1: %\documentclass{emulateapj}

2: 

3: \documentclass[12pt,preprint]{aastex}

4: 

5: %\usepackage{natbib}

6: 

7: \shortauthors{Hao, H. et.~al}

8: 

9: \shorttitle{Variable Intervening Fe~II and Mg~II Absorber}

10: 

11: \begin{document}

12: 

13: \title{Strongly Variable $z=1.48$  Fe~II and Mg~II Absorption in the

14: Spectra of $z=4.05$ GRB\,060206\altaffilmark{1}}

15: 

16: \author{

17: H.~Hao\altaffilmark{2}, K.~Z.~Stanek\altaffilmark{3},

18: A.~Dobrzycki\altaffilmark{4}, T.~Matheson\altaffilmark{5},

19: M.~C.~Bentz\altaffilmark{3}, J.~Kuraszkiewicz\altaffilmark{2},

20: P.~M.~Garnavich\altaffilmark{6}, J.~C.~Howk\altaffilmark{6},

21: M.~L.~Calkins\altaffilmark{7}, G.~Worthey\altaffilmark{8},

22: M.~Modjaz\altaffilmark{2}, J.~Serven\altaffilmark{8} }

23: 

24: 

25: \altaffiltext{1}{Based on data from the F.~L. Whipple Observatory,

26:   which is operated by the Smithsonian Astrophysical Observatory.}

27: 

28: \altaffiltext{2}{Harvard-Smithsonian Center for Astrophysics, 60

29: Garden Street, Cambridge, MA 02138; {hhao@cfa.harvard.edu},

30: {jkuraszkiewicz@cfa.harvard.edu}, {mmodjaz@cfa.harvard.edu}}

31: 

32: \altaffiltext{3}{Department of Astronomy, The Ohio State

33: University, 140 W. 18th Avenue, Columbus, OH 43210;

34: {kstanek@astronomy.ohio-state.edu},

35: {bentz@astronomy.ohio-state.edu}}

36: 

37: \altaffiltext{4}{European Southern Observatory,

38: Karl-Schwarzschild-Strasse 2, Garching, D-85748, Germany;

39: {adam.dobrzycki@eso.org}}

40: 

41: \altaffiltext{5}{National Optical Astronomy Observatory, 950 North

42: Cherry Avenue, Tucson, AZ 85719; {matheson@noao.edu}}

43: 

44: \altaffiltext{6}{Department of Physics, University of Notre Dame,

45: 225 Nieuwland Science Hall, Notre Dame, IN 46556;

46: {pgarnavi@nd.edu}, {jhowk@nd.edu}}

47: 

48: \altaffiltext{7}{F.~L.~Whipple Observatory, 670 Mt.~Hopkins Road,

49: P.O.~Box 97, Amado, AZ 85645; {mcalkins@cfa.harvard.edu}}

50: 

51: \altaffiltext{8}{Department of Physics and Astronomy, P.~O.~Box

52: 642814, Washington State University, Pullman, WA 99164;

53: {gworthey@wsu.edu}, {jdogg@wsu.edu}}

54: 

55: \begin{abstract}

56: 

57: We report on the discovery of strongly variable \ion{Fe}{2} and

58: \ion{Mg}{2} absorption lines seen at $z=1.48$ in the spectra of

59: the $z=4.05$ gamma-ray burst (GRB) 060206 obtained between 4.13 to

60: 7.63 hours (observer frame) after the burst.  In particular, the

61: \ion{Fe}{2} line equivalent width (EW) decayed rapidly from $1.72

62: \pm 0.25$~\AA\ to $0.28 \pm 0.21$~\AA, only to increase to $0.96

63: \pm 0.21$~\AA\ in a later spectrum. The \ion{Mg}{2} doublet shows

64: even more complicated evolution: the weaker line of the doublet

65: drops from $2.05\pm 0.25$~\AA\ to $0.92\pm 0.32$~\AA, but then

66: more than doubles to $2.47\pm 0.41$~\AA\ in later data.  The ratio

67: of the EWs for the \ion{Mg}{2} doublet is also variable, being

68: closer to $1\!:\!1$ (saturated regime) when the lines are stronger

69: and becoming closer to $2\!:\!1$ (unsaturated regime) when the

70: lines are weaker, consistent with expectations based on atomic

71: physics.  We have investigated and rejected the possibility of any

72: instrumental or atmospheric effects causing the observed strong

73: variations. Our discovery of clearly variable intervening

74: \ion{Fe}{2} and \ion{Mg}{2} lines immediately indicates that the

75: characteristic size of intervening patches of \ion{Mg}{2}

76: ``clouds'' is comparable to the GRB beam size, i.e., about

77: $10^{16}\;$cm. We discuss various implications of this discovery,

78: including the nature of the \ion{Mg}{2} absorbers, the physics of

79: GRBs, and measurements of chemical abundances from GRB and quasar

80: absorption lines.

81: 

82: \end{abstract}

83: 

84: \keywords{galaxies: ISM, gamma rays: bursts, intergalactic medium}

85: 

86: \section{Introduction}

87: 

88: The optical spectrum of a gamma-ray burst (GRB) afterglow is

89: generally a power-law with absorption lines superimposed.  These

90: absorption lines correspond to intervening material between the

91: source and the observer. Typically, the highest-redshift lines are

92: thought to be associated with the host galaxy \citep[e.g., see

93: discussion in][]{jha01}, and thus provide a measurement of the

94: redshift of the burst. There might be several lower-redshift

95: absorption lines caused by intervening systems.  Spectra of

96: luminous GRB afterglows have opened a new window on

97: absorption-line studies along lines of sight that are not

98: associated with quasars \citep[cf.][]{vreeswijk03}.

99: 

100: Recent studies of GRB absorption spectra have produced an

101: unexpected result. \cite{prochter06} reported statistically

102: significant evidence for a much higher incidence of strong

103: \ion{Mg}{2} absorbers along GRB sight lines than that expected

104: from studies of quasars of comparable redshift. They considered

105: and rejected three possible scenarios to produce such an effect,

106: namely dust associated with the \ion{Mg}{2} absorbers, ejection of

107: \ion{Mg}{2} by the GRB, and strong gravitational lensing by

108: galaxies associated with the emission.

109: 

110: \cite{frank06} proposed a simple geometric solution to the quandary,

111: with resulting predictions that are readily testable.  Namely, if the

112: \ion{Mg}{2} absorbers seen in both GRB and quasar spectra are truly

113: intervening, then the Prochter et al.\ result could be caused by

114: different sizes of the GRB and quasar emission regions.  If the

115: \ion{Mg}{2} absorbers have physical sizes comparable to the emitting

116: regions of quasars and GRBs, then a background source with a larger

117: emission region would ``see'' a smaller effective absorption column

118: density than a background source with a smaller emission region (see

119: Frank et al.\ for details of their proposed scenario).

120: 

121: One immediate prediction of the Frank et al.\ scenario is that for

122: strongly dynamic objects like GRB afterglows, the strength and

123: structure of intervening absorption lines in GRB spectra should vary

124: in some, if not all, cases. Motivated by this prediction, we have

125: examined our existing, multi-epoch spectroscopic data of GRB\,060206

126: and have discovered such a behavior.  In this paper, we provide the

127: first observational evidence of this prediction. In fact, as far as we

128: know this is the first report ever of clearly evolving intervening

129: (i.e., non-intrinsic) ground state absorption lines seen in spectra of

130: any cosmological object. There is evidence for variation in fine

131: structure lines in the GRB host environment \citep[cf.][]{vreeswijk06,

132: zavadsky06}.

133: 

134: GRB\,060206 triggered {\em Swift}-BAT on 2006 Feb 6 at 04:46:53 UT

135: \citep{morris2006}. A likely afterglow was identified by

136: \citet{fynbo06}, who also determined that the afterglow was at

137: high redshift ($z=4.048$).  The evolution of a bright and complex

138: afterglow was reported by several groups, and \cite{wozniak06},

139: \cite{stanek06}, and \cite{monfardini06} have published detailed

140: photometric data for this ``anomalous'' event. The burst also

141: showed complex behavior in the X-rays as seen with {\em Swift}-XRT

142: instrument, but the overall X-ray evolution follows that seen in

143: the optical \citep{stanek06}.

144: 

145: 

146: \section{Spectroscopic Observations}

147: 

148: The brightness of the optical transient (OT) associated with

149: GRB\,060206 allowed us to take spectra for several hours after the

150: burst.  Spectra of the OT were obtained with the FAST spectrograph

151: \citep{fabricant98} at the Cassegrain focus of the 1.5m Tillinghast

152: telescope at the F.~L. Whipple Observatory at 08:25, 09:02, 09:33,

153: 10:05, 10:41, 11:37, 12:09 (all times UT).  The OT was visible on the

154: telescope acquisition camera and could be identified and placed on the

155: spectrograph slit for the observations. Each spectrum was taken at or

156: near the parallactic angle \citep{filippenko82}. The spectra were

157: reduced in the standard manner with IRAF\footnote{IRAF is distributed

158: by the National Optical Astronomy Observatory, which is operated by

159: the Association of Universities for Research in Astronomy, Inc., under

160: cooperative agreement with the National Science Foundation.} using the

161: optimal extraction package of \citet{horne86}.  Night-sky emission

162: lines were removed by fitting a line to two sky regions on either side

163: of the OT column-by-column; the value determined from the linear

164: function at the position of the OT extraction was then subtracted.

165: Wavelength calibration was provided with an HeNeAr lamp taken

166: immediately after each OT spectrum.  Minor adjustments to the

167: wavelength solution were made based upon night-sky lines in the OT

168: frames. There were clouds during some of the exposure, but the

169: signal-to-noise ratio (S/N) is comparable for all the spectra. The

170: resolution of each spectrum is $~6.2$~\AA\, as measured from the FWHM

171: of night-sky emission lines.

172: 

173: We flux-calibrated the spectra with spectrophotometric standard

174: stars using our own software.  Relative fluxes in our OT spectrum

175: are accurate to $~4\%$ over the observed wavelength range.

176: Following the arguments of \cite{wade88} and \cite{matheson00},

177: telluric correction was accomplished by dividing the GRB spectra

178: with a normalized spectrum of the flux standard, which was set to

179: unity for regions unaffected by atmospheric features. For our

180: spectra, telluric correction was only applied over the regions

181: 6250-6360~\AA\ and 6840-7410~\AA. We scaled the strength of the

182: absorption by the difference of the airmasses to the $0.6$ power

183: to correct the saturated B band properly \citep{wade88}. Most of

184: the correction is $~2\%$ or less, with a $~5\%$ correction in the

185: 6250-6360~\AA\ region and $~25\%$ in the strong region of the B

186: band.

187: 

188: 

189: In the S/N weighted average spectrum, shown in Figure \ref{grb-all}, a

190: damped Lyman-alpha absorption system (DLA) is present at

191: $\lambda=6145$~\AA. A number of strong and weak absorption metal line

192: systems are superimposed on the continuum, identified as systems at

193: $z=1.48$ and $z=4.05$ \citep[reported by ][]{fynbo06}. The very strong

194: DLA system at $\lambda=6145$~\AA\ indicates that the other absorbers

195: at $z=4.05$ most likely represent the host galaxy of the GRB. In the

196: intervening metal line system at $z=1.48$, our data show absorption

197: from the transitions \ion{Fe}{2} $\lambda$2600 and \ion{Mg}{2}

198: $\lambda\lambda$2796, 2803. These are fairly typical in absorption

199: line systems \citep[e.g., ][]{steidel92}. Two absorption lines at

200: $\lambda=6331.3$~\AA\ and $6338.3$~\AA\ represent a \ion{Si}{4} system

201: at $z=3.54$, with a strong corresponding Ly$\alpha$ absorption line at

202: $5525.9$~\AA.

203: 

204: \section{Spectroscopic Monitoring}

205: 

206: Motivated by the prediction of Frank et al. (2006) that the

207: absorption lines along GRB sight lines should vary in some cases,

208: we examined our time series of GRB~060206 spectra to look for this

209: effect.  Variation of the intervening $z=1.48$ \ion{Fe}{2} and

210: \ion{Mg}{2} lines is obvious even upon casual inspection of the

211: spectra, as shown in Figure~\ref{compare}.  These lines are seen

212: to vary in strength, profile, and possibly velocity.  We have

213: checked the original 2-D spectra and no cosmic ray features are

214: apparent that would cause this variation. In addition, night-sky

215: emission lines and telluric features were carefully removed with

216: several variations in technique to ensure that reduction processes

217: did not introduce the observed variability. In fact, there are no

218: sky lines or telluric features at all near the strongly variable

219: $z=1.48$ \ion{Fe}{2}.

220: 

221: 

222: To measure the magnitude of variations in the equivalent widths

223: (EWs) of the $z=1.48$ absorption lines, we fit each of the

224: \ion{Mg}{2} lines with multiple Gaussians and the \ion{Fe}{2} line

225: with a single Gaussian. The error of the EWs are calculated from

226: the conservative fit uncertainties. The resulting EW measurements

227: are listed in Table~\ref{MgII} and plotted in Figure~\ref{MgIIf}.

228: The first three spectra show a gradual decrease in the EW of the

229: \ion{Mg}{2} $\lambda 2796$ line accompanied by a sharp decline in

230: the EW of the weaker \ion{Mg}{2} $\lambda 2803$ line.  As a

231: result, the ratio of the doublet increases from $1.2\!:\!1$ in the

232: first spectrum to about $2\!:\!1$ in the third.  This is followed

233: in the fourth spectrum by a sharp rise in the EW of both

234: components of the doublet and a sharp decrease in the doublet

235: ratio to $1.3:1$. While the EW of the \ion{Fe}{2} line also

236: decreases throughout the first three spectra, it seems to

237: completely disappear in the fourth spectrum. The two small

238: features in this spectrum at the location of the \ion{Fe}{2}

239: feature are consistent with noise (see Figure~\ref{compare}).  The

240: \ion{Fe}{2} absorption line gradually returns throughout the last

241: three spectra, but it never regains its original strength.  The

242: only other spectral feature to show possible variability is the

243: \ion{Si}{4} $\lambda 1393.76$ absorption line at the presumed

244: redshift of the host galaxy. The variability in this line does not

245: follow the behavior of either the \ion{Mg}{2} or \ion{Fe}{2}

246: systems at $z = 1.48$. We are still investigating the nature and

247: significance of variability in this line.

248: 

249: 

250: Variation in the EW of an absorption line is a direct consequence

251: of variation in the column density of the absorber as seen by the

252: background source.  The \ion{Fe}{2} and \ion{Mg}{2} systems are

253: not necessarily expected to be distributed homogeneously

254: throughout an absorbing ``cloud.''  As the beam size of the GRB

255: increases due to its physical expansion, the beam could include

256: various dense or dilute regions of the ``cloud,'' leading to

257: different behaviors in the measured EWs of the absorbers.

258: Additionally, the \ion{Fe}{2} and \ion{Mg}{2} may not be

259: co-spatial, so their behaviors may not be directly comparable as

260: seen here.

261: 

262: To estimate the statistical significance of these observed

263: variations, we apply $\chi^2$ statistics to the ``null

264: hypothesis'' that the \ion{Fe}{2} and \ion{Mg}{2} lines are not

265: variable.  Fitting the multiple measurements of the \ion{Fe}{2} EW

266: with a constant value gives $\chi^2=38.8$ with six degrees of

267: freedom.  Such a value of $\chi^2$ results in a probability of

268: $p=0.0000794\%$ ($\sim 5 \sigma$ Gaussian significance) that the

269: \ion{Fe}{2} line is not varying.  A similar analysis for the

270: \ion{Mg}{2} $\lambda 2803$ line results in $\chi^2 = 14.0$, which

271: gives $p = 2.93\%$ ($2.2 \sigma$ Gaussian significance) for lack

272: of variation in the line EW.  Finally, for the \ion{Mg}{2}

273: $\lambda 2796$ line, we find $\chi^2=7.4$, resulting in $p=28.9\%$

274: for the assumption that the EW does not vary. The combined data clearly show statistically

275: significant evidence for variability.  Note that while we only

276: test here the significance of variations in the line EWs, it is

277: also immediately apparent from the spectra that the line profiles

278: are varying as well (see Figure~\ref{compare}).  Each of these

279: lines is likely composed of several components that are blended in

280: our spectra, and it is variations in individual components that

281: are causing the observed variations in the total EWs and line

282: profiles.

283: 

284: To further investigate the variability of the lines, we use the

285: spectral subtraction algorithm developed by Hartman \& Stanek (2007,

286: in preparation), which uses the entire spectrum in the wavelength range

287: $6250\AA \sim 7200\AA$ to obtain the convolution kernel for matching individual

288: observed spectra to the reference spectrum. We choose the first

289: observed spectrum as the reference, then calculate the difference

290: between the matched spectra and the observed spectra for the rest of

291: the observed spectra, divided by the variance spectra we can calculate

292: the significance of the difference. The \ion{Fe}{2} shows a

293: $4.9\sigma$ variation in the line region $(6439\AA\sim6450\AA)$ and

294: the \ion{Mg}{2} $\lambda 2796$ line shows a $2.5\sigma$ variation in

295: the line region $(6926\AA\sim6942\AA)$.

296: 

297: \section{Discussion}

298: 

299: Variable intervening \ion{Fe}{2} and \ion{Mg}{2} was predicted by

300: \cite{frank06} as a direct consequence of their geometric solution to

301: the \ion{Mg}{2} absorption over-abundance problem discovered by

302: \cite{prochter06}. Our result immediately provides constraints on the

303: physical size of the \ion{Mg}{2} absorbers that are several orders of

304: magnitude stronger than those determined from studies of lensed

305: quasars --- $\sim 0.01$~pc \citep{frank06} vs. $< 200-300$~pc

306: \citep{rauch02, petitjean00}. Observations of lensed quasars such as

307: those proposed by \cite{dong06} may provide additional constraints on

308: the physical sizes of the absorbers.  Significant spectral variability

309: on the timescales of months to years induced by microlensing can be

310: used to probe the $10^{14}-10^{16}\;$cm cloudlets along the sightlines

311: of lensed quasars. Properties of the cloudlets (physical size,

312: sub-structure, etc.) can be studied statistically if such spectral

313: variations are seen.

314: 

315: We can also investigate the average physical density of the

316: absorbers. Assuming the ``cloud'' is optically thin, we can obtain

317: a lower limit on the \ion{Mg}{2} column density $N_{\rm MgII}=

318: 1.7\times10^{13}\;{\rm{cm}}^{-2}$. To evaluate this result, we

319: have used our third spectrum taken $\sim 5\;$hours after the

320: burst, where the \ion{Mg}{2} line ratio is close to $2\!:\!1$,

321: indicating the lines are not strongly saturated (see

322: Table~\ref{MgII}). The beam size of the GRB during that epoch can

323: be estimated \citep{waxman97,loeb98} via

324: \begin{equation}

325: R_s(t) = 4.1\times10^{15}\left(\frac{E_{52}}{n_1}\right)^{\frac 1

326: 8}

327:          (1+z_s)^{-5/8}(t/hour)^{5/8}\; {\rm{cm}}

328: \end{equation}

329: where $z_s$ is the source redshift, $E_{52}$ is the

330: ``isotropic-equivalent'' of the energy release in units of

331: $10^{52}~{\rm erg~s^{-1}}$, and $n_1$ is the ambient gas density in

332: units of ~$1~{\rm cm^{-3}}$. For GRB~060206, $E_{52}=5.8$

333: \citep{palmer06}, and assuming $n_1 = 1\;{\rm{cm}}^{-3}$, then

334: $R_s=5.1\times 10^{15}\;$cm for our third spectrum. With the physical

335: size of the individual \ion{Mg}{2} ``clumps'' constrained to be

336: comparable to the GRB beam size (i.e., $\sim 10^{16}\;$cm), one can

337: immediately derive the average physical density of \ion{Mg}{2} atoms

338: to be about $0.0015~\rm{cm}^{-3}$. This is the first time that such a

339: strong constraint has been obtained. Given its large \ion{Mg}{2}

340: equivalent width, the z=1.48 absorber is likely a Lyman limit system

341: with $\log{N(HI)} > 18$, although there is a $\sim 20-30$\%

342: probability that it is a damped Lyman-alpha system (Rao et

343: al. 2006). Then the physical density of the hydrogen of this cloud is

344: $>200~\rm{cm}^{-3}$. The estimate of physical density of \ion{Mg}{2}

345: absorbers is very important for understanding the cloud origins and

346: properties, and more generally the metal lines and metal enrichment

347: (see discussion in Schaye et al 2007).

348: 

349: Another consequence of this discovery relates to GRBs themselves.  In

350: the case of GRB\,060206, we see the \ion{Mg}{2} EW initially

351: decreasing, but then it shows a dramatic increase.  This behavior is

352: either a result of the GRB beam becoming much smaller in a short time

353: (possibly due to an episode of energy injection) or part of the

354: afterglow ring suddenly brightening. Our discovery points the way to a

355: new method for studying GRB phenomena by monitoring afterglows with

356: high-resolution optical spectroscopy.

357: 

358: Additional implications relate to chemical abundances for intervening

359: systems as measured using absorption lines. In our first spectrum, the

360: ratio of EWs for \ion{Fe}{2} compared to \ion{Mg}{2} is about 0.7.  In

361: our fourth spectrum, where the \ion{Fe}{2} is only marginally present,

362: and \ion{Mg}{2} is very strong, that ratio is less than 0.1. Estimates

363: of relative abundances for these two elements would be dramatically

364: different when analyzing these two low-resolution spectra. If the

365: system is at the lower end of this column density range$\log{N(HI)} >

366: 18$, some of the difference in \ion{Mg}{2}/\ion{Fe}{2} ratio could be

367: the result of spatially-varying ionization levels.  However, it may be

368: that there are intrinsic variations in the relative metal abundances

369: (although such variations are not seen in damped Lyman-alpha systems:

370: Prochaska 2003, Rodriguez et al 2006).

371: 

372: The observations presented in this paper open up the exciting

373: possibility of studying time-variability of intervening absorption

374: lines in the spectra of cosmological objects. Strongly and rapidly

375: variable objects, such as GRBs and blazars, are obvious targets for

376: further studies, but quasars (including both lensed and non-lensed)

377: should also be investigated. Variable narrow absorption lines have

378: been observed in spectra of quasars (e.g., Hamann et al. 1997), but

379: they have been interpreted as originating in gas physically associated

380: with the quasar. In fact, variability of narrow absorption lines is

381: now used as one of the criteria to select intrinsic absorbers

382: associated with quasars.  Results presented in this paper suggest that

383: using this criterion should be re-examined---quasars are optically

384: variable sources, so size of their continuum emission region is bound

385: to vary, just like it does for GRBs. As already discussed by Frank et

386: al. (2006), quasars with smaller continuum emitting regions (i.e.,

387: intrinsically fainter ones), should have higher incidence of the

388: strong \ion{Mg}{2} absorbers.

389: 

390: Note that throughout this paper we considered the $z=1.48$ variable

391: absorption system not to be physically associated with the $z=4.05$

392: GRB, i.e., to be truly intervening due to cosmological expansion. For

393: these two objects to be physically associated would require

394: \ion{Fe}{2} and \ion{Mg}{2} absorbing gas to have ejection velocity

395: $v\sim 0.6\;$c, while maintaining relatively low temperature in order

396: not to be completely ionized.  We consider such scenario most

397: unlikely. Detection of a galaxy associated with this particular

398: $z=1.48$ variable absorption system would settle this issue beyond any

399: reasonable doubt.  As discussed in Prochter et al. (2006), such

400: foreground galaxies were already identified for a number of strong,

401: intervening \ion{Mg}{2} absorbers seen in GRB spectra.

402: 

403: \acknowledgments

404: 

405: We thank J.~Hartman, P.~Martini, S.~Gaudi, S.~Mathur, R.~Pogge, S.~Frank,

406: X.~Dai, J.~Prieto, D.~An, M.~Elvis, R.~Kirshner, and R.~Narayan

407: for helpful comments and discussions.  M.~C.~Bentz is supported by

408: a National Science Foundation Graduate Fellowship.

409: 

410: \clearpage

411: 

412: \begin{thebibliography}{}

413: 

414: \bibitem[Dessauges-Zavadsky et al (2006)]{zavadsky06}

415: Dessauges-Zavadsky, M., et al.~2006, \apj, 648, L89

416: 

417: \bibitem[Dong (2006)]{dong06} Dong, S. 2006, \apj, submitted (astro-ph/0612403)

418: 

419: \bibitem[Fabricant et al.~(1998)]{fabricant98} Fabricant, D., et al.~1998, \pasp,110,79

420: 

421: \bibitem[Filippenko (1982)]{filippenko82} Filippenko, A. V.~1982, \pasp, 94, 715

422: 

423: \bibitem[Frank et al.~(2006)]{frank06} Frank, S., et al.~2006,

424: \apj, submitted (astro-ph/0605676)

425: 

426: \bibitem[Fynbo et al.~(2006)]{fynbo06} Fynbo, J.~P.~U., et al.~2006,

427:  \aap, 451, L47

428: 

429: \bibitem[Hamann et al.~(1997)]{hamann97} Hamann, F., Barlow, T., A., \&

430: Junkkarinen, V. 1997, \apj, 478, 87

431: 

432: \bibitem[Horne (1986)]{horne86} Horne, K. 1986, \pasp, 98, 609

433: 

434: \bibitem[Jha et al.~(2001)]{jha01} Jha, S., et al.~2001,

435: \apj, 554, L155

436: 

437: \bibitem[Loeb \& Perna (1998)]{loeb98} Loeb, A., \& Perna, R. 1998,

438: \apj, 495, 597

439: 

440: \bibitem[Matheson et al.~(2000)]{matheson00} Matheson, T.,~et al.~2000, \aj, 120, 1499

441: 

442: \bibitem[Monfardini et al.~(2006)]{monfardini06} Monfardini, A., et al.~2006, \apj, 648, 1125

443: 

444: \bibitem[Morris et al.~(2006)]{morris2006} Morris, D.~C., et al.~2006,

445: GCN 4682

446: 

447: \bibitem[Palmer et al.~(2006)]{palmer06} Palmer, D., et al. 2006, GCN Circ. 4697

448: 

449: \bibitem[cf. Petitjean et al.~(2000)]{petitjean00} Petitjean, P., et

450: al. 2000, \aap, 359, 457

451: 

452: \bibitem[Prochaska (2003)]{prochaska03} Prochaska, J. X. 2003,  \apj,  582,  49

453: 

454: \bibitem[Prochter et al.~(2006)]{prochter06} Prochter, G. E., et al.~2006, \apj, 648, L93

455: 

456: \bibitem[Rao et al.~(2006)]{rao06} Rao, S. M., Turnshek, D. A., \&

457: Nestor, D. B. 2006, \apj, 636, 610

458: 

459: \bibitem[Rauch et al.~(2002)]{rauch02} Rauch, M., Sargent, W. L. W.,

460: Barlow, T. A., \& Simcoe, R.A. 2002, \apj, 576, 45

461: 

462: \bibitem[Rodriguez et al.~(2006)]{rodriguez06} Rodriguez, E., et

463: al. ~2006, A\&A, 446, 791

464: 

465: \bibitem[Schaye et al.~(2007)]{schaye07} Schaye, J., Carswell, R.

466: F., \& Kim, T. S. MNRAS submitted (astro-ph/0701761)

467: 

468: \bibitem[Stanek et al.~(2006)]{stanek06} Stanek, K.~Z., et al.

469: 2006, \apjl, in press (astro-ph/0602495)

470: 

471: \bibitem[Steidel \& Sargent (1992)]{steidel92} Steidel,~C.~C., \&

472: Sargent, W. L. W. 1992, \apjs, 80,1

473: 

474: \bibitem[Vreeswijk et~al.(2003)]{vreeswijk03} Vreeswijk, P.~M.,

475: M{\o}ller, P., \& Fynbo, J. P. U. 2003, \aap, 409, L5

476: 

477: \bibitem[Vreeswijk et~al.(2006)]{vreeswijk06} Vreeswijk, P.~M., et

478: al. 2006, \aap, submitted (astro-ph/0611690)

479: 

480: \bibitem[Wade \& Horne(1988)]{wade88} Wade, R.~A., \& Horne, K.~D.,

481: 1988, \apj, 324, 411

482: 

483: \bibitem[Waxman (1997)]{waxman97} Waxman, E. 1997, \apj, 491, L19

484: 

485: \bibitem[Wozniak et al.(2006)]{wozniak06} Wozniak, P. R., et al.~2006,

486: \apj, 642, L99

487: 

488: \end{thebibliography}

489: 

490: \clearpage

491: 

492: \begin{deluxetable}{llccc}

493: \tablewidth{0pt} \tablenum{1} \tablecaption{EVOLUTION OF

494: EQUIVALENT WIDTH OF THE $z=1.48$ ABSORPTION LINES \label{MgII}}

495: \tablehead{\colhead{$t_{obs}$\tablenotemark{a}} &

496: \colhead{Age\tablenotemark{b}} &

497: \colhead{EW\tablenotemark{c}$_{\rm{Mg~II}~2796}$} &

498: \colhead{EW$_{\rm{Mg~II}~2803}$} &

499: \colhead{EW$_{\rm{Fe~II}~2600}$} } \startdata

500:  2.86384 & 4.13 & 2.53 $\pm$ 0.23 & 2.05 $\pm$ 0.25 & 1.72 $\pm$ 0.25 \\

501:  2.88946 & 4.50 & 2.14 $\pm$ 0.25 & 1.50 $\pm$ 0.27 & 1.30 $\pm$ 0.15 \\

502:  2.91110 & 5.02 & 2.06 $\pm$ 0.30 & 0.92 $\pm$ 0.33 & 0.60 $\pm$ 0.20 \\

503:  2.93323 & 5.55 & 3.14 $\pm$ 0.44 & 2.47 $\pm$ 0.41 & 0.28 $\pm$ 0.21 \\

504:  2.95857 & 6.15 & 2.24 $\pm$ 0.29 & 1.93 $\pm$ 0.34 & 0.36 $\pm$ 0.17 \\

505:  2.99745 & 7.09 & 1.90 $\pm$ 0.44 & 2.34 $\pm$ 0.58 & 0.96 $\pm$ 0.29 \\

506:  3.01990 & 7.63 & 1.96 $\pm$ 0.35 & 1.41 $\pm$ 0.42 & 0.67 $\pm$ 0.40

507: \enddata

508: \tablenotetext{a}{Heliocentric Julian Date at middle of exposure

509: minus 2,453,770.}

510: 

511: \tablenotetext{b}{Age of the middle of exposure in hours from the

512: detection of the burst at 2006 Feb 06 04:46:53 UT

513: \citep{morris2006}.}

514: 

515: \tablenotetext{c}{Observer-frame equivalent width.}.

516: 

517: \end{deluxetable}

518: 

519: \clearpage

520: 

521: \begin{figure}

522: \plotone{f1.eps} \caption{Spectra of the OT associated with

523: GRB\,060206. Prominent absorption features associated with three

524: redshift systems are indicated: $z = 1.48$ (\emph{black long

525: dashed vertical lines}); $z = 4.05$ (\emph{red solid vertical

526: lines}); $z=3.54$ (\emph{blue short dashed vertical lines}).

527: \label{grb-all}}

528: \end{figure}

529: 

530: \clearpage

531: 

532: \begin{figure}

533: \plotone{f2.eps}

534: \caption{Spectra of the GRB\,060206 taken at 4.13

535: hours, 5.02 hours

536:          and 5.55 hours after the burst trigger in the $z=1.48$

537:          \ion{Fe}{2} and \ion{Mg}{2} region. Each spectrum has been

538:          offset by a constant flux value for clarity.

539: \label{compare}}

540: \end{figure}

541: 

542: \clearpage

543: 

544: \begin{figure}

545: \plotone{f3.eps}

546: \caption{Time evolution of the observed EW of the

547: absorption lines of \ion{Mg}{2} $\lambda$2796 (\emph{filled

548: circles}) \ion{Mg}{2} $\lambda$2803 (\emph{open circles}), and

549: \ion{Fe}{2} $\lambda$2600 (\emph{filled squares}). The EW

550: measurements for the different lines have been slightly offset in

551: time for clarity.\label{MgIIf}}

552: \end{figure}

553: 

554: \end{document}

555: