1: \documentclass[12pt,preprint]{aastex}
2: %\documentclass[manuscript]{aastex}
3: %\documentclass[preprint2]{aastex}
4: \usepackage{emulateapj5}
5:
6: \newcommand{\pfrac}[2]{\left(\frac{#1}{#2}\right)}
7: \def\Mesz{M\'esz\'aros~}
8: \def\etal{{et al.~}}
9: \def\eps{\epsilon}
10:
11:
12: \slugcomment{Submitted to ApJL}
13:
14: \shorttitle{Radio re-brightening of GRB afterglows}
15: \shortauthors{Li \& Song}
16:
17: \begin{document}
18:
19: \title{Late-Time Radio Re-Brightening of Gamma-Ray Burst Afterglows:\\ Evidence for Double-Sided Jets}
20:
21: \author{Zhuo Li\altaffilmark{1} and L. M. Song\altaffilmark{1}}
22:
23: \altaffiltext{1}{Particle Astrophysics Center, Institute of High
24: Energy Physics, CAS, Beijing 100039, China}
25:
26: \begin{abstract}
27: The central engine of gamma-ray bursts (GRBs) is believed to eject
28: double-sided ultra-relativistic jets. For an observed GRB, one of
29: the twin jets should point toward us, and is responsible for the
30: prompt gamma-ray and subsequent afterglow emission. We consider in
31: this Letter the other receding jet, which will give rise to
32: late-time radio re-brightening (RRB) when it becomes
33: non-relativistic (NR) and radiative isotropic. The RRB peaks at a
34: time $5t_{NR}=2(E_{j,51}/n)^{1/3}$~yr after the GRB, where
35: $t_{NR}$ is the observed NR timescale for the preceding jet, $E_j$
36: is the jet energy and $n$ is the ambient medium density. The peak
37: flux is comparable to the preceding-jet emission at $t_{NR}$. We
38: expect the RRB of GRB 030329 1.7 yr after the burst with a flux
39: $\sim0.6$~mJy at 15~GHz. The cases of GRBs 970508 and 980703 have
40: also been discussed. The detection of RRB, which needs dense
41: monitoring campaign even a few years after a GRB, will be the
42: direct evidence for the existence of double-sided jets in GRBs,
43: and prove the black hole-disk system formation in the cores of
44: progenitors.
45: \end{abstract}
46:
47: \keywords{gamma-rays: bursts --- ISM: jets and outflows --- radio continuum:
48: general}
49:
50: \section{Introduction}
51: A standard shock model of gamma-ray burst (GRB) afterglows has now
52: been well established, in which the GRB outflow (jet) drives a
53: relativistic expanding blast-wave that sweeps up and heats the GRB
54: ambient medium to produce the long-term X-ray, optical and radio
55: afterglow by synchrotron/inverse-Compton (IC) emission (see Zhang
56: \& \Mesz 2004 or Piran 2004 for recent reviews).
57:
58: The remarkable supernova signature in GRB 030329 afterglow (Hjorth
59: \etal 2003) has confirmed that GRBs, at least the long-duration
60: ones, are originated from explosions of massive stars. Their
61: progenitors are believed to form black hole-disk systems in their
62: cores, and produce double-sided jets, which are responsible for
63: the GRB emission, along the spinning axis of the black hole (\Mesz
64: 2002). For the observed hard GRB photons to escape freely,
65: avoiding electron/positron pair formation, the outflows are
66: reguired to be ultra-relativistic with Lorentz factor
67: $\Gamma\ga100$. Because of the relativistic beaming effect, the
68: GRB emission is confined in a narrow cone, and can only be
69: observed when the light of sight is within the cone.
70: Interestingly, it was this beaming effect that made many authors
71: expect the present of ``orphan afterglows'' (Rhoads 1997; Dalal,
72: Griest, \& Pruet 2002; Granot \etal 2002; Granot \& Loeb 2003)
73: which correspond to off-axis GRBs without gamma-rays associated,
74: but with afterglows detected when the shocks decelerate.
75:
76: The previous works, except for Granot \& Loeb (2003), only
77: discussed the emission from the preceding jet (PJ), which is
78: pointing toward us, and neglected that from the other receding jet
79: (RJ), since it always points away from us. However, in this Letter
80: we discuss specially the emission from the RJ, which will emerge
81: when the RJ become non-relativistic (NR) and the radiation becomes
82: isotropic. Because of the light travel time delay, the RJ
83: contribution will overlay above the late-time, already decayed PJ
84: emission. In the NR phase the afterglow peak emission has moved to
85: radio frequencies, so the re-brightening from the RJ is only
86: expected in late-time radio afterglows. We predict the late-time
87: radio re-brightening (RRB) of GRB 030329 and discuss the long-term
88: radio observations for two GRBs 970508 and 980703. The observation
89: of RRBs will be particularly important due to its straightforward
90: implication of the existence of double-sided jets in GRBs.
91:
92: \section{Emission from the receding jet}
93: Consider two collimated jets, that are ejected in opposite
94: directions from the central engine of a GRB. For an observed GRB,
95: the observer should be almost on the jet axis. We assume the two
96: jets have the same characteristics, such as the half opening angle
97: $\theta_j$, the initial Lorentz factor $\Gamma_0$, and the kinetic
98: energy $E_j/2$ (or with equivalent isotropic energy
99: $E_{iso}=2E_j/\theta_j^2$). Since most afterglow fits seem to
100: favor a uniform medium (Panaitescu \& Kumar 2002), we assume a
101: constant particle density $n$ for both jets.
102:
103: In the above assumptions we have take the standard view of a
104: homogeneous jet with sharp edge (Rhoads 1999). Recently, a
105: structured jet model (\Mesz \etal 1998; Dai \& Gou 2001; Rossi
106: \etal 2002; Zhang \& \Mesz 2002) was proposed, but detailed fits
107: to afterglow data are still needed to determine whether it is
108: consistent with current observations (Granot \& Kumar 2003). Since
109: the standard jet model have been well successful in the data fits
110: (e.g., Panaitescu \& Kumar 2002), we only discuss homogeneous jets
111: here.
112:
113: \subsection{Jet dynamics}
114: We first discuss the PJ. The PJ evolution can be divided into four
115: phases. (1) Initially, $\Gamma\gg1/\theta_j$, the transverse size
116: of the jet is larger than that of the causally connected region,
117: therefore the jet evolves as if it was a conical section of a
118: spherical relativistic blast wave. The PJ undergoes first a
119: coasting phase with $\Gamma=\Gamma_0$. (2) After the jet-induced
120: shock sweeps up enough medium, the jet kinetic energy is mostly
121: transferred into the shocked medium, and the jet begins to
122: decelerate significantly. This occurs at the deceleration radius
123: $r_d=3E_{iso}/(4\pi n\Gamma_0^2m_pc^2)$ at an observer time
124: $t_d=r_d/(2\Gamma_0^2c)$, where a relation $dr\approx2\Gamma^2cdt$
125: has been taken for surperluminal motion. At the deceleration
126: phase, the dynamics could be well described by the Blandford \&
127: McKee (1976) self-similar solution, where $\Gamma\propto
128: r^{-3/2}\propto t^{-3/8}$. (3) As the jet continues to
129: decelerate, the Lorentz factor drops to $1/\theta_j$ at a time
130: $t_j=t_d(\Gamma_0\theta_j)^{8/3}$, corresponding to a radius
131: $r_j=r_d(\Gamma_0\theta_j)^{2/3}$. From the time on the transverse
132: size of causally connected regions exceeds that of the jet,
133: therefore the sphere approximation breaks down, and the jet starts
134: expanding sideways. The dynamical evolution in this stage depends
135: on the degree of sideways expansion. If the lateral velocity in
136: the comoving frame equals to the local sound speed, the jet will
137: spread quickly with the opening angle increasing and $\Gamma$
138: dropping exponentially (Rhoads 1999). At this stage, the radius
139: hardly increases and may be regarded as a constant $r\approx r_j$,
140: and $\Gamma\propto t^{-1/2}$. (4) Finally the jet goes into NR
141: phase with $\Gamma\sim1$ at $t_{NR}=t_j/\theta_j^2$, where the
142: evolution can be well described by Sedov-Taylor solution
143: $\beta\propto r^{-3/2}\propto t^{-3/5}$, with
144: $\beta=(1-1/\Gamma^2)^{1/2}$. In summary, the evolution is:
145: \begin{equation}
146: \begin{array}{lll}
147: \mbox{coasting:} & t<t_d & \Gamma=\Gamma_0,~ r\propto t \\
148: \mbox{sphere:} & t_d<t<t_j & \Gamma\propto t^{-3/8},~ r\propto t^{1/4} \\
149: \mbox{spreading:} & t_j<t<t_{NR} & \Gamma\propto t^{-1/2},~ r\approx r_j \\
150: \mbox{NR:} & t>t_{NR} & \beta\propto t^{-3/5}~(\Gamma\sim1),~
151: r\propto t^{2/5}
152: \end{array}
153: \end{equation}
154: Using the above simplified dynamical relation, the NR time and
155: radius, defined as where $\Gamma=1$, are able to be calculated as
156: \begin{equation}\label{eq:nr}
157: t_{NR}=\frac1{2c}\pfrac{3E_j}{2\pi m_pc^2n}^{1/3},~ r_{NR}\approx
158: r_j=\pfrac{3E_j}{2\pi m_pc^2n}^{1/3}.
159: \end{equation}
160: They are independent of $\Gamma_0$ and $\theta_j$. With typical
161: values $E_j=10^{51}E_{j,51}$~ergs and $n=1$~cm$^{-3}$ (e.g., Frail
162: \etal 2001; Panaitescu \& Kumar 2001), we have
163: \begin{equation}\label{eq:tnr}
164: t_{NR}=130\pfrac{E_{j,51}}{n}^{1/3}{\rm d}.
165: \end{equation}
166: Note that this $t_{NR}$ refers to the PJ.
167:
168: As for the RJ, it also transits into NR phase at a radius $r_{NR}$.
169: However, due to the light travel time the observer time epoch is delayed
170: by a time $2r_{NR}/c$, so the observed NR time for the RJ is
171: \begin{equation}\label{eq:rjtnr}
172: t_{NR}^{RJ}=t_{NR}+\frac {2r_{NR}} c=5t_{NR},
173: \end{equation}
174: 5 times that of the PJ\footnote{It can be shown that the factor 5
175: is also valid for the wind case where the external density
176: decreases as the square of the distance from the source (Dai \& Lu
177: 1998; Chevalier \& Li 2000).}.
178:
179: There are uncertainties about the lateral spreading of jets, and
180: it might be that little spreading occurs, as shown in some
181: numerical simulations (e.g., Cannizzo \etal 2004). If we assume
182: the extreme case that the lateral velocity is zero, then the jet
183: remains conical geometry and continues to evolve as if it was a
184: conical section of a spherical relativistic blast wave after
185: $\Gamma<1/\theta_j$. In this case, the NR time for the PJ is
186: $t_{NR}=(1/2)(3E_j/2\pi
187: m_pc^5n\theta_j^2)^{1/3}=610(E_{j,51}/n\theta_{j,-1}^2)^{1/3}$~d,
188: with $\theta_{j,-1}=\theta_j/10^{-1}$, and the corresponding
189: radius is $r_{NR}=2ct_{NR}$, therefore for the RJ, the NR time is
190: $t_{NR}^{RJ}=t_{NR}+2r_{NR}/c=5t_{NR}=8.4(E_{j,51}/n\theta_{j,-1}^2)^{1/3}$~yr.
191: These may be regarded as upper limits, while the values in eqs.
192: (\ref{eq:nr}-\ref{eq:rjtnr}) for the fastest spreading cases as
193: lower limits.
194:
195: \subsection{Non-relativistic phase emission}
196: For the detailed discussion of afterglow emission in NR phase, the readers
197: might refer to Frail, Waxman \& Kulkarni (2000), Livio \& Waxman
198: (2000) and Waxman (2004), we here only give brief introduction.
199:
200: Consider also the PJ emission first. After $t_{NR}$, the evolution
201: is described by the spherical NR self-similar evolution, with the
202: PJ radius given by $r=r_{NR}(t/t_{NR})^{2/5}$, and the swept-up
203: particle number is $N_e\simeq(4/3)\pi nr^3$. Simply assume that
204: the shocked material forms a thin and uniform shell, with the
205: width $\Delta=r/\eta\ll r$ and the post-shock thermal energy
206: density $U=E_j/(4\pi r^2\Delta)$. The electron and magnetic-field
207: energy densities are assumed to be constant fraction $\eps_e$ and
208: $\eps_B$, respectively, of the post-shock thermal energy density
209: $U$. The electrons are believed to be accelerated to form a power
210: law distribution, $dn_e/d\gamma_e\propto\gamma_e^{-p}$ for
211: $\gamma_e\ge\gamma_m$. Based on the above assumptions, it can be
212: found that the magnetic field strength and the minimum electron
213: Lorentz factor evolve as $B\propto t^{-3/5}$ and $\gamma_m\propto
214: t^{-6/5}$, respectively.
215:
216: The accelerated electrons will give rise to synchrotron
217: emission. The characteristic synchrotron frequency of electrons
218: with $\gamma_m$ is $\nu_m=eB\gamma_e^2/(2\pi m_ec)$, i.e.,
219: \begin{equation}
220: \nu_m=0.3\eps_{e,-1}^2\eps_{B,-1}^{1/2}n^{1/2} \pfrac
221: t{t_{NR}}^{-3}\mbox{GHz},
222: \end{equation}
223: where $\eps_{e,-1}=\eps_e/10^{-1}$, and
224: $\eps_{B,-1}=\eps_B/10^{-1}$. The peak specific luminosity is
225: $L_{\nu_m}=N_eP/\nu_m$ where
226: $P=(4/3)\sigma_Tc\gamma_m^2(B^2/8\pi)$ is the synchrotron power of
227: an electron with $\gamma_m$. The observed radio frequency is
228: usually beyond $\nu_m$ at NR stage, the specific luminosity is
229: therefore $L_\nu=L_{\nu_m}(\nu/\nu_m)^{(1-p)/2}\propto
230: t^{21/10-3p/2}$, and given by
231: \begin{eqnarray}
232: L_\nu=4\times10^{30}\eps_{e,-1}\eps_{B,-1}^{3/4}n^{3/4}E_{j,51}\nonumber\\
233: \pfrac\nu{10~ {\rm GHz}}^{-1/2}\pfrac t{t_{NR}}^{-9/10}\mbox{ergs
234: s$^{-1}$Hz$^{-1}$},
235: \end{eqnarray}
236: where $p=2$ has been taken. The NR light curve flattens, compared
237: with the decay, $\propto t^{-p}$, in the spreading phase (Rhoads
238: 1999; Sari, Piran \& Halpern 1999).
239:
240: As for the RJ, when it is still relativistic the emission is
241: always beamed away from us and hence invisible. At the time
242: $t_{NR}^{RJ}$ after the GRB, the RJ becomes non-relativistic and
243: the emission becomes isotropic. The observed RJ luminosity at this
244: point should be comparable to that of the PJ at $t_{NR}$, because
245: of the same characteristics of the PJ and RJ. If neglect the NR
246: jet velocity with respect to the light speed, the time delay
247: between the emission from the PJ and RJ could be roughly fixed as
248: $2r/c\approx2r_{NR}/c=4t_{NR}$ (for $r$ is not $\gg r_{NR}$). The
249: RJ specific luminosity, then, could be written as, for $t\geq
250: t_{NR}^{RJ}$,
251: \begin{equation}\label{eq:LnuRJ}
252: L_\nu^{RJ}(t)\simeq L_\nu(t-4t_{NR})\propto \pfrac
253: {t-4t_{NR}}{t_{NR}}^{-9/10}.
254: \end{equation}
255: A schematic plot is given in figure 1 for both PJ and RJ radio
256: light curves. After rapid increase in flux, the RJ emission
257: reaches the peak at around $t_{NR}^{RJ}$, and then decreases as a
258: delayed light curve of the NR-phase PJ, but showing a steeper
259: decline in the logarithmic plot. Note that the transition to NR
260: phase might be gradual, and the jet edge might be not sharp,
261: therefore both these can smooth the RRB feature on the plot.
262:
263:
264:
265: \section{Observation}
266:
267: \subsection{GRB 970508 and GRB 980703}
268: As discussed above, if a GRB can be monitored for a long time in radio bands,
269: e.g., a few years, the RRB may be able to be
270: detected. So far, there are two GRBs that have been reported
271: in the literatures with radio monitoring longer than one year:
272: GRB 970508 and GRB 980703 (Frail, Waxman \& Kulkarni 2000;
273: Berger, Kulkarni \& Frail 2001; Frail \etal 2003).
274:
275: Frail, Waxman \& Kulkarni (2000) report an extensive monitoring
276: of the radio afterglow of GRB 970508, lasting 450 days after the
277: burst. In the data analysis they found that the spectral and
278: temporal radio behavior indicate a transition to NR expansion at
279: $t_{NR}\approx100$~d, therefore, we should expect the RRB at
280: around day 500. However, this is unfortunately behind the last
281: observation of this burst, so we might have missed that apparent
282: phenomena.
283:
284: GRB 980703 was monitored even longer, up to $\sim10^3$~d, as shown
285: in figure 1 of Frail \etal (2003). Two obvious flattening appear
286: in the radio light curves: while the late-time transition to a
287: constant flux is thought to be the host galaxy contribution, the
288: earlier flattening at $\approx40$~d is attributed to the
289: transition into NR expansion (Berger, Kulkarni \& Frail 2001).
290: Modelling the $\Gamma$ evolution has also inferred a similar value
291: of $t_{NR}=30-50$~d (Frail \etal 2003). Therefore, the expected
292: RRB should arise at $\sim200$~d, and then after another time
293: interval of 200~d, around 400~d from the burst, the radio flux
294: should decline to a value similar to just before the RRB. At first
295: glance of figure 1 of Frail \etal (2003), the radio light curves,
296: somewhat more extensive at 8.5 and 4.8~GHz, seem to show only
297: decays without any re-brightening. However, we notice that there
298: is no observation data collected between 210~d and $\sim400$~d,
299: therefore the RRB may be missed once again due to the lack of
300: observation.
301:
302: \subsection{Prediction for GRB 030329}
303: GRB 030329 is the nearest cosmological burst ($z=0.1685$), and has
304: a bright afterglow at all wavelengths. Berger \etal (2003) have
305: reported its bright radio afterglow up to $\sim70$~d after the
306: burst. Since no radio flattening (due to transition to NR
307: evolution) appears yet\footnote{Although the X-ray afterglow shows
308: a flattening around day 37 (Tiengo \etal 2004), we believe this is
309: not relevant to the relativistic to NR transition of the radio
310: jet, because of no simultaneous radio flattening (Berger \etal
311: 2003), and also conflicted with the still superluminal expansion
312: in the radio angular size measurement (Taylor \etal 2004). The
313: X-ray flattening may need other explanations, e.g. an IC
314: component; NR transition for the narrow jet in two-component model
315: (Berger \etal 2003).}, we should calculate $t_{NR}$ using $E_j/n$
316: derived from available data.
317:
318: The early breaks ($\sim0.5$~d) in the R band (Price \etal 2003)
319: and X-ray (Tiengo \etal 2004) light curves infer a small jet
320: opening angle, leading to a small jet-corrected energy release in
321: GRB 030329, which is more than one order of magnitude below the
322: average value around which most GRBs are narrowly clustered (Frail
323: \etal 2001; Panaitescu \& Kumar 2001). Two models have been
324: proposed to solve this problem: a refreshed-shock model with 10
325: times more energy injected later (Granot, Nakar \& Piran 2003);
326: and a two-component model with most energy in a wider jet
327: component (Berger \etal 2003). Both models have the same result:
328: the kinetic energy corresponding to the late-time radio afterglow
329: is typical for all GRBs. Based mainly on the early multi-frequency
330: data, a typical value for medium density, $n=1$~cm$^{-3}$, was
331: also derived by Willingale \etal (2004). Therefore, the estimated
332: value of $t_{NR}$ should be typical as in eq. (\ref{eq:tnr}),
333: hence $t_{NR}\sim130$~d. In fact, the radio angular size
334: measurement of this burst shows relativistic expanding 83~d after
335: the burst (Taylor \etal 2004), consistent with the estimation
336: here.
337:
338: In addition, Berger \etal (2003) have inferred
339: $E_{j,51}/n=0.14\nu_{c,13}^{-1/2}$ from a snapshot spectrum at
340: $t_j\approx10$~d, where $\nu_c=\nu_c/10^{13}$~Hz is the cooling
341: frequency. This parameter value leads to a too early NR phase,
342: $t_{NR}\approx68$~d (from eq. \ref{eq:tnr}), conflicted with the
343: observed superluminal expansion (Taylor \etal 2004). However, in
344: the parameter derivation, the assumption of jet break time at
345: $t_j\approx10$~d might be incorrect, because the millimeter
346: observation shows that the light curves have already steepened at
347: $\approx5$~d at both 100 and 250~GHz (Sheth \etal 2003), implying
348: that $t_j\la5$~d. Moreover, the angular size evolution can also
349: give constraints to the model parameters. For GRB 030329, Granot,
350: Ramirez-Ruiz \& Loeb (2004) derive $E_{j,51}/n\sim0.8$ (see also
351: Oren, Nakar \& Piran 2004) from the observation by Taylar \etal
352: (2004). This yields a later transition at $t_{NR}\sim120$~d.
353:
354: We will take the more reasonable value $t_{NR}\approx120$~d (It
355: should be noticed that this might still be the lower limit since
356: the most rapid sideways spreading has been assumed), and hence the
357: expected RRB should arise at $\sim1.7$~yr after GRB 030329, with
358: the fluxes comparable to that at the time $t_{NR}$. Assuming the
359: same temporal slope for late time, an extrapolation of the
360: available radio light curves (Berger \etal 2003) to $\sim120$~d
361: gives a flux $\sim0.6$~mJy at 15~GHz, or $\sim0.3$~mJy at 44~GHz.
362: Dense monitoring campaign around $\sim1.7$~yr after GRB 030329 is
363: required to obtain a well-observed RRB profile.
364:
365: \section{Summary and discussion}
366: In this work we suggest that RRBs are common for GRBs with
367: double-sided jets, which come from the RJ transition to NR
368: evolution. If we assume the same properties for both PJ and RJ and
369: also for the ambient medium on both sides, the RRB would arise at
370: a time $t_{NR}^{RJ}=5t_{NR}=1.8(E_{j,51}/n)^{1/3}$~yr after the
371: burst, with a flux comparable to that at the time $t_{NR}$. After
372: $t_{NR}^{RJ}$, the afterglow behaves as a delayed emission of that
373: behind $t_{NR}$, with the time lag of $4t_{NR}$ (eq.
374: \ref{eq:LnuRJ}). Unfortunately, no radio data could be collected
375: when RRBs occur for GRB 970508 and GRB 980703, two longest
376: observed GRBs so far. We suppose the RRB of GRB 030329 around
377: 1.7~yr after the burst with $\sim0.6$~(0.3)~mJy at 15~(44)~GHz,
378: and urge dense monitoring campaign during that time. It should be
379: noticed that for weak jet-spreading, the RRB might be more delayed
380: since $t_{NR}^{RJ}$ is much larger (see in the end of \S 2.1).
381:
382: There has been growing evidence for collimated jets in GRBs over
383: the past several years, which is coming mainly from observations
384: of achromatic breaks in the afterglow light curves (e.g., Kulkarni
385: et al. 1999; Stanek et al. 1999). However, there are still other
386: explanations for the light curve breaks, for example, the
387: transition from the relativistic to NR phase of the blast wave
388: at a few days due to highly dense medium (Dai \& Lu 1999; Wang,
389: Dai \& Lu 2000); the effects of IC scattering flattening or
390: steepening the light curves (Wei \& Lu 2000); a sudden drop in the
391: external density (Kumar \& Panaitescu 2000); a break in the energy
392: spectrum of radiating electrons (Li \& Chevalier 2001). On the
393: other hand, though black hole-disk system with twin jets is
394: generally assumed in GRBs, magnetized, rapidly rotating neutron
395: stars remain contenders, and an off-center dipole could lead to a
396: one-sided jet. Since RRBs are only associated with collimated,
397: double-sided, relativistic outflows, the detection of RRBs would
398: provide straightforward evidence of double-sided jets in GRBs, and
399: prove the black hole-disk system formation in the cores of
400: progenitors. This make RRB observations significantly important.
401:
402: \acknowledgments
403:
404: Z. Li thanks R. F. Shen for discussions, and L. J. Gou for
405: comments. This work was supported by the National 973 Project and
406: the Special Funds for Major State Basic Research Projects.
407:
408:
409: \begin{thebibliography}{}
410:
411: \bibitem[Berger et al.(2003)]{2003Natur.426..154B} Berger, E., et al.\
412: 2003, \nat, 426, 154
413:
414: \bibitem[Berger, Kulkarni, \& Frail(2001)]{2001ApJ...560..652B} Berger, E.,
415: Kulkarni, S.~R., \& Frail, D.~A.\ 2001, \apj, 560, 652
416:
417: \bibitem[Cannizzo, Gehrels, \& Vishniac(2004)]{2004ApJ...601..380C}
418: Cannizzo, J.~K., Gehrels, N., \& Vishniac, E.~T.\ 2004, \apj, 601, 380
419:
420: \bibitem[]{422} Chevalier, R. A., \& Li, Z.-Y.\ 2000, ApJ, 536, 195
421:
422: \bibitem[Dai \& Gou(2001)]{2001ApJ...552...72D} Dai, Z.~G.~\& Gou, L.~J.\
423: 2001, \apj, 552, 72
424:
425: \bibitem[]{427} Dai, Z.~G.~\& Lu, T.\ 1998, \mnras, 298, 87
426:
427: \bibitem[Dai \& Lu(1999)]{1999ApJ...519L.155D} Dai, Z.~G.~\& Lu, T.\ 1999,
428: \apjl, 519, L155
429:
430: \bibitem[Dalal, Griest, \& Pruet(2002)]{2002ApJ...564..209D} Dalal, N.,
431: Griest, K., \& Pruet, J.\ 2002, \apj, 564, 209
432:
433: \bibitem[Frail et al.(2001)]{2001ApJ...562L..55F} Frail, D.~A., et al.\
434: 2001, \apjl, 562, L55
435:
436: \bibitem[Frail et al.(2003)]{2003ApJ...590..992F} Frail, D.~A., et al.\
437: 2003, \apj, 590, 992
438:
439: \bibitem[Frail, Waxman, \& Kulkarni(2000)]{2000ApJ...537..191F} Frail,
440: D.~A., Waxman, E., \& Kulkarni, S.~R.\ 2000, \apj, 537, 191
441:
442: \bibitem[]{444} Granot, J.~\& Kumar, P.\ 2003, \apj, 591, 1086
443:
444: \bibitem[Granot \& Loeb(2003)]{2003ApJ...593L..81G} Granot, J.~\& Loeb, A.\
445: 2003, \apjl, 593, L81
446:
447: \bibitem[Granot, Nakar, \& Piran(2003)]{2003Natur.426..138G} Granot, J.,
448: Nakar, E., \& Piran, T.\ 2003, \nat, 426, 138
449:
450: \bibitem[Granot, Panaitescu, Kumar, \& Woosley(2002)]{2002ApJ...570L..61G}
451: Granot, J., Panaitescu, A., Kumar, P., \& Woosley, S.~E.\ 2002,
452: \apjl, 570, L61
453:
454: \bibitem[]{453} Granot, J., Ramirez-Ruiz, E., \& Loeb, A.\ 2004, (astro-ph/0407182)
455:
456: \bibitem[]{456} Hjorth, J., \etal \ 2003, \nat, 423, 847
457:
458: \bibitem[Kulkarni et al.(1999)]{1999Natur.398..389K} Kulkarni, S.~R., et
459: al.\ 1999, \nat, 398, 389
460:
461: \bibitem[Kumar \& Panaitescu(2000)]{2000ApJ...541L..51K} Kumar, P.~\&
462: Panaitescu, A.\ 2000, \apjl, 541, L51
463:
464: \bibitem[Li \& Chevalier(2001)]{2001ApJ...551..940L} Li, Z.-Y.~\& Chevalier,
465: R.~A.\ 2001, \apj, 551, 940
466:
467: \bibitem[Livio \& Waxman(2000)]{2000ApJ...538..187L} Livio, M.~\& Waxman,
468: E.\ 2000, \apj, 538, 187
469:
470: \bibitem[]{470} M{\' e}sz{\' a}ros, P. 2002, \araa, 40, 137
471:
472: \bibitem[Meszaros, Rees, \& Wijers(1998)]{1998ApJ...499..301M} M{\' e}sz{\' a}ros,
473: P., Rees, M.~J., \& Wijers, R.~A.~M.~J.\ 1998, \apj, 499, 301
474:
475: \bibitem[]{475} Oren, Y., Nakar, E., \& Piran, T.\ 2004, preprint (astro-ph/0406277)
476:
477: \bibitem[Panaitescu \& Kumar(2001)]{2001ApJ...560L..49P} Panaitescu, A.~\&
478: Kumar, P.\ 2001, \apjl, 560, L49
479:
480: \bibitem[Panaitescu \& Kumar(2002)]{2002ApJ...571..779P} Panaitescu, A.~\&
481: Kumar, P.\ 2002, \apj, 571, 779
482:
483: \bibitem[]{483} Piran, T. 2004, Rev. Mod. Phys., accepted (astro-ph/0405503)
484:
485: \bibitem[Price et al.(2003)]{2003Natur.423..844P} Price, P.~A., et al.\
486: 2003, \nat, 423, 844
487:
488: \bibitem[]{488}{{Rhoads}, J.~E.}\ 1997, \apj, 487, L1
489:
490: \bibitem[Rhoads(1999)]{1999ApJ...525..737R} Rhoads, J.~E.\ 1999, \apj, 525,
491: 737
492:
493: \bibitem[Rossi, Lazzati, \& Rees(2002)]{2002MNRAS.332..945R} Rossi, E.,
494: Lazzati, D., \& Rees, M.~J.\ 2002, \mnras, 332, 945
495:
496: \bibitem[Sari, Piran, \& Halpern(1999)]{1999ApJ...519L..17S} Sari, R.,
497: Piran, T., \& Halpern, J.~P.\ 1999, \apjl, 519, L17
498:
499: \bibitem[Sheth et al.(2003)]{2003ApJ...595L..33S} Sheth, K., Frail, D.~A.,
500: White, S., Das, M., Bertoldi, F., Walter, F., Kulkarni, S.~R., \&
501: Berger, E.\ 2003, \apjl, 595, L33
502:
503: \bibitem[Stanek et al.(1999)]{1999ApJ...522L..39S} Stanek, K.~Z.,
504: Garnavich, P.~M., Kaluzny, J., Pych, W., \& Thompson, I.\ 1999,
505: \apjl, 522, L39
506:
507: \bibitem[Taylor, Frail, Berger, \& Kulkarni(2004)]{2004ApJ...609L...1T}
508: Taylor, G.~B., Frail, D.~A., Berger, E., \& Kulkarni, S.~R.\ 2004,
509: \apjl, 609, L1
510:
511: \bibitem[Tiengo et al.(2004)]{2004A&A...423..861T} Tiengo, A., Mereghetti,
512: S., Ghisellini, G., Tavecchio, F., \& Ghirlanda, G.\ 2004, \aap,
513: 423, 861
514:
515: \bibitem[Wang, Dai, \& Lu(2000)]{2000MNRAS.317..170W} Wang, X.~Y., Dai,
516: Z.~G., \& Lu, T.\ 2000, \mnras, 317, 170
517:
518: \bibitem[Waxman(2004)]{2004ApJ...602..886W} Waxman, E.\ 2004, \apj, 602,
519: 886
520:
521: \bibitem[Wei \& Lu(2000)]{2000A&A...360L..13W} Wei, D.~M.~\& Lu, T.\ 2000,
522: \aap, 360, L13
523:
524: \bibitem[Willingale et al.(2004)]{2004MNRAS.349...31W} Willingale, R.,
525: Osborne, J.~P., O'Brien, P.~T., Ward, M.~J., Levan, A., \& Page,
526: K.~L.\ 2004, \mnras, 349, 31
527:
528: \bibitem[Zhang \& M{\' e}sz{\' a}ros(2002)]{2002ApJ...571..876Z} Zhang,
529: B.~\& M{\' e}sz{\' a}ros, P.\ 2002, \apj, 571, 876
530:
531: \bibitem[]{541} Zhang, B.~\& M{\' e}sz{\' a}ros, P.\ 2004, Int. J. Mod. Phys. A, 19, 2385
532:
533: \end{thebibliography}
534:
535: \begin{figure}[ht]
536: \epsscale{.7} \plotone{f1.eps} \caption{The schematic log-log plot
537: for the radio light curve at $\nu>\nu_m$. The solid and dashed
538: lines correspond to the contribution from the PJ and RJ,
539: respectively. The time scalings are marked, assuming $p=2$. After
540: a steep decline in the spreading phase, the PJ light curve flatten
541: at $t_{NR}$. The RJ emission rapidly increases to a peak around
542: $t_{NR}^{RJ}=5t_{NR}$, and then declines following the way of the
543: PJ unless a time delay of $4t_{NR}$, but exhibiting steep light
544: curve in the log-log plot. The dashed-dot line indicates
545: $L_{\nu}(t_{NR})=L_\nu^{RJ}(t_{NR}^{RJ})$.}
546: \end{figure}
547:
548:
549: \end{document}
550: