1: %%

2: %% Beginning of file 'sample.tex'

3: %%

4: %% Modified 2004 January 9

5: %%

6: %% This is a sample manuscript marked up using the

7: %% AASTeX v5.x LaTeX 2e macros.

8: 

9: %% The first piece of markup in an AASTeX v5.x document

10: %% is the \documentclass command. LaTeX will ignore

11: %% any data that comes before this command.

12: 

13: %% The command below calls the preprint style

14: %% which will produce a one-column, single-spaced document.

15: %% Examples of commands for other substyles follow. Use

16: %% whichever is most appropriate for your purposes.

17: %%

18: 

19: %% define

20: \def\eiso{E_{\rm iso}}

21: \def\egamma{E_{\gamma}}

22: \def\ep{E_{\rm peak}}

23: \def\epo{E^{\rm obs}_{\rm peak}}

24: \def\eps{E^{\rm src}_{\rm peak}}

25: \def\lpiso{L^{\rm peak}_{\rm iso}}

26: 

27: %% v2 (12/01/06)

28: %% Joe and Ryo's comments are implemented.  

29: %%

30: %% v3 (12/05/06)

31: %% Comments from Joe, ..

32: %%

33: %% v4 (12/??/06)

34: %% circulate to Co-Is

35: %%

36: %% v5 (12/22/06)

37: %% Fixed the XRT analysis part (Joe), 

38: %% Finalized the discussion part (Taka & Ryo)

39: %% Fixed the comments from internal reviewers 

40: %%

41: %% v6 (03/20/07)

42: %% Revised draft including referee's comments

43: %% - Liang's curvature fit (ref comment 5)

44: %% - Add paragraph about disadvantage of our approach in Ep<15 keV case 

45: %%    (ref comment 1)

46: %% - Correct ref comment 2-1, 2-2

47: %% - Correct ref comment 4

48: %% 

49: %% v9 (06/18/07)

50: %% Joe's correction of English

51: 

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

53: 

54: %% manuscript produces a one-column, double-spaced document:

55: 

56: %% \documentclass[manuscript]{aastex}

57: 

58: %% preprint2 produces a double-column, single-spaced document:

59: 

60: %% \documentclass[preprint2]{aastex}

61: 

62: %% Sometimes a paper's abstract is too long to fit on the

63: %% title page in preprint2 mode. When that is the case,

64: %% use the longabstract style option.

65: 

66: %% \documentclass[preprint2,longabstract]{aastex}

67: 

68: %% If you want to create your own macros, you can do so

69: %% using \newcommand. Your macros should appear before

70: %% the \begin{document} command.

71: %%

72: %% If you are submitting to a journal that translates manuscripts

73: %% into SGML, you need to follow certain guidelines when preparing

74: %% your macros. See the AASTeX v5.x Author Guide

75: %% for information.

76: 

77: \newcommand{\vdag}{(v)^\dagger}

78: \newcommand{\myemail}{takanori@milkyway.gsfc.nasa.gov}

79: 

80: %% You can insert a short comment on the title page using the command below.

81: 

82: \slugcomment{Not to appear in Nonlearned J., 45.}

83: 

84: %% If you wish, you may supply running head information, although

85: %% this information may be modified by the editorial offices.

86: %% The left head contains a list of authors,

87: %% usually a maximum of three (otherwise use et al.).  The right

88: %% head is a modified title of up to roughly 44 characters.

89: %% Running heads will not print in the manuscript style.

90: 

91: \shorttitle{Gamma-Ray;burst}

92: \shortauthors{Sakamoto et al.}

93: 

94: %% This is the end of the preamble.  Indicate the beginning of the

95: %% paper itself with \begin{document}.

96: 

97: \begin{document}

98: 

99: %% LaTeX will automatically break titles if they run longer than

100: %% one line. However, you may use \\ to force a line break if

101: %% you desire.

102: 

103: \title{Evidence of 

104: %the 

105: Exponential Decay Emission \\in the {\it \bfseries Swift}  Gamma-ray Bursts}

106: 

107: %% Use \author, \affil, and the \and command to format

108: %% author and affiliation information.

109: %% Note that \email has replaced the old \authoremail command

110: %% from AASTeX v4.0. You can use \email to mark an email address

111: %% anywhere in the paper, not just in the front matter.

112: %% As in the title, use \\ to force line breaks.

113: 

114: \author{T. Sakamoto\altaffilmark{1,2}, 

115: J. E. Hill\altaffilmark{1,3,4},

116: R. Yamazaki\altaffilmark{5},

117: L. Angelini\altaffilmark{1},

118: H. A. Krimm\altaffilmark{1,3,4},

119: G. Sato\altaffilmark{1,6},\\

120: S. Swindell\altaffilmark{7}, 

121: K. Takami\altaffilmark{5}, 

122: J. P. Osborne\altaffilmark{8}}

123: %% BAT team member..

124: %L. Barbier\altaffilmark{1}, 

125: %S. D. Barthelmy\altaffilmark{1}, 

126: %J. R. Cummings\altaffilmark{1,2}, 

127: %E. E. Fenimore\altaffilmark{3},

128: %N. Gehrels\altaffilmark{1}, 

129: %D. Hullinger\altaffilmark{4}, 

130: %C. B. Markwardt\altaffilmark{1,4},

131: %D. M. Palmer\altaffilmark{3},

132: %A. M. Parsons\altaffilmark{1},

133: %J. Tueller\altaffilmark{1},

134: 

135: \altaffiltext{1}{NASA Goddard Space Flight Center, Greenbelt, MD 20771}

136: \altaffiltext{2}{Oak Ridge Associated Universities, P.O. Box 117, 

137:  Oak Ridge, Tennessee 37831-0117}

138:  \altaffiltext{3}{CRESST NASA Goddard Space Flight Center, Greenbelt, MD 20771}

139: \altaffiltext{4}{Universities Space Research Association, 10211 Wincopin 

140: 	Circle, Suite 500, Columbia, MD 21044-3432} 

141: \altaffiltext{5}{Department of Physics, Hiroshima University, Higashi-Hiroshima, 

142: 	Hiroshima, 739-8526, Japan}

143: \altaffiltext{6}{Institute of Space and Astronautical Science, 

144: JAXA, Kanagawa 229-8510, Japan}

145: \altaffiltext{7}{Department of Physics, North Carolina Agricultural

146: and Technical State University, 1601 East Market Street,Greensboro, North Carolina 27411}

147: \altaffiltext{8}{Department of Physics and Astronomy, University of

148: Leicester, LE1, 7RH, UK}

149: %%% 

150: %\altaffiltext{3}{Los Alamos National Laboratory, P.O. Box 1663, Los

151: %Alamos, NM, 87545}

152: %\altaffiltext{4}{Department of Physics, University of Maryland, 

153: %	College Park, MD 20742}

154: 

155: %% Mark off your abstract in the ``abstract'' environment. In the manuscript

156: %% style, abstract will output a Received/Accepted line after the

157: %% title and affiliation information. No date will appear since the author

158: %% does not have this information. The dates will be filled in by the

159: %% editorial office after submission.

160: 

161: %% Start writing the draft (10/14/06)

162: %% T. Sakamoto

163: %%

164: %% v3 send for Swift internal review (12/11/06)

165: %% T. Sakamoto

166: %%

167: %% v4 impliment the comments for internal reviewers (12/20/06)

168: %% T. Sakamoto

169: %% 

170: %% v5 modified the discussion part (12/21/06)

171: %%    final modification before submission (12/26/06)

172: %% T. Sakamoto

173: 

174: \begin{abstract}

175: 

176: We present a systematic study of the steep decay emission 

177: from gamma-ray bursts (GRBs) observed by the {\it Swift} X-Ray 

178: Telescope (XRT).  In contrast to the analysis described in recent 

179: literature, we produce composite Burst Alert Telescope (BAT) and 

180: XRT light curves by extrapolating the XRT data (2--10 keV) into 

181: the BAT energy range (15--25 keV) rather than extrapolating the 

182: BAT data into the XRT energy band (0.3--10 keV).  Based on the fits 

183: to the composite light curves, we have confirmed the existence of 

184: an exponential decay component which smoothly connects the BAT 

185: prompt data to the XRT steep decay for several GRBs.  We also 

186: find that the XRT steep decay for some of the bursts can be well 

187: fit by a combination of a power-law with an exponential decay model.  

188: We discuss this exponential component within the frame work of both 

189: the internal and the external shock model.  

190: 

191: \end{abstract}

192: 

193: %% Keywords should appear after the \end{abstract} command. The uncommented

194: %% example has been keyed in ApJ style. See the instructions to authors

195: %% for the journal to which you are submitting your paper to determine

196: %% what keyword punctuation is appropriate.

197: 

198: %% Authors who wish to have the most important objects in their paper

199: %% linked in the electronic edition to a data center may do so in the

200: %% subject header.  Objects should be in the appropriate "individual"

201: %% headers (e.g. quasars: individual, stars: individual, etc.) with the

202: %% additional provision that the total number of headers, including each

203: %% individual object, not exceed six.  The \objectname{} macro, and its

204: %% alias \object{}, is used to mark each object.  The macro takes the object

205: %% name as its primary argument.  This name will appear in the paper

206: %% and serve as the link's anchor in the electronic edition if the name

207: %% is recognized by the data centers.  The macro also takes an optional

208: %% argument in parentheses in cases where the data center identification

209: %% differs from what is to be printed in the paper.

210: 

211: \keywords{Gamma-ray Burst}

212: 

213: %% Plan

214: %% 1. Introduction

215: %%    - GRB tail emission (BATSE tail, XRT steep decay (O'brian, Nousek), 

216: %%	curveture effect (Kumar, Liang))

217: %%    - Systematic problem in BAT extrapolation of 0.5-10 keV band.  

218: %%    - Our approach: XRT -> 15-25 keV band 

219: %% 2.1 BAT data analysis

220: %%    - HEASOFT

221: %% 2.2 XRT data analysis

222: %%    - Joe will write...

223: %% 2.3 Our sample

224: %%     - Selection criteria

225: %% 3. Results 

226: %% 

227: %% 4. Discussion

228: %% 

229: %% Table 

230: %% - BAT-XRT decay index

231: %% - XRT Spectral analysis

232: %% - BAT spectral analysis (is this needed?)

233: %% Figure 

234: %% - XRT-BAT light curve in 15-25 keV (middle pannel shows the XRT hardness ratio, 

235: %%   bottom panel shows the residual) 

236: %%   for each GRBs

237: %% - Histogram of decay index (T0@last peak, T0@Yamazaki def., T0@free)

238: 

239: \section{Introduction}

240: 

241: The transition between the gamma-ray burst (GRB) prompt emission and 

242: afterglow emission has generated great interest within the scientific 

243: community.  It is generally accepted that the GRB prompt

244: emission is due to internal shocks originating from the collision 

245: of faster and slower moving shells, whereas, the afterglow is believed 

246: to originate from an external shock resulting from 

247: the relativistic fireball colliding with a circum-burst medium 

248: \citep{rm1994,sp1997,mr1997}.  During the GRB episode there should be a

249: transition from one phase to the other, however, it is still 

250: not well understood as to when this transition occurs.  

251: %

252: The Burst and Transient Source Experiment (BATSE) observation 

253: of GRB 980923 showed long lasting tail emission, $\sim$400 s, which 

254: is best described by a power-law temporal decay \citep{giblin1999}.  

255: Based on the spectral and temporal characteristics of this burst, 

256: \citet{giblin1999} concluded that the tail emission was a part of 

257: the afterglow emission, thus, the external shock could be generated 

258: during the prompt $\gamma$-ray phase.  \citet{giblin2002} performed 

259: a systematic study of the prompt tail emission using 40 GRBs observed 

260: by BATSE.  They found that the temporal decays are best described by a 

261: power-law with a decay index of $-2$ rather than an exponential.  

262: There are several other analysis of BATSE data which reach the same 

263: conclusion \citep[e.g.,][]{rs2002}.  

264: %

265: According to the BeppoSAX observations, the late time afterglow smoothly 

266: connects with the prompt emission if the onset time of the light curve 

267: is defined as the start time of the last pulse observed in the Wide 

268: Field Camera (2--30 keV) \citep[e.g.,][]{pian2001,piro2005}.  

269: These observations support the idea that the late X-ray spike represents 

270: the onset of the afterglow.  However, the delay of a few hours to a 

271: few days before the narrow field X-ray instrument is pointed to the 

272: GRB position, weakens the discussion concerning the transition from 

273: the prompt emission to the afterglow.  

274: 

275: As a result of the revolutionary features of {\it Swift} 

276: \citep{gehrels2004}, our understanding of the X-ray properties of 

277: GRBs has been improved dramatically.  With the combination of the 

278: accurate on-board calculation of the GRB position by the Burst Alert Telescope 

279: \citep[BAT:15-150 keV;][]{barthelmy2005} and the fast slewing capability 

280: of the spacecraft, {\it Swift} can begin a highly sensitive X-ray 

281: observation with the X-Ray Telescope \citep[XRT:0.2-10 keV;][]{burrows2005} 

282: within a few tens of seconds to a few hundred seconds after the 

283: burst trigger.  According to the XRT observations, the X-ray properties 

284: of the GRB emission have very complex features

285: \citep{nousek2006,zhang2006a}.  One of the most unexpected discoveries 

286: by the XRT is the existence of the steep decay component during the 

287: initial phase of the X-ray light curve.  The origin of this steep decay 

288: component is generally considered to be a result of the delayed 

289: prompt emission from different viewing latitudes of the jet, the so 

290: called ``curvature effect'' \citep[e.g.][]{fenimore1996,kumar2000,zhang2006a}.  

291: \citet{tagliaferri2005} and \citet{barthelmy2006} 

292: investigated the steep decay component with the composite BAT and XRT 

293: light curves for several GRBs.  To generate the composite light curve, 

294: both papers performed an extrapolation of the BAT mask-weighted 

295: (background subtracted) light curve into the XRT 0.2--10 keV energy band 

296: using a best-fit power-law photon index.  The authors found that GRB 

297: 050126 and GRB 050219A do not show continuous emission from the BAT to 

298: the XRT light curve, however, GRB 050315 and GRB 050319 do display 

299: a smooth continuation from the BAT to the XRT light curve. 

300: \citet{obrien2006} performed a systematic study of the early X-ray emission 

301: using a sample of 40 {\it Swift} GRBs.  They constructed a composite BAT 

302: and XRT light curve in the 0.3--10 keV band.  In order to extrapolate 

303: the BAT data points, the BAT mask-weighted count rate was converted to flux 

304: in the 0.3--10 keV band using the mean of the best-fit photon indices 

305: obtained from a simple power-law fit to both the BAT and the XRT spectrum.  

306: The authors found that a fit of the 40 superimposed GRB light curves 

307: in the 0.3-10 keV band could be described by an exponential decay 

308: followed by a power-law decay.  \citet{willingale2007} investigated the 

309: X-ray emission including the late time light curve data.  

310: They found that the X-ray light curve data can be modeled with the 

311: superposition of an early (``prompt'') and a late time (``afterglow'')

312: component.  

313: 

314: In this paper, we describe the analysis 

315: of the tail emission using an alternative approach. 

316: We generate the composite BAT and XRT light curve in the 15--25 keV BAT 

317: energy band by extrapolating the 2--10 keV XRT count rate into the 

318: 15-25 keV band.  Either approach, extrapolating the BAT count rate down 

319: to the XRT band (hereafter BAT-to-XRT extrapolation) or extrapolating 

320: the XRT count rate up to the BAT band (hereafter XRT-to-BAT extrapolation), 

321: can encounter similar systematic problems.  Most of the GRB prompt 

322: emission spectra is well fit by the Band function \citep{band1993} with 

323: a low-energy and a high-energy photon index of $\sim 1$ and $\sim 2.3$, 

324: respectively \citep[e.g.][]{sakamoto2005,kaneko2006}.  However, it is a 

325: well known characteristic that the peak energy of the spectrum, $\ep$, 

326: not only evolves during the burst \citep[e.g.,][]{lloyd2002,frontera1999} 

327: but also changes from burst-to-burst \citep[e.g.,][]{sakamoto2005}.  

328: Here, we discuss the issues with regard to both types of extrapolation

329: by considering five cases depending on the value of $\ep$ 

330: (shown in figure \ref{fig:prob_extrapolation_bat_xrt} left is the BAT-to-XRT 

331: extrapolation and right is the XRT-to-BAT extrapolation).  If we 

332: consider the energy bands of the BAT (15-150 keV) and the XRT (0.3-10 keV), 

333: the five cases 

334: are defined by: 1) $\ep > 150$ keV, 2) 15 keV $< \ep <$ 150 keV, 

335: 3) 10 keV $< \ep <$ 15 keV, 4) 0.3 keV $< \ep <$ 10 keV, and 5) $\ep <$ 0.3 keV.  

336: For cases 1 and 5, both the BAT-to-XRT and the XRT-to-BAT extrapolations 

337: should provide the correct flux because the photon index of the

338: extrapolated energy band is the same as the observed energy band 

339: (``1'' and ``5'' of the left and the right panels of figure 

340: \ref{fig:prob_extrapolation_bat_xrt}).  

341: For cases 3 and 4, the BAT-to-XRT extrapolation would over estimate 

342: the flux in the XRT energy band, since the photon index in the observed 

343: energy band is steeper 

344: than that of the extrapolated energy band (``3'' and ``4'' of the 

345: left panel of figure \ref{fig:prob_extrapolation_bat_xrt}).  

346: Similarly, for case 2 and 3, the XRT-to-BAT extrapolation would 

347: over estimate the flux in the BAT energy band, because the photon 

348: index in the observed energy band is shallower than that of the 

349: extrapolated energy band (``2'' and ``3'' of the right panel of figure 

350: \ref{fig:prob_extrapolation_bat_xrt}).  

351: For case 2, using the BAT-to-XRT extrapolation (``2'' of the left 

352: panel of figure \ref{fig:prob_extrapolation_bat_xrt}), and for case 4, 

353: using the XRT-to-BAT  extrapolation (``4'' of the right panel of figure 

354: \ref{fig:prob_extrapolation_bat_xrt}), 

355: the flux would be overestimated in the XRT and the BAT energy band respectively, 

356: because the simple power-law fit is 

357: a tangential line to the curved spectrum as a result of the narrow 

358: energy bands of the BAT and the XRT (e.g. Sakamoto et al. in

359: preparation).   

360: 

361: To date, most of the BAT and XRT composite light curves in the 

362: literature have been produced from a BAT-to-XRT extrapolation.  

363: However, we believe that the XRT-to-BAT extrapolation described here  

364: may minimize the systematic errors, especially with respect to 

365: investigating GRB tail emission, for the following three reasons.  

366: First, we can account for the spectral evolution during the prompt 

367: emission if we extract the flux from the time-resolved spectral 

368: analysis of the BAT data.  As previously mentioned, $\ep$ shifts 

369: from hundreds of keV to a few keV during the prompt emission.  If in 

370: the analysis, one does not account for the spectral evolution 

371: during the prompt emission, which is the case for the majority of the 

372: published XRT and BAT composite light curves using the BAT-to-XRT extrapolation, 

373: the systematic error in the extrapolated flux in the X-ray range 

374: could be significant.  Second, since the BAT mask-weighted 

375: count rate does not correct for the energy dependence of each photon, 

376: the count rate of the source in the off-axis direction will be 

377: systematically smaller than the on-axis case.  This effect becomes 

378: an issue when the BAT-to-XRT extrapolation has been performed by 

379: converting the BAT mask-weighted count-rate into flux using a fixed 

380: photon index obtained from a simple power-law fit.  According to the 

381: BAT Crab observation, the count rate from the Crab is $\sim$15\% smaller 

382: in the 45 degree off-axis case.  This off-axis effect is correctly 

383: taken into account in the BAT energy response matrix, but not in 

384: the BAT mask-weighted count rates.  Therefore, unless one applies an

385: additional off-axis correction to the BAT mask-weighted count rates, a 

386: systematically smaller flux is obtained if the source is in the off-axis

387: direction, which is always the case for the BAT GRB data prior to the 

388: spacecraft slewing to the GRB position.  Third, according to the 

389: GRB synchrotron shock model \citep{sari1998b}, if the observed spectrum 

390: has a photon index steeper than $2$, and a power-law index of an electron 

391: distribution, $p$, in the range of 2 $\le$ $p$ $<$ 3 (where

392: N$(\gamma_{e}) \propto \gamma_{e}^{-p}$, and $\gamma_{e}$ is the 

393: Lorentz factor of the electrons), then the observed frequency should 

394: be above the synchrotron critical frequency for electrons with a minimum 

395: Lorentz factor ($\nu_{m}$) in the fast cooling phase or above the 

396: cooling frequency ($\nu_{c}$) in the slow cooling phase.  In this case, 

397: since there is no characteristic frequency above $\nu_{m}$ (in the fast 

398: cooling case) and $\nu_{c}$ (in the slow cooling case), it is reasonable 

399: to extrapolate upwards in energy.  The electron power-law index 

400: $2 \le p < 3$ is typical for both the prompt emission 

401: \citep[e.g.,][]{kaneko2006} and the afterglow of GRBs 

402: \citep[e.g.,][]{pk2002,yost2003}.  Thus, if we select the burst samples 

403: which have an observed photon index steeper than $2$ in the XRT data, 

404: the systematic error in the XRT-to-BAT extrapolation should be reduced.  

405: 

406: Following these arguments, we describe in this paper the XRT-to-BAT 

407: extrapolation to investigate the nature of the XRT steep decay

408: component.  As we mentioned, in principle the XRT-to-BAT extrapolation 

409: has similar issues to the BAT-to-XRT extrapolation.  

410: However, for certain bursts the XRT-to-BAT extrapolation could 

411: greatly reduce the systematic error.  More importantly, a different 

412: approach to the analysis may provide an alternate view of the problem.  

413: 

414: \section{Analysis}

415: 

416: X-ray light curves were produced from the XRT data for all 

417: {\it Swift} GRBs detected between June 2005 \footnote{Following updates 

418: to the on-board software which compensate for the uncontrolled 

419: temperature due to failure of the cooling control system 

420: \citep{kennea2005}, bright Earth effects and micrometeriod 

421: damage \citep{modes2}.} and September 2006.  54 GRBs with an early 

422: phase power-law decay index steeper than $-2$ 

423: ($t_{0}$ taken as the BAT trigger time) were identified.  

424: From the spectral analysis of the X-ray data from these bursts in 

425: the 0.3 -- 10 keV band, we found that many bursts exhibited a significant 

426: difference in the best-fit spectral parameters between the early ($<$ 1000

427: seconds) data and the later ($>$ 1000 seconds) data.  This is 

428: likely due to the fact that $\ep$ is moving through XRT energy range early 

429: in the observation but one cannot measure this from the XRT data

430: alone because of the narrow energy band \citep[e.g.,][]{butler2007}.  To 

431: minimize the effects of the spectral evolution and also the absorption, 

432: only data above 2 keV was used in

433: these analysis.  We further refined the burst sample to those which satisfied the

434: following criteria: 1) Greater than four data bins containing at least 

435: 20 counts in the 2--10 keV light curve during the early phase ($<$ 1000

436: seconds). This criteria reduced the number of bursts in the sample

437: significantly, due to much fewer counts being detected above 2 keV.  

438: 2) No more than a single flare present in the early XRT light 

439: curve\footnote{Bursts exhibiting flares in the BAT were not excluded.}.  

440: 3) The joint spectral analysis of the Photon Counting (PC) and 

441: Windowed Timing (WT) data must include a photon index of $2$ in the 90\% 

442: confidence interval.  13 GRBs satisfied the screening criteria.  Since the 2-10 keV 

443: joint spectral fit to the PC and WT data includes a best fit photon index steeper 

444: than $2$ for all the GRBs in the sample, $\ep$ should be below 2 keV.  

445: This will reduce the systematic error of the extrapolation because there should 

446: be no spectral evolution as a result of the shift of $\ep$ during the burst.  

447: 

448: As mentioned in \citet{zhang2006a}, the definition of the 

449: offset time ($t_{0}$) is critical when performing fits to 

450: the early phase light curve.  Traditionally, $t_{0}$ is defined as 

451: the trigger time of the GRB instrument; when the count rate exceeds 

452: some background level (rate trigger).  However, the definition of 

453: the BAT trigger time is different.  The BAT trigger time is 

454: the start time of the foreground time interval of the image from 

455: which the GRB is detected on-board.  Thus, to be comparable to a 

456: rate trigger time, we define $t_{0}$ as the start 

457: time of the prompt emission (start time of $t_{100}$ interval) for 

458: the whole sample.  

459: 

460: \subsection{BAT analysis}

461: 

462: The BAT analysis was performed using HEAsoft version 6.1.1 and CALDB 

463: version 2006-05-30.  The event-by-event data were used for these analysis.  

464: The non-linear energy correction for each event was applied by {\tt bateconvert}.  

465: The mask-weighting factors were calculated by {\tt batmaskwtevt} using 

466: the on-board position.  The detector enable and disable map was created by 

467: {\tt bathotpix} combining the enable and disable map generated by the flight 

468: software.  We created the BAT light curve by {\tt batbinevt} in the full energy range 

469: (15-350 keV\footnote{The coded mask is transparent to photons above 150 keV.  

470: Thus, photons above 150 keV are treated as background in the mask-weighted 

471: method.  The effective upper boundary is $\sim$ 150 keV.}) 

472: in 4 ms bin except for GRB 051109A (64 ms), GRB 060427 (1 s) and 

473: GRB 060923C (64 ms).  We used larger binning for these three GRBs 

474: because of the low signal-to-noise ratio of the emission.  The duration and 

475: the time intervals based on the Bayesian Block algorithm \citep{scargle1998} 

476: were calculated by {\tt battblocks}.  

477: The spectrum of each time interval was extracted by {\tt batbinevt}.  

478: The energy response file was created by {\tt batdrmgen}.  If the time 

479: interval was during the spacecraft slew, we updated the keywords in the 

480: spectral file related to the energy response process by {\tt batupdatephakw} 

481: and then created 

482: the energy response file for the time interval by {\tt batdrmgen}.  

483: We applied systematic error vectors to the spectrum using {\tt

484: batphasyserr}  prior to doing the spectral analysis.  The spectral

485: analysis was performed using Xspec 11.3.2.  

486: 

487: The energy flux in the 15--25 keV band was calculated for each time 

488: interval directly from the spectral fitting process.  The spectra 

489: from each time interval were fitted with a simple power-law model.  

490: According to the BAT GRB catalog (Angelini et al. in preparation), the 

491: detection threshold of the BAT in the 15--25 keV band is $\sim$ 10$^{-9}$ 

492: ergs cm$^{-2}$ s$^{-1}$.  Based on this result, the 15--25 keV flux 

493: was treated as an upper limit when the calculated 15--25 keV flux was 

494: less then 10$^{-9}$ ergs cm$^{-2}$ s$^{-1}$.  

495: The upper limit was estimated from using the event-by-event data from the Crab 

496: nebula on-axis observation collected on 2005 March 24 (observation ID: 

497: 00050100016).  According to this observation, the BAT can detect the Crab 

498: nebula in the 15--25 keV band at 5 $\sigma$ in a one second exposure.  

499: Assuming that the BAT sensitivity scales as the square-root of the 

500: exposure time \citep{markwardt2005} and a canonical Crab flux of $5.3

501: \times 10^{-9}$ ergs cm$^{-2}$ s$^{-1}$ in the 15--25 keV band, 

502: we calculated the BAT 5 $\sigma$ 

503: upper limit in the 15--25 keV band from the following equation, 

504: \begin{equation}

505: F(15-25 {\rm keV})_{5\sigma} = \frac{5.3 \times 10^{-9}}{f_{pcode}}\, t_{exp}^{-0.5} \;({\rm ergs\,cm^{-2}\,s^{-1}}).

506: \end{equation}

507: Here $t_{exp}$ is the exposure time and $f_{pcode}$ is the partial 

508: coded fraction.  Since our estimation of the 5$\sigma$ upper limit is based on the on-axis Crab 

509: observation, $f_{pcode}$ will correct for a source observed in

510: the off-axis direction.  

511: 

512: For the time-averaged spectral analysis, we use the time interval from 

513: the emission start time to the emission end time ($t_{100}$ interval).  

514: When the spacecraft slew occurred during the time interval, we created the 

515: response matrices for each five second period, taking into account the position 

516: of the GRB in detector coordinates.  We then weighted these response 

517: matrices by the five second count rates and created the averaged 

518: response matrices.  Since the spacecraft slews about one degree per one 

519: second in response to a GRB trigger, we choose five second intervals to calculate 

520: the energy response for every five degrees.  

521: 

522: \subsection{XRT analysis}

523: 

524: The 13 bursts meeting the criteria outlined in section 2 were  

525: processed using the HEAsoft tools version 6.1.1 including version 2.5a 

526: of the {\it Swift} software.  The level 2 cleaned event files were  

527: produced from the {\tt xrtpipeline} task (version 10.4) using the  

528: standard screening criteria.  Version 8 of the response matrices in 

529: CALDB and the corresponding ancillary response produced from 

530: {\tt xrtmkarf} were used.  The standard grades \citep{burrows2005} 

531: were used in the analysis; grades 0-2 and 0-12 for WT and PC mode, respectively.

532: 

533: Both WT and PC mode data \citep{modes,modes2} from the first observation 

534: segment (000) were analyzed.  The source and background extraction

535: regions were nominally a 40-pixel square and an annulus with a 3 pixel 

536: inner radius and 25 pixel outer radius for WT and for PC mode,

537: respectively.  The extraction regions were modified for the piled-up

538: cases in WT mode in accordance with \citet{pileup2}, eliminating 

539: an inner square centered on the source.  It is well documented that 

540: pile-up in PC mode causes a redistribution of single pixel events to higher grades 

541: \citep{pileup}. In order to account for pile-up in PC mode, the grade 

542: distributions were analyzed for each burst during periods where the

543: count-rate exceeded 0.5 counts/sec.  For each burst, the

544: percentage of single pixel events were plotted versus the radius of 

545: the center of the annulus for increasing radii (0--8 pixels).  

546: The annulus with smallest radius at which the percentage of single 

547: pixel events became constant was used for the PC mode analysis.  Nominally, 

548: for a spectrum where the average energy is 1.5\ keV, 78\% of the events 

549: will be single pixel events \citep{moretti2004}.  

550: 

551: The hardness ratio ((2.0-10.0)keV/(0.5-2.0) keV) was examined for each

552: burst to ensure that there was no significant spectral evolution.  

553: An exposure map was created from {\tt xrtexpomap} to correct for the 

554: dead columns and hot pixels. {\tt xrtmkarf} was used to create the 

555: ancillary response files.  This includes corrections for losses in 

556: the wings of the point spread function and the center of the annulus, 

557: for the exposure and for vignetting. 

558: 

559: The spectrum file was binned with a minimum of 20 counts/bin 

560: in order for $\chi^2$ statistics to be valid for the spectral fitting.  

561: Xspec version 11.3.2 was used to perform a joint spectral fit of the 

562: WT and PC data from 2--10 keV using a simple power law model.  

563: The 15--25 keV normalization obtained from the $pegpwrlw$ model 

564: was used to extrapolate the XRT count rate into flux in the 

565: BAT 15--25 keV energy range.  Only 

566: light curve bins with greater than 90\% exposure were used in order to 

567: limit errors due to dead-time incurred by instrument mode switching. 

568: The light curve was binned to have greater than 20 counts/bin.

569: 

570: \subsection{Fitting a composite BAT and XRT light curve}

571: 

572: To investigate the connection between the prompt and the 

573: afterglow emission in the composite light curves, we first fit 

574: the XRT light curve, only using data where the hardness ratio 

575: showed no strong spectral evolution.   

576: We then fit both the BAT and XRT light curves jointly.  Both fits 

577: were performed using a power-law model with an offset time (PLO), 

578: \begin{equation}

579: F_{15-25 \rm keV} = K_{pow}\,(t-t_{0}^{pow})^{-\alpha}, 

580: \label{eq:curvature}

581: \end{equation}

582: where $t_{0}^{pow}$ is a offset time, $\alpha$ is a decay index, 

583: and $K_{pow}$ is the normalization, and with an exponential model (EXP), 

584: \begin{equation}

585: F_{15-25 \rm keV} = K_{exp}\,\exp(-\frac{t}{w}), 

586: \label{eq:exp}

587: \end{equation}

588: where $w$ is the decay constant and $K_{exp}$ is the normalization.  

589: For the XRT only fit, we fixed $t^{pow}_{0}$ to zero.

590: Finally, we fit the BAT and XRT light curves simultaneously using a 

591: combination of a power-law model with an exponential decay component (PLEXP), 

592: \begin{equation}

593: F_{15-25 \rm keV} =  K_{pow}\,(t-t_{0}^{pow})^{-\alpha} +

594:  K_{exp}\,\exp(-\frac{t}{w}).  

595: \label{eq:comb}

596: \end{equation}

597: 

598: For the fit to the XRT only light curve, the time interval 

599: was from the first XRT data point to the last data point before 

600: showing a deficit from the PLO model using an offset time of zero.  

601: For the joint BAT and XRT fit, 

602: the time interval was from the first BAT data point to the 

603: last XRT data point before showing a residual from the PLO model.  

604: Any other definitions used for the time interval are stated as a 

605: footnote in table \ref{tab:bat_xrt_lc_para}.  The values of the time 

606: intervals used in the light curve fitting are shown in the fourth 

607: column of table \ref{tab:bat_xrt_lc_para}.  

608: The best-fit model was selected based on the $\chi^{2}$ of the fit.  

609: However, because a PLO model will not fit the data before $t_{0}^{pow}$, 

610: the judgment between PLO and PLEXP is based on visual inspection 

611: as to whether the model fit both the BAT and XRT data simultaneously or not.  

612: 

613: \section{Results}

614: 

615: The left panels of figures \ref{fig:bat_xrt_lc1}--\ref{fig:bat_xrt_lc5} show the 

616: composite BAT (black open circles) and XRT (red open 

617: triangles) light curve in the 15-25 keV band overlaid with the best-fit 

618: light curve model.  The light curve models are PLO (eq. (\ref{eq:curvature})),  

619: EXP (eq. (\ref{eq:exp})), and PLEXP (eq. (\ref{eq:comb})) from top to bottom.  

620: In the bottom figure of PLEXP, both PLO and EXP  components are also shown as 

621: a dashed and a dash-dotted line, respectively.  

622: The best-fit parameters of the light curves are summarized in 

623: table \ref{tab:bat_xrt_lc_para}.  The best-fit of these models is 

624: labeled in blue.  The right panels of figures 

625: \ref{fig:bat_xrt_lc1}--\ref{fig:bat_xrt_lc5} from top to bottom 

626: show the BAT light curve in the 15-150 keV band, the BAT photon 

627: index based on the time-resolved spectral analysis, the XRT 2--10 keV 

628: count rate, and the XRT count rate ratio (2.0--10.0) keV/(0.5--2.0) keV.  

629: The best-fit spectral parameters based on the 2--10 keV joint fit to 

630: the XRT WT and PC mode data are summarized in table \ref{tab:xrt_spec_para}.

631: 

632: From the initial steep decay phase of the XRT light curve, it is difficult 

633: to distinguish between PLO and EXP from the XRT data alone.  Both models 

634: fit equally well for all of the bursts.  However, the difference and the 

635: importance of the individual components become clear when the BAT data

636: are included in the fit.  An EXP model fits well for GRB 050814, GRB 050915B,

637: GRB 060427 and GRB 060428B.  A PLO or a PLEXP model is not required for

638: these GRBs.  For GRB 060923C, a PLO is the model best represented by 

639: the composite light curve.  A PLEXP is the best model for the remaining 7 GRBs.  

640: The best-fit parameters which we used in the systematic study presented 

641: in this section, are shown in bold font in table \ref{tab:bat_xrt_lc_para}.  

642: 

643: First, we investigated the possibility of the curvature effect for those 

644: GRBs which have a PLO component in the composite BAT and XRT light curve fit.  

645: According to the curvature effect \citep{fenimore1996,kumar2000,zhang2006a}, 

646: the relation between the decay index, $\alpha$, and the XRT photon

647: index, $\Gamma_{XRT}$, should be described by, $\alpha$ =

648: 1+$\Gamma_{XRT}$, if the curvature effect is the cause of the XRT steep 

649: decay.  Figure \ref{fig:curvature} shows the correlation between 

650: $\alpha$ and $\Gamma_{XRT}$ for our sample.  The dashed line is the 

651: expected relationship from the curvature effect ($\alpha$ = 1+$\Gamma_{XRT}$).  

652: Although GRB 060923C may be consistent with the curvature effect, 

653: the majority of the bursts in the sample do not satisfy the expected relation.  

654: The inconsistency with the curvature effect could be due to neglecting 

655: the spectral evolution during the steep decay in our analysis of the XRT data.  

656: Looking at the time evolution of the count rate ratio of our sample, we 

657: find a hard to soft evolution from 0.6 to 0.5 and from 0.6 to 0.4 

658: for GRB 060202 and GRB 060211A.  These changes correspond to 

659: an evolution of the photon index from 1.5 to 1.7 and from 1.5 to 1.9, 

660: respectively, according to the calculation by the Xspec $fakeit$ command 

661: using the detector and the ancillary response files created for each source region.  

662: If this spectral evolution is taken into account, the steep decay could be 

663: consistent with the expectation of the curvature effect for GRB 060211A 

664: and GRB 060202.  However, we do not see a strong spectral 

665: evolution for the other GRBs with the exception of GRB 060418.  

666: Note that for GRB 060418, there is a strong evolution in the hardness 

667: ratio during the episode at $t_{0}$+150 s which may cause an error in 

668: the extrapolated flux.  

669: 

670: We find that most of the sample requires an EXP component to fit 

671: the BAT and XRT light curves simultaneously.  Therefore, we can conclude that 

672: some of the early steep decay observed by XRT is a continuation of the 

673: exponential decay tail of the prompt emission.  Interesting 

674: characteristics can be found for the bursts where a PLEXP 

675: model is the most representative model for the composite light curve.  

676: The dominant component of the fit to the XRT light curve 180 s 

677: after $t_{0}$ for GRB 060202 is an EXP.  For GRB 050803 

678: and GRB 060109 there is almost equal contribution from 

679: the EXP and PLO components in the initial XRT data .  Whereas, a PLO 

680: is the dominant component for GRB 051109A, GRB 060111B GRB 060211A, GRB 060306 

681: and GRB 060418B.  This result clearly demonstrates that the XRT steep decay 

682: could be composed of at least two different components.  

683: Without careful consideration of both the BAT and the XRT data simultaneously, 

684: it is not possible to distinguish between these two different

685: components.  It is important to note that \citet{obrien2006} also 

686: reached a similar conclusion; that the BAT and XRT composite light 

687: curve is composed of an exponential decay which relaxes to a power-law decay.  

688: 

689: For the bursts that exhibit an EXP component, we investigated the 

690: correlation between the exponential decay index, $w$, and the prompt emission 

691: properties derived from the BAT data (table \ref{tab:bat_grb_para}).  

692: The results are summarized in figure \ref{fig:cor_w}.  

693: No correlation is found for the properties of the prompt emission 

694: except between $w$ and the BAT T$_{90}$ which is expected because both 

695: parameters are related to the duration of the bursts.

696: 

697: Based on our study, there is a strong indication that the steep decay 

698: component observed by the XRT is part of the prompt emission \citep[e.g., also][]

699: {nousek2006,obrien2006}.  Thus, we calculate the fluence which may be 

700: below the sensitivity limit of the BAT.  This fluence was calculated by 

701: accumulating the flux from the best-fit composite BAT and XRT light 

702: curve model from the end of the emission as detected by the BAT to 1000 

703: seconds after $t_{0}$.  Figure \ref{fig:BAT_XRT_fluence_hist} shows 

704: the ratio of the percentage of the fluence in the tail emission which 

705: is below the BAT sensitivity limit and the fluence recorded by the BAT.  

706: For 7 out of the 13 GRBs in the sample, the fluence of the tail component 

707: is less than 15\% of the fluence recorded by BAT.  However, more than 15\%

708: of the fluence may be radiated below the BAT sensitivity limit for 

709: GRB 050915B, GRB 051109A, GRB 060202, GRB 060211A, GRB 060427, and 

710: GRB 060428B.  This result gives rise to the question as to

711: whether the fluence measured by the $\gamma$-ray instrument reflects 

712: the true fluence of the prompt emission.  

713: 

714: \section{Discussion}

715: 

716: We have presented the BAT and XRT composite light curves, derived 

717: extrapolating the XRT 2-10 keV flux up to the BAT 15--25 keV energy 

718: range for GRBs which have a steep decay component in the initial XRT 

719: light curve.  Based on the simultaneous fit of 

720: both the BAT and XRT light curves, we have confirmed the existence of an EXP 

721: component which smoothly connects the BAT prompt emission to the 

722: XRT steep decay for several GRBs.  We have also found that the XRT  

723: steep decay for some of the bursts can be fit well by a PLEXP 

724: model.  In the following sections, we discuss the possible origins of the 

725: PLO and EXP components.

726: 

727: \subsection{Origin of the PLO component}

728: 

729: A PLO component most likely originates from an internal shock 

730: \citep[so called, curvature radiation or high-latitude emission

731: associated with the last bright spike.][] {kumar2000,nousek2006,zhang2006a,yamazaki2006}.  

732: Our results support this idea, because for most bursts the 

733: XRT steep decay component smoothly connects with the last 

734: bright episode detected by the BAT (e.g. GRB 050803).  The instantaneous 

735: emission from a uniform jet produces a decay index of 

736: $\alpha = 1 + \Gamma$ \citep{kumar2000}.  This formula was examined 

737: using the power-law decay index derived from the PLO model and the 

738: photon index based on the joint WT and PC spectral analysis of our 

739: sample.  For the majority of the sample, we find that the decay index 

740: is not consistent with the formula.  One of the possible reasons for 

741: this inconsistency could be the spectral evolution during the 

742: early XRT observation in some bursts \citep{zhangbb2006}.  

743: However, as  discussed in section 3,  the inconsistency cannot always 

744: be associated with spectral evolution.  Another possible reason is 

745: the choice of the time-zero \citep{zhang2006a,yamazaki2006}. 

746: \citet{liang2006} investigated the curvature effect as an origin 

747: of the XRT steep decay using the data set of \citet{obrien2006}.  

748: They made the assumption that the XRT steep decay component is due 

749: to the curvature effect and investigated whether the time-zero is 

750: consistent with the beginning of the bright episode.  They concluded 

751: that for most of the sample, the time-zero was consistent with this picture.   

752: The main issue with their approach is that for the fixed power-law decay 

753: index, as expected from the formula $\alpha = 1+\Gamma$,  

754: by shifting the offset time it is possible to fit the decay index 

755: to most early XRT data.  This is because in their approach, the fitting 

756: parameters are not only the offset time but also the normalization, 

757: allowing extra freedom in the fit.  Here, we demonstrate this problem using 

758: GRB 050803 and GRB 050814.   The dash-dotted lines in figure 

759: \ref{fig:linag_prob} show the best fit PLO model by changing 

760: $\alpha$ from 1 to 5 for GRB 050803 and from 2 to 3 for GRB 050814, 

761: and varying $t_{0}^{pow}$.  Only the XRT data, shown in red triangles, 

762: are used in the fitting process as in \citet{liang2006}.  As seen in the figures, 

763: the choice of $\alpha$ and $t_{0}^{pow}$ is not unique if one only try to fit 

764: the XRT data.  Moreover, as it is clearly demonstrated in the case of GRB 050814, 

765: even if $t_{0}^{pow}$ is chosen as the start time of the GRB pulse, the intensity 

766: of the pulse expected from the model is an order of magnitude brighter than 

767: the data.  Therefore, if a bright episode in 

768: the BAT data, which could contribute to the steep decay component, is 

769: simultaneously fit with the XRT data, as in our approach, both the 

770: offset time and the decay index will be uniquely constrained by the data.  

771: It may be difficult to test the curvature effect or the relation 

772: $\alpha = 1+\Gamma$ definitely without fitting the XRT and BAT 

773: data simultaneously.  The third possibility is to abandon the assumption

774: of the uniform jet emission.  Our results may suggest 

775: that the structure of the jet is much more complex than a uniform jet

776: \citep[e.g.][]{yamazaki2006}.  

777: 

778: \subsection{Origin of EXP component}

779: 

780: Since it is difficult to explicitly state the origin of an EXP component, 

781: we will discuss the possibilities of both an internal shock 

782: and an external shock as the origin of an EXP component.  

783: 

784: \subsubsection{External shock scenario for EXP component}

785: 

786: One interpretation of an EXP component is the presence of the 

787: external shock emission during the prompt phase.  \citet{fenimore1999} 

788: studied the case of the co-existence of the emission from the external shock 

789: (deceleration of the initial shell) during the emission from the internal 

790: shocks.  They showed that the smooth long lasting soft emission which arose 

791: from an external shock could overlay the light curve of the 

792: prompt emission.  Furthermore, \citet{fenimore1999} showed that the efficiency 

793: of converting the bulk energy to radiation is 85\% in this case,   

794: whereas, internal shocks without deceleration only convert 

795: about 1\% \citep{kumar1999,panaitescu1999}.  \citet{zhang2006b} 

796: calculated the radiation efficiency using the {\it Swift} X-ray afterglow 

797: data and show that about half of the GRBs have an 

798: efficiency $\gtrsim$ 1\% which provides a challenge for producing the prompt 

799: emission from internal shocks alone.  Furthermore, according to the 

800: optical observation by Rapid Telescopes for Optical Response (RAPTOR) 

801: during the prompt emission of GRB 050820A, 

802: smoothly decaying emission which does not correlate with the prompt 

803: spikes was found \citep{raptor_050820a}.  If this emission is from 

804: the external shock, we might be observing the deceleration of the 

805: outflow during the prompt phase.  

806: 

807: Following the argument of \citet{fenimore1996} and \citet{fenimore1999}, 

808: we estimate the bulk Lorentz factor of our sample, assuming 

809: an EXP component is purely due to external shock emission.  

810: Let us assume that an external shock starts its emission at the radius, $R_{0}$.  

811: The external shock will be decelerated because of sweeping up the 

812: inter-stellar medium (ISM).  The total energy of the central engine, 

813: $E_{0}$, can be expressed as, 

814: \begin{equation}

815: E_{0} = (4 \pi / 3) R_{0}^{3} n_{\rm ISM} m_{p}c^{2} \gamma_{0}^{2}

816: \label{eq:e0}

817: \end{equation}

818: where $n_{\rm ISM}$ is a density of the ISM, $m_{p}$ is the proton mass, c is 

819: the speed of light, and $\gamma_{0}$ is the bulk Lorentz factor at $R_{0}$.  

820: The duration of the emission ($\Delta T$) is determined by the radial time 

821: scale \citep{piran1999} which is the difference in the arrival time of the 

822: photons emitted between $R_{0}$ and $aR_{0}$ ($a>1$)  as measured by the observer, 

823: \begin{displaymath}

824: \Delta T = [(a^{4}-1)/4] (R_{0}/2 \gamma_{0}^{2}c).  

825: \end{displaymath}

826: Therefore, the bulk Lorentz factor can be expressed as, 

827: \begin{equation}

828: \gamma_{0} = (32 \pi m_{p} c^{2} / 3)^{-1/8} [(a^{4}-1)/4]^{3/8}

829:  (E_{0}/n_{ISM})^{1/8} (c \Delta T)^{-3/8}.  

830: \label{eq:gamma0}

831: \end{equation}

832: The relationship between $\Delta T$ and the full width at half maximum 

833: (FWHM) of the pulse ($T_{1/2}$) is described by, 0.22 $\Delta T$ = $T_{1/2}$, which is 

834: valid for a pulse shape of a fast rise, exponential decay (FRED) 

835: \citep{fenimore1996}.   Thus, once we know the redshift, since $T_{1/2}$ can be 

836: estimated from the best-fit parameters of an EXP component, we can calculate 

837: the bulk Lorentz factor as a function of $E_{0}$/$n_{\rm ISM}$.  In the following 

838: arguments, we use a typical value of 2 for the parameter ``$a$'' 

839: (in the $a=2$ case, $\Delta T$ is the radial time scale from $R_{0}$ to 2$R_{0}$).  

840: 

841: For bursts with unknown redshifts (all bursts in the sample except 

842: GRB 050814, GRB 051109A and GRB 060418),  we used the mean redshift of

843: 2.4 obtained from the {\it Swift} long GRBs\footnote{http://heasarc.gsfc.nasa.gov/docs/swift/archive/grb\_table/}.  

844: We can derive a reasonable range of

845: $\gamma_{0}$ from 143 to 350 and $R_{0}$ of $\sim$ 10$^{16}$ cm 

846: assuming $E_{0}/n_{\rm ISM} = 1 \times 10^{52}$ erg cm$^{3}$ 

847: for the 12 GRBs in our sample which have an EXP component in the best

848: fit model (table \ref{tab:bat_grb_para}).  According 

849: to the calculation of $\gamma_{0}$ by the Rapid Eye Mount (REM) telescope 

850: using the peak time of the early afterglow data, $\gamma_{0}$ is about 

851: 400 for GRB 060418 and GRB 060607A \citep{molinari2006}.  

852: Their $\gamma_{0}$ value agrees within a factor of two of our estimates.  

853: For GRB 060418, $E_{0}$ will be $\sim$ $1 \times 10^{54}$ 

854: ergs if we assume $\gamma_{0}$ of 400 as derived from the REM observation and 

855: $n_{\rm ISM}$ of 1 cm$^{-3}$.  $E_{0}$ of $\sim$ $10^{54}$ ergs is also 

856: a typical value according to the calculation of \citet{zhang2006b} 

857: using X-ray afterglow data observed by XRT.  

858: 

859: It would be interesting to look for a correlation between 

860: $\gamma_{0}$ and $\ep$ in the GRB rest frame 

861: ($\eps$) although it is difficult to calculate for our sample because 

862: we do not have measurements of both $\ep$ and redshift.  

863: One of the advantages of using the {\it Swift} sample for this 

864: study is that soft GRBs, so called X-ray Flashes (XRFs), are included 

865: in the sample because of the relatively softer energy response of 

866: BAT compared to that of BATSE.  In the dirty fireball model 

867: \citep{dermer1999}, $\eps$ has a strong dependency on the bulk Lorentz 

868: factor ($\eps \propto \gamma_{0}^{4}$).  The unified jet models for 

869: XRFs and GRBs, such as the structured-jet model \citep{rossi2002}, 

870: and the variable jet opening angle model \citep{lamb2006,tim2006}, 

871: expect a positive correlation between $\eps$ and the bulk Lorentz 

872: factor.  In the off-axis jet model \citep{yamazaki2004}, 

873: no correlation is expected between $\gamma_{0}$ and $\eps$ because the Doppler factor 

874: will change as a function of a viewing angle but not $\gamma_{0}$.  

875: Another interesting theoretical model to discuss is 

876: the case of a very high Lorentz factor \citep{mochkovitch2003,barraud2005}.  

877: According to this model, XRFs can be produced in a condition with 

878: a very high $\gamma_{0}$ (so called ``clean fireball''), while 

879: classical GRBs have a moderate $\gamma_{0}$.  In this case, we would expect 

880: a negative correlation between $\eps$ and the bulk Lorentz factor.  

881: Some additional effort, for example, to estimate $\ep$ from the {\it Swift} BAT data 

882: (Sakamoto et al. in preparation) and to estimate a redshift from the {\it Swift} data 

883: \citep[e.g.,][]{grupe2006} is encouraged in order to discuss 

884: the correlation between $\gamma_{0}$ and $\eps$ and the origin of XRFs.  

885: 

886: We can derive another constraint on $\gamma_0$ which is independent of 

887: the previous discussion.  Our observational results suggest that the photons

888: originating from an internal shock via the curvature effect (PLO component) 

889: and photons from an external shock (EXP component) 

890: arrive at the observer almost simultaneously.  The observed time of the 

891: photon from an internal shock, $T_{int}$, can be expressed as, 

892: \begin{displaymath}

893: T_{int} \sim (R_{int}/2c\gamma_{int}) [1+(\gamma_{int}\theta)^{2}], 

894: \end{displaymath}

895: where $R_{int}$ is the radius where an internal shock emits,

896: $\gamma_{int}$ is the bulk Lorentz factor at $R_{int}$, and $\theta$ 

897: is the jet opening half-angle.  On the other hand, the observed time of 

898: an external shock emission, $T_{ext}$, is expressed as, 

899: \begin{displaymath}

900: T_{ext} \sim R_{ext}/(4c\gamma_{ext}^2), 

901: \end{displaymath}

902: where $R_{ext}$ is the radius where an external shock emits and $\gamma_{ext}$ is a 

903: bulk Lorentz factor at $R_{ext}$ \citep{sari1998a}.  If we assume 

904: $T_{int} \sim T_{ext}$ and also $\gamma_{int} \sim \gamma_{ext} \sim \gamma_{0}$, 

905: we have, 

906: \begin{displaymath}

907: 2R_{int}[1+(\gamma_{0}\theta)^{2}] \sim R_{ext}.  

908: \end{displaymath}

909: $\eps$ can be 

910: written as a function of $\theta$ and $\gamma_{0}$, 

911: \begin{displaymath}

912: \eps(\theta) \sim (2\gamma_{0} h \nu_{0}^{\prime})[1+(\gamma_{0} \theta)^{2}] \sim 

913: [\eps (\theta = 0)] / (1 + (\gamma_{0} \theta)^{2}), 

914: \end{displaymath}

915: where $\eps (\theta = 0)$ 

916: is $\ep$ observed by the on-axis observer.  In this case, $\eps (\theta = 0)$ 

917: corresponds to the observed $\ep$ multiplied by (1+z).  

918: Therefore, the relationship 

919: between $R_{int}$, $R_{ext}$ and $\ep$ is given by, 

920: \begin{displaymath}

921: R_{ext}/2 R_{int} \sim [\ep (\theta = 0)]/\ep(\theta).  

922: \end{displaymath}

923: Since the observed photon index of the XRT steep decay emission of 

924: our sample is $\sim 2$, which suggests that the observation of the 

925: spectrum is above $\ep$, it is reasonable to assume that the upper 

926: limit of $\ep(\theta)$ is in the few keV range.  Hence, the condition will be, 

927: \begin{displaymath}

928: R_{ext}/2 R_{int} > [\ep (\theta = 0)]/\ep(\theta^{\prime}) 

929: \end{displaymath}

930: where $\ep(\theta^{\prime}$) is the upper limit of a few keV.  

931: If we use the angular spreading time ($\Delta T_{ang}$) \citep{piran1999} 

932: as the time scale of an internal shock, then, 

933: \begin{displaymath}

934: R_{int} \sim 2 c \gamma_{0} \Delta T_{ang}.  

935: \end{displaymath}

936: From Eq.~(\ref{eq:e0}), the radius of an external shock can be expressed as 

937: \begin{displaymath}

938: R_{ext} = (4 \pi m_{p}c^{2}/3)^{-1/3} (E_{0}/n_{ISM})^{1/3} \gamma_{0}^{-2/3}.  

939: \end{displaymath}

940: Thus, $\gamma_{0}$ can be written as, 

941: \begin{equation}

942: \gamma_{0} > (1/4)^{3/8} (4 \pi m_{p}c^{2}/3)^{-1/8} (E_{0}/n_{ISM})^{1/8} (c \Delta T_{ang})^{-3/8} 

943: [\ep (\theta = 0)/\ep(\theta^{\prime})]^{-3/8}.  

944: \end{equation}

945: 

946: In the case of GRB 050803, $\Delta T_{ang}$ can be derived as the

947: duration of the last spike (the pulse at $t_{0}$+90 s), $\Delta T_{ang}$ 

948: $\sim$ 3 s, if one takes into account the time dilation effect.  

949: Although it is not possible to extract the information about the $\ep$ 

950: of this pulse from the BAT data, it is reasonable to assume $\ep$ $>$ 

951: 100 keV since the photon index from a simple power-law fit to the BAT 

952: data is $1.2 \pm 0.2$ which is close to the low energy photon index 

953: of the typical GRB spectrum ($\ep$ should be around or 

954: above the BAT upper energy range of 150 keV).  Therefore, we use 

955: $\ep (\theta = 0)/\ep(\theta^{\prime}) > 100$.  The 

956: lower limit of $\gamma_{0}$ of GRB 050803 is estimated to be $<$ 50 

957: assuming a redshift of 2.4 (the mean redshift of the {\it Swift} 

958: long GRBs) and $E_{0}/n_{ISM}$ of $1 \times 10^{52}$ ergs cm$^{3}$.  

959: Applying the same assumptions for the GRB spectral parameters 

960: ($\ep (\theta = 0)/\ep(\theta^{\prime}) > 100$), we estimate 

961: a lower limit of $<$ 60 for GRB 060418 

962: using the measured redshift 

963: of 1.489, $\Delta T_{ang}$ of 2 s (the pulse at $t_{0}$+50 s), and 

964: $E_{0}/n_{ISM}$ of $1 \times 10^{52}$ ergs cm$^{3}$.  The estimated 

965: lower limit of $\gamma_{0}$ for GRB 060418 is not contradicted by the value based 

966: on the REM observation.  

967: %

968: In summary, for the EXP component, one can provide a reasonable bulk Lorentz 

969: factor, $\gamma_{0}$, within the external shock scenario, which is consistent 

970: with other measurements.  

971: %Further detailed observations of the broad-band spectrum 

972: %during the prompt emission is needed to resolve these two components.  

973: 

974: \subsubsection{Internal shock scenario for EXP component}

975: 

976: As we discussed in \S 4.1, a simple interpretation based on a uniform

977: jet model could contradict the prediction of the curvature effect.  

978: Here we discuss an internal shock scenario for both EXP and PLO

979: components based on an inhomogeneous jet model.  

980: \citet{yamazaki2006} investigated the GRB prompt emission 100-1000 

981: seconds after the GRB trigger within the frame work of a multiple sub-shell model.  

982: According to their study, despite an angular inhomogeneity of the jet, 

983: the tail emission has a monotonic decay which resembles the XRT steep decay.  

984: In this context, if the jet has a core in which the emission energy is 

985: densely confined compared with the outer region, the PLO decay 

986: component arising from an on-beam sub-shell may be overlaid by 

987: the off-beam core emission which causes the EXP decaying component.   

988: \citet{takami2006} further extended their study, and in order to 

989: investigate the unknown jet structure, they proposed unique 

990: definitions of the decay index derived by unique definitions of 

991: the time-zero and of the fitting interval of the observed light curve.

992: They found that the decay index in their definitions should have 

993: a wide scatter in the case of a power-law like structured jet. 

994: Here, we calculated the decay index using our BAT and XRT composite 

995: light curve based on the definitions of \citet{takami2006}.  Because of the 

996: difficulty of using exactly the same definition of time-zero 

997: as proposed in \citet{takami2006} ($T_{*}$ in their paper), we define 

998: the end of the BAT emission 

999: as $T_{*}$.  We can fit our light curves using the proposed fitting interval 

1000: for five GRBs in our sample ($\alpha_{tail}$ as a decay index, $\chi^{2}$ and 

1001: d.o.f. of the fits are shown in the last two columns of table \ref{tab:bat_grb_para}).  

1002: Our results, based on a very small sample, show that the decay index 

1003: ranges from 0.6 to 3.0.  However, both the size of the sample and 

1004: the number of data points included in the light 

1005: curve fit are very small because of using XRT data above 2 keV.  

1006: There is an additional issue further reducing the sample; the 

1007: appearance of a shallow decay in the XRT data in the fitting interval.  

1008: Unfortunately, it is hard to conclude the structure of a jet with our 

1009: limited sample.  Once the GRB sample suitable for fitting a light curve 

1010: with the unique definitions of \citet{takami2006} can be increased, 

1011: we may be able to draw a conclusion about the jet structure of 

1012: GRBs using the XRT steep decay component.  

1013: 

1014: \section{Summary}

1015: 

1016: In this paper, we presented a systematic study of the steep decay 

1017: emission observed by the XRT.  We constructed composite light curves 

1018: in the 15--25 keV band extrapolating the XRT data (2--10 keV) up to 

1019: the BAT energy range (15--25 keV).  Based on the simultaneous fitting 

1020: of the BAT and XRT data, we confirmed the existence of an EXP component 

1021: for the majority of the bursts in the sample.  We found that for the 

1022: PLO component, the majority of the GRBs in our sample are inconsistent 

1023: with the relationship of the curvature effect, $\alpha = \Gamma + 1$, 

1024: which is only valid in the case of the uniform jet.  We also found that more 

1025: than 15\% of the prompt fluence may be radiated below the BAT 

1026: sensitivity limit for half of our sample.  We argue that 

1027: the EXP component could be the emission from the external shock 

1028: which may indicate the deceleration of the initial shell by ISM during 

1029: the prompt phase.  We discuss the case of the prompt tail emission 

1030: from the structured jet as an origin of the XRT steep decay but the

1031: sample is too small for a solid conclusion.  

1032: 

1033: \acknowledgements

1034: We would like to thank A. P. Beardmore, G. Chincarini, C. Guidorzi, 

1035: P. T. O'Brien and K. L. Page for valuable comments.  We also would like 

1036: to thank the anonymous referee for comments and suggestions that materially 

1037: improved the paper.  

1038: This research was performed while T.~S. held a NASA Postdoctoral Program 

1039: administered by Oak Ridge Associated Universities at NASA Goddard Space 

1040: Flight Center.  R.~Y. was supported in part by Grants-in-Aid for Scientific Research

1041: of the Japanese Ministry of Education, Culture, Sports, Science,

1042: and Technology 18740153.  The material of the paper has been improved by

1043: the discussions during the workshop ``Implications of {\it Swift's}

1044: Discoveries about Gamma-Ray Bursts'' at the Aspen Center for Physics.  

1045: 

1046: \begin{thebibliography}{99}

1047: \bibitem[Band et al.(1993)]{band1993}

1048: 	Band, D.L., et al. 1993, ApJ, 413, 281

1049: \bibitem[Barraud et al.(2005)]{barraud2005}

1050: 	Barraud, C. et al. 2005, A\&A, 440, 809

1051: \bibitem[Barthelmy et al.(2005)]{barthelmy2005}

1052: 	Barthelmy, S.D. et al. 2005, Space Sci. Rev., 120, 143

1053: \bibitem[Barthelmy et al.(2006)]{barthelmy2006}

1054: 	Barthelmy, S.D. et al. 2006, ApJ, 635, L133

1055: \bibitem[Burrows et al.(2005)]{burrows2005}

1056: 	Burrows, S.D. et al. 2005, Space Sci. Rev., 120, 165

1057: \bibitem[Butler \& Kocevski(2007)]{butler2007}

1058: 	Butler, N.R., Kocevski, D. 2007, Submitted to ApJ (astro-ph/0702638)

1059: \bibitem[Dermer et al.(1999)]{dermer1999}

1060: 	Dermer, C.D., Chiang, J., \& M$\ddot{a}$ttcher, M. 1999, ApJ, 656

1061: \bibitem[Donaghy (2006)]{tim2006}

1062: 	Donaghy, T. Q. 2006, ApJ, 645, 436

1063: \bibitem[Dupree et al.(2006)]{dupree2006}

1064: 	Dupree, A.K. et al. 2006, GCN Circ. 4969, http://gcn.gsfc.nasa.gov/gcn3/4969.gcn3

1065: \bibitem[Fenimore et al.(1996)]{fenimore1996}

1066: 	Fenimore, E.E., Madras, C.D., \& S. Nayakshin 1996, ApJ, 473, 998

1067: \bibitem[Fenimore \& Ramirez-Ruiz(1999)]{fenimore1999}

1068: 	Fenimore, E.E. \& Ramirez-Ruiz, E. 1999, Submitted to ApJ (astro-ph/9909299)

1069: \bibitem[Frontera et al.(1999)]{frontera1999}

1070: 	Frontera, F. et al. 1999, ApJS, 127, 59

1071: \bibitem[Gehrels et al.(2004)]{gehrels2004}

1072: 	Gehrels, N. et al. 2004, ApJ, 611, 1005

1073: \bibitem[Giblin et al.(1999)]{giblin1999}

1074: 	Giblin, T.W. et al. 1999, ApJ, 524, L47

1075: \bibitem[Giblin et al.(2002)]{giblin2002}

1076: 	Giblin, T.W. et al. 2002, ApJ, 570, 537

1077: \bibitem[Grupe et al.(2006)]{grupe2006}

1078: 	Grupe, D. et al. 2007, \aj, 133, 2216

1079: \bibitem[Hill et al.(2004)]{modes} Hill, J.E.,  et al. 2004,  

1080: 	\procspie, 5165, 217

1081: \bibitem[Hill et al.(2005)]{modes2} Hill, J.E.,  et al. 2005,  

1082: 	\procspie, 5898, 313

1083: \bibitem[Jakobsson et al.(2006)]{jakobsson2006}

1084: 	Jakobsson, P. et al. 2006, in AIP Conf. Proc. 836, Gamma-ray bursts in the 

1085: 	{\it Swift} era, ed. S.S.Holt, N. Gehrels, and J.A.Nousek (New York: AIP), 552

1086: \bibitem[Kaneko et al.(2006)]{kaneko2006}

1087: 	Kaneko, Y. et al. 2006, ApJS, 166, 298

1088: \bibitem[Kennea et al. (2005)]{kennea2005}

1089: 	Kennea, J.A. et al. 2005, \procspie, 5898, 341

1090: \bibitem[Kumar (1999)]{kumar1999}

1091: 	Kumar, P. 1999, ApJ, 523, L113

1092: \bibitem[Kumar \& Panaitescu(2000)]{kumar2000}

1093: 	Kumar, P. \& Panaitescu, A. 2000, ApJ, 541, L51

1094: \bibitem[Lamb et al.(2006)]{lamb2006}

1095: 	Lamb, D.Q., Donaghy, T.Q., \& Graziani, C. 2006, ApJ, 520, 335

1096: \bibitem[Lloyd-Ronning and Petrosian (2002)]{lloyd2002}

1097: 	Lloyd-Ronning, N.M., Petrosian, V. 2002, ApJ, 565, 182

1098: \bibitem[Liang et al.(2006)]{liang2006}

1099: 	Liang, E.W. et al. 2006, ApJ, 646, 351

1100: \bibitem[Markwardt et al.(2005)]{markwardt2005}

1101: 	Markwardt, C.B. et al. 2005, ApJ, 633, L77

1102: \bibitem[M\'esz\'aros \& Rees(1997)]{mr1997}

1103: 	M\'esz\'aros, P., \& Rees, M. 1997, ApJ, 476, 232

1104: \bibitem[Mochkovitch et al.(2003)]{mochkovitch2003}

1105: 	Mochkovitch, R., Daigne, F., Barraud, C., \& Atteia, J. L.

1106:         2003, in APS Conf. Ser. 312, Third Rome Workshop on Gamma-Ray Bursts in

1107:         the Afterglow Era, ed. M. Feroci et al. (San Francisco: ASP), 381

1108: \bibitem[Molinari et al.(2006)]{molinari2006}

1109: 	Molinari, E. et al. 2007, A\&A in press (astro-ph/0612607)

1110: \bibitem[Moretti et al.(2004)]{moretti2004}

1111:         Moretti, A. et al. 2004, \procspie, 5165, 232 

1112: \bibitem[Nousek et al.(2006)]{nousek2006}

1113: 	Nousek, J.A. et al. 2006, ApJ, 642, 389

1114: \bibitem[O'Brien et al.(2006)]{obrien2006}

1115: 	O'Brien et al. 2006, ApJ, 647, 1213

1116: \bibitem[Panaitescu et al.(1999)]{panaitescu1999}

1117: 	Panaitescu, A., Spada, M., \& M\'esz\'aros, P. 1999, 

1118: 	ApJ, 522, L105

1119: \bibitem[Panaitescu \& Kumar(2002)]{pk2002}

1120: 	Panaitescu, A., \& Kumar, P. 2002, ApJ, 571, 779

1121: %\bibitem[P\'elangeon et al.(2006)]{pelangeon2006}

1122: %	P\'elangeon, A. 2006, in AIP Conf. Proc. 836, Gamma-ray bursts in the 

1123: %	Swift era, ed. S.S.Holt, N. Gehrels, and J.A.Nousek (New York: AIP), 149

1124: \bibitem[Pian et al.(2001)]{pian2001}

1125: 	Pian, E. et al. 2001, A\&A, 372, 456

1126: \bibitem[Piran(1999)]{piran1999}

1127: 	Piran, T. 1999, Physics Reports, 314, 575

1128: \bibitem[Piro et al.(2005)]{piro2005}

1129: 	Piro, L. et al. 2005, ApJ, 623, 314

1130: \bibitem[Quimby et al.(2005)]{quimby2005}

1131: 	Quimby, R. et al. 2005, GCN Circ. 4221, http://gcn.gsfc.nasa.gov/gcn3/4221.gcn3

1132: \bibitem[Rees \& M\'esz\'aros(1994)]{rm1994}

1133: 	Rees, M., \& M\'esz\'aros, P. 1994, ApJ, 430, L93

1134: \bibitem[Romano et al.(2006)]{pileup2} Romano, P., et al. 2006, A\&A, 456, 917

1135: \bibitem[Rossi et al.(2002)]{rossi2002}

1136:         Rossi, E., Lazzati, D., \& Rees, M.J. 2002, MNRAS, 332, 945

1137: \bibitem[Ryde \& Svensson(2002)]{rs2002}

1138: 	Ryde, F. \& Svensson, R. 2002, ApJ, 566, 210

1139: \bibitem[Sakamoto et al.(2005)]{sakamoto2005}

1140: 	Sakamoto et al. 2005, ApJ, 629, 311

1141: %\bibitem[Sakamoto et al.(2006)]{sakamoto2006}

1142: %	Sakamoto et al. 2006, submitted to ApJ

1143: \bibitem[Sari \& Piran(1997)]{sp1997}

1144: 	Sari, R., \& Piran, T. 1997, ApJ, 485, 270

1145: \bibitem[Sari(1998)]{sari1998a}

1146: 	Sari, R. 1998, ApJ, 494, L49

1147: \bibitem[Sari et al.(1998)]{sari1998b}

1148: 	Sari, R., Piran, T., \& Narayan, R. 1998, ApJ, 497, L17

1149: \bibitem[Scargle(1998)]{scargle1998}

1150: 	Scargle, J. D. 1998, ApJ, 504, 405

1151: \bibitem[Tagliaferri et al.(2005)]{tagliaferri2005}

1152: 	Tagliaferri, G. et al. 2005, Nature, 436, 985

1153: \bibitem[Takami et al.(2007)]{takami2006}

1154: 	Takami, K., Yamazaki, R., Sakamoto, T. \& Sato, G 2007, ApJ in press (arXiv:0704.1055)

1155: \bibitem[Vaughan et al.(2006)]{pileup} Vaughan, S., et al. 2006,  

1156: 	\apj, 638, 920

1157: \bibitem[Vestrand et al.(2006)]{raptor_050820a}

1158: 	Vestrand, W.T. et al. 2006, Nature, 442, 172

1159: \bibitem[Willingale et al.(2007)]{willingale2007}

1160:         Willingale, R. et al. 2007, ApJ in press (astro-ph/0612031)

1161: \bibitem[Yamazaki et al.(2004)]{yamazaki2004}

1162: 	Yamazaki, R., Ioka, K., \& Nakamura, T. 2004, ApJ, 607, L103

1163: \bibitem[Yamazaki et al.(2006)]{yamazaki2006}

1164: 	Yamazaki, R., Toma, K., Ioka, K., \& Nakamura, T. 2006, MNRAS, 369, 311

1165: \bibitem[Yost et al.(2003)]{yost2003}

1166: 	Yost, S. A., Harrison, F. A., Sari, R., \& Frail, D. A. 2003, 

1167: 	ApJ, 597, 459

1168: \bibitem[Zhang et al.(2006a)]{zhang2006a}

1169: 	Zhang, B. et al. 2006a, ApJ, 642, 354

1170: \bibitem[Zhang et al.(2006b)]{zhang2006b}

1171: 	Zhang, B. et al. 2006b, ApJ, 655, 989

1172: \bibitem[Zhang, Liang, \& Zhang(2006)]{zhangbb2006}

1173: 	Zhang, B.B., Liang, E., Zhang, B. 2006, ApJ in press (astro-ph/0612246)

1174: \end{thebibliography}

1175: 

1176: \clearpage

1177: 

1178: %% tables

1179: %%% XRT fitting

1180: \newpage

1181: \begin{deluxetable}{ccccccccccc}

1182: \tabletypesize{\scriptsize}

1183: \rotate

1184: \tablecaption{Parameters of the light curve fits. Errors quoted at 68\%

1185:  confidence level.  See text for details (section 2.3 and 3).\label{tab:bat_xrt_lc_para}}

1186: %\label{tab:bat_xrt_lc_para}

1187: \tablewidth{0pt}

1188: \tablehead{

1189: \colhead{GRB} & 

1190: \colhead{$t_{0}$} &

1191: \colhead{Data$^{\ddagger}$} & 

1192: \colhead{Fitting} & 

1193: \multicolumn{4}{c}{Power-law} &

1194: \multicolumn{3}{c}{Exponential}\\\cline{5-11}

1195: \colhead{} &

1196: \colhead{UT} &

1197: \colhead{} &

1198: \colhead{[s]} &

1199: \colhead{t$_{0}^{pow}$} & 

1200: \colhead{$\alpha$} &

1201: \colhead{$K_{\rm pow}$} & 

1202: \colhead{$\chi^{2}$/dof} & 

1203: \colhead{$w$} & 

1204: \colhead{$K_{\rm exp}$} & 

1205: \colhead{$\chi^{2}$/dof}

1206: }

1207: \startdata

1208: 050803 & 2005-08-03 & XRT & 100--147 & 0$^{\dagger}$ & $5.3 \pm 0.7$ &

1209:   17.4 & 6.6 / 5 & $22 \pm 3$ & $3.8_{-1.9}^{+4.0} \times 10^{-8}$ & 7.3 / 5 \\

1210:        &    19:14:59.3                  & PL/EX & 0--406 & $87.2_{-0.4}^{+0.3}$ &

1211:  $1.18 \pm 0.07$ & $(1.1 \pm 0.3) \times 10^{-8}$ & 14.4 / 10 & $32.3 \pm 0.9$ &

1212:  $(8.4 \pm 0.7) \times 10^{-9}$ & 122.2 / 13\\

1213:        &                         & PLE & 0--406 & {\boldmath $88.3_{-0.05}^{+0.03}$} 

1214: & {\boldmath $0.87 \pm 0.03$} & {\boldmath $2.0 \times

1215:  10^{-9}$$^{\dagger}$} & {\bf --} & {\boldmath $26 \pm 1$} & {\boldmath

1216:  $1.0 \times 10^{-8}$$^{\dagger}$} & {\bf 22.0 / 12} \\\hline

1217: 050814 & 2005-08-14 & XRT & 167--466 & 0$^{\dagger}$ & $3.6 \pm 0.3$ & 

1218:  $4.1_{-2.4}^{+6.2} \times 10^{-2}$ & 11.7 / 20 & $71 \pm 4$ & $4.0 \times

1219:  10^{-9}$ & 23.7 / 20 \\

1220:        &   11:38:55.4  & PL/EX & 0--466 & 70.3 & 2.5 & $5.3 \times

1221:  10^{-5}$ & 10.3 / 21 & {\boldmath $66 \pm 2$} & {\boldmath $(5.0 \pm

1222:  0.5) \times 10^{-9}$} & {\bf 27.7 / 23}\\

1223:        &               & PLE & 0--466 & 44.5 & $1.6$ & $5.1 \times

1224:  10^{-8}$$^{\dagger}$ & -- & $62 \pm 1$ & $5.8 \times 10^{-9}$$^{\dagger}$ & 24.2 / 22 \\\hline

1225: 050915B & 2005-09-15 & XRT & 158--228 & 0$^{\dagger}$ & $5.3

1226:  \pm 0.8$ & $1.3 \times 10^{2}$ & 10.7 / 5 & $36_{-5}^{+6}$ & 

1227:  $2.1_{-1.2}^{+2.7} \times 10^{-8}$ & 11.7 / 5 \\

1228:         & 21:22:56.6 & PL/EX & 7--228 & 21.3 & $2.1$ & $4.0 \times

1229:  10^{-6}$ & 27.0 / 7 & {\boldmath $33.2 \pm 0.6$} & {\boldmath $(3.0 \pm

1230:  0.1) \times 10^{-8}$} & {\bf 27.6 / 8}\\

1231:         &                         & PLE & 7--228 & $-29_{-10}^{+11}$ &

1232:  $3.8^{+0.3}_{-0.2}$ & $2.0 \times 10^{-2}$$^{\dagger}$ & -- & 

1233:  $35_{-4}^{+2}$ & $1.7 \times 10^{-8}$$^{\dagger}$ & 16.4 / 7\\\hline

1234: 051109A & 2005-11-09 & XRT & 131--196 & 0$^{\dagger}$ & $2.2 \pm 0.7$ 

1235:  & $3.2 \times 10^{-6}$ & 0.4 / 4 & $74_{-19}^{+36}$ &

1236:  $5.0_{-2.5}^{+5.2} \times 10^{-10}$ & 0.3 / 10 \\

1237:         & 01:12:17.6 & PL/EX & 0--196 & $1.9$ & -1.6 & $2.2 \times 10^{-7}$ & 7.8 / 6

1238:  & $28 \pm 1$ & $(1.5 \pm 0.2) \times 10^{-8}$ & 50.8 / 8\\

1239:         &  & PLE & 0--196 & {\boldmath $4.78 \pm 0.02$} & {\boldmath $0.79 \pm 0.03$} 

1240: & {\boldmath $2.5 \times 10^{-9}$$^{\dagger}$} & {\bf --} & {\boldmath $21 \pm 3$} 

1241: & {\boldmath $1.5 \times 10^{-8}$$^{\dagger}$} & {\bf 2.8 / 7}\\\hline

1242: 060109 & 2006-01-09 & XRT & 110--200 & 0$^{\dagger}$ & $4.3 \pm 0.4$ & 

1243: $2.6 \times 10^{-1}$ & 7.3 / 8 & $33 \pm 3$ &

1244:  $8.6_{-2.7}^{+4.0} \times 10^{-9}$ & 8.5 / 8\\

1245:        & 16:54:41.2 & PL/EX & 0--862 & $88 \pm 1$ & $1.6 \pm 0.1$ &

1246:  $7.1_{-2.2}^{+3.7} \times 10^{-8}$ & 11.1 / 11 & $42 \pm 2$ & $(3.6 \pm

1247:  0.4) \times 10^{-9}$ & 109.1 / 12\\

1248:        &                         & PLE & 0--862 & {\bf 90 $\pm$ 0.3} &

1249:  {\boldmath $1.42 \pm 0.03$} & {\boldmath $1.8 \times

1250:  10^{-8}$$^{\dagger}$} & {\bf --} & {\boldmath $32 \pm 2$} & {\boldmath $3.5 \times

1251:  10^{-9}$$^{\dagger}$} & {\bf 8.4 / 11} \\\hline

1252: 060111B & 2006-01-11 & XRT & 89--149 & 0$^{\dagger}$  &$2.8

1253:  \pm 0.7$ & $2.1 \times 10^{-5}$ & 2.6 / 3 & $43_{-9}^{+14}$ &

1254:  $6.5_{-3.3}^{+6.2} \times 10^{-10}$ & 3.2 / 3\\

1255:         & 20:15:41.2 & PL/EX & 0--149 & 50 & $2.4$ & $9.3 \times

1256:  10^{-7}$ & 13.0 / 6 & $23 \pm 1$ & $(5.7 \pm 0.7) \times 10^{-9}$ &

1257:  133.3 / 7\\

1258:         &                         & PLE & 0--149 & {\boldmath $50.1 \pm 0.3$} &

1259: {\boldmath $2.38 \pm 0.03$} & {\boldmath $7.5 \times 10^{-7}$$^{\dagger}$} & {\bf

1260:  --} & {\boldmath $13 \pm 2$} & {\boldmath $7.2 \times 10^{-9}$$^{\dagger}$} & {\bf 14.5 / 6}\\\hline

1261: 060202 & 2006-02-02 & XRT & 250--350 & 0$^{\dagger}$ & $2.2 \pm 0.3$ & 

1262:  $8.7 \times 10^{-5}$ & 14.8 / 18 & $138_{-16}^{+20}$ &

1263:  $3.5_{-0.8}^{+1.1} \times 10^{-9}$ & 15.2 / 18\\

1264:        & 08:40:29.9 & PL/EX & 0--350$^{a}$ & $135_{-8}^{+6}$ & $1.2 \pm 0.2$

1265:  & $1.8 \times 10^{-7}$ & 20.0 / 24 & $141 \pm 6$ & $(3.4 \pm 0.3)

1266:  \times 10^{-9}$ & 118.2 / 27\\

1267:        &                         & PLE & 0--350 & {\boldmath $127 \pm

1268:  5$} & {\boldmath $2.4 \pm 0.1$} & {\boldmath $1.3 \times 10^{-5}$$^{\dagger}$} &

1269:  {\bf --} & {\boldmath $141 \pm 5$} & {\boldmath $2.8 \times 10^{-9}$$^{\dagger}$} & {\bf 22.1 / 26}\\\hline

1270: 060211A & 2006-02-11 & XRT & 232--312 & 0$^{\dagger}$ & $2.1 \pm 0.9$ 

1271:  & $1.6 \times 10^{-5}$ & 5.0 / 6 & $127_{-39}^{+96}$ &

1272:  $8.3_{-5.0}^{+12.2} \times 10^{-10}$ & 5.2 / 6\\

1273:         & 09:39:59.9  & PL/EX & 0--913 & 72 & $2.3$ &

1274:  $1.7 \times 10^{-5}$ & 43.8 / 29 & $81 \pm 2$ & 

1275:  $(3.6 \pm 0.2) \times 10^{-9}$ &  333.0 / 32\\

1276:         &             & PLE & 0--913 & {\bf 72} & {\boldmath $2.29 \pm 0.02$} & 

1277: {\boldmath $1.4 \times 10^{-5}$$^{\dagger}$} & {\bf --} 

1278: & {\boldmath $54 \pm 4$} & {\boldmath $3.1 \times 10^{-9}$$^{\dagger}$}

1279:  & {\bf 39.9 / 31}\\\hline

1280: 060306  & 2006-03-06 & XRT & 97--147 & 0$^{\dagger}$ & $3.7 \pm 0.8$ &

1281:  $2.2 \times 10^{-3}$ & 3.1 / 2 & $31_{-6}^{+9}$ &

1282:  $1.7_{-1.0}^{+2.2} \times 10^{-9}$ & 3.3 / 2 \\

1283:         & 00:49:09.3 & PL/EX & 35--256 & 40 & $2.3$ & $1.1 \times 10^{-6}$ & 31.4 / 4 & 

1284: $17.8 \pm 0.7$ & $(3.1 \pm 0.7) \times 10^{-8}$ & 145.6 / 6\\

1285:         &            & PLE & 35--256 & {\bf 40} & {\boldmath $2.33

1286:  \pm 0.02$} & {\boldmath $1.0 \times 10^{-6}$$^{\dagger}$} & {\bf --} & 

1287: {\boldmath $13 \pm 1$} & {\boldmath $2.5 \times 10^{-8}$$^{\dagger}$} &

1288:  {\bf 25.4 / 5}\\\hline

1289: 060418 & 2006-04-18 & XRT & 178--400 & 0$^{\dagger}$ & $2.6 \pm 0.1$ &

1290:  $3.1_{-1.7}^{+3.8} \times 10^{-4}$ & 27.1 / 28 & $101 \pm 6$ &

1291:  $(2.0 \pm 0.3) \times 10^{-9}$ & 31.6 / 28\\

1292:        & 03:05:49.2$^{b}$ & PL/EX & 108--797$^{c}$ & $146 \pm 2$ & 

1293: $1.07 \pm 0.08$ & $2.2_{-0.8}^{+1.2} \times 10^{-8}$ & 25.0 / 30 & $262 \pm 23$ & 

1294: $(3.0 \pm 0.5) \times 10^{-10}$ & 123.0 / 31\\

1295:        &           & PLE & 108--797$^{b}$ & {\boldmath $148.7 \pm 0.3$} & 

1296: {\boldmath $0.89 \pm 0.01$} & {\boldmath $7.5 \times 10^{-9}$$^{\dagger}$} & {\bf --} 

1297: & {\boldmath $43 \pm 3$} & {\boldmath $1.5 \times 10^{-8}$$^{\dagger}$} & {\bf 23.4 / 30}\\\hline

1298: 060427 & 2006-04-27 & XRT & 148--198 & 0$^{\dagger}$ &

1299:  $4.4 \pm 1.3$ & $4.8 \times 10^{-1}$ & 3.7 / 3 & $41_{-10}^{+20}$ & $5.1_{-0.3}^{+14.8}

1300:  \times 10^{-9}$ & 4.0 / 3 \\

1301:        &  11:43:01.0 & PL/EX  & 0--218 & 66 & $2.8$ & $3.7 \times

1302:  10^{-5}$ & 3.5 / 3 & {\boldmath $47 \pm 2$} & {\boldmath $(2.9 \pm 0.4)

1303:  \times 10^{-9}$} & {\bf 5.1 / 6}\\

1304:        &             & PLE & 0--218 & 114 & $6.9$ & 2.7$^{\dagger}$ 

1305: & -- & 47 & $2.8 \times 10^{-9}$$^{\dagger}$ & 3.3 / 5\\\hline

1306: 060428B & 2006-04-28 & XRT & 235--340 & 0$^{\dagger}$ & $5.0 \pm 0.7$ & 

1307:  74 & 7.9 / 7 & $55_{-7}^{+10}$ & $4.1_{-2.1}^{+4.6} \times 10^{-9}$ & 9.4 / 7 \\

1308:         &  08:54:15.2 & PL/EX & 0--340 & 229 & $0.7$ & $3.5 \times

1309:  10^{-10}$ & 2.9 / 6 & {\boldmath $48 \pm 0.9$} & {\boldmath $8.9 \times

1310:  10^{-9}$} & {\bf 10.4 / 8} \\

1311:         &             & PLE & 0--340 & $236_{-2}^{+1}$ & $0.40 \pm

1312:  0.05$ & $6.8 \times 10^{-11}$$^{\dagger}$ & 2.7 / 7 & $39 \pm 2$ & $1.0 \times

1313:  10^{-8}$$^{\dagger}$ & 2.7 / 7\\\hline

1314: 060923C & 2006-09-23 & XRT & 205--529 & 0$^{\dagger}$ & $2.7 \pm 0.3$ &

1315:  $1.7 \times 10^{-4}$ & 3.5 / 5 & $115_{-14}^{+17}$ & $4.0 \times

1316:  10^{-10}$ & 5.4 / 5\\

1317:         & 13:33:10.8 & PL/EX & 0--529 & {\boldmath $-3 \pm 2$} &

1318:  {\boldmath $2.76 \pm 0.02$} & {\boldmath $2.0 \times

1319:  10^{-4}$$^{\dagger}$} & {\bf 3.5 / 6} & $43 \pm 1$ & $(1.3 \pm 0.2)

1320:  \times 10^{-8}$ & 45.3 / 6 \\

1321:         &  & PLE & 0--529 & $-2.9 \pm 6$ & $2.42 \pm 0.03$ & $2.6

1322:  \times 10^{-5}$$^{\dagger}$ & 3.2 / 5 & $33_{-10}^{+4}$ & $9.5 \times 10^{-9}$$^{\dagger}$ &

1323:  3.2 / 5\\\hline

1324: \enddata

1325: \tablenotetext{\ddagger}{XRT: fitting only the XRT data with a PLO model and an EXP model, PL/EX: joint BAT and XRT fit with a PLO model and an EXP model, PLE: joint BAT and XRT fit with a PLEXP model.}

1326: \tablenotetext{\dagger}{Fixed value.}

1327: \tablenotetext{a}{All the BAT data points and the XRT data from $t_{0}$+224 s to 

1328: $t_{0}$+350 s are used in the fit.}

1329:  \tablenotetext{b}{Although {\it battblocks} found a time interval 

1330:  which is 63 second before 

1331:  $t_{0}$, we concluded that this interval is due to the contamination of

1332:  Sco X-1 in the BAT field of view based on the BAT image analysis.}

1333: \tablenotetext{c}{The fit to the XRT data is from $t_{0}$+108 s to $t_{0}$+148 s, 

1334: and from $t_{0}$+797 s to $t_{0}$+400 s.  The BAT data points from $t_{0}$+146 s 

1335: to $t_{0}$+156 s are also included in the fit.}

1336: 

1337: \end{deluxetable}

1338: 

1339: 

1340: %% XRT spectral paramters

1341: \newpage

1342: \begin{deluxetable}{ccccc}

1343: \tablecaption{XRT spectral parameters based on a joint fit to WT and PC 

1344:  data using data above 2 keV. The error in the photon index is quoted at 

1345: the 90\% confidence level.  \label{tab:xrt_spec_para}}

1346: \tablewidth{0pt}

1347: \tablehead{

1348: \colhead{GRB} & 

1349: \colhead{WT Fitting Range} & 

1350: \colhead{PC Fitting Range} & 

1351: \colhead{$\Gamma_{XRT}$} & 

1352: \colhead{$\chi^{2}$/dof}\\

1353: \colhead{} &

1354: \colhead{[s]} &

1355: \colhead{[s]} &

1356: \colhead{} &

1357: \colhead{}}

1358: \startdata

1359: 050803  & 100.1--184.1   & 185.5--12976 & $1.9 \pm 0.2$ & 33.0 / 37\\

1360: 050814  & 166.6--384.9 & 386.5--14133 & $2.1 \pm 0.2$ & 33.5 / 35\\

1361: %050819  & 153.6--209.8 & WT & ${\bf 2.13}_{-0.25}^{+0.57}$ & 5.61 / 7\\

1362: %        & 211.2--10131.4 & WT & $2.77_{-0.8}^{+1.8}$ & 1.34 / 1\\

1363: 050915B & 158.4--313.3 & 229.9--13791 & $1.8 \pm 0.3$ & 9.3 / 11\\

1364: 051109A & 131.4--214.3 & 3444--17540 & $2.2 \pm 0.2$ & 19.0 / 31\\

1365: %        & 3444--17540  & PC & ${\bf 2.17}_{-0.12}^{+0.13}$ & 58.6 / 71\\

1366: %060108  & 110.6--5603 & PC & ${\bf 1.74}_{-0.18}^{+0.81}$ & 4.5 / 2\\

1367: 060109  & 109.8--200.2 & 201.6--12559 & $2.0 \pm 0.2$ & 23.7 / 21\\

1368: 060111B & 89.2--426.7 & 155.4--12655 & $2.2 \pm 0.3$ & 10.0 / 14 \\

1369: 060202  & 175.1--1025 & 1028--13027 & $1.99_{-0.04}^{+0.05}$ & 364.6 /  337\\

1370: 060211A & 137.1--330.9 & 332.5--13054 & $1.8 \pm 0.1$ & 68.6 / 66\\

1371: 060306  & 96.9--174.5 & 175.8--12874 & $2.40 \pm 0.03$ & 9.8 / 15\\

1372: 060418  & 166.1--697.7 & 699.9--12561 & $2.00_{-0.06}^{+0.07}$ & 211.7 / 209\\

1373: 060427  & 148.2--236.5 & 237.9--12691. & $2.0 \pm 0.4$  & 7.7 / 10\\

1374: %060428A & 83.4--122.7 & 124.1--13362 & $2.0 \pm 0.2$ & 35.9 / 26\\

1375: 060428B & 235.6--440.8 & 442.7--12773 & $2.6 \pm 0.3$ & 16.6 / 17 \\

1376: %060502A & 84.3--149.0 & 150.4--6276 & $2.0 \pm 0.5$ & 3.2 / 6\\

1377: %060605  & -- & 126.9--74028 & $2.2 \pm 0.3$ & 10.1 / 10\\

1378: %060708  & 70.5--151.3 & 152.8--6164 & 3.34 & 10.5 / 5\\

1379: %060714  & 120.3--260.8 & 262.3--5833 & $2.0 \pm 0.1$ & 75.2 / 75\\

1380: %060729  & 129.1--353.8 & 355.5--12232 & $2.7 \pm 0.1$ & 122.8 / 114\\

1381: 060923C & 204.6--267.4 & 268.7--12796 & $2.1 \pm 0.4$ & 3.0 / 7

1382: \enddata

1383: \end{deluxetable}

1384: 

1385: \newpage

1386: \begin{deluxetable}{cccccccccccc|cc}

1387: \tabletypesize{\scriptsize}

1388: \rotate

1389: \tablecaption{BAT prompt emission properties, and the estimated Bulk Lorentz 

1390: factor, $\gamma_{0}$, and the radius of the external shock, R$_{0}$ based on 

1391: the external shock model.  The XRT decay index based on the definition

1392:  of \citet{takami2006}, $\alpha_{tail}$, and $\chi^{2}$ of the fit are also shown.  

1393: Errors quoted the 68\% confidence level.  \label{tab:bat_grb_para}}

1394: \tablewidth{0pt}

1395: \tablehead{

1396: \colhead{GRB} & 

1397: \colhead{T$_{90}$} &

1398: \colhead{Model$\star$} & 

1399: \colhead{$\Gamma_{BAT}$} & 

1400: \colhead{$\ep$} &

1401: \colhead{S$_{E}$$^{a}$} &

1402: \colhead{F$^{peak}_{E}$$^{b}$} &

1403: \colhead{S3/S2$^{c}$} &

1404: \colhead{z} &

1405: \colhead{$T_{1/2}$} &

1406: \colhead{$\gamma_{0}^{\ddagger}$} & 

1407: \colhead{$R_{0}$} & 

1408: \colhead{$\alpha_{tail}$} &

1409: \colhead{$\chi^2$/d.o.f.}\\

1410: \colhead{} &

1411: \colhead{[s]} &

1412: \colhead{} &

1413: \colhead{} &

1414: \colhead{[keV]} &

1415: \colhead{} &

1416: \colhead{} &

1417: \colhead{} &

1418: \colhead{} &

1419: \colhead{} &

1420: \colhead{} &

1421: \colhead{[cm]} &

1422: \colhead{} &

1423: \colhead{}}

1424: \startdata

1425: 050803 & 87.9 & PL & $1.39 \pm 0.07$ & -- & $22 \pm

1426:  1$ & $8.1 \pm 0.7$ & $1.5 \pm 0.1$ & -- & 18 & 270 & $2.8 \times 10^{16}$ & -- & --\\

1427: 050814 & 140.6 & PL & $1.8 \pm 0.1$ & -- & $19 \pm 

1428:  1$ & $6.2 \pm 1.2$ & $1.2 \pm 0.1$ & 5.3$^{1}$ & 45 & 240 & $3.0 \times 10^{16}$ & -- & --\\

1429: 050915B & 40.9 & CPL & $1.4 \pm 0.2$ & $60_{-5}^{+7}$ & $34 \pm 1$

1430:  & $17 \pm 1$ & $1.05 \pm 0.04$ & -- & 24 & 242 & $3.0 \times 10^{16}$ & -- & --\\

1431: 051109A & 37.2 & PL & $1.5 \pm 0.1$ & -- & $22 \pm

1432:  2$ & $29 \pm 3$ & $1.4 \pm 0.2$ & 2.346$^{2}$ & 15 & 287 & $2.7 \times 10^{16}$ & $0.59 \pm 0.05$ & 9.7 / 14\\

1433: 060109 & 115.4 & PL & $1.9 \pm 0.1$ & -- & $6.6 \pm

1434:  0.6$ & $3.4 \pm 0.1$ & $1.0 \pm 0.2$ & -- & 23 & 246 & $3.0 \times 10^{16}$ & -- & --\\

1435: 060111B & 58.8 & PL & $1.0 \pm 0.1$ & -- & $16 \pm 1$ & 

1436: $14 \pm 2$ & $2.0 \pm 0.1$ & -- & 9 & 350 & $2.4 \times 10^{16}$ &  $0.9 \pm 0.1$ & 1.0 / 3\\

1437: 060202 & 198.9 & PL & $1.8 \pm 0.1$ & -- & $22 \pm

1438:  1$ & $3.7 \pm 0.8$ & $1.2 \pm 0.1$ & -- & 98 & 143 & $4.3 \times 10^{16}$ & -- & --\\

1439: 060211A & 126.3 & CPL & $0.9 \pm 0.3$ & $58_{-5}^{+8}$ & $16 \pm

1440:  1$ & $3.3 \pm 0.1$ & $1.1 \pm 0.1$ & -- & 24 & 242 & $3.0 \times 10^{16}$ 

1441: & $1.3 \pm 0.2$ & 1.0 / 2\\

1442: 060306 & 61.2 & PL & $1.80 \pm 0.05$ & -- & $21 \pm

1443:  1$ & $47 \pm 2$ & $1.1 \pm 0.1$ & -- & 46 & 190 & $3.5 \times 10^{16}$ & $1.2 \pm 0.2$ & 2.4 / 3\\

1444: 060418 & 103.1 & PL & $1.64 \pm 0.03$ & -- & $80

1445:  \pm 1$ & $49 \pm 2$ & $1.28 \pm 0.03$ & 1.489$^{3}$ & 21 & 226 & $3.1 \times 10^{16}$  & -- & --\\

1446: 060427 & 64 & PL & $1.9 \pm 0.2$ & -- & $5.0 \pm 0.5$ & 

1447: $1.7 \pm 0.7$ & $1.1 \pm 0.2$ & -- & 32 & 217 & $3.2 \times 10^{16}$ & -- & --\\

1448: 060428B & 57.9 & PL & $2.6 \pm 0.1$ & -- & $8.2 \pm 0.5$ & 

1449: $3.4 \pm 0.6$ & $0.7 \pm 0.1$ & -- & 27 & 232 & $3.1 \times 10^{16}$  & $3.0 \pm 0.3$ & 4.7 / 8\\

1450: 060923C & 75.8 & PL & $2.3 \pm 0.1$ & -- & $16 \pm

1451:  1$ & $5.0 \pm 1.5$ & $0.8 \pm 0.1$ & -- & -- & -- & -- & -- & --\\

1452: \enddata

1453: \tablenotetext{\star}{PL: power-law (dN/dE $\sim$ E$^{-{\Gamma_{BAT}}}$),

1454:  CPL: cutoff power-law (dN/dE $\sim$ E$^{-{\Gamma_{BAT}}}$ $\exp(-(2-\Gamma_{BAT})E/\ep)$)} 

1455: %\tablenotetext{*}{Estimated $\ep$ using $\ep$-$\Gamma$ relation.}

1456: %\tablenotetext{\dagger}{Pseudo-z}

1457: \tablenotetext{\ddagger}{Calculated bulk Lorentz factor assuming

1458:  $E_{0}$/n = $1 \times 10^{52}$ erg cm$^{3}$.}

1459: \tablenotetext{a}{Energy fluence in the 15--150 keV band [10$^{-7}$ erg cm$^{-2}$]}

1460: \tablenotetext{b}{1-s peak energy flux in the 15--150 keV band [10$^{-8}$ erg cm$^{-2}$ s$^{-1}$]}

1461: \tablenotetext{c}{Fluence ratio between S3(50--150 keV)/S2(25--50 keV)}

1462: \tablenotetext{1}{\citet{jakobsson2006}}

1463: \tablenotetext{2}{\citet{quimby2005}}

1464: \tablenotetext{3}{\citet{dupree2006}}

1465: \end{deluxetable}

1466: 

1467: %% figures

1468: %%% difference of actual flux vs. extrapolated BAT flux

1469: \newpage

1470: \begin{figure}

1471: \centerline{

1472: \includegraphics[width=8.5cm]{f1a.eps}

1473: \vspace{0.5cm}

1474: \includegraphics[width=8.5cm]{f1b.eps}}

1475: \caption{Schematic figures of the observed spectra with different 

1476:  $\ep$ energies (case 1: $\ep > 150$ keV, case 2: 15 keV $< \ep <$ 150 keV, 

1477: case 3: 10 keV $< \ep <$ 15 keV, case 4: 0.3 keV $< \ep <$ 10 keV, 

1478: and case 5: $\ep <$ 0.3 keV) demonstrating issues with the different 

1479:  extrapolations (left: extrapolating the BAT data down to the XRT energy

1480:  band (BAT-to-XRT extrapolation); 

1481: right: extrapolating the XRT data up to the BAT energy range (XRT-to-BAT 

1482:  extrapolation)).  

1483: The light blue hatched regions show the extrapolated energy band.  

1484: The extrapolated spectra shown with red dotted lines indicate issue with 

1485: the extrapolation.  There should be no issue with the extrapolation 

1486: of the spectra shown in case 1 and 5 (blue dotted lines).  See text for 

1487: details  (section 1).}

1488: \label{fig:prob_extrapolation_bat_xrt}

1489: \end{figure}

1490: 

1491: %%% Fit light curve and BAT/XRT light curve

1492: \newpage

1493: \begin{figure}

1494: \centerline{

1495: \includegraphics[width=7cm,angle=-90]{f2a.eps}

1496: \hspace{0.5cm}

1497: \includegraphics[width=7cm,angle=-90]{f2b.eps}}

1498: \vspace{0.5cm}

1499: \centerline{

1500: \includegraphics[width=7cm,angle=-90]{f2c.eps}

1501: \hspace{0.5cm}

1502: \includegraphics[width=7cm,angle=-90]{f2d.eps}}

1503: \vspace{0.5cm}

1504: \centerline{

1505: \includegraphics[width=7cm,angle=-90]{f2e.eps}

1506: \hspace{0.5cm}

1507: \includegraphics[width=7cm,angle=-90]{f2f.eps}}

1508: \caption{The BAT and XRT composite light curve.  See text for details

1509:  (section 3).}

1510: \label{fig:bat_xrt_lc1}

1511: \end{figure}

1512: 

1513: \newpage

1514: \begin{figure}

1515: \centerline{

1516: \includegraphics[width=7cm,angle=-90]{f3a.eps}

1517: \hspace{0.5cm}

1518: \includegraphics[width=7cm,angle=-90]{f3b.eps}}

1519: \vspace{0.5cm}

1520: \centerline{

1521: \includegraphics[width=7cm,angle=-90]{f3c.eps}

1522: \hspace{0.5cm}

1523: \includegraphics[width=7cm,angle=-90]{f3d.eps}}

1524: \vspace{0.5cm}

1525: \centerline{

1526: \includegraphics[width=7cm,angle=-90]{f3e.eps}

1527: \hspace{0.5cm}

1528: \includegraphics[width=7cm,angle=-90]{f3f.eps}}

1529: \caption{continued.}

1530: \label{fig:bat_xrt_lc2}

1531: \end{figure}

1532: 

1533: \newpage

1534: \begin{figure}

1535: \centerline{

1536: \includegraphics[width=7cm,angle=-90]{f4a.eps}

1537: \hspace{0.5cm}

1538: \includegraphics[width=7cm,angle=-90]{f4b.eps}}

1539: \vspace{0.5cm}

1540: \centerline{

1541: \includegraphics[width=7cm,angle=-90]{f4c.eps}

1542: \hspace{0.5cm}

1543: \includegraphics[width=7cm,angle=-90]{f4d.eps}}

1544: \vspace{0.5cm}

1545: \centerline{

1546: \includegraphics[width=7cm,angle=-90]{f4e.eps}

1547: \hspace{0.5cm}

1548: \includegraphics[width=7cm,angle=-90]{f4f.eps}}

1549: \vspace{0.5cm}

1550: \caption{continued.}

1551: \label{fig:bat_xrt_lc3}

1552: \end{figure}

1553: 

1554: \newpage

1555: \begin{figure}

1556: \centerline{

1557: \includegraphics[width=7cm,angle=-90]{f5a.eps}

1558: \hspace{0.5cm}

1559: \includegraphics[width=7cm,angle=-90]{f5b.eps}}

1560: \vspace{0.5cm}

1561: \centerline{

1562: \includegraphics[width=7cm,angle=-90]{f5c.eps}

1563: \hspace{0.5cm}

1564: \includegraphics[width=7cm,angle=-90]{f5d.eps}}

1565: \vspace{0.5cm}

1566: \centerline{

1567: \includegraphics[width=7cm,angle=-90]{f5e.eps}

1568: \hspace{0.5cm}

1569: \includegraphics[width=7cm,angle=-90]{f5f.eps}}

1570: \caption{continued.}

1571: \label{fig:bat_xrt_lc4}

1572: \end{figure}

1573: 

1574: \newpage

1575: \begin{figure}

1576: \centerline{

1577: \includegraphics[width=7cm,angle=-90]{f6a.eps}

1578: \hspace{0.5cm}

1579: \includegraphics[width=7cm,angle=-90]{f6b.eps}}

1580: \caption{continued.}

1581: \label{fig:bat_xrt_lc5}

1582: \end{figure}

1583: 

1584: % curvature effect check 

1585: \newpage

1586: \begin{figure}[pt]

1587: \centerline{

1588: \includegraphics[width=10cm,angle=-90]{f7.eps}}

1589: \caption{The relationship between the power-law decay index, $\alpha$, of the 

1590: best-fit light curve with a PLO component and a photon index, $\Gamma_{XRT}$.  

1591: The dashed line is the 

1592: expected relationship from the curvature effect ($\alpha = 1+ \Gamma_{XRT}$).  

1593: }

1594: \label{fig:curvature}

1595: \end{figure}

1596: 

1597: \newpage

1598: \begin{figure}[pt]

1599: \centerline{

1600: \includegraphics[width=6.5cm,angle=-90]{f8a.eps}

1601: \hspace{0.5cm}

1602: \includegraphics[width=6.5cm,angle=-90]{f8b.eps}}

1603: \caption{The relationship between the decay constant, $w$, of the 

1604:  exponential model and the BAT prompt emission properties.  Top left: 

1605:  $w$ vs. BAT T$_{90}$, Bottom left: $w$ vs. the fluence ratio between the 

1606:  50-100 keV and 25-50 keV band (the dotted and dash-dotted lines are 

1607: the calculation assuming $\ep=30$ keV and $\ep=100$ keV, respectively,

1608:  with a low and a high energy photon index of 1 and 2.5 in the Band function), 

1609: Top right: $w$ vs. the fluence in the 15-150 keV band, 

1610: Bottom right: $w$ vs. the 1-s peak flux in the 15-150 keV band.}

1611: \label{fig:cor_w}

1612: \end{figure}

1613: 

1614: \newpage

1615: \begin{figure}

1616: \centerline{

1617: \includegraphics[width=12cm,angle=-90]{f9.eps}}

1618: \caption{A histogram of the ratio, in percent, between the fluence 

1619: accumulated from the end of the emission as detected by the BAT to 1000 s  

1620: after $t_{0}$ (S(Tail:15-25 keV)) and the fluence recorded by the BAT 

1621: (S(BAT:15-25 keV)).}

1622: \label{fig:BAT_XRT_fluence_hist}

1623: \end{figure}

1624: 

1625: \newpage

1626: \begin{figure}[pt]

1627: \centerline{

1628: \includegraphics[width=10cm,angle=-90]{f10a.eps}}

1629: \vspace{1cm}

1630: \centerline{

1631: \includegraphics[width=10cm,angle=-90]{f10b.eps}}

1632: \caption{The BAT and XRT composite light curves of GRB 050803 (top) and 

1633: GRB 050814 (bottom) overlaid with the best-fit PLO model for different 

1634: choices of $\alpha$ and $t_{0}^{pow}$.  The labels in the parentheses are 

1635: ($t_{0}^{pow}$, $\alpha$).  The model which represents \citet{liang2006} 

1636: is shown with a blue dashed line.}

1637: \label{fig:linag_prob}

1638: \end{figure}

1639: 

1640: 

1641: %% User Guide for details.

1642: %%

1643: %% No more than seven \figcaption commands are allowed per page,

1644: %% so if you have more than seven captions, insert a \clearpage

1645: %% after every seventh one.

1646: 

1647: %% Tables should be submitted one per page, so put a \clearpage before

1648: %% each one.

1649: 

1650: %% Two options are available to the author for producing tables:  the

1651: %% deluxetable environment provided by the AASTeX package or the LaTeX

1652: %% table environment.  Use of deluxetable is preferred.

1653: %%

1654: 

1655: %% Three table samples follow, two marked up in the deluxetable environment,

1656: %% one marked up as a LaTeX table.

1657: 

1658: %% In this first example, note that the \tabletypesize{}

1659: %% command has been used to reduce the font size of the table.

1660: %% We also use the \rotate command to rotate the table to

1661: %% landscape orientation since it is very wide even at the

1662: %% reduced font size.

1663: %%

1664: %% Note also that the \label command needs to be placed

1665: %% inside the \tablecaption.

1666: 

1667: %% This table also includes a table comment indicating that the full

1668: %% version will be available in machine-readable format in the electronic

1669: %% edition.

1670: %%

1671: \clearpage

1672: 

1673: \end{document}

1674: