1: \documentstyle[emulateapj,psfig]{article}
2: %\documentstyle[emulateapj,apjfonts,psfig]{article}
3: %\documentclass[12pt,preprint]{aastex}
4: %\documentstyle[11pt,epsfig,aaspp4]{article}
5: %\documentstyle[12pt,epsfig,aasms4]{article}
6: %\documentstyle[11pt,epsfig,aaspp4]{article}
7: %\documentstyle[aaspp4]{article}
8: %\documentstyle[aasms4]{article}
9: %\input{epsf}
10: %\voffset=0truein
11:
12: %\newif\ifAMStwofonts
13: %\AMStwofontstrue
14:
15: \begin{document}
16: \title{Temporal Profiles and Spectral Lags of XRF 060218}
17: \author{
18: En-Wei Liang$^{1,2}$, Bin-Bin Zhang $^{1,3,4}$, Mike Stamatikos$^{5}$,
19: Bing Zhang$^1$, Jay Norris$^5$, Neil Gehrels$^5$, Jin Zhang $^{3,4}$,
20: Z. G. Dai$^{6,1}$}
21:
22: \affil{$^1$Department of Physics, University of Nevada, Las Vegas, NV
23: 89154, USA; \\lew@physics.unlv.edu;bzhang@physics.unlv.edu\\
24: $^2$Department of Physics, Guangxi University, Nanning 530004, China\\
25: $^3$National Astronomical Observatories/Yunnan Observatory, CAS,
26: Kunming 650011, China\\
27: $^4$The Graduate School of the Chinese Academy of Sciences,
28: Beijing 100039, China\\
29: $^5$NASA Goddard Space Flight Center, Greenbelt, MD 20771\\
30: $^6$Department of Astronomy, Nanjing University, Nanjing 210093, China}
31:
32:
33: \begin{abstract}
34: The spectral and temporal properties of the non-thermal emission of
35: the nearby XRF 060218 in 0.3-150 keV band are studied. We show that
36: both the spectral energy distribution and the light curve properties
37: suggest the same origin of the non-thermal emission detected by {\em
38: Swift} BAT and XRT. This event has the longest pulse duration
39: and spectral lag observed to date among the known GRBs. The pulse
40: structure and its energy dependence are analogous to typical GRBs. By
41: extrapolating the observed spectral lag to the {\em CGRO/BATSE} bands
42: we find that the hypothesis that this event complies with the same
43: luminosity-lag relation with bright GRBs cannot be ruled out at
44: $2\sigma$ significance level. These intriguing facts, along with its
45: compliance with the Amati-relation, indicate that XRF 060218 shares
46: the similar radiation physics as typical GRBs.
47: \end{abstract}
48:
49: \keywords{gamma-rays: bursts---method: statistical}
50:
51: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
52:
53: \section{Introduction}
54: X-ray flashes (XRFs), cosmic explosions with lower spectral peak energies ($E_p$) in $\nu
55: F_\nu$ spectra than typical gamma-ray bursts (GRBs) (Heise et al. 2001; Kippen et al.
56: 2003), are thought to be the low energy extension of typical GRBs (Lamb et al. 2005;
57: Sakamoto et al. 2004, 2006; Cui et al. 2005). They may be GRB jets (uniform or
58: structured) viewed at off-axis directions (e.g. Zhang et al. 2003; Yamazaki et al. 2004;
59: Zhang et al. 2004) or intrinsically different events (Lamb et al. 2005; Soderberg et al.
60: 2005). It is long speculated that long duration GRBs originate from a relativistic jet
61: emerging from a collapsing massive star progenitor (Woosley et al. 1993; Paczynski 1998;
62: MacFadyen et al. 1999; Zhang et al. 2003), and associations of core-collapsing supernovae
63: (SNe) with long GRB afterglows have been spectroscopically identified in a number of
64: systems, including GRB980425/SN1998bw (Galama et al. 1998), GRB030329/SN2003dh (Stanek et
65: al. 2003; Hjorth et al. 2003), and GRB031203/SN 2003lw (Malesani et al. 2004). The firm
66: spectroscopic association of the nearby XRF 060218 (at z=0.0331, Mirabal et al. 2006)
67: detected by Swift (Campana et al. 2006) with the Type Ic SN 2006aj was established
68: recently (Modjaz et al. 2006; Pian et al. 2006; Sollerman et al. 2006; Mirabal et al.
69: 2006; Cobb et al. 2006; Soderberg et al. 2006). This suggests that the progenitors of
70: both GRBs and XRFs are related to the death of massive stars and XRFs are the low energy
71: extension of the more ``standard'' GRBs.
72:
73: Some empirical relations have been discovered from typical GRBs. It is interesting to
74: verify whether XRF 060218 satisfies these relations as an approach to access its
75: ``standardness''. It has been reported that XRF 060218 complies with the isotropic energy
76: vs. spectral peak energy ($E_{iso}-E_p$) relation derived from long GRBs (Amati et al.
77: 2002) and other XRFs (Amati et al. 2006). The pulses in GRB light curves usually display
78: a fast-rise-exponential-decay (FRED) shape and it has been found that the pulse width is
79: related to the energy of the observational band as $\omega \propto E^{-0.4}$ (Fenimore et
80: al. 1995; Norris et al. 2005). A correlation between the isotropic luminosity and the
81: spectral lag ($L_{\rm iso}-\tau$ relation) of light curves was also discovered with
82: typical GRBs as $L_{\rm iso} \propto \tau^{-1.18}$ (Norris et al. 2000). Both $\omega-E$
83: and $L_{\rm iso}-\tau$ relations are related to the structure of light curves. Since most
84: photons of an XRF are in the X-ray band, previous GRB missions did not observe the real
85: light curves of XRFs. This makes it difficult to identify the temporal structure and the
86: spectral lag of XRFs. XRF 060218 is long and soft, and {\em Swift} BAT and XRT
87: simultaneously collected the data in the gamma-ray to X-ray bands. This makes it possible
88: to measure its temporal structure and to examine whether it complies with the same
89: $\omega-E$ and $L_{\rm iso}-\tau$ relations derived from GRBs. In this {\em Letter} we
90: focus on this issue. Throughout the paper $H_0=71$ km s$^{-1}$ Mpc$^{-1}$,
91: $\Omega_{\rm{m}}=0.3$, and $\Omega_\Lambda=0.7$ are adopted.
92:
93: \section{Data}
94: XRF 060218 was detected with the {\em Swift}/BAT on 2006 February
95: 18.149 UT. It is a long burst, with a duration $T_{90}\sim 2000$
96: seconds in the 15-150 keV band. Swift slewed autonomously to the burst
97: and the X-ray telescope (XRT) and UV/Optical Telescope (UVOT) began
98: collecting data 159 s after the burst trigger.
99:
100: XRF 060218 was an image trigger. The BAT event data lasted only until $t\sim 300$ seconds
101: after the trigger. The survey data are used to derive the BAT light curve. XRF 060218 is
102: very soft, with most of the emission in the BAT band being lower than 50 keV (Campana et
103: al. 2006). In order to obtain a high level of signal-to-noise ratio of the BAT light
104: curve, we use the light curve in whole BAT band, i.e., 15-150 keV. The first orbit of the
105: XRT data is fully in the Windowed Timing (WT) mode. We extract the light curves and
106: spectrum of the XRT data with the Xselect package. The spectrum is grouped with the tool
107: {\em grppha}, and the spectral fitting is carried out with the {\em Xspec} package.
108: Following Campana et al. (2006) we fit the XRT spectrum in the first orbit with a model
109: combining a black body component with temperature $kT$ and a cut-off power law ($F\propto
110: E^{-\Gamma}e^{-E/E_{c}}$) component. The absorption in both Milky Way\footnote{The $N_H$
111: of our Galaxy (Dickey et al. 1990) for this burst is $\sim 1.1\times 10^{21}$ cm$^{-2}$}
112: and the GRB host galaxy ($N_H^{host}$) is incorporated. We obtain
113: $N_H^{host}=0.63^{+0.03}_{-0.03}\times 10^{22}$ cm$^{-2}$, $kT=0.122^{+0.003}_{-0.004}$
114: keV, $\Gamma=1.78^{+0.08}_{-0.08}$, and $E_{c}=21.8^{+15.8}_{-6.7}$ keV (corresponding to
115: $E_p\sim 5$ keV), with reduced $\chi^2=1.55$ for 769 degrees of freedom. We derive the
116: unabsorbed light curves in the 0.3-2 keV, 2-5 keV, and 5-10 keV bands from the XRT data.
117: As reported by Campana et al. (2006), a thermal component exists, and is likely of
118: different origin (e.g. the shock breakout emission associated with SN 2006aj; c.f. Li
119: 2006; Ghisellini et al. 2006b) from the non-thermal component. We therefore subtract the
120: contribution of this component from the observed light curves. The $kT$ and the radiation
121: radii ($R_{BB}$) evolve with time during the first orbit. We read the values of $kT$ and
122: $R_{BB}$ from Campana et al.(2006) and calculate the light curves of this component in
123: 0.3-2, 2-5, and 5-10 keV bands. Since the temperature of the black body component is
124: below $0.2$ keV, the light curves in the energy band higher than 2 keV is essentially not
125: contaminated by the thermal component. In the 0.3-2 keV band, the derived light curve of
126: the thermal component continuously increase with time, which could be well fitted by
127: $\log F_{0.3-2}=(-10.69\pm 0.09)+(0.66\pm 0.03)\log t$ in cgs units. We thus subtract the
128: contribution of this component from the 0.3-2 keV band light curve according to this
129: fitting result. The derived non-thermal light curves in the three XRT bands as well as
130: the one in the BAT band are shown in Fig.\ref{LC}(a). To clearly view the pulse width and
131: spectral lag dependences with energy, we also plot in Fig.\ref{LC}(b) the normalized
132: lightcurves. The characteristics of the light curves along with the average photon energy
133: ($\bar{E}$) of each energy band are reported in Table 1.
134:
135: \section{Results}
136: \subsection{Spectral energy distribution}
137: One important question is whether the XRT non-thermal emission and the
138: BAT emission are of the same origin. In order to clarify this we first
139: study the spectral energy distribution (SED) of the burst using joint
140: BAT-XRT data. Since this event is image trigger, the BAT event file
141: contains only the data in the first 300 seconds since the BAT
142: trigger. XRT began collecting data 159 seconds after the BAT
143: trigger. We thus only obtain simultaneous observations of the two
144: instruments from 160 to 300 seconds with the event files. The
145: joint-fit SED in this period is shown in Fig.\ref{SED}, which is well
146: fitted by the BB+CPL model (Campana et al. 2006) with the following
147: parameters: $N_H^{host}=0.65^{+0.03}_{-0.03}\times 10^{22}$ cm$^{-2}$,
148: $kT=0.13^{+0.01}_{-0.01}$ keV, $\Gamma=1.56^{+0.08}_{-0.07}$ and
149: $E_{c}=122^{+160}_{-46}$ keV (corresponding to $E_p\sim 54$ keV), with
150: reduced $\chi^2=0.97$ for 327 degrees of freedom. The $E_p$ strongly
151: evolves with time, from $54$ keV at the beginning down to $\sim
152: 5$ keV at later times. This is consistent with that reported by Campana
153: et al. (2006) and Ghisellini et al. (2006a). The model
154: fitting results are shown in Fig.\ref{SED}. One can observe that the
155: BAT-component is a good extrapolation of the XRT non-thermal
156: component. This result implies that the non-thermal emissions
157: detected by XRT and BAT are of the same origin.
158:
159: \subsection{Pulse width and Energy Dependence}
160: As shown in Fig.\ref{LC}, the light curves in different energy bands
161: can be all modeled by a single FRED-pulse. Kocevski et al. (2003)
162: developed an empirical expression to fit a FRED-like pulse, which
163: reads,
164: \begin{equation}\label{Kocevski}
165: F(t)={F_m}(\frac{t+t_0}{t_m+t_0})^r[\frac{d}{d+r}+\frac{r}{d+r}(\frac{t+t_0}
166: {t_m+t_0})^{(r+1)}]^{-\frac{r+d}{r+1}},
167: \end{equation}
168: where $t_{m}$ is the time of the maximum flux ($F_{m}$), $t_0$ is the
169: offset time, and $r$ and $d$ are the rising and decaying power-law
170: indices, respectively. We fit the light curves with
171: Eq.(\ref{Kocevski}) and then measure the pulse width, rising and
172: decaying times at the full-width half-maximum (FWHM) of the fitting
173: light curves, and the rising-to-decaying time ratio ($\varphi$). The
174: errors of these quantities are derived from simulations by assuming a
175: normal distribution of the errors of the fitting parameters. The
176: reported errors are at $1\sigma$ confidence level. The results are
177: tabulated in Table 1. We show $\omega$ as a function of $\bar{E}$ in
178: Fig.\ref{FWHM_LAG}(a). Apparently the two quantities are correlated. A
179: best fit yields $\omega\propto E^{-0.31\pm 0.03}$. The $\varphi$
180: parameter ranges from $0.43$ to 0.59. It is found that XRF 060218 roughly
181: satisfies the same $\omega - E$ relation (Fenimore et al. 1995; Norris
182: et al. 2005) and its $\varphi$ values are also well consistent with
183: that observed in typical GRBs (e.g. Norris et al. 1996; Liang et
184: al. 2002), although it has a much longer pulse width than other
185: single-pulse GRBs. These results imply that XRF 060218 may be an
186: extension of GRBs to the extremely long and soft regime.
187:
188: \subsection{Spectral Lag and Energy Dependence}
189: The light curves shown in Fig.\ref{LC}(a) display a significant
190: spectral lag ($\tau$), with soft photons lagging behind the hard
191: photons, as usually seen in long GRBs (Norris et al. 2000; Yi et
192: al. 2006). We illustrate this lag behavior with the
193: intensity-normalized light curves in Fig.\ref{LC}(b). The light curves
194: peak at $405\pm 25$, $735\pm 9$, $919\pm 7$, and $1082\pm 13$ seconds,
195: respectively, in a sequence of high energy band to low energy band as
196: shown in Fig. \ref{FWHM_LAG}(b). The best fit to the correlation
197: between the peak time ($t_{peak} $) and the average photon energy
198: yields,
199: \begin{equation}\label{T_E}
200: \log t_{peak}=(3.04\pm 0.04)-(0.25\pm0.05) \log E.
201: \end{equation}
202: A simple estimate of the lags between any pairs of the 4 light curves obtains
203: $\tau_{peak}=163\sim 677$ seconds, being consistent with that shown in Gehrels et al.
204: (2007). We note XRF 060218 becomes the new record-holder of the long-lag, wide-pulse
205: GRBs. The previous record holder was GRB 971208, with $\tau\sim 58$ seconds and
206: $\omega=395$ seconds (Norris et al. 2005).
207:
208: We also calculate the lags with the cross correlation function (CCF) method. The errors
209: of lags are evaluated by simulations. The results are also reported in Table 1. The lag
210: derived by the CCF method ($\tau_{\rm CCF}$) is strongly correlated with $\tau_{\rm
211: peak}$, but is systematically lower\footnote{We here derive $\tau_{\rm CCF}$ from the
212: peak of the CCF without considering the side lobe contribution of the CCF. A fit to the
213: CCF with a cube or quartic function gives a larger lag by considering the side lobe
214: contribution (Norris et al. 2000). However, this method strongly depends on the
215: artificially-selected range of CCF for the fitting. Since the light curves are a smooth
216: pulse and their lags are significantly larger than the time bin, the peaks of CCFs are
217: robust to estimate the lags.} than $\tau_{\rm peak}$ [Fig. \ref{Liso_lag}(a)]. A best fit
218: gives $\tau_{\rm CCF}=(-100\pm 17)+(0.91\pm 0.08)\tau_{\rm peak}$.
219:
220: The $L_{\rm iso}-\tau$ relation was discovered with six bright BATSE
221: GRBs (Norris et al. 2000) and the spectral lag was defined by the
222: light curves in the 25-50 keV and 100-300 keV bands. We investigate
223: whether the lag behavior of XRF 060218 is consistent with the $L_{\rm
224: iso}-\tau$ relation. Since XRF 060218 is a soft XRF and the emission
225: in the 100-300 keV band is too weak to derive a light curve,
226: we assume that $t_{peak}$ of the light curve in the 100-300
227: keV band follows the $t_{peak}-E$ relation (Eq.\ref{T_E}) and
228: perform the extrapolation. With the extrapolated $t_{peak}$ we
229: then estimate $\tau_{peak}$ for the light curves in the 25-50 keV
230: (average energy 30 keV) and 100-300 keV (average energy 200 keV)
231: bands. We obtain $\tau_{peak}=177\pm 16$ seconds. Since
232: $\tau_{\rm CCF}$ is more reliable, we use the $\tau_{\rm peak}
233: - \tau_{\rm CCF}$ relation [Fig. \ref{Liso_lag}(a)] to derive
234: $\tau_{\rm CCF}=61\pm 26$ seconds. This lag is used in the $L-\tau$
235: relation analysis. Using the peak fluxes in the BAT and XRT band,
236: we estimate $L_{iso}=1.2\times 10^{47}$ ergs s$^{-1}$.
237:
238: Figure \ref{Liso_lag}(b) shows the $L_{iso}-\tau$ relation derived by Norris et al.
239: (2000) compared against XRF 060218 as well as two other nearby GRBs, 980425 and 031203.
240: The data of the previous GRBs are taken from Norris et al. (2000) and Sazonov et al.
241: (2004)\footnote{The peak photon fluxed in the 50-300 keV bands of these GRBs are
242: converted to energy fluxes with the spectral index presented in Friedman \& Bloom (2005),
243: and their $L_{iso}$'s are recalculated with the cosmological parameters used in this
244: paper.}. The grey band and the two dashed lines mark the best fits at the $1\sigma$ and
245: $2\sigma$ confidence level, respectively, and the solid line is the regression line for
246: the 6 GRBs that were used to draw the $L-\tau$ correlation, i.e. $\log L_{\rm
247: iso}=(50.22\pm 0.32)-(1.21\pm 0.21)\times \log \tau$ (errors are at the $1\sigma$ level).
248: We can see that XRF 060218 is definitely inside the $2\sigma$ region and is marginally at
249: the $1\sigma$ region boundary. Therefore, the hypothesis that XRF 060218 follows the
250: $L-\tau$ relation cannot be ruled out at the $2\sigma$ significance level. We caution
251: that the $\tau$ is inferred from the extrapolation of the $t_{\rm peak}-E$ relation. This
252: introduces uncertainties in deriving the lag. The other two nearby GRBs, 980425 and
253: 031203, are out of the $2 \sigma$ region, which are identified as significant outliers of
254: this relation (e.g. Sazonov et al. 2004).
255:
256: \section{Conclusions and discussion}
257:
258: We have investigated the non-thermal emission of XRF 060218. The early SED of this event
259: from 0.3-150 keV observed by BAT and XRT suggests that the non-thermal emission detected
260: by the two instruments are the same component. By subtracting the contribution of the
261: thermal emission we derive the light curves of the non-thermal emission. They are
262: composed of a broad single pulse, and the energy dependences of the widths and the
263: rising-to-decaying-time ratio of the pulses are roughly consistent with those derived in
264: typical GRBs. The light curves show significant spectral lags, with a well-defined peak
265: time sequence from high energy band to low energy bands, i.e. $t_{\rm peak}\propto
266: E^{-0.25\pm0.05}$. We infer the spectral lag in the BATSE bands and find that the
267: hypothesis that this event complies with the $L_{\rm iso}-\tau$ relation with typical
268: GRBs cannot be ruled out at the $2\sigma$ significance level.
269:
270: These intriguing facts, along with its compliance with the Amati-relation, strongly
271: suggest that GRB 060218 is a ``standard'' burst at the very faint, long, and soft end of
272: the GRB distribution. Since all these relations concern the temporal and spectral
273: properties of emission, they are likely related to the radiation mechanisms. The results
274: therefore imply that XRF 060218 and other XRFs may share the similar radiation physics
275: (e.g. synchrotron or inverse Compton scattering in internal shocks, M\'{e}sz\'{a}ros,
276: 2002; Zhang \& M\'{e}sz\'{a}ros 2004; Piran 2005; M\'{e}sz\'{a}ros 2006) with harder GRBs
277: .
278:
279:
280: As discovered by Norris (2002), the proportion of long-lag bursts within long-duration
281: bursts increases from negligible among bright BATSE bursts to $\sim 50\%$ at the trigger
282: threshold, and their peak fluxes are $\sim 2$ orders of magnitude lower than those of the
283: brightest bursts. This argues that they are intrinsically under-luminous. Taken together
284: with the fact that three nearby GRBs, 980425, 031203, and 060218, are long-lagged and
285: under-luminous, an intuitive speculation is that long-lag bursts are probably relatively
286: nearby (e.g., Norris et al. 2005). The local GRB rate of these GRBs thus should be much
287: higher than that expected from the high luminosity GRBs (Liang et al. 2006; see also Cobb
288: et al. 2006; Pian et al. 2006; Soderberg et al. 2006). A possible scenario to explain
289: their wide-pulse, long-lag, and under-luminous features is the off-axis viewing angle
290: effect (e.g. Nakamura 1999; Salmonson 2000; Ioka \& Nakamura 2001). Another scenario is
291: that these features are intrinsic, being due to their lower Lorentz factors (Kulkarni et
292: al. 1998; Woosley \& MacFadyen 1999; Salmonson 2000; Dai et al. 2006; Wang et al. 2006).
293: They might be from a unique GRB population (Liang et al. 2006) having a different type of
294: central engine (e.g. neutron stars rather than black holes) from bright GRBs (e.g.
295: Mazzali et al. 2006; Soderberg et al. 2006).
296:
297:
298:
299: \acknowledgments
300: We thank the anonymous referee for helpful
301: suggestions, and S. Campana, D. Burrows, J. Nousek, K. Page,
302: T. Sakamoto, X.-Y. Wang, and Z. Li for discussion. This work was
303: supported by NASA under grants NNG06GH62G and NNG05GB67G, and the
304: National Natural Science Foundation of China under grants
305: 10463001(EWL).
306:
307:
308:
309: \begin{thebibliography}{99}
310: \bibitem[]{321}Amati, L., et al., 2002, A\&A, 390, 81
311: \bibitem[]{322}Amati, L., et al., 2006, GCN 4846.
312: \bibitem[]{323}Campana, S., et al. 2006, Nature, 442, 1008
313: \bibitem[]{324}Cobb, B. E., et al. 2006, ApJ, 645, L113
314: \bibitem[]{325}Cui, X. H., Liang, E. W, \& Lu R. J., 2005, ChJAA, 5, 151
315: \bibitem[]{326}Dai, Z. G, Zhang, B., \& Liang, E. W. 2006, astro-ph/0604510
316: \bibitem[]{327}Dickey, J. M.\& Lockman, F. J. 1990, ARA\&A, 28,215
317: \bibitem[]{328}Fenimore, E. E., et al. 1995, ApJ, 448, L101
318: \bibitem[]{3281}Friedman, A. S., \& Bloom, J. S. 2005, ApJ, 627, 1
319: \bibitem[]{329}Galama, T. J. et al. 1998, Nature, 395, 670
320: \bibitem[]{3481}Gehrels, G., et al. 2006, Nature, in press (astro-ph/0610635).
321: \bibitem[]{3291}Ghisellini, G., et al. 2006a, MNRAS, in press (astro-ph/0605431)
322: \bibitem[]{3292}Ghisellini, G., Ghirlanda, G., Tavecchio, F. 2006b, MNARS, submitted (astro-ph/0608555)
323: \bibitem[]{330}Heise, J., et al. 2001, in Gamma- Ray Bursts in the
324: Afterglow Era, ed. E. Costa, F. Frontera, \& J. Hjorth (Berlin: Springer), 16
325: \bibitem[]{332}Hjorth, J., et al. 2003, Nature, 423, 847
326: \bibitem[]{333}Ioka, K., Nakamura, T. 2001, ApJ, 554, L163
327: \bibitem[]{334}Kippen, R. M., et al. 2003, in AIP Conf. Proc. 662,
328: Gamma-Ray Burst and Afterglow Astronomy 2001, ed. G. R. Ricker \&
329: R. K. Vanderspek (Melville: AIP), 244
330: \bibitem[]{337}Kocevski, D., Ryde, F., and Liang, E. 2003, ApJ, 596, 389
331: \bibitem[]{3371}Kulkarni, S. R., et al. 1998, Nature, 395, 663
332: \bibitem[]{3371}Li, L. X. 2006, MNARS, submitted (astro-ph/0605387)
333: \bibitem[]{338}Liang, E. W., Xie, G. Z., Su, C. Y. 2002, PASJ, 54, 1
334: \bibitem[]{339}Liang, E. W., Zhang, B., Virgili, F. \& Dai, Z. G. 2006,
335: astro-ph/0605200
336: \bibitem[]{341}Lamb, D. Q., Donaghy, T. Q., Graziani, C. 2005, ApJ, 620, 355
337: \bibitem[]{342}Mazzali, P. A. et al. 2006, Nature, 442, 1011
338: \bibitem[]{343}M\'{e}sz\'{a}ros, P. 2002, ARA\&A, 40, 137
339: \bibitem[]{344}M\'{e}sz\'{a}ros, P. 2006, Rep. Prog. Phys, 69, 2259
340: \bibitem[]{345}MacFadyen, A. I.\& Woosley, S. E. 1999, ApJ, 524, 262
341: \bibitem[]{346}Malesani et al. 2004,ApJ, 609, L5
342: \bibitem[]{3461}Nakamura, T. 1999, ApJ, 522, L101
343: \bibitem[]{347}Mirabal, N., et al. 2006, ApJ, 643, L99
344: \bibitem[]{348}Modjaz, M., et al. 2006, ApJ, 645, L21
345: \bibitem[]{3481}Norris, J. P. et al. 1996, ApJ, 459, 393
346: \bibitem[]{3482}Norris, J. P., 2002, ApJ, 579, 386
347: \bibitem[]{349}Norris, J. P. et al. 2005, ApJ, 627, 324
348: \bibitem[]{350}Norris, J. P., Marani, G. F., Bonnell, J. T. 2000, ApJ, 534, 248
349: \bibitem[]{351}Pacz\'{y}nski, B. 1998, ApJ, 494, L45
350: \bibitem[]{352}Pian, E., et al. 2006, Nature, 442, 1011
351: \bibitem[]{353}Piran, T. 2005, Rev. Mod. Phy., 76, 1143
352: \bibitem[]{354}Sakamoto, T., et al. 2004, ApJ, 602, 875
353: \bibitem[]{355}Sakamoto, T., et al. 2006, ApJ, 636, L73
354: \bibitem[]{356}Salmonson, J. D. 2000, ApJ, 544, L115
355: \bibitem[]{357}Sazonov, S. Yu., Lutovinov, A. A., Sunyaev, R. A.,
356: 2004, Nature, 430, 646
357: %\bibitem[]{359}Schaefer, B. E. 2003, ApJ, 583, L67
358: \bibitem[]{360}Soderberg, A. M., et al. 2005,ApJ, 627, 877
359: \bibitem[]{361}Soderberg, A. M., et al. 2006, Nature, 442, 1014
360: \bibitem[]{362}Sollerman, J., et al. 2006, A\&A, 454, 503
361: \bibitem[]{363}Stanek, K. Z., et al. 2003, ApJ, 591, L17
362: \bibitem[]{364}Wang, X.-Y., Li, Z., Waxman, E. \& M\'esz\'aros, P. 2006,
363: astro-ph/0608033
364: \bibitem[]{366}Woosley, S. E., et al. 1993, ApJ, 405, L273
365: \bibitem[]{3661}Woosley, S. E., \& MacFadyen, A. I. 1999, A\&AS, 138, 499
366: \bibitem[]{367}Yamazaki, R., Ioka, K., Nakamura, T. 2004, ApJ, 607, L103
367: \bibitem[]{368}Yi, T. F., Liang, E. W., Qin, Y. P., Lu, R. J. 2006,
368: MNRAS, 367, 1751
369: \bibitem[]{370}Zhang, B. \& M\'{e}sz\'{a}ros, P., 2004, Int. J. Mod.
370: Phy. A, 19, 2385
371: \bibitem[]{372}Zhang, B. et al. 2004, ApJ, 601, L119
372: \bibitem[]{373}Zhang, W., Woosley, S. E., \& MacFadyen, A. I.2003,
373: ApJ, 586, 356
374: \end{thebibliography}
375:
376: \clearpage
377:
378: \begin{table}
379: \caption{Temporal structures and spectral lags of the light curves} Temporal Structures
380:
381: \begin{tabular}{lllllll}
382: \hline
383: Band & Peak(s) &$\omega$ (s) & Rising time (s)& Decaying time (s) & $\varphi$&$\bar{E}$ (keV)\\
384: \hline
385: (1)15-150 keV &405(25) & 889 (244) & 311 (28) & 578(185) & 0.54(0.18) & 36.9 \\
386: (2)5-10 keV &735(9) & 1278 (45) & 475 (12) & 803(35) & 0.59(0.03) & 6.9\\
387: (3)2-5 keV &919(7) & 1707 (40) & 624 (8) & 1084(34) & 0.58(0.02) & 3.1\\
388: (4)0.3-2 keV &1082(13) & 2625 (125) & 794 (14) & 1831 (112)& 0.43(0.03) & 0.7\\
389: \hline
390: \end{tabular}
391: \\
392: \hspace{6pt}
393:
394: Spectral Lags
395:
396: \begin{tabular}{llllllll}
397: \hline
398: Bands & $\Delta E$(keV)&$\tau_{\rm peak}$(s)&$\tau_{\rm CCF}$(s)&Bands & $\Delta E$(keV) &$\tau_{\rm peak}$(s)&$\tau_{\rm CCF}$(s)\\
399: %&Bands & $\Delta E$ &$\tau_{\rm peak}$&$\tau_{\rm CCF}$\\
400: %& (keV) &(s)&(s)& & (keV) &(s)&(s)& (keV) &(s)&(s)\\
401: \hline
402: (1)-(2)& 30.0 &330(26)&249(37) &(1)-(3)&33.8 &514(26)& 389(47) \\
403: (1)-(4)&36.2 &677(28)&518(70) &(2)-(3)& 3.8 &184(11)&81(12)\\
404: (2)-(4)& 6.2 &347(16)&173(25) &(3)-(4)&2.4 &163(15)&43(8) \\
405: \hline
406: \end{tabular}
407:
408: Note: all the errors are derived by simulations and in $1\sigma$ significance level.
409: \end{table}
410:
411: \clearpage
412:
413: \begin{figure}
414: \plotone{f1.eps}
415: \caption{(a) Unabsorbed light curves of the non-thermal
416: gamma-rays/X-rays in the energy bands of 15-150 keV, 5-10 keV, 2-5
417: keV, and 0.3-2 keV, respectively. The fitting curves with Eq.(1) are
418: plotted. (b) Normalized light curves from the empirical model fitting.
419: \label{LC}}
420: \end{figure}
421:
422: \clearpage
423:
424: \begin{figure}
425: \plotone{f2.eps}
426: \caption{The BAT-XRT joint spectral energy distribution from 160 to
427: 300 seconds since the BAT trigger. The BB+CutoffPL fitting model is
428: also shown.}
429: \label{SED}
430: \end{figure}
431:
432: \clearpage
433:
434: \begin{figure}
435: \epsscale{1.0}
436: \plotone{f3.eps}
437: \caption{The pulse duration (panel a) and the peak
438: time (panel b) as a function of the average photon energy of the
439: non-thermal emission. The solid lines in both panels are the best
440: fits.}
441: \label{FWHM_LAG}
442: \end{figure}
443:
444: \clearpage
445:
446: \begin{figure}
447: \plotone{f4.eps}
448: \caption{Panel (a): Comparison of the spectral lags derived from the
449: peak times and from the CCF method. The solid line is the best
450: fit. Panel (b): Isotropic gamma-ray luminosity as a function of
451: spectral lag. The spectral lags of typical GRBs and GRB 980425 are
452: calculated with the light curves in the 25-50 keV and 100-300 keV
453: bands observed by CGRO/BATSE. The lag of GRB 031203 is calculated with
454: the light curves in the 20-50 keV and 100-200 keV bands. The grey band
455: and the two dashed lines mark the best fits at the $1\sigma$ and $2\sigma$
456: confidence level, respectively, and the solid line is the regression
457: line for the six typical GRBs presented in Norris et al. (2000).}
458: \label{Liso_lag}
459: \end{figure}
460:
461: \end{document}
462: