1: % LaTeX file
2: %28-April-2004
3:
4: %\documentstyle[12pt,/home/fiore/testi/macros/aaspp4,epsfig]{article}
5: \documentstyle[12pt,aaspp4,epsfig]{article}
6:
7: \def\ls{{_<\atop^{\sim}}}
8: \def\gs{{_>\atop^{\sim}}}
9: \def\cgs{{ erg cm$^{-2}$ s$^{-1}$}}
10:
11: %%%%%%%%%%%%%%%%
12: \begin{document}
13: %%%%%%%%%%%%%%%%
14:
15: \title{A flash in the dark: UVES/VLT high resolution spectroscopy
16: of GRB afterglows\footnote{Based on observations collected at the European
17: Southern Observatory, ESO, the VLT/Kueyen telescope, Paranal, Chile, in the
18: framework of programs 69.A--0516(B) and 70.A--0599(B)}}
19:
20: \author{F. Fiore$^1$, V. D'Elia$^1$, D. Lazzati$^{2,3}$, R. Perna$^{3,4}$,
21: L. Sbordone$^{1,5}$, G. Stratta$^{1,6}$, E.J.A. Meurs$^7$, P. Ward$^7$,
22: L.A. Antonelli$^1$,
23: G. Chincarini$^8$, S. Covino$^9$, A. Di Paola$^1$, A. Fontana$^1$,
24: G. Ghisellini$^9$, G. Israel$^1$, F. Frontera$^{10}$, G. Marconi$^{1,5}$,
25: L. Stella$^1$, M. Vietri$^{11}$, F. Zerbi$^9$\\
26: $^1$INAF--Osservatorio Astronomico di Roma, Via Frascati 33, I--00044
27: Monteporzio Catone, Italy;\\
28: $^2$ Institute of Astronomy, University of Cambridge, Madingley Road,
29: Cambridge, UK;\\
30: $^3$ Department of Astrophysical and Planetary Science, CU Boulder, Boulder,
31: 80309, USA;\\
32: $^4$ Princeton University Observatory, Princeton, NJ 08544--1001, USA;\\
33: $^5$ European Southern Observatory, Casilla 19001, Santiago, Chile;\\
34: $^6$ Dipartimento di Fisica, Universita' di Roma La Sapienza, Italy;\\
35: $^7$ Dunsink Observatory, Castleknock, Dublin 15, Ireland;\\
36: $^8$ Universita' di Milano Bicocca, Piazza della Scienza 3, 20126 Milano,
37: Italy\\
38: $^9$ INAF, Osservatorio Astronomico di Brera, via E. Bianchi 46, 23807
39: Merate (LC), Italy.\\
40: $^{10}$ Dipartimento di Fisica, Universita' di Ferrara, via Paradiso 12,
41: 44100 Ferrara, Italy;\\
42: $^{11}$ Scuola Normale Superiore, I--56100 Pisa, Italy.
43: }
44:
45: \author{\tt (version: 2 February 2005) }
46:
47:
48: \begin{abstract}
49: We present the first high resolution (R=20000--45000, corresponding to
50: 14 km/s at 4200\AA\ to 6.6 km/s at 9000\AA\ ) observations of the
51: optical afterglow of Gamma Ray Bursts. GRB 020813 and GRB 021004 were
52: observed by UVES@VLT 22.19 hours and 13.52 hours after the trigger,
53: respectively. These spectra show that the inter--stellar matter of the
54: GRB host galaxies is complex, with many components contributing to
55: each main absorption system, and spanning a total velocity range of up
56: to about 3000 km/s. Several narrow components are resolved down to a
57: width of a few tens of km/s. In the case of GRB 021004 we detected
58: both low and high ionization lines. Combined with photoionization
59: results obtained with CLOUDY, the ionization parameters of the various
60: systems are consistent with a remarkably narrow range with no clear
61: trend with system velocity. This can be interpreted as due to density
62: fluctuations on top of a regular $R^{-2}$ wind density profile.
63: \end{abstract}
64:
65: \keywords{gamma rays: bursts -- cosmology: observations -- galaxies:
66: abundances -- ISM}
67:
68: \section{Introduction}
69:
70: For a few hours after their onset, Gamma Ray Bursts (GRBs) are the
71: brightest beacons in the far Universe, offering a superb opportunity
72: to investigate both GRB physics and high redshift galaxies. Tens of
73: minutes after a GRB, its optical afterglow can be as bright as
74: magnitude 13--15; a few hours later a magnitude of 16--19 is often
75: achieved. Bright examples are the cases of GRB 990123 (z=1.600),
76: for which the reverse shock reached R=9-10 mag 1 minute after the GRB
77: and the optical afterglow reached R=14 mag 12 minutes after the GRB
78: (e.g. Akerlof et al. 1999, Galama et al. 1999), and GRB 030329 (z=0.1685),
79: for which the optical afterglow reached R=12.7 mag at 1.5
80: hours from the GRB and decreased down to R=19 mag after $\sim 10$ days
81: (e.g. Price et al. 2003; Stanek et al. 2003). High resolution (a few
82: tens of km/s in the optical band), high quality (signal to noise $>10$
83: per resolution element) spectra can therefore be gathered, provided
84: that the afterglow is observed rapidly by 8m class telescopes.
85: Since GRBs are associated with the collapse of massive stars (see
86: Woosley 1993, Paczynski 1998, MacFadyen \& Woosley 1999, Vietri \&
87: Stella 1998 for theoretical reasons and Galama et al. 1998,
88: Stanek et al. 2003, Hjorth et al. 2003 and Della Valle et al. 2003 for
89: observational reasons), this is expected to open a new window in
90: the study of the environment in which intermediate to high redshift
91: star--formation occurs and, in particular, on the physical, chemical
92: and dynamical state of the inter--stellar matter (ISM) of the GRB host
93: galaxies.
94:
95: The study of z$\gs1$ galaxies has so far mostly relied upon
96: Lyman--break galaxies (LBGs) at z$=3--4$ (see e.g. Steidel et al.
97: 1999) and galaxies which happen to be along the line of sight to
98: bright background QSOs. Some of these systems are associated with
99: Damped Ly$\alpha$ systems (DLA, see e.g. Pettini et al. 1997).
100: However, LBGs are characterized by pronounced star--formation and
101: their inferred chemical abundances may be related to these regions
102: rather than being representative of typical high z galaxies.
103: Metal line systems associated to DLAs along the line of sight to
104: quasars probe mainly galaxy haloes, rather than their bulges or discs.
105: Furthermore, it is not clear whether galaxies associated with DLAs are
106: truly representative of the whole high--z galaxy population. GRB
107: afterglows provide an independent tool to study the ISM of high z
108: galaxies. Savaglio, Fall \& Fiore (2003) studied the metal abundances
109: in three GRB host galaxies using low--medium resolution spectroscopy
110: and a curve of growth analysis. They found metal column densities
111: higher that in QSO--DLAs and a strong inverse correlation between
112: [Si/Zn], [Cr/Zn] and [Fe/Zn] and the Zn column density, indicating a
113: dense environment and large dust depletion. On the other hand,
114: Vreeswijk et al. (2004) found indications for a relatively low
115: metallicity and low dust content in the ISM of the z=3.372 host galaxy
116: of GRB 030323, using FORS2 low and intermediate resolution
117: spectroscopy. Using again a curve of growth analysis, and taking
118: advantage of ultra--deep {\it Gemini} multi--object spectrograph
119: observations, Savaglio et al. (2004) studied the ISM of a sample of
120: faint K band selected galaxies at 1.4$<$z$<$2.0, finding \ion{Mg}{2}
121: and \ion{Fe}{2} abundances much higher than in QSO--DLAs and similar
122: to those in GRB host galaxies.
123: %While
124: %similar procedures were applied successfully in the past to QSO--DLAs
125: %(e.g. Turnshek et al. 1989), they are not free from possible systematic
126: %errors.
127: A much better job can be done with high resolution ($R>20000$)
128: observations because: (i) lines can be separated; (ii) metal
129: column densities can be measured through a fit to the line profile;
130: (iii) fainter lines can be measured; (iv) information on the gas
131: dynamics in the GRB host galaxies can be derived. Furthermore, as
132: suggested by the comparison of the Vreeswijk et al. (2004) and the
133: Savaglio et al. (2004) studies, GRB afterglows allow us to probe
134: galaxy ISM at much higher redshifts than even ultra--deep, standard
135: galaxy spectroscopy, such as the Gemini Deep Deep Survey.
136:
137: For all these reasons we started a pilot program to observe bright GRB
138: afterglows of promptly localized GRBs with UVES@VLT.
139: The program has been conceived and designed to make
140: full use of the
141: %
142: %The aim of this
143: %program, in addition to the scientific motivations briefly outlined
144: %above, is to set up procedures which can help to maximize the
145: %scientific return of high resolution follow--up observations of
146: GRB afterglows discovered by the GRB--dedicated {\it Swift} mission,
147: launched on November 20 2004\footnote{http://swift.gsfc.nasa.gov}.
148: {\it Swift} will provide GRB positions ($<4\arcmin$ precision) in
149: $<10$~s, X--ray afterglow positions ($3\arcsec$ precision) in
150: $<100$~s, and an optical finding chart in $<300$~s, with a few arcsec
151: to sub--arcsec position accuracy, thus revolutionizing fast response
152: multiwavelength studies of GRBs.
153: %$\approx30\%$ of the Swift localized GRB (i.e.
154: %a dozen per semester) should will be visible from Paranal, and occur
155: %during the Paranal night. We expect that roughly half of these will be
156: %bright enough to provide high quality, high resolution spectra, also
157: %thanks to the development of a dedicated Rapid Response observing Mode
158: %for the VLT telescopes.
159:
160:
161: This article concentrates on high resolution spectroscopy of two GRB
162: afterglows. The HETE--2 FREGATE, WXM and SXC instruments detected
163: GRB 021004 on 2002 Oct. 4 12:06:13.57 UT (Shirasaki et al. 2002) and
164: GRB 020813 on 2002 Aug. 13 02:44:19.17 UT (Villasenor et al. 2002).
165: The WXM flight localization software produced a reliable and
166: relatively accurate position 49 seconds after the burst for GRB 021004
167: (error box of 30$\arcmin$ radius) and 4 minutes after the burst for
168: GRB 020813 (error box of 14$\arcmin$ radius). Bright optical
169: counterparts were identified 10 minutes and 1.9 hours after the
170: triggers by Fox et al. (2002, 2003), and Fox, Blake \& Price (2002)
171: respectively. Ground analysis of the HETE2 data performed a few hours
172: after each GRB improved the localization of the event, providing error
173: boxes of 2$\arcmin$ and 1$\arcmin$ radius, respectively. UVES
174: observations started 13.52 hours after the GRB 021004 trigger,
175: (Savaglio et al. 2002) and 22.19 hours after the GRB 020813 trigger
176: (Fiore et al. 2002).
177:
178: Optical low to intermediate resolution spectroscopy of the
179: afterglow of GRB 021004 was obtained by Moller et al. (2003), Mirabal et
180: al. (2003), Matheson et al. (2003), Schaefer et al. (2003). One of the main
181: conclusion of these works is that density fluctuations on top of a
182: regular wind density profile are able to explain both the presence of
183: strong, blue--shifted, high ionization absorption line systems and
184: irregularities in the optical light curve. Schaefer et al. (2003)
185: propose a scenario where a clumpy wind originates from a massive star
186: progenitor (a Wolf--Rayet star), while Mirabal et al. (2003) suggest
187: that the ionized absorption takes place in a fragmented shell nebula
188: around the GRB progenitor. Optical spectropolarimetry of the
189: afterglow of GRB 020813 has been presented by Barth et al. (2003) who
190: report the detection of strong absorption systems at z=1.223 and
191: z=1.255. A further analysis of the Keck LRIS spectrum
192: is presented by Savaglio et al. (2004).
193: The present paper reports the results of higher resolution
194: spectroscopy of both afterglows.
195:
196: Both GRBs were observed in the X--rays too by the Chandra High Energy
197: Transmission Grating System, which has a resolution of $\sim 1000$ at
198: 1 keV (Butler et al. 2003 and references therein).
199: Both observations lasted about 80 ks and started 21 hours and
200: 20 hours after the GRB event for GRB 020813 and GRB 021004 respectively.
201: Fading X--ray afterglows were detected in both cases with decay indices
202: of $-1.38\pm0.06$ and $-0.9\pm0.1$, consistent with the values
203: reported for the optical afterglows by Covino et al. (2003) and
204: Holland et al. (2003). The time averaged fluxes were
205: $2.2\times10^{-12}$ \cgs and $6.3\times10^{-13}$\cgs respectively. The
206: X--ray spectra are consistent with power laws of energy index of
207: $-0.85\pm0.04$ and $-1.01\pm0.08$ reduced at low energy by Galactic
208: column density (which is rather high in the direction of GRB 020813,
209: $N_{HGal}=7.5\times10^{20}$ cm$^{-2}$, and is
210: $N_{HGal}=4.2\times10^{20}$ cm$^{-2}$ in the direction of GRB 021004).
211: The upper limits to the rest frame X--ray absorbing column are of the
212: order of a few$\times10^{20}$ in both cases. Butler et al. (2003)
213: report the detection of an emission line at 1.3 keV at the 3.3$\sigma$
214: confidence level, interpreted as a SXVI l$\alpha$ line blue--shifted by
215: 0.12c.
216:
217:
218: \section{Observations and data reduction}
219:
220: In the framework of ESO programs 69.A--0516 and 70.A--0599 we observed
221: the afterglows of GRB 020813 and GRB 021004 with the high resolution
222: UV--visual echelle spectrograph (UVES, Dekker et al. 2000), mounted at
223: the VLT--UT2 telescope. Table 1 gives the log of the observations
224: used in this paper. A further 1 hour dichroic 1 observation is
225: available in the ESO archive, but we were not able to extract a
226: reliable spectrum from this observation, which has therefore been
227: discarded. In order to maximize the signal to noise ratio the CCD was
228: rebinned $2\times2$ pixels. The data reduction has been performed
229: using the UVES
230: pipeline\footnote{http://www.eso.org/observing/dfo/quality/UVES/pipeline/}.
231: The final useful spectra extend from about 4250\AA\ to about 9400\AA\
232: and were rebinned to 0.1\AA\ to increase the signal to noise. The
233: resulting resolution element, set to two pixels, ranges then from 14
234: km/s at 4200\AA\ to 6.6 km/s at 9000\AA\ . The noise spectrum, used
235: to determine the errors on the best fit line parameters, has been
236: calculated from the real background subtracted and rebinned spectrum
237: using line free regions. This therefore takes into account both
238: statistical errors and systematic errors in the pipeline processing
239: and background subtraction. Table 2 gives the signal to noise ratio
240: per 0.1\AA\ pixel at different wavelengths for the 4 spectra in Table
241: 1. Figure 1 plots the full UVES spectrum of GRB 021004, for which both
242: UVES dichroics, and both red and blue arms, were used, allowing us to
243: obtain a particularly wide spectral coverage, extending from
244: $\sim$4250\AA\ to $\sim$9400\AA\ . A further dichroic 1, blue arm
245: image, covering the band 3400\AA\ -- 3900 \AA\ was obtained, but the
246: signal to noise is too low to allow a reliable extraction of the
247: spectrum in this band. Figure \ref{spe020813} plots the full UVES
248: spectrum of GRB 020813. This spectrum, as well as the spectrum in
249: figure 1, was smoothed with a gaussian function with $\sigma=1.5$
250: pixels. Tables 3 and 4 give the equivalent width of all lines detected
251: in the UVES spectra for the two GRB afterglows, along with their
252: identification. The equivalent widths are computed at the
253: redshift of the absorption systems. For the GRB host galaxy identified
254: systems and the 2 main intervening systems at z=1.60 and 1.38 we
255: report also faint lines with (signal to noise $>1$). Unidentified
256: lines are reported only if the signal to noise is $>3$. Their
257: equivalent widths are computed at zero redshift.
258:
259: \begin{table}
260: \caption{\bf Journal of observations}
261: \footnotesize
262: \begin{tabular}{lccccccc}
263: \tableline\tableline
264: Date UT & Dichroic & B. Arm C.W. & R. Arm C.W. & slit$^{a}$
265: & seeing$^{a}$ & exposure$^{b}$ & time since GRB$^{c}$\\
266: \hline
267: 14/08/02 00:55:25 & 1 & 3460\AA\ & 5800\AA\ & 1 & $\ls1$ & 60 & 22.19 \\
268: 05/10/02 01:37:37 & 2 & 4370\AA\ & 8600\AA\ & 1 & $\ls1$ & 30 & 13.52 \\
269: 05/10/02 02:10:53 & 1 & 3460\AA\ & 5800\AA\ & 1 & $\ls1$ & 30 & 14.08 \\
270: 05/10/02 04:09:54 & 1 & 3460\AA\ & 5800\AA\ & 1 & $\ls1$ & 60 & 16.06 \\
271: \tableline
272: \end{tabular}
273: \normalsize
274:
275: $^{a}$arcsec;
276: $^{b}$min;
277: $^{c}$hr
278: \end{table}
279:
280: \begin{table}
281: \caption{\bf Signal to noise ratio}
282: \begin{tabular}{lccc}
283: \tableline\tableline
284: $\lambda$ & 14/08/02 00hr & 05/10/02 01hr-02hr & 05/10/02 04hr \\
285: \hline
286: 4250\AA\ & & 3.6 & \\
287: 5000\AA\ & 4.7 & 5.9 & 9.3 \\
288: 6000\AA\ & 5.6 & 5.4 & 11.2 \\
289: 7000\AA\ & 5.9 & 9.8 & 10.2 \\
290: 7800\AA\ & & 6.8 & \\
291: 9250\AA\ & & 2.9 & \\
292: \tableline
293: \end{tabular}
294: %
295: %$^{a}$arcsec;
296: %$^{b}$min
297:
298: \end{table}
299:
300: \begin{table}
301: \caption{\bf GRB 021004 UVES line identifications}
302: \footnotesize
303: \begin{tabular}{lccc}
304: \hline
305: \hline
306: $\lambda$(\AA\ ) & $W_{rest}$(\AA\ )$^a$& ID & z \\
307: \hline
308: 4298.40 & 0.13$\pm$0.02 & ?\ion{Si}{2}$\lambda$1304.37? & ?2.2953? \\
309: 4301.73 & 0.06$\pm$0.03 & \ion{Si}{2}$\lambda$1808.01 & 1.3793-8 \\
310: 4304.53 & 0.09$\pm$0.03 & \ion{Si}{2}$\lambda$1808.01 & 1.3807 \\
311: 4311.20 & 0.16$\pm$0.08 & & \\
312: 4346.30 & 0.27$\pm$0.08 & \ion{Al}{2}$\lambda$1670.79 & 1.6014 \\
313: 4347.25 & 0.18$\pm$0.03 & \ion{Al}{2}$\lambda$1670.79 & 1.6019 \\
314: 4348.60 & 0.11$\pm$0.03 & \ion{Al}{2}$\lambda$1670.79 & 1.6028 \\
315: 4350.95 & 0.53$\pm$0.08 & & \\
316: 4358.28 & 0.35$\pm$0.08 & & \\
317: 4398.34 & 0.06$\pm$0.03 & \ion{C}{2}$\lambda$1334.53 & 2.2958 \\
318: 4401.68 & 0.24$\pm$0.02 & \ion{C}{2}$\lambda$1334.53 & 2.2981-4 + \\
319: & & ?\ion{C}{2}$^*\lambda$1335.70? & ?2.2958? \\
320: 4415.50 & 0.10$\pm$0.03 & \ion{Al}{3}$\lambda$1854.71 & 1.3807 \\
321: 4441.34 & 0.04$\pm$0.02 & \ion{C}{2}$\lambda$1334.53 & 2.328 \\
322: 4445.35 & 0.09$\pm$0.02 & \ion{C}{2}$^*\lambda$1335.70 & 2.328 \\
323: 4593.59 & 0.15$\pm$0.02 & \ion{Si}{4}$\lambda$1393.76 & 2.2958 \\
324: 4596.71 & 0.02$\pm$0.02 & \ion{Si}{4}$\lambda$1393.76 & 2.2981-4\\
325: 4623.48 & 0.21$\pm$0.02 & \ion{Si}{4}$\lambda$1402.77 & 2.2958 \\
326: 4626.41 & 0.02$\pm$0.02 & \ion{Si}{4}$\lambda$1402.77 & 2.2981-4\\
327: 4629.72 & 0.40$\pm$0.02 & \ion{Si}{4}$\lambda$1393.76 & 2.321 \\
328: 4631.35 & 0.30$\pm$0.07 & & \\
329: 4635.85 & 0.65$\pm$0.07 & & \\
330: 4637.38 & 3.63$\pm$0.07 & & \\
331: 4638.01 & 0.31$\pm$0.02 & \ion{Si}{4}$\lambda$1393.76 & 2.328 \\
332: 4643.19 & 0.25$\pm$0.07 & & \\
333: 4658.07 & 0.27$\pm$0.02 & \ion{Si}{4}$\lambda$1402.77 & 2.321 \\
334: 4666.28 & 1.07$\pm$0.07 & & \\
335: 4668.38 & 0.56$\pm$0.02 & \ion{Si}{4}$\lambda$1393.76 & 2.328 \\
336: 5102.54 & 0.50$\pm$0.02 & \ion{C}{4}$\lambda$1548.20 & 2.2958 \\
337: 5106.50 & 0.35$\pm$0.02 & \ion{C}{4}$\lambda$1548.20 & 2.2981-4\\
338: 5111.04 & 0.33$\pm$0.02 & \ion{C}{4}$\lambda$1550.77 & 2.2958 \\
339: 5114.82 & 0.16$\pm$0.02 & \ion{C}{4}$\lambda$1550.77 & 2.2981-4\\
340: 5141.11 & 1.74$\pm$0.02 & \ion{C}{4}$\lambda$1448.20 & 2.321 \\
341: 5150.4 & 2.10$\pm$0.02 & \ion{C}{4}$\lambda$1550.77 & 2.321 + \\
342: & & \ion{C}{4}$\lambda$1448.20 & 2.328 \\
343: 5159.66 & 1.57$\pm$0.02 & \ion{C}{4}$\lambda$1550.77 & 2.328 \\
344: 5510.62 & 0.05$\pm$0.02 & \ion{Al}{2}$\lambda$1670.79 & 2.2981-4\\
345: 5560.30 & 0.47$\pm$0.02 & \ion{Al}{2}$\lambda$1670.79 & 2.328 \\
346: 5577.71 & 1.47$\pm$0.03 & \ion{Fe}{2}$\lambda$2344.21 & 1.3793-8\\
347: 5580.81 & 0.07$\pm$0.03 & \ion{Fe}{2}$\lambda$2344.21 & 1.3807 \\
348: 5592.94 & 0.14$\pm$0.06 & & \\
349: 5649.59 & 0.31$\pm$0.03 & \ion{Fe}{2}$\lambda$2374.21 & 1.3793-8\\
350: 5669.68 & 0.67$\pm$0.03 & \ion{Fe}{2}$\lambda$2382.77 & 1.3793-8\\
351: 6064.56 & 0.60$\pm$0.06 & & \\
352: 6098.32 & 0.22$\pm$0.06 & & \\
353: \hline
354: \end{tabular}
355: \normalsize
356:
357: Errors are 67\% confidence intervals.
358: $^a$ the equivalent width of unidentified lines is computed at
359: zero redshift;
360: ? = uncertain identification. In many cases several components contribute to
361: a single line in one entry of the table. In these cases either the redshift
362: is given with only 4 digits or it is given as a range.
363: A + sign indicates that
364: a line is partly blended with the following entry.
365: \end{table}
366:
367:
368: \setcounter{table}{2}
369:
370:
371: \begin{table}
372: \caption{\bf GRB 021004 UVES line identifications, continued.}
373: \footnotesize
374: \begin{tabular}{lccc}
375: \hline
376: \hline
377: $\lambda$(\AA\ ) & $W_{rest}$(\AA\ )$^a$& ID & z \\
378: \hline
379: 6099.33 & 0.19$\pm$0.02 & \ion{Fe}{2}$\lambda$2344.21 & 1.6019 \\
380: 6101.32 & 0.09$\pm$0.02 & \ion{Fe}{2}$\lambda$2344.21 & 1.6028 \\
381: 6154.64 & 0.52$\pm$0.03 & \ion{Fe}{2}$\lambda$2586.65 & 1.3793-8\\
382: 6158.03 & 0.01$\pm$0.03 & \ion{Fe}{2}$\lambda$2586.65 & 1.3807 \\
383: 6160.76 & 0.17$\pm$0.06 & & \\
384: 6172.42 & 0.35$\pm$0.02 & \ion{Al}{3}$\lambda$1854.72 & 2.328 \\
385: 6176.74 & 0.21$\pm$0.06 & & \\
386: 6177.88 & 0.17$\pm$0.02 & \ion{Fe}{2}$\lambda$2374.46 & 1.6019 \\
387: 6180.04 & 0.14$\pm$0.05 & & \\
388: 6186.51 & 0.25$\pm$0.02 & \ion{Al}{3}$\lambda$1862.79 & 2.321 + \\
389: 6186.99 & 0.11$\pm$0.02 & \ion{Fe}{2}$\lambda$2600.17 & 1.3793-8\\
390: 6187.81 & 0.05$\pm$0.02 & \ion{Fe}{2}$\lambda$2600.17 & 1.3798 \\
391: 6190.23 & 0.12$\pm$0.02 & \ion{Fe}{2}$\lambda$2600.17 & 1.3807 \\
392: 6199.36 & 0.11$\pm$0.02 & \ion{Al}{3}$\lambda$1862.79 & 2.328 + \\
393: 6198.4 & & & \\
394: 6199.53 & 0.29$\pm$0.02 & \ion{Fe}{2}$\lambda$2382.77 & 1.6019 \\
395: 6201.66 & 0.28$\pm$0.02 & \ion{Fe}{2}$\lambda$2382.77 & 1.6028 \\
396: 6653.88 & 0.89$\pm$0.02 & \ion{Mg}{2}$\lambda$2796.35 & 1.3793-8\\
397: 6655.88 & 0.24$\pm$0.02 & \ion{Mg}{2}$\lambda$2796.35 & 1.3802 \\
398: 6657.60 & 0.47$\pm$0.02 & \ion{Mg}{2}$\lambda$2796.35 & 1.3807 \\
399: 6671.00 & 0.78$\pm$0.02 & \ion{Mg}{2}$\lambda$2803.53 & 1.3793-8\\
400: 6672.73 & 0.06$\pm$0.02 & \ion{Mg}{2}$\lambda$2803.53 & 1.3802 \\
401: 6674.77 & 0.32$\pm$0.02 & \ion{Mg}{2}$\lambda$2803.53 & 1.3807 \\
402: 6730.16 & 0.25$\pm$0.02 & \ion{Fe}{2}$\lambda$2586.65 & 1.6019 \\
403: 6732.55 & 0.09$\pm$0.02 & \ion{Fe}{2}$\lambda$2586.65 & 1.6028 \\
404: 6750.64 & 0.08$\pm$0.02 & \ion{Mn}{2}$\lambda$2594.50 & 1.6019 \\
405: 6764.18 & 0.43$\pm$0.04 & & \\
406: 6765.43 & 0.25$\pm$0.02 & \ion{Fe}{2}$\lambda$2600.17 & 1.6019 \\
407: 6767.79 & 0.17$\pm$0.02 & \ion{Fe}{2}$\lambda$2600.17 & 1.6028 \\
408: 6768.78 & 0.18$\pm$0.04 & & \\
409: 6781.81 & 0.04$\pm$0.02 & \ion{Mn}{2}$\lambda$2606.46 & 1.6019 \\
410: 6788.06 & 0.03$\pm$0.02 & \ion{Mg}{1}$\lambda$2852.94 & 1.3793-8\\
411: 7003.91 & 0.12$\pm$0.03 & & \\
412: 7041.91 & 0.12$\pm$0.03 & & \\
413: 7042.52 & 0.11$\pm$0.03 & & \\
414: 7043.25 & 0.10$\pm$0.03 & & \\
415: 7274.52 & 0.26$\pm$0.01 & \ion{Mg}{2}$\lambda$2796.35 & 1.6014 \\
416: 7275.83 & 0.33$\pm$0.01 & \ion{Mg}{2}$\lambda$2796.35 & 1.6019 \\
417: 7277.28 & 0.49$\pm$0.01 & \ion{Mg}{2}$\lambda$2796.35 & 1.6024 \\
418: 7278.39 & 0.24$\pm$0.01 & \ion{Mg}{2}$\lambda$2796.35 & 1.6028 \\
419: 7293.12 & 0.27$\pm$0.01 & \ion{Mg}{2}$\lambda$2803.35 & 1.6014 \\
420: 7294.58 & 0.36$\pm$0.01 & \ion{Mg}{2}$\lambda$2803.53 & 1.6019 \\
421: 7295.91 & 0.04$\pm$0.01 & \ion{Mg}{2}$\lambda$2803.53 & 1.6024 \\
422: 7296.85 & 0.23$\pm$0.01 & \ion{Mg}{2}$\lambda$2803.53 & 1.6028 \\
423: 7422.05 & 0.04$\pm$0.01 & \ion{Mg}{1}$\lambda$2852.94 & 1.6014 \\
424: 7423.01 & 0.23$\pm$0.01 & \ion{Mg}{1}$\lambda$2852.94 & 1.6019 \\
425: \hline
426: \end{tabular}
427: \normalsize
428:
429: $^a$ the equivalent width of unidentified lines is computed at
430: zero redshift.
431: ? = uncertain identification. In many cases several components contribute to
432: a single line in one entry of the table. In these cases either the redshift
433: is given with only 4 digits or it is given as a range.
434: A + sign indicates that
435: a line is partly blended with the following entry.
436: \end{table}
437:
438: \setcounter{table}{2}
439:
440:
441: \begin{table}[ht!]
442: \caption{\bf GRB 021004 UVES line identifications, continued.}
443: \footnotesize
444: \begin{tabular}{lccc}
445: \hline
446: \hline
447: $\lambda$(\AA\ ) & $W_{rest}$(\AA\ )$^a$& ID & z \\
448: \hline
449: 7929.74 & 0.75$\pm$0.05 & & \\
450: 7929.53 & 0.08$\pm$0.02 & \ion{Fe}{2}$\lambda$2382.77 & 2.328 \\
451: 7930.42 & 0.04$\pm$0.02 & \ion{Fe}{2}$\lambda$2382.77 & 2.328 \\
452: 7931.03 & 0.03$\pm$0.02 & \ion{Fe}{2}$\lambda$2382.77 & 2.328 \\
453: 8228.16 & 0.37$\pm$0.05 & & \\
454: 9178.45 & 0.75$\pm$0.08 & & \\
455: 9205.45 & 0.70$\pm$0.09 & & \\
456: 9216.41 & 0.04$\pm$0.03 & \ion{Mg}{2}$\lambda$2796.35 & 2.2958 \\
457: 9222.70 & 0.09$\pm$0.03 & \ion{Mg}{2}$\lambda$2796.35 & 2.2981 \\
458: 9223.62 & 0.18$\pm$0.03 & \ion{Mg}{2}$\lambda$2796.35 & 2.2984 \\
459: 9246.33 & 0.09$\pm$0.03 & \ion{Mg}{2}$\lambda$2803.53 & 2.2981 \\
460: 9247.39 & 0.12$\pm$0.03 & \ion{Mg}{2}$\lambda$2803.53 & 2.2984 \\
461: 9330.33 & 0.57$\pm$0.03 & \ion{Mg}{2}$\lambda$2803.53 & 2.328 \\
462: \hline
463: \end{tabular}
464: \normalsize
465:
466: $^a$ the equivalent width of unidentified lines is computed at
467: zero redshift.
468: \end{table}
469:
470: \begin{table}[ht!]
471: \caption{\bf GRB 020813 UVES line identifications}
472: \footnotesize
473: \begin{tabular}{lccc}
474: \hline
475: \hline
476: $\lambda$(\AA\ ) & $W_{rest}$(\AA\ )$^a$& ID & z \\
477: \hline
478: 4824.64 & 1.06$\pm$ 0.05 & & \\
479: 4824.64 & 0.59$\pm$ 0.05 & & \\
480: 4998.03 & 0.47$\pm$ 0.05 & & \\
481: 5212.23 & 0.18$\pm$ 0.03 & \ion{Fe}{2}$\lambda$2344.21 & 1.2234 \\
482: 5236.74 & 1.64$\pm$ 0.07 & & \\
483: 5281.33 & 0.09$\pm$ 0.03 & \ion{Fe}{2}$\lambda$2344.21 & 1.2529 \\
484: 5285.19 & 1.70$\pm$ 0.03 & \ion{Fe}{2}$\lambda$2344.21 & 1.255 \\
485: 5297.11 & 0.85$\pm$ 0.03 & \ion{Fe}{2}$\lambda$2382.77 & 1.2234 \\
486: 5353.61 & 1.31$\pm$ 0.03 & \ion{Fe}{2}$\lambda$2374.46 & 1.255 \\
487: 5368.21 & 0.28$\pm$ 0.03 & \ion{Fe}{2}$\lambda$2382.77 & 1.2529 \\
488: 5372.23 & 1.63$\pm$ 0.03 & \ion{Fe}{2}$\lambda$2382.77 & 1.255 \\
489: 5840.73 & 0.44$\pm$ 0.06 & & \\
490: 5850.86 & 0.73$\pm$ 0.06 & & \\
491: 5858.24 & 0.23$\pm$ 0.02 & \ion{Fe}{2}$\lambda$2600.17 & 1.2529 \\
492: 5862.33 & 1.82$\pm$ 0.02 & \ion{Fe}{2}$\lambda$2600.17 & 1.255 \\
493: 5896.25 & 0.40$\pm$ 0.06 & & \\
494: 6067.05 & 2.30$\pm$ 0.06 & & \\
495: 6216.93 & 1.34$\pm$ 0.03 & \ion{Mg}{2}$\lambda$2796.35 & 1.2234 \\
496: 6232.85 & 1.36$\pm$ 0.03 & \ion{Mg}{2}$\lambda$2803.53 & 1.2234 \\
497: 6300.13 & 0.55$\pm$ 0.03 & \ion{Mg}{2}$\lambda$2796.35 & 1.2529 \\
498: 6304.72 & 2.38$\pm$ 0.03 & \ion{Mg}{2}$\lambda$2796.35 & 1.255 \\
499: 6310.38 & 0.30$\pm$ 0.06 & & \\
500: 6316.13 & 0.30$\pm$ 0.03 & \ion{Mg}{2}$\lambda$2803.53 & 1.2529 \\
501: 6321.11 & 2.19$\pm$ 0.03 & \ion{Mg}{2}$\lambda$2803.53 & 1.255 \\
502: 6343.49 & 0.11$\pm$ 0.03 & \ion{Mg}{1}$\lambda$2852.13 & 1.2234 \\
503: 6427.36 & 0.24$\pm$ 0.03 & \ion{Mg}{1}$\lambda$2852.13 & 1.2529 \\
504: 6432.12 & 1.34$\pm$ 0.03 & \ion{Mg}{1}$\lambda$2852.13 & 1.255 \\
505: 6686.87 & 0.27$\pm$ 0.03 & \ion{Fe}{1}$\lambda$2967.76 & 1.2529 \\
506: \hline
507: \end{tabular}
508: \normalsize
509:
510: Errors are 67\% confidence intervals.
511: $^a$ the equivalent width of unidentified lines is computed at
512: zero redshift.
513: \end{table}
514:
515:
516: \begin{figure*}
517: \centerline{
518: \vbox{
519: \psfig{figure=f1_a.eps,height=16cm,width=5.5cm,angle=-90}
520: \psfig{figure=f1_b.eps,height=16cm,width=5.5cm,angle=-90}
521: \psfig{figure=f1_c.eps,height=16cm,width=5.5cm,angle=-90}
522: \psfig{figure=f1_d.eps,height=16cm,width=5.5cm,angle=-90}
523: }
524: }
525: \label{spe021004}
526: \caption {UVES spectrum of the GRB 021004 afterglow smoothed with a
527: gaussian function with $\sigma=1.5$ pixel . The dotted line is the
528: error spectrum}
529: \end{figure*}
530:
531: \setcounter{figure}{0}
532:
533: \begin{figure*}
534: \centerline{
535: \vbox{
536: \psfig{figure=f1_e.eps,height=16cm,width=5.5cm,angle=-90}
537: \psfig{figure=f1_f.eps,height=16cm,width=5.5cm,angle=-90}
538: \psfig{figure=f1_g.eps,height=16cm,width=5.5cm,angle=-90}
539: \psfig{figure=f1_j.eps,height=16cm,width=5.5cm,angle=-90}
540: }
541: }
542: \caption {continued}
543: \end{figure*}
544:
545: \setcounter{figure}{0}
546:
547: \begin{figure*}
548: \centerline{
549: \vbox{
550: \psfig{figure=f1_h.eps,height=16cm,width=5.5cm,angle=-90}
551: \psfig{figure=f1_k.eps,height=16cm,width=5.5cm,angle=-90}
552: \psfig{figure=f1_i.eps,height=16cm,width=5.5cm,angle=-90}
553: \psfig{figure=f1_l.eps,height=16cm,width=5.5cm,angle=-90}
554: }
555: }
556: \caption {continued}
557: \end{figure*}
558:
559:
560: \begin{figure*}
561: \centerline{
562: \vbox{
563: \psfig{figure=f2_a.eps,height=16cm,width=5.5cm,angle=-90}
564: \psfig{figure=f2_b.eps,height=16cm,width=5.5cm,angle=-90}
565: \psfig{figure=f2_c.eps,height=16cm,width=5.5cm,angle=-90}
566: \psfig{figure=f2_d.eps,height=16cm,width=5.5cm,angle=-90}
567: }
568: }
569: \label{spe020813}
570: \caption {UVES spectrum of the GRB 020813 afterglow smoothed with a
571: gaussian function with $\sigma=1.5$ pixel . The dotted line is the
572: error spectrum}
573: \end{figure*}
574:
575:
576:
577: \section{Column densities}
578:
579: In this paper we focus on the systems which are likely to be
580: associated with the GRB host galaxies. The line fitting was
581: performed using the MIDAS package FITLYMAN (Fontana \& Ballester
582: 1995). This uses a Voigt profile and yields independently
583: the column density N and the Doppler parameter $b$ for each absorption
584: component. For each absorption system several lines, spread over the
585: entire spectral range covered by the UVES observations, were
586: fitted simultaneously, using the same number of components for each
587: line, and the same redshift and $b$ value for each component.
588:
589:
590: \subsection{GRB 021004}
591:
592: For this GRB we consider the absorption systems at the following
593: redshifts and velocities (in km/s) with respect to the Ly$\alpha$
594: emission of the host galaxy (Mirabal et al. 2003): z=2.328, v=0
595: (system A in figure \ref{sys}); z=2.328, v=--139 km/s (system B);
596: z=2.328, v=--224 km/s (system C); z=2.321; z=2.298; z=2.296. For
597: these systems we detected \ion{C}{4}, \ion{C}{2}, \ion{Si}{4},
598: \ion{Al}{2}, \ion{Al}{3}, \ion{Mg}{2} and \ion{Fe}{2} lines (see Table
599: 3). We report in the table also the tentative identification of
600: a \ion{Si}{2}$\lambda$1304 line at z=2.2953. This is close but not
601: coincident with the redshift of systems z=2.296. We therefore consider
602: this identification uncertain, and we do not consider this line in
603: the following analysis.
604:
605: Unfortunately, the Ly${\alpha}$
606: absorption associated with the z=2.328 systems falls exactly in the gap
607: between the dichroic 1 and 2 blue arm spectra, and therefore is not
608: covered in this study. \ion{Zn}{2} and \ion{Cr}{2} absorption lines
609: for the main z=2.328 and z=2.321 systems fall in the region affected
610: by atmospheric telluric features (see figure 1) and are
611: therefore not accessible.
612:
613: Figure \ref{sys} shows the \ion{C}{4} and \ion{Si}{4} doublets of the
614: above 6 systems. Note that the \ion{C}{4}$\lambda$1548 lines of the
615: z=2.328 A,B,C systems are strongly blended with the
616: \ion{C}{4}$\lambda$1550 line of the z=2.321 system, while the
617: \ion{Si}{4}$\lambda$1404 line of the z=2.298 system is blended with
618: the \ion{Si}{4}$\lambda$1393 line of the z=2.321 system. Each of the
619: six systems actually comprises several components, within a velocity
620: range of several tens of km/s. For this reason the identification of
621: the different systems is somewhat subjective, the true message being
622: that the geometry and kinematics of the ISM clouds probed by the GRB
623: line of sight are complex. Nevertheless, sticking to the above system
624: identifications will be useful.
625:
626:
627: There are 13 lines in Table 3 associated with the z=2.328
628: systems, 5 lines associated with the z=2.321 system, and up to 17 lines
629: associated to the z=2.296-2.298 systems. Some of these lines can be
630: split further in several components. To test the robustness of the
631: fit, in terms of accuracy and stability of the results, we performed
632: many fits, using several combination of lines/systems. We found that
633: the fits presented below are a good compromise between increasing the
634: statistical precision of the fit, obtained increasing the number of
635: lines fitted simultaneously, and the stability/repeatibility of the
636: results, which degraded increasing the number of fitted parameters, due
637: to the increasingly complex shape of the $\chi^2$ hyper-surface in the
638: parameter space, that may contain many local minima.
639:
640: For the z=2.328 A,B,C and z=2.321 systems we fitted simultaneously the
641: \ion{C}{4}$\lambda \lambda$1550, 1548, \ion{C}{2}$\lambda$1334,
642: \ion{C}{2}$^*\lambda$1335, \ion{Si}{4}$\lambda
643: \lambda$1404, 1393 \ion{Al}{2} $\lambda1670$ and the
644: \ion{Al}{3}$\lambda$1854 line. We excluded from the fit the z=2.328
645: \ion{Al}{3}$\lambda$1862 line because its blue wing is strongly
646: blended with another line, and the z=2.321 \ion{Al}{3}
647: $\lambda$1862 line because it is strongly blended with the
648: \ion{Fe}{2}$\lambda$2600 line of one of the components of the
649: z=1.38 intervening system. For the same systems we also fitted the
650: \ion{Fe}{2}$\lambda$2382, \ion{Fe}{2}$\lambda$2374
651: \ion{Fe}{2}$\lambda$2344 \ion{Fe}{2}$\lambda$1608,
652: \ion{Mg}{2}$\lambda$2803 and \ion{Mg}{2}$\lambda$2796 lines
653: simultaneously. Figure \ref{z1} shows the line spectra in velocity
654: space, along with the best fit model, while Table 5 presents the best
655: fit abundances along with the velocity shift of each system with
656: respect to the redshift of the host galaxy, assumed to be 2.328 for
657: GRB 021004 (Mirabal et al. 2003). We used 2 components for the
658: z=2.328\_A system, and 1 component for the other 3 systems. The best
659: fit doppler parameter b ranges from a minimum of 12$\pm$6 km/s (one of
660: the components of the z=2.328\_A system) up to a maximum of 107$\pm$10
661: km/s (system z=2.321). Of course different b values are obtained
662: using an higher (or lower) number of components. Unfortunately the
663: signal to noise ratio of our spectrum is not good enough to
664: unambiguously identify the different components. On the other hand, we
665: verified that the total best fit column density of each system is
666: stable, within the statistical errors, changing the number of
667: components in each system.
668:
669: Similar series of fits were performed for the z=2.296 and z=2.298
670: systems, for which we used 1 and 2 components, respectively (see
671: figure \ref{z2}). Table 5 gives again the best fit abundances for
672: these 2 systems. The best fit doppler parameter b ranges from
673: 9$\pm$7 km/s to 40$\pm$13 km/s.
674:
675:
676:
677:
678: \begin{figure*}
679: \centerline{
680: \vbox{
681: \psfig{figure=f3a.eps,height=14cm,width=10cm,angle=-90}
682: \psfig{figure=f3b.eps,height=14cm,width=10cm,angle=-90}
683: }
684: }
685: \caption{\ion{C}{4} (upper panel) and \ion{Si}{4} (lower panel) absorption
686: systems in the UVES spectra of GRB 021004 for z=2.328 (see Table 5).
687: }
688: \label{sys}
689: \end{figure*}
690:
691: \begin{figure*}
692: \centerline{
693: \vbox{
694: \psfig{figure=f5a.eps,width=10cm,angle=-90}
695: \psfig{figure=f5b.eps,width=10cm,angle=-90}
696: }
697: }
698: \caption{GRB 021004. Upper panel: UVES spectrum near the \ion{C}{4}$\lambda
699: \lambda$1550, 1548, \ion{C}{2}$\lambda$1334, \ion{C}{2}$^*\lambda$1335,
700: \ion{Si}{4}$\lambda \lambda$1404, 1393,
701: \ion{Al}{2}$\lambda$1670 and \ion{Al}{3}$\lambda$1854
702: lines for the z=2.328 and z=2.321 systems in velocity space, along
703: with the best fit model (solid line) and residuals (grey). Lower
704: panel: same for the \ion{Fe}{2}$\lambda$2382,
705: \ion{Fe}{2}$\lambda$2374, \ion{Fe}{2}$\lambda$2344,
706: \ion{Fe}{2}$\lambda$1608 and \ion{Mg}{2}$\lambda$2803 lines. The zero
707: in the velocity scale refers to the redshift of the host
708: galaxy. }
709: \label{z1}
710: \end{figure*}
711:
712: \begin{figure*}
713: \centerline{
714: \vbox{
715: \psfig{figure=f7a.eps,width=10cm,angle=-90}
716: \psfig{figure=f7b.eps,width=10cm,angle=-90}
717: }
718: }
719: \caption{GRB 021004. Upper panel: UVES spectrum near the \ion{C}{4} $\lambda
720: \lambda$ 1550, 1548, \ion{C}{2}$\lambda$1334, \ion{C}{2}$^*\lambda$1335,
721: \ion{Si}{4}$\lambda \lambda$1404, 1393
722: lines for the z=2.298 and z=2.296 systems in
723: velocity space, along with the best fit model (solid line) and
724: residuals (grey). Lower panel: same for the \ion{Fe}{2}$\lambda$2382,
725: \ion{Fe}{2}$\lambda$2374, \ion{Fe}{2}$\lambda$2344,
726: \ion{Fe}{2}$\lambda$1608, \ion{Mg}{2}$\lambda$2803 and
727: \ion{Mg}{2}$\lambda$2796 lines. The zero in the velocity scale
728: refers to z=2.296.}
729: \label{z2}
730: \end{figure*}
731:
732:
733: \begin{table}
734: \caption{\bf GRB 021004. Logarithmic ion column densities in cm$^{-2}$}
735: \footnotesize
736: \begin{tabular}{lcccccccc}
737: \tableline\tableline
738: System & v. shift$^{a}$ & \ion{Si}{4} & \ion{C}{4} & \ion{C}{2} & \ion{C}{2}$^*$ & \ion{Al}{2} & \ion{Fe}{2} & \ion{Mg}{2} \\
739: \hline
740: 2.328\_A & 0 & 15.30$\pm$0.56 & $>$15.2 & 13.40$\pm$0.45 & 13.90$\pm$0.58 & 13.55$\pm$0.35$^c$ & 13.34$\pm$0.15 & 13.82$\pm$0.24\\
741: 2.328\_B & --139 & 14.27$\pm$0.16 & $>$14.4 & $<$13.2 & $<$13.2 & 12.23$\pm$0.35 & 12.53$\pm$0.24 & 12.93$\pm$0.60\\
742: 2.328\_C & --224 & 13.24$\pm$0.17 & 14.40$\pm$0.11 & $<$13.2 & $<$13.2 & $<$12.0 & $<$12.3 & $<$12.8 \\
743: 2.321 & --632 & 14.11$\pm$0.08 & 15.04$\pm$0.05 & $<$13.2 & $<$13.2 & $<$12.0 & $<$12.3 & $<$12.8 \\
744: 2.298 & --2729 & 13.43$\pm$0.33 & 14.16$\pm$0.20 &14.28$\pm$0.20$^b$&$<$13.2 & $<$12.0 & 12.68$\pm$0.20 & 13.22$\pm$0.10\\
745: 2.296 & --2913 & 13.79$\pm$0.16 & 14.72$\pm$0.25 & 13.43$\pm$0.40 & $^b$ & $<$12.0 & $<$12.3 & 12.68$\pm$0.21\\
746: \tableline
747: \end{tabular}
748: \normalsize
749:
750: errors, upper and lower limits are 90\% confidence intervals;
751: $^{a}$km/s; $^b$ the z=2.296 \ion{C}{2}$^*$ line is completely blended with the z=2.298 \ion{C}{2} line;
752: $^c$ for system z=2.328\_A we measuered also a column density of \ion{Al}{3} of 13.66$\pm$0.30.
753:
754: \end{table}
755:
756:
757: \begin{figure}[h]
758: \centering
759: \includegraphics[width=10cm]{f9.eps}
760: \caption{The logarithmic ratio between the \ion{C}{4}, \ion{C}{2}, \ion{Fe}{2},
761: \ion{Al}{2} and \ion{Mg}{2} column densities to that of \ion{Si}{4}
762: for the six absorption systems as a function of the negative velocity
763: shift with respect to the redshift of the host galaxy. }
764: \label{relden}
765: \end{figure}
766:
767:
768: The detection of both high and low ionization lines, feasible thanks
769: to the extremely wide spectral coverage achieved by UVES, allows us to
770: obtain constraints on the ionization status of the gas responsible for
771: the UV absorption, by comparing ion column density ratios with the
772: predictions of photoionization codes. Unfortunately we were not able
773: to measure column densities of different ions of the same element,
774: except for \ion{C}{2}/\ion{C}{4} and \ion{Al}{2}/\ion{Al}{3} for
775: system z=2.328\_A, and \ion{C}{2}/\ion{C}{4} for system z=2.298. In the
776: other cases we are forced to use ratios of column densities of
777: different ions of different elements in this analysis. These estimates
778: of the gas ionization parameter are therefore somewhat degenerate with
779: respect to relative element abundances. We used the Grevesse \&
780: Anders (1989) meteoritic abundances with extensions by Grevesse et
781: al. (1993).
782:
783:
784:
785: \begin{figure}
786: \vbox{
787: \hbox{
788: \psfig{figure=f10_a.eps,height=6.8cm,width=8cm}
789: \psfig{figure=f10_b.eps,height=6.8cm,width=8cm}
790: }
791: \hbox{
792: \psfig{figure=f10_c.eps,height=6.8cm,width=8cm}
793: \psfig{figure=f10_d.eps,height=6.8cm,width=8cm}
794: }
795: \hbox{
796: \psfig{figure=f10_e.eps,height=6.8cm,width=8cm}
797: \psfig{figure=f10_f.eps,height=6.8cm,width=8cm}
798: }
799: }
800: \vspace{-0.3cm}
801: \caption{Ion ratios from CLOUDY models for the six absorption
802: systems, assuming a gas density of 1 cm$^{-3}$:
803: [\ion{Mg}{2}/\ion{Si}{4}] = solid red curves,
804: [\ion{Fe}{2}/\ion{Si}{4}]= dashed blue curves,
805: [\ion{Al}{2}/\ion{Si}{4}] = dotted magenta curves and
806: [\ion{C}{4}/\ion{Si}{4}] and [\ion{C}{2}/\ion{Si}{4}]= dot-dashed black
807: curves. The line indicating the model values is thick when it is
808: consistent within 90\% with the measured ion ratios. The vertical
809: solid lines mark the range of allowed U values in each panel.}
810:
811: \label{ion2}
812: \end{figure}
813:
814: \begin{figure}
815: \vbox{
816: \hbox{
817: \psfig{figure=f11_a.eps,height=6.8cm,width=8cm}
818: \psfig{figure=f11_b.eps,height=6.8cm,width=8cm}
819: }
820: \hbox{
821: \psfig{figure=f11_c.eps,height=6.8cm,width=8cm}
822: \psfig{figure=f11_d.eps,height=6.8cm,width=8cm}
823: }
824: \hbox{
825: \psfig{figure=f11_e.eps,height=6.8cm,width=8cm}
826: \psfig{figure=f11_f.eps,height=6.8cm,width=8cm}
827: }
828: }
829: \caption{ Ion ratios from CLOUDY photoionization models for the
830: six absorption systems assuming a gas density of 10$^8$ cm$^{-3}$.
831: Curves as in figure \ref{ion2}. The line indicating the model values
832: is thick when it is consistent within 90\% with the measured ion
833: ratios, as in figure \ref{ion2}. The vertical solid lines mark the
834: allowed U values in each panel.}
835: \label{ion2_8}
836: \end{figure}
837:
838:
839: Figure \ref{relden} shows the logarithmic ratio between the
840: \ion{C}{4}, \ion{C}{2},\ion{Fe}{2}, \ion{Al}{2} and \ion{Mg}{2}
841: column densities and that of \ion{Si}{4} for the six absorption
842: systems as a function of the velocity shift with respect to the
843: redshift of the host galaxy. No large variation of the ion ratios are
844: seen for the six systems. We compared these line ratios to the
845: predictions obtained by simulating a gas cloud illuminated by an
846: ionizing continuum. We used Cloudy (vs 90.04, Ferland et al. 2002) to
847: build grids of photoionization models as a function of U, the
848: ionization parameter. U is defined as the ratio between the ionizing
849: photon density and the electron density of the gas. It is computed
850: assuming a constant density profile throughout the cloud, and a plane
851: parallel geometry. We studied gas densities between 1 cm$^{-3}$ and
852: $10^{8}$ cm$^{-3}$. The ionizing continuum was assumed to be a power
853: law, $F(E)=E^{-\Gamma}$ photons cm$^{-2}$ s$^{-1}$, with cutoffs at
854: low and high energies. The high energy cutoff is fixed at $10^{21}$Hz,
855: while we run the simulations for different low energy cutoffs, from
856: $10^{10}$Hz to $10^{14}$Hz. We produced several sets of simulations
857: with $\Gamma$ in the range 1--2. The ionizing continuum is constant
858: in time. GRBs are highly variable sources and the ionization structure
859: of the gas can be better studied using a time--dependent
860: photoionization code, such as those of Nicastro et al. (1999) and
861: Perna \& Lazzati (2002). Nevertheless, our simpler approach is
862: instructive in identifying general trends.
863:
864: \begin{figure}
865: \centering
866: \includegraphics[width=10cm]{f12.eps}
867: \caption{The logarithm of the best fit ionization parameters from
868: figure \ref{ion2} (n=1 cm$^{-3}$ filled circles) and from figure
869: \ref{ion2_8} (n=$10^8$ cm$^{-3}$, open
870: triangles, shifted by logU=-1.5), as a function of the velocity shift
871: with respect to the redshift of the host galaxy. Error-bars and upper
872: limits represent the purely statistical 90\% confidence intervals.}
873: \label{uv}
874: \end{figure}
875:
876: The predicted photoionization equilibrium ion ratios of
877: [\ion{Mg}{2}/\ion{Si}{4}], [\ion{Fe}{2}/\ion{Si}{4}],
878: [\ion{Al}{2}/\ion{Si}{4}] and [\ion{C}{4}/\ion{Si}{4}] are shown in
879: figure \ref{ion2} for $\Gamma=2$ and n=1 cm$^{-3}$ and figure
880: \ref{ion2_8} for $\Gamma=2$ and n=$10^8$ cm$^{-3}$. For each of the
881: six redshift systems, thick segments superimposed on these curves
882: indicate our 90\% confidence determinations for these ratios. In
883: figure \ref{ion2} (for n=1 cm$^{-3}$) the [\ion{Mg}{2}/\ion{Si}{4}],
884: [\ion{Fe}{2}/\ion{Si}{4}] and [\ion{Al}{2}/\ion{Si}{4}] ion ratios for
885: the systems z=2.328\_A, z=2.328\_B, z=2.328\_C, and z=2.296 are all
886: consistent with the same ionization status. For the system z=2.328\_A
887: we measured the column density of \ion{Al}{2} and \ion{Al}{3}. Also
888: their ratio is consistent with the ionization parameter inferred from
889: the above ion ratios, see figures \ref{ion2} and \ref{ion2_8}.
890: For the same system we measured the column density of \ion{C}{2}. The
891: [\ion{C}{2}/\ion{Si}{4}] ion ratio indicates a ionization parameter
892: somewhat higher than that indicated by the other ion ratios for n=1
893: cm$^{-3}$ (see figure \ref{ion2}), while for n=$10^8$ cm$^{-3}$ the
894: ionization parameter is fully consistent with that indicated by the
895: other ion ratios (figure \ref{ion2_8}). The system at z=2.298
896: shows a ionization parameter lower than that of system z=2.328\_A,
897: while the system at z=2.321 shows a higher ionization parameter.
898: However, all ionization parameters systems are consistent with a
899: remarkably narrow range: $10^{-1.7}<U<10^{-1}$, for n=1 cm$^{-3}$ and
900: $10^{-1.2}<U<10^{-1}$, for n=$10^8$ cm$^{-3}$ see figure \ref{uv},
901: which plots the ionization parameters as a function of the velocity
902: for the six systems. We note that in these ranges of U the
903: [\ion{Fe}{2}/\ion{Mg}{2}] ratio varies by less than 30\%, and the
904: value found for the three systems in which we detect \ion{Fe}{2} and
905: \ion{Mg}{2} lines ([\ion{Mg}{2}/\ion{Fe}{2}]$\approx0.4$) is
906: consistent with what is expected from meteoritic abundances (and as
907: was adopted for the CLOUDY calculations). This value is also in the
908: range found for the Galactic ISM 0.1$<$[Mg/Fe]$<$0.8. It is
909: somewhat higher than that found by Savaglio et al. (2004) in a sample
910: of faint K--band selected galaxies at 1.4$<$z$<$2.0
911: -0.84$<$[Mg/Fe]$<$0.13. Assuming meteoritic abundances the hydrogen
912: column density of the densest system (z=2.328\_A) is
913: $\gs5\times10^{19}$ cm$^{-2}$, where the equality yields if all Si is
914: in the form of \ion{Si}{4}. This column is so low that would have
915: easily escaped detection in the Chandra X--ray spectrum (Butler et
916: al. 2003) and it is consistent with the upper limit given by
917: Moller et al. (2003).
918:
919: Qualitatively similar results are drawn from the analysis of the
920: n=$10^8$ cm$^{-3}$ curves, although the range of acceptable ionization
921: parameters for the 6 systems is narrower, see figure \ref{ion2_8}.
922: Qualitatevely similar results were also obtained for the $\Gamma=1$
923: models.
924:
925: \subsubsection{The z=2.328\_A system}
926:
927: For this system we detect the \ion{C}{2}$^*\lambda$1335 fine structure
928: line. Unfortunately, the \ion{C}{2}$\lambda$1334 ground state line is
929: barely visible (with a signal to noise of just 2). The presence
930: of a strong \ion{C}{2}$^*$ line suggests either a high gas density or a
931: strong radiation field (Srianand \& Petitjean 2000, 2001, Silva \&
932: Viegas 2002). \ion{C}{2}$^*$ and \ion{Si}{2}$^*$ fine structure lines
933: have been detected in two other GRBs: GRB 020813 (Savaglio \& Fall 2004)
934: and GRB 030323 (Vreeswijk et al. 2004). Their ratio can be used to
935: constrain the gas density or the radiation field (if both
936: lines are detected one may solve for both variables, as shown by
937: Srianand \& Petitjean 2000), so their detection in GRB spectra looks
938: very promising to constrain the physical properties of the
939: absorbing gas.
940:
941: For this system the \ion{C}{4} line is strongly saturated and so it
942: cannot be used together with the \ion{C}{2} to constrain the gas
943: ionization status. As discussed in the previous section, within
944: rather large uncertainties, the [\ion{C}{2}/\ion{Si}{2}] would suggest
945: for n=1 cm$^{-3}$ a ionization parameter somewhat higher than that
946: indicated by all other line ratios, assuming meteoritic
947: abundances. This may be explained by a slight under-abundance of C
948: with respect to Si in this system. A similar conclusion applies to
949: system z=2.321, see below.
950:
951: The \ion{C}{4}$\lambda1548$ lines of the z=2.328 A,B and C systems
952: are blended with the \ion{C}{4}$\lambda1550$ of the
953: z=2.321 system, suggesting the presence of line locking,
954: as first reported by Savaglio et al. (2002). Line locking
955: is usually interpreted as the signature of line radiation
956: pressure acceleration (see e.g. Foltz et al. 1987,
957: Srianand \& Petitjean 2000).
958:
959:
960:
961: \subsubsection{The z=2.321 system}
962:
963: For this system (as for z=2.328\_C) we were able to detect \ion{C}{4}
964: and \ion{Si}{4} absorption lines only. The upper limits on the
965: \ion{Mg}{2} and \ion{Fe}{2} transition allow us to put a significant
966: lower limit on the ionization parameter of U$\gs10^{-1}$ for this
967: system for a gas density of 1 cm$^{-3}$ and U$\gs10^{-1.2}$ for a gas
968: density of 10$^8$ cm$^{-3}$. For n=1, this is higher than the values
969: for the other systems at a confidence level better than 90\%.
970: The [\ion{C}{4}/\ion{Si}{4}] ratio is formally inconsistent with this
971: higher ionization parameter, but the disagreement is marginal, taking
972: into account that the \ion{C}{4}/\ion{Si}{4} curves in figures
973: \ref{ion2} and \ref{ion2_8} are very flat, and therefore small
974: differences in the obsterved \ion{C}{4}/\ion{Si}{4} ratio would
975: translate in large differences in the ionization parameter. Also in
976: this case the disagreement may be explained by a slight under-abundance
977: of C with respect to Si.
978:
979: \subsubsection{The z=2.296 and z=2.298 systems}
980:
981: Given the relatively high velocity shift of about 3000 km/s between
982: the z=2.296 and z=2.298 systems and the host galaxy redshift, there is
983: the possibility, at least in principle, that these systems are not
984: associated with the GRB host galaxy but rather are intervening
985: systems. We evaluated the probability of finding intervening \ion{C}{4}
986: systems within 3000 km/s from the GRB host galaxy, from the density
987: distribution of \ion{C}{4} systems given in D'Odorico et al. (1998) and
988: Petitjean \& Bergeron (1994). The probability of finding by chance
989: two systems with column densities equal or larger than those given in
990: Table 5 for these two lowest redshift systems is $1.2\times10^{-5}$,
991: while the probability for only one of these systems is 0.1\% for the
992: z=2.296 system and 0.5\% for the z=2.298 system.
993:
994: The z=2.298 system has a ionization parameter $<10^{-1.7}$ at the
995: 90\% confidence level for a gas density of 1 cm$^{-3}$ (around
996: $10^{-1.2}$ for a gas density of 10$^8$ cm$^{-3}$). This is the
997: lowest of the ionization parameters found in the 6 systems.
998: Furthermore, for this system we detect a rather strong
999: \ion{C}{2}$\lambda$1334 ground state line but not the
1000: \ion{C}{2}$^*\lambda$1335 high excitation line, see figure \ref{z2},
1001: again indicating a lower radiation field (or a lower gas density).
1002:
1003: The observed wavelength of the z=2.298 system \ion{C}{2}$\lambda$1334
1004: line coincides with the wavelength of the z=2.296 \ion{C}{2}$^*\lambda$1335
1005: line (see Table 3). This may be a further indication of line locking.
1006:
1007:
1008: %\subsubsection {\ion{C}{4} column density}
1009:
1010: %\ion{C}{4} lines for the systems z=2.328\_A and z=2.328\_B are
1011: %strongly saturated and therefore the column densities in Table 5
1012: %should be regarded as lower limits. The column densities of the other
1013: %four \ion{C}{4} systems are reasonably consistent with each other
1014: %within the errors (see figure \ref{relden}). Assuming meteoritic
1015: %abundances, the [\ion{C}{4}/\ion{Si}{4}] line ratios of these four
1016: %systems are also consistent with the ionization parameter suggested by
1017: %the other line ratios, within their rather large errors, if
1018: %$\Gamma=2$. The only possible exception is the system at z=2.321, but
1019: %also in this case the disagreement is marginal, taking into account
1020: %that the \ion{C}{4}/\ion{Si}{4} curves in figures \ref{ion2} and
1021: %\ref{ion2_8} are very flat, and therefore small differences in the
1022: %obsterved \ion{C}{4}/\ion{Si}{4} ratio would translate in large
1023: %differences in the ionization parameter.
1024:
1025: \subsubsection{Intervening systems}
1026:
1027: Two main intervening systems are present along the line
1028: of sight to GRB 021004, one around z=1.60 and the other around z=1.38
1029: (see eg. Mirabal et al. 2003, Moller et al 2003). The UVES high
1030: resolution spectrum of GRB 021004 allows an accurate study of these systems,
1031: which are split in several components (see Table 3).
1032:
1033: The system at z=1.38 consists at least 3 components at z=1.3807
1034: (z=1.38\_A), z=1.3802 (z=1.38\_B), z=1.3795 (z=1.38\_C). The latter
1035: system is also split in several parts in the range 1.3793-1.3798. The
1036: total velocity range spanned by the system is $\sim180$ km/s. We
1037: detected the following lines belonging to this system:
1038: \ion{Si}{2}$\lambda$1808, \ion{Al}{3}$\lambda$1854,
1039: \ion{Fe}{2}$\lambda$2344,\ion{Fe}{2}$\lambda$2374,
1040: \ion{Fe}{2}$\lambda$2586,\ion{Mg}{2}$\lambda \lambda$2796,2803, and
1041: \ion{Mg}{1}$\lambda$2852.
1042:
1043: The system at z=1.60 is made by at least 4 components at z=1.6028
1044: (z=1.60\_A), z=1.6024 (z=1.60\_B), z=1.6019 (z=1.60\_C) and z=1.6014
1045: (z=1.60\_D). The total velocity range spanned by the system is of
1046: $\sim160$ km/s. For this system we detected the following lines:
1047: \ion{Al}{2}$\lambda$1670,
1048: \ion{Fe}{2}$\lambda$2344,\ion{Fe}{2}$\lambda$2374,
1049: \ion{Fe}{2}$\lambda$2382,
1050: \ion{Fe}{2}$\lambda$2586,\ion{Mn}{2}$\lambda$2594,
1051: \ion{Fe}{2}$\lambda$2600,\ion{Mn}{2}$\lambda$2606, \ion{Mg}{2}$\lambda
1052: \lambda$2796,2803, and \ion{Mg}{1}$\lambda$2852.
1053:
1054: We estimated the column densities of the ions of the
1055: intervening systems using the same fitting procedure used
1056: for the six systems associated with the GRB host galaxy.
1057: They are given in Table 6
1058:
1059: \begin{table}
1060: \caption{\bf Log. ion column densities in cm$^{-2}$ of the GRB 021004
1061: intervening systems}
1062: \footnotesize
1063: \begin{tabular}{lccccc}
1064: \tableline\tableline
1065: System & v. shift$^a$ &\ion{Fe}{2}& \ion{Mg}{2} & \ion{Si}{2} & \ion{Al}{3}\\
1066: \hline
1067: z=1.38\_A & 0 & 15.68$\pm$0.16 & 13.23$\pm$0.07 & 15.12$\pm$0.20 & 13.02$\pm$0.20 \\
1068: z=1.38\_B & --63 & -- & 12.92$\pm$0.32 & -- & -- \\
1069: z=1.38\_C & --150 & 14.75$\pm$0.40 & 15.20$\pm$0.70 & -- & -- \\
1070: \hline
1071: System & v. shift$^a$ &\ion{Fe}{2}& \ion{Mg}{1} & \ion{Mg}{2} & \ion{Al}{2}\\
1072: \hline
1073: z=1.60\_A & 0 & 13.69$\pm$0.16 & 11.33$\pm$0.26 & 13.54$\pm$0.21 & 12.66$\pm$0.27 \\
1074: z=1.60\_B & --46 & 12.63$\pm$0.16 & 11.64$\pm$0.18 & 12.95$\pm$0.10 & -- \\
1075: z=1.60\_C & --104 & 14.35$\pm$0.18 & 12.80$\pm$0.27 & 15.45$\pm$0.80 & 13.24$\pm$0.50 \\
1076: z=1.60\_D & --160 & 13.57$\pm$0.10 & 11.82$\pm$0.14 & 13.47$\pm$0.06 & 12.58$\pm$0.25 \\
1077: \tableline
1078: \end{tabular}
1079: \normalsize
1080:
1081: errors, upper and lower limits are 90\% confidence intervals;
1082: $^{a}$km/s
1083:
1084: \end{table}
1085:
1086: \begin{table}
1087: \caption{\bf GRB 020813, Logarithmic ion column densities in cm$^{-2}$}
1088: \footnotesize
1089: \begin{tabular}{lccc}
1090: \tableline\tableline
1091: System & velocity shift$^{a}$ & \ion{Fe}{2} & \ion{Mg}{2} \\
1092: \hline
1093: z=1.255\_A & 0. & 15.20$\pm$0.29 & $>16.0$ \\
1094: z=1.255\_B & --106 & 15.47$\pm$0.27 & $>15.5$ \\
1095: z=1.255\_C & --306 & 13.58$\pm$0.10 & 13.79$\pm$0.21 \\
1096: z=1.2234 & --4204 & 13.96$\pm$0.18 & $>16.0$ \\
1097: \tableline
1098: \end{tabular}
1099: \normalsize
1100:
1101: $^{a}$km/s
1102:
1103: \end{table}
1104:
1105:
1106: \subsection{GRB 020813}
1107:
1108: For this GRB we considered the absorption systems at the following
1109: velocities with respect to the redshift of the host galaxy, which was
1110: assumed to be 1.255 (Barth et al. 2003): v=0 km/s (system A in figure
1111: \ref{grb020813}), v=--106 km/s (system B) and v=--306 km/s (system C).
1112: Note that the system A and system B lines are strongly blended with
1113: the exception of the \ion{Fe}{2}2374 line. We also consider the
1114: system at z=1.2234.
1115:
1116: We fitted simultaneously the 6 lines of the 3 absorption systems in
1117: figure \ref{grb020813}. Table 7 gives for the 3 systems the best fit
1118: \ion{Fe}{2} and \ion{Mg}{2} abundances along with the velocity shift
1119: of each system with respect to the redshift of the host galaxy.
1120: \ion{Mg}{2} lines of systems A and B are strongly saturated and
1121: therefore their column density estimates are more uncertain.
1122: Unfortunately the UVES spectrum covers a wavelength range much smaller
1123: than the Keck LRIS spectrum and several of the lines studied by Barth
1124: et al. (2003) and Savaglio \& Fall (2004), in particular \ion{Zn}{2},
1125: \ion{Cr}{2}, \ion{Si}{2}, are not accessible.
1126:
1127: We performed similar fits to the 5 lines associated with the
1128: z=1.2234 system, split in two components.
1129: The results are in figure \ref{grb020813_12234} and in Table 7.
1130:
1131: Unfortunately in this case the redshift of the GRB is not high enough
1132: to have strong high ionization lines in the spectrum, and therefore we
1133: do not have a direct way to constrain the ionization status of the gas
1134: responsible for the UV absorption.
1135:
1136: The [\ion{Fe}{2}/\ion{Mg}{2}] ratio is consistent with a constant
1137: value in the four systems. This ratio is also consistent with
1138: meteoritic abundances, assuming that the ionization parameter is in
1139: the range $10^{-2.5} - 10^{-1}$.
1140:
1141:
1142:
1143: \subsection{The z=1.2234 system}
1144:
1145: The system at z=1.2234 is shifted by about 4200 km/s from the redshift
1146: of the host galaxy. The probability to find an \ion{Mg}{2} intervening
1147: system with $W_{\lambda 2796.35}=1.3\pm0.03\AA\ $ (see Table 7) within
1148: this velocity range is $\ls1\%$, using the distribution of \ion{Mg}{2}
1149: systems given by Steidel \& Sargent (1992). Although not as
1150: conclusive as in the case of the z=2.296 and z=2.298 systems in the
1151: spectrum of GRB 021004, we remark that this probability is rather
1152: small, and that the velocity shift with respect to the redshift of the
1153: host galaxy is intriguingly similar to that of the GRB 021004 systems.
1154:
1155: \begin{figure}
1156: \centerline{
1157: \psfig{figure=f13.eps,width=10cm,angle=-90}
1158: }
1159: \caption{UVES spectrum of GRB 020813 near the \ion{Fe}{2}$\lambda$2600,
1160: \ion{Fe}{2}$\lambda$2382, \ion{Fe}{2}$\lambda$2374,
1161: \ion{Fe}{2}$\lambda$2344, \ion{Mg}{2}$\lambda$2803 and
1162: \ion{Mg}{2}$\lambda$2796 lines for the three absorption systems, in
1163: velocity space, along with the best fit model, solid line and
1164: residuals. The zero of the velocity scale refers to z=1.2545}
1165: \label{grb020813}
1166: \end{figure}
1167:
1168: \begin{figure}
1169: \centerline{
1170: \psfig{figure=f14.eps,width=10cm,angle=-90}
1171: }
1172: \caption{UVES spectrum of GRB 020813 near the
1173: \ion{Fe}{2}$\lambda$2382,
1174: \ion{Fe}{2}$\lambda$2344, \ion{Mg}{2}$\lambda$2803,
1175: \ion{Mg}{2}$\lambda$2796 and \ion{Mg}{1}$\lambda$2852 lines
1176: for the z=1.2234 absorption systems, in
1177: velocity space, along with the best fit model, solid line and
1178: residuals. The zero of the velocity scale refers to z=1.2234}
1179: \label{grb020813_12234}
1180: \end{figure}
1181:
1182:
1183: \section{Discussion}
1184:
1185: GRB 021004 shows absorption systems which span $\sim3000$~km/s in
1186: velocity towards the observer (no receding system is detected). The
1187: systems are most likely local to the GRB since the probability of
1188: finding two or more different absorption systems along the line of
1189: sight by random fluctuations is negligible.
1190:
1191: It has been suggested that radiative acceleration by the prompt GRB
1192: emission may be responsible for the high detected speeds (Schaefer et
1193: al. 2003). However, this seems unlikely since low ionization ions such
1194: \ion{Fe}{2} and \ion{Mg}{2} are present in the high velocity absorber.
1195: The radiative acceleration is dominated by the GRB and early afterglow
1196: radiation. It is therefore unlikely that recombination can be so rapid
1197: to have any influence on the acceleration of the absorbing
1198: material. In the photoionization phase each nucleon receives an amount
1199: of energy comparable to the ionization potential of the $K$--shell
1200: electron. For a solar metallicity plasma the acceleration is regulated
1201: by H atoms and, at the end of the photoionization, the absorbing
1202: material has an outward velocity of:
1203: %
1204: \begin{equation}
1205: v_{\rm{ion}}\sim\sqrt{\frac{h\nu_{\rm{ion}}}{m_p}}\sim50\quad{\rm km/s}
1206: \end{equation}
1207: %
1208: Once the gas is fully ionized, acceleration proceeds due to radiation
1209: pressure on free electrons through Inverse Compton (IC)
1210: interactions. For a burst of isotropic equivalent energy
1211: $E_{\rm{iso}}$ the IC acceleration is obtained by momentum
1212: conservation:
1213: %
1214: \begin{equation}
1215: v_{\rm{IC}}=\frac{E_{\rm{iso}}\,\sigma_T}{
1216: 4\pi\,R^2\,m_p\,c} \sim 0.6\,R_{18}^{-2}\quad{\rm km/s}
1217: \end{equation}
1218: %
1219: we conclude that at the minimum distance of the absorbing medium
1220: $R\sim10^{18}$~cm (Lazzati et al. 2002; Heyl \& Perna 2003; in order
1221: to be ahead of the fireball at the time at which absorption is
1222: detected) radiative acceleration is unable to propel the absorber to
1223: an outflowing speed comparable to the value(s) observed.
1224:
1225: Possible explanations for the large outflowing speed are either a
1226: supernova exploded several years prior to the GRB (Vietri \& Stella
1227: 1998) or a high velocity wind from the progenitor Wolf--Rayet (WR)
1228: star (MacFadyen \& Woosley 1999; Schaefer et al. 2003; Mirabal et
1229: al. 2003). The large velocity spread with similar ionization parameter
1230: is however difficult to account for in a SNR scenario, where values
1231: around a typical expansion velocity would be expected. On the other
1232: hand, WR winds are known to be clumpy and velocities up to
1233: $\sim4000$~km/s were detected from P--Cygni profiles (Niedzlieski \&
1234: Sk\'orzy\'nski 2002).
1235:
1236: In this work we derived physical parameters of the absorbers under the
1237: assumption of equilibrium conditions which, $\sim0.5$~days after the
1238: GRB explosion, are attained only if the electron density in the
1239: absorber is $n\sim 10^{7-8}$~cm$^{-3}$. Assuming equilibrium
1240: conditions, the total column density of the absorber can be computed
1241: from the ion column density corrected for the ionization
1242: fraction. Consider the \ion{Si}{4} line in the A system. The density
1243: is given by:
1244: %
1245: \begin{equation}
1246: n=\frac{N_{\rm{Si}}}{A_{\rm{Si}}\,R\,\eta}
1247: \sim\frac{10^{15.3}}{4\times10^{-5}\,10^{18}\,R_{18}\,10^{-6}\eta_{-6}}
1248: \approx5\times10^7 \, R_{18}^{-1}\, \eta_{-6}^{-1} \qquad {\rm cm}^{-3}
1249: \end{equation}
1250: %
1251: where $R$ is the fireball radius at the time the lines were observed,
1252: $A_{\rm{Si}}$ is the Si abundance
1253: and $\eta$ is the ratio of the width of the absorbing shell over its
1254: radius. The ionization parameters derived above and implications
1255: discussed here are therefore relevant only in the case of an extremely
1256: clumpy wind or SNR.
1257:
1258: The photoionization results of CLOUDY yield an ionization parameter
1259: constrained in a relatively small range $10^{-1.7}<U<10^{-1}$. In a
1260: single explosion GRB model, the ionization parameter scales with the
1261: square of the outflow velocity, since the absorber's distance scales
1262: with velocity as well. Even though the ionization parameter depends on
1263: the electron density, it seems unlikely that density variations are
1264: such as to compensate for the large velocity difference. In a wind
1265: environment, on the other hand, the ionization parameter is constant
1266: since the photon density and particle density scale with the same
1267: power of the distance. The relatively small variations in the inferred
1268: $U$, which do not show any clear trend with velocity, can therefore be
1269: interpreted as density fluctuations on top of a regular $R^{-2}$ wind
1270: density profile (as already discussed by e.g. Schaefer et al. 2003
1271: and Mirabal et al. 2003).
1272:
1273: Finally, we note that the \ion{Fe}{2} and \ion{Mg}{2} column densities
1274: found in GRB 020813 are 10--100 times higher than those in GRB 021004.
1275: This is likely due to a much higher ionization of the gas in the
1276: latter case, rather than a large difference in the total absorbing
1277: column of gas or to highly non solar metal abundances (note that in
1278: both cases the [\ion{Fe}{2}/\ion{Mg}{2}] ratio is consistent with
1279: meteoritic abundances). We also note the similarity between the
1280: velocity shift between the z=1.2234 system and the GRB 020813 host
1281: galaxy and the shift of the z=2.296 and z=2.298 systems with respect
1282: to the GRB 021004 host galaxy. Although the probability for a chance
1283: occurrence of the z=1.2234 system is not as low as in the cases of the
1284: two shifted systems in GRB 021004, and that the z=1.2234 ionization
1285: status is probably much lower than that of the z=2.296 and z=2.298
1286: systems, this similarity might suggest a common scenario for the two
1287: GRBs.
1288:
1289: \section{Conclusions}
1290:
1291: One of the straightforward results of our UVES high resolution
1292: observations of two GRB afterglows is that the ISM of the host
1293: galaxies is complex, and many components contribute to each main
1294: absorption systems. These components span a total velocity range of up
1295: to about 3000 km/s. Several narrow components are resolved down to a
1296: width of a few tens of km/s. The UVES wide band coverage allowed us
1297: to investigate simultaneously both high ionization lines such as
1298: \ion{C}{4} and \ion{Si}{4}, and low ionization lines such as
1299: \ion{Mg}{2} and \ion{Fe}{2} in GRB 021004. This allowed us to constrain
1300: the ionization parameter of the gas of the differerent absorption
1301: systems. Combined with photoionization results obtained with CLOUDY
1302: the ionization parameters appear to lie in a relatively small range,
1303: with no clear trend with the system velocity. This can be interpreted
1304: as density fluctuations on top of a regular $R^{-2}$ wind density
1305: profile. The [\ion{Mg}{2}/\ion{Fe}{2}] ratio of $\approx0.4$ found for
1306: systems z=2.328\_A, z=2.328\_B, and z=2.298 is consistent with what is
1307: expected from meteoritic abundances. Assuming these abundances, the
1308: measured \ion{Si}{4} column density for system z=2.328\_A implies a
1309: lower limit of $\gs5\times10^{19}$ cm$^{-2}$ to the system total
1310: hydrogen column density.
1311:
1312: Indeed, our study shows that rapid reaction to the GRB triggers, high
1313: resolution and wide spectral coverage are the key ingredients to study
1314: GRB host galaxies. Today observations of this kind are still difficult
1315: and rather episodic. However, they should become routine after the
1316: launch of the {\it Swift} satellite, also thanks to the development of
1317: a dedicated Rapid Response observing Mode for the VLT telescopes.
1318: This mode will make possible automatic follow--up of GRBs (or other
1319: transient events) with response times in the 10m--1hr range, therefore
1320: helping in gathering spectra of unprecedented quality of medium to
1321: high redshift GRB host galaxies.
1322:
1323: \acknowledgments
1324: We acknowledge support from contract ASI/I/R/390/02
1325: and MIUR grant Cofin--2003--41. We thank Fabrizio Nicastro and
1326: Emanuele Giallongo for usefull discussions, Sandra Savaglio for her
1327: early work on this program, and an anonymous referee for comments
1328: that helped to improve the presentation.
1329:
1330: \begin{thebibliography}{}
1331:
1332: \bibitem{} Akerlof, C., Balsano, R., Barthelemy, S. et al. 1999,
1333: Nature, 398, 400
1334:
1335: \bibitem{} Barth, A.J., Sari, R., Cohen, M.H. 2003, ApJ, 584, L47
1336:
1337: \bibitem{} Butler, N.R., Marshall, H.L, Ricker, G.R. et al. 2003,
1338: ApJ, 597, 1010
1339:
1340: \bibitem{} Covino, S., Malesani, D. Tavecchio, F. et al. 2003,
1341: A\&A, 404, L5
1342:
1343: \bibitem{} Dekker, H., D'Odorico, S., Kaufer, A.,
1344: Delabre, B., \& Kotzlowski, H. 2000, The International
1345: Society for Optical Engineering, 4008, 534
1346:
1347: \bibitem{} Della Valle, M. Malesani, D., Benetti, S. et al. 2003,
1348: A\&A, 406, L33
1349:
1350: \bibitem{} D'Odorico, V., Cristiani, S., D'Odorico, S., Fontana, A.,
1351: Giallongo, E. 1998, A\&AS, 127, 217
1352:
1353: \bibitem{} Ferland, G.S. 2002, University of Kentucky Dept. of Physics
1354: and Astronomy, Internal Report
1355:
1356: \bibitem{} Fiore, F. et al. 2002, GCN 1524
1357:
1358: \bibitem{} Foltz, C.B., Weymann, R.J., Morris, S.L., Turnshek, D.A. 1987,
1359: ApJ, 317, 450
1360:
1361: \bibitem{} Fontana, A. \& Ballester, P. 1995, The ESO Messenger, 80, 37
1362:
1363: \bibitem{} Fox, D.W., Blake, C., Price, P.A. 2002, GCN Circ. 1470
1364:
1365: \bibitem{} Fox, D.W. 2002, GCN Circ. 1564
1366:
1367: \bibitem{} Fox, D.W., Yost, S, Kulkarni, S.R.
1368: et al. 2003, Nature, 422, 284
1369:
1370: \bibitem{} Galama, T.J., Vreeswijk, P. M., van Paradijs, J. et al.
1371: 1998, Nature, 395, 670
1372:
1373: \bibitem{} Galama, T.J., Briggs, M.S., Wijers, R.A.M.
1374: et al. 1999, Nature, 398, 394
1375:
1376: \bibitem{} Gladders, M. \& Hall, P. 2002, GCN Circ. 1472
1377:
1378: \bibitem{} Grevesse, N. \& Anders, E. 1989, AIP Conference Proceedings, Vol. 183.
1379: New York, American Institute of Physics, p. 1-8.
1380:
1381: \bibitem{} Grevesse, N., Noel, A. \& Sauval, A.J., 1993, A\&A, 271, 587
1382:
1383: \bibitem{} Heyl, J. S. \& Perna, R. 2003, ApJ, 586, L13
1384:
1385: \bibitem{} Hjorth, J., Sollerman, J., Moller, P. et al. 2003, Nature, 423, 847
1386:
1387: \bibitem{} Holland, S.T., Weidinger, M., Fynbo, J.P.U. et al. 2003,
1388: AJ, 125, 2291
1389:
1390:
1391: \bibitem{} Lazzati, D., Rossi, E., Covino, S., Ghisellini, G.,
1392: \& Malesani, D.\ 2002, A\&A, 396, L5
1393:
1394: \bibitem{} MacFadyen, A.I. \& Woosley, S.E. 1999, ApJ, 524, 262
1395:
1396: \bibitem{} Mirabal, N., Halpern, J. P., Chornock, R. et al. 2003,
1397: ApJ, 595, 935
1398:
1399: \bibitem{} Matheson, T., Garnavich, P. M., Foltz, C., et al. 2003,
1400: ApJL, 582, L5
1401:
1402: \bibitem{} Moller, P. et al. 2003, A\&A, 396, L21
1403:
1404: \bibitem{} Nicastro, F., Fiore, F., Perola, G.C., Elvis, M. 1999,
1405: ApJ, 512, 184
1406:
1407: \bibitem{} Niedzlieski, A. \& Sk\'orzy\'nski, W., 2002, Acta Astronomica,
1408: 52, 81
1409:
1410: \bibitem[Paczynski(1998)]{pac98} Paczynski, B. 1998, ApJL, 494, L45
1411:
1412: \bibitem{} Perna, R. \& Lazzati, D. 2002, ApJ, 580, 261
1413:
1414: \bibitem{} Petitjean, P. \& Bergeron, J. 1994, A\&A 283, 759
1415:
1416: \bibitem{} Pettini, M., Smith, L.J., King, D.L., Hunstead, R.W. 1997,
1417: ApJ, 486, 665
1418:
1419: \bibitem[Price et al. (2003)]{pri03} Price, P.A. et al. 2003, Nature, 423, 844
1420:
1421: \bibitem{} Savaglio, S. et al. 2002, GCN 1633
1422:
1423: \bibitem{} Savaglio, S., Fall, S.M., Fiore, F. 2003, ApJ, 585, 638
1424:
1425: \bibitem{} Savaglio, S., Glazebrook, K., Crampton, D. et al. 2004,
1426: ApJ, 602, 51
1427:
1428: \bibitem{} Savaglio, S., Fall, S.M. 2004, ApJ, 614, 293
1429:
1430: \bibitem{} Silva, A.I., \& Viegas, S.M. 2002, MNRAS, 329, 135
1431:
1432: \bibitem{} Shirasaki, Y. et al. 2002 GCN 1565
1433:
1434: \bibitem{} Srianand, R. \& Petitjean, P. 2000, A\&A, 357, 414
1435:
1436: \bibitem{} Srianand, R. \& Petitjean, P. 2001, A\&A, 373, 816
1437:
1438: \bibitem[Stanek et al. (2003)]{sta03} Stanek, K.Z., Matheson, T.,
1439: Garnavich, P. M. et al. 2003, ApJ, 591L, 17
1440:
1441: \bibitem{} Steidel, C.C. \& Sargent, W.L.W. 1992, ApJS, 80, 1
1442:
1443: \bibitem{} Steidel, C.C., Adelberger, K.L., Giavalisco, M., Dickinson, M.,
1444: Pettini, M. 1999, ApJ, 519, 1
1445:
1446: \bibitem{} Schaefer, B.E., Gerardy, C.L., Höflich, P. et al.,
1447: 2003, ApJ, 588, 387
1448:
1449: \bibitem{} Turnshek, D.A., Wolfe, A.M., Lanzetta, K.M. et al. 1989,
1450: ApJ, 344, 567
1451:
1452: \bibitem{} Vietri, M. \& Stella, L. 1998, ApJ, 507, L45
1453:
1454: \bibitem{} Villasenor et al. 2002, GCN 1471
1455:
1456: \bibitem{} Vreeswijk, P.M., Ellison S.L., Ledoux, C. et al. 2004, A\&A,
1457: 419, 927
1458:
1459: \bibitem[Woosley(1993)]{woo93} Woosley, S. E. 1993, \apj, 405, 273
1460:
1461: \end{thebibliography}
1462:
1463:
1464: \end{document}
1465: