]{f2.ps}
47: \caption{Examples of the deblending procedure in
48: the IRAC $3.6\mu$m image as performed separately for the
49: first ({\it top row}) and second ({\it bottom row}) 
50: epoch images from GOODS. 
51: We fit the original images ({\it left}) using the high resolution
52: NICMOS data to construct models for the nearby neighbors ({\it middle})
53: which are subtracted resulting in the cleaned images ({\it right}).
54: We recover consistent fluxes for both epochs
55: after subtraction. The panel sizes are $4\farcs1 \times 4\farcs1$. }
56: \end{figure}
57: }
58: 
59: %  \includegraphics[angle=0,width=8.0cm]{f3a.ps}  
60: %  \includegraphics[angle=0,width=8.0cm]{f3b.ps}
61: \def\figc{
62: \begin{figure}
63:   \includegraphics[angle=0,width=8.0cm]{f3.ps}  
64: \caption{The observed SEDs of the IRAC-detected $z\approx$7 galaxies and the best-fit 
65: Solar metallicity models for each star formation history. The vertical
66: and horizontal bars mark the $1\sigma$ flux uncertainty and the
67: width of the passband. The arrows are $1\sigma$ upper limits.
68: The sources show evidence for a break between the near-infrared and mid-infrared fluxes,
69: which is well fit by a fairly evolved stellar population. The fits generally
70: require little reddening, although galaxy 1417 also allows for moderate 
71: amounts of reddening and fairly large star formation rates (up to $A_V=0.4$ and $SFR\sim25~$M$_\sun~$yr$^{-1}$).}
72: \end{figure}
73: \leavevmode
74: \vspace{-0.2cm}
75: }
76: 
77: \def\taba{
78: \begin{table} 
79: \begin{center}
80: \caption{Corrected Photometry of $z_{850}$-dropouts in the UDF.}
81: \begin{tabular}{ccccccc}
82: \tableline\tableline
83: ID & $z_{850}$ & $J_{110}$  & $H_{160}$ & $K_s$ & 3.6$\mu$m & 4.5$\mu$m \\ 
84: %\multicolumn{1}{c}{$P$\tablenotemark{a}} & $P R_{maj}$ & $P R_{min}$ &
85: \tableline
86: \phs964 &  $>29.6$   & 26.8$\pm$0.2 & 26.8$\pm$0.2 & 27.0$\pm$0.9 &    26.1$\pm$0.2 & 26.4$\pm$0.5\\
87: 1417 & 28.0$\pm$0.2 & 26.7$\pm$0.2 & 26.3$\pm$0.2 & 26.1$\pm$0.3 & 25.3$\pm$0.1 & 25.5$\pm$0.2\\
88: \phs950 &  $>29.6$  & 26.9$\pm$0.2 & 26.6$\pm$0.2 &     $>27.2$  &     27.0$\pm$0.5 & 26.4$\pm$0.5\\
89: 1125 & 29.2$\pm$0.7 & 27.3$\pm$0.4 & 27.2$\pm$0.3 &  $>27.2$  & 26.9$\pm$0.4 &  $>27.2$  \\
90: \tableline
91: \end{tabular}
92: \tablecomments{Object IDs are from B04. 
93: All magnitudes are in the AB system. Optical/near-infrared 
94: fluxes were measured in 0\farcs9 diameter apertures and IRAC fluxes were measured in
95: 2\farcs5 diameter apertures. We corrected the fluxes for missing light outside 
96: the aperture assuming stellar profiles. The corrections are  $5-20$\% in the 
97: optical/near-infrared and a factor 1.75, 1.81 for IRAC $3.6\mu$m, $4.5\mu$m,
98: respectively. We adjusted the NICMOS $J_{110}$ and $H_{160}$ zeropoints 
99: by -0.16 and -0.04 following \citet{Jo06}. In the improved NICMOS reduction (B06)
100: object 1417 falls within the $z_{850}$-dropout selection criteria, whereas 
101: in B04 it fell just outside. The upper limits are $1\sigma$. None of the 
102: galaxies are detected in $V_{606}, i_{775}$, and $5.8\mu$m to $1\sigma$ limits 
103: of 30.6, 30.5, and 25.1, respectively.}
104: \end{center}
105: \leavevmode
106: \vspace{-0.5cm}
107: \end{table}
108: }
109: 
110: \def\tabb{
111: \begin{table}
112: \begin{center}
113: \caption{Best-fit Model Parameters}
114: \setlength{\tabcolsep}{0.025in}
115: \scriptsize
116: \begin{tabular}{ccrrrcc}
117: \tableline\tableline
118:  & \multicolumn{6}{c}{Solar metallicity}  \\
119: ID & \phs\phs$z_{phot}$ & Mass  & Age$_w$ & $A_V$ & SFR & $\chi^2_{red}$ \\
120: & & $(10^9M_\sun)$ & (Myr) & & (\msun~yr$^{-1}$)  \\
121: \tableline
122: \multicolumn{6}{c}{Single Age Burst (SSP)} \\
123: \tableline
124:          964 & \ \ \  7.2 $^{+0.1}_{-0.2}$ &  1.8 $^{+0.7}_{-0.4}$ &    40 & \ \ \ \ \ 0.0 &  0.0 &  1.5  \\
125:         1417 & \ \ \  6.7 $^{+0.2}_{-0.2}$ &  5.8 $^{+0.9}_{-1.0}$ &    81 &  0.1 &  0.0 &  0.8  \\
126:          950 & \ \ \  7.2 $^{+0.1}_{-0.1}$ &  1.0 $^{+0.2}_{-0.1}$ &    25 &  0.0 &  0.0 &  1.4  \\
127:         1125 & \ \ \  7.0 $^{+0.2}_{-0.3}$ &  0.7 $^{+0.4}_{-0.1}$ &    32 &  0.0 &  0.0 &  0.9  \\
128: \tableline
129: \multicolumn{6}{c}{Exponentially Declining SFR (e-folding time $\tau=100$~Myr; $\tau_{100}$)}  \\
130: \tableline
131:          964 &  \ \ \ 7.2 $^{+0.1}_{-0.1}$ &  3.4 $^{+0.8}_{-0.7}$ &   133 &  0.0 &  5.2 &  1.2  \\
132:         1417 &  \ \ \ 6.8 $^{+0.1}_{-0.2}$ &  9.3 $^{+1.3}_{-2.2}$ &   176 &  0.2 & 10.7 &  0.7  \\
133:          950 &  \ \ \ 7.3 $^{+0.1}_{-0.2}$ &  0.6 $^{+0.5}_{-0.3}$ &    35 &  0.0 &  9.4 &  1.1  \\
134:         1125 &  \ \ \ 7.1 $^{+0.1}_{-0.2}$ &  0.9 $^{+0.8}_{-0.6}$ &    77 &  0.0 &  4.4 &  0.8  \\
135: \tableline
136: \multicolumn{6}{c}{Constant Star Formation (CSF)} \\
137: \tableline
138:          964 &\ \ \  7.4 $^{+0.1}_{-0.2}$ &  4.2 $^{+2.4}_{-2.1}$ &   320 &  0.0 &  7.3 &  1.3  \\
139:         1417 &\ \ \  6.8 $^{+0.2}_{-0.2}$ & 11.5 $^{+4.5}_{-3.6}$ &   254 &  0.4 & 25.5 &  0.9  \\
140:          950 &\ \ \  7.3 $^{+0.1}_{-0.2}$ &  0.7 $^{+0.6}_{-0.4}$ &    40 &  0.0 &  9.9 &  1.1  \\
141:         1125 &\ \ \  7.0 $^{+0.2}_{-0.2}$ &  1.0 $^{+1.3}_{-0.6}$ &   127 &  0.0 &  4.5 &  0.8  \\
142: \tableline
143: \end{tabular}
144: \tablecomments{
145: The best-fit parameters for Bruzual \& Charlot (2003) stellar population models 
146: and a \citet{Cal00} obscuration law. We omit the $B$ and $8\mu$m band from the 
147: fit as they are not deep enough to constrain the models, leaving 9 filters
148: ($V_{606},i_{775},z_{850},J_{110},H_{160},K_{s},3.6\mu$m,
149: $4.5\mu$m, and 5.8$\mu$m), and 4 free parameters (redshift, stellar mass, age, 
150: and visual extinction $A_V$). We assume Solar metallicity and a 
151: Salpeter IMF from $0.1-100$~\msun; adopting a Chabrier 
152: IMF would result in similar ages but $\sim1.7 \times$ lower stellar masses.
153: Sub-Solar ($1/50~Z_\sun$) metallicities result in 30\% higher ages
154: and 20\% higher masses. We consider redshifts from 0 to 10 and $A_V$ from 0 to 2.
155: We explore 3 star formation histories (SFHs): SSP, $\tau_{100}$, and
156: CSF. The 68\% confidence intervals are obtained from
157: Monte-Carlo simulations and the reduced $\chi^2_{red}$ values assume 
158: 5 degrees of freedom. We impose a minimum best-fit age of 25Myr 
159: to avoid unrealistically young solutions.
160: Finally, we weight the best-fit age with the SFH to better represent
161: the age of the stars comprising the bulk of the stellar mass (Age$_w$)
162: For CSF the correction is 0.5, for exponentially declining SFHs it is 
163: $[age-\tau+\tau exp(-age/\tau)]/[1-exp(-age/\tau)]$, while SSP
164: requires no correction \citep{Fo04}.}
165: \end{center}
166: \leavevmode
167: \vspace{-0.5cm}
168: \end{table}
169: }
170: %, allowing better comparison between different SFHs. 
171: 
172: 
173: %\slugcomment{}
174: 
175: \shorttitle{Spitzer IRAC confirmation of $z\approx7$ galaxies in the UDF}
176: \shortauthors{Labb\'e et al.}
177: 
178: 
179: %% This is the end of the preamble.  Indicate the beginning of the
180: %% paper itself with \begin{document}.
181: 
182: \begin{document}
183: 
184: %% LaTeX will automatically break titles if they run longer than
185: %% one line. However, you may use \\ to force a line break if
186: %% you desire.
187: 
188: \title{{\it Spitzer} IRAC confirmation of $z_{850}$-dropout galaxies in the Hubble Ultra Deep Field: \\
189:        stellar masses and ages at $z\approx7$\altaffilmark{1}}
190: 
191: %% Use \author, \affil, and the \and command to format
192: %% author and affiliation information.
193: %% Note that \email has replaced the old \authoremail command
194: %% from AASTeX v4.0. You can use \email to mark an email address
195: %% anywhere in the paper, not just in the front matter.
196: %% As in the title, use \\ to force line breaks.
197: 
198: \author{Ivo labb\'e\altaffilmark{2,3},Rychard Bouwens\altaffilmark{4}, G.D. Illingworth\altaffilmark{4},
199: M. Franx\altaffilmark{5}}
200: 
201: %% Notice that each of these authors has alternate affiliations, which
202: %% are identified by the \altaffilmark after each name.  Specify alternate
203: %% affiliation information with \altaffiltext, with one command per each
204: %% affiliation.
205: 
206: \altaffiltext{1}{Based on observations with the {\em Spitzer Space Telescope}, which is 
207: operated by the Jet Propulsion Laboratory, California Institute of Technology under 
208: NASA contract 1407.  Support for this work was provided by NASA through contract 
209: 125790 issued by JPL/Caltech. Based on observations with the
210: NASA/ESA {\em Hubble Space Telescope}, obtained at the Space
211: Telescope Science Institute which is operated by AURA, Inc.,
212: under NASA contract NAS5-26555. Based on service mode observations collected at 
213: the European Southern Observatory, Paranal, Chile (ESO Program 073.A-0764A).
214: }
215: 
216: \altaffiltext{2}{Carnegie Observatories, 813 Santa Barbara Street, Pasadena, CA 91101 [e-mail: {\tt ivo@ociw.edu}]}
217: \altaffiltext{3}{Carnegie fellow}
218: \altaffiltext{4}{Astronomy Department,University of California, Santa Cruz, CA 95064}
219: \altaffiltext{5}{Leiden Observatory, Postbus 9513, 2300 RA Leiden, Netherlands}
220: 
221: \begin{abstract}
222: Using {\it Spitzer} IRAC mid-infrared imaging from the Great Observatories Origins 
223: Deep Survey, we study $z_{850}$-dropout 
224: sources in the Hubble Ultra Deep Field. After carefully removing
225: contaminating flux from foreground sources, we clearly detect 
226: two $z_{850}$-dropouts at $3.6\mu$m and $4.5\mu$m, while two others 
227: are marginally detected. 
228: The mid-infrared fluxes strongly support their interpretation as galaxies at 
229: $z\approx7$, seen when the Universe was only 750~Myr old. 
230: The IRAC observations allow us
231: for the first time  to constrain the rest-frame optical colors, stellar masses, 
232: and ages of the highest redshift galaxies.
233: %
234: Fitting stellar population models to the spectral energy 
235: distributions, we find photometric 
236: redshifts in the range $6.7-7.4$, rest-frame colors $U-V=0.2-0.4$, $V$-band
237: luminosities $L_V=0.6-3 \times 10^{10}~L_\sun$, 
238: stellar masses $1-10 \times 10^{9}$~\msun,
239: stellar ages $50-200$~Myr, star formation
240: rates up to $\sim25$~\msun~yr$^{-1}$, and low reddening $A_V<0.4$.
241: Overall, the $z=7$ galaxies appear substantially less 
242: massive and evolved than Lyman break galaxies or Distant
243: Red Galaxies at $z=2-3$, but fairly similar to recently 
244: identified systems at $z=5-6$. 
245: %
246: The stellar mass density inferred from our $z=7$ sample is 
247: $\rho_* = 1.6^{+1.6}_{-0.8}\times 10^{6}$~\msun Mpc$^{-3}$ (to 0.3$L^*_{z=3}$), in 
248: apparent agreement with recent cosmological hydrodynamic simulations,
249: but we note that incompleteness and sample variance may introduce 
250: larger uncertainties.
251: %
252: The ages of the two most massive galaxies suggest they 
253: formed at $z\gtrsim8$, during the era of cosmic reionization,
254: but the star formation rate density derived from their stellar masses
255: and ages is not nearly sufficient to reionize the universe. 
256: The simplest explanation for this deficiency is that lower-mass galaxies beyond our
257:  detection limit reionized the universe.
258: 
259: \end{abstract}
260: 
261: %% Keywords should appear after the \end{abstract} command. The uncommented
262: %% example has been keyed in ApJ style. See the instructions to authors
263: %% for the journal to which you are submitting your paper to determine
264: %% what keyword punctuation is appropriate.
265: 
266: \keywords{galaxies: evolution --- galaxies: high redshift --- infrared: galaxies }
267: 
268: %% Authors who wish to have the most important objects in their paper
269: %% linked in the electronic edition to a data center may do so by tagging
270: %% their objects with \objectname{} or \object{}.  Each macro takes the
271: %% object name as its required argument. The optional, square-bracket 
272: %% argument should be used in cases where the data center identification
273: %% differs from what is to be printed in the paper.  The text appearing 
274: %% in curly braces is what will appear in print in the published paper. 
275: %% If the object name is recognized by the data centers, it will be linked
276: %% in the electronic edition to the object data available at the data centers  
277: 
278: \section{Introduction}
279: Observations of massive galaxies at high redshift
280:  with the {\it Hubble Space Telescope} and the 
281: {\it Spitzer Space Telescope} are revolutionizing 
282: our knowledge of the early formation history of stars and galaxies. 
283: Blue star forming galaxies at $z=2-3$ with stellar masses
284:  $10^{10}-10^{11}$\msun\ are routinely identified 
285:  from optical imaging (Steidel et al. 1996a,b,2004),
286: and have been studied in detail 
287: \citep[e.g.,][Shapley et al. 2001,2005]{PDF01}, while
288: near-infrared surveys at $z=2-3$ have uncovered substantial numbers of 
289: redder, more evolved galaxies with larger stellar masses
290: $>10^{11}$\msun\  \citep{Fr03,Ya04,Da05}. Some of these red galaxies 
291: appear to have stellar ages $> 1.5$ Gyr,
292: implying that they formed most of their stars before $z\sim5$ \citep{La05},
293: and suggesting that massive galaxies should exist well beyond these redshifts.
294: Direct detection of such galaxies would place strong 
295: constraints on galaxy formation models (e.g., \citealt{So01,Na05}).
296: 
297: Tantalizingly, the most recent surveys with the Advanced Camera for Surveys 
298:  (ACS) and the Near-Infrared Camera and Multiobject Spectrograph (NICMOS) 
299: on the {\it Hubble Space Telescope} have identified 
300: sources at $z=5-6.5$ with fairly evolved stars and
301: stellar masses of $1-4 \times 10^{10}$\msun\ \citep[e.g.,][]{Ya05,Ey05,Do05} or perhaps 
302: more \citep{Mo05}. Critical for these results was access to the rest-frame 
303: wavelengths longward of the Balmer/4000 \AA\ break offered by the 
304: InfraRed Array Camera (IRAC; \citealt{Fa04}) on {\it Spitzer}. 
305: Without mid-infrared photometry to very faint magnitudes the
306: stellar ages and masses of $z\gtrsim5$  galaxies are  poorly constrained.
307: 
308: In this Letter, we extend mass estimates to $z=7-8$
309: by analyzing the mid-infrared fluxes of 6 $z_{850}$-dropout
310: candidates found in the Hubble Ultra Deep Field (UDF)
311: by Bouwens et al. (2004, hereafter B04). These candidates 
312: were selected from exceptionally deep optical ACS 
313: (Beckwith et al. in prep) and near-infrared NICMOS 
314: imaging \citep{Th05}, and when combined with the 
315: ultradeep IRAC data available there, offer us an 
316: ideal opportunity to verify their reality and to study
317: their stellar populations. The stellar masses and ages of $z_{850}$-dropout galaxies 
318: would provide us with the first direct look at galaxy formation at $z\gtrsim7$,
319: building on the comprehensive $z\sim6$ study (Bouwens et al. 2006b).
320: Where necessary, we assume an 
321: $\Omega_M=0.3, \Omega_\Lambda=0.7,$ cosmology with 
322: $H_0=70$~km~s$^{-1}$Mpc$^{-1}$.
323: Magnitudes are expressed in the AB photometric system.  
324: 
325: \ifemulate\figa\fi
326: 
327: \section{Observations}
328: The Great Observatories Origins Deep Survey (GOODS; Dickinson et al., in prep) 
329: observed the UDF with IRAC in two epochs, 
330: each time integrating for $\approx23.3$ hours in the 
331: $3.6, 4.5, 5.8$, and $8.0\mu$m channels.\footnote{This paper uses 
332: data release DR3 of epoch 1 and data release DR2 of epoch 2, available 
333: from \url{http://data.spitzer.caltech.edu/popular/goods/}}
334: We estimate limiting depths
335: in the combined IRAC images by measuring 
336: the effective flux variation in random apertures on 
337: empty background regions. The
338: limits for point sources are 27.7, 27.2, 25.1, and
339: 24.9 in the four channels (1$\sigma$, total, 2\farcs5 diameter
340: aperture).
341: We supplement the observations with deep $K_s$-band
342: data from the {\it Very Large Telescope} and {\it Magellan}
343: (Labb\'e et al. in prep) and we use an independent reduction of 
344: UDF NICMOS data with improved noise properties and fewer
345: artifacts \citep[][hereafter B06]{Bo06}. The new NICMOS images revealed that 
346: 2 of the original 6 $z_{850}$-dropouts were electronic ghosts 
347: of nearby bright stars; hence we removed them from the sample.
348: 
349: Matching ACS/NICMOS and IRAC photometry is challenging
350: because of the much larger size and extended wings of the 
351: IRAC point-spread functions (PSFs) resulting in
352: flux contamination by nearby foreground sources. Visual
353: inspection shows that 2 $z_{850}$-dropouts are substantially
354: blended and all 4 are likely to contain at least some flux 
355: from nearby objects.
356: We have developed a technique to robustly subtract the 
357: contaminating flux (Labb\'e et al., in prep). 
358: Briefly, we first detect sources
359: with SExtractor \citep{BA96} in a summed NICMOS $J_{110} + H_{160}$ image
360: to determine the light distributions at high resolution using the pixels 
361: in the ``segmentation'' maps. We then convolve these template
362: images individually with a carefully constructed kernel to match it 
363: to the IRAC PSF.  Third, we fit all detected sources, including  
364: neighbors, simultaneously to the registered IRAC image, leaving 
365: only the flux scalings as free parameters. 
366: Finally, we subtract the best-fit images of all neighboring sources. 
367: 
368: \ifemulate\taba\fi
369: 
370: After cleaning the IRAC images, we performed conventional
371: aperture photometry in $3.6,4.5,5.8$, and $8.0\mu$m bands
372: in 2\farcs5 diameter apertures. 
373: %We note that this procedure is more robust than using the fitted fluxes directly. 
374: Photometry in the ACS $B_{435},V_{606},i_{775},z_{850}$, 
375: NICMOS $J_{110},H_{160}$, and  $K_s$ bands was done in 
376: 0\farcs9 diameter apertures and we obtained
377: magnitudes and limits consistent with B04.
378: We summarize the photometry in Table~1. 
379: 
380: 
381: \section{Mid-Infrared Fluxes of $z_{850}$-dropout sources}
382: Figure 1 shows the {\it HST}/ACS+NICMOS and the combined
383:  {\it Spitzer}/IRAC images of the $z_{850}$ dropouts. 
384:  Two objects (ID 964 and 1417) are unambiguously detected 
385: in $3.6\mu$m ($5-8 \sigma$) and in the slightly shallower 
386: $4.5\mu$m ($2-3 \sigma)$. Two others 
387: (ID 950 and 1125) are only marginally detected, but 
388: are probably real as the sum of their $3.6$ and $4.5\mu$m 
389: images reveals a visible source. Unfortunately, the IRAC 
390: observations are not deep enough to definitively 
391: confirm or reject the reality of the undetected sources.
392: None of the candidates are visible at $5.8$ and $8.0\mu$m.
393: To evaluate the robustness of the deblending photometry, we performed
394: the procedure independently on the first and second epoch 
395: IRAC data (see Figure~2). Reassuringly, we measure consistent fluxes 
396: and we detect the brightest,
397: most promising sources 964 and 1417 in each dataset individually.
398:  
399: The $3.6\mu$m magnitudes are faint, ranging from 25.3 to 27.0, 
400: with $H_{160}-3.6\mu$m colors in the range -0.4 to 1.0. The IRAC detected
401: objects 964 and 1417 are the reddest, showing a 
402: factor of $\sim2$ rise in $f_\nu$ flux densities between $H_{160}$ 
403: and $3.6\mu$m, while the SEDs are flatter at $1.1-2.2\mu$m and 
404: $3.6-4.5\mu$m (see Table~1). The rise at $3.6\mu$m is similar in 
405: strength to what is found in spectroscopically confirmed galaxies 
406: at $z=6$ \citep{Ya05,Ey05}, and suggests the presence of 
407: a substantial redshifted Balmer break, indicative of evolved 
408: stellar populations.  
409: 
410: \ifemulate\figb\fi
411: 
412: \section{Stellar Populations of $z\approx7$ galaxies}
413: We fit stellar population synthesis models of \citet{BC03} 
414: and a \citet{Cal00} obscuration law to the broadband fluxes to 
415: constrain the stellar populations. The models assume Solar metallicity
416: and a Salpeter initial mass function (IMF) between 0.1 and 100~\msun. 
417: %We use simple  models, 
418: %exploring Solar and  sub-Solar ($1/50~Z_{\sun}$) metallicities 
419: We explore three different star formation histories (SFHs): 
420: a single age burst (SSP), an exponentially declining star 
421: formation rate (SFR) with a timescale of 100~Myr ($\tau_{100}$), and constant 
422: star formation (CSF). We leave the redshift, mass, age, 
423: and extinction as free parameters.
424: 
425: We find acceptable fits for all sources (see Table~2 and Figure~3) and 
426: obtain confidence intervals on the parameters with 
427: Monte Carlo simulations. %, where we randomly vary the fluxes 
428: %200 times within their 1$\sigma$ errors and refit the models.
429: The best-fit redshifts vary from 6.7 to 7.4 and
430: most Monte Carlo solutions are in a narrow range 
431: around the best fit indicating that the redshift is well constrained.
432: %the strong $z_{850}-J_{110}$ break constrains
433: Old stellar populations at $z\sim1$ fit the data
434: poorly as they do not reproduce the strong
435: break across the $i_{775},z_{850}$, and 
436: $J_{110}$ bands and the blue near-infrared continuum. 
437: Even so, we note that 3\%  and 11\% of the solutions 
438: for object 1417 and 1125, allowed a redshift 
439: of $z\sim1$ when the random flux variations 
440: ``conspired'' to suppress the break. The best-fit models have
441: average rest-frame optical colors of $U-V=0.4~(0.2)$ and 
442: $V-$band luminosities $L_V=2.3(1.0) \times 10^{10}L_\sun$ for the IRAC-detected 
443: (undetected) sources.
444: 
445: \ifemulate\tabb\fi
446: \ifemulate\figc\fi
447: 
448: Determinations of the stellar population age and mass
449:  depend on the assumed star formation history and 
450:  metallicity. For the whole sample, $\tau_{100}$
451: models fit the best, with ages of $40-180$~Myr and instantaneous 
452: SFRs of $4-11$~\msun~yr$^{-1}$.
453: Converting the rest-frame $1500~$\AA\ luminosities directly 
454: into (absorption-corrected) SFRs \citep{Ma98} results 
455: in similar values. Nevertheless, the degeneracy between age and dust
456: prevents us from placing firm limits on the ages and SFRs.
457: The stellar masses are generally better constrained. As expected,
458: the IRAC-detected galaxies 964 ($z_{ph}=7.4$) and 1417 ($z_{ph}=6.8$)
459: are the reddest, most massive, and oldest in the sample. 
460: The average uncertainties on the masses are approximately 
461: a factor of $2-3$. 
462: 
463: The extreme SSP and CSF models set lower and 
464: upper boundaries to the masses, ages, and SFRs, whereas
465: assuming sub-Solar metallicities
466: ($1/50~Z_{\sun}$) instead of Solar results in 30\%
467: higher ages and 20\% higher masses.
468: Because all SFHs and metallicities provide equally acceptable 
469: fits to the data, we will hereafter adopt the mean of the 
470: SSP and CSF models and both metallicities 
471: as our fiducial values. We then find ages of $50-200$~Myr, 
472: masses of $1-10 \times 10^9$~\msun, low reddening $A_V<0.4$, 
473: and star formation rates of $3-12$~\msun~yr$^{-1}$.
474: 
475: 
476: \section{Discussion}
477: 
478: %Using deep {\it Spitzer}/IRAC imaging from
479: %GOODS (Dickinson et al. in prep), 
480: Using the GOODS dataset (Dickinson et al. in prep),
481: we have estimated and analyzed the {\it Spitzer}/IRAC 
482: mid-infrared fluxes of 4 $z_{850}$-dropouts
483: candidates, which were identified in the UDF by B04 
484: and remeasured more accurately in B06. The sources
485: are rare, with a surface density of 0.7~arcmin$^{-2}$, 
486: and very faint, with observed magnitudes of $H_{160}=26-27$ 
487: and $3.6\mu=25-27$, placing them well beyond the 
488: spectroscopic capabilities of current telescopes, but 
489: in reach of future facilities such as JWST and ALMA. 
490: IRAC directly confirms the reality of two sources, while 
491: two others are marginally detected. Modeling of the broadband 
492: fluxes strongly supports their interpretation 
493: as $z\approx7$ galaxies with substantial stellar masses 
494: $1-10 \times 10^9$ and ages $50-200$~Myr. 
495: 
496: 
497: Using the redshift selection function for the
498: $z_{850}$-dropout sample (B04), we obtain an effective volume of 
499: 9000 Mpc$^{3}$ to $0.3L^*_{z=3}$ and we infer a stellar mass 
500: density of $\rho_* = 1.6^{+1.6}_{-0.8}\times 10^{6}$~\msun Mpc$^{-3}$.
501: Comparing to the stellar mass density at lower redshifts, computed using 
502: similar techniques to similar luminosities, we find a continuing decrease where 
503: the density at $z\approx7$ is 95\%, 22\% of that at $z=6,5$  \citep{Ya06,Sta06}.
504: Recent Smooth Particle
505: Hydrodynamics (SPH) simulations in a $\Lambda$CDM universe 
506:  predict stellar mass- and number densities for massive ($>1.8 \times 10^{9}$M$_\sun$)
507: galaxies of $0.9 \times 10^{6}$~\msun~Mpc$^{-3}$ and 
508: $2.5 \times 10^{-4}$~Mpc$^{-3}$ (SPH G6 run, \citealt{Na05}),
509: remarkably close to the minimum estimates for our sample  
510: ($0.7 \times 10^{6}$~\msun Mpc$^{-3}$ and $2.2 \times 10^{-4}$~Mpc$^{-3}$ to the same mass limit).
511: However, we cannot exclude additional stellar mass residing
512: in massive non-starforming or dust-enshrouded galaxies, 
513: which the $z_{850}$-dropout criteria would have missed. 
514: %but which could dominate the high-mass end as is the case at $z=2-3$ \citep{vD06}.
515: 
516: The ages of the two most massive (IRAC-detected) $z_{850}$-dropouts 
517: suggest that the bulk of their stellar mass formed at 
518: even higher redshifts $z\gtrsim8$, during the epoch 
519: of cosmic reionization \citep{Sp06}. 
520: We can place a simple upper limit on the contribution
521: of high-mass galaxies to reionization,
522: by calculating the maximum SFR densities implied by the 
523: observed stellar masses and ages for these objects.
524: Taking the maximum masses and dividing it by 
525: the minimum ages ($\sim60$Myr), we infer a substantial 
526: SFR density 0.04~M$_\sun$yr$^{-1}$~Mpc$^{-3}$. This is higher than calculated 
527: directly from the rest-frame UV luminosities at $z=6-7$  
528: to the same limits \citep{Bo06b}, but  still more than $3$ times too small to reionize 
529: the universe at the lowest probable redshift $z_{reion}=8.6$ \citep{Sp06} 
530: for canonical assumptions \citep[][see also Yan et al. 2006]{Ma99,Bo06b}. 
531: 
532: The simplest explanation for the lack of ionizing
533: photons provided by high-mass galaxies at $z=7$
534: is that low-mass galaxies beyond our current
535: detection limits were primarily responsible for reionization
536: \citep{LB03,YW04,Bo06b}.
537: Other possibilities are that massive galaxies are missing 
538: from current surveys and are unaccounted for in the models,
539: or that the observed galaxies had a top heavy IMF, which would 
540: increase the ionizing efficiency per unit stellar mass.
541: Obviously, incompleteness, sample variance, and 
542: large scale structure dominate the uncertainties in our
543: results. Larger, very deep near-infrared surveys from the ground
544: and from space will address this issue in more detail.
545: 
546: \acknowledgments
547: \vskip -0.2cm
548: We are most grateful for the efforts of the GOODS team who provided the
549: data used in this analysis. We thank Ken Nagamine for providing his
550: numerical simulations and the referee for helpful comments. 
551: IL is supported by a fellowship from the Carnegie Institution
552: of Washington. RJB and GDI acknowledge support from NASA grant HST-GO09803.05-A 
553: and NAG5-7697.
554: 
555: %% After the acknowledgments section, use the following syntax and the
556: %% \facility{} macro to list the keywords of facilities used in the research
557: %% for the paper.  Each keyword will be checked against the master list during
558: %% copy editing.  Individual instruments or configurations can be provided 
559: %% in parentheses, after the keyword, but they will not be verified.
560: 
561: %{\it Facilities:} \facility{Spitzer (IRAC)},  \facility{HST (ACS/NICMOS)}, \facility{VLT (ISAAC)}.
562: 
563: \begin{thebibliography}{}
564: \bibitem[Bertin \& Arnouts(1996)]{BA96} Bertin, E.~\& Arnouts, S.\ 1996, \aaps, 117, 393
565: \bibitem[Bouwens et al.(2004)]{Bo04} Bouwens, R.~J., et al.\ 2004, \apjl, 616, L79  (B04)
566: %\bibitem[Bouwens et al.(2005)]{Bo05} Bouwens, R.~J., et al.\ 2005, \apjl, 624, L5 
567: \bibitem[Bouwens \& Illingworth(2006)]{Bo06} Bouwens, R.~J.,  \& Illingworth, G.~D., \ 2006, Nature, in press \\ astro-ph/0607087 (B06)
568: \bibitem[Bouwens et al.(2006b)]{Bo06b} Bouwens et al.\ 2006b, \apj, in press, astro-ph/0509641
569: \bibitem[Bruzual \& Charlot(2003)]{BC03}  Bruzual, G.~\& Charlot, S.\ 2003, \mnras, 344, 1000 (BC03)
570: \bibitem[Calzetti et al.(2000)]{Cal00} Calzetti, D., et al.\ 2000, \apj, 533, 682 
571: \bibitem[Daddi et al.(2005)]{Da05} Daddi, E., et al.\ 2005, \apj, 626, 680 
572: \bibitem[Dow-Hygelund et al.(2005)]{Do05} Dow-Hygelund, C.~C., et al.\ 2005, \apjl, 630, L137 
573: \bibitem[de Jong et al.(2006)]{Jo06} de Jong, R.~S., et al.\ 2006, NICMOS ISR 2006-001
574: \bibitem[Eyles et al.(2005)]{Ey05} Eyles, L.~P., et al.\ 2005, \mnras, 364, 443 
575: \bibitem[Fazio et al.(2004)]{Fa04} Fazio, G.~G., et al.\  2004, \apjs, 154, 10 
576: \bibitem[F\"orster Schreiber et al.(2004)]{Fo04} F\"orster Schreiber, N.~M., et al.\ 2004, \apj, 616, 40 
577: \bibitem[Franx et al.(2003)]{Fr03} Franx, M.~et al.\ 2003, \apjl, 587, L79
578: \bibitem[Labb{\'e} et al.(2005)]{La05} Labb{\'e}, I., et al.\ 2005, \apjl, 624, L81 
579: \bibitem[Lehnert \& Bremer(2003)]{LB03} Lehnert, M.~D., \& Bremer, M.\ 2003, \apj, 593, 630 
580: \bibitem[Madau et al.(1998)]{Ma98} Madau, P., Pozzetti, L., \& Dickinson, M.\ 1998, \apj, 498, 106 
581: \bibitem[Madau et al.(1999)]{Ma99} Madau, P., Haardt, F., \& Rees, M.~J.\ 1999, \apj, 514, 648 
582: \bibitem[Mobasher et al.(2005)]{Mo05} Mobasher, B., et al.\ 2005, \apj, 635, 832 
583: \bibitem[Nagamine et al.(2005)]{Na05} Nagamine, K., et al.\ 2005, \apj, 627, 608 
584: \bibitem[Papovich, Dickinson, \& Ferguson(2001)]{PDF01} Papovich, C., Dickinson, M., \& Ferguson, H.~C.\ 2001, \apj, 559, 620
585: \bibitem[Stark et al.(2006)]{Sta06} Stark et al., 2006, submitted to ApJ, astro-ph/0604250 
586: \bibitem[Steidel et al.(1996a)]{St96a} Steidel, C.~C., et al. \ 1996, \aj, 112, 352
587: \bibitem[Steidel et al.(1996b)]{St96b} Steidel, C. C., et al. 1996b, \apj, 462, L17
588: \bibitem[Steidel et al.(2004)]{St04} Steidel, C.~C., et al.\ 2004, \apj, 604, 534 
589: \bibitem[Somerville et al.(2001)]{So01} Somerville, R.~S., Primack, J.~R., \& Faber, S.~M.\ 2001, \mnras, 320, 504 
590: \bibitem[Shapley et al.(2001)]{Sh01} Shapley, A.~E., et al.\ 2001, \apj, 562, 95 
591: \bibitem[Shapley et al.(2005)]{Sh05} Shapley, A.~E., et al.\ 2005, \apj, 626, 698 
592: \bibitem[Spergel et al.(2006)]{Sp06} Spergel et al., 2006, submitted to \apj
593: \bibitem[Thompson et al.(2005)]{Th05} Thompson, R.~I., et al.\ 2005, \aj, 130, 1 
594: \bibitem[Yan et al.(2004)]{Ya04} Yan, H., et al.\ 2004, \apj, 616, 63 
595: \bibitem[Yan \& Windhorst(2004)]{YW04} Yan, H., \& Windhorst, R.~A.\ 2004, \apjl, 600, L1 
596: \bibitem[Yan et al.(2005)]{Ya05} Yan, H., et al.\ 2005, \apj, 634, 109 
597: \bibitem[Yan et al.(2006)]{Ya06} Yan, H., et al.\ 2006, \apj, in press
598: \end{thebibliography} 
599: 
600: \ifemulate\else
601: %  \clearpage
602:   \figa
603: %  \clearpage
604:   \figb
605: %  \clearpage
606:   \figc
607:   \clearpage
608:   \taba
609:   \clearpage
610:   \tabb  
611: \fi
612: 
613: 
614: \end{document}
615: 
616: