1: %\documentclass[12pt,preprint]{aastex}
2: \documentclass[apjl]{emulateapj}
3: \usepackage{apjfonts}
4: \usepackage{amsmath}
5: %\usepackage{emulateapj5}
6: %\usepackage{onecolfloat}
7: %\usepackage{graphicx}
8: %\usepackage{fancyheadings}
9: %\usepackage{amstex}
10: %\usepackage{ulem}
11: \usepackage{rotating}
12: \usepackage{lscape}
13:
14: \def\arcmin{\hbox{$^\prime$}}
15: \def\arcsec{\hbox{$^{\prime\prime}$}}
16: \def\arcdeg{\mbox{$^\circ$}}%
17: \def\fd{\hbox{$~\!\!^{\rm d}$}}
18: \def\fh{\hbox{$~\!\!^{\rm h}$}}
19: \def\fm{\hbox{$~\!\!^{\rm m}$}}
20: \def\fs{\hbox{$~\!\!^{\rm s}$}}
21:
22: \newcommand{\aapg}{\gtrsim}
23: \newcommand{\aapl}{\lesssim}
24: \newcommand{\cmjj}{\mbox{${\rm cm^{-2}}$}}
25: \newcommand{\etal}{et al.}
26: \newcommand{\hI}{\mbox{${\rm H\ I}$}}
27: \newcommand{\kms}{\mbox{km\ s${^{-1}}$}}
28: \newcommand{\lya}{\mbox{${\rm Ly}\alpha$}}
29: \newcommand{\lyb}{\mbox{${\rm Ly}\beta$}}
30: \newcommand{\lyc}{\mbox{${\rm Ly}\gamma$}}
31: \newcommand{\civ}{\mbox{${\rm C\ IV}$}}
32: \newcommand{\cii}{\mbox{${\rm C\ II}$}}
33: \newcommand{\oi}{\mbox{${\rm O\ I}$}}
34: \newcommand{\ovi}{\mbox{${\rm O\ VI}$}}
35: \newcommand{\nii}{\mbox{${\rm N\ II}$}}
36: \newcommand{\nv}{\mbox{${\rm N\ V}$}}
37: \newcommand{\siiv}{\mbox{${\rm Si\ IV}$}}
38: \newcommand{\siii}{\mbox{${\rm Si\ II}$}}
39: \newcommand{\siiii}{\mbox{${\rm Si\ III}$}}
40: \newcommand{\feiii}{\mbox{${\rm Fe\ III}$}}
41: \def\N#1{{N({\rm #1})}}
42: \def\cm#1{\, {\rm cm^{#1}}}
43: \def\rtp{\, \right ) }
44: \def\ltp{\left ( \,}
45: \def\perd{\;\;\; .}
46: \newcommand{\ibid}{\underline{\makebox[0.5in]{}}.}
47: \newcommand{\civof}{\ion{C}{4}$_{15}$}
48:
49: \special{papersize=8.5in,11in}
50:
51: \begin{document}
52:
53: %\twocolumn[%
54: \lefthead{Chen \etal}
55: \righthead{}
56:
57: \slugcomment{Accepted for Publication in the Astrophysical Journal}
58: %\slugcomment{Submitted to the Astrophysical Journal}
59:
60: \title{High-Redshift Starbursting Dwarf Galaxies Revealed by $\gamma$-ray Burst Afterglows$^{1,2}$}
61: \author{HSIAO-WEN CHEN\altaffilmark{3},
62: DANIEL A.\ PERLEY\altaffilmark{4},
63: LINDSEY K.\ POLLACK\altaffilmark{5},
64: JASON X.\ PROCHASKA\altaffilmark{5},
65: JOSHUA S.\ BLOOM\altaffilmark{4,11},
66: MIROSLAVA DESSAUGES-ZAVADSKY\altaffilmark{6},
67: MAX PETTINI\altaffilmark{7},
68: SEBASTIAN LOPEZ\altaffilmark{8},
69: ALDO DALL'AGLIO\altaffilmark{9}, and
70: GEORGE D.\ BECKER\altaffilmark{10}
71: }
72:
73: \altaffiltext{1}{Based in part on observations made with the NASA/ESA
74: Hubble Space Telescope, obtained at the Space Telescope Science
75: Institute, which is operated by the Association of Universities for
76: Research in Astronomy, Inc., under NASA contract NAS 5-26555. }
77:
78: \altaffiltext{2}{Observations reported here were obtained in part at the
79: Magellan telescopes, a collaboration between the Observatories of the Carnegie
80: Institution of Washington, University of Arizona, Harvard University,
81: University of Michigan, and Massachusetts Institute of Technology.}
82:
83: \altaffiltext{3}{Dept.\ of Astronomy \& Astrophysics and
84: Kavli Institute for Cosmological Physics,
85: University of Chicago, Chicago, IL, 60637, U.S.A. \\
86: {\tt hchen@oddjob.uchicago.edu}}
87:
88: \altaffiltext{4}{Department of Astronomy, 601 Campbell Hall, University of
89: California, Berkeley, CA 94720 }
90:
91: \altaffiltext{5}{UCO/Lick Observatory; University of California, Santa
92: Cruz, Santa Cruz, CA 95064}
93:
94: \altaffiltext{6}{Observatoire de Gen\`eve, 51 Ch. des Maillettes,
95: 1290 Sauverny, Switzerland}
96:
97: \altaffiltext{7}{Institute of Astronomy, Madingley Rd., Cambridge, CB3 0HA, UK}
98:
99: \altaffiltext{8}{Departamento de Astronom\'ia, Universidad de Chile, Casilla 36-D, Santiago, Chile}
100:
101: \altaffiltext{9}{Astrophysikalisches Institut Potsdam, An der Sternwarte 16, D-14482 Potsdam, Germany}
102:
103: \altaffiltext{10}{Carnegie Observatories, 813 Santa Barbara St., Pasadena, CA 91101}
104:
105: \altaffiltext{11}{Sloan Research Fellow}
106:
107: %\newpage
108:
109: \begin{abstract}
110:
111: We present a study of 15 long-duration $\gamma$-ray burst (GRB) host
112: galaxies at $z>2$. The GRBs are selected with available early-time
113: afterglow spectra in order to compare interstellar medium (ISM)
114: absorption-line properties with stellar properties of the host
115: galaxies. In addition to five previously studied hosts, we consider
116: new detections for the host galaxies of GRB\,050820 and GRB\,060206
117: and place 2-$\sigma$ upper limits to the luminosities of the remaining
118: unidentified hosts. We examine the nature of the host galaxy
119: population and find that (1) the UV luminosity distribution of GRB
120: host galaxies is consistent with expectations from a UV luminosity
121: weighted random galaxy population with a median luminosity of $\langle
122: L(UV)\rangle=0.1\,L_*$; (2) there exists a moderate correlation
123: between UV luminosity and Si\,II $\lambda\,1526$ absorption width,
124: which together with the observed large line widths of $W(1526)> 1.5$
125: \AA\ for a large fraction of the objects suggests a galactic outflow
126: driven velocity field in the host galaxies; (3) there is tentative
127: evidence for a trend of declining ISM metallicity with decreasing
128: galaxy luminosity in the star-forming galaxy population at $z=2-4$;
129: (4) the interstellar UV radiation field is found $\approx
130: 35-350\times$ higher in GRB hosts than the Galactic mean value; and
131: (5) additional galaxies are found at $\aapl 2''$ from the GRB host in
132: all fields with known presence of strong Mg\,II absorbers, but no
133: additional faint galaxies are found at $\aapl 2''$ in fields without
134: strong Mg\,II absorbers. Our study confirms that the GRB host
135: galaxies (with known optical afterglows) are representative of
136: unobscured star-forming galaxies at $z>2$, and demonstrates that high
137: spatial resolution images are necessary for an accurate identification
138: of GRB host galaxies in the presence of strong intervening absorbers.
139:
140: \end{abstract}
141:
142: \keywords{gamma rays: bursts---ISM: abundances---ISM:
143: kinematics---intergalactic medium}
144: %]
145:
146: \section{INTRODUCTION}
147:
148: Gamma-ray bursts (GRBs) are among the most energetic events in the
149: universe. In particular, long-duration GRBs are believed to originate
150: in the catastrophic death of massive stars (e.g.\ Woosley 1993;
151: Paczy\'nski 1998; Bloom \etal\ 2002; Stanek \etal\ 2003; Hjorth \etal\
152: 2003; see Woosley \& Bloom 2006 for a recent review). Since massive
153: stars evolve rapidly, long-duration GRBs should probe instantaneous
154: star formation out to the highest redshifts (e.g.\ Wijers \etal\ 1998)
155: with the afterglows serving as signposts to starburst galaxies in the
156: distant universe.
157:
158: Many bursts are followed by optical afterglows that can briefly
159: exceed the absolute brightness of any known quasar by orders of
160: magnitude (e.g.\ Akerlof \etal\ 1999; Kann \etal\ 2007; Bloom \etal\
161: 2008) and serve as bright background sources for probing intervening
162: gas along the line of sight. Early-time, high-resolution spectroscopy
163: of GRB afterglows have revealed numerous absorption features produced
164: by ground-state and excited-state ions in the interstellar medium
165: (ISM) of the host galaxies (e.g.\ Vreeswijk \etal\ 2004; Prochaska,
166: Chen, \& Bloom 2006; Vreeswijk \etal\ 2007; D'Elia \etal\ 2008).
167: Detailed studies based on comparisons of absorption-line strengths
168: have yielded accurate constraints on the host ISM properties,
169: including gas density, temperature, chemical composition, and
170: kinematics of the GRB host environment (e.g.\ Fynbo \etal\ 2006;
171: Savaglio 2006; Prochaska \etal\ 2007a). Specifically, roughly 50\% of
172: known GRBs at $z>2$ are found in ISM of neutral gas column density
173: $N(\hI)>10^{21}$ \cmjj\ (Jakobsson \etal\ 2006; Chen \etal\ 2007a). In
174: addition, the host ISM generally exhibit moderate chemical enrichment with a
175: median metallicity of $> 1/10$ solar, although with a substantial
176: scatter over the range from $1/100$ to $\sim 1/2$ solar values (Fynbo
177: \etal\ 2006a; Savaglio 2006; Prochaska \etal\ 2007a). Comparisons of
178: different ionic abundances also show that there exists a large
179: differential depletion with $[{\rm Zn}/{\rm Fe}]> +0.6$ dex
180: and $\alpha/{\rm Fe} \aapg 0.4$ dex, confirming the presence of a large
181: amount of gas mass and a chemical enrichment history dominated by
182: massive stars (Savaglio 2006; Prochaska \etal\ 2007a). Finally, there
183: is a lack of molecular gas despite the presence of a large $N(\hI)$
184: (Fynbo \etal\ 2006; Tumlinson \etal\ 2007).
185:
186: Interpretations of these absorption-line data are not
187: straightforward. In particular, the observed low-metal content in the
188: GRB host ISM, as opposed to optically selected luminous starburst
189: galaxies (e.g.\ Shapley \etal\ 2004), can be explained if the
190: progenitor stars originate in the outskirts of a luminous galaxy or if
191: the host galaxies are underluminous and have on average lower
192: metallicities. A local-galaxy analogue of this is seen with the
193: so-called luminosity--metallicity relation (e,g.\ Tremonti \etal\ 2004
194: for nearby galaxies). Recently, Fynbo \etal\ (2008) considered both
195: scenarios and showed that the metallicity distribution of GRB hosts is
196: consistent with the expectation that these host galaxies represent a
197: weighted star-forming galaxy population according to the on-going star
198: formation rate (SFR). In addition, the large atomic gas column
199: density in contrast to the lack of molecular gas may be due to either
200: an enhanced UV radiation field in the star-forming regions near the
201: GRB progenitor or a relatively low dust and metal content indicated in
202: the absorption-line data (e.g.\ Tumlinson \etal\ 2007; Whalen \etal\
203: 2008). Recent detections of the 2175-\AA\ dust absorption feature in
204: GRBs\,070802 (Kr\"uhler \etal\ 2008; Eliasdottir \etal\ 2008) and
205: 080805 (Jakobsson \etal\ 2008) offer an important test for this
206: scenario. Regardless, these issues have direct impact on our
207: understanding of both the GRB progenitors and star formation physics.
208: Supplemental imaging and spectroscopic observations of the host
209: galaxies are necessary for accurate interpretations of the
210: absorption-line measurements.
211:
212: The transient nature of optical afterglows allows deep imaging and
213: spectroscopic studies of galaxies close to the lines of sight,
214: including the hosts, when the afterglows disappear (c.f.\ M{\o}ller
215: \etal\ 2002a; Chen \& Lanzetta 2003 for searches of damped \lya\
216: absorbing galaxies along quasar sightlines). At $z>2$, where accurate
217: ISM abundance measurements are available based on afterglow
218: absorption-line spectroscopy, only four host galaxies have been
219: unambiguously identified (see Savaglio \etal\ 2008 for a compilation).
220: Comparison studies between known ISM properties from afterglow
221: absorption spectroscopy and the observed morphology and luminosity of
222: the host galaxies have been published individually for GRB\,000926
223: (Castro \etal\ 2003), GRB\,011211 (Vreeswijk \etal\ 2006), GRB\,021004
224: (Fynbo \etal\ 2005), and GRB\,030323 (Vreeswijk \etal\ 2004). The
225: four GRB host galaxies together show an order-of-magnitude scatter in
226: their rest-frame UV luminosity and ISM metallicity.
227:
228: At $z<2$, where the majority of common ISM absorption features occur
229: at ultraviolet wavelengths and spectroscopic observations become
230: challenging on the ground, $>40$ GRB host galaxies have been
231: identified in late-time imaging follow-up. These known host galaxies
232: provide important insights for understanding the nature of GRB
233: progenitors at intermediate redshifts. First, Chary \etal\ (2002)
234: compared the estimated SFR and total stellar mass for 12 GRB host
235: galaxies and found that the host galaxies have on average higher SFR
236: per unit stellar mass than local starburst galaxies. Le Floc'h \etal\
237: (2003) examined the optical and near-infrared $R-K$ colors and
238: rest-frame $B$-band luminosity function of 15 GRB host galaxies at
239: $\langle z\rangle\approx 1$.
240: % including 971214 at z=3.42; 000131 is not detected.
241: They found that these host galaxies have on average bluer colors and
242: fainter luminosity ($\langle L_{B}\rangle \approx 0.1\,L_*$) than
243: random star-forming galaxies at the same redshift range. Additional
244: mid-infrared imaging observations by these authors showed that this is
245: not due to dust extinction (Le Floc'h \etal\ 2006). Similarly,
246: Christensen \etal\ (2004) studied the optical and near-infrared
247: spectral energy distributions (SEDs) of ten GRB host galaxies at
248: $\langle z\rangle=0.85$.
249: % with three at z>1; they were 000926 at z=2.037, 990123 at z=1.600,
250: % and 000418 at z=1.118
251: They found based on a comparison with $\sim 1000$ galaxies identified
252: at a similar redshift range in the Hubble Deep Fields that GRB host
253: galaxies have on average younger stellar age and shorter
254: characteristic star-forming time scale. In addition, Fruchter \etal\
255: (2006) compared HST images of $>40$ GRB host galaxies at $\langle
256: z\rangle\approx 1$ with the host galaxies of core-collapse supernovae
257: (SNe). They found that the majority of GRB host galaxies exhibit
258: irregular morphology, unlike the host galaxies of core-collapse SNe.
259: Recently, independent studies by Castro Cer\'on \etal\ (2008) and
260: Savaglio \etal\ (2008) show that $\langle z\rangle\aapl 1$ GRB host
261: galaxies contain on average lower stellar mass than field star-forming
262: galaxies. Together, these results show that GRB host galaxies at
263: $\langle z\rangle\approx 1$ represent a relatively young dwarf
264: population that have experienced recent on-going star-forming
265: episodes.
266:
267: While GRB host galaxies based on the intermediate-redshift sample
268: appear to be underluminous and low mass systems, it is not clear
269: whether the long-duration GRBs originate preferentially in relatively
270: metal dificient star-forming regions (c.f.\ Wolf \& Podsiadlowski
271: 2007; Modjaz \etal\ 2008). A low-metallicity environment is favored by
272: popular progenitor models so that the progenitor star can preserve
273: high spin and a massive stellar core to produce a GRB (e.g.\ Hirschi
274: \etal\ 2005; Yoon \& Langer 2005; Woosley \& Heger 2006).
275: %Constraints
276: %on the chemical content in the ISM of distant starburst galaxies
277: %selected by GRBs therefore offer important insights both for testing
278: %GRB progenitor models and for
279: Chemical abundance measurements for $z\aapl 2$ galaxies have been based
280: primarily on emission line observations of associated H\,II regions.
281: A subset of the host galaxies at $z<1$ have been followed up
282: spectroscopically for measuring emission-line fluxes. A mean
283: metallicity of roughly $1/4$ solar is found but with a large scatter
284: (e.g.\ Sollerman \etal\ 2005; Modjaz \etal\ 2008; Savaglio \etal\
285: 2008). The accuracy of emission-line-based abundance estimates
286: depends sensitively on the accuracy of the calibrations between
287: different line diagnostics (e.g.\ Kewly \& Dopita 2002; Skillman
288: \etal\ 2003; Kennicutt \etal\ 2003). Whether or not there exists a
289: maximum metallicity for forming a GRB remains an open question.
290:
291: We have carried out an optical and near-infrared imaging survey of
292: fields around 15 GRBs at $z>2$. The GRBs are selected to have
293: early-time afterglow spectra in order to compare ISM absorption-line
294: properties with stellar properties. The goal is to identify the host
295: galaxies and constrain their rest-frame UV and optical luminosities.
296: The primary objectives are (1) to quantify the luminosity distribution
297: of the GRB host galaxy populations and investigate whether or not the
298: GRB host galaxies trace the typical star-forming galaxies at high
299: redshift and (2) to examine whether there exists a correlation between
300: the ISM metal content and host luminosity. The starburst nature of
301: GRB hosts makes this galaxy sample a unique laboratory for studying
302: star formation physics and stellar feedback at high redshift. We
303: adopt a $\Lambda$CDM cosmology, $\Omega_{\rm M}=0.3$ and
304: $\Omega_\Lambda = 0.7$, with a dimensionless Hubble constant $h =
305: H_0/(100 \ {\rm km} \ {\rm s}^{-1}\ {\rm Mpc}^{-1})$ throughout the
306: paper.
307:
308: %\renewcommand{\thefootnote}{\fnsymbol{footnote}}
309: \setcounter{footnote}{0}
310:
311: \section{The GRB SAMPLE}
312:
313: \begin{center}
314: \begin{small}
315: \begin{deluxetable*}{p{1in}cccccccrc}
316: %\tabletypesize{\tiny}
317: %\rotate
318: \tablewidth{0pt}
319: \tablecaption{Summary of the Optical and Near-infrared Imaging Data}
320: \tablehead{\multicolumn{1}{c}{Field} & \colhead{RA(J2000)} & \colhead{Dec(J2000)} & \colhead{$z_{\rm GRB}$} & \colhead{$\log\,N(\hI)$} & \colhead{References\tablenotemark{a}} & \colhead{Instrument} & \colhead{Filter} & \colhead{EXPTIME (s)} & \colhead{FWHM (\arcsec)}}
321: \startdata
322: GRB\,011211 \dotfill & 11:15:17.98 & $-$21:56:56.2 & 2.140 & $20.4\pm 0.2$ & (1) & HST/STIS\tablenotemark{b} & Clear & 19544 & 0.1 \nl
323: & & & & & & Magellan/PANIC & $H$ & 12960 & 0.4 \nl
324: GRB\,020124 \dotfill & 09:32:50.81 & $-$11:31:10.6 & 3.198 & $21.7\pm 0.2$ & (2) & HST/STIS\tablenotemark{c} & Clear & 24798 & 0.1 \nl
325: & & & & & & Magellan/PANIC & $H$ & 13860 & 0.4 \nl
326: GRB\,030323 \dotfill & 11:06:09.38 & $-$21:46:13.3 & 3.372 & $21.90\pm 0.07$ & (3) & HST/ACS/WFC\tablenotemark{d} & F606W & 5928 & 0.1 \nl
327: & & & & & & Magellan/PANIC & $H$ & 14400 & 0.4 \nl
328: GRB\,030429 \dotfill & 12:13:07.50 & $-$20:54:49.7 & 2.658 & $21.6\pm 0.2$ & (4) & Magellan/PANIC & $H$ & 7740 & 0.4 \nl
329: GRB\,050401 \dotfill & 16:31:28.81 & $+$02:11:14.2 & 2.899 & $22.6 \pm 0.3$ & (5) & Magellan/PANIC & $H$ & 6300 & 0.6 \nl
330: & & & & & & Keck/LRIS & $g$ & 2540 & 1.1 \nl
331: & & & & & & Keck/LRIS & $R_c$ & 2460 & 1.0 \nl
332: GRB\,050730 \dotfill & 14:08:17.13 & $-$03:46:16.7 & 3.968 & $22.15 \pm 0.10$ & (6) & Magellan/MagIC & $i'$ & 2700 & 0.6 \nl
333: & & & & & & Keck/LRIS & $g$ & 3900 & 1.2 \nl
334: & & & & & & Keck/LRIS & $R_c$ & 3900 & 1.0 \nl
335: GRB\,050820A \dotfill & 22:29:38.11 & $+$19:33:37.1 & 2.615 & $21.0 \pm 0.1 $ & (7) & HST/ACS/WFC\tablenotemark{e} & F625W & 2238 & 0.1 \nl
336: & & & & & & HST/ACS/WFC & F775W & 4404 & 0.1 \nl
337: & & & & & & HST/ACS/WFC & F850LP & 14280 & 0.1 \nl
338: & & & & & & Magellan/PANIC & $H$ & 8460 & 0.6 \nl
339: & & & & & & Keck/LRIS & $g$ & 2620 & 0.7 \nl
340: & & & & & & Keck/LRIS & $R_c$ & 2500 & 1.9 \nl
341: GRB\,050908 \dotfill & 01:21:50.75 & $-$12:57:17.2 & 3.343 & $17.55 \pm 0.10$ & (8) & Magellan/PANIC & $H$ & 12150 & 0.5 \nl
342: GRB\,050922C \dotfill & 21:09:33.08 & $-$08:45:30.2 & 2.199 & $21.5 \pm 0.1$ & (7) & Magellan/PANIC & $H$ & 12560 & 0.5 \nl
343: & & & & & & Keck/LRIS & $g$ & 3870 & 1.2 \nl
344: & & & & & & Keck/LRIS & $R_c$ & 3660 & 1.0 \nl
345: GRB\,060206 \dotfill & 13:31:43.42 & $+$35:03:03.6 & 4.048 & $20.85 \pm 0.10$ & (9) & HST/ACS/WFC\tablenotemark{f} & F814W & 9886 & 0.1 \nl
346: GRB\,060607 \dotfill & 21:58:50.40 & $-$22:29:46.7 & 3.075 & $16.85 \pm 0.10$ & (10) & Magellan/PANIC & $H$ & 18540 & 0.5 \nl
347: & & & & & & Keck/LRIS & $g$ & 2550 & 1.0 \nl
348: & & & & & & Keck/LRIS & $R_c$ & 2430 & 1.5 \nl
349: GRB\,070721B \dotfill & 02:12:32.97 & $-$02:11:40.4 & 3.626 & $21.50\pm 0.20$ & (11) & Magellan/PANIC & $H$ & 10320 & 0.5 \nl
350: \hline
351: \multicolumn{10}{c}{Previously Published Fields} \nl
352: \hline
353: GRB\,000301C \dotfill & 16:20:18.60 & $+$29:26:36.0 & 2.040 & $21.2\pm 0.5$ & (12) & HST/STIS\tablenotemark{g} & Clear & 16422 & 0.1 \nl
354: GRB\,000926 \dotfill & 17:04:09.62 & $+$51:47:11.2 & 2.038 & $21.3\pm 0.3$ & (13) & \multicolumn{4}{l}{Castro \etal\ (2003)} \nl
355: GRB\,021004 \dotfill & 00:26:54.68 & $+$18:55:41.6 & 2.329 & $19.5\pm 0.5$ & (14) & \multicolumn{4}{l}{Fynbo \etal\ (2005)}
356: \enddata
357: \tablenotetext{a}{(1): Vreeswijk \etal\ (2006); (2) Hjorth \etal\ (2003); (3) Vreeswijk \etal\ (2004); (4) Jakobsson \etal\ (2004); (5) Watson \etal\ (2006); (6) Chen \etal\ (2005); (7) Prochaska \etal\ (2007); (8) Fynbo, private communication; (9) Fynbo \etal\ (2006); (10) Chen \etal\ (2007), see also \S\ 4.11; (11) Malesani \etal\ (2007); (12) Jensen \etal\ (2001); (13) Fynbo \etal\ (2002); (14) Fynbo \etal\ (2005), but see \S\ 4.13 for discussion.}
358: \tablenotetext{b}{The imaging data were retrieved from the HST data archive; PID $= 8867$. The host galaxy was clearly detected in these images that have been analyzed and published in Jakobsson \etal\ (2003).}
359: \tablenotetext{c}{The imaging data were retrieved from the HST data archive; PID $= 9180$. A detailed analysis of these space images is published in Berger \etal\ (2002). No emission from the host galaxy was found.}
360: \tablenotetext{d}{The imaging data were retrieved from the HST data archive; PID $= 9405$. The host galaxy was found in the first-epoch images (exptime $=$ 1920 s), which have been analyzed and published by Vreeswijk \etal\ (2004).}
361: \tablenotetext{e}{The imaging data were retrieved from the HST data archive; PID $= 10551$.}
362: \tablenotetext{f}{The imaging data were obtained through our own program before ACS failed; PID $= 10817$.}
363: \tablenotetext{g}{The imaging data were retrieved from the HST data archive; PID $= 8189$. The host galaxy was marginally detected in these images that have been analyzed and published in Fruchter \etal\ (2006).}
364: \end{deluxetable*}
365: \end{small}
366: \end{center}
367:
368: We generated a sample of 15 GRBs at $z_{\rm GRB}>2$ for studying the
369: nature of high-redshift GRB host galaxies and for comparing the host
370: properties with those of field galaxies in deep surveys. The GRBs
371: were selected to have early-time, moderate-to-high resolution
372: afterglow spectra available for measuring the underlying neutral
373: hydrogen column density. In addition, eight of the GRBs have
374: sufficiently high-spectral resolution for constraining the chemical
375: abundances of the host ISM. Only five of these fields have late-time
376: images published in the literature. To identify the stellar
377: counterpart of the host galaxies, we have carried out a near-infrared
378: imaging survey of these fields using PANIC (Martini \etal\ 2004) and
379: the $H$ filter on the Magellan Baade telescope on Las Campanas, Chile.
380: In addition, we have obtained and analyzed late-time deep optical
381: images available for some of these GRB fields from either our own
382: observations using the Advance Camera for Surveys (ACS; Ford \etal\
383: 1998) or unpublished data found in the Hubble Space Telescope (HST)
384: data archive\footnote{http://archive.stsci.edu/hst/.}. A summary of
385: the fields is presented in Table 1, where we list the GRBs, their
386: coordinates, redshift, and inferred neutral hydrogen column density,
387: $N(\hI)$, in columns (1) through (5). The corresponding references
388: for the redshift and $N(\hI)$ measurements are listed in column (6).
389:
390: This GRB host sample presented here is the largest sample of GRBs at
391: $z_{\rm GRB}>2$ for which both ISM absorption properties and
392: constraints on the emission properties of the host galaxies are
393: available. It offers a unique opportunity to carry out a systematic
394: study to understand the nature of starburst galaxies hosting GRBs at
395: high redshift. The redshifts of the GRBs in our sample span a range
396: from $z=2.04$ to $z=4.05$ (left panel of Figure 1). The neutral
397: hydrogen column densities of the GRB host ISM span a range from
398: $\log\,N(\hI)=16.9$ to $\log\,N(\hI)=22.6$ (right panel of Figure 1).
399: The sample size is restricted by the amount of available observing
400: time. In comparison to GRBs with known redshift or $N(\hI)$ in the
401: literature, we show in Figure 1 that our sample is representative of
402: the spectroscopically identified GRB population in the redshift and
403: $N(\hI)$ parameter space. Note that three of the 15 known GRB host
404: absorbers do not contain high $N(\hI)$ ($\log\,N(\hI)\ge 20.3$) that
405: would qualify them as a damped \lya\ absorber (e.g.\ Wolfe \etal\
406: 2005). The observed low $N(\hI)$ indicates that the ISM in front of
407: these GRBs is mostly ionized.
408:
409: \begin{figure*}
410: \begin{center}
411: \includegraphics[scale=0.35,angle=0]{nz.eps}
412: \includegraphics[scale=0.35,angle=0]{nh.eps}
413: \end{center}
414: %\plottwo{f1a.eps}{f1b.eps}
415: \caption[]{{\it Left}: Redshift distribution of the 15 GRBs in our
416: sample (shaded histogram), in comparison to the distributions of all
417: GRBs with known redshifts (open histogram) and of those found by the
418: {\it Swift} Satellite (Gehrels \etal\ 2004; dashed histogram). {\it
419: Right}: Neutral hydrogen column density distribution of the GRBs in
420: our sample in shaded histograms, in comparison to the distribution of
421: known host $N(\hI)$ for GRBs found prior to July 2007 (see Chen \etal\
422: 2007a).}
423: \end{figure*}
424:
425: \section{IMAGING OBSERVATIONS AND DATA ANALYSIS}
426:
427: To constrain the star formation and/or stellar population of the
428: host galaxies, we have carried out an optical and near-infrared
429: imaging survey of nine GRB fields that have not been studied before.
430: We have also obtained new near-infrared images of the fields around
431: GRB\,011211 (previously studied by Jakobsson \etal\ 2003) and
432: GRB\,030323 (previously studied by Vreeswijk \etal\ 2004), and
433: analyzed additional HST images of the field around GRB\,030323 that
434: were not included in Vreeswijk \etal\ (2004). At $z<3$, near-infrared
435: images offer valuable constraints for the intrinsic luminosity of the
436: host galaxies at rest-frame optical wavelengths, while optical images
437: provide constraints for their rest-frame UV luminosities. Here we
438: describe relevant imaging observations and data processing.
439:
440: \subsection{Optical Images from the Hubble Space Telescope}
441:
442: High spatial resolution and high sensitivity optical images obtained
443: using the Wide Field Channel (WFC) of Advance Camera for Surveys (ACS)
444: on board the Hubble Space Telescope (HST) are available for
445: GRBs\,030323, 050820A, and 060206. The field surrounding GRB\,030323
446: was observed under program ID 9405 (PI: Fruchter) using ACS/WFC and
447: the F606W filter during July 2003 and December 2003. The observations
448: were carried out in a sequence of four exposures of between 480 and
449: 522 s each. The images were retrieved from the HST data archive. The
450: field surrounding GRB\,050820A was observed under program ID 10551
451: (PI: Kulkarni) using ACS/WFC and the F625W, F775W, and F850LP filters
452: during September 2005 and June 2006. The observations were carried
453: out in a sequence of two to four exposures of between 400 and 807 s
454: each. The images were retrieved from the HST data archive. The field
455: surrounding GRB\,060206 was observed under our own program (PID=10817;
456: PI: Chen) using ACS/WFC and the F814W filter during November and
457: December 2006. The observations were carried out in the standard
458: ``ACS-WFC-DITHER-BOX'' pattern of exposures between 1215 and 1256 s
459: each.
460:
461: Individual exposures were reduced using standard pipeline
462: techniques, corrected for geometric distortion using drizzle,
463: registered to a common origin, filtered for deviant pixels based on a
464: $5\,\sigma$ rejection criterion, and combined to form a final stacked
465: image. A summary of the optical imaging observations is presented in
466: columns (7) through (10) of Table 1, which lists for each field the
467: instrument and filter used, total exposure time, and the full width at
468: half maximum (FWHM) of the median point spread function (PSF) as
469: determined from point sources. We note that in addition to the GRBs
470: listed in Table 1 we will include in the following analysis previous
471: HST imaging observations of GRB\,000926 ($z_{\rm GRB}=2.038$) and
472: GRB\,021004 ($z_{\rm GRB}=2.323$) published by Fynbo \etal\ (2002),
473: Castro \etal\ (2003) and Fynbo \etal\ (2005), respectively.
474:
475: \subsection{Optical and Near-infrared Images from the Magellan and Keck Telescopes}
476:
477: Optical $i'$ images of GRB\,050730 were obtained using MagIC on the
478: Magellan Baade telescope in June 2008. The observations were carried
479: out in one set of three exposures, 900 s in duration. Dither offsets
480: of 15\arcsec\ were applied between exposures. The sky condition was
481: photometric with a mean seeing of 0.6\arcsec. Individual exposures
482: were first corrected for pixel-to-pixel variation using a flat-field
483: image formed by median filtering sky images obtained during even
484: twilight. Fringes are not apparent in the MagIC-$i'$ images.
485: Next,the processed individual images obtained on the same night were
486: registered to a common origin, filtered for deviant pixels based on a
487: $5\,\sigma$ rejection criterion and a bad pixel mask formed using the
488: flat-field frames, and stacked according to a weighting factor that is
489: proportional to the inverse of the sky variance. The photometric zero
490: points were determined using five SDSS southern standard stars (Smith
491: \etal\ 2002) observed during the night.
492:
493: Optical $R_c$ and $g$ images of GRBs\,050401, 050730, 050820A,
494: 050922C, and 060607 were obtained using the Low Resolution Imaging
495: Spectrometer (LRIS; Oke \etal\ 1995) on the the Keck I telescope in
496: May and July of 2006. Integration times varied from source to source
497: and are given in Table 1, but generally consisted of dithered
498: exposures of about 600~s each and total integration times of 40 to 65
499: minutes per field. Individual exposures were reduced and combined
500: using standard techniques. Photometric calibration was performed
501: using a series of exposures on the Landolt field Markarian A at
502: three different elevations during the July run and a single visit to
503: the PG\,2213 field during the May run. The observation conditions
504: were photometric. Astrometry was performed using a large sample of
505: USNO B1.0 standard stars in each field.
506:
507: Near-infrared $H$ images of GRBs\,011211, 030323, 050401, 050820A,
508: and 050922c were obtained using PANIC on the Magellan Baade telescope
509: in February 2004, May 2006, and August 2007. The observations were
510: carried out in five or nine sets of three to four exposures, 45 to 60
511: s in duration. Dither offsets of eight to 15 arcsec were applied
512: between different sets of exposures in a slanted square pattern.
513:
514: Individual exposures were first corrected for pixel-to-pixel
515: variation using a flat-field image formed by median filtering all the
516: images obtained on the same night. Next, we corrected for geometric
517: distortion in individual flat-field images using the IRAF {\it geomap}
518: task, according to a 2D distortion map provided by the PANIC
519: instrument team. Next,the processed individual images obtained on the
520: same night were registered to a common origin, filtered for deviant
521: pixels based on a $5\,\sigma$ rejection criterion and a bad pixel mask
522: formed using the flat-field frames, and stacked according to a
523: weighting factor that is proportional to the inverse of the sky
524: variance. Images obtained during non-photometric nights were scaled
525: to match the fluxes of common objects observed during photometric
526: conditions. Individual stacked images from different nights were
527: combined to form a final image of each field. The photometric zero
528: points were determined using three to five IR standard stars (Persson
529: \etal\ 1998) observed under photometric conditions. A summary of the
530: near-infrared imaging observations is presented in Table 1.
531:
532: \subsection{Astrometry}
533:
534: An accurate astrometric solution is necessary for the images
535: obtained at late times, in order to correctly identify the host
536: galaxies of the GRBs. To obtain an accurate astrometric solution for
537: each final stacked image, we first calibrate the astrometry using
538: $\aapl\, 12$ USNO stars with a low-order polynomial fit. Next, we
539: refine the astrometric solution using $\aapg 2$ 2MASS stars in the
540: image by adjusting field offsets and rotation. We find that the final
541: astrometric solution is accurate to am r.m.s.\ of 0.2\arcsec.
542:
543: \section{DESCRIPTION OF INDIVIDUAL FIELDS}
544:
545: Including previously published images around GRBs 000301C, 000926
546: and 021004, we have now collected deep optical and/or near-infrared
547: images surrounding 15 GRBs at $z>2$. All of these GRBs have
548: early-time afterglow spectra available that allow us to determine both
549: the gas properties of the GRB host galaxies and the line-of-sight
550: properties of absorbers foreground to the hosts. We will show that
551: known line-of-sight properties from afterglow spectra are important
552: for accurately identifying the GRB host galaxies (e.g.\ GRB\,030429,
553: GRB\,060206, GRB\,070721B, and see Pollack \etal\ 2008 for a rich
554: galaxy field around GRB\,060418). Here we describe contraints on the
555: emission properties of individual host galaxies from available imaging
556: data, together with a brief description of known absorption-line
557: properties from afterglow spectroscopy. A summary of the emission and
558: absorption properties of these $z>2$ GRB host galaxies is presented in
559: Table 3.
560:
561: \subsection{GRB\,011211 at $z=2.142$}
562:
563: \begin{figure}
564: \begin{center}
565: \includegraphics[scale=0.32,angle=0]{g011211_tot.ps}
566: \end{center}
567: %\plotone{f2.ps}
568: \caption[]{Registered optical (left three panels) and near-infrared
569: $H$ (right panel) images of the field around GRB\,011211 at $z_{\rm
570: GRB}=2.142$ after the afterglow had faded. The epoch during which the
571: images were taken is indicated at the bottom of each panel. The host
572: appears to consist of three compact regions, all of which are
573: confirmed to be at the host redshift based on the presence of \lya\
574: emission in ground-based narrow-band imaging follow-up (Fynbo \etal\
575: 2003a). The GRB is found by Jakobsson \etal\ (2003) to originate in
576: the fainter region, southwest of the center blob, as indicated by the
577: fading optical transient (arrow in the left panel). The mean FWHM of
578: the PSF in the $H$ image was found to be $\approx 0.37\arcsec$ based
579: on an average of 10 stars across the PANIC field. Two of the compact
580: regions are detected in the $H$-band image, but the region directly
581: associated with the burst does not exhibit observable flux. We
582: measure an $H$-band magnitude of $AB(H)=25.0\pm 0.3$ over a 1\arcsec\
583: diameter aperture for the host galaxy.}
584: \end{figure}
585:
586: This burst was detected by BeppoSAX and the optical transient was
587: found $\approx 10$ hours later with $R=19$ (Grav \etal\ 2001).
588: Low-resolution (${\rm FWHM}\equiv\delta\,v \approx 680$ \kms) optical
589: spectra of the afterglow were taken by different groups, which
590: revealed multiple absorption features indicating a source redshift of
591: $z=2.142$ (Fruchter \etal\ 2001; Gladders \etal\ 2001). Analyzing
592: existing afterglow spectra obtained using FORS2 on the VLT telescopes,
593: Vreeswijk \etal\ (2006) reported a total neutral hydrogen column
594: density of $\log\,N(\hI)=20.4\pm 0.2$. In addition, the authors
595: applied a curve-of-growth (COG) analysis over a series of absorption
596: features found in the afterglow spectra and derived $[{\rm Si}/{\rm
597: H}]=-0.9_{-0.4}^{+0.6}$ and $[{\rm Fe}/{\rm H}]=-1.3\pm
598: 0.3$\footnote{Chemical abundances are measured relative to solar
599: values and defined as $[{\rm M}/{\rm H}]\equiv\log ({\rm M}/{\rm H}) -
600: \log ({\rm M}/{\rm H})_\odot$.}. The large $N(\hI)$ indicates that
601: even the weakest absorption features detected in the afterglow
602: spectrum may be saturated due to the low resolution and low $S/N$ of
603: the data. Here we consider the abundance measurements based on the
604: COG analysis lower limits to the intrisic ISM metallicity and adopt
605: $[{\rm Si}/{\rm H}]> -1.3$ for the ISM of the host galaxy. However,
606: we note that the absence of relatively weak transitions such as Si\,II
607: $\lambda\,1808$ in the afterglow spectrum indicates $[{\rm Si}/{\rm
608: H}]< -0.7$. Finally, this line of sight exhibits no Mg\,II absorbers
609: of rest-frame absorption equivalent width $W(2796)>1$ \AA\ between
610: $z=0.359$ and $z=2$ (Prochter \etal\ 2006).
611:
612: Imaging follow-up of the field around GRB\,011211 was carried out
613: during four different epochs with STIS and the clear filter on board
614: HST. The imaging data were reduced and analyzed by Jakobsson \etal\
615: (2003), who reported the host galaxy has an $R$-band magnitude of
616: $AB(R)=25.15\pm 0.11$ over a 1\arcsec-diameter aperture. In addition,
617: \lya\ emission was detected in a ground-based narrow-band imaging
618: survey by Fynbo \etal\ (2003a), who measured a total flux of
619: $f(\lya)=(2.8\pm 0.8)\times 10^{-17}$ erg s$^{-1}$ cm$^{-2}$ and
620: derived an SFR of $0.8\pm 0.2$ M$_\odot$ yr$^{-1}$.
621:
622: We have observed this field using PANIC on Magellan in February
623: 2004, and obtained a total integration of 216 minutes. The mean FWHM
624: of the PSF was found to be $\approx 0.37\arcsec$ based on an average
625: of 10 stars across the PANIC field. The final stacked image is
626: presented in Figure 2, together with HST STIS images obtained roughly
627: 14, 26, and 32 days after the initial burst. The optical and $H$
628: images are registered to a common origin. The host galaxy is marked
629: by a circle of 1\arcsec\ radius. The host appears to consist of three
630: compact (presumably star-forming) regions in the STIS images, all of
631: which are confirmed to be at the host redshift based on the presence
632: of \lya\ emission in the ground-based narrow-band images presented in
633: Fynbo \etal\ (2003a). The GRB is found to originate in the faintest of
634: the three blobs, southeast of the center one. Two of the compact
635: regions are detected in the $H$-band image, but the region directly
636: associated with the burst does not exhibit any detectable flux. We measure
637: an $H$-band magnitude of $AB(H)=25.0\pm 0.3$ over a 1\arcsec\ diameter
638: aperture for the host galaxy.
639:
640: At $z=2.142$, the observed $R$-band magnitude corresponds to a
641: rest-frame absolute magnitude at 2000 \AA\ of
642: $M(2000)-5\,\log\,h=-19.2\pm 0.1$. The observed $H$-band magnitude
643: allows us to derive a rest-frame absolute $B$-band magnitude of
644: $M_{AB}(B) - 5\,\log\,h=-19.1\pm 0.3$ for the GRB host galaxy.
645:
646: \subsection{GRB\,020124 at $z_{\rm GRB}=3.198$}
647:
648: This burst was detected by HETE (Ricker \etal\ 2002) and the optical
649: transient was found $\approx 22$ hours later with $R\sim 18.5$ (Price
650: \etal\ 2002). Low-resolution ($\delta\,v \approx 680$ \kms) optical
651: spectra of the afterglow were obtained using FORS1 on the VLT Melipal
652: telescope, which revealed multiple absorption features indicating a
653: source redshift of $z=3.198$ (Hjorth \etal\ 2003). The total neutral
654: hydrogen column density derived based on the observed \lya\ absorption
655: line is $\log\,N(\hI)=21.7\pm 0.2$ (Hjorth \etal\ 2003). The low
656: resolution and low $S/N$ ($\approx\, 3-4$) of the data did not allow
657: an accurate measurement of the chemical content in the host ISM. No
658: information is available for the line-of-sight properties of
659: additional intervening absorbers.
660:
661: \begin{figure}
662: \begin{center}
663: \includegraphics[scale=0.3,angle=270]{grb020124_sed.eps}
664: \end{center}
665: \caption[]{Schematic diagram to illustrate the photometric zero point
666: offset neccessary to be included for sources at $z=3.2$, from a clear
667: bandpass that covers a spectral range over $\lambda=2000-10,000$ \AA\
668: to a typical Johnson $R$ filter. The historgram represents a typical
669: starburst spectral template with additional absorption at
670: $\lambda_{\rm rest} \le 1215$ \AA\ due to the IGM \lya\ forest and the
671: ISM internal to the host.}
672: \end{figure}
673:
674: Imaging follow-up of the field around GRB\,020124 was carried out
675: roughly 18 and 21 days after the burst with STIS and the clear filter
676: on board HST. The imaging data were reduced and analyzed by Berger
677: \etal\ (2002). While the OT was detected in the first epoch image, no
678: detectable flux was found at the OT position in the second epoch image.
679: Berger \etal\ (2002) placed an upper limit for the $R$-band magnitude
680: of the host galaxy at $R>29.5$. We have retrieved the imaging data
681: from the HST archive and determined a 5-$\sigma$ limit of the second
682: epoch image at $AB({\rm clear})=29.4$ over a 0.5\arcsec\ diameter
683: aperture. To derive the corresponding detection limit, we take into
684: account the bandpass difference between STIS/clear and $R$ and the IGM
685: opacity that reduces most of the light at $\lambda_{\rm obs}\le 4000$
686: \AA\ for sources at $z=3.198$ (Figure 3). We apply a 0.4 mag offset
687: in the photometric zero point and conclude that the host galaxy is
688: fainter than $AB(R)=29$ at the 5-$\sigma$ level of significance.
689:
690: We have also observed this field using PANIC on Magellan in February
691: 2004, and obtained a total integration of 231 minutes. The mean FWHM
692: of the PSF was found to be $\approx 0.4\arcsec$. The final stacked
693: image is presented in Figure 4, together with available HST STIS
694: images. The optical and $H$ images are registered to a common origin.
695: The $H$-band image has been smoothed using a Gaussian kernel of ${\rm
696: FWHM}=0.4\arcsec$. The position of the GRB is marked by a circle of
697: 0.5\arcsec radius. No detectable flux is found at the location of
698: the GRB, but some emission features are observed with $AB(H)\approx
699: 26.4$ at $\Delta\,\theta=0.4\arcsec$ angular distance away. We place
700: a 2-$\sigma$ limit in the observed $H$-band magnitude of $AB(H)=26.1$
701: over a 0.5\arcsec\ diameter aperture for the host galaxy. At
702: $z=3.198$, the observed $H$-band magnitude limit allows us to derive a
703: limiting rest-frame absolute magnitude of $M_{AB}(3960) -
704: 5\,\log\,h>-18.8$ for the GRB host galaxy.
705:
706: \begin{figure}
707: \begin{center}
708: \includegraphics[scale=0.32,angle=0]{g020124_tot.ps}
709: \end{center}
710: \caption[]{Registered optical (left two panels) and near-infrared $H$
711: (right panel) images of the field around GRB\,020124 at $z_{\rm
712: GRB}=3.198$. The epoch during which the images were taken is
713: indicated at the bottom of each panel. The $H$-band image has been
714: smoothed using a Gaussian kernel of ${\rm FWHM}=0.4\arcsec$, which is
715: roughly the size of the PSF. While the OT is still visible in the
716: first-epoch image (left panel), the host is not detected in either of
717: the two late-time images. We place a 2-$\sigma$ $H$-band limiting
718: magnitude of $AB(H)=26.1$ over a 0.5\arcsec\ diameter aperture for the
719: host galaxy.}
720: \end{figure}
721:
722: \subsection{GRB\,030323 at $z_{\rm GRB}=3.372$}
723:
724: This burst was detected by HETE-2 (Graziani \etal\ 2003) and the
725: optical transient was found $\approx 7.6$ hours later with $R=18.7$
726: (Gilmore \etal\ 2003a). Moderate-resolution ($\delta\,v \approx 150$
727: \kms) optical spectra of the afterglow, covering $\lambda=4560-7310$
728: \AA, were taken using FORS2 on the VLT Yepun telescope by Vreeswijk
729: \etal\ (2004), revealing multiple absorption features from ions in
730: both ground states and excited states that are consistent with a
731: source redshift of $z=3.372$. These authors estimated the total
732: neutral hydrogen column density of $\log\,N(\hI)=21.90\pm 0.07$ in the
733: host ISM and chemical abudances of $[{\rm S}/{\rm H}]=-1.26\pm 0.2$
734: and $[{\rm Fe}/{\rm H}]=-1.47\pm 0.11$. The observed large column
735: densities of various ions suggest that these reported values represent
736: only lower limits to the intrinsic abundances of these ions. We
737: therefore adopt $[{\rm S}/{\rm H}]> -1.26$ for the ISM of the host
738: galaxy. This line of sight exhibits no strong Mg\,II absorbers at
739: $z=0.824-1.646$ (Prochter \etal\ 2006).
740:
741: Imaging follow-up of the field around GRB\,030323 was carried out
742: during two different epochs with ACS and the F606W filter on board
743: HST. The first epoch imaging data with a total exposure time of 1920
744: s were reduced and analyzed by Vreeswijk \etal\ (2004), who identified
745: the host galaxy at $\Delta\,\theta=0.14\arcsec$ from the position of
746: the OT and measured $AB({\rm F606W})=28.0\pm 0.3$ over a 0.3\arcsec\
747: diameter aperture. Additional imaging data were obtained five month
748: later with HST using the same instrument setup. A stack of all
749: available imaging data from the HST data archive shows a clear
750: detection of the host galaxy at $\Delta\,\theta=0.22\arcsec$ from the
751: position of the OT. This position is consistent with the position of
752: Vreeswijk \etal\ to within the astrometric uncertainties. After
753: correcting for the Galactic extinction ($E(B-V)=0.049$ according to
754: Schlegel \etal\ 1998), we measure a total flux of $AB({\rm
755: F606W})=27.4\pm 0.1$ over a 0.5\arcsec\ diameter aperture.
756:
757: We have also observed this field using PANIC on Magellan in February
758: 2004, and obtained a total integration of 240 minutes. The mean FWHM
759: of the PSF was found to be $\approx 0.4\arcsec$. The final stacked
760: image is presented in Figure 5, together with a stack of available
761: ACS/F606W images. The optical and $H$ images are registered to a
762: common origin. The host galaxy is marked by a circle of 0.5\arcsec\
763: radius. The host appears to be extended in the ACS image, but not
764: detected in the stacked $H$ image. We place a 2-$\sigma$ limit in the
765: observed $H$-band magnitude of $AB(H)=26.2$ over a 0.5\arcsec\ diameter
766: aperture for the host galaxy.
767:
768: \begin{figure}
769: \begin{center}
770: \includegraphics[scale=0.38,angle=0]{g030323_tot2.ps}
771: \end{center}
772: \caption[]{Registered optical (left panel) and near-infrared $H$
773: (right panel) images of the field around GRB\,030323 at $z_{\rm
774: GRB}=3.372$. The epoch during which the images were taken is
775: indicated at the bottom of each panel. The host galaxy is identified
776: at $\Delta\,\theta=0.22\arcsec$ from the position of the OT with
777: $AB({\rm F606W})=27.4\pm 0.1$ over a 0.5\arcsec\ diameter aperture.
778: The $H$-band image has been smoothed using a Gaussian kernel of ${\rm
779: FWHM}=0.4\arcsec$, which is roughly the size of the PSF. The host is
780: not detected in the $H$ image. We place a 2-$\sigma$ $H$-band
781: limiting magnitude of $AB(H)=26.2$ over a 0.5\arcsec\ diameter aperture
782: for the host galaxy.}
783: \clearpage
784: \end{figure}
785:
786: At $z=3.372$, the observed F606W magnitude corresponds to
787: $M_{AB}(1400) - 5\,\log\,h=-17.6\pm 0.1$ for the GRB host galaxy. The
788: observed $H$-band magnitude limit allows us to derive a limiting
789: rest-frame absolute magnitude of $M_{AB}(3800) - 5\,\log\,h>-18.8$ for
790: the GRB host galaxy.
791:
792: \subsection{GRB\,030429 at $z_{\rm GRB}=2.658$}
793:
794: This burst was detected by HETE-2 (Doty \etal\ 2003) and the optical
795: transient was found $\approx 3.5$ hours later (Gilmore \etal\ 2003b)
796: at $\Delta\,\theta\approx 1.2\arcsec$ southeast of an extended source
797: of $R\approx 24$ (Fynbo \etal\ 2003b). Low-resolution ($\delta\,v
798: \approx 680$ \kms) optical spectra of the afterglow and the extended
799: source were obtained using FORS1 on the VLT Melipal telescope. The
800: spectrum of the afterglow exhibits multiple absorption features that
801: are consistent with $z=2.658$ (Jakobsson \etal\ 2004). In contrast,
802: the spectrum of the extended source exhibits a single emission feature
803: at $\lambda=6858$ \AA. Jakobsson \etal\ (2004) identified this
804: emission as [O\,II] at $z=0.841$, which is supported by the presence
805: of an absorption feature in the afterglow spectrum if interpreted as a
806: Mg\,II absorber at $z=0.841$. The GRB host galaxy therefore remains
807: unidentified. These authors also reported a total neutral hydrogen
808: column density of $\log\,N(\hI)=21.6\pm 0.2$ based on the observed
809: \lya\ absorption feature. The low resolution of the spectrum did not
810: allow an accurate measurement of the chemical content in the host ISM.
811:
812: \begin{figure}
813: \begin{center}
814: \includegraphics[scale=0.38,angle=0]{g030429_tot.ps}
815: \end{center}
816: \caption[]{Near-infrared $H$ image of the field around GRB\,030349 at
817: $z_{\rm GRB}=2.658$. The mean FWHM of the PSF is 0.35\arcsec. A
818: smooth version of the $H$-band image is presented in the right panel.
819: No detectable flux is found at the immediate location of the GRB, but
820: extended emission features (pointed by an arrow in the left panel) are
821: clearly visible at $\approx 1\arcsec$ southwest of the OT, or
822: 1.3\arcsec\ south of the galaxy at $z=0.841$. We measure
823: $AB(H)=20.57\pm 0.05$ for the foreground galaxy, and $AB(H)=24.4\pm
824: 0.1$ for the host candidate.}
825: \clearpage
826: \end{figure}
827:
828: We have observed the field around GRB\,030429 using PANIC on
829: Magellan in February 2004, and obtained a total integration of 129
830: minutes. The mean FWHM of the PSF was found to be $\approx
831: 0.35\arcsec$. The final stacked image is presented in Figure 6,
832: together with a smoothed version with a Gaussian kernel of ${\rm
833: FWHM}=0.4\arcsec$. The position of the GRB is marked by a circle of
834: 0.5\arcsec radius. No detectable flux is found at the immediate
835: location of the GRB, but additional emission features are present at
836: $\approx 1\arcsec$ southwest of the OT, or 1.3\arcsec\ south of the
837: galaxy at $z=0.841$. We measure $AB(H)=20.57\pm 0.05$ for the
838: foreground galaxy, and $AB(H)=24.4\pm 0.1$ for the host candidate.
839:
840: The emission morphology resembles those found for the hosts of
841: GRB\,000926 and GRB\,011211, with the OT originating in the faintest
842: of the three emission blobs. The close proximity to the foreground
843: galaxy suggests that the host may be gravitationally lensed by this
844: foreground object, although the large angular separation implies that
845: the foreground galaxy would have to be $\approx\,4\times$ as massive
846: as the Milky Way (see also Jakobsson \etal\ 2004). Additional images
847: at optical wavelengths are necessary to investigate this lensing
848: hypothesis further.
849:
850: If the extended feature is associated with the host galaxy, then the
851: OT would be at a projected distance of $\rho=5.4\ h^{-1}$ kpc from the
852: center of the galaxy. In the following discussion, we consider the
853: observed brightness as an upper limit to the brightness of the GRB
854: host and derive a rest-frame $B$-band magnitude of $M_{AB}(B) -
855: 5\,\log\,h>-20.1$ for the GRB host galaxy.
856:
857: \subsection{GRB\,050401 at $z_{\rm GRB}=2.899$}
858:
859: This burst was detected by {\it Swift} (Barbier \etal\ 2005) and
860: prompt localization of the source was reported by Angelini \etal\
861: (2005) from the X-ray afterglow. The optical transient was found
862: $\approx 1$ hour later with $R=20.3$ (McNaught \& Price 2005).
863: Low-resolution ($\delta\,v \approx 680$ \kms) optical spectra of the
864: afterglow were taken using FORS2 on the VLT Antu telescope (Watson
865: \etal\ 2006), revealing multiple absorption features from ions in both
866: ground states and excited states that are consistent with a source
867: redshift of $z=2.899$. Watson \etal\ (2006) estimated the total
868: neutral hydrogen column density of $\log\,N(\hI)=22.6\pm 0.3$ in the
869: host ISM and metallicity of $[{\rm Zn}/{\rm H}]=-1.0\pm 0.4$.
870: Similiar to the other GRB hosts where only low-resolution afterglow
871: spectra are available, the observed large column densities of various
872: ions suggest that these reported values represent only lower limits to
873: the intrinsic abundances of these ions. We therefore adopt $[{\rm
874: Zn}/{\rm H}]> -1.0$ for the ISM of the host galaxy. An additional set
875: of ionic transitions (such as Al\,II and Fe\,II) was also reported at
876: $z=2.5$ by Watson \etal\ (2006).
877:
878: We have observed the field around GRB\,050401 using PANIC on
879: Magellan in May 2006, and obtained a total integration of 105 minutes.
880: The mean FWHM of the PSF was found to be $\approx 0.6\arcsec$. No
881: detectable flux is seen at the immediate location of the GRB. We
882: place a 2-$\sigma$ limit in the observed $H$-band magnitude of
883: $AB(H)=25.1$ over a 1\arcsec\ diameter aperture for the host galaxy.
884: At $z=2.899$, the observed $H$-band magnitude limit allows us to
885: derive a limiting rest-frame absolute magnitude of $M_{AB}(B) -
886: 5\,\log\,h>-19.6$ for the GRB host galaxy.
887:
888: Optical images of the field around GRB\,050401 have also been
889: obtained using LRIS and the $g$ and $R_c$ filters on the Keck I
890: telescope in May 2006. The mean FWHM of the PSF was found to be
891: $\approx 1.1\arcsec$ in the combined $g$ image and $\approx
892: 1.0\arcsec$ in the combined $R_c$ image. The stacked images are
893: presented in Figure 7, which have been smoothed with a Gaussian kernel
894: of ${\rm FWHM}=1\arcsec$. The position of the GRB is marked by a
895: circle of 2\arcsec\ radius. At the location of the afterglow reported
896: by (McNaught \& Price 2005), we detect faint emission in both images
897: at roughly 2-$\sigma$ significance level. We estimate a total
898: brightness of $AB(g)=27.5\pm 0.5$ and $AB(R_c)=27.3\pm 0.4$ over a
899: 2\arcsec\ diameter aperture for the object. Due to the presence of a
900: strong absorber at $z=2.5$, the identification of the observed faint
901: emission is uncertain. We note, however, that at $z=2.899$ the IGM
902: \lya\ forest is expected to absorb a large fraction of flux in the $g$
903: band, resulting in $g-R_c\approx 0.2-0.4$ mag for a flat spectrum
904: source. Additional HST imaging data are necessary to confirm the host
905: identification. At $z=2.899$, the observed $R_c$-band magnitude would
906: imply a rest-frame absolute magnitude of $M_{AB}(1600) -
907: 5\,\log\,h=-17.4\pm 0.4$ for the candidate host galaxy.
908:
909: \begin{figure}
910: \begin{center}
911: \includegraphics[scale=0.31,angle=0]{g050401_lris.ps}
912: \end{center}
913: %\caption[]{A smoothed $H$ image of the field around GRB\,050401 at
914: %$z_{\rm GRB}=2.899$. No detectable flux is seen at the location of
915: %the OT. We place a 2-$\sigma$ $H$-band limiting magnitude of
916: %$AB(H)=25.1$ over a 1\arcsec diameter aperture for the host galaxy.}
917: \caption[]{Optical $g$ (left panel) and $R_c$ (right panel) images of
918: the field around GRB\,050401 at $z_{\rm GRB}=2.899$. The mean seeing
919: is 1.1\arcsec\ and 1.0\arcsec\ in $g$ and $R_c$, respectively. At the
920: location of the afterglow reported by (McNaught \& Price 2005), we
921: detect faint emission in both images at roughly 2-$\sigma$
922: significance level. The circle indicates a 2\arcsec\ angular radius
923: around the afterglow position.}
924: \end{figure}
925:
926: \subsection{GRB\,050730 at $z_{\rm GRB}=3.968$}
927:
928: This burst was detected by {\it Swift} (Holland \etal\ 2005). An
929: optical transient was found promptly using the Ultraviolet-Optical
930: Telescope (UVOT) on board Swift with $V = 17.6$ about three minutes
931: after the burst trigger (Holland \etal\ 2005). We obtained an echelle
932: spectrum of the afterglow, using the MIKE spectrograph (Bernstein
933: \etal\ 2003) on the Magellan Clay telescope, four hours after the
934: initial trigger. Descriptions of the data were presented in Chen
935: \etal\ (2005) and Prochaska \etal\ (2007b). The spectrum covers a
936: full spectral range from 3300 \AA\ through 9400 \AA\ with a spectral
937: resolution of $\delta\,v \approx 10$ \kms\ at wavelength
938: $\lambda=4500$ \AA\ and $\delta\,v \approx 12$ \kms\ at $\lambda=8000$
939: \AA. The host of the GRB exhibits a strong damped \lya\ absorption
940: feature with $\log\,N(\hI)=22.15\pm 0.05$, and metallicity $[{\rm
941: S}/{\rm H}]=-2.26\pm 0.1$ and $[{\rm S}/{\rm Fe}]> +0.24\pm 0.11$ (see
942: also Starling \etal\ 2005; D'Elia \etal\ 2007). The low metallicity
943: and low $\alpha$-element enhancement implies a nearly dust free
944: medium. No trace of H$_2$ is found despite the large $N(\hI)$. The
945: high-resolution, high $S/N$ echelle data allow us to place a sensitive
946: limit on the molecular fraction of the host ISM at $f_{\rm H_2}\equiv
947: 2N({\rm H_2})/[N(\hI)+2\,N({\rm H_2})] < 10^{-7.1}$ (Tumlinson \etal\
948: 2007). Additional strong absorbers are found at $z=3.56$, $z=3.02$,
949: $z=2.25$, and $z=1.77$ (Chen \etal\ 2005).
950:
951: We have observed the field around GRB\,070530 using MagIC and the
952: $i'$ filter on Magellan in June 2008, and obtained a total integration
953: of 45 minutes. The mean FWHM of the PSF was found to be $\approx
954: 0.6\arcsec$. The final stacked image is presented in Figure 8, which
955: has been smoothed using a Gaussian kernel of ${\rm FWHM}=0.6\arcsec$.
956: The position of the GRB is marked by a circle of 1\arcsec\ radius. No
957: detectable flux is seen at the immediate location of the GRB. We
958: place a 2-$\sigma$ limit in the observed $i'$-band magnitude of
959: $AB(i')=26.6$ over a 1\arcsec\ diameter aperture for the host galaxy.
960: At $z=3.968$, observed $i'$-band magnitude limit allows us to derive a
961: limiting rest-frame absolute magnitude of $M_{AB}(1500) -
962: 5\,\log\,h>-18.6$ for the GRB host galaxy.
963:
964: Optical images of the field around GRB\,050730 have also been
965: obtained using LRIS and the $g$ and $R_c$ filters on the Keck I
966: telescope in May 2006. The mean FWHM of the PSF was found to be
967: $\approx 1.2\arcsec$ in the combined $g$ image and $\approx
968: 1.0\arcsec$ in the combined $R_c$ image. No detectable flux is
969: seen at the immediate location of the GRB. We place 2-$\sigma$
970: limits of $AB(g)=26.6$ and $AB(R_c)=26.4$ over a 2\arcsec\ diameter
971: aperture for the host galaxy. Taking into account that at $z=3.968$
972: the IGM \lya\ forest and the large amount of neutral gas in the host
973: ISM absorb a large fraction of flux in the $g$ and $R_c$ bands, we
974: find that the flux limits derived from the LRIS images are consistent
975: with the flux limit seen in the MagIC $i'$-band data.
976:
977: \begin{figure}
978: \begin{center}
979: \includegraphics[scale=0.4,angle=0]{g050730.ps}
980: \end{center}
981: \caption[]{A smoothed $i'$ image of the field around GRB\,050730 at
982: $z_{\rm GRB}=3.968$. No detectable flux is seen at the location of
983: the OT. We place a 2-$\sigma$ $i'$-band limiting magnitude of
984: $AB(H)=26.6$ over a 1\arcsec\ diameter aperture for the host galaxy.}
985: \end{figure}
986:
987: \subsection{GRB\,050820A at $z_{\rm GRB}=2.615$}
988:
989: This burst was detected by {\it Swift} (Page \etal\ 2005a). An
990: optical transient, reported less than 1 hour after the GRB, was
991: identified in data taken shortly after the trigger (Fox \& Cenko 2005;
992: Vestrand \etal\ 2006). High-resolution ($\delta\,v \approx 7-10$
993: \kms) echelle spectra of the afterglow obtained shortly after the
994: burst are available from both our own observations using HIRES (Vogt
995: \etal\ 1994) on the Keck I telescope and the ESO data archive for UVES
996: (D'Odorico \etal\ 2000). These two spectra together cover a full
997: spectral range from 3300 \AA\ through 10,000 \AA, allowing accurate
998: estimates of chemical abudances in the ISM of the GRB host galaxy.
999: Based on multiple absorption features from both ground-state and
1000: excited-state ions, we determine a source redshift of $z=2.6147$ and a
1001: total neutral hydrogen column density of $\log\,N(\hI)=21.0\pm 0.1$
1002: (Prochaska \etal\ 2007b; Ledoux \etal\ 2005). An absorption-line
1003: analysis of various ions shows that $[{\rm S}/{\rm H}]= -0.63\pm 0.11$
1004: and $[{\rm S}/{\rm Fe}]= +0.97\pm 0.09$ (Prochaska \etal\ 2007b), from
1005: which we derive a dust-to-gas ratio that is comparable to what is seen
1006: in the Small Magellanic Cloud (SMC). Adopting the SMC dust-to-gas
1007: ratio (Gordon \etal\ 2003), we estimate a visual extinction in the
1008: host ISM of $A_V\approx 0.08$. Finally, no trace of H$_2$ is found
1009: despite the large $N(\hI)$ and moderate metallicity. The
1010: high-resolution, high $S/N$ echelle data allow us to place a sensitive
1011: limit on the molecular fraction of the host ISM at $f_{\rm H_2}\equiv
1012: 2N({\rm H_2})/[N(\hI)+2\,N({\rm H_2})] < 10^{-6.5}$ (Tumlinson \etal\
1013: 2007).
1014:
1015: Imaging follow-up of the field around GRB\,050820A was carried out
1016: with ACS and the F625W, F775W, F850LP filters on board HST in two
1017: epochs, roughly 37 days, and nine months after the burst
1018: (PID$=$10551). We retrieved the imaging data from the HST data
1019: archive and processed the images following the descriptions in \S\ 3.
1020: Stacked images from the two epochs are presented in Figure 9. The OT
1021: was clearly detected in the first epoch image, but faded in the second
1022: epoch image which reveals faint extended emission features of the host
1023: galaxy. In addition to the extended low-surface feature seen at the
1024: position of the OT, which is identified as the host galaxy, we
1025: identify two compact sources at $\Delta\,\theta=1.3\arcsec$ north of
1026: the OT (Object $A$) and $0.4\arcsec$ south of the OT (Object $B$) from
1027: the afterglow lines of sight. We measure isophotal magnitudes of the
1028: host and Objects $A$ and $B$. The isophotal apertures are defined
1029: based on the extent of objects found in a ``white light'' image, which
1030: is a stack of the F625W, F775W, and F850LP images. This ``white
1031: light'' image has the optimal $S/N$ for recovering faint emission
1032: features and allows us to determine a common aperture for every object
1033: across different bandpasses. We correct the observed brightness to
1034: account for intrinsic absorption in the Milky Way ($E(B-V)=0.044$
1035: along the line of sight according to Schlegel \etal\ 1998). The
1036: photometric measurements are presented in Table 2.
1037:
1038: \begin{figure*}
1039: \begin{center}
1040: \includegraphics[scale=0.5,angle=0]{g050820_acs.ps}
1041: \end{center}
1042: \caption[]{Images of the field around GRB\,050820A at $z_{\rm
1043: GRB}=2.6147$. The first-epoch image (left panel), taken $\approx 37$
1044: days after the burst with ACS and the F850LP filter, clearly shows the
1045: OT (indicated by the arrow). The second-epoch image (right panel),
1046: taken nine months later, reveals extended faint blue emission of the
1047: host galaxy. The false-color image was formed by combining stacks
1048: of F625W, F775W, and F850LP images. The yellow cross and circle mark
1049: the position of the x-ray afterglow and associated error reported by
1050: Page \etal\ (2005b). In addition to the host galaxy, we also point out
1051: two compact sources $A$ and $B$ at $\Delta\,\theta=1.3\arcsec$ and
1052: $0.4\arcsec$, respectively, from the afterglow lines of sight.}
1053: \end{figure*}
1054:
1055: We have also observed this field using PANIC on Magellan in August
1056: 2007, and obtained a total integration of 141 minutes. The mean FWHM
1057: of the PSF was found to be $\approx 0.5\arcsec$. The final stacked
1058: image is presented in Figure 10, together with the second-epoch
1059: ACS/F850LP image. The optical and $H$ images are registered to a
1060: common origin, and the $H$-band image has been smoothed using a
1061: Gaussian kernel of ${\rm FWHM}=0.5\arcsec$. The position of the GRB
1062: is marked by a circle of 0.5\arcsec\ radius. Extended emission at the
1063: position of the host galaxy is detected in the $H$ image at $\approx
1064: 3\,\sigma$ level of significance. We measure an $H$-band magnitude of
1065: $AB(H)=25.3\pm 0.3$ within the isophotal aperture defined for the host
1066: galaxy in the ``white light'' image described above.
1067:
1068: At $z=2.6147$, the observed F775W magnitude of the host galaxy
1069: corresponds to a rest-frame absolute magnitude at 2000 \AA\ of
1070: $M(2000)-5\,\log\,h=-18.2\pm 0.06$. The observed $H$-band magnitude
1071: allows us to derive a rest-frame absolute $B$-band magnitude of
1072: $M_{AB}(B) - 5\,\log\,h=-19.2\pm 0.3$ for the GRB host galaxy.
1073: Adopting $A_V=0.08$ estimated above from absorption-line abundance
1074: ratios and the SMC extinction law (Gordon \etal\ 2003), we derive
1075: extinction corrected rest-frame absolute magnitudes of $M_{AB}(B) -
1076: 5\,\log\,h=-19.3\pm 0.3$ and $M(2000)-5\,\log\,h=-18.5\pm 0.06$.
1077:
1078: \begin{figure*}
1079: \begin{center}
1080: \includegraphics[scale=0.6,angle=0]{g050820_nir.ps}
1081: \end{center}
1082: \caption[]{Registered ACS F850LP image (left panel) and near-infrared
1083: $H$ (right panel) image of the field around GRB\,050820A at $z_{\rm
1084: GRB}=2.6147$. The $H$-band image has been smoothed using a Gaussian
1085: kernel of ${\rm FWHM}=0.5\arcsec$, which is roughly the size of the
1086: PSF. The host is detected in the $H$ image with $AB(H)=25.3\pm 0.3$.
1087: Neither Object $A$ or $B$ exhibits faint emission in the $H$ image. We
1088: place a 2-$\sigma$ limit of $AB(H)>26$ for the two sources.}
1089: \clearpage
1090: \end{figure*}
1091:
1092: Neither Object $A$ or $B$ exhibits detectable flux in the $H$ image.
1093: We place a 2-$\sigma$ limit of $AB(H)>26$ for the two sources. The
1094: $H$-band photometric measurements are also presented in Table 2. The
1095: observed optical and near-infrared colors of Objects $A$ and $B$ are
1096: relatively bluer than those of the GRB host galaxy and inconsistent
1097: with the expectations for $z>2$ star-forming objects. This suggests
1098: that these are likely foreground galaxies. At these small angular
1099: separations ($\Delta\,\theta < 1.4$), objects at $z<2.6$ have
1100: projected distances of $\rho< 8\ h^{-1}$ kpc to the afterglow line of
1101: sight, and are expected to imprint strong Mg\,II
1102: $\lambda\lambda\,2796, 2803$ absorption features in the afterglow
1103: spectrum (e.g.\ Chen \& Tinker 2008).
1104:
1105: Incidentally, two strong Mg\,II absorbers are found at $z=0.692$ and
1106: $z=1.430$ with $W(2796)=2.99\pm 0.03$ \AA\ and $W(2796)=1.9\pm 0.1$
1107: \AA\ in the rest frame, respectively. The $z=0.692$ Mg\,II absorber
1108: exhibits a complex kinematic profile with multiple absorption
1109: components spreading over a line-of-sight velocity interval of
1110: $\Delta\,v\approx 500$ \kms, and a non-negligible amount of Ca$^+$
1111: ions (see Figure 13 in Prochaska \etal\ 2007b). The complex kinematic
1112: profile of the absorber and the presence of Ca$^+$ together may be
1113: explained by a sightline passing through an interacting system,
1114: simliar to the Milky Way and the Magellanic Stream (e.g.\ Gibson
1115: \etal\ 2000; Putman \etal\ 2003). Attributing both Objects $A$ and
1116: $B$ to the $z=0.692$ absorber implies a total absolute $B$-band
1117: magnitude of $M_{AB}(B) - 5\,\log\,h=-16.27\pm 0.05$. Follow-up
1118: near-infrared H$\alpha$ spectroscopy is necessary to provide
1119: conclusive identifications of these two sources.
1120:
1121: Optical images of the field around GRB\,050820A have also been
1122: obtained using LRIS and the $g$ and $R_c$ filters on the Keck I
1123: telescope in July 2006. The mean FWHM of the PSF was found to be
1124: $\approx 0.7\arcsec$ in the combined $g$ image and $\approx
1125: 1.9\arcsec$ in the combined $R_c$ image. The effective seeing was
1126: significantly compromised in the $R_c$ images due to a focus problem
1127: with the LRIS red-side. Faint emission is clearly detected at the
1128: position of the afterglow, but we are unable to obtain accurate
1129: measurements of the host magnitude due to contaminating light from
1130: galaxies $A$ and $B$ (Figure 9) in the ground-based images.
1131:
1132: \begin{figure}
1133: \begin{center}
1134: \includegraphics[scale=0.45,angle=0]{h050820a.ps}
1135: \end{center}
1136: \caption[]{Constraints on the galaxy age and star formation history.
1137: Top: The broad-band spectral energy distribution of the host galaxy of
1138: GRB\,050820A at $z=2.6147$ (points with errorbars), in comparison to
1139: the best-fit template (solid histogram) and the predicted broad-band
1140: photometric points (open squares). The best-fit SED is charaterized
1141: by a declining SFR of e-folding time $\tau=300$ Myr, $1/5$ solar
1142: metallicity, and intrinsic dust extinction of $E(B-V)=0.05$ following
1143: the SMC extinction law. Bottom: The likelihood function of the
1144: stellar age for the host galaxy, suggesting that a recent starburst
1145: occurred about $400-700$ Myr prior to the GRB explosion.}
1146: \end{figure}
1147:
1148: For the host galaxy, the optical and $H$-band photometric
1149: measurements presented in Table 2 allow us to examine the star
1150: formation history and constrain the stellar mass based on the observed
1151: spectral energy distribution (SED). To constrain the underlying
1152: stellar population, we consider a suite of synthetic stellar
1153: population models generated using the Bruzual \& Charlot (2003)
1154: spectral library. We adopt a Salpeter initial mass function with a
1155: range of metallicity from $1/5$ solar to $2\times$ solar and a range
1156: of star formation history from a single burst to exponentially
1157: declining SFR of e-folding time $\tau=300$ Myr or 1 Gyr. We include
1158: intrinsic dust extinction that follows the SMC extinction law.
1159: Comparing the observed SED with model predictions allows us to
1160: constrain the stellar age. The results are presented in Figure 11,
1161: where the observed SED of the galaxy is shown in the top panel
1162: together with the best-fit model. It is clear that the $H$-band
1163: photometry provides the necessary measurement for constraining the
1164: stellar population based on the 4000-\AA\ flux decrement. The bottom
1165: panel of Figure 10 shows the likelihood distribution function versus
1166: stellar age, indicating that the last major episode of star formation
1167: occurred at $\approx 400-700$ Myr ago.
1168:
1169: Adopting the best-fit stellar synthetic model, we infer a total
1170: stellar mass of $M_*= 0.9_{-0.4}^{+2.5}\times 10^9\ h^{-2}$ M$_\odot$.
1171: We argue that our inferred stellar mass is accurate, despite an
1172: absence of rest-frame near-infrared flux measurements. This is
1173: supported by previous studies, which show that stellar masses
1174: determined based on the observed 4000-\AA\ flux decrement are
1175: consistent with those determined based on rest-frame near-infrared
1176: luminosity to within the uncertainties (cf.\ Chen \& Marzke 2004 and
1177: Yan \etal\ 2004 for red galaxies identified in the Hubble Ultra Deep
1178: Field and see also Shapley \etal\ 2005 for a detailed comparison based
1179: on 72 star-forming galaxies at $\langle z\rangle=2.3$). The inferred
1180: stellar mass for the host galaxy of GRB\,050820A at $z=2.6147$ is
1181: comparable to those derived for $z\approx 1$ GRB host galaxies (e.g.\
1182: Castro Cer\'on \etal\ 2006, 2008; Savaglio \etal\ 2008) but falls in
1183: the bottom 2\% of the UV luminous galaxies studied by Shapley \etal\
1184: (2005).
1185:
1186: \begin{center}
1187: \begin{small}
1188: \begin{deluxetable*}{lccccc}
1189: \tablewidth{0pt}
1190: \tablecaption{A Summary of the Observed Isophotal Magnitudes of Objects at $\Delta\,\theta\le 1.5\arcsec$ from GRB\,050820A}
1191: \tablehead{\multicolumn{1}{c}{Objects} & \colhead{$\Delta\,\theta$ (\arcsec)} & \colhead{$AB$(F625W)} & \colhead{$AB$(F775W)} & \colhead{$AB$(F850LP)} & \colhead{$AB$($H$)}}
1192: \startdata
1193: Host & ... & $26.24\pm 0.28$ & $26.30\pm 0.06$ & $26.04\pm 0.05$ & $25.3\pm 0.3$ \nl
1194: $A$ & $1.34\pm 0.05$ & $26.08\pm 0.06$ & $26.11\pm 0.05$ & $25.81\pm 0.05$ & $>26.0$ \nl
1195: $B$ & $0.44\pm 0.05$ & $26.35\pm 0.23$ & $26.21\pm 0.05$ & $26.25\pm 0.05$ & $>26.0$ \nl
1196: \enddata
1197: \end{deluxetable*}
1198: \end{small}
1199: \end{center}
1200:
1201: \subsection{GRB\,050908 at $z_{\rm GRB}=3.344$}
1202:
1203: \begin{figure}
1204: \begin{center}
1205: \includegraphics[scale=0.4,angle=0]{g050908_tot.ps}
1206: \end{center}
1207: \caption[]{A smoothed $H$ image of the field around GRB\,050908 at
1208: $z_{\rm GRB}=3.344$. No detectable flux is seen at the location of
1209: the OT. We place a 2-$\sigma$ $H$-band limiting magnitude of
1210: $AB(H)=26$ over a 0.5\arcsec\ diameter aperture for the host galaxy.}
1211: \end{figure}
1212:
1213: This burst was detected by {\it Swift} (Goad \etal\ 2005) and the
1214: optical afterglow was reported by Torii (2005) 14 minutes after the
1215: trigger to have $R\approx 18.8$. We obtained moderate-resolution
1216: ($\delta\,v \approx 40-150$ \kms) optical spectra of the afterglow,
1217: using GMOS on the Gemini north telescope (Foley \etal\ 2005) and
1218: DEIMOS on the Keck II telescope (Prochaska \etal\ 2005). The two
1219: spectra together provides contiguous spectral coverage over
1220: $\lambda=5070-9000$ \AA\ (Chen \etal\ 2007b). The afterglow spectra
1221: exhibit a suite of absorption features, consistent with a source
1222: redshift of $z=3.3437$. Contrary to the majority of GRB host
1223: galaxies, the line-of-sight toward GRB\,050908 displays only a modest
1224: amount of neutral gas in the host galaxy. Our initial estimate based
1225: on the observed \lya\ absorption line suggest $\log\,N(\hI)\sim 19.2$.
1226: However, an additional spectrum obtained by Fugazza \etal\ (2005) and
1227: analyzed by Fynbo and collaborators displays non-negligible flux at
1228: $\lambda_{\rm obs} < 4000$ \AA\ (corresponding to rest-frame
1229: wavelength range $\lambda_{\rm rest} < 912$ \AA\ at $z=3.34$),
1230: indicating $\log\,N(\hI)=17.55\pm 0.1$ (Johan Fynbo 2007, private
1231: communication). This source repressents one of a few GRB sightlines
1232: observed so far that do not pass through neutral gas clouds in the
1233: host galaxies. Other known sources are GRB\,021004 at $z_{\rm
1234: GRB}=2.329$ with $\log\,N(\hI)=19.5\pm 0.5$, GRB\,060526 at $z_{\rm
1235: GRB}=3.221$ with $\log\,N(\hI)=20.00\pm 0.15$ (Jakobsson \etal\ 2006),
1236: and GRB\,060607 at $z_{\rm GRB}=3.075$ with $\log\,N(\hI)=16.85\pm
1237: 0.10$ (see further discussion in \S\S\ 4.11 and 4.13). A strong
1238: Mg\,II absorber is found at $z=1.548$ in the afterglow spectrum
1239: (Prochter \etal\ 2006).
1240:
1241: We have observed the field around GRB\,050908 using PANIC on
1242: Magellan in August 2007, and obtained a total integration of 203
1243: minutes. The mean FWHM of the PSF was found to be $\approx
1244: 0.5\arcsec$. The final stacked image is presented in Figure 12, which
1245: has been smoothed using a Gaussian kernel of ${\rm FWHM}=0.5\arcsec$.
1246: The position of the GRB is marked by a circle of 1\arcsec\ radius. No
1247: detectable flux is seen at the immediate location of the GRB. We
1248: place a 2-$\sigma$ limit in the observed $H$-band magnitude of
1249: $AB(H)=26.0$ over a 0.5\arcsec\ diameter aperture for the host galaxy.
1250: At $z=3.344$, the observed $H$-band magnitude limit allows us to derive a
1251: limiting rest-frame absolute magnitude of $M_{AB}(3825) -
1252: 5\,\log\,h>-18.9$ for the GRB host galaxy.
1253:
1254: \subsection{GRB\,050922C at $z_{\rm GRB}=2.199$}
1255:
1256: \begin{figure}
1257: \begin{center}
1258: \includegraphics[scale=0.4,angle=0]{g050922c_tot.ps}
1259: \end{center}
1260: \caption[]{A smoothed $H$ image of the field around GRB\,050922C at
1261: $z_{\rm GRB}=2.199$. No detectable flux is seen at the location of
1262: the OT. We place a 2-$\sigma$ $H$-band limiting magnitude of
1263: $AB(H)=26$ over a 0.5\arcsec\ diameter aperture for the host galaxy.}
1264: \end{figure}
1265:
1266: This burst was detected by {\it Swift} (Norris \etal\ 2005) and the
1267: optical afterglow was reported by Rykoff \etal\ (2005) ten minutes
1268: after the trigger to have $R\approx 16.0$ mag. Spectroscopic
1269: follow-up was carried out roughly 1.5 hours after the trigger by
1270: Jakobsson \etal\ (2006), who reported $z_{\rm GRB}=2.198$ based on a
1271: series of absorption features. A foreground damped \lya\ absorber
1272: (DLA) was also found at $z=2.07$ along this sightline. Additional
1273: high-resolution ($\delta\,v \approx 7$ \kms) echelle spectra of the
1274: afterglow were obtained roughly 3.5 hours after the trigger using UVES
1275: on the VLT Kueyen telescope (D'Elia \etal\ 2005). The spectra were
1276: retrieved from the ESO data archive and processed using our own
1277: reduction software.
1278:
1279: The combined echelle spectrum covers a spectral range over
1280: $\lambda=3300-10,000$ \AA, allowing accurate estimates of chemical
1281: abudances in the ISM of the GRB host galaxy. Based on multiple
1282: absorption features from both ground-state and excited-state ions, we
1283: determine a total neutral hydrogen column density of
1284: $\log\,N(\hI)=21.5\pm 0.1$. An absorption-line analysis of various
1285: ions shows that $[{\rm S}/{\rm H}]= -2.03\pm 0.15$, $[{\rm Zn}/{\rm
1286: H}]= -2.3\pm 0.3$ and $[{\rm Fe}/{\rm H}]= -2.6\pm 0.1$ (Prochaska
1287: \etal\ 2007b; Piranomonte \etal\ 2008). These measurements together
1288: imply a dust-to-gas ratio roughly $1/20$ of what is seen in the SMC.
1289: Adopting the SMC dust-to-gas ratio (Gordon \etal\ 2003), we estimate a
1290: visual extinction in the host ISM of $A_V\approx 0.01$ (see also
1291: Prochaska \etal\ 2007b). Searches for H$_2$ absorption features in
1292: the echelle spectrum has also yielded null results, placing a
1293: 4-$\sigma$ upper limit on the ISM molecular fraction of the host at
1294: $f_{\rm H_2} < 10^{-6.8}$ (Tumlinson \etal\ 2007). Additional strong
1295: absorbers are found at $z=2.077$, $z=2.01$ and $z=1.99$ (Piranomonte
1296: \etal\ 2008).
1297:
1298: We have observed the field around GRB\,050922C using PANIC on
1299: Magellan in August 2007, and obtained a total integration of 209
1300: minutes. The mean FWHM of the PSF was found to be $\approx
1301: 0.5\arcsec$. The final stacked image is presented in Figure 13, which
1302: has been smoothed using a Gaussian kernel of ${\rm FWHM}=0.5\arcsec$.
1303: The position of the GRB is marked by a circle of 2\arcsec\ radius. No
1304: detectable flux is seen at the immediate location of the GRB. We
1305: place a 2-$\sigma$ limit in the observed $H$-band magnitude of
1306: $AB(H)=26.0$ over a 0.5\arcsec\ diameter aperture for the host galaxy.
1307: At $z=2.199$, observed $H$-band magnitude limit allows us to derive a
1308: limiting rest-frame absolute magnitude of $M_{AB}(B) -
1309: 5\,\log\,h>-18.2$ for the GRB host galaxy.
1310:
1311: Optical images of the field around GRB\,050922c have also been
1312: obtained using LRIS and the $g$ and $R_c$ filters on the Keck I
1313: telescope in July 2006. The mean FWHM of the PSF was found to be
1314: $\approx 1.2\arcsec$ in the combined $g$ image and $\approx
1315: 1.0\arcsec$ in the combined $R_c$ image. No detectable flux is seen
1316: at the immediate location of the GRB. We place 2-$\sigma$ limits of
1317: $AB(g)=27.0$ and $AB(R_c)=26.4$ over a 2\arcsec\ diameter aperture for
1318: the host galaxy. At $z=2.199$, observed $g$-band magnitude limit
1319: allows us to derive a limiting rest-frame absolute magnitude of
1320: $M_{AB}(1500) - 5\,\log\,h>-17.2$ for the GRB host galaxy. This is
1321: comparable to the flux limit in the $H$-band imaging data based on a
1322: mean color of ${\rm UV}-B=1.22$ for $z=2-3$ star-forming galaxies
1323: (Shapley \etal\ 2005).
1324:
1325: \subsection{GRB\,060206 at $z_{\rm GRB}=4.048$}
1326:
1327: \begin{figure}
1328: \begin{center}
1329: \includegraphics[scale=0.5,angle=0]{g060206_acs.ps}
1330: \end{center}
1331: \caption[]{HST ACS/F814W image of the field around GRB\,060206 at
1332: $z_{\rm GRB}=4.048$. The host is detected with $AB({\rm
1333: F814W})=27.6\pm 0.1$. Object $A$ at $\Delta\,\theta\approx 1\arcsec$
1334: southwest of the host has $AB({\rm F814W})=26.22\pm 0.05$ and was
1335: initially noted by Th\"one \etal\ (2008a) as a possible host
1336: candidate. Given its offset from the OT position, we consider galaxy
1337: A a foreground galaxy likely associated with one of the Mg\,II
1338: absorbers at $z=1.48$ or $z=2.26$ found in the afterglow spectra.}
1339: \end{figure}
1340:
1341: This burst was detected by {\it Swift} (Morris \etal\ 2006). An
1342: optical transient was nearly instantaneously identified with $V\approx
1343: 16.7$ (Fynbo \etal\ 2006b; Boyd \etal\ 2006). Spectroscopic follow-up
1344: of the afterglow was carried out by multiple groups (Fynbo \etal\
1345: 2006a; Prochaska \etal\ 2006; Aoki \etal\ 2006; Hao \etal\ 2007). The
1346: host galaxy at $z=4.048$ is found to have $\log\,N(\hI)=20.85\pm 0.10$
1347: and $[{\rm S}/{\rm H}]= -0.85\pm 0.10$ based on moderate resolution
1348: ($\delta\,v \approx 40$ \kms) afterglow spectra (Fynbo \etal\ 2006a;
1349: Th\"one \etal\ 2008). No H$_2$ is detected to a 4-$\sigma$ limit of
1350: $f_{\rm H_2} < 10^{-3.6}$ (c.f.\ Fynbo \etal\ 2006; Tumlinson \etal\
1351: 2007). In addition to the GRB host, two strong Mg\,II absorbers are
1352: found along the line of sight at $z=1.48$ and $z=2.26$ with
1353: $W(2796)=0.95\pm 0.1$ \AA\ and $W(2796)=1.5\pm 0.1$ \AA, respectively
1354: (Aoki \etal\ 2006; Hao \etal\ 2007; Th\"one \etal\ 2008).
1355:
1356: We obtained late-time images of the field around GRB\,060206 on
1357: November 25, 2006, using ACS and the F814W filter on board HST
1358: (PID$=$10817). The images were processed and registered using the
1359: standard pipeline technique. A stacked ACS/F814W image is presented
1360: in Figure 14. At the position of the OT (marked by a circle of
1361: 0.5\arcsec\ radius), we clearly detect a faint source of $AB({\rm
1362: F814W})=27.6\pm 0.1$. We identify the source as the host galaxy of
1363: GRB\,060206. At $z=4.048$, the observed F814W magnitude corresponds
1364: to a rest-frame absolute magnitude of $M_{AB}(1600) -
1365: 5\,\log\,h=-17.7\pm 0.1$ for the GRB host galaxy.
1366:
1367: At $\Delta\,\theta=0.96\pm 0.02\arcsec$ southwest of the host, we
1368: note the presence of galaxy $A$ with $AB({\rm F814W})=26.22\pm 0.05$.
1369: This galaxy was also seen in a deep, ground-based $r'$-band image
1370: published in Th\"one \etal\ (2008). The authors initially identified
1371: galaxy $A$ as a candidate for the GRB host galaxy, but revised the
1372: identification after analyzing the HST images. Given the large
1373: angular distance to the OT (corresponding to $\rho\approx 4.8\ h^{-1}$
1374: kpc at $z=4.408$) and the presence of at least two strong Mg\,II
1375: absorbers in the afterglow spectra, we conclude that galaxy $A$ is
1376: likely a foreground galaxy associated with one of the Mg\,II
1377: absorbers.
1378:
1379: \subsection{GRB\,060607 at $z_{\rm GRB}=3.075$}
1380:
1381: \begin{figure}
1382: \begin{center}
1383: \includegraphics[scale=0.4,angle=0]{g060607_tot.ps}
1384: \end{center}
1385: \caption[]{A smoothed $H$ image of the field around GRB\,060607 at
1386: $z_{\rm GRB}=3.075$. No detectable flux is seen at the location of
1387: the OT. We place a 2-$\sigma$ $H$-band limiting magnitude of
1388: $AB(H)=26.5$ over a 0.5\arcsec\ diameter aperture for the host galaxy.}
1389: \end{figure}
1390:
1391: This burst was detected by {\it Swift} and the optical afterglow was
1392: detected one minute after the trigger by UVOT and the white filter
1393: ($1600-6500$ \AA) on board the satellite (Ziaeepour \etal\ 2006). The
1394: initial brightness was estimated $\approx 15.7$ mag (Ziaeepour \etal\
1395: 2006). A series of high-resolution ($\delta\,v \approx 7$ \kms)
1396: optical spectra of the afterglow were obtained using UVES on the VLT
1397: telescope by Ledoux \etal\ (2006), starting 7.5 minutes after the
1398: trigger. These authors identified the GRB host at $z=3.082$ and noted
1399: two probable DLAs at $z=2.937$ and $z=3.05$. We retrieved the echelle
1400: spectra from the ESO data archive and processed the data using our own
1401: reduction software.
1402:
1403: The afterglow spectrum of GRB\,060607 exhibits a number of unusual
1404: features. First, the GRB host appears to arise in a weak \lya\
1405: absorber at $z=3.075$, with associated Si\,IV and C\,IV absorption
1406: feature but no trace of low-ionization species. We determine
1407: $\log\,N(\hI)=16.85\pm 0.10$ based on a simultaneous Voigt profile
1408: analysis of the Lyman absorption series, using the
1409: VPFIT\footnote{http://www.ast.cam.ac.uk/\char'176rfc/vpfit.html}
1410: software package. This system represents the only optically thin
1411: absorber detected in a GRB host galaxy (Jakobsson \etal\ 2006; Chen
1412: \etal\ 2007a). Second, the strong \lya\ absorber at $z=3.05$,
1413: $\approx 1840$ \kms\ blueshifted from the host, is a Lyman limit
1414: system of $\log\,N(\hI)=19.2\pm 0.1$ with weak metal absorption lines
1415: detected, implying a metallicity $< 1/100$ solar. Third, the strong
1416: \lya\ absorber at $z=2.937$ exhibits complex kinematic profiles in
1417: various ionic transitions, spreading over $\Delta\,v\approx 600$ \kms.
1418: We measure $\log\,N(\hI)=19.6\pm 0.1$.
1419:
1420: We have observed the field around GRB\,060607 using PANIC on
1421: Magellan in August 2007, and obtained a total integration of 309
1422: minutes. The mean FWHM of the PSF was found to be $\approx
1423: 0.5\arcsec$. The final stacked image is presented in Figure 15, which
1424: has been smoothed using a Gaussian kernel of ${\rm FWHM}=0.5\arcsec$.
1425: The position of the GRB is marked by a circle of 2\arcsec\ radius.
1426: Despite the discovery of three strong \lya\ absorbers at
1427: $z=2.93-3.08$, no detectable flux is seen at the immediate location
1428: of the GRB. We place a 2-$\sigma$ limit in the observed $H$-band
1429: magnitude of $AB(H)=26.5$ over a 0.5\arcsec\ diameter aperture for the
1430: host galaxy. At $z=3.075$, observed $H$-band magnitude limit allows
1431: us to derive a limiting rest-frame absolute magnitude of $M_{AB}(B) -
1432: 5\,\log\,h>-18.3$ for the GRB host galaxy.
1433:
1434: Optical images of the field around GRB\,060607 have also been
1435: obtained using LRIS and the $g$ and $R_c$ filters on the Keck I
1436: telescope in July 2006. The mean FWHM of the PSF was found to be
1437: $\approx 1.0\arcsec$ in the combined $g$ image and $\approx
1438: 1.5\arcsec$ in the combined $R_c$ image. The effective seeing was
1439: compromised in the $R_c$ images due to a focus problem with the LRIS
1440: red-side. No detectable flux is found at the immediate location of
1441: the GRB. We place 2-$\sigma$ limits of $AB(g)=26.8$ and
1442: $AB(R_c)=26.5$ over a 2\arcsec\ diameter aperture for the host galaxy.
1443: At $z=3.075$, observed $R_c$-band magnitude limit allows us to derive
1444: a limiting rest-frame absolute magnitude of $M_{AB}(1600) -
1445: 5\,\log\,h>-18.3$ for the GRB host galaxy.
1446:
1447: \subsection{GRB\,070721B at $z_{\rm GRB}=3.626$}
1448:
1449: \begin{figure}
1450: \begin{center}
1451: \includegraphics[scale=0.4,angle=0]{g070721b_tot.ps}
1452: \end{center}
1453: %\plotone{f16.ps}
1454: \caption[]{A smoothed $H$ image of the field around GRB\,070721B at
1455: $z_{\rm GRB}=3.626$. No detectable flux is seen at the location of
1456: the OT (marked by the cross). An extended source is seen at
1457: $\Delta\theta=0.9\arcsec$ sourtheast of the OT with $AB(H)=23.7\pm
1458: 0.1$, which has been confirmed to be at $z=3.09$ and associated with
1459: the foreground DLA along the line of sight (Fynbo 2008, private
1460: communication; Milvang-Jensen \etal\ 2008, in preparation). We place
1461: a 2-$\sigma$ $H$-band limiting magnitude of $AB(H)=25.8$ over a
1462: 0.5\arcsec\ diameter aperture for the host galaxy.}
1463: \end{figure}
1464:
1465: This burst was detected by {\it Swift} (Ziaeepour \etal\ 2007) and
1466: the optical afterglow was detected one minute after the trigger by
1467: UVOT and the white filter ($1600-6500$ \AA) on board the satellite
1468: (Schady \etal\ 2007). The initial brightness was estimated $\approx
1469: 15.9$ mag (Schady \etal\ 2007). Spectroscopic follow-up was carried
1470: out by Malesani \etal\ (2007), who reported $z_{\rm GRB}=3.626$ based
1471: on the presence of a strong \lya\ feature and a series of metal
1472: absorption lines. The host galaxy is found to have
1473: $\log\,N(\hI)=21.50\pm 0.20$ based on moderate resolution afterglow
1474: spectra (Jakobsson \etal\ 2008, in preparation). A additional
1475: intervening damped \lya\ absorber is found at $z=3.09$.
1476:
1477: We have observed the field around GRB\,070721B using PANIC on
1478: Magellan in August 2007, and obtained a total integration of 172
1479: minutes. The mean FWHM of the PSF was found to be $\approx
1480: 0.5\arcsec$. The final stacked image is presented in Figure 16, which
1481: has been smoothed using a Gaussian kernel of ${\rm FWHM}=0.5\arcsec$.
1482: The position of the GRB is marked by the cross within a circle of
1483: 2\arcsec\ radius. No detectable flux is seen at the immediate
1484: location of the GRB. However, an extended source is seen at
1485: $\Delta\theta=0.9\arcsec$ sourtheast of the OT with $AB(H)=23.7\pm
1486: 0.1$.
1487: %. This is likely the adjacent object next to the OT, as pointed
1488: %out in Malesani \etal\ (2007). At $z=3.626$, the projected distance
1489: %between the OT and the extended source would be $\rho=4.6\ h^{-1}$
1490: %kpc.
1491: This source has been confirmed to be at $z=3.09$, associated with the
1492: foreground DLA along the line of sight (Fynbo 2008, private
1493: communication; Milvang-Jensen \etal\ 2008, in preparation). This
1494: represents by far the most luminous DLA galaxy found at $z>2$ (e.g.\
1495: M{\o}ller \etal\ 2002a). At $z=3.09$, the projected distance between
1496: the foreground DLA and the extended source would be $\rho=4.8\ h^{-1}$
1497: kpc. We therefore consider the GRB host galaxy missing in our
1498: $H$-band image and place a 2-$\sigma$ limit of $AB(H)=25.8$ over a
1499: 0.5\arcsec\ diameter aperture for the host galaxy. At $z=3.626$,
1500: observed $H$-band magnitude limit allows us to derive a limiting
1501: rest-frame absolute magnitude of $M_{AB}(3600) - 5\,\log\,h>-19.3$ for
1502: the GRB host galaxy.
1503:
1504: \subsection{GRBs 000301C, 000926, \& 021004}
1505:
1506: \begin{figure}
1507: \begin{center}
1508: \includegraphics[scale=0.32,angle=0]{g000301c_tot.ps}
1509: \end{center}
1510: %\plotone{f17.ps}
1511: \caption[]{Optical images of the field around GRB\,000301C at $z_{\rm
1512: GRB}=2.0404$. The epoch during which the images were taken is
1513: indicated at the bottom of each panel. The late-epoch image (right
1514: panel) has been smoothed using a Gaussian kernel of ${\rm
1515: FWHM}=0.15\arcsec$, which is roughly the size of the PSF. At the
1516: location of the OT, we detect faint emission of the host galaxy in the
1517: late-time image and estimate $AB({\rm clear})=28.9\pm 0.5$ over a
1518: 0.5\arcsec\ diameter aperture (consistent with what was reported in
1519: Fruchter \etal\ 2006).}
1520: \end{figure}
1521:
1522: Early-time imaging and spectroscopic observations of the afterglow
1523: of GRB\,000301C have been presented in Jensen \etal\ (2001), who
1524: reported based on low-resolution afterglow spectra that the source is
1525: at $z=2.0404$. The host ISM in front of the afterglow was found to
1526: have $\log\,N(\hI)=21.2\pm 0.5$. Multi-epoch imaging follow-up was
1527: carried out using STIS on board HST and the clear filter under program
1528: \# 8189 (PI: Fruchter; see also Fruchter \etal\ 2006). We retrieved
1529: available imaging data from the HST data archive and analyzed the
1530: images ourselves. In the left panel of Figure 17, we present a
1531: stacked image of the field obtained in April 2000 with a total
1532: exposure time of 9391 s. The optical transient is clearly visible in
1533: this early epoch image. In contrast, the images (total exposure time
1534: of 7031 s) obtained in February 2001 (right panel of Figure 17)
1535: exhibit only faint emission at the location of the GRB. Attributing
1536: the faint emission to the host galaxy, we measure a total flux of
1537: $AB({\rm clear})=28.9\pm 0.5$ over a 0.5\arcsec\ diameter aperture.
1538: This is consistent with the measurement reported by Fruchter \etal\
1539: (2006). Accounting for the bandpass difference (e.g.\ Figure 3), we
1540: derive $AB(R)=28.8\pm 0.5$ and $M_{AB}(2200) - 5\,\log\,h=-15.2\pm
1541: 0.5$.
1542:
1543: Both GRB\,000926 and GRB\,021004 have been studied extensively by
1544: previous authors. Their host galaxies are identified in early effort
1545: to carry out imaging follow-up with HST. Here we briefly review known
1546: emission and absorption properties of the two host galaxies to
1547: complete the discussion of individual GRB hosts.
1548:
1549: Detailed spectroscopic and imaging studies of GRB\,000926 are
1550: presented in Harrison \etal\ (2001), Fynbo \etal\ (2001), and Castro
1551: \etal\ (2003). The afterglow spectrum displays a DLA of
1552: $\log\,N(\hI)=21.3\pm 0.3$ and a series of metal absorption features
1553: at $z=2.0385$ with an estimated metallicity of $[{\rm Zn}/{\rm
1554: H}]=-0.17\pm 0.15$ and metal abundance ratio of $[{\rm Zn}/{\rm
1555: Fe}]=+1.3\pm 0.15$ (Castro \etal\ 2003). Adopting the dust-to-gas
1556: ratio and the SMC extinction law, we derive a visual extinction of
1557: $A_V=0.15$ and $E(B-V)=0.06$.
1558:
1559: The host galaxy is identified both in \lya\ emission on the ground
1560: (Fynbo \etal\ 2002) and in space images obtained using HST WFPC2 and
1561: the F606W filter (Castro \etal\ 2003). It displays a disturbed
1562: morphology at rest-frame UV wavelengths with extended emission over a
1563: $\approx 15\ h^{-1}$ kpc projected size. Additional near-infrared
1564: $J$-band images show that the host has $AB(J)=24.1_{-0.4}^{+0.7}$
1565: (Christensen \etal\ 2004) which, together with available WFPC2
1566: photometry, constrains its absolute magnitudes at rest-frame UV and
1567: optical wavelengths. Including the intrinsic extinction correction
1568: estimated from absorption line analysis, we find $M_{AB}(2000) -
1569: 5\,\log\,h=-19.63\pm 0.07$ and $M_{AB}(B) -
1570: 5\,\log\,h=-20.26_{-0.4}^{+0.7}$. Comparing the observed SED with a
1571: suite of stellar synthetic models described in \S\ 4.6, we find the
1572: host galaxy is best decribed with a mean stellar age of $\approx 570$
1573: Myr and an exponentially declined SFR of e-folding time 1 Gyr (left
1574: panels of Figure 18). Adopting the best-fit stellar synthetic model,
1575: we can constrain the total underlying stellar mass to be $M_*=
1576: 2.1_{-1.9}^{+4.0}\times 10^9\ h^{-2}$ M$_\odot$, consistent with
1577: $M_*= (1.6\pm 3.1)\times 10^9\ h^{-2}$ M$_\odot$ estimated by Savaglio
1578: \etal\ (2008).
1579:
1580: Detailed spectroscopic and imaging studies of GRB\,021004 are
1581: presented in M{\o}ller \etal\ 2002b, Mirabal \etal\ 2003, Schaefer
1582: \etal\ 2003, Fiore \etal\ 2005, and Fynbo \etal\ 2005. The afterglow
1583: spectrum displays a combination of emission and absorption due to the
1584: hydrogen \lya\ transition and a series of metal absorption lines at
1585: $z=2.329$. The presence of \lya\ emission in the afterglow spectrum
1586: makes it difficult to determine $N(\hI)$ precisely. Based on the
1587: absence of Lyman limit photons, Fynbo \etal\ (2005) derive
1588: $\log\,N(\hI)=19.5\pm 0.5$. This is therefore one of the four GRB
1589: Lyman limit absorbers published so far (including GRB\,060526 in
1590: Jakobsson \etal\ 2006). We have retrieved available echelle spectra
1591: of the OT from the ESO data archive. Based on the absorption
1592: strengths of Si\,II and Fe\,II observed in the combined spectrum, we
1593: take into account necessary corrections for the ionization fraction of
1594: the gas and place an upper limit to the metallicity of the gas at
1595: $[{\rm Si}/{\rm H}]<-1$.
1596:
1597: Optical and near-infrared photometry of the host galaxy of
1598: GRB\,021004 are reported and analyzed in Fynbo \etal\ (2005). The
1599: authors estimate a mean stellar age of $\approx 42$ Myr. Adopting the
1600: optical and near-infrared photometric measurements of Fynbo \etal\
1601: (2005), we derive a rest-frame absolute magnitude of $M_{AB}(B) -
1602: 5\,\log\,h=-20.38\pm 0.15$ or $M_{AB}(2000) - 5\,\log\,h=-19.83\pm
1603: 0.07$. Comparing the observed SED with a suite of stellar synthetic
1604: models described in \S\ 4.6, we find the host galaxy is best decribed
1605: by a single starburst episode that occurred $\approx 39$ Myr ago, and
1606: constrain the total underlying stellar mass to be $M_*=
1607: 1.3_{-0.5}^{+2.9}\times 10^9\ h^{-2}$ M$_\odot$. This is smaller than
1608: $M_*= (7.8\pm 0.7)\times 10^9\ h^{-2}$ M$_\odot$ estimated by Savaglio
1609: \etal\ (2008). The discrepancy may be understood by the poorly
1610: constrained star formation history, displayed in the right panels of
1611: Figure 18.
1612:
1613: We note that two strong Mg\,II absorbers are found at $z=1.38$ and
1614: $z=1.60$ with $W(2796)=1.6$ \AA\ and $W(2796)=1.4$ \AA, respectively.
1615: The large absorption width of these foreground absorbers imply
1616: possible contaminating light in the host imaging observation (c.f.\
1617: Pollack \etal\ 2008). Indeed, the host galaxy is clearly resolved
1618: into two compact clumps in available HST images (e.g.\ Fynbo \etal\
1619: 2005; Chen \etal\ 2007b), but extended \lya\ emission is detected in
1620: ground-based narrow-band imaging follow-up (Jakobsson \etal\ 2005).
1621: In the following discussion, we consider the derived absolute
1622: magnitude as an upper limit to the intrinsic luminosity of this host
1623: galaxy.
1624:
1625: \begin{figure}
1626: \begin{center}
1627: \includegraphics[scale=0.35,angle=0]{h000926_sed.ps}
1628: \includegraphics[scale=0.35,angle=0]{h021004_sed.ps}
1629: \end{center}
1630: \caption[]{Constraints on the galaxy age and star formation history
1631: based on a stellar synthesis model analysis. Left: The broad-band
1632: spectral energy distribution of the host galaxy of GRB\,000926 at
1633: $z=2.0379$ (points with errorbars), in comparison to the best-fit
1634: template (solid histogram) and the predicted broad-band photometric
1635: points (open squares). The best-fit SED is charaterized by a
1636: declining SFR of e-folding time $\tau=300$ Myr, solar metallicity, and
1637: intrinsic dust extinction of $E(B-V)=0.025$ following the SMC
1638: extinction law. The likelihood function of the stellar age for the
1639: host galaxy is display in the bottom left panel, suggesting that a
1640: recent starburst occurred about $650$ Myr prior to the GRB explosion.
1641: Right: A summary of the stellar synthesis model analysis for the host
1642: galaxy of GRB\,021004 at $z=2.329$.}
1643: \end{figure}
1644:
1645: \section{DISCUSSION}
1646:
1647: We have carried out an optical and near-infrared imaging survey of the
1648: fields around 15 GRBs at $z>2$. The GRBs are selected to have
1649: early-time afterglow spectra in order to compare ISM absorption-line
1650: properties with stellar properties. The redshifts of the GRBs span a
1651: range from $z=2.04$ to $z=4.05$, and the neutral hydrogen column
1652: densities of the GRB host ISM span a range from $\log\,N(\hI)=16.9$ to
1653: $\log\,N(\hI)=22.6$ (Figure 1).
1654:
1655: In addition to the five previously known GRB host galaxies, we report
1656: new detections for the host galaxies of GRB\,050820 and GRB\,060206.
1657: The seven identified GRB host galaxies have rest-frame UV absolute
1658: magnitudes spanning from $M_{AB}(U)-5\,\log\,h=-15.2$ to
1659: $M_{AB}(UV)-5\,\log\,h=-19.8$ mag. For sources with a known
1660: dust-to-gas ratio from afterglow absorption-line analysis, we correct
1661: the rest-frame UV magnitude for dust extinction assuming an SMC
1662: extinction law. The inferred SFR spans a range from 0.6 to 3.8
1663: $h^{-2}$ M$_\odot$ yr$^{-1}$. We are able to constrain the underlying
1664: stellar populations for three host galaxies, GRB\,000926, GRB\,021004,
1665: and GRB\,050820, based on comparisons of the observed optical and
1666: near-infrared broad-band colors and a suite of stellar population
1667: synthetic models . We estimate total stellar masses of between
1668: $M_*=(0.9-2.1)\times 10^9\ h^{-2}$ M$_\odot$ for these hosts. We also
1669: place 2-$\sigma$ upper limits for the rest-frame luminosities of the
1670: remaining eight GRB host galaxies based on the depths in available
1671: optical and near-infrared images. Finally, high spatial resolution
1672: images from the HST allow us to deblend the GRB host galaxies from
1673: foreground absorbers and to unveil a range of rest-frame UV morphology
1674: between compact (e.g.\ GRB\,060206) and extended (e.g.\ GRB\,050820)
1675: emission features of the hosts.
1676:
1677: A summary of known absorption-line properties and stellar properties
1678: of the 15 GRB host galaxies is presented in Table 3. Combining
1679: early-time, high-resolution afterglow spectra and the results of
1680: late-time imaging survey of the GRB fields allows us to address a
1681: number of issues regarding both the nature of GRB progenitor
1682: environment and star-forming physics in distant starburst galaxies.
1683:
1684: \subsection{The Luminosity Distribution of GRB Host Galaxies}
1685:
1686: First, we examine the luminosity distribution of GRB host galaxies
1687: above $z=2$ based on the survey result of our sample. The goal is to
1688: characterize the nature of GRB host galaxies based on comparisons of
1689: their luminosity distribution and the luminosity distribution of a
1690: randomly selected sample from the field galaxy population. We
1691: incorporate the non-detections in our follow-up imaging survey by
1692: evaluating the cumulative maximum fraction ${\cal F}_{\rm max}$ of GRB
1693: host galaxies that are fainter than a given UV absolute magnitude
1694: $M_{\rm max}(UV)$ (solid histogram in Figure 19). For the host of
1695: GRB\,030429, we have a constraint only for the rest-frame $B$-band
1696: magnitude. We infer its corresponding UV magnitude based on the mean
1697: color of $\langle {\rm UV}-B\rangle=1.22$ (with an r.m.s. scatter of
1698: 0.3 mag) observed for luminous starburst galaxies at $z=2-3$ in
1699: Shapley \etal\ (2005). For the host of GRB\,060607, we have estimated
1700: limiting magnitudes both in the rest-frame UV and $B$ bands. We adopt
1701: the more sensitive limit based on the conversion of ${\rm
1702: UV}-B=1.22$\footnote{We note that while there is no apparent trend
1703: between $UV-B$ and $M_{UV}$ in the luminous starburst sample of
1704: Shapley \etal\ (2005), it is possible that that fainter dwarf
1705: starburst galaxies may be bluer. Given that only two fields
1706: (GRB\,030429 and GRB\,060607) are affected by this conversion, we find
1707: that the results presented in this section are not sensitive to the
1708: adopted $UV-B$ color.}. The empirical observations of the sample of
1709: 15 GRB fields confirm the previous understanding for $z\sim 1$ GRB
1710: hosts (e.g.\ Le Floc'h \etal\ 2003): as much as 70\% of long-duration
1711: GRBs may originate in galaxies fainter than $0.1\,L_*$.
1712:
1713: At $z=2-4$, the field galaxy population is now well characterized by a
1714: Schechter luminosity function $\phi(M)$ at rest-frame UV wavelengths
1715: with $M_{AB*}-5\,\log\,h=-20.2\pm 0.1$, $\phi_*=(4\pm 0.6)\times
1716: 10^{-3}\ h^{3}\,{\rm Mpc}^{-3}$, and a faint-end slope $\alpha=-1.7\pm
1717: 0.1$ (Bouwens \etal\ 2007; Reddy \etal\ 2008). If the GRB host
1718: galaxies are representative of the field galaxy population, then we
1719: expect that the fraction of host galaxies found in a luminosity
1720: interval is proportional to the space density of galaxies in the
1721: luminosity range. The expected cumulative maximum fraction of the
1722: host galaxies versus UV magnitude can be estimated following
1723: \begin{equation}
1724: {\cal F}_{\rm max}[M>M_{\rm max}]\propto \left [ \sum_{i=1}^{n} \phi(M_i)\,dM
1725: + \sum_{j=1}^{m} \int_{M_{min}}^{M_{\rm max}} \phi(M) dM \right ],
1726: \end{equation}
1727: where the first term extends over $n$ known host galaxies that have
1728: $M_{AB}(UV)>M_{\rm max}$, the second term extends over $m$
1729: unidentified host galaxies that may be as luminous as $M_{\rm max}$,
1730: and $\phi(M)$ is the galaxy luminosity function. The luminosity
1731: distribution is expected to resemble the galaxy luminosity function
1732: with a dominant fraction attributed to faint dwarf galaxies (e.g.\
1733: Jakobsson \etal\ 2005; Fynbo \etal\ 2008).
1734:
1735: We test this hypothesis by calculating ${\cal F}_{\rm max}$ for a
1736: sample of 15 random galaxies that share the known luminosities of the
1737: seven identified GRB host galaxies and the empirical 2-$\sigma$ upper
1738: limits for the remaining eight GRB host galaxies. We experiment with
1739: different $M_{\rm min}$ (with corresponding minimum luminosity $L_{\rm
1740: min}$) in Equation (1). The results for $L_{\rm min}=0.01\,L_*$ and
1741: for $L_{\rm min}=0.005\,L_*$ are shown as the dashed curves in Figure
1742: 19. Smaller $L_{\rm min}$ models predict a larger contributions from
1743: fainter galaxies. The model expections for different adopted $L_{\rm
1744: min}$ clearly deviate from the observed distribution. The hypothesis
1745: that the host galaxies of long-duration GRBs trace random field
1746: galaxies is rejected at $> 99$\% confidence level based on a
1747: Kolmogorov--Smirnov (KS) test.
1748:
1749: Next, under the assumption that GRBs trace instantaneous star
1750: formation, galaxies with higher on-going SFR are expected to have
1751: higher probability to host a GRB event. Applying the rest-frame UV
1752: luminosity as a measure of the on-going SFR, we modify Equation (1) to
1753: include a UV luminosity weighting, $\int L\,\phi(L)\,d\,(L)$, for
1754: calculating the expected ${\cal F}_{\rm max}$,
1755: \begin{eqnarray}
1756: {\cal F}_{\rm max}[M>M_{\rm max}] &\propto&
1757: \sum_{i=1}^{n} 10^{-0.4\,(M_i-M_*)}\,\phi(M_i)\,dM \nonumber \\ &+&
1758: \sum_{j=1}^{m} \int_{M_{min}}^{M_{\rm max}} 10^{-0.4\,(M-M_*)}\,\phi(M) dM .
1759: \end{eqnarray}
1760: %\begin{equation}
1761: %{\cal F}_{\rm max}[M>M_{\rm max}]\propto
1762: %\left [ \sum_{i=1}^{n} 10^{-0.4\,(M_i-M_*)}\,\phi(M_i)\,dM
1763: %+ \sum_{j=1}^{m} \int_{M_{min}}^{M_{\rm max}} 10^{-0.4\,(M-M_*)}\,\phi(M) dM \right ].
1764: %\end{equation}
1765: The results are shown as solid and dotted curves in Figure 19. The
1766: observations are best described under this hypothesis for $L_{\rm
1767: min}=0.001\,L_*$ (solid curve) based on a KS test, but we cannot rule
1768: out models with $L_{\rm min}=0.01\,L_*$ (lower dotted curve) or
1769: $L_{\rm min}=0.0001\,L_*$ (upper dotted curve).
1770:
1771: \begin{figure}
1772: \begin{center}
1773: \includegraphics[scale=0.3,angle=270]{magUV_mod.eps}
1774: \end{center}
1775: \caption[]{Cumulative distribution of the maximum fraction of GRB host
1776: galaxies that are fainter than a given UV absolute magnitude (solid
1777: histogram), based on the seven identified host galaxies and eight
1778: fields with upper limits in our imaging sample. The curves show model
1779: expectations based on different hypothesis for the origin of GRB host
1780: galaxies. We consider two scenarios: (1) GRB host galaxies are
1781: representative of the general galaxy population and (2) GRB host
1782: galaxies originate preferencially in galaxies of higher star formation
1783: rate. Adopting the best-fit UV luminosity function of star-forming
1784: galaxies at $z=3-4$ from Bouwens \etal\ (2007) and Reddy \etal\
1785: (2008), we show the expectations for scenario (1) in dashed curves for
1786: two different minimum luminosity cutoffs, $L_{\rm min}=0.01\,L_*$
1787: (bottom) and $L_{\rm min}=0.005\,L_*$ (top). The expectation for
1788: scenario (2) with $L_{\rm min}=0.001\,L_*$ is shown in the solid
1789: curve, and the dotted curves are for $L_{\rm min}=0.01\,L_*$ (lower)
1790: and $L_{\rm min}=0.0001\,L_*$ (upper). The comparisons provide
1791: empirical evidence supporting the expectation that GRB host galaxies
1792: form an SFR weighted field galaxy sample.}
1793: \end{figure}
1794:
1795: In summary, the sample of seven known GRB host galaxies and eight upper
1796: limits for unidentified hosts allows us to determine the cumulative
1797: maximum fraction of the $z>2$ host galaxy population as a function of
1798: rest-frame UV magnitude. Adopting rest-frame UV luminosity as a
1799: measure of on-going SFR, we find that the empirical sample is best
1800: described by a SFR-weighted sample of the field galaxy population.
1801: Models that do not include SFR weighting can be ruled out at $>99$\%
1802: confidence level. Based on the best-fit SFR-weighted model, we
1803: estimate a median luminosity for the GRB host galaxies at $\approx
1804: 0.1\,L_*$.
1805:
1806: A similar analysis has been presented in Jakobsson \etal\ (2005), who
1807: derived constraints for the luminosity function of GRB host galaxies
1808: under the assumption that the observed brightness distribution of the
1809: hosts follows a luminosity weighted field galaxy population.
1810: % The
1811: %authors compiled a list of known GRB hosts and upper limits published
1812: %in the literature and found that the host luminosity function may be
1813: %characterized by a faint-end slope of $\alpha=-1.55$, consistent with
1814: %what was found for luminous star-forming galaxies by Adelberger \&
1815: %Steidel (2000) with large errorbars. The agreement has become worse,
1816: %given the well-determined UV luminosity function of star-forming
1817: %galaxies at $z>2$ in recent years (e.g.\ Bouwens \etal\ 2007; Reddy
1818: %\etal\ 2008) that shows a steeper faint-end slope, $\alpha=-1.7$.
1819: Our study differs from the approach of Jakobsson \etal\ in that our
1820: analysis is based on a uniform set of photometric measurements from
1821: our own imaging survey. Then we adopt the known UV luminosity
1822: function of the field galaxy population (with a faint-end slope of
1823: $\alpha=-1.7$) and examine different hypotheses for generating the
1824: observed GRB host galaxy sample. We confirm that the host galaxy
1825: population is representative of a UV luminosity weighted sample. The
1826: difference between the host luminosity function of Jakobsson \etal\
1827: and the best-fit luminosity of star-forming galaxies at $z>2$ may be
1828: due to uncertainties in published photometric data of the host
1829: galaxies in their sample and uncertainties in the line-of-sight
1830: absorption properties.
1831:
1832: We note that although the selection criterion of our GRB sample is
1833: based on available afterglow spectra that presumably includes only
1834: GRBs with relatively bright optical afterglows, the broad range in the
1835: isotropic energy release of the GRBs (see Table 1 in Chen \etal\
1836: 2007a) indicates that the GRBs in our sample are not an overly biased
1837: portion of the long-duration GRB population. That is, we have
1838: selected a representative subsample of the unobscured GRB host galaxy
1839: population. However, the presence of 20\% dark bursts that do not
1840: have optical afterglows found (e.g.\ Tanvir \& Jakobsson 2007)
1841: suggests that some GRBs originate in heavily obscured star-forming
1842: regions that are not included in our sample and are likely missed in
1843: the UV selected field galaxy sample as well. To constrain the
1844: fraction of dust obscured star-forming galaxies requires an
1845: independent study of the fields around dark bursts.
1846:
1847:
1848: \subsection{The Luminosity--Metallicity Relation in GRB Host Galaxies}
1849:
1850: \begin{figure}
1851: \begin{center}
1852: \includegraphics[scale=0.4,angle=0]{gdla_w1526.eps}
1853: \end{center}
1854: \caption[]{Absolute UV magnitude of GRB host galaxies versus
1855: rest-frame absorption equivalent width of the Si\,II $\lambda\,1526$
1856: transition in the hosts as observed in afterglow spectra. For this
1857: study, we consider only host DLAs because the lyman limit absorbers do
1858: not include neutral ISM in the host galaxies. The Si\,II
1859: $\lambda\,1526$ transition is typically saturated at rest-frame
1860: absorption equivalent width $W(1526)>0.3$ \AA. The line width
1861: therefore provides a measure of the velocity field of cool clouds
1862: along the line of sight through a halo (e.g.\ Prochaska \etal\ 2008).
1863: We find a moderate trend with stronger $W(1526)$ appearing in more
1864: luminous hosts at nearly 95\% confidence level.}
1865: \end{figure}
1866:
1867: We combine known absorption-line properties with estimated galaxy UV
1868: luminosity to investigate the gas kinematics and chemical enrichment
1869: in the ISM of GRB host galaxies. The goals of this study are (1) to
1870: examine the physical processes that determine the observed
1871: absorption-properties in the host ISM, (2) to investigate whether
1872: there is a metallicity cutoff in GRB host galaxies as favored by
1873: various theoretical models (e.g.\ Hirschi \etal\ 2005; Yoon \& Langer
1874: 2005; Woosley \& Heger 2006), and (3) to probe the
1875: luminosity--metallicity relation below the magnitude limit of most
1876: previous studies at $z>2$ (e.g.\ Erb \etal\ 2006; Maiolino \etal\ 2008).
1877:
1878: We first compare rest-frame absorption equivalent widths of the Si\,II
1879: $\lambda\,1526$ transition $W(1526)$ observed in the host with the
1880: absolute UV magnitude. For this study, we consider only host DLAs
1881: because the lyman limit absorbers do not include neutral ISM in the
1882: host galaxies. The Si\,II $\lambda\,1526$ transition is typically
1883: saturated at $W(1526)>0.3$ \AA. The line width provides a measure of
1884: the velocity field of cool clouds along the line of sight through a
1885: galactic halo, rather than the total gas column density (e.g.\
1886: Prochaska \etal\ 2008). We adopt the $W(1526)$ measurements published
1887: in Prochaska \etal\ (2008). Including additional measurement for
1888: GRB\,060206 in public afterglow spectra from Subaru Science Data
1889: Archive (Aoki \etal\ 2006), we have assembled eight GRB hosts galaxies
1890: for this study.
1891:
1892: Figure 20 shows that there exists a relatively significant trend with
1893: stronger Si\,II transitions appearing in more luminous hosts.
1894: Including upper limits, we find based on a generalized Kendall test
1895: that the probability of a positive correlation between $M_{UV}$ and
1896: $W(1526)$ is nearly 95\%. While the positive correlation is similar
1897: to the mass-metallicity relation seen in field galaxies (e.g.\ Erb
1898: \etal\ 2006; Maiolino \etal\ 2008) , we note that half of the host galaxies have
1899: $W(1526)\aapg 1.5$ \AA, corresponding to a velocity width of $\sim 300$
1900: \kms. Attributing the large line width to gravitational motion of
1901: clouds within the host dark matter halos would require a halo mass $>
1902: 10^{11}\ h^{-1}\,{\rm M_\odot}$, comparable to the mass scale found
1903: for halos that host luminous starburst galaxies at $z\sim 2$ (e.g.\
1904: Conroy \etal\ 2008) but at odds with expectations for starburst
1905: galaxies of low stellar mass and low luminosity. But because $M_{UV}$
1906: serves as a measure of the on-going SFR, the observed $M_{UV}$ versus
1907: $W(1526)$ correlation implies that the velocity field observed in GRB
1908: host galaxies is driven by galactic outflows.
1909:
1910: %Recall that an afterglow line of sight passes through only part of the
1911: %host galaxy and therefore the measurement of $W(1526)$ from afterglow
1912: %spectra does not represent a global value of the velocity field in the
1913: %host. Under a simple assumption that sightlines through a larger
1914: %$N(\hI)$ cover a larger pathlength in the host galaxies, we
1915: %re-evaluate the correlation in Figure 19 by applying a scaling
1916: %relation with the observed $N(\hI)$. We find based on a maximum
1917: %likelihood analysis that including $N(\hI)$ scaling does not improve
1918: %the significance of the observed $M_{UV}$ versus $W(1526)$
1919: %correlation.
1920:
1921: Next, we examine whether there exists a correlation between host
1922: luminosity and ISM metallicity, in comparison to what is known for
1923: luminous starburst galaxies published by Erb \etal\ (2006). We note
1924: two apprarent caveats in our study. First, chemical abundances of the
1925: GRB host galaxies in our sample are determined for the cold neutral
1926: medium using absorption-line techniques, whereas the metallicities of
1927: field galaxies are determined from the integrated emission-line fluxes
1928: of their H\,II regions. Although extensive studies have yet to be
1929: undertaken to examine possible systematic differences between
1930: absorption- and emission-line abundance measurements, available
1931: evidence based on limited studies of nearby starburst galaxies have
1932: yielded consistent metallicity measurements using either
1933: absorption-line or emission-line techniques (see Russell \& Dopita
1934: 1992 and Welty \etal\ 1997, 1999 for the Large and Small Magellanic
1935: Clouds; Lecavelier des Estangs \etal\ 2004 for I\,Zw 18; and
1936: Schulte-Ladbeck \etal\ 2005 and Bowen \etal\ 2005 for
1937: SBS\,1543$+$593). In the few cases where emission and absorption line
1938: measurements have been compared in galaxies at $z > 2$, the two
1939: methods have been found to give consistent answers to within a factor
1940: of $\sim 2$ (e.g.\ Pettini 2006).
1941:
1942: Second, absorption-line measurements are for the ISM along the
1943: afterglow line of sight, whereas emission-line observations are
1944: integrated measurements averaged over the entire galaxies. A large
1945: metallicity gradient is commonly seen in nearby galaxies with a slope
1946: varying between $d\,Z/d\,r \approx -0.02$ dex / kpc to $-0.07$ dex /
1947: kpc (e.g.\ Zaritsky \etal\ 1994; van Zee \etal\ 1998; Kennicutt \etal\
1948: 2003). Therefore, a metallicity measurement for individual sightlines
1949: may not be representative of the global mean value of the entire host
1950: galaxy. We note, however, that $z>2$ galaxies are relatively compact
1951: with typical half light radii of $\approx 2$ kpc (Bouwens \etal\ 2004;
1952: Law \etal\ 2007). Previous observations (e.g.\ Bloom \etal\ 2002;
1953: Fruchter \etal\ 2006) and current findings for GRB\,050820 and
1954: GRB\,060206 show that the GRBs occur within $\aapl 2$ kpc radius of
1955: their host galaxies. This is also consistent with the large $N(\hI)$
1956: observed in the afterglow spectra with the exception of four non-DLAs.
1957: Spatial variation of the observed metallicity across distant
1958: star-forming galaxies is therefore not expected to exceed 0.2 dex. We
1959: proceed with an analysis that compares the ISM metal content of GRB
1960: galaxies with those of known starburst galaxies at $z\approx 2$.
1961:
1962: Figure 21 shows the the luminosity--metallicity relation reproduced
1963: from Erb \etal\ (2006) for starburst galaxies at $\langle
1964: z\rangle=2.3$ (open squares). The oxygen abundances are measured
1965: using the $N2$ index, which is a measure of the flux ratio between the
1966: observed [N\,II] $\lambda\,6584$ to H$\alpha$ lines\footnote{See Erb
1967: \etal\ 2006 for a discussion of possible systematic uncertainties
1968: associated with the $N2$ index. While it is known that the $N2$ index
1969: saturates at solar metallicity, the saturation does not affect our
1970: comparison because GRB host galaxies appear to have mostly sub-solar
1971: abundances (e.g.\ Prochaska \etal\ 2007a).}. The rest-frame $B$-band
1972: magnitudes published in Erb \etal\ have been converted to UV magnitude
1973: according to $UV=B+1.22$ to facilitate comparisons with the GRB host
1974: galaxy population and with predictions from numerical simulations
1975: (crosses in Figure 21, reproduced from Fynbo \etal\ 2008). The $UV-B$
1976: color conversion is estimated based on the mean $R-K$ colors for these
1977: galaxies in Shapley \etal\ (2005).
1978:
1979: We include in Figure 21 measurements for nine GRB host galaxies for
1980: which measurments (solid points) or constraints (open circles) on the
1981: ISM metallicity are available. Only known host DLAs are considered in
1982: the comparison here. Left arrows represent the fields (GRB\,050401,
1983: GRB\,050922c and GRB\,050730) for which no hosts have been found and
1984: 2-$\sigma$ upper limits to the intrinsic luminosity are shown. The
1985: lower limits represent those GRB fields for which only moderate
1986: resolution afterglow spectra are available and the observed
1987: absorption-line strengths represent a lower limit to the intrinsic
1988: values (e.g.\ Prochaska 2006).
1989:
1990: \begin{figure}
1991: \begin{center}
1992: \includegraphics[scale=0.4,angle=0]{magz_hiz.eps}
1993: \end{center}
1994: \caption[]{The luminosity--metallicity relation in distant
1995: galaxies. Solid points with errorbars are GRB host galaxies with
1996: accurate absorption-line metallicity measurements, while open circles
1997: are GRB host galaxies with constraint on their ISM metallicities due to
1998: either insufficient spectral resolution in the afterglow spectra or
1999: unknown ionization fraction of the ISM. Open squares are
2000: emission-line metallicity measurements for luminous starburst galaxies
2001: at $z\approx 2$ from Erb \etal\ (2006). The oxygen abundance was
2002: evaluated using the observed [N\,II] to H$\alpha$ line ratio. We have
2003: converted the rest-frame $B$-band magnitude to UV magnitude according
2004: to $UV=B+1.22$, which is estimated based on the mean $R-K$ colors for
2005: these galaxies in Shapley \etal\ (2005). Crosses represent the
2006: predicted luminosity--metallicity relation for starburst galaxies at
2007: $z\sim 3$ (Fynbo \etal\ 2008) based on numerical simulations by
2008: Sommer-Larsen \& Fynbo (2008). }
2009: \end{figure}
2010:
2011: Three interesting features are seen in Figure 21. First, the majority
2012: of the GRB host galaxies are fainter than the faintest starburst
2013: galaxies studied in magnitude-limited surveys (open squares) and offer
2014: a promising probe to extend the studies of ISM metal enrichment to
2015: fainter luminosity limits than existing faint galaxy surveys
2016: (Djorgovski \etal\ 2004). Second, the metallicities found in GRB host
2017: galaxies span two orders of magnitude from $-2.2$ dex below solar to
2018: $\approx -0.2$ below solar, while the rest-frame UV luminosity of the
2019: galaxies extends from $\approx 1/2\, L_*$ to fainter than $0.05\,L_*$.
2020: There is no apparent upper metallicity cutoff in the sample of GRB
2021: host galaxies at $z>2$. Finally, we include predictions from
2022: numerical simulations performed by Sommer-Larsen \& Fynbo (2008) that
2023: include SNe feedback. The crosses are reproduced from Fynbo \etal\
2024: (2008) for comparison with observations. Despite the presence of
2025: lower limits in the metallicities of three GRB hosts, there is a
2026: moderate trend that suggests a steeper luminosity--metallicity
2027: relation than what is seen in simulations. We note that this is
2028: unlikely due to a selection bias against metal enriched faint galaxies
2029: based on the study presented in \S\ 5.1, which shows that the GRB host
2030: galaxies are consistent with an SFR selected field galaxy sample.
2031:
2032: Different theoretical studies have been carried out to understand the
2033: key astrophysical processes that determine the luminosity--metallicity
2034: relation found in galaxies from different epochs (e.g.\ Dekel \& Woo
2035: 2003; Tassis \etal\ 2008; Brooks \etal\ 2007; Finlator \& Dav\'e
2036: 2008). The steep decline in ISM metallicity toward fainter magnitudes
2037: may be interpreted as a signature of SNe driven outflows that removes
2038: metals more effectively in lower-mass halos than in more massive ones
2039: (e.g.\ Dekel \& Woo 2003), or as a signature of inefficient star
2040: formation in low mass halos due to low gas density (e.g.\ Tassis
2041: \etal\ 2008; Robertson \& Kravtsov 2008). The model expectations
2042: presented in Figure 20 includes SNe feedback but no ISM radiation
2043: field to account for subsequent destruction of molecules in
2044: star-forming regions. The observed steeper luminosity--metallicity
2045: relation in faint galaxies and the absence of molecular gas in GRB
2046: host ISM (Tumlinson \etal\ 2007) suggests that an enhanced ISM
2047: radiation field from young stars may be important to effectively
2048: reduce subsequent star formation and ISM chemical enrichment in dwarf
2049: galaxies.
2050:
2051: \subsection{Empirical Constraints for the ISM Radiation Field in GRB Host Galaxies}
2052:
2053: To understand the lack of molecular gas in GRB host ISM, Tumlinson
2054: \etal\ (2007) performed a comparison between observations and
2055: predictions from a grid of models that cover a parameter space spanned
2056: by four unknown properties. These include gas density, metallicity,
2057: clouds size and interstellar radiation field. The authors showed that
2058: the lack of molecular gas in the host ISM can be understood by a
2059: combination of low metallicity (low grain production rate) and high
2060: interstellar UV radiation field in the host ISM. Recently, Whalen
2061: \etal\ (2008) carried out a set of numerical simulations to examine
2062: whether the absence of molecular gas is due to the afterglow radiation
2063: field or an enhanced interstellar UV radiation field from a
2064: pre-existing H\,II region where the progenitor star resides. Taking
2065: into account known metallicities from afterglow absorption analysis,
2066: Whalen \etal\ concluded that, similar to what is found in the
2067: Magellanic Clouds, ISM radiation fields of intensities up to 100 times
2068: the Galactic mean are necessary to explain the absence of H$_2$ in the
2069: host ISM.
2070:
2071: With the resolved galaxy morphologies seen in the HST images of
2072: GRB\,000926, GRB\,030323, GRB\,050820A, and GRB\,060206, we can
2073: measure directly the mean radiation field in the ISM of the host
2074: galaxies at near UV wavelengths ($\approx 1500-2000$ \AA). The mean
2075: UV radiation intensity is determined by averaging the observed UV flux
2076: over the extent of each host galaxy. For comparison with models, we
2077: convert from the observed mean radiation intensity at near UV
2078: wavelengths to the far UV Lyman-Warner band ($11.8-13.6$ eV) based on
2079: the Galactic UV radiation spectrum of Gondhalekar \etal\ (1980). The
2080: estimated far UV radiation fields $I_0$ of individual host galaxies
2081: are presented in Table 3. Adopting a mean Galactic radiation field at
2082: far UV wavelengths of $I_0=2.5\times 10^{-8}$ photons cm$^{-2}$
2083: s$^{-1}$ Hz$^{-1}$ (Gondhalekar \etal\ 1980), we find that, despite a
2084: moderate SFR, the ISM radiation field in the GRB host galaxies spans a
2085: range over $35 - 350 \times$ the Galactic mean value due to a
2086: relatively compact size. The observations provide empirical support
2087: for the previous theoretical understanding (e.g.\ Whalen \etal\ 2008)
2088: that strong UV radiation fields in low-metallicity ISM environment may
2089: be a dominant factor for suppressing the formation of molecules and
2090: therefore subsequent star formation in dwarf galaxies.
2091:
2092: \subsection{Properties of Foreground Mg\,II Absorbing Galaxies}
2093:
2094: A puzzling observation of the absorption properties along GRB lines
2095: of sight is the apparent overdensity of strong (rest-frame absorption
2096: equivalent width $W(2796)>1$ \AA) Mg\,II absorbers (Prochter \etal\
2097: 2006). Afterglow spectra exhibit on average $\approx 4\times$ more
2098: strong Mg\,II absorbers than random QSO spectra (Prochter \etal\ 2006;
2099: Sudilovsky \etal\ 2007), although such over abundance is not seen in
2100: C\,IV absorbers at somewhat higher redshift (Sudilovsky \etal\ 2007;
2101: Tejos \etal\ 2007). Various scenarios have been considered to explain
2102: this discrepancy, including dust extinction due to the presence of
2103: these absorbers that biases observations of QSO sightlines, the GRBs
2104: being gravitationally lensed by the foreground absorbers, and the
2105: absorbers being intrinsic to the GRBs (Prochter \etal\ 2006; Porciani
2106: \etal\ 2007). However, none of these scenarios alone is found
2107: sufficient to explain the observed overabundance along afterglow
2108: sightlines. A different scenario has been proposed by Frank \etal\
2109: (2007), who consider different beam sizes between QSOs and GRB
2110: afterglows as a possible explanation, but two important consequences
2111: are associated with this scenario. First, a partial covering of
2112: Mg\,II gas is expected along QSO sightlines. Second, a skewed
2113: frequency distribution of Mg\,II absorbers is also expected along GRB
2114: sightlines. None is confirmed in empirical data (e.g.\ Pontzen \etal\
2115: 2007). We refer the readers to Porciani \etal\ (2007) for a detailed
2116: discussion.
2117:
2118: Late-time HST images of the fields around known GRBs offer a
2119: detailed view of the galaxy environment around known foreground Mg\,II
2120: absorbers along the sightlines toward the GRBs (e.g.\ Pollack \etal\
2121: 2008). Six of the 15 GRBs sightlines in our sample have known strong
2122: Mg\,II absorbers of $W(2796) > 1$ \AA\ in the foreground (e.g.\
2123: Prochter \etal\ 2006), three of which have deep HST images available.
2124: Including the field around GRB\,060418 (Pollack \etal\ 2008), we find
2125: that in all four cases the HST images have uncovered at least one
2126: galaxy at angular separation $\Delta\,\theta \aapl 1''$ from the
2127: afterglow sightlines. In contrast, such galaxies at small
2128: $\Delta\,\theta$ are clearly absent in fields with no known strong
2129: Mg\,II absorbers (Figure 22). A summary of the known Mg\,II absorbers
2130: and possible candidate absorbing galaxies is presented in Table 4.
2131:
2132: \begin{figure}
2133: \begin{center}
2134: \includegraphics[scale=0.35,angle=270]{gaden.ps}
2135: \end{center}
2136: \caption[]{Cumulative number of galaxies versus angular distance to
2137: the afterglow line of sight for GRB fields with (top) and without
2138: (bottom) intervening strong Mg\,II absorbers found in afterglow
2139: spectra. All seven fields have high spatial resolution HST images
2140: available for identifying galaxies brighter than $AB(R)=28$. In all
2141: four fields with known intervening strong Mg\,II galaxies, we find
2142: additional objects at $\Delta\,\theta<2\arcsec$ from the GRB afterglow
2143: position. In contrast, none is seen at this small angular separation
2144: in fields without known Mg\,II absorbers.}
2145: \end{figure}
2146:
2147: Although only one of the candidate galaxies in Table 4 has been
2148: spectroscopically confirmed (the Mg\,II absorber at $z=0.84$ toward
2149: GRB\,030429), the finding of additional faint galaxies along the GRB
2150: sightlines strongly disfavors the absorbers being intrinsic to the
2151: GRBs. In addition, these candidate galaxies exhibit a broad range of
2152: optical-infrared colors (c.f.\ the blue colors of galaxies toward
2153: GRB\,050820 shown in Table 2 and Figure 9 and those found toward
2154: GRB\,060418 in Pollack \etal\ 2008). Dust extinction does not appear
2155: to be a uniform factor across different GRB sightlines that results in
2156: the higher incidence of strong Mg\,II absorbers. Comparisons of the
2157: absorbing galaxy properties found along afterglow and QSO sightlines
2158: (O'Meara \etal\ 2006; Becker \etal\ in preparation) should provide
2159: further insights for understanding the differential incidences of
2160: Mg\,II absorbers. The morphology of additional faint emission near
2161: the absorbing galaxy toward GRB\,030429 (Figure 6) is suggestive of a
2162: lensed event. Additional optical imaging and spectroscopic data are
2163: necessary to test the lensing hypothesis for this source.
2164:
2165: A conclusive answer to the observed overdensity of foreground Mg\,II
2166: absorbers along GRB sightlines requires a larger sample of imaging and
2167: spectroscopic data of the candidate absorbing galaxies. Based on the
2168: current finding, however, we caution that the presence of intervening
2169: galaxies at small angular distances to the GRBs introduces
2170: non-negligible contamination for identifying GRB host galaxies based
2171: on imaging data alone. High spatial resolution images are crucial for
2172: resolving the host galaxies from foreground absorbers. While the
2173: properties of the Mg\,II absorbers are beyond the scope of this paper,
2174: we emphasize that in the absence of spectroscopic observations it
2175: will be necessary to take into account line-of-sight absorption-line
2176: properties in afterglow spectra when evaluating the uncertainty of an
2177: imaging identification of GRB host galaxies.
2178:
2179: \section{Summary}
2180:
2181: We present a study of faint galaxies uncovered along GRB lines of
2182: sight, based on an optical and near-infrared imaging survey of the
2183: fields around 15 GRBs at $z>2$. The GRBs are selected with available
2184: early-time afterglow spectra in order to compare ISM absorption-line
2185: properties with stellar properties. The redshifts of the GRBs span a
2186: range from $z=2.04$ to $z=4.05$. The neutral hydrogen column
2187: densities of the GRB host ISM span a range from $\log\,N(\hI)=16.9$ to
2188: $\log\,N(\hI)=22.6$. Our analysis differs from previous studies in
2189: that we have obtained a uniform set of photometric data from our own
2190: imaging survey, reducing systematic uncertainties in photometric
2191: measurements of the host galaxies.
2192:
2193: In addition to the five previously studied GRB host galaxies, we
2194: consider new detections for the host galaxies of GRB\,050820 and
2195: GRB\,060206. We also place 2-$\sigma$ upper limits for the rest-frame
2196: luminosities of the remaining eight GRB host galaxies based on the
2197: depths in available optical and near-infrared images. Combining
2198: early-time, high-resolution afterglow spectra and late-time imaging
2199: survey of the GRB fields allows us to address a number of issues
2200: regarding both the nature of GRB progenitor environment and
2201: star-forming physics in distant starburst galaxies. The line-of-sight
2202: properties uncovered in the afterglow spectra have also proven to be
2203: valuable for filtering potential contaminations due to foreground
2204: galaxies. The results of our study are summarized as the following:
2205:
2206: 1. GRB host galaxies exhibit a broad range of rest-frame UV
2207: absolute magnitudes spanning from $M_{AB}(U)-5\,\log\,h\aapg -15$ to
2208: $M_{AB}(UV)-5\,\log\,h=-19.8$. The distribution of rest-frame UV
2209: luminosities shows that the GRB host galaxy population is best
2210: described by a UV luminosity weighted random galaxy population with a
2211: median luminosity of $\langle L(UV)\rangle=0.1\,L_*$. Models that
2212: include no luminosity weighting are ruled out at $>99$\% confidence
2213: level. This result demonstrates that GRB host galaxies are
2214: representative of unobscured star-forming galaxies at $z>2$.
2215:
2216: 2. There exists a relatively significant correlation between UV
2217: luminosity and Si\,II $\lambda\,1526$ transition in GRB host galaxies.
2218: A generalized Kendall test including upper limits indicates that the
2219: probability of a positive correlation between $M_{UV}$ and $W(1526)$
2220: is nearly 95\%. Attributing the observed large line widths of
2221: $W(1526)\aapg 1.5$ \AA\ to gravitational motions of gaseous clouds
2222: would require massive halos that are also more rare at $z>2$.
2223: Adopting $M_{UV}$ as a measure of on-going SFR and the lack of
2224: correlation between SFR and total stellar mass in star-forming
2225: galaxies at $z\approx 2$, we therefore interpret the observed $M_{UV}$
2226: versus $W(1526)$ correlation as indicating that the velocity field
2227: observed in GRB host galaxies is driven by galactic outflows.
2228:
2229: 3. GRB host ISM exhibit a broad range of chemical enrichment, from
2230: $<1/100$ solar to $\sim 1/2$ solar. No apparent metallicty cutoff is
2231: seen in the high-redshift host galaxy population. Similar to nearby
2232: galaxies, a tentative trend of declining ISM metallicity toward
2233: fainter ($< 0.1\,L_*$) luminosities is seen in the star-forming galaxy
2234: population at $z=2-4$. The slope is steeper than what is expected in
2235: some numerical simulations that incorporate supernovae feedback.
2236: Together with an absence of molecular gas and the presence of large
2237: amounts atomic gas, the observed luminosity--metallcity relation may
2238: be explained by low star formation efficiency in dwarf galaxies.
2239:
2240: 4. We measure the interstellar radiation field using resolved
2241: rest-frame UV morphologies in available HST images. We find that the
2242: UV radiation field in GRB host ISM spans a range over $\approx
2243: 35-350\times$ higher than the Galactic mean value. The strong ISM
2244: radiation field observed in GRB host galaxies is expected to increase
2245: the formation threshold of molecules and suppress subsequent star
2246: formation, supporting the hypothesis that star formation efficiency in
2247: dwarf galaxies is reduced due to the strong radiation field of
2248: existing H\,II regions.
2249:
2250: 5. We examine the galaxy environment in available HST images of
2251: strong Mg\,II absorbers found along GRB sightlines. In all fields
2252: with known strong absorbers, we identify at least one faint galaxy at
2253: $\aapl 1''$ from the afterglow position. In contrast, such galaxies at
2254: small $\Delta\,\theta$ are clearly absent in fields with no known
2255: strong Mg\,II absorbers. The finding of additional faint galaxies
2256: along the GRB sightlines strongly disfavors the strong absorbers being
2257: intrinsic to the GRBs. Additional imaging and spectroscopic data of
2258: the candidate absorbing galaxies are necessary to investigate the
2259: effect of dust and gravitational magnification, but the presence of
2260: intervening galaxies at small angular distances to the GRBs increases
2261: the ambiguity of identifying GRB host galaxies. It is necessary to
2262: combine high spatial resolution images and early-time afterglow
2263: spectra for accurate identifications of the host galaxies.
2264:
2265:
2266: \acknowledgments
2267:
2268: The authors thank N. Gnedin, R. Kennicutt, B. M\'enard,
2269: E. Ramirez-Ruiz, C. Thom for important discussions. We also thank
2270: J.\ Fynbo and P.\ Jakobsson for constructive comments that helped
2271: improve the presentation of the paper. Support for the HST program
2272: \#10817 was provided by NASA through a grant from the Space Telescope
2273: Science Institute, which is operated by the Association of
2274: Universities for Research in Astronomy, Inc., under NASA contract NAS
2275: 5-26555. H.-W.C. acknowledges partial support from an NSF grant
2276: AST-0607510 and HST-GO-10817.01A.
2277:
2278: \begin{references}
2279:
2280: \vskip 0.2in
2281:
2282: \reference{} Adelberger, K. L. \& Steidel, C. C. 2000, ApJ, 544, 218
2283:
2284: \reference{} Akerlof, C. \etal\ 1999, Nature, 398, 400
2285:
2286: \reference{} Angelini, L. \etal\ 2005, GCN Circ. 3161
2287:
2288: \reference{} Aoki, K. \etal\ 2006, GCN Circ. 4703
2289:
2290: \reference{} Barbier, L. \etal\ 2005, GCN Circ. 3162
2291:
2292: \reference{} Berger, E. \etal\ 2002, ApJ, 581, 981
2293:
2294: \reference{} Bernstein, R. \etal\ 2003, Proc. SPIE, 4841, 1694
2295: %, Shectman, S.~A., Gunnels, S.~M., Mochnacki, S., \&
2296: % Athey, A.~E. 2003, Proc. SPIE, 4841, 1694
2297:
2298: \reference{} Bloom, J.~S., Kulkarni, S.~R., \& Djorgovski, S.~G. 2002, AJ, 123, 1111
2299:
2300: \reference{} Bloom, H.~S. \etal\ 2008, ApJ submitted (arXiv: 0803.3215)
2301:
2302: \reference{} Bouwens, R. J., Illingworth, G. D., Blakeslee, J. P., Broadhurst, T. J., \& Franx, M. 2004, ApJ, 611, L1
2303:
2304: \reference{} Bouwens, R. J., Illingworth, G. D., Franx, M., \& Ford, H. 2007, ApJ, 670, 928
2305:
2306: \reference{} Bowen, D. V., Jenkins, E. B., Pettini, M., \& Tripp, T. M. 2005, ApJ, 635, 880
2307:
2308: \reference{} Boyd, P., Hunsberger, S., \& Gronwall, C. 2006, GCN Circ. 4684
2309:
2310: \reference{} Brooks, A. M., Governato, F., Booth, C. M., Willman, B., Gardner, J. P., Wadsley, J., Stinson, G., \& Quinn, T. 2007, ApJ, 655, L17
2311:
2312: \reference{} Bruzual, G., \& Charlot, S. 2003, MNRAS, 344, 1000
2313:
2314: \reference{} Castro, S. \etal\ 2003, ApJ, 586, 128
2315:
2316: \reference{} Castro Cer\'on, J. M., Michalowski, M J., Hjorth, J., Watson, D., Fynbo, J. P. U., \& Gorosabel, J. 2006, ApJ, 653, L85
2317:
2318: \reference{} Castro Cer\'on, J. M., Michalowski, M J., Hjorth, J., Malesani, D., Gorosabel, J., Watson, D., \& Fynbo, J. P. U. 2008, ApJ submitted (arXiv:0803.2235)
2319:
2320: \reference{} Chary, R., Becklin, E. E., \& Armus, L. 2002, AJ, 566, 229
2321:
2322: \reference{} Chen, H.-W. \& Lanzetta, K. M. 2003, ApJ, 597, 706
2323:
2324: \reference{} Chen, H.-W. \& Marzke, R. O. 2004, ApJ, 615, 603
2325:
2326: \reference{} Chen, H.-W., Prochaska, J. X., Bloom, J. S., \& Thompson, I. B. 2005, ApJ, 634, L25
2327: \reference{} Chen, H.-W., Prochaska, J. X., \& Gnedin, N. 2007a, ApJ, 667, L125
2328:
2329: \reference{} Chen, H.-W., Prochaska, J. X., Ramirez-Ruiz, E., Bloom, J. S., Dessauges-Zavadsky, M., \& Foley, R. J. 2007b, ApJ, 663, 420
2330:
2331: \reference{} Chen, H.-W. \& Tinker, J. L. 2008, ApJ in press (arXiv:0801.2169)
2332:
2333: \reference{} Christensen, L., Hjorth, J., \& Gorosabel, J. 2004., A\&A, 425, 913
2334:
2335: \reference{} Conroy, C., Shapley, A. E., Tinker, J. L., Santos, M. R., \& Lemson, G.\ 2008, ApJ, 679, 1192
2336:
2337: \reference{} Dekel, A. \& Woo, J. 2003, MNRAS, 344, 1131
2338:
2339: \reference{} D'Elia, V. \etal\ 2005, GCN Circ. 4044
2340:
2341: \reference{} D'Elia, V. \etal\ 2007, A\&A, 467, 629
2342:
2343: \reference{} D'Elia V. \etal\ 2008, ApJL submitted (arXiv:0804.2141)
2344:
2345: \reference{} Djorgovski, S. G., et al.\ 2004, in ASP Conf. Ser., Gamma-Ray Bursts in the Afterglow Era, ed. L. Piro \& M. Feroci (San Francisco: ASP), p.\ 249
2346:
2347: \reference{} D'Odorico, S., Cristiani, S., Dekker, H., et~al. 2000, in SPIE 4005,
2348: Discoveries and Research Prospects from 8- to 10-Meter-Class Telescopes,
2349: ed. J. Bergeron, 121
2350:
2351: \reference{} Doty, J. \etal\ 2003, GCN Circ. 2211
2352:
2353: \reference{} Eliasdottir \etal\ 2008, ApJ submitted
2354:
2355: \reference{} Erb, D. K., Shapley, A. E., Pettini, M., Steidel, C. C., Reddy, N. A., \& Adelberver, K. L. 2006, ApJ, 644, 813
2356:
2357: \reference{} Finlator, K. \& Dav\'e, R. 2008, MNRAS, 385, 2181
2358:
2359: \reference{} Fiore, F. \etal\ 2005, ApJ, 624, 853
2360:
2361: \reference{} Foley, R.~J., Chen, H.-W., Bloom, J.~S., \& Prochaska, J.~X. 2005, GCN Circ. 3949
2362:
2363: \reference{} Ford, H. C., et al.\ 1998, Proc. SPIE, 3356, 234
2364:
2365: \reference{} Fox, D.~B. \& Cenko, S.~B. 2005, GCN Circ. 3829
2366:
2367: \reference{} Frank, S., Bentz, M. C., Stanek, K. Z., Mathur, S., Dietrich, M., Peterson, B. M., \& Atlee, D. W. 2007, Ap\&SS, 312, 325
2368:
2369: \reference{} Fruchter, A., Vreeswijk, P., Rhoads, J., \& Burud, I. 2001, GCN Circ. 1200
2370:
2371: \reference{} Fruchter, A. \etal\ 2006, Nature, 441, 463
2372:
2373: \reference{} Fugazza, D. \etal\ 2005, GCN Circ. 3948
2374:
2375: \reference{} Fynbo, J. P. U. \etal\ 2001, A\&A, 373, 796
2376:
2377: \reference{} Fynbo, J. P. U. \etal\ 2002, A\&A, 388, 425
2378:
2379: \reference{} Fynbo, J.~P.~U. \etal\ 2003a, A\&A, 406, L63
2380:
2381: \reference{} Fynbo, J.~P.~U. \etal\ 2003b, GCN Circ. 2185
2382:
2383: \reference{} Fynbo, J.~P.~U. \etal\ 2005, ApJ, 633, 317
2384:
2385: %\reference{} Fynbo, J. P. U. \etal\ 2005b, GCN Circ. 3176
2386:
2387: \reference{} Fynbo, J. P. U. \etal\ 2006a, A\&A, 451, L47
2388:
2389: \reference{} Fynbo, J. P. U. \etal\ 2006b, GCN Circ. 4683
2390:
2391: \reference{} Fynbo, J. P. U., Prochaska, J. X., Sommer-Larsen, J., Dessauges-Zavadsky, M., \& M{\"o}ller, P. 2008, ApJ in press (arXiv:0801.3273)
2392:
2393: \reference{} Gehrels, N. \etal\ 2004, ApJ, 611, 1005
2394:
2395: \reference{} Gibson, B. K., Giroux, M. L., Penton, S. V., Putman, M. E., Stocke, J. T., \& Shull, J. M. 2000, AJ, 120, 1830
2396:
2397: \reference{} Gilmore, A., Kilmartin, P., \& Henden, A. 2003a, GCN Circ. 1949
2398:
2399: \reference{} Gilmore, A., Kilmartin, P., \& Henden, A. 2003b, GCN Circ. 2184
2400:
2401: \reference{} Gladders, M., Holland, S., Garnavich, P. M., Jha, S., Stanek, K. Z., Bersier, D., \& Barrientos, L. F., 2001, GCN Circ. 1209
2402:
2403: \reference{} Goad, M. \etal\ 2005, GCN Circ. 3942
2404:
2405: \reference{} Gondhalekar, P. M., Phillips, A. P., \& Wilson, R. 1980, A\&A, 85, 272
2406:
2407: \reference{} Gordon, K. D., \etal\ 2003, ApJ, 594, 279
2408:
2409: \reference{} Grav, T. \etal\ 2001, GCN Circ. 1191
2410:
2411: \reference{} Graziani, C. \etal\ 2003, GCN Circ. 1956
2412:
2413: \reference{} Hao, H. \etal\ 2007, ApJ, 659, L99
2414:
2415: \reference{} Harrison, F.~A. \etal\ 2001, ApJ, 559, 123
2416:
2417: \reference{} Hirschi, R., Meynet, G., \& Maeder, A. 2005, A\&A, 443, 581
2418:
2419: \reference{} Hjorth, J. \etal\ 2003, ApJ, 597, 699
2420:
2421: \reference{} Holland, S.~T. \etal\ 2005, GCN Circ. 3704
2422:
2423: \reference{} Jakobsson, P. \etal\ 2003, A\&A, 408, 941
2424:
2425: \reference{} Jakobsson, P. \etal\ 2004, A\&A, 427, 785
2426:
2427: \reference{} Jakobsson, P. \etal\ 2005, MNRAS, 362, 245
2428:
2429: %\reference{} Jakobsson, P. \etal\ 2005, GCN Circ. 4017
2430:
2431: \reference{} Jakobsson, P. \etal\ 2006, A\&A, 460, L13
2432:
2433: \reference{} Jakobsson, P., Fynbo, J.P.U., Vreeswijk, P. M., \& de Ugarte Postigo, A. 2008, GCN Circ. 8077
2434:
2435: \reference{} Jensen, B. L. \etal\ 2001, A\&A, 370, 909
2436:
2437: \reference{} Kann, D. A., Masetti, N., \& Klose, S. 2007, AJ, 133, 1187
2438:
2439: \reference{} Kennicutt, R. C., Jr., Bresolin, F., \& Garnett, D. R. 2003, ApJ, 591, 801
2440:
2441: \reference{} Kewley, L. J. \& Dopita, M. A. 2002, ApJS, 142, 35
2442:
2443: \reference{} Kr\"uhler, T., K\"upc\"u Yolda\c{s}, A., Greiner, J., Clemens, C., McBreen, S., Primak,N., Savaglio, S., Yolda\c{s}, A., \& Szokoly, G. P. 2008, ApJ in press (arXiv:0805.2824)
2444:
2445: \reference{} Law, D. R., Steidel, C. C., Erb, D. K., Pettini, M., Reddy, N. A., Shapley, A. E., Adelberger, K. L., \& Simenc, D. J. 2007, ApJ, 656, L1
2446:
2447: \reference{} Lecavelier des Etangs, A., D\'esert, J.-M., Kunth, D., Vidal-Madjar, A., Callejo, G., Ferlet, R., H\'ebrard, G., \& Lebouteiller, V. 2004, A\&A, 413, 131
2448:
2449: \reference{} Ledoux, C. \etal\ 2005, GCN Circ. 3860
2450:
2451: \reference{} Ledoux, C. \etal\ 2006, GCN Circ. 5237
2452:
2453: \reference{} Le Floc'h, E., et al.\ 2002, A\&A, 400, 499
2454:
2455: \reference{} Le Floc'h, E., Charmandaris, V., Forrest, W. J., Mirabel, I. F., Armus, L., \& Devost, D. 2006, ApJ, 642, L636
2456:
2457: \reference{} Maiolino, R. \etal\ 2008, A\&A in press (arXiv:0806.2410)
2458:
2459: \reference{} Martini, P., Persson, S.E. Murphy, D.C., Birk, C., Shectman, S.A., Gunnels, S.M., \& Koch, E. 2004, Proc. SPIE, 5492, 1653
2460:
2461: \reference{} McNaught, R. \& Price, P, A. 2005, GCN Circ. 3163
2462:
2463: \reference{} Malesani, D. \etal\ 2007, GCN Circ. 6713
2464:
2465: \reference{} Modjaz, M., Kewley, L., Kirshner, R. P., Stanek, K. Z., Challis, P., Garnavich, P. M., Greene, J. E., Kelly, P. L., \& Prieto, J. L. 2008, AJ, 135, 1136
2466:
2467: \reference{} M{\o}ller, P., Warren, S.J., Fall, S.M., Fynbo, J.U., \&
2468: Jakobsen, P. 2002a, ApJ, 574, 51
2469:
2470: \reference{} M{\o}ller, P. \etal\ 2002b, A\&A, 396, L21
2471:
2472: \reference{} Morris, D. \etal\ 2006, GCN Circ. 4682
2473:
2474: \reference{} Norris, J. \etal\ 2005, GCN Circ. 4013
2475:
2476: \reference{} Oke, J. B., et al.\ 1995, PASP, 107, 375
2477:
2478: \reference{} O'Meara, J. M., Chen, H.-W., \& Kaplan, D. L. 2006, ApJ, 642, L9
2479:
2480: \reference{} Paczy\'nski, B. 1998, ApJ, 494, 45
2481:
2482: \reference{} Page, M. \etal\ 2005a, GCN Circ. 3830
2483:
2484: \reference{} Page, M. \etal\ 2005b, GCN Circ. 3837
2485:
2486: \reference{} Persson, S.E., Murphy, D.C., Krzeminski, W., Roth, M., \& Rieke,
2487: M.J. 1998, AJ, 116, 2475
2488:
2489: \reference{} Pettini, M. 2006 in The Fabulous Destiny of Galaxies: Bridging Past and Present, V. LeBrun, A. Mazure, S. Arnouts \& D. Burgarella eds., (Paris: Frontier Group), 319 (astro-ph/0603066)
2490:
2491: \reference{} Piranomonte, S. \etal\ 2008, A\&A submitted
2492:
2493: \reference{} Pollack, L. K., Chen, H.-W., Prochaska, J. X., \& Bloom, J. S. 2008, ApJ submitted
2494:
2495: \reference{} Porciani, C., Viel, M., \& Lilly, S. J. 2007, ApJ, 659, 218
2496:
2497: \reference{} Pontzen, A., Hewett, P., Carswell, R., \& Wild, V.\ 2007, MNRAS, 381, L99
2498:
2499: \reference{} Price, P. A. \etal\ 2002, GCN Circ. 1221
2500:
2501: \reference{} Prochaska, J.~X. \etal\ 2005, GCN Circ. 3971
2502:
2503: \reference{} Prochaska, J. X., Chen, H.-W., \& Bloom, J. S. 2006, ApJ, 648, 95
2504:
2505: \reference{} Prochaska, J. X. 2006, ApJ, 650, 272
2506:
2507: \reference{} Prochaska, J. X. \etal\ 2006, GCN Circ. 4701
2508:
2509: \reference{} Prochaska, J. X., Chen, H. -W., Dessauges-Zavadsky, M., \& Bloom, J. S. 2007a, ApJ, 666, 267
2510:
2511: \reference{} Prochaska, J. X., Chen, H. -W., \& Bloom, J. S. \etal\ 2007b, ApJ, 168, 231
2512:
2513: \reference{} Prochter, G. E., Prochaska, J. X., Chen, H.-W., Bloom, J. S., Dessauges-Zavadsky, M., Foley, R. J., Lopez, S., Pettini, M., Dupree, A. K., \& Guhathakurta, P. 2006, ApJ, 648, L93
2514:
2515: \reference{} Putman, M. E., Staveley-Smith, L., Freeman, K. C., Gibson, B. K., \& Barnes, D. G. 2003, ApJ, 586, 170
2516:
2517: \reference{} Reddy, N. A., Steidel, C. C., Pettini, M., Adelberger, K. L., Shapley, A. E., Erb, D. K., \& Dickinson, M. 2008, ApJS, 175, 48
2518:
2519: \reference{} Ricker, G. R., et al.\ 2002, GCN Circ. 1220
2520:
2521: \reference{} Robertson, B. E. \& Kravtsov, A. V. 2008, ApJ, 680, 1083
2522:
2523: \reference{} Russell, S. C. \& Dopita, M. A. 1992, ApJ, 384, 508
2524:
2525: \reference{} Rykoff, E. S., Yost, S. A., \& Rujopakarn, W. 2005, GCN Circ. 4011
2526:
2527: \reference{} Savaglio, S. 2006, New Journal of Physics, 8, 195
2528:
2529: \reference{} Savaglio, S., Glazebrook, K., \& Le Borgne, D. 2008, ApJ submitted (arXiv:0803.2718)
2530:
2531: \reference{} Schady, P. \etal\ 2007, GCN Circ. 6641
2532:
2533: \reference{} Schaefer, B. \etal\ 2003, ApJ, 588, 387
2534:
2535: \reference{} Schlegel, D. J., Finkbeiner, D. P., \& Davis, Marc 1998, ApJ, 500, 525
2536:
2537: \reference{} Schulte-Ladbeck, R. E. \etal\ 2005, ApJ, 625, L79
2538:
2539: \reference{} Shapley, A. E., Erb, D. K., Pettini, M., Steidel, C. C., \& Adelberger, K. L. 2004, ApJ, 612, 108
2540:
2541: \reference{} Shapley, A. E., Steidel, C. C., Erb, D. K., Reddy, N. A., Adelberver, K. L., Pettini, M., Barmby, P., \& Huang, J. 2005, ApJ, 626, 698
2542:
2543: \reference{} Skillman, E. D., C{\^o}t\'e, S., \& Miller, B. W. 2003, AJ, 125, 610
2544:
2545: \reference{} Smith, J. A. \etal\ 2002, AJ, 123, 2121
2546:
2547: \reference{} Sollerman, J., \"Ostlin, G., Fynbo, J.P.U., Hjorth, J., Fruchter, A., \& Pedersen, K. 2005, New Astronomy, 11, 103
2548:
2549: \reference{} Sommer-Larsen, J. \& Fynbo, J. P. U. 2008, MNRAS, 385, 3
2550:
2551: \reference{} Stanek, K. Z. \etal\ 2003, ApJ, 591, L17
2552:
2553: \reference{} Starling, R. L. C., \etal\ 2005, A\&A, 442, L21
2554:
2555: \reference{} Sudilovsky, V., Savaglio, S., Vreeswijk, P., Ledoux, C., Smette, A., \& Greiner, J. 2007, ApJ, 669, 741
2556:
2557: \reference{} Tanvir, N. R. \& Jakobsson, P. 2007, Phil. Trans. Royal Society A, 366, 1377
2558:
2559: \reference{} Tassis, K., Kravtsov, A. V., \& Gnedin, N. Y. 2008, ApJ, 672, 888
2560:
2561: \reference{} Tejos, N., Lopez, S., Prochaska, J. X., Chen, H.-W., \& Dessauges-Zavadsky, M. 2007, ApJ, 671, 622
2562:
2563: \reference{} Th\"one, C. C. \etal\ 2008, A\&A in press (arXiv:0708.3448)
2564:
2565: %\reference{} Th\"one, C. C. \etal\ 2008b, A\&A submitted (arXiv:0806.1182)
2566:
2567: \reference{} Torii, K. 2005, GCN Circ. 3943
2568:
2569: \reference{} Tremonti, C. A. \etal\ 2004, ApJ, 613, 898
2570:
2571: \reference{} Tumlinson, J., Prochaska, J. X., Chen, H.-W., Dessauges-Zavadsky, M., \& Bloom, J. S. 2007, ApJ, 668, 667
2572:
2573: \reference{} van Zee, L., Salzer, J. J., Haynes, M. P., O'Donoghue, A. A., \&
2574: Balonek, T. J. 1998, AJ, 116, 2805
2575:
2576: \reference{} Vestrand, W.~T. \etal\ 2006, Nature, 442, 172
2577:
2578: \reference{} Vogt, S. S., et al. 1994, Proc. SPIE, 2198, 362
2579:
2580: \reference{} Vreeswijk, P. \etal\ 2004, A\&A, 419, 927
2581:
2582: \reference{} Vreeswijk, P. \etal\ 2006, A\&A, 447, 145
2583:
2584: \reference{} Vreeswijk, P. M. \etal\ 2007, A\&A, 468, 83
2585:
2586: \reference{} Watson, D. \etal\ 2006, ApJ, 652, 1011
2587:
2588: \reference{} Welty, D. E., Lauroesch, J. T., Blades, J. C., Hobbs, L. M., \& York, D. G. 1997, ApJ, 489, 672
2589:
2590: \reference{} Welty, D. E., Frisch, P. C., Sonneborn, G., \& York, D. G. 1999, ApJ, 512, 636
2591:
2592: \reference{} Whalen, D., Prochaska, J. X., Heger, A., \& Tumlinson, J. 2008, ApJ in press (arXiv:0802. 0737)
2593:
2594: \reference{} Wijers, R. A. M. J., Bloom, J. S., Bagla, J. S., \& Natarajan, P. 1998, MNRAS, 294, L13
2595:
2596: \reference{} Wolf, C. \& Podsiadlowski, P. 2007, MNRAS, 375, 1049
2597:
2598: \reference{} Wolfe, A. M., Gawiser, E., \& Prochaska, J. X. 2006, ARA\&A, 43, 861
2599:
2600: \reference{} Woosley, S.~E. 1993, ApJ, 405, 273
2601:
2602: \reference{} Woosley, S. E. \& Heger, A. 2006, ApJ, 637, 914
2603:
2604: \reference{} Woosley, S.~E. \& Bloom, J.~S. 2006, ARA\&A, 44, 507
2605:
2606: \reference{} Yan, H. \etal\ 2004, ApJ, 616, 63
2607:
2608: \reference{} Yoon, S.-C. \& Langer, N. 2005, A\&A, 443, 643
2609:
2610: \reference{} Zaritsky, D., Kennicutt, R.C. Jr., \& Huchra, J.P. 1994, ApJ,
2611: 420, 87
2612:
2613: \reference{} Ziaeepour, H. Z. \etal\ 2006, GCN Circ. 5233
2614:
2615: \reference{} Ziaeepour, H. Z. \etal\ 2007, GCN Circ. 6640
2616:
2617: \end{references}
2618:
2619: \newpage
2620:
2621: \begin{landscape}
2622: \begin{deluxetable}{lcccccrcccccc}
2623: \tabletypesize{\small}
2624: %\begin{deluxetable}{p{1in}cccccrcccccc}
2625: %\tabletypesize{\scriptsize}
2626: %\rotate
2627: \tablewidth{0pt}
2628: \tablecaption{Summary of known ISM and Stellar Properties of GRB Host Galaxies at $z>2$}
2629: \tablehead{ \colhead{} & \colhead{} & \multicolumn{6}{c}{Absorption Properties} & \colhead{} & \multicolumn{4}{c}{Galaxy Properties} \\
2630: \cline{3-8}
2631: \cline{10-13} \\
2632: \colhead{} & \colhead{} &\colhead{} & \colhead{} & \colhead{$W(1526)^a$} & \colhead{} & \colhead{} & \colhead{} & \colhead{} & \colhead{$M_{AB}(UV)$} & \colhead{$I_0^e$} & \colhead{$M_{AB}(B)^f$} & \colhead{$M_*$} \\
2633: \colhead{Field} & \colhead{} &\colhead{$z_{\rm GRB}$} & \colhead{$\log\,N(\hI)$} & \colhead{(\AA)} & \colhead{$\left [\frac{\rm M}{\rm H}\right ]_{\rm ISM}^b$} & \colhead{$\left [\frac{\rm M}{\rm Fe}\right ]_{\rm ISM}^c$} & \colhead{$A_V^d$} & \colhead{} & \colhead{$-5\,\log\,h$} & (photons/cm$^2$/s/Hz) & \colhead{$-5\,\log\,h$} & \colhead{($\times 10^9\ h^{-2}$ M$_\odot$)}}
2634: \startdata
2635: GRB\,000301C\dotfill & & 2.040 & $21.2 \pm 0.5$ & ... & ... & ... & ... & & $-15.2\pm 0.5$ & ... & ... & ... \nl
2636: GRB\,000926 \dotfill & & 2.038 & $21.3 \pm 0.3$ & $2.40\pm 0.14$ & $-0.17\pm 0.15$ & $+1.32\pm 0.15$ & 0.15 & & $-19.63\pm 0.07$ & $3.3\times 10^{-6}$ & $-20.26_{-0.4}^{+0.7}$ & $2.1_{-1.9}^{+4.0}$ \nl
2637: GRB\,011211 \dotfill & & 2.140 & $20.4 \pm 0.2$ & $1.35\pm 0.19$ & $> -1.3$ & ... & ... & & $-19.2\pm 0.1$ & ... & $-19.1\pm 0.3$ & ... \nl
2638: GRB\,021004 \dotfill & & 2.329 & $19.5 \pm 0.5$ & ... & $< -1.0$ & ... & ... & & $-19.83\pm 0.07$ & ... & $-20.38\pm 0.15$ & $2.1_{-1.9}^{+4.0}$ \nl
2639: GRB\,020124 \dotfill & & 3.198 & $21.7 \pm 0.2$ & ... & ... & ... & ... & & $>-15.1$ & ... & ... & ... \nl
2640: GRB\,030323 \dotfill & & 3.372 & $21.90 \pm 0.07$ & $0.75\pm 0.03$ & $> -1.26$ & ... & ... & & $-17.6\pm 0.1$ & $8.8\times 10^{-7}$ & 18.8 & ... \nl
2641: GRB\,030429 \dotfill & & 2.658 & $21.6 \pm 0.2$ & ... & ... & ... & ... & & ... & ... & $>-20.1$ & \nl
2642: GRB\,050401 \dotfill & & 2.899 & $22.6 \pm 0.3$ & $2.31\pm 0.26$ & $> -1.57$ & ... & ... & & $>-17.4$ & ... & $>-19.6$ & ... \nl
2643: GRB\,050730 \dotfill & & 3.968 & $22.15 \pm 0.10$ & $0.37\pm 0.05$ & $-2.26\pm 0.14$ & $+0.24\pm 0.11$ & 0.00 & & $>-18.6$ & ... & ... & ... \nl
2644: GRB\,050820A \dotfill & & 2.615 & $21.0 \pm 0.1 $ & $1.65\pm 0.05$ & $-0.63\pm 0.11$ & $+0.97\pm 0.15$ & 0.08 & & $-18.50\pm 0.06$ & $5.6\times 10^{-6}$ & $-19.3\pm 0.3$ & $0.9_{-0.4}^{+2.5}$ \nl
2645: GRB\,050908 \dotfill & & 3.343 & $17.55 \pm 0.10$ & ... & ... & ... & ... & & $>-18.9$ & ... & ... & ... \nl
2646: GRB\,050922C \dotfill & & 2.199 & $21.5 \pm 0.1$ & $0.52\pm 0.05$ & $-2.03\pm 0.15$ & $+0.60\pm 0.10$ & 0.01 & & $>-17.1$ & ... & $>-18.2$ & ... \nl
2647: GRB\,060206 \dotfill & & 4.048 & $20.85 \pm 0.10$ & $1.15\pm 0.05$ & $-0.85\pm 0.15$ & ... & ... & & $-17.7\pm 0.1$ & $9.1\times 10^{-6}$ & ... & ... \nl
2648: GRB\,060607 \dotfill & & 3.075 & $16.85 \pm 0.10$ & ... & ... & ... & ... & & $>-18.3$ & ... & $>-18.3$ & ... \nl
2649: GRB\,070721B \dotfill & & 3.626 & $21.50\pm 0.20$ & ... & ... & ... & ... & & $>-19.3$ & ... & ... & ... \nl
2650: \enddata
2651:
2652: \tablenotetext{a}{Rest-frame absorption equivalent width of Si\,II
2653: $\lambda1526$ in the host ISM. The measurements of all but the one
2654: for GRB\,060206 are adopted from Prochaska \etal\ (2008). The
2655: measurement for GRB\,060206 is determined using public spectra in
2656: Subaru Science Data Archive (Aoki \etal\ 2006). }
2657: \tablenotetext{b}{All measurements but the one for GRB\,000926 are
2658: based on the observed S\,II absorption strength. The measurement for
2659: GRB\,000926 is based on the observed Si\,II absorption strength,
2660: because S\,II transitions are not covered in available moderate
2661: resolution spectra of the afterglow. All lower limits are based on
2662: the observed Zn abundance in low-to-moderate resolution afterglow
2663: spectra. The upper limit for GRB\,021004 marks the maximum Si
2664: abundance prior to an ionization fraction correction.}
2665: \tablenotetext{c}{The relative abundances are measured based on sulfur
2666: wherever sulfur is available. For GRB\,00096, the reported value is
2667: based on silicon. } \tablenotetext{d}{The values are derived,
2668: assuming the SMC extinction law (Gordon \etal\ 2003).}
2669: \tablenotetext{e}{ISM far UV radiation field estimated from resolved
2670: host galaxy images in HST data.} \tablenotetext{f}{The rest-frame
2671: $B$-band absolute magnitude is derived based on available $H$-band
2672: photometry in the observed frame. For sources at $z>3$ with detected
2673: emission, we infer $M_{AB}(B)$ from the measured $M_{AB}(UV)$ and the
2674: mean color of starburst galaxies at $z=2-3$ from Shapeley \etal\
2675: (2005).}
2676: \end{deluxetable}
2677: \clearpage
2678: \end{landscape}
2679:
2680: \begin{deluxetable*}{p{1in}ccccrccc}
2681: \tabletypesize{\scriptsize}
2682: \tablewidth{0pt}
2683: \tablecaption{Summary of Candidate Galaxies Associated with Foreground $W(2796)>1$ \AA\ Mg\,II Absorbers}
2684: \tablehead{ \colhead{} & \colhead{} & \colhead{} & \multicolumn{2}{c}{Mg\,II Properties} & \colhead{} & \multicolumn{3}{c}{Galaxy Properties} \\
2685: \cline{4-5}
2686: \cline{7-9} \\
2687: \colhead{Field} & \colhead{} &\colhead{$z_{\rm GRB}$} & \colhead{$z_{\rm Mg\,II}$} & \colhead{$W(2796)$\tablenotemark{a}} & \colhead{} & \colhead{$AB$} & \colhead{$\Delta\,\theta (\arcsec)$} & \colhead{Filter}}
2688: \startdata
2689: GRB\,021004 \dotfill & & 2.329 & 1.38 & 1.81 & & $> 24.4$ & $\apl 0.3$ & F606W \nl
2690: & & & 1.60 & 1.53 & & $> 24.4$ & $\apl 0.3$ & F606W \nl
2691: GRB\,030429 \dotfill & & 2.658 & 0.84 & 3.30 & & $20.57\pm 0.05$ & $\approx 1.3''$ & $H$ \nl
2692: GRB\,050820A \dotfill & & 2.615 & 0.69 & 2.99 & & $26.30\pm 0.05$ & $\approx 0.4''$ & F625W \nl
2693: & & & 1.43 & 1.89 & & $26.20\pm 0.05$ & $\approx 1.3''$ & F625W \nl
2694: GRB\,060206 \dotfill & & 3.548 & 2.26 & 1.60 & & $26.22\pm 0.55$ & $\approx 1.0''$ & F814W \nl
2695: GRB\,070721B \dotfill & & 3.626 & 3.09 & ... & & $23.7\pm 0.1$ & $\approx 0.9''$ & $H$ \nl
2696: \enddata
2697:
2698: \tablenotetext{a}{Measurements of Mg\,II absorbers along GRB\,021004,
2699: GRB\,030429, and GRB\,050820A are adopted from Prochter \etal\ (2006).
2700: The Mg\,II absorber toward GRB\,060206 is identified based on our own
2701: analysis of available FOCAS spectra. The strong absorber toward
2702: GRB\,070721B is identified based on a DLA feature by Fynbo and
2703: collaborators (private communication).}
2704:
2705: \end{deluxetable*}
2706:
2707:
2708: \end{document}
2709: