1: %\documentclass[manuscript]{aastex}
2: \documentclass[12pt,preprint]{aastex}
3: %\documentclass{article}
4: %\usepackage{emulateapj,psfig}
5: %\documentclass[preprint2]{emulateapj}
6: %\usepackage{psfig}
7:
8: %\usepackage{times}
9: %\usepackage{graphicx}
10: %\usepackage[british]{babel}
11: %\usepackage[applemac]{inputenc}
12:
13:
14: \newcommand \gsim{ \lower .75ex \hbox{$\sim$} \llap{\raise .27ex \hbox{$>$}} }
15: \newcommand \lsim{ \lower .75ex\hbox{$\sim$} \llap{\raise .27ex \hbox{$<$}} }
16: \newcommand \ghi{$E_{\rm peak}-E_{\gamma}$}
17: \newcommand \lz{ $E_{\rm peak}-E_{\rm iso}-t_{\rm jet}$}
18: \newcommand \ama{$E_{\rm peak}-E_{\rm iso}$}
19: \newcommand \amaf{$E^{\rm obs}_{\rm peak}$--Fluence}
20: \newcommand \sw{{\it Swift}}
21: \newcommand \sax{{\it Beppo}SAX}
22: \newcommand \he{{\it Hete-II}}
23: \newcommand \ko{{\it Konus-Wind}}
24: \newcommand \ba{BATSE}
25: \newcommand \ep{$E_{\rm peak}$}
26: \newcommand \epo{$E^{\rm obs}_{\rm peak}$}
27: \newcommand \epof{$E^{\rm obs}_{\rm peak}$--Fluence}
28: \newcommand \eiso{$E_{\rm iso}$}
29: \newcommand \eg{$E_{\rm \gamma}$}
30: \newcommand \tj{$\theta_{\rm jet}$}
31: \newcommand \tbo{$t_{\rm b,opt}$}
32:
33: %\shorttitle{GRB precursors}
34: %\shortauthors{Burlon et al.}
35:
36: \begin{document}
37: \title{Precursors in \sw\ Gamma Ray Bursts with redshift}
38:
39: \normalsize \author{D. Burlon\altaffilmark{1,5}, G. Ghrlanda\altaffilmark{1},
40: G. Ghisellini\altaffilmark{1}, D. Lazzati\altaffilmark{2}, L. Nava\altaffilmark{1,3},
41: M. Nardini\altaffilmark{4}, A. Celotti\altaffilmark{4}}
42: \affil{
43: INAF --
44: Osservatorio Astronomico di Brera, via Bianchi 46, I--23807 Merate, Italy\\
45: Universit\`a degli Studi di Milano Bicocca, P.za della Scienza 3, I-20126 Milano, Italy}
46: \altaffiltext{1}{INAF --
47: Osservatorio Astronomico di Brera, via Bianchi 46, I--23807 Merate, Italy}
48: \altaffiltext{2}{JILA, University of Colorado, 440 UCB, Boulder, CO 80309-0440, USA}
49: \altaffiltext{3}{Univ. dell'Insubria, V. Valleggio, 11, I--22100, Como, Italy}
50: \altaffiltext{4}{SISSA/ISAS, Via Beirut 2/4, I-34014 Trieste, Italy}
51: \altaffiltext{5}{Universit\`a degli Studi di Milano Bicocca, P.za della Scienza 3, I-20126 Milano, Italy}
52:
53: %\author{D. Burlon\altaffilmark{1,2}, G. Ghirlanda\altaffilmark{1}, G. Ghisellini\altaffilmark{1},
54: % D. Lazzati\altaffilmark{3}, L. Nava\altaffilmark{1,4}, M. Nardini\altaffilmark{5},
55: % A. Celotti\altaffilmark{5}} \email{E-mail: davide.burlon@brera.inaf.it}
56: %\slugcomment{E-mail: davide.burlon@brera.inaf.it}
57: %\affil{Osservatorio Astronomico di Brera, via E. Bianchi 46, I-23807 Merate, Italy\\
58: %Universit\`a degli Studi di Milano Bicocca, P.za della Scienza 3, I-20126 Milano, Italy}
59:
60: %\and{G. Ghirlanda\altaffilmark{1}}
61: %%\affil{Osservatorio Astronomico di Brera, via E. Bianchi 46, I-23807 Merate, Italy}
62:
63: %\and{G. Ghisellini\altaffilmark{1}}
64: %%\affil{Osservatorio Astronomico di Brera, via E. Bianchi 46, I-23807 Merate, Italy}
65:
66: %\and{D. Lazzati\altaffilmark{3}}
67: %%\affil{JILA, University of Colorado, 440 UCB, Boulder, CO 80309-0440, USA}
68:
69: %\and{L. Nava\altaffilmark{1,4}}
70: %%\affil{Osservatorio Astronomico di Brera, via E. Bianchi 46, I-23807 Merate, Italy\\
71: %%Universit\`a degli Studi dell'Insubria, Via Valleggio 11, I-22100 Como, Italy}
72:
73: %\and{M. Nardini\altaffilmark{5}}
74: %%\affil{SISSA/ISAS, Via Beirut 2/4, I-34014 Trieste, Italy}
75:
76: %\and{A. Celotti\altaffilmark{5}}
77: %%\affil{SISSA/ISAS, Via Beirut 2/4, I-34014 Trieste, Italy}
78:
79: %\altaffiltext{1}{Osservatorio Astronomico di Brera, via E. Bianchi 46, I-23807 Merate, Italy}
80: %\altaffiltext{2}{Universit\`a degli Studi di Milano Bicocca, P.za della Scienza 3, I-20126 Milano, Italy}
81: %\altaffiltext{3}{JILA, University of Colorado, 440 UCB, Boulder, CO 80309-0440, USA}
82: %\altaffiltext{4}{Universit\`a degli Studi dell'Insubria, Via Valleggio 11, I-22100 Como, Italy}
83: %\altaffiltext{5}{SISSA/ISAS, Via Beirut 2/4, I-34014 Trieste, Italy}
84:
85:
86: \begin{abstract}
87: We study a sample of Gamma--Ray Bursts detected by the Swift satellite
88: with known redshift which show a precursor in the {\it Swift}--BAT
89: light curve. We analyze the spectra of the precursors and compare
90: them with the time integrated spectra of the prompt emission. We find
91: neither a correlation between the two slopes nor a tendency for the
92: precursors spectra to be systematically harder or softer than the
93: prompt ones. The energetics of the precursors are large: on average,
94: they are just a factor of a few less energetic (in the source rest
95: frame energy range 15--150 keV) than the entire bursts. These
96: properties do not depend upon the quiescent time between the end of
97: the precursor and the start of the main event. These results suggest
98: that what has been called a ``precursor'' is not a phenomenon distinct
99: from the main event, but is tightly connected with it, even if,
100: in some case, the quiescent time intervals can be longer than 100 seconds.
101: \end{abstract}
102:
103: \keywords{Gamma rays: bursts --- Radiation mechanisms: non-thermal --- X--rays: general}
104:
105: \section{Introduction}
106:
107: What happens in a Gamma--Ray Burst (GRB) before the main $\gamma$--ray
108: prompt event is still an open question.
109: Koshut et al. (1995, hereafter K95) searched in the \ba\ sample for
110: precursors defined as pulses with a peak intensity lower than that
111: of the main GRB and separated from it by a quiescent phase at least as
112: long as the main burst duration. They found that a precursor was
113: present in $\sim 3$\% of the 995 GRBs detected up to May 1994: their
114: duration appeared weakly correlated with those of the main GRBs and on
115: average shorter than that of the burst. The spectral properties of
116: the precursors (based on hardness ratios) showed no relation with
117: those of the GRB, being both softer and harder.
118:
119: Lazzati (2005; L05 hereafter) searched for precursors as weak events
120: {\it preceding} the \ba\ trigger. He found them in $\sim$20\% of the
121: bursts. These precursors account for only a small fraction of the
122: total GRB counts, and their duration show a weak correlation with that
123: of the burst. In contrast to those of K95, the precursors studied by
124: L05 are softer than the main events.
125:
126: K95 also found that the typical precursor--to--burst separation time is
127: $\Delta t \simeq$100 s, whereas L05 showed that most precursors were
128: closer to the onset of the main event (with $\Delta t \simeq$ 30 s).
129: These timescales are computed in the observer frame since we do not know
130: the redshift for most of the \ba\ bursts.
131: This difference could be due to the different definition of
132: precursor--to--burst separation adopted (peak--to--peak separation and
133: interval between the onset times for K95 and L05, respectively). Both
134: analysis revealed, however, that the minimum separation is of the
135: order of 10 s.
136:
137: The main limitations of the above studies were:
138: i) the lack of known distances, that prevented to quantify
139: the absolute energy of the precursor event;
140: ii) the poor spectral characterization of the precursor, that
141: was based on low resolution spectra: typically the spectrum was
142: described by either the hardness ratio (K95) or by a fit to a 3--4
143: channel broad band spectrum (L05).
144:
145: Theoretical models for precursors can be separated into three classes:
146: the ``fireball precursor'' (Li 2007; Lyutikov \& Blandford, 2003;
147: Meszaros \& Rees, 2000; Daigne \& Mochkovitch, 2002; Ruffini et al. 2001);
148: the ``progenitor precursor'' models (Ramirez--Ruiz,
149: MacFadyen \& Lazzati, 2002; Lazzati \& Begelman 2005) and the
150: ``two step engine'' model (Wang \& Meszaros 2007).
151: In the first class the precursor is associated to the
152: initially trapped fireball radiation being released when transparency
153: is reached. In the second class, based on the collapsar scenario, the
154: precursor is identified with the interaction of a weakly relativistic
155: jet with the stellar envelope. A strong terminal shock, breaking out of
156: the envelope, is expected to produce transient emission. In both
157: classes of models the precursors emission is predicted to be thermal,
158: characterized by a black--body spectrum.
159: In the third class the collapse
160: of the progenitor star leads to the formation of a neutron star whose
161: emission would be responsible for the precursor;
162: subsequent accretion onto the neutron star causes its collapse onto a
163: black hole, originating the GRB prompt.
164:
165:
166: Since the launch of the \sw\ satellite (Gehrels et al. 2004),
167: redshifts have been measured for a considerable number of GRBs.
168: It is thus possible to investigate the rest frame properties and
169: calculate the energetics of any precursor associated with them.
170: These of course are key physical quantities for the understanding of
171: their origin.
172:
173: A major issue however is the very definition of ``precursor'', as
174: there is no obvious objective criterion. For this reason we have
175: considered a ``loose'' operative definition for their ``selection'',
176: allowing ourselves to sub--select, {\it a--posteriori}, those
177: events sharing some characterizing property.
178: Thus, in our operative approach, a precursor is defined simply
179: as an initial signal for which:
180: %
181: \begin{itemize}
182: \item the peak flux is smaller than that of the main event in
183: the \sw--BAT 15--350 keV light curve;
184: \item the flux falls below the background level before the start of
185: the main event.
186: \end{itemize}
187: %
188: Our definition is quite similar to that adopted by K95 although we do
189: not require the precursor to precede the bursts by a time delay at
190: least as long as the main burst duration and, differently from L05, we
191: do not impose the condition that a precursor did not trigger the
192: detector.
193: Indeed, such a definition might comprise initial pulses
194: with very short time separation from the start of the main event,
195: making them effectively undistinguishable from first pulses of the
196: burst prompt emission.
197: But how long should the temporal separation be to let us distinguish among
198: the two? Or are there, instead, other (e.g. spectral) properties
199: which can neatly distinguish precursors from main events,
200: independently of the time separation? And are they really physically
201: different? These are some of the issues we are trying to investigate.
202:
203: We adopt a $\Lambda$CDM cosmology with $\Omega_\Lambda=h_0=0.7$ and
204: $\Omega_{\rm M}=0.3$.
205:
206:
207:
208: \section{Already known precursors with redshift}
209:
210: In the literature, there are 5 bursts with precursors and known redshift.
211: GRB~011121 (Piro et al. 2005) and GRB~030329 (Vanderspek et al. 2004)
212: show two possible precursors each, preceding the burst trigger by a
213: few sec. However, these two GRBs do not match our selection criteria
214: as the precursor candidates occur during the rising of the main event.
215: GRB~050820A (Cenko et al. 2006), GRB~060124 (Romano et al. 2006) and
216: GRB~061121 (Page et al. 2007) have precursors which triggered \sw--BAT
217: and preceded the main events by several tens of sec. Although \sw--BAT
218: could not completely follow the main events of GRB~060124 and
219: GRB~050820A, due to the limited burst--mode memory buffer and to the
220: passage over the South Atlantic Anomaly respectively, the \ko\ data
221: complete the light curve allowing to study the main event.
222: In Table~1 their main temporal and spectral properties and energetics are
223: reported.
224:
225: %%----------------------------------------------------
226: %\begin{figure}
227: %\vskip -0.2 true cm
228: %\hskip -0.7 true cm
229: %{\psfig{figure=dist_ene.ps,width=9cm,height=6cm}}
230: %\vskip -0.2 true cm
231: %\caption{
232: %Distribution of isotropic equivalent energy in the
233: % 15--150 keV rest frame band. Different colours correspond to
234: % different samples: grey: total Swift GRBs with $z$ (excluding the 15
235: % with precursors - from Sakamoto et al. 2008); red: precursors; blue:
236: % GRBs main event, after the precursor. }
237: %\label{distro}
238: %\end{figure}
239:
240: %%\begin{figure}
241: %%\vskip -0.5 true cm
242: %%\hskip -0.7 true cm
243: %%{\psfig{figure=z.ps,width=9cm,height=7cm}}
244: %%\vskip -0.5 true cm
245: %%\caption{ Redshift distribution of the 16 GRBs with precursors studied
246: %% in this work compared with the redshift distribution of \sw\ long
247: %% bursts (89 GRBs). {\bf DA TAGLIARE SE NON SERVE... } }
248: %%\label{z}
249: %%\end{figure}
250: %%----------------------------------------------------
251:
252: \section{Precursors in the \sw\ sample}
253:
254: We have searched for precursor activity in all GRBs with measured
255: redshift detected by \sw\ up to March 2008, comprising 105 GRBs. By
256: applying our definition criteria, we found 15 GRBs with a precursor,
257: including the three GRBs (061121, 060124 and 050820A) already
258: discussed in the literature. All of them are long GRBs, i.e.
259: $T_{90} > 2$ s.
260: Since GRB 070306 has two precursors, this implies a total of 16
261: precursors in the \sw\ sample.
262: %
263: % Among the 105 events, 4 cases (060505, 060605, 061210 and
264: % 071112C) did not have available burst--mode event data files, from the
265: % published literature (Hullinger et al. 2006,
266: % for GRB~060505; Sato et al. 2006 for GRB~060605;
267: % Cannizzo et al. 2006
268: % for GRB~061210; Stratta et al. 2007 for 071112C),
269: % we can exclude the presence of precursors associated to them. We
270: % point out that for one event (out the 105), namely GRB 050709, the
271: % trigger was given by \he\ so that \sw\ observations started too late
272: % for any prompt detection.
273: %The found 15 bursts are all GRBs with redshift having an initial pulse
274: %fulfilling our definition of precursor.
275: % Adding GRB 030329 and GRB
276: % 011121, detected by \he\ and {\it Beppo}SAX, respectively, we have a
277: % sample of 17 GRBs.
278: % The \sw--BAT data (15--150 keV) allow to study the precursor spectrum
279: % and compare it with the spectrum of the main GRB event.
280: %
281: We have applied the standard \sw--BAT data reduction pipeline (v.2.8)
282: to extract light curves and spectra for the GRBs in our sample.
283: We computed the
284: precursor and main GRB duration $T_{90}$ from the background
285: subtracted 15--350 keV light curves binned at 1 s. The precursor and
286: main GRB spectra were obtained with the standard procedure, taking into
287: account the energy dependent systematic errors.
288: The spectral analysis was performed with {\it Xspec v11.3.2}. The
289: spectra were fit with a single (PL) and a cutoff (CPL) power--law
290: model. For the precursors the PL model provides the best fit, i.e. the
291: fit with the CPL model does not statistically improve
292: (at the 3$\sigma$ confidence level).
293:
294: Since, in some theoretical model, the precursor emission is expected to have a thermal
295: origin, we also fit their spectra with a black--body (BB). For
296: the 9 precursors with the largest ($>$ 12) signal--to--noise ratio S/N
297: (integrated over 15--150 keV), the BB representation is excluded at
298: more than 3$\sigma$ in 6 cases and between 2 and 3$\sigma$ in 3 cases. In
299: GRB 060115 and GRB 071010B (S/N$>$10) the residuals of the BB fit show
300: systematic deviations at low and high energies. For these two precursors,
301: an hybrid BB+PL model (Ryde 2005, but see Ghirlanda et al. 2007)
302: yielded a BB component contributing $\sim$50\%
303: of the total flux, but this model was only 1$\sigma$ significantly
304: better than the single PL model. For the remaining 5 precursors the
305: low S/N ($<$10) does not allow to discriminate between the BB, PL, or
306: other models with the same number of free parameters.
307:
308: \section{Results}
309:
310: In Fig.~\ref{alfa} (left panel) the photon spectral indices of
311: precursor ($\alpha_{\rm prec}$) and main GRB ($\alpha_{\rm GRB}$) are
312: compared.
313: % The bursts already published in the literature are shown
314: % with filled symbols while the \sw\ bursts analysed in this paper are
315: % shown with open symbols.
316: There is no clear tendency for the precursor emission to be harder or softer
317: than the prompt. The typical photon index
318: distributions of precursors and main events are both fully consistent
319: with that for the whole \sw\ sample recently published (Sakamoto et
320: al. 2008).
321:
322: As all of the precursor spectra are best fitted by a single power--law,
323: it is not possible to determine either the peak energy in $\nu F_{\nu}$
324: or the bolometric energy $E_{\rm iso}$.
325: As the best possible proxy for the latter, we consider
326: the energy emitted, in the rest frame, between 15 and 150 keV.
327: As shown in Fig.~\ref{alfa} (right panel) the precursor
328: isotropic energy is on average $\sim$1/3 of
329: that of the corresponding main GRB event.
330:
331: %For comparison {\bf ??} we estimated $E_{\rm iso}$ in the 15-150 keV
332: %band for GRB 011121 and GRB 030329 using the spectral information
333: %available in Piro et al. (2005) and Vanderspek et al. (2004). Their
334: %precursor spectra are softer than the precursor spectra in our sample;
335: %also they are $\sim$30 times less energetic than the main events.
336:
337: % Exceptions are GRB
338: % 030329 and GRB 011121: in these cases the precursor is $\sim$30 time
339: % less energetic than the main event and it has the smallest time delay
340: % of the sample. The spectra of the precursors for of these two bursts
341: % are the softest among the bursts in our sample and also with respect
342: % to their prompt.
343: % However we have to point out that if we had applied our
344: % standard selection criteria, both GRB 030329 and GRB 011121 would have been
345: % rejected as candidates; we decided to put them in the sample because they
346: % were already published as ``precursor candidates''.
347:
348: In order to examine the possible role of the duration of the quiescent
349: time, i.e. the time delay $\Delta t$ between the end of the precursor
350: and the start of the main event, we have divided the sample into three
351: subsets, according to $\Delta t$ calculated in the source rest frame
352: [$\Delta t \equiv (T_{\rm 1, main} -T_{\rm 2, prec})/(1+z)$], $T_{\rm
353: 1, main}$ and $T_{\rm 2, prec}$ are reported in Tab.~1). $\Delta t$
354: is broadly distributed between a few sec and a few tens of sec with an
355: average value $\sim$10 s. By comparing the behavior of the
356: precursors belonging to the three subgroups we can check whether our
357: sample is ``contaminated'' by initial ``pulses'' that possibly have
358: properties and origin different from those of ``true'' precursors.
359: Both panels of Fig. \ref{alfa} -- where the events are coded according
360: to $\Delta t$ (i.e. $\Delta t < 15$ s, 15 s $<\Delta t<40$ s and
361: $\Delta t> 40 $ s) -- show that there is no clear separation among
362: them.
363: In terms of energetics, a K--S test on the distributions of $E_{\rm
364: iso}$ for \sw\ GRBs with redshift (adapted from Sakamoto et al. 2008)
365: with and without precursors indicates that they are consistent with
366: being drawn from the same distribution (null hypothesis probability $P$=3\%).
367: As expected the corresponding distribution for the precursors
368: is shifted towards lower $E_{\rm iso}$.
369:
370: % In Fig. \ref{z} we compare the redshift distributions of bursts with
371: % precursor and the total $z$ distribution of \sw\ bursts. One can see
372: % that there is no marked preference, for precursor GRBs, to be closer,
373: % in agreement with the fact that their total energetics is no different
374: % from the other GRBs.
375:
376: Finally, an analysis of the rest frame pulse durations $T_{90}$
377: supports the finding by L05, namely the existence of a tentative
378: (1$\sigma$ significant) correlation between the $T_{90}$ of precursors and
379: the $T_{90}$ of main events.
380:
381: %--------------------------------------------------
382: %\begin{figure*}
383: %\hskip -0.5 true cm
384: %{\psfig{figure=light.ps,width=18cm,height=14.5cm}}
385: %\caption{ Light curves of the GRBs studied in this paper. Rate is in
386: % units of counts per seconds per illuminated detector. Vertical dashed
387: % lines corresponds to the precursor and main event integration time
388: % for the spectral analysis. Time is the observed time. }
389: %\label{lc}
390: %\end{figure*}
391: %----------------------------------------------------
392:
393: \section{Conclusions}
394:
395: Our results point to a clear but puzzling conclusion: the spectra and
396: energetics of the selected initial pulses, being them bona fide
397: precursors or not, are indistinguishable from those of the main event.
398: While this could be not surprising for ``precursors'' which were in
399: fact the initial pulse of the main event, in cases like GRB 060124 and
400: GRB 050820A the precursor precedes the main event by $\sim$100 s (rest
401: frame time): yet they behave as the rest of the main emission, like
402: ``normal'' initial pulses.
403:
404: This forces us to re--consider what the very same precursor phenomenon
405: is. Our finding contrasts with that by L05, who found precursors much
406: fainter and significantly softer than the main event. However the
407: precursor selection criteria are different and in particular the
408: requirement by L05 that the precursors did not trigger \ba\ obviously
409: biased the sample against strong precursors. Our result are instead
410: more consistent with the Koshut et al. (1995) one, whose selection
411: criteria is similar to ours.
412: We therefore cannot exclude that there are two kinds of ``precursors'':
413: one as strong as and spectrally similar to the main event and the other being
414: softer and dimmer.
415: {\it But -- independent of that -- both can occur $\sim$100 s
416: before the main event.}
417: Indeed, this long delay is both the most intriguing feature and the
418: main difficulty for all the proposed progenitor interpretation. As
419: discussed by Wang \& Meszaros (2007), the progenitor class of
420: models cannot explain delays longer than $\sim$10 s.
421:
422: The origin of quiescent times has been discussed by Ramirez--Ruiz,
423: Merloni \& Rees (2001), who considered the possibility that a
424: temporal modulation in Lorentz factor of ejected shells/relativistic
425: outflow would lead to time dependent emission via dissipation in
426: internal shocks. ``Fireball'' models predict too short quiescent
427: timescales if the main prompt emission mechanism is internal shocks
428: taking place at typical radii $R= 10^{13}R_{13}$ cm, since $\Delta
429: t\sim R/(c\Gamma^2)\sim 0.03 R_{\rm 13}/\Gamma^2_2$ s.
430: %
431: % Both in the ``fireball'' and
432: % ``progenitor'' class of models, in fact,
433: % one expects delays of the order of a second, or even less, between
434: % the precursor and the main event, since in both cases the precursor
435: % is associated either with the fireball itself or the breaking
436: % out of the fireball through the star surface.
437: % If this fireball produces the first part of the main event at
438: % the canonical radii of the order of $R\sim 10^{13} R_{13}$ cm,
439: % then the associated delay must be of the order of
440: % $\Delta t\sim R/(c\Gamma^2)\sim 0.03 R_{\rm 13}/\Gamma^2_2$ s,
441: % where $\Gamma=10^2\Gamma_2$ is the bulk Lorentz factor.
442: %
443: External shocks occurring at $R\sim 10^{16}$ cm can lead to
444: % instead $\Delta t\sim 33 R_{\rm 16}/\Gamma^2_2$ s,
445: time delays similar to what observed, but -- in the case of a
446: homogeneous fireball interacting with an homogeneous interstellar
447: medium (ISM) -- this process hardly accounts for fast prompt
448: variability, suppressed by the curvature effect. More complex
449: external shock scenarios can overcome this problem (e.g. Dermer et
450: al. 1999), but in turn require a strongly clumped ISM.
451:
452: A second issue emerging from our results concerns the spectral shape
453: of the precursor. The non--thermal appearance of the spectra is not
454: the chief problem, as this may arise as convolution of black-body
455: emission at different temperatures and/or from different locations,
456: consistently with the predicted thermal character. What
457: remains puzzling (or revealing) is that, on average, the power--law fit
458: spectral indices are very similar to those of the main event.
459: The large energetics of the precursors studied here is also difficult
460: to explain within the precursor models proposed so far as, whatever the
461: progenitor nature, they rival the main event energetics.
462:
463: In the collapsar model, the precursor photons may be produced in a
464: region emerged from the progenitor star. Indeed, heated cocoon
465: material has been proposed as responsible for the precursor
466: (Ramirez-Ruiz, McFadyen \& Lazzati 2002), but the expected energetics
467: would be low compared to our findings. This also applies to the
468: scenario proposed by Lazzati \& Begelman (2005), where the jet opening
469: angle increases in time, so an observer off axis could detect the
470: prompt emission after the precursor, when the jet angle becomes equal
471: to the viewing angle.
472: % ``Fireball models'', instead, must take into account that
473: % we must add, to the precursor energy we see, also the energy spent to
474: % emerge from the stellar interior. This can amount to a few $\times
475: % 10^{51}$ erg (Lazzati ???? Woosley???)
476: The ``two steps'' engine model (Wang \& Meszaros 2007) envisages that
477: the precursor is associated with the cooling phase of the
478: proto--neutron star and the delay time should correspond to the
479: accretion phase which ultimately leads to the collapse of the neutron
480: star to a black hole, when the ``normal'' GRB activity begins.
481:
482: An alternative possibility is that precursors do not represent any
483: distinct physical process, but are simply a manifestation of the same
484: phenomenon producing the prompt emission, which sometimes does give
485: raise to quiescent intervals between emission peaks.
486: We can put an upper limit on the energy emitted during these quiescent times
487: % Assuming
488: % that the central engine may be active for such long time intervals
489: % we estimated the upper limits on the energy that can be released
490: % during ``quiescence'' in the form of extended emission.
491: considering that the BAT sensitivity for a 5$\sigma$ detection is
492: $2\times 10^{-10} (dT/20~ {\rm ks})^{-0.5}$ erg cm$^{-2}$ s$^{-1}$
493: (Markwardt et al. 2007), where $dT$ is the exposure time. Using the
494: delay times reported in Tab. 1 we estimate the mean value of the
495: 1$\sigma$ upper limit to the energy emitted during the quiescence. The
496: ratio of these limits to the precursors energy ranges from 0.012 for
497: GRB~061007 to 0.25 for GRB~060124, with an average of 0.14.
498:
499: Finally, we applied the same selection criterion, adopted for
500: precursors, to search for emission episodes (``postcursors'')
501: following the main bursts and separated by a quiescent phase. Within
502: the 15 GRBs with precursors, GRB~060210 and GRB~0508020A show two and
503: three pulses after the main burst\footnote{We recall
504: that for GRB~050820A, \sw\ entered the {\it SAA}
505: during the main event and therefore we could not perform the
506: spectral analysis of the BAT data.}. For GRB~060210 the two
507: postcursors (separated by 60 and 150 s from the end of the main burst)
508: have spectral indices $-1.76 \pm 0.28$ and $-1.83\pm0.39$ and
509: energetics $(7.31 \pm 2.29)\times 10^{51}$ erg and
510: $(5.04\pm2.14)\times 10^{51}$ erg, respectively. The spectra are
511: softer and the energetics smaller than the main event and
512: the precursor. Since this is the only burst, in our sample, having
513: both a precursor and a postcursor, we cannot draw any strong
514: conclusion. We plan to study spectra and energetics of postcursors by
515: relaxing the condition of having also a precursor in a forthcoming
516: paper.
517:
518: %We intend to investigate this
519: %possibility in a forthcoming paper {\bf extending the postcursor
520: % sample to the complete parent population of \sw\ GRBs with z. We
521: % find 23 GRBs with postcursor emission}.
522:
523: % This is probably the most promising model for explaining the
524: % origin of the precursors, and we plan to investigate it further
525:
526: % Previous and our results, concerning the long $\Delta t$, and the
527: % spectral results and energetics derived in this paper are difficult to
528: % explain in the framework of existing models. Thus one is tempted to
529: % associate the found ``precursors'' to the beginning of the prompt
530: % emission. In this case the central engine, in a fraction ($\sim$15\%)
531: % of bursts, becomes to be active, then stops, and after some time it
532: % starts again. Since the ``quiescent'' time can be quite long, this
533: % gives us the impression that there is some separate phenomenon
534: % occurring much earlier than the start of the main event. In other
535: % words, in this case the ``precursor'' is nothing more than the first
536: % pulse of the burst, produced by the same central engine. If true,
537: % this implies that in some cases the central engine can be quiescent
538: % for a very long time (i.e. hundreds of seconds) before re--starting.
539: % The problem, in this case, is to explain why the central engine can
540: % stop for such a long time. Is it a phase of no injection of energy or
541: % merely a phase in which the energy is not transformed into radiation?
542: % On energetic grounds, the first option is to be preferred, since if
543: % the injected energy continues unaltered during quiescence, then we
544: % easily arrive to really huge total energetics.
545:
546: %We believe that our results add puzzling features to an already
547: %puzzling issue, with no clear way out. {\it In a companion paper we
548: % plan to characterize even better the precursors found in this work,
549: % investigating the possibility that they are responsible for the
550: % optical flux observed (when it was possible) between the precursor
551: % and the main event, and discussing yet another theoretical idea for
552: % the origin of the ``precursors" we see. }
553: %{\bf SI PUO' METTERE O TOGLIERE....}
554:
555:
556:
557: \acknowledgements
558: We thank the referee for constructive comments.
559: This research was partly supported by PRIN--INAF 2008 and ASI
560: I/088/06/0 grants. We acknowledge the use of public data from the
561: \sw\ data archive.
562:
563:
564:
565: \begin{thebibliography}{}
566: % \bibitem[]{} Cannizzo J.K. et al., 2006, GCN rep. 20.1 % GRB~061210;
567: \bibitem[]{} Cenko S.B., et al., 2006, ApJ, 652, 490
568: \bibitem[]{} Daigne F., Mochkovitch R., 2002, MNRAS, 336, 1271
569: \bibitem[]{} Dermer C., B\"ottcher M. \& Chiang J., 1999, ApJ, 515, L49 % external shocks
570: \bibitem[]{} Gehrels N., et al., 2004, ApJ, 611, 1005
571: \bibitem[]{} Ghirlanda G., Bosnjak Z., Ghisellini G., Tavecchio F. \& Firmani C.,
572: 2007, MNRAS, 379, 73
573: \bibitem[]{} Ghirlanda G., Nava L., Ghisellini, G., Firmani C. \& Cabrera J.I.,
574: 2008, MNRAS, in press (arXiv:0804.1675)
575: % \bibitem[]{} Hullinger D. et al., 2006, GCN 5142 % GRB~060505
576: \bibitem[]{} Koshut T., et al., 1995, ApJ, 452, 145 (K95)
577: \bibitem[]{} Lazzati D., 2005, MNRAS, 357, 722 (L05)
578: \bibitem[]{} Lazzati D. \& Begelman M.C., 2005, ApJ, 629, 903
579: \bibitem[]{} Li L.--X., 2007, MNRAS, 380, 621
580: \bibitem[]{} Lyutikov M. \& Blandford R.D., 2003, preprint (astro-ph/0312347)
581: \bibitem[]{} Markwardt C.B.,
582: %Barthelmy S.D., Cummings J.C.,
583: et al., 2007, ``The Swift BAT Software Guide'' (v. 6.3) (swift.gsfc.nasa.gov/docs/swift/analysis)
584: \bibitem[]{} Meszaros P. \& Rees M.J., 2000, ApJ, 530, 292
585: \bibitem[]{} Page K.L., et al., 2007, ApJ, 663, 1125
586: \bibitem[]{} Piro L., et al., 2005, ApJ, 623, 314
587: \bibitem[]{} Ramirez--Ruiz E., Merloni A. \& Rees M.J., 2001, MNRAS, 324, 1147
588: \bibitem[]{} Ramirez--Ruiz E., MacFadyen A.I. \& Lazzati, 2002, MNRAS, 331, 197
589: \bibitem[]{} Romano P. et al., 2006, A\&A, 456, 917
590: \bibitem[]{} Ruffini R., Bianco C.L., Fraschetti F., Xue S.-S., Chardonnet P., 2001,
591: ApJ, 555, L113
592: \bibitem[]{} Ryde F., 2005, ApJ, 625, 95
593: \bibitem[]{} Sakamoto T., et al., 2008, ApJS, 175, 179
594: %\bibitem[]{} Sato G. et al., 2006, GCN 5231 % GRB~060605
595: % \bibitem[]{} Stratta G. et al., 2007, GCN rep. 104.2
596: \bibitem[]{} Vanderspek R., et al. 2004, ApJ, 617, 1251
597: \bibitem[]{} Wang X.--Y. \& Meszaros P., 2007, ApJ, 670, 1247
598:
599: \end{thebibliography}
600:
601:
602: \clearpage
603:
604:
605: \setcounter{table}{0}
606: \begin{table*}
607: \hskip -1.7 true cm
608: \begin{footnotesize}
609: %\begin{small}
610: \begin{tabular}{|llllllllllll|}
611:
612: \hline
613: GRB &~~$z$ & \vline & && Main pulse & &\vline & & & Precursors & \\
614: & & \vline &$T_1$ &$T_2$ &~~~~~~$\alpha$ &~~~~~~$E_{\rm iso}^a$ &\vline&$T_1$ &$T_2$ &~~~~~~$\alpha$ &~~~~~~$E_{\rm iso}^a$ \\
615: & &\vline &s &s & &~~~~~~erg &\vline&s &s & &~~~~~~erg \\
616: \hline
617: 050820A$^*$ &2.6 & \vline &225 &553 &--1.12$^{+0.13}_{-0.15}$ &1.07($\pm$0.23)E53$^a$&\vline
618: &--17 &22 &--1.74$\pm$0.08 &1.14($\pm$0.16)E52$^a$\\
619: 060124$^+$ &2.297 & \vline &301.2 &811.2&--1.48$\pm$0.02 &1.02($\pm$0.14)E53$^b$&\vline
620: &--1.5 &13.5 &--1.80$\pm$0.20 &4.57($\pm$0.75)E51$^b$\\
621: 061121$^{+ \#}$ &1.314 & \vline &61.8 &83.38&--1.32$\pm$0.05 &4.19($\pm$0.67)E52$^c$&\vline
622: &--5 &10 &--1.68$\pm$0.09 &1.37($\pm$0.22)E51$^d$\\
623: \hline
624: 071010B &0.947 &\vline &--1.5 &15.23 &--2.03$\pm$0.04 &5.40($\pm$0.13)E51 &\vline
625: &--30.0 &--12.77 &--1.76$\pm$0.19 &7.21($\pm$0.11)E51\\
626: 070411 &2.954 &\vline &49.3 &98.3 &--1.65$\pm$0.11 &1.41($\pm$0.93)E51 &\vline
627: &--19.7 &31.3 &--1.64$\pm$0.14 &1.28($\pm$0.09)E52\\
628: 070306 &1.49 &\vline &83.5 & 154.5 &--1.64$\pm$0.06 &1.10($\pm$0.05)E52 &\vline
629: &--118.5 &--103.5 &--1.40$\pm$0.65 &3.77($\pm$2.34)E50\\
630: & &\vline & & & & &\vline
631: &--12.48 &40.51 &--1.59$\pm$0.26 &2.99($\pm$0.82)E51\\
632: 061007 &1.261 &\vline &27.2 &71.2 &--0.94$\pm$0.03 &6.06($\pm$0.08)E52 &\vline
633: &--2.8 &12.2 &--1.07$\pm$0.06 &7.08($\pm$0.18)E51\\
634: 060729 &0.54 &\vline &56.9 &123.9 &--1.74$\pm$0.11 &1.23($\pm$0.14)E51 &\vline
635: &--1.1 &29.4 &--1.8$\pm$20.75 &2.34($\pm$1.36)E50\\
636: 060714 &2.711 &\vline &69.9 &116.9 &--1.30$\pm$0.47 &1.90($\pm$0.16)E52 &\vline
637: &--13.1 &43.9 &--1.86$\pm$0.20 &1.35($\pm$0.25)E52\\
638: 060707 &3.425 &\vline &--7.3 &49.7 &--1.70$\pm$0.15 &1.97($\pm$0.14)E52 &\vline
639: &--48.3 &--23.3 &--1.69$\pm$0.35 &4.00($\pm$1.82)E51\\
640: 060210 &3.91 &\vline &--72.3 &21 &--1.39$\pm$0.08 &5.15($\pm$0.34)E52 &\vline
641: &--236.3 &--200.3 &--1.40$\pm$0.33 &1.07($\pm$0.20)E52\\
642: 060115 &3.53 &\vline &78.9 &129.9 &--1.63$\pm$0.11 &1.99($\pm$0.24)E52 &\vline
643: &--22.1 &30.9 &--1.82$\pm$0.19 &1.34($\pm$0.34)E52\\
644: 050401 &2.90 &\vline &20.7 & 29.71 &--1.43$\pm$0.12 &1.18($\pm$0.14)E52 &\vline
645: &--7.28 &6.71 &--1.45$\pm$0.10 &2.01($\pm$0.20)E52\\
646: 050318 &1.44 &\vline &22.9 &29.9 &--1.94$\pm$0.09 &1.83($\pm$0.11)E51 &\vline
647: &--1.1 &5.8 &--2.11$\pm$0.24 &9.63($\pm$2.59)E50\\
648: 050315 &1.949 &\vline &--6.4 &52. &--2.16$\pm$0.09 &2.42($\pm$0.10)E52 &\vline
649: &--57.5 &--25.5 &--1.72$\pm$0.30 &2.37($\pm$0.11)E51\\
650: \hline
651: \end{tabular}
652: \caption{Data from Swift/BAT except for: $^*$ Konus-Wind; $^+$
653: Konus--Wind, precursor from Swift.
654: $T_1$ and $T_2$ are in the observer frame.
655: $E_{\rm iso}$ is computed in the 15--150
656: keV rest frame band, except for $^*$(20-1000 keV), $^+$(20-2000 keV)
657: and $^\#$(20-5000 keV).
658: Peak energies $E_{\rm peak}$ (keV) of main
659: pulses: 367$^{+95}_{-62}$ (050820A); 193$^{+78}_{-39}$ (060124);
660: 557$\pm$66 (061121); 41.0$\pm$8.5 (060714). This corresponds to a
661: cut--off power-law model for all bursts. Errors are given at $90\%$
662: confidence level. References: (a) Cenko et al. (2006); (b) Romano et
663: al. (2006); (c) Ghirlanda et al. (2008, and references therein); (d)
664: Page et al. (2007). }
665: \label{tab1}
666: \end{footnotesize}
667: %\end{small}
668: \end{table*}
669:
670: \clearpage
671:
672: %--------------------------------------------------
673: \begin{figure}
674: %\begin{center}
675: \hskip -2.1 true cm
676: %\vskip -0.5 true cm
677: \scalebox{1.31}{\plottwo{f1a.eps}{f1b.eps}}
678: %\end{center}
679: \caption{
680: {\it Left panel}:
681: Precursor versus burst photon spectral index.
682: Different symbols correspond to
683: different {\it rest frame} time separation $\Delta t$, between the
684: precursor end and the start of the main event:
685: filled circles: $\Delta t>40$ s;
686: grey stars: $15<\Delta t<40$ s;
687: open squares: $\Delta t< 15$ s.
688: {\it Right panel}:
689: Precursor versus burst isotropic energy in the 15--150 keV rest frame band.
690: Same symbols as in the left panel.
691: The solid and dashed lines correspond to precursors having equal and
692: 1/10 the energetic of the main event, respectively.}
693: \label{alfa}
694: \end{figure}
695: %----------------------------------------------------
696:
697:
698:
699:
700:
701: \end{document}
702:
703: