1: 

2: 

3: 

4: 

5: 

6: %        aa.dem

7: % AA vers. 4.01, LaTeX class for Astronomy & Astrophysics 

8: % demonstration file 

9: %                                                (c) Springer-Verlag HD   

10: %------------------------------------------------------ 

11: % 

12: %\documentclass[referee]{aa} 

13: % for a referee version 

14: % 

15: \documentclass[article]{aa} 

16: \usepackage{epsfig,deluxe} 

17: % 

18: \begin{document} 

19:  

20:  

21: \newcommand{\gsim}{\hbox{\rlap{$^>$}$_\sim$}} 

22: % A&A Section 6: Form. struct. and  evolut. of stars} 

23: %  \thesaurus{06     % A&A Section 6: Form. struct. and evolut. of stars   

24: \authorrunning{S. Dado, A. Dar \& A. De R\'ujula} 

25: \titlerunning{Radio Afterglows of GRBs} 

26: \title{On the Radio Afterglow of Gamma Ray Bursts} 

27: \author{Shlomo Dado$^{^1}$, Arnon Dar$^{^1}$ and 

28: A. De R\'ujula$^{^2}$} 

29: \institute{1. Physics Department and Space Research Institute, Technion\\   

30:                Haifa 32000, Israel\\ 

31:            2. Theory Division, CERN, CH-1211 Geneva 23, Switzerland} 

32:  

33: \maketitle 

34:  

35: \begin{abstract} 

36:  

37: We use the cannonball (CB) model of gamma ray bursts (GRBs) to predict the

38: spectral and temporal behaviour of their radio afterglows (AGs). A single

39: simple expression describes the AGs at all times and frequencies; its

40: high-frequency limit reproduces the successful CB model predictions for

41: optical and X-ray AGs. We analyze all of the observed radio AGs of GRBs

42: with known redshifts, including those of the exceptionally close-by GRB

43: 980425.  We also study in detail the time-evolution of the AGs' spectral

44: index. The agreement between theory and observations is excellent, even

45: though the CB model is extremely frugal in the number of parameters

46: required to explain the radio observations. We propose to use the

47: scintillations in the radio AGs of GRBs to verify and measure the

48: hyperluminal speed of their jetted CBs, whose apparent angular velocity is

49: of the same order of magnitude as that of galactic pulsars, consistently

50: measured directly, or via scintillations. 

51:  

52: \end{abstract} 

53: 

54: 

55: \section{Introduction} 

56:  

57: The Cannonball Model is based on the hypothesis that GRBs and their 

58: afterglows are made in supernova explosions by the jetted ejection of 

59: relativistic plasmoids: ``cannonballs'' made of ordinary baryonic matter   

60: (Dar and De R\'ujula 2000a), similar to the ones observed in quasars and   

61: microquasars (e.g., Mirabel and Rodriguez 1994; 1999 and references 

62: therein). The name cannonball (CB) originates in the contention that 

63: ---due to a mechanism that we have explicitly discussed in Dado et al.   

64: 2001--- the ejected plasmoids stop expanding very early in the afterglow   

65: (AG) phase. 

66:  

67: The CB paradigm gives a good description of the properties of 

68: the $\gamma$-rays 

69: in a GRB, that we modelled in simple approximations in 

70: Dar and De R\'ujula 2000b. 

71: It suggests an alternative (Dar and De R\'ujula 2001a),

72: which is rather promising (Dado et al. 2002), to 

73: the ``Fe-line'' interpretation of the spectral lines observed 

74: in some X-ray afterglows (GRB 970508: Piro et al. 1998; 

75: GRB 970828: Yoshida et al. 1999, 2001; GRB 991216: Piro et al. 2000; 

76: GRB 000214: Antonelli et al. 2000). The model also provides 

77: an extremely simple and successful description of the spectrum, 

78: and of the shape and absolute magnitude of the light curves of the 

79: optical and X-ray afterglows of {\it all} GRBs of known redshift, 

80: at {\it all} observed times (Dado et al. 2001, hereafter called DDD 2001). 

81: This description is universal, it encompasses the early optical 

82: flash of GRB 990123, the very peculiar optical and X-ray AG of GRB 970508,   

83: and all of the properties of GRB 980425, associated with SN1998bw. 

84:  

85: In this paper we derive the CB model's predictions for radio afterglows,   

86: and compare them to {\it all} radio observations in GRBs of known   

87: redshift. We also study the evolution of the spectral index of AGs

88: as a function of time. The CB model ---in parameter-thrifty

89: and very simple terms--- passes these tests with 

90: flying colours. 

91:  

92: \section{Summary} 

93:  

94: In the CB model a GRB jet consists of $\rm n_{_{CB}}$ cannonballs, 

95: typically a few, 

96: each of them generating a prominent 

97: pulse in the $\gamma$-ray signal, as they 

98: reach the transparent outskirts of the shell of their associated supernova   

99: (SN).  We assume CBs to be made of ordinary matter, mainly hydrogen,

100: and to enclose a  magnetic field maze, as is the case for the 

101: observed ejections from quasars and microquasars.

102: The interstellar medium (ISM) the CBs traverse in the AG phase 

103: has been previously partially 

104: ionized by the GRB radiation and is fully 

105: ionized by Coulomb interactions as it enters the CB. 

106: In analogy to processes occurring in quasar and microquasar 

107: ejections, the ionized ISM particles are multiple 

108: scattered, in a ``collisionless'' way, by the CBs' turbulent magnetic   

109: fields. 

110: In the rest system of the CB the ISM swept-up nuclei are isotropically   

111: re-emitted, exerting 

112: an inwards force on the CB's surface. This allows one to compute {\it   

113: explicitly} 

114: the CB's radius as a function of time (DDD 2001). 

115: The radius, for typical parameters, and in minutes of observer's

116: time, reaches a constant  $\rm R_{max}$ of a few times $10^{14}$ cm.

117: 

118:  

119: The ISM nuclei (mainly protons) that a CB scatters also decelerate its

120: flight: its Lorentz factor, $\rm\gamma(t)$, is calculable.  Travelling at

121: a large $\gamma$ and viewed at a small angle $\theta$, the CB's emissions

122: are strongly relativistically aberrant:  in minutes of observer's time,

123: the CBs are parsecs away from their source.  For a constant CB radius and

124: an approximately constant ISM density, $\rm\gamma(t)$ has an explicit

125: analytical expression, as discussed in Appendix I.  Typically

126: $\rm\gamma=\gamma(0)/2$ at a distance of order 1 kpc from the source, and

127: $\gamma(0)\sim 10^3$.  Due to a limited observational sensitivity, GRBs

128: have been detected only up to angles $\theta$ of a few times

129: $1/\gamma(0)$. 

130: 

131: 

132: The ISM electrons entering a CB are caught up and bounce off 

133: its enclosed magnetic domains acquiring a predictable power-law energy   

134: spectrum, as we argue in Section 3. In the CB's rest system

135: $\rm dn_e/dE\propto E^{-2}$ below an energy 

136: $\rm E_b(t)\simeq \gamma(t)\, m_e\, c^2$, steepening to 

137: $\rm dn_e/dE\propto E^{-(p+1)}$, with $\rm p\simeq 2.2$, 

138: above this energy\footnote{In our previous work   

139: $\rm p$ was called $\rm\beta_p$, referring to the proton spectrum, 

140: for which radiative losses are negligible.}.

141: The energy $\rm E_b$ does not correspond to the conventional

142: synchrotron ``cooling break'' but to the {\it injection bend} at the

143: energy at which electrons enter the CB with a Lorentz factor $\rm\gamma(t)$.

144: In Section 4 we discuss the observational 

145: support of the existence of the injection bend, which is strong.

146: Given the very large magnetic and radiation energy densities  

147: in the CB, the usual cooling break (at the energy at which 

148: the energy-loss rate due to synchrotron emission and inverse Compton 

149: scattering equals that due to bremsstrahlung, 

150: adiabatic losses and escape) happens only at subrelativistic energies, 

151: as discussed in Appendix II. 

152:  

153: The magnetic energy-density in a CB (DDD 2001) is: 

154: \begin{equation}\rm 

155: U_B= {B^2\over8\, \pi}\sim {1\over 4}\,\gamma^2\,n_p\, m_p\, c^2,

156: \label{emag}

157: \end{equation}

158: with $\rm n_p$ the 

159: ISM baryon density (seen as $\rm \gamma\, n_p$ by the CB in its rest 

160: system). Thus, the magnetic field is:

161: \begin{equation} 

162: \rm B(t)\sim 3 \;\left[{n_p\over 10^{-3}\,cm^3}\right]^{1/2}

163: \left[{\gamma(t)\over 10^3}\right]\; Gauss.

164: \label{mag}

165: \end{equation}

166: The pitch-angle averaged characteristic synchrotron-radiation 

167: frequency of electrons of energy $\rm E=E_b$ is (Rybicki and Lightman 1979):

168: \begin{equation}

169: \rm \nu_b(t)\sim 0.29\,{3\over 4}\; \gamma(t)^2\,\nu_L

170: \label{nub}

171: \end{equation} 

172: where $\rm\nu_L=e\,B/(2\,\pi\,m_e\,c)$ is the Larmor frequency in the  

173: CB enclosed magnetic field $\rm B$. 

174: To a good approximation, in the CB rest system and prior to 

175: cumulation, absorption

176: and limb-darkening corrections, the synchrotron radiation has a spectral 

177: shape: 

178: \begin{eqnarray} 

179: \rm \nu\,{dn_\gamma\over d\,\nu} &\propto& \rm 

180: f_{sync}(\nu,t) \equiv  

181: {K(p)\over \nu_b(t)}{[\nu/\nu_b(t)]^{-1/2}\over

182: \sqrt{1+[\nu/\nu_b(t)]^{(p-1)}}}

183: \nonumber \\ \rm 

184:  \rm K(p)&\equiv&\rm

185: {\sqrt{\pi}\over   \Gamma\left[{2\, p-1\over 2(p-1)}\right]

186:                                    \, \Gamma\left[{2\, p-3\over 2(p-1)}\right]}

187: \simeq{p-2\over 2\,(p-1)}\; , 

188: \label{sync} 

189: \end{eqnarray}

190: where we have normalized $\rm f_{sync}(\nu)$ to a unit integral 

191: over all frequencies and the approximation

192: is good to better than 8\% precision

193: in the range $\rm 2<p\leq 2.6$. Note that, for $\rm \nu\gg\nu_b$,

194: $\rm f_{syn}\propto \nu_b^{(p-2)/2}$  

195: i.e., it is independent of $\rm \nu_b$ for $\rm p=2$. For $\rm p\sim 2.2\, ,$  

196: the extremely weak dependence of $\rm f_{syn}$  

197: on $\rm \nu_b$ in the optical and X-ray bands was

198: neglected in DDD 2001. 

199: 

200: For the first $\sim 10^3$ seconds of observer's time,

201: a CB is still cooling fast and emitting

202: via thermal bremsstrahlung (DDD 2001), but after that

203: the CB emissivity integrated over frequency is 

204: equal to the energy deposition rate of the ISM 

205: electrons in the CB\footnote{The kinetic energy of a CB is mainly lost to 

206: the ISM protons it scatters; only a fraction $\rm\leq m_e/m_p$ is 

207: re-emitted by electrons, as the AG.}:

208: \begin{equation} \rm {dE\over dt} 

209: \simeq \eta\, \pi\, R_{max}^2\, n_e\, m_e\, c^3\, \gamma(t)^2 , 

210: \label{cbemissivity} 

211: \end{equation} 

212: where $\rm n_e\gamma$ is the ISM 

213: electron density in the CB rest system and $\eta$ is the fraction

214: of ISM electrons that enter the CB

215: and radiate there the bulk of their incident energy.  

216: In the early afterglow Eq.~(\ref{cbemissivity}) 

217: must be modified to account for the fact that  the bulk of the radio emission

218: by the incoming ISM electrons

219: is delayed by the time it takes them to cool down to energies

220: much lower than their initial one.

221: This implies that Eq.~(\ref{cbemissivity}) must

222: be modified by a multiplicative ``cumulation factor'' $\rm C(\nu,t)$,

223: which is $\approx 1$ at optical and X-ray wavelengths, as discussed in detail

224: in Section 5. Two other factors,  discussed in Sections 6 and 7, 

225: distinguish radio

226: waves from higher-frequency emissions: attenuation by self-absorption 

227: and limb darkening; they introduce two extra factors $\rm A_{_{CB}}[\nu]$

228: and $\rm L_{_{CB}}(\nu,\theta_{_{CB}})$, with $\rm \theta_{_{CB}}$ a

229: direction of emission relative to the CB's velocity vector.

230: Normalized as in Eq.~(\ref{cbemissivity}) and corrected by all these factors,

231: the afterglow energy flux density of a CB is:  

232: \begin{eqnarray} 

233: \rm F_{_{CB}}[\nu,t,\theta_{_{CB}}] &\simeq&  \rm 

234:                \eta\, \pi\,R_{max}^2\, n_e\, m_e\, c^3 \gamma(t)^2  

235:  \;f_{sync}(\nu,t)\nonumber \\  &\times & \rm

236: C(\nu,t)\, A_{_{CB}}[\nu]\, L_{_{CB}}(\nu,\theta_{_{CB}})\, , 

237: \label{Fnucb} 

238: \end{eqnarray} 

239: to be summed over $\rm n_{_{CB}}$ for a jet with that number of 

240: cannonballs. This expression, for $\rm\nu\gg\nu_b$ and the second

241: row set to unity, reproduces the optical and X-ray AG result 

242: discussed in DDD 2001.  

243: 

244: An observer in the GRB progenitor's rest system,

245: viewing a CB at an angle $\theta$ (corresponding to

246: $\rm\theta_{_{CB}}$ in the CB's proper frame), sees its radiation 

247: Doppler-boosted by a factor $\delta$:

248: \begin{eqnarray} 

249: \rm 

250: \delta(t)&\equiv&\rm

251: {1\over\gamma(t)\,(1-\beta(t)\cos\theta)}

252: \simeq {2\,\gamma(t)\over 1+\theta^2\gamma(t)^2}\; , 

253: \nonumber \\ \rm

254: \cos\theta_{_{CB}}&=&\rm

255: {\cos\theta-\beta(t)\over 1-\beta(t)\,\cos\theta}

256: \simeq{1-\theta^2\gamma(t)^2\over1+\theta^2\gamma(t)^2}

257: \label{doppler} 

258: \end{eqnarray} 

259: where the approximations are valid in the domain of interest for GRBs: large $\gamma$ and small $\theta$. Since the CB is catching-up\footnote{For

260: $\theta\gamma>1$, the observer sees the back of the CB as it is coming

261: forth towards her: $\rm \cos\theta_{_{CB}}<1$. The CB would actually

262: hit or pass by the observer before its back is unveiled, were it not for the 

263: fact that its motion is decelerated.}

264: with the radiation it emits, $\delta$ is also the relative time aberration:

265: $\rm dt_{obs}=dt_{_{CB}}/\delta$. The observed spectral energy 

266: density is modulated by a factor $\delta^3$, two powers of $\delta$ reflecting

267: the relativistic forward collimation of the radiation emitted in the CB's

268: rest system.  The AG spectral energy density $\rm F_{obs}$

269: seen by a cosmological observer at a redshift $\rm z$ 

270: (Dar and De R\'ujula, 2000a), is:

271: \begin{equation} 

272: \rm F_{obs}[\nu,t]\simeq 

273: \rm {A_{Gals}\, (1+z)\,\delta^3 

274:                     \over 4\, \pi\, D_L^2}\, 

275: F_{_{CB}}\left[{(1+z)\,\nu\over\delta(t)},{\delta(t)\,t\over 1+z} 

276: \right]\, , 

277: \label{Fnuobser} 

278: \end{equation} 

279: where $\rm A_{Gals}$ represents the absorption in the host galaxy and the

280: Milky Way,   $\rm F_{_{CB}}$

281: is as in Eq.~(\ref{Fnucb}), and $\rm D_L$ is the luminosity distance

282: (we use throughout a cosmology with $\Omega=1$ and 

283: $\Omega_\Lambda=0.7$).  In the CB model, the extinction

284: in the host galaxy may be time dependent: in a day or so,

285: CBs typically move to kiloparsec distances from their birthplace,

286: where the extinction should have drastically diminished.

287: 

288: In DDD 2001 we fit, in the CB model, the R-band AG light curves of 

289: GRBs. The fit involves five parameters per GRB:

290: the overall normalization;  $\theta$: 

291: the viewing angle; $\gamma_0$: the $\rm t=0$ value of the

292: Lorentz factor; $\rm x_\infty$: the ``deceleration''

293: parameter of the CBs in the ISM; and the

294: spectral index $\rm p$.  The value of $\rm p$, obtained

295: from the temporal shape of the afterglow, is in every case

296: very close to the expectation $\rm p=2.2$, and ---within

297: the often large uncertainties induced by absorption--- with

298: the observed spectra from optical frequencies to X-rays (DDD 2001).

299: 

300: In this paper we complete our previous work by making broad-band

301: fits to the data at all available radio and optical frequencies.

302: In so doing, we need to introduce {\it just one} new ``radio''

303: parameter:  an ``absorption frequency'' $\rm \nu_a$, corresponding

304: to unit CB opacity at a reference frequency. We set $\rm p=2.2$ so

305: that the extension to a broad-band analysis does not involve an

306: increase in the total number of parameters.  We have to refer  very

307: often to the values of the parameters that our previous experience

308: with the CB model made us choose as reference values.  For convenience,

309: these are listed in Table I.

310:  

311: 

312: The predictions of the CB model, for typical parameters,

313: are summarized in Fig.~(\ref{figCBpreds}). The energy density spectra 

314: at radio to optical frequencies are shown, at various times after

315: the GRB, in the upper panel. The spectral slopes before and soon after

316: the peak frequency are $\rm 3/2$ and $\rm -(p-1)/2$, as indicated.

317: The spectra peak at a frequency at which self-attenuation in the CBs

318: results in an opacity of $\cal{O}$(1).

319: At frequencies well above the frequency $\rm \nu_b$ characterizing

320: the injection bend, the spectrum steepens to a slope $\rm -p/2$.

321: In the figure's lower panel

322: we show light curves at various radio frequencies. At large times 

323: and for $\rm \nu\gg \nu_b$ ---which is the case at all frequencies

324: in the example of Fig.~(\ref{figCBpreds}), whose parameters

325: are close to those of GRB 000301c--- they

326: tend to $\rm t^{-2\,(p+1)/3}$, this behaviour being reached

327: at earlier times, the higher the frequency. 

328: For $\rm \nu\ll \nu_b$, the corresponding limiting behaviour

329: is $\rm \approx t^{-4/3}$, observable at low frequencies

330: in the cases of GRBs 991216, 991208 and 000418. All of the above

331: predictions are robust: they do not depend on the

332: detailed form of the attenuation, cumulation and limb-darkening

333: factors. The early rise

334: of the light curves does depend on such details, on which

335: we shall have to invest a disproportionate effort in Sections 5 to 7.

336: 

337: %  estaba la primera figura 

338: 

339: The CB model provides an 

340: excellent description of the data, as discussed in Sections 4, 8 and 9.

341: In the case of GRB 980425, for which the optical AG is dominated

342: by SN1998bw, we used the parameters that fit

343: its X-ray afterglow (DDD 2001) and the GRB's fluence (Dar and De R\'ujula

344: 2000a) to argue that they are not exceptional.

345: The CB-model's description of the

346: radio data for this GRB/SN pair is excellent: there is nothing peculiar

347: about this GRB, nor about its associated supernova, as we discuss 

348: in detail in Section 9,  along with the question of the

349: angular separation in the sky of the SN and the associated CBs,

350: which may have been, or may still be, observable.

351: 

352: The apparent sky velocities of cosmological CBs are extremely superluminal

353: and their angular velocities happen to be of the same order of magnitude

354: as those of galactic pulsars. This implies that CB velocities can possibly

355: be extracted from their observed radio scintillations, as discussed in

356: Section 12.

357:  

358:  

359: \section{The  electron spectrum}

360: 

361: \subsection{Numerical simulations}

362: 

363: The acceleration of charged particles by a moving CB is not substantially

364: different from some of the cases already studied in the literature,

365: the acceleration of cosmic rays and electrons having

366: attracted an enormous amount of attention since Fermi's first

367: analysis in 1949 (for an excellent introduction, see Longair 1994).

368: The most efficient and thus promising mechanism is the ``first-order'' 

369: acceleration of particles by fast-moving shocks, extensively

370: studied analitically  and numerically since the pioneering

371: works of  Axford et al. (1977), Krymsky (1977), Bell (1978) and Blandford

372: and Ostriker (1978). The analysis closest to the case at hand

373: is that of Ballard and Heavens (1992), who studied

374: acceleration by relativistic shocks, with the charged particles 

375: deflected by highly disordered magnetic fields, rather than, as it is 

376: generally assumed, by small irregularities in an otherwise uniform field.

377: The ``relativistic'' and ``chaotic'' inputs are what make this work

378: particularly relevant to the case of particle acceleration by 

379: and within CBs.

380: 

381: Ballard and Heavens study numerically, for various values of a

382: moving discontinuity's Lorentz factor ranging up to $\rm\gamma_s=5$,

383: the result of its collision with an isotropic ensemble of particles

384: with $\rm \gamma_p=100$. They find that, for $\rm\gamma_s=5$,

385: the resulting particle energy distribution has a break (in this reference

386: system) at $\rm\gamma\sim 10\, \gamma_p$, at which point it steepens.

387: The particles below the break have a dominantly

388: very forward motion: they are the ones which have been upscattered

389: just once. Given this hint, it is easy to reproduce the numerical

390: results in an analytic approximation. In the shock's rest system,

391: the energy of the particles that have been scattered only once

392: is equal to their incoming energy: the break in the spectrum

393: seen in the simulations is a kinematical break occurring roughly

394: at the injection energy. ``Observed'' in the system in which the

395: shock is travelling at $\rm\gamma_s=5$, this {\it injection bend}

396: is very reminiscent of the familiar synchrotron-cooling ``break'',

397: but it has little to do with it; indeed, in the simulations of Ballard and 

398: Heavens (1992) cooling was entirely neglected.

399: 

400: \subsection{A simple analysis}

401: 

402: Consider the CB in its rest system and temporarily postpone the

403: discussion of cooling. The ISM electrons impinge on the CB in a fixed direction 

404: with a Lorentz factor equal to that of the CB in the GRB progenitor's rest system,  

405: $\rm\gamma_e=\gamma(t)$. The electrons not having ``bounced back'' 

406: off the CB's strong magnetic field, or having

407: done it only once, retain the incoming energy, $\rm E_b(t)=\gamma(t)\,m_e\, c^2$,   

408: so that their energy distribution is:

409: $\rm dn_e/dE\propto \delta[E-E_b(t)]$.

410: A very robust (i.e. detail independent) feature of the studies

411: of acceleration by relativistic shocks is that the particles

412: having bounced more than once acquire a spectrum 

413: $\rm dn_e/dE\propto E^{-p}$,

414: with $\rm p=2$ in analytical approximations and $\rm p\sim 2.2$

415: in numerical simulations. A few bounces are sufficient to attain

416: such a spectrum. The CB is a system of finite transverse dimensions

417: and the magnetic field contrast between its interior and its exterior

418: is very large. Thus, we do not expect the same electron to bounce

419: many times off the CB, as the latter catches up with it. The

420: acceleration should occur mainly within the CB as charged particles

421: bounce off its chaotically moving magnetic domains, and  it should be very

422: fast and efficient, since the injection is highly relativistic and

423: there is no distinction between ``first and second order Fermi'' processes.

424: The overall ``source''

425: spectrum of relativistic electrons is:

426: \begin{eqnarray}

427: \rm {dn_e^s\over dE}&\sim&\rm A_1(t)\,E_b(t)\,\delta[E-E_b(t)] 

428: \nonumber\\ &+& \rm

429: A_2(t)\,\Theta[E-E_b(t)] \left[{E\over E_b(t)}\right]^{-p}\!\!\! ,

430: \label{source}

431: \end{eqnarray}

432: with $\rm A_1$ and $\rm A_2$ of comparable magnitude and a time 

433: dependence which is that of the rate, 

434: $\rm\eta\,\pi\,R_{max}^2\,c\,n_e\,\gamma(t)$,

435:  at which electrons enter the CB.

436: 

437: \subsection{The spectrum of cooled electrons} 

438: 

439: The electron energy loss by synchrotron radiation is:

440: \begin{eqnarray}

441: \rm -{dE\over dt}&=&\rm A_S\,\beta^2\,E^2 ,\nonumber \\ \rm

442: A_S &\equiv& \rm 

443: {B^2\over 6\,\pi}\,{\sigma_T\,c\over (m_e\,c^2)^2}\, ,

444: \label{Eloss}

445: \end{eqnarray}

446: with $\beta\approx 1$ for the relativistic energies of interest and

447: $\rm \sigma_T=0.665$ barn the Thomson cross-section.

448: Let the rate at which fresh electrons are supplied by the ISM be called $\rm R$.  

449: The electron source distribution of Eq.~(\ref{source}) ``ages'' by cooling so that:

450: \begin{equation}\rm

451: {\partial\over\partial t}\left[{dn_e\over dE}\right]=

452: {d\over dE}

453: \left[{dE\over dt}\,{dn_e\over dE}\right]+

454: R\,{dn_e^s\over dE}\, .

455: \label{nequilib}

456: \end{equation}

457: At times longer than the synchrotron cooling time, the electron

458: distribution tends to a time-independent $\rm {dn_e/ dE}$,

459: obtained by equating to zero the l.h.s. of Eq.~(\ref{nequilib})

460: and integrating it with the source function of Eq.~(\ref{source}):

461: \begin{eqnarray}

462: \rm {dn_e\over dE}&\sim&\rm A_1(t)\,\Theta[E_b(t)-E]\,{1\over E^2} 

463: \nonumber\\

464: &+& \rm {A_2(t) \over p-1}\,\Theta[E-E_b(t)]\,\left[{E\over 

465: E_b(t)}-1\right]^{-(p+1)}\! . 

466: \label{result}

467: \end{eqnarray}

468: 

469: Admittedly, the process of acceleration that we have discussed is

470: not well understood, our derivation is heuristic and Eq.~(\ref{result})

471: is not even a continuous function (the step function in Eq.~(\ref{source})

472: should not be so abrupt, the magnetic energy in Eq.~(\ref{Eloss}) 

473: should not have a fixed value). All we want to conclude from this

474: exercise is that,  when the probability of

475: an electron to have been ``kicked'' only once is not negligible

476: ($\rm A_1$ comparable to $\rm A_2$), 

477: the electron spectrum  has an injection bend at

478: $\rm E\sim E_b(t)$, around which its spectral index changes by 

479: $\sim 1$

480: from $\rm \sim 2$ to $\rm \sim p+1$. We choose to characterize this

481: behaviour by the function:

482: \begin{equation}

483: \rm  {dn_e\over dE} \propto {E^{-2}\over \sqrt{1 + [E/E_b(t)]^{2\,(p-1)}}}\, .

484: \label{approx1}

485: \end{equation}

486: Note how similar the injection bend is to a cooling break

487: (also a spectral steepening by roughly one unit)

488: even though their origins are so different.

489: The observational evidence for an injection bend at the

490: injection energy turns out to be strong, as we proceed to show.

491: 

492: \section{Evidence for an injection bend}

493: 

494: The injection bend induces the gradual transition

495: in the spectral energy distribution described by Eq.~(\ref{sync}), occurring

496: at a ``bend'' frequency:

497: \begin{equation}

498: \rm \nu_b \simeq {1.87\times 10^{15} \over 1+z}\, 

499: {[\gamma(t)]^3\, \delta(t)\over 10^{12}}\,

500: \left[{n_p\over 10^{-3}\;cm^3}\right]^{1/2}

501: Hz,

502: \label{nubend}

503: \end{equation}

504: where we have used the characteristic synchrotron frequency 

505: of Eq.~(\ref{nub}) for

506: the magnetic field of Eq.~(\ref{mag}), and transposed the result to the 

507: observer's frame.

508: 

509: %estaba la segunda figura

510: 

511: For the reference CB parameters and $\rm z=1$,

512:  $\rm\nu_b(t=0)\simeq 0.93\times 10^{15}$ Hz,

513: above the optical band. Since the product $\gamma^3\, \delta$ typically

514: declines by more than an order of magnitude within a couple of days,

515: the bend frequency in many GRBs

516: crosses the optical band into the NIR during the early afterglow.

517: In Fig.~(\ref{figinjection}) we present the time dependence of 

518: $\rm\nu_b(t)$  for $\gamma_0=1250$ and 750, characterizing the range

519: of the observations, for various angles $\theta$, $\rm z=1$,

520: and the rest of the parameters at their reference values of Table I. 

521: The figures show that, depending on the parameters, the bend frequency

522: in the early AG may be above or below the optical band, and, if it is above,

523: it will cross it later.

524: 

525: The bend frequency of the CB model is not the

526: break frequency of the traditional fireball model. 

527: The time evolution of the former is given by Eq.~(\ref{nubend}),

528: and is different from that of the latter, which,  

529: prior to the ``break'' in the AG light-curve, can be shown to be

530: $\rm t^{-1/2}$ (Granot and Sari, 2002).

531: 

532: The evolution predicted by Eq.(~\ref{sync}) from a $\nu^{-0.5\pm 0.1}$ 

533: to a $\nu^{-1.1\pm 0.1}$ spectral behaviour is affected by extinction.

534: The early behaviour corresponds to times when CBs are not yet

535: very far from their progenitors:  extinction in the host galaxy

536: may steepen the spectrum. After a day or more, when the CBs

537: are further away, we do not expect strong extinction in the host.

538: So the prediction (after extinction in the Galaxy is corrected for)

539: is an evolution from a behaviour close to ---or steeper than---

540: $\nu^{-0.5\pm 0.1}$, to a more universal $\nu^{-1.1\pm 0.1}$ at

541: later times.

542: 

543: 

544: The predicted spectral behaviour has been  observed,

545: with varying degrees of significance, in

546: the AG of several GRBs, listed in Table II. 

547: The first column is the bend frequency $\rm \nu_b^0$ at $\rm t=0$,

548: computed with Eq.~(\ref{nubend}) and the optical AG parameters of 

549: Table III

550: (the density $\rm n_p$ is extracted from the measured

551: $\rm x_\infty$ with use of Eq.~(\ref{range})

552: and our reference $\rm R_{max}$ and $\rm N_{_{CB}}$). For the listed

553: GRBs the bend frequency is above the visible band at $\rm t=0$ and the

554: early AG measurements result in effective spectral slopes, $\rm \beta(t_1)$,

555: not far from the expectation $-0.5\pm 0.1$, or somewhat steeper. 

556: A few days later, the measured

557: values, $\rm \beta(t_2)$, are compatible with the expectation ${-1.1\pm 0.1}$.

558: The second entry on GRB 990510 in Table II

559: (Beuermann et al. 1999)

560: requires an explanation. These authors argue that  

561: $\rm \beta(t_2)=0.55\pm 0.10$, a result that assumes a strong

562: extinction correction in the host galaxy. But, after a day or so,

563: we do not expect such an extinction. For the latest points measured by  Beuermann et al. (1999),

564: at day 3.85 (well after the bend), $\rm B-R=0.98\pm 0.07$

565: and $\rm R-I=0.49 \pm 0.06$. Converting these results ---without

566: extinction--- to a spectral slope yields $\rm \beta(t_1)=1.11\pm 0.12$,

567: in agreement with  expectation.

568: 

569: The evolution from a

570: softer to a harder spectrum should be a gradual change in time, rather

571: than a sharp break, so that an AG's optical spectrum, if ``caught'' as

572: the injection bend is ``passing'' should have an index evolving

573: from $-0.5\pm 0.1$ to ${-1.1\pm 0.1}$ with the time dependence

574: described by  Eqs.~(\ref{sync}) and (\ref{nubend}).

575: In Fig.~(\ref{index2}) we test this prediction in the case of GRB 970508, for 

576: the time-dependent value of the ``effective'' slope

577:  $\rm \alpha \simeq \, \Delta\, [log\, F_\nu]/\Delta\,[log\,\nu]$, 

578: constructed from the theoretical expectation in the

579: same frequency intervals used by the observers.

580: The actual predicted $\rm \nu_b(t)$ in Eq.~(\ref{nubend}) is obtained by use of

581: the optical-AG fitted parameters ($\theta$, $\rm \gamma_0$ and

582: $\rm x_\infty$) that determine $\rm \gamma(t)$ and $\rm\delta(t)$, and

583: the density $\rm n_p$  deduced\footnote{GRB 970508 has a peculiar

584: AG, whose CB-model interpretation requires an ISM density

585: change at $\rm t\sim 1.2$ observer's days (DDD 2001).} from

586: $\rm x_\infty$ and the reference $\rm N_{_{CB}}$ and $\rm R_{max}$.

587: The data are gathered by Galama et al. (1998a) from 

588: observations in the U, B, V, $\rm R_c$ and $\rm I_c$ bands

589: (Castro-Tirado et al. 1998, Galama et al. 1998b;

590: Metzger et al. 1997; Sokolov et al. 1998; Zharikov et al. 1998), 

591: by Chary et al. (1998)

592: for K band results, and by Pian et al. 1998 for the H band.

593: 

594: In spite of considerable uncertainties in the 

595: spectral slopes deduced from observations (Galama et al. 1998a), the

596: results shown in Fig.~(\ref{index2}) are satisfactory: the observed crossing

597: of the injection bend is in agreement with the theoretical {\it prediction},

598: based on the fit in DDD 2001 to the overall R-band light curve from which

599: the GRB 970508 AG parameters have been fixed; 

600: {\it no extra parameters have been fit.}

601: A couple of points in the lower panel do not agree with the prediction,

602: but they do not agree with the observations at very nearby frequencies reported

603: in the upper panel, either.

604: 

605: % estaba la tercera figura

606: 

607: A complementary analysis to that in the previous paragraph is the

608: study of an AG's optical spectrum at a fixed time at which

609: the injection bend is crossing the observed frequency range, 

610: or is nearby. A spectral ``snapshot'' at such time should have

611: the intermediate slope

612: and curvature described by Eq.~(\ref{sync}) for $\rm\nu\sim\nu_b$.

613: To test this prognosis, 

614: we compare in Fig.~(\ref{index}) the predicted spectral 

615: shape of the optical/NIR AG of GRB 000301c around March 4.45 UT 

616: ($\sim 3$ days after burst) to its measured shape (Jensen et al. 2001).

617: We have selected this GRB because its extinction correction in the galactic ISM

618: is rather small: $\rm E(B - V)=0.05$ (Schlegel et al. 1998),

619: and there is no evidence for  significant extinction in the

620: host galaxy (Jensen et al. 2001). The theoretical line in Fig.~(\ref{index})

621: is given by Eq.~(\ref{sync}) with the observer's $\rm \nu_b$

622: of Eq.~(\ref{nubend}) ($\rm \nu_b\,(1+z)=1.75\times 10^{14}$ Hz at

623: $\rm t=3$ days, for the density deduced from the value of $\rm x_\infty$

624: of this GRB, and the reference values of $\rm N_{_{CB}}$ and $\rm R_{max}$).

625: In the figure the theory's normalization is arbitrary but the (slightly

626: evolving) slope of the theoretical curve is an absolute {\it prediction}:

627: it is based on the fit in DDD 2001 to the overall R-band light curve and,

628: once more, {\it no extra parameters have been fit}.

629: The result is astonishingly good, even for the curvature which 

630: ---given the figure's aspect ratio as chosen by 

631: the observers--- is not easily visualized (a look at a slant angle helps).  

632: The late-time spectral slope deduced 

633: from the HST observations (Smette et al. 2001) around day 

634: 33.5 after burst indicated a slope of $\sim -1.1$, again in agreement

635: with our expectation. 

636: 

637: % estaba la cuarta figure

638: 

639: We conclude that the evidence is very strong for a spectral injection bend at

640: the time-dependent frequency, Eq.~(\ref{nubend}), predicted in the CB model.

641: As illustrated in Fig.~(\ref{figCBpreds}) and contrasted with data

642: in Section 8, further evidence for the injection bend is provided

643: by the fact that it is essential to the description of the

644: observed broad-band spectra of GRB afterglows.

645: 

646: \section{The cumulation factor}

647: 

648: Three factors that are irrelevant in the optical and X-ray domains

649: play a role in the description of the longer radio wavelengths

650: and the early radio AG. In this section we discuss the first one of them.

651: 

652: Electrons that enter a CB with an injection Lorentz factor  $\rm \gamma(t)$

653: are rapidly Fermi accelerated to a distribution that we have argued

654: to be roughly that of Eq.~(\ref{source}). On a longer time scale, they

655: lose energy by synchrotron radiation, and their

656: energy distribution evolves as in Eq.~(\ref{nequilib}). 

657: Electrons with a large $\rm \gamma\sim {\cal{O}}\,[\gamma(t)]$ emit

658: synchrotron radiation, with no significant time-delay, 

659: at the observer's optical and X-ray wavelengths.

660: But the emission of radio is delayed by the time it takes the electrons

661: to ``descend'' to an energy at which their characteristic emission

662: is in the observer's radio band. At the start of the afterglow, when 

663: equilibrium conditions have not yet been reached, this implies a

664: dearth of radio emission relative to the higher-frequency bands.

665: This introduces a ``cumulation factor'' $\rm C(\nu,t)$ in Eq.~(\ref{Fnucb}).

666: 

667: Consider a fixed observed radio frequency $\rm \nu_{obs}$. 

668: It corresponds to a time changing frequency 

669: $\rm \nu=(1+z)\nu_{obs}/\delta(t)$ in the CB system.

670: The CB electrons preferentially emitting at this frequency (over an

671: unconstrained range of pitch angles) are those

672: whose Lorentz factor $\rm\gamma_e$ satisfies the relation

673: $\rm \nu \sim 0.22\,\gamma_e^2\,\nu_L$, in analogy to Eq.~(\ref{nub}). To

674: estimate\footnote{We can solve Eq.~(\ref{nequilib}) exactly for a given

675: source spectrum by the Mellin transform

676: methods so familiar in Quantum Chromodynamics, but this would 

677: be unjustified: the acceleration mechanism is not understood well

678: enough for the study of exact cooling solutions to be currently justifiable.} 

679: the time $\rm \Delta t$ it takes an electron to decelerate from

680: $\rm\gamma\sim \gamma(t)$ to $\rm\gamma=\gamma_e$, substitute the magnetic

681: energy density of Eq.~(\ref{emag}) into the electron energy loss of Eq.(\ref{Eloss})

682: and integrate, to obtain

683: \begin{eqnarray}\rm

684: \Delta t&=&\rm {3\,m_e\over n_p\,m_p\,\sigma_T\,c}\,{1\over\gamma^2}\,

685: \left({1\over \gamma_e}-{1\over \gamma}\right) \nonumber \\

686: &=&\rm [8.27\times 10^7\,s]\left[{10^{-3}\,cm^{-3}\over n_p}\right]

687: {10^6\over \gamma^2}\,

688: \left({1\over \gamma_e}-{1\over \gamma}\right)\, .

689: \label{Deltat}

690: \end{eqnarray}

691: The function $\gamma$ is given by Eq.~(\ref{gammacb}) of Appendix I,

692: which we may rewrite as:

693: \begin{eqnarray}\rm

694: {1\over \gamma^2}&=&\rm{1\over \gamma_0^2}\left[1+{t\over t_0}\right]

695: \nonumber\\ \rm

696: t_0&\equiv&\rm [5.14\times 10^7\, s]

697:     \left[{x_\infty\over 1\;Mpc}\right]

698: \left[{10^3\over \gamma_0}\right]^2

699: \label{1overg2}

700: \end{eqnarray}

701: 

702: The electrons emitting the observed radio frequencies have 

703: $\rm\gamma_e\!\sim\! {\cal{O}}(1)$, so that the proper CB times 

704: $\rm\Delta t$ and $\rm t_0$ are of ${\cal{O}}(1)$ year,

705: corresponding to observer's times ---foreshortened

706: by a factor $\rm(1+z)/\delta$--- of ${\cal{O}}(1)$ day. 

707: For optical and radio observations $\rm\gamma_e\!\sim\! {\cal{O}}(\gamma)$

708: there is no significant delay in their emission. Moreover, the 

709: electron accumulation rate ($\rm \eta \pi\,R_{max}^2\,n_e\,c\,\gamma$ in the

710: CB system)  is orders of magnitude larger than the characteristic

711: synchrotron cooling time $\rm E/(dE/dt)$ of Eq.~(\ref{Eloss}), even for

712: $\gamma\sim 10^3$. Thus, the optical and X-ray AG emission

713: starts as soon as the CB is transparent to its enclosed radiation: for 

714: each CB, a few observer's seconds after the corresponding $\gamma$-ray

715: pulse (DDD 2001). The radio signal, on the other hand, must await 

716: a time $\rm \Delta t$ for the

717: cumulated electrons to cool down. 

718: 

719: The simple way to parametrize the frequency-dependent

720: ``cumulation effect'' is to use the expression for the total number

721: of electrons $\rm N(t)$ incorporated by the CB up to time $\rm t$ 

722: (Eq.~(\ref{acctot}) of Appendix I) and to posit\footnote{The rapid onset

723: of the radio signals from SNe is not understood, see e.g.

724: Weiler et al. 2001. Perhaps electron cumulation also plays a role there.}:

725: \begin{equation}\rm

726: C(\nu,t)={N(t-\Delta t)\over N(t)}\,\Theta(t-\Delta t),

727: \label{cumul}

728: \end{equation}

729: where the frequency dependence is via $\rm \Delta t=\Delta t(\nu,t)$,

730: and the sharp start at $\rm t>\Delta t$ is an artifact of our simplifications.

731: For optical and X-ray frequencies, $\rm \Delta t=0$ and $\rm C(\nu,t)=1$.

732: In practice we find that, except for GRB 980425 whose viewing angle

733: is exceptionally large, one may also use {\it within} the radio band an

734: approximation to Eq.~(\ref{cumul}): 

735: \begin{equation}\rm

736: C(\nu,t)\sim C(t)=\left[1-{\gamma(t)\over\gamma_0}\right]^{1/2},

737: \label{cumulapp}

738: \end{equation}

739: which is also frequency independent. 

740: 

741: 

742: 

743: \section{The attenuation factor}

744: 

745: At optical and X-ray frequencies the CB is transparent and, for the spectrum

746: of Eq.~(\ref{sync}), the bulk of the radiation's energy is emitted

747: around the bend frequency $\rm\nu_b$. At such relatively high

748: frequencies, as illustrated in Fig.~(\ref{figCBpreds}), absorption is 

749: unimportant. Thus, for optical and

750: X-ray afterglows (DDD 2001) it suffices to know that all of the incoming

751: electron's energy is reradiated, the spatial distribution of the

752: radiating electrons within the CB is irrelevant. But in the radio,

753: where absorption is important, the location of these electrons

754: inevitably plays a role. In the next sections

755: we argue that it is plausible

756: that the radiating electrons be close to the surface ``illuminated''

757: by the ISM (\S 6.1), and that the values of the CB's plasma frequency (\S 6.2)

758: and free-free absorption coefficient (\S 6.3) actually suggest that they may 

759: be relatively close to that surface. In \S 6.4 we deduce the final form

760: of the attenuation factor in the CB model, characterized by a single parameter.

761: 

762: \subsection{Electron penetration}

763: 

764: Using numerical simulations, Achterberg al. (2001) have shown 

765: that for simple geometries  the bulk of highly relativistic particles 

766: encountering a collisionless shock escape before

767: they undergo diffusive shock acceleration.  In reality, the

768: geometry of the CB, its density distribution and its magnetic field 

769: distribution are very complicated, making  

770: the fraction of the ISM electrons that penetrate inside the CB, and

771: their distribution there, very uncertain.

772: 

773: Several length scales play a role in discussing the fate of an electron that 

774: enters the CB with $\rm\gamma_e=\gamma(t)$. The Larmor radius  

775: is $\rm R_L=m_e\,c^2\,\gamma/(e\,B)$,

776: which is independent of $\gamma$ for $\rm B$ scaling as in

777: Eq.~(\ref{mag}). For our reference parameters, $\rm R_L\sim 6$ km is many

778: orders of magnitude smaller than the CB's radius and does not play

779: a crucial role. The length of an electron's curled-up trajectory as 

780: it radiatively loses energy  is $\rm c\,E/(dE/dt)$ or $\sim 2.6\times 

781: 10^{15}$ cm for the cooling rate of Eq.~(\ref{Eloss}) and an initial 

782: $\gamma=10^3$. This is only an order of 

783: magnitude larger than the reference CB's radius $\rm R_{max}$.

784: We have no way to estimate the typical coherence size of a CB's 

785: magnetic domain $\rm L_B$, but the depth 

786: $\rm D\sim (c\,\tau_\gamma\;L_B)^{1/2}$ to which an electron penetrates, 

787: even for a relatively simple magnetic mess ($\rm L_B$ not much smaller than

788: $\rm R_{max}$) is smaller than the CB's radius.

789: For $\rm L_B$ as small as $\rm R_L$,

790: $\rm D\sim 4\times 10^{10}$ cm, some four orders of magnitude smaller than 

791: $\rm R_{max}$. Even this concrete value is uncertain, for it depends on the

792: surface magnetic field as $\rm B^{-3/4}$, and  the surface $\rm B$-value

793: may be different from that of Eq.~(\ref{mag}), which is a volume average.

794: 

795: In addition to all of the above uncertainties, it is possible that a CB's

796: illuminated working surface be turbulent, and harbour fast plasma motions, 

797: if only to establish local charge neutrality, which is disrupted as electrons

798: and protons penetrate the CB to different depths. 

799: We conclude that the the fraction of ISM electrons that enter inside

800: the CB {\it may} be small  and the synchrotron-radiating

801: electrons {\it may} be concentrated close to the CB's surface, as opposed

802: to be acquiring a uniform distribution over the CB's 

803: volume. 

804: 

805: \subsection{The plasma frequency}

806: 

807: The plasma frequency in a CB with an average free inner electron density

808: $\rm \bar n_e^{free}$ is: 

809: \begin{equation} 

810: \rm \nu_p = \left[{\bar

811: n_e^{free}\, e^2\over\pi\, m_e}\right]^{1/2}\simeq 28\, \left[{\bar

812: n_e^{free}\over 10^7\;cm^{-3}}\right]^{1/2}\, MHz. 

813: \label{plasma}

814: \end{equation} 

815: For a fully ionized CB $\rm \bar n_e^{free}=\bar n_e$, to

816: whose reference value we normalized the above result (the fraction of

817: electrons swept up from the ISM is small, relative to the total number in

818: the CB --to whose free fraction Eq.~(\ref{plasma}) refers--

819: but the CB is highly ionized, as shown in Appendix III). 

820: 

821: For $\rm \nu\!<\!\nu_p$

822: the radio emission is completely damped within a typical 

823: length $\rm \sim c/[\nu_p^2-\nu^2]^{1/2}$, much smaller than the CB's radius.

824: At very early times, $\delta\sim 10^3$ and the Doppler-boosted value 

825: of $\rm \nu_p$ falls in the low end of the observed range of radio signals,

826: where a sharp cutoff is not observed. We must conclude that the (small

827: fraction of)

828: radiating electrons is located in a CB surface layer whose total

829: electron density (dominated by the thermal electron constituency) is 

830: smaller than our reference average value, 

831: a one order of magnitude reduction being

832: comfortably sufficient to move the value

833: of $\rm \nu_p$ to a position below the currently observed frequencies.

834: We have explicitly checked that our 

835: fits do not improve significantly with the inclusion of $\rm \nu_p$ 

836: as a free parameter: the minimization procedure always ``gets rid''

837: of the fit $\rm \nu_p$ by choosing it somewhat below the reference value of

838: Eq.~(\ref{plasma}), and below the lowest measured frequencies.

839:  

840: 

841: \subsection{Free-free attenuation}

842: 

843: At the MHz frequencies in the CB system corresponding to

844: the observed radio frequencies, the 

845: synchrotron emission is strongly attenuated by free-free absorption 

846: (inverse bremsstrahlung) in the CB; free-free absorption dominates over self-synchrotron absorption, as shown in Appendix 4. For a 

847: hydrogenic plasma, the free-free

848: absorption coefficient at radio frequencies is: 

849: \begin{equation} 

850: \rm \chi_\nu \simeq 0.018\, g_{ff}\,\bar n_e\, \bar n_i\, T^{-3/2}\,   

851: \nu^{-2}\, cm^{-1}\, , 

852: \label{chifree} 

853: \end{equation} 

854: where $\rm \bar n_i\simeq \bar n_e$ is the free ion 

855: density in the CB, in units of $\rm 

856: cm^{-3}$, $\rm T$ is the plasma temperature in degrees Kelvin,

857: $\nu$ is in Hertz

858: and $\rm g_{ff}$ is a Gaunt factor for free-free emission, of $\cal{O}$(10)

859: at the relevant frequencies.

860: 

861: 

862: The opacity $\tau_\nu$ of a surface layer of depth $\rm D$ is:

863: \begin{equation} 

864: \rm \tau_\nu(D,t) =\int_{R_{max}-D}^{R_{max}}\chi_\nu \, dr\equiv 

865: \bar \chi_\nu[t]\,D\, . 

866: \label{taufree} 

867: \end{equation}

868: Equilibrium between photoionization of atomic hydrogen in the CB by 

869: synchrotron radiation and 

870: recombination of free electrons and protons to hydrogen 

871: keeps the CB plasma partially ionized during the observed AG. 

872: The Coulomb relaxation rate in the CB is very fast because of its 

873: high density. Consequently, the CB plasma is 

874: approximately in quasi thermal equilibrium. Because of the exponential 

875: dependence of the Saha equation 

876: on temperature, and the high ionization rate, the CB's temperature is kept 

877: practically constant around a few eV, and the ion density and free 

878: electron density become proportional to $\rm \gamma(t)$ (Appendix III).  

879: Using $\rm T_0= 10^5\, K$  and the reference

880: average densities $\rm \bar n_e=\bar n_i\simeq 10^7\, cm^{-3}$ 

881: in Eqs.~(\ref{chifree}) and (\ref{taufree}), we obtain:

882: \begin{equation}

883: \rm

884: \tau_\nu(D,t)\sim 1.4\times 10^2 \left[{D\over R_{max}}\right]

885: \left[{1\;GHz\over \nu}\right]^2 \,

886: \left[{T\over 10^5}\right]^{-3/2}\!\! ,

887: \label{taubis}

888: \end{equation}

889: which is very large for $\rm D\sim {\cal{O}}(R_{max})$. A reduction in surface

890: or average CB

891: density of one order of magnitude or more ---which, as we have seen, renders

892: unobservable the unobserved plasma-frequency cutoff---

893: reduces $\tau_\nu$ by two orders of magnitude, or more. 

894: For $\rm D\sim R_{max}/10$, this would make $\rm \tau(D)$,

895: as required, of order unity at the peak frequency 

896: $\sim 10^2$ GHz of the early-time spectrum of Fig.~(\ref{figCBpreds}) 

897: (at which time the observed and CB frequencies differ by a factor 

898: $\delta/(1+z)\sim 10^3$). 

899: 

900: 

901: The conclusion is that a reasonable deviation of the properties

902: of the CB from their reference bulk average values (a reduction

903: of the total number-density of free electrons in 

904: a synchrotron-emitting surface layer) implies, not only that the

905: plasma-frequency break is not observable in the current data, but also

906: that the magnitude of the free-free attenuation is the required one.

907:  Our ignorance of the depth, temperature and density of ions and

908: electrons in the radio-emitting surface of a CB can be absorbed into

909: a single parameter: a characteristic absorption frequency,

910: $\rm \nu_a$, in the opacity of 

911: Eqs.~(\ref{chifree},  \ref{taufree}, \ref{taubis}):

912: \begin{equation}\rm

913: \tau_\nu\equiv\left[{\nu\over\nu_a}\right]^{-2}\,\left[{\gamma(t)\over

914: \gamma_0}\right]^{2}.

915: \label{absfreq}

916: \end{equation}

917: The frequency dependence of the free-free attenuation, 

918: $\chi_\nu\propto\nu^{-2}$, is fairly well supported by the observed

919: radio spectra at their lowest frequencies, as our comparisons

920: with observations in Sections 8 and 9 demonstrate.

921: 

922: \subsection{Attenuation in slabs and spheres}

923: 

924: We do not know a priori  the geometry  of the working surface

925: from which a CB's synchrotron radiation is emitted. In the case of

926: optical AGs this is immaterial, for the CB is transparent to radiation

927: at the corresponding CB-system wavelengths: the bulk of the radiation

928: energy is emitted at these frequencies.

929: For the case of radio AGs, attenuation is important

930: and the shape of the emitting surface layers plays a role: the expression

931: for attenuation as a function of opacity is geometry-dependent.

932: 

933: For a planar-slab geometry, the familiar expression for the attenuation is:

934: \begin{equation} 

935: \rm 

936: A[\nu]={1-e^{-\tau_\nu}\over \tau_\nu}\, . 

937: \label{Attslab} 

938: \end{equation} 

939: For the emission from a sphere of constant properties, we obtain:

940: \begin{equation} 

941: \rm 

942: A[\nu]= {3\, \tau_\nu^2-6\, \tau_\nu+6\, [1-e^{-\tau_\nu}]\over 

943: \tau_\nu^3}\, , 

944: \label{Attsphere1} 

945: \end{equation} 

946: while for the emission from a thin spherical surface, the result

947: of Eq.~(\ref{Attslab}) is recovered.

948: 

949: For the sake of definiteness, we adhere to CBs that are spherical

950: in their rest system. This means that, as the frequencies increase

951: and the CB evolves from being opaque to being transparent, we

952: should use an attenuation evolving from Eq.~(\ref{Attsphere1})

953: to Eq.~(\ref{Attslab}). Rather than doing that, we have checked

954: explicitly that our results are insensitive to the use of one

955: or the other form, and used the simpler one.

956: 

957: 

958: \section{The illumination and limb-darkening factor}

959: 

960: Consider a spherical CB in its rest system. It is ``illuminated'' by  

961: incoming ISM electrons only in its ``front'' hemisphere. If observed

962: at an angle $\rm\theta_{_{CB}}\neq 0$, a fraction of the ``dark'' CB

963: is also exposed to the observer, like the Moon in phases other than

964: totality. For radio waves ---to which the CB is not transparent--- these

965: geometrical facts play a non-trivial role.

966: 

967: Place the direction of the CB motion, or of its illumination, at

968: $(\theta,\phi)=(0,0)$; at a direction $\rm \vec n_i=(0,0,1)$ in Cartesian

969: coordinates. The normal to a sphere's surface point at $(\theta,\phi)$

970: is $\rm \vec n_s=(\cos\theta\sin\phi,\sin\theta,\cos\theta\cos\phi)$. 

971: The observer is in the direction $\rm (0,\theta_{_{CB}})$, where we have taken

972: the liberty to label ``$\theta$'' what in this parametrization is 

973: an azimuthal angle; the corresponding

974: unit vector is $\rm \vec n_{_{CB}}=(\sin\theta_{_{CB}},0,\cos\theta_{_{CB}})$.

975: The relation between $\rm \theta_{_{CB}}$ and the terrestrial observer's

976: viewing angle is that of Eq.~(\ref{doppler}).

977: 

978: When attenuation plays a significant role, an element of a CB's 

979: surface reemits an amount of energy proportional to the cosine of

980: the illumination angle: $\rm \vec n_i \cdot \vec n_s$.

981: Because of the limb-darkening effect, the reemitted radiation

982: depends on the cosine of the angle between the surface element

983: and the observer: $\rm \vec n_s \cdot \vec n_{_{CB}}\equiv \mu$.

984: A simple characterization of the functional form of the limb darkening effect

985: (see e.g. Shu, 1991) is:

986: \begin{equation}\rm

987: F(\mu)=\left({2\over 5}+{3\over 5}\,\mu\right)\,\Theta[\mu]\, .

988: \label{limb}

989: \end{equation}

990: The combined effect of illumination and limb darkening is an emitted

991: radiation proportional to:

992: \begin{eqnarray}\rm

993: E(\cos\theta_{_{CB}})&=&\rm

994: \int^1_{-1} d\cos\theta\int^{\pi/2}_{\theta_{_{CB}}-{\pi/ 2}}

995: d\phi\,F(\vec n_s \cdot \vec n_{_{CB}})\, \vec n_i \cdot \vec n_s\nonumber\\

996: &=&\rm {1\over 5}\,[2+(2+\pi-\theta_{_{CB}})\cos\theta_{_{CB}}

997: +\sin\theta_{_{CB}}].

998: \label{ild}

999: \end{eqnarray}

1000: An excellent and simple approximation to Eq.~(\ref{ild}) is:

1001: \begin{equation}\rm

1002: E(x)=E(1)\,{1\over 10}\,

1003: [4+x][1+x]\, ,

1004: \label{cosapp}

1005: \end{equation}

1006: with $\rm x=\cos\theta_{_{CB}}$.

1007: 

1008: For negligible self-attenuation $\rm A[\nu]=1$, as in the optical, 

1009: there is no limb darkening and illumination effect. As absorption

1010: becomes increasingly important for longer wavelengths, the

1011: effect becomes fully relevant. We interpolate between these two

1012: extremes by writing:

1013: \begin{equation}\rm

1014: L_{_{CB}}(\nu,\cos\theta_{_{CB}})\simeq A[\nu]+(1-A[\nu])\,

1015: {E(\cos\theta_{_{CB}})\over E(1)}

1016: \label{illlumdar}

1017: \end{equation}

1018:  to obtain the overall illumination and limb-darkening

1019: correction factor to the energy flux density of Eq.~(\ref{Fnucb}).

1020: 

1021: 

1022: 

1023: \section{Broad band spectra: radio and optical results}

1024: 

1025: In practice, it is not an effortless task to test

1026: a prediction for an AG's spectrum extending, as in the

1027: upper panel of Fig.~(\ref{figCBpreds}),

1028: to all measured wavelengths from radio to X-rays. The problem

1029: is not related to the model, but to the data. First, the

1030: corrections due to absorption, particularly in the host

1031: galaxy, are frequency-dependent and notoriously difficult

1032: to ascertain with confidence. Second, the integration

1033: times employed in the radio observations are long,

1034: so that the theoretical prediction varies within the time

1035: window, and so do the optical energy flux densities,

1036: measured over much shorter periods, as well as some of

1037: the radio observations themselves. Unavoidably, this

1038: will make our spectral figures look a bit peculiar,

1039: with two theoretical curves bracketing the expectations,

1040: and various observational points at the same frequency.

1041: 

1042: 

1043: We study the AG light-curves and broad-band spectra of all

1044: GRBs with known-redshift whose AG was measured both in the

1045: radio and optical bands\footnote{The domain extending from the optical to

1046: the X-ray regime ---but for the early injection bend discussed in

1047: Section 4--- is compatible  with the expected behaviour $\rm

1048: \nu^{-p/2}\sim\nu^{-1.1}$ (DDD 2001).}. 

1049: 

1050: Our predictions are given by Eq.~(\ref{Fnuobser}), 

1051: fit to the optical and radio observations.

1052: The fitted parameters are the overall 

1053: normalization, $\gamma(0)$, $\theta$, the

1054: deceleration parameter $\rm x_\infty$ (whose meaning and role are

1055: reviewed in Appendix I) and the CB self absorption frequency $\rm \nu_a$ of

1056: Eq.~(\ref{absfreq}).  We found in DDD 2001 that $\rm p$ is very narrowly

1057: distributed around its theoretical value $\rm p=2.2$, and we

1058: fix it to that value for all GRBs in the current analysis.

1059: Thus, {\it the total number of parameters in our broad-band fits

1060: is the same as we used in

1061: DDD 2001 to describe just the R-band light-curve}.

1062: 

1063: The values of the parameters, listed in Table III, are very similar 

1064: to those deduced in DDD

1065: 2001 by fitting only the R-band optical data with the high-$\nu$ limit of

1066: Eq.~(\ref{Fnuobser}). The small differences are due 

1067: not only to the use of radio data and optical bands other than R,

1068: but also to the inclusion of the effects of the injection bend in the CB

1069: synchrotron AG global formula, Eq.~(\ref{Fnuobser}), and (to a small

1070: extent) to the use of a fixed $\rm p=2.2$. The results show that the theory 

1071: agrees with observations both at radio and optical wavelengths. 

1072: For some GRBs a slightly 

1073: better fit to the radio data is obtained if the absorption frequency 

1074: $\rm\nu_a$ is best fitted to the radio data alone or if a fitted power-law

1075: dependence on time is used for the CB opacity instead of 

1076: Eq.~(\ref{absfreq}), with all other parameters taken from the global fit.

1077: Because of scintillations, and of the very detail-dependent character of our

1078: prediction for the time dependence of a CB's opacity,

1079: it is difficult to assess whether or not

1080: the slightly improved $\chi^2$ values are significant or not.

1081: 

1082: Notice in Table III that the distributions of parameters are fairly

1083: narrow, in particular for $\gamma_0$. Of particular interest, since

1084: it can be predicted, is the distribution in $\gamma \times\theta$.

1085: Since AGs are discovered at optical and X-ray frequencies,

1086: the angular distribution is that of the high frequency limit of

1087: Eq.~(\ref{Fnuobser}). For small $\theta$, 

1088: $\rm d\,N/d\theta\propto \theta/(1+\gamma^2\theta^2)^{4.1}$.

1089: This distribution has a maximum at $\gamma\theta\sim 0.37$

1090: and a median at $\sim 0.5$. In Table III there are four cases

1091: with $\gamma\theta$ below the median and five above. The

1092: worst ``outlier'' in $\theta$ is much less so in $\gamma\theta$.

1093: The conclusion that this distribution is perfectly compatible

1094: with the expectation can also be reached from Fig.~(39) of

1095: DDD 2001, whose results were obtained from only

1096: optical data, but for which the statistics is a bit better.

1097: 

1098: We first discuss the broad-band spectra and light curves of three

1099: representative GRBs: 000301c, 000926 and 991216.

1100: The optical AG of GRB 000301c is practically unextinct, 

1101: that of GRB 000926 has strong extinction in the host galaxy 

1102: (e.g., Fynbo et al. 2001) and that of GRB

1103: 991216 has strong extinction both in the host galaxy and in

1104: ours (e.g., Halpern et al. 2000). We discuss GRBs 991208, 000418,

1105: 000510, 990123 and 970508 in slightly less detail. The apparently

1106: special case of GRB 980425 is discussed separately in the next chapter.

1107:    

1108: \subsection{GRB 000301c}

1109: 

1110: For this GRB we fit the radio data of Berger et al. (2000)

1111: and the optical data of Garnavich et al. (2000b), Jensen et al.

1112: (2001), Masetti et al., (2000), Rhoads and Fruchter (2001) and

1113: Sagar et al. (2001). Our results for the light curves at all observed

1114: optical and radio frequencies are gathered in Fig.~(\ref{all301}),

1115: which is representative of the trends seen in all GRBs.

1116: The narrowly spaced lines in the figure are the optical light curves

1117: for ---from top to bottom--- the K, J, I, R, V, B and U bands.

1118: Their very satisfactory comparison with data is reported in 

1119: Fig.~(\ref{K301}). The results for the

1120: radio AG are the more spaced lines in Fig.~(\ref{all301}), which

1121: correspond ---from top to bottom at the figure's left side--- to 

1122: frequencies of 1.43, 4.86, 8.46, 15, 22.5, 100, 250 and 350 GHz.

1123: Their very satisfactory comparison with observations is reported

1124: in Figs.~(\ref{figr030101}) to (\ref{figr030104}). Notice that all 

1125: features of the

1126: data have precisely the trends summarized in Fig.~(\ref{all301}).

1127: In Figs.~(\ref{rad-opt301}) and (\ref{rad-opt301b}) we present the

1128: complementary information, by comparing

1129: our fits with the observations for the radio-to-optical spectra

1130: of GRB 000301c in four radio time-integration brackets;

1131: 1 to 5, 5 to 10, 10 to 20, and 20 to 30 days. 

1132: The pronounced peaks are at  (observer's) frequencies 

1133: for which the opacity of Eq.~(\ref{absfreq}) is 

1134: $\rm \tau_\nu\sim{\cal{O}}(1)$. The injection bend at a higher

1135: frequency is clearly visible, it is responsible for the agreement

1136: between the radio and optical magnitudes and frequency trends.

1137: The two curves in these figures, and many later ones, refer to the

1138: expectation at the two times which bracket the actual radio observation.

1139: The results are quite satisfactory.

1140: 

1141: \subsection{GRB 000926}

1142: 

1143: We have made a global fit to the NIR/optical data (Di Paola et al. 2000; 

1144: Fynbo et al. 2001, Harrison et al. 2001; Price et al. 2001; Sagar et al.  

1145: 2001) and the radio data (Harrison et al. 2001) on this GRB. In 

1146: Fig.~(\ref{opt926}) 

1147: we compare the fitted CB-model predictions with the measured light

1148: curves for the I, R, V, B and U bands, after subtraction of the

1149: host galaxy and SN contributions (DDD 2001).

1150: The theoretical predictions were corrected for galactic 

1151: extinction E(B - V)=0.0235 (Schlegel et al. 1998) and for the estimated

1152: extinction in the host galaxy, E(B - V)=0.40 (Harrison et al. 2001).

1153: In Figs.~(\ref{figr092601}) to (\ref{figr092603})

1154: we present the radio light curves for

1155: six frequencies ranging from 98.48 to 1.43 GHz.

1156: 

1157: In Figs.~(\ref{rad-opt926}) to (\ref{rad-opt926c}) we

1158: make the complementary comparison of theory and

1159: observations for the radio-to-optical spectra,

1160: in six time intervals extending from 0.8 to 100 days. The results, in spite

1161: of the crude estimate of extinction in

1162: the host galaxy and the scintillations so clearly visible in the radio

1163: light curves, are satisfactory.

1164: 

1165: \subsection{GRB 991216} 

1166: 

1167: The NIR/optical data for this GRB are from

1168: Halpern et al. (2000) and Garnavich et al. (2000a); the

1169: radio data from Frail et al. (2000b). 

1170: In Fig~(\ref{opt216}) we present the comparison between

1171: the measured light curves for the K, J, I, R bands, after

1172: subtraction of the host galaxy and SN contributions (DDD 2001), and the

1173: fitted CB model predictions. The predictions were corrected for

1174: extinction in the host galaxy and ours, as estimated by Halpern

1175: et al. (2000): E(B - V)=0.40. 

1176: In Figs.~(\ref{figr121602}) to (\ref{figr121603}) 

1177: we present the radio light curves at six frequencies from

1178: 350 to 1.43 GHz.

1179: In Figs.~(\ref{rad-opt216}) to

1180: (\ref{rad-opt216c}) we make the complementary comparison of

1181: theory and observations 

1182: for the radio to optical spectra,

1183: in six time intervals extending

1184: from 0.44 to 80 days. The results are once again satisfactory.

1185: 

1186: \subsection{GRB 991208}

1187: 

1188: 

1189: We fit the 

1190: NIR/optical data (Castro-Tirado et al. 2001; Sagar et al. 2000) and

1191: the radio data (Galama et al. 2000)  on the AG of GRB 991208.

1192: In Fig~(\ref{opt208}) 

1193: we present the comparison between

1194: the measured light curves for the I, R, V and B bands, and the

1195: fitted CB model predictions, after 

1196: subtraction of the host galaxy and SN contributions (DDD 2001).

1197: The theoretical predictions were corrected only

1198: for the small galactic extinction E(B - V)=0.016 (Schlegel et al. 1998) 

1199: in the direction of this GRB, there being no spectral evidence for optical

1200: extinction in the host galaxy. 

1201:  In Figs.~(\ref{figr120801}) to

1202: (\ref{figr120804}) we also present the radio light curves at 

1203: 100, 86.14, 30, 22.49, 14.97, 8.46, 4.86 and 1.43 GHz.

1204: In Figs.~(\ref{rad-opt208}) and

1205: (\ref{rad-opt208b}) we make the complementary comparison 

1206: for the radio to optical spectra in three 

1207: time intervals extending

1208: from 2 to 14.3 days. The results are satisfactory.

1209: 

1210: 

1211: 

1212: \subsection{GRB 000418}

1213: 

1214: 

1215: The NIR/optical data are from Klose et al. (2000)

1216: and the radio data from Berger et al. (2001a).

1217: In Fig.~(\ref{opt418})  we compare

1218: the fitted CB-model predictions with

1219: the measured light curves for the R-band, after

1220: subtraction of the host galaxy and SN contribution (DDD 2001).

1221: The theoretical predictions were corrected for galactic extinction

1222: and for extinction in the host galaxy as estimated by

1223: Berger et al. (2001a): E(B - V)=0.40.

1224: In Figs.~(\ref{figr041801}) and  (\ref{figr041802})

1225: we also present the radio light curves at 22.5, 15, 8.46 and

1226: 4.86 GHz. In Fig.~(\ref{rad-opt418}) we

1227: make the complementary comparison for the radio to optical spectra,

1228: in two time

1229: intervals extending from 9.5 to 100 days.

1230: The results are satisfactory.

1231: 

1232: \subsection{GRB 990510}

1233: 

1234: The NIR/optical data were gathered by Beuermann et al. (1999),

1235: Harrison et al. (1999) and Stanek et al. (1999) and the radio data 

1236: by Harrison et al. (1999). In Fig.~(\ref{opt510}) we present the comparison between

1237: the measured light curves for the I, R, V, B bands, 

1238: after subtraction of the

1239: host galaxy and SN contribution (DDD 2001), and the

1240: fitted CB model predictions, corrected for Galactic extinction

1241: (E(B - V)=0.203, Schlegel et al. 1998) and for extinction in the host

1242: galaxy as estimated by Stanek et al. (1999).

1243: In Figs.~(\ref{figr051001}) and (\ref{figr051002})  

1244: we present the radio light curves at 13.7, 8.6 and  4.8 GHz.

1245: In Figs.~(\ref{figr051002}) and (\ref{rad-opt510})  we also 

1246: make the complementary comparison of theory and

1247: observations for the radio to optical spectra   

1248: three time intervals

1249: extending from 1 to 40 days.

1250: The agreement between theory and observations is very good although

1251: its significance is limited by the sparse radio data.

1252: 

1253: 

1254: 

1255: \subsection{GRB 990123}

1256: 

1257: 

1258: We have fit the NIR/optical data (Castro Tirado 1999; Fruchter et al. 1999;

1259: Galama et al. 1999, Holland et al. 2000; Kulkarni et al. 1999a) 

1260: and the radio data (Galama et al. 1999; Kulkarni et al. 1999b)

1261: for this GRB. In Fig.~(\ref{opt123}) we present the comparison between

1262: the fitted CB model predictions ---assuming a constant ISM density after 0.1 

1263: observer's days\footnote{the earlier optical data are discussed in 

1264: DDD 2001.} and after subtraction of  the host galaxy and SN contributions---

1265: with the measured light curves  for the K, I, R, V, B and U bands.

1266: The theoretical predictions were corrected for the small Galactic extinction

1267: in the GRB direction (E(B - V)=0.016, Schlegel et al. 1998) but not for 

1268: extinction in the host

1269: galaxy, since there is no spectral evidence for significant extinction 

1270: there. In Fig.~(\ref{figr012301}) we  present the radio light curves at 

1271: 15 and 8.46 GHz. In Figs.~(\ref{rad-opt123a}) to 

1272: (\ref{rad-opt123b}) we make the complementary comparison of theory and

1273: observations for the radio to optical spectra,

1274: in four time intervals extending from 0.1 to 20 days. 

1275: The agreement between theory and observations is good despite 

1276: the limited available data on the radio AG and its    

1277: modulation by scintillations. 

1278: 

1279: \subsection{GRB 970508}

1280: 

1281: 

1282: The optical (and X-ray) AG of GRB 970508 is the only one so far that has

1283: been seen to rise and fall very significantly (e.g., 

1284: Garcia et al. 1998; Galama et al. 1998b;  Pedersen et al. 1998;

1285: Schaefer et al. 1997; Sokolov et al. 1998; Zharikov et al. 1998).  In DDD 

1286: 2001 we have shown

1287: that a CB model fit to this AG fails, if one assumes ---like in all our other

1288: fits---  a constant ISM density.  However, we have argued there that GRB

1289: progenitors are presumably located in super-bubbles of 0.1 to 0.5 kpc

1290: size. There may be instances in which the jet of CBs, after travelling for

1291: such a distance, does not continue onwards to a similarly low-density halo

1292: region, but encounters a higher-density domain. Indeed, we have shown that

1293: a fairly satisfactory fit to the optical (and X-ray) AG is obtained

1294: upon assuming an upwards jump in density by a factor

1295: $\sim 2.2$ at $\rm t\sim 1.1$ day after burst.

1296: This jump occurs before the first available

1297: data points on the radio AG (Galama et al. 1998a; Frail et al. 2000b).

1298: Therefore, we have fitted the optical data and the radio data 

1299: with the ISM density profile that was fitted

1300: to the R-band light curve. 

1301: 

1302: In Fig.~(\ref{opt508}) we present the comparison between

1303: the measured light curve  for the I,R,V and B  bands after

1304: subtraction of the host galaxy and SN contribution (DDD 2001). 

1305: The theoretical predictions were corrected for the small galactic extinction

1306: in the GRB direction (E(B - V)=0.016, Schlegel et al. 1998) but not for 

1307: extinction in the host

1308: galaxy, since there is no spectral evidence for significant extinction 

1309: there. In Figs.~(\ref{figr050801}) and (\ref{figr050802}) we also present 

1310: the radio light curves at 8.46, 4.86 and 1.43 GHz.

1311: In Figs.~(\ref{figr050802}) to (\ref{rad-opt508b}) we

1312: make the complementary comparison of theory and

1313: observations for the radio to optical spectra,

1314: in five time intervals

1315: extending from 0.12 to 470 days. The results are quite satisfactory.

1316: 

1317: \subsection{Commentary}

1318: 

1319: 

1320: In DDD 2001 we demonstrated that, in the CB model, the spectral index

1321: in the optical to X-ray domain could be extracted from the 

1322: time-dependence of the optical

1323: light curves. The fits resulted in $\rm \alpha=p/2\simeq 1.1$ for

1324: all GRBs of known redshift. This result is in good agreement

1325: with the observed late spectral observations.

1326: We have learned in this section that the CB model also provides an

1327: excellent description of the AG spectra in the broader band that

1328: includes the radio data. Only one new parameter, $\rm \nu_a$, is

1329: involved in the extension to the broader band. And this fitted 

1330: parameter and the injection bend ---at its {\it predicted} frequency and 

1331: time-dependent position---  bring about

1332: the agreement between the different magnitudes and spectral trends

1333: of the radio and optical domains.

1334:   

1335: In some of our fits to broad band spectra, such as the earliest data

1336: on GRBs 000301c,  991216 and 990123

1337: in the upper panels of Figs.~(\ref{rad-opt301}), (\ref{rad-opt216})

1338: and (\ref{rad-opt123a}), respectively, the theoretical curve is

1339: an underestimate of the low-energy spectral intensity.

1340: In other cases, such as GRBs 000926, 991208 and 980425,

1341: the spectral fits are excellent at all times. The lowest frequencies and

1342: earliest times are the most dependent on our simplifications

1343: concerning the GRB geometry, density profile, self-absorption,

1344: cumulation and limb-darkening. We would have been surprised

1345: if these simplifications worked even better than they do, and

1346: the fits do improve if we remove our  approximation

1347: of a fixed spectral index $\rm p=2.2$. But our aim in this paper

1348: is not to obtain spectacularly good fits, but to demonstrate that, 

1349: even in the simplest approximations, the CB model provides

1350: a good description of the broad-band data. The analysis of the

1351: lowest radio frequencies at the earliest times brings forth a

1352: plethora of details that are not of fundamental interest: our

1353: ultimate goal is not to understand these details, but

1354: to investigate what the origin of GRBs actually is.

1355:  

1356: 

1357: \section{SN1998bw and GRB 980425} 

1358:  

1359: The time and position of the peculiar gamma ray burst 980425 (Soffita et  

1360: al. 1998) coincided with supernova SN1999bw (Tinney et al. 1998) in the  

1361: spiral galaxy ESO 184-G82,  at a nearby $\rm z=0.0085$ (Tinney et al. 

1362: 1998; Sadler et al. 1998; Galama et al. 1998c; Lidman et al. 1998; 

1363: Iwamoto et al. 1998).  Iwamoto et al. (1998) estimated the 

1364: the core collapse of SN1998bw to have happened within 

1365: $-$2 to +7 days 

1366: of GRB 980425. The BeppoSAX Narrow Field Instrument (NFI) located 

1367: 10h after burst an X-ray source coincident in position with SN1999bw that  

1368: declined slowly with time between April and November 1998 (Pian et 

1369: al. 2000). A posteriori statistics indicate a very low chance probability  

1370: ($\leq 10^{-4}$) of a GRB being so nearly coincident in position.  But  

1371: despite how close  ---if it was associated with SN1999bw--- the progenitor  

1372: of GRB980425 was to us (38 Mpc for  $\rm H=65\, km\,Mpc^{-1}\, s^{-1}$),  

1373: the $\gamma$-ray fluence indicated only $7\times 

1374: 10^{48}$ erg equivalent spherical energy release in $\gamma$ rays, much  

1375: smaller than $\sim 3\times 10^{53}$ erg, the mean value for 

1376:  the score of other GRBs with known ---cosmological--- redshifts.

1377:  

1378: \subsection{SN1998bw: the accepted lore}

1379: 

1380: Like its accompanying GRB,

1381: SN1998bw was also claimed to be a very peculiar radio supernova 

1382: (e.g. Kulkarni et al. 1998). Over the past twenty years approximately two  

1383: dozen  SNe have been detected in the radio: 2 

1384: Type Ib, 5 Type Ic, and the rest Type II. A much larger list of 

1385: more than 100 additional SNe have low radio upper limits (for a 

1386: review see, e.g., Weiler et al. 2000 and references therein). Type Ib/c  

1387: SNe are fairly homogeneous in their radio properties, but 

1388: SN1998bw\footnote{SN1998bw was 

1389: classified initially as  Type Ib (Sadler et al. 1998), then 

1390: Type Ic (Patat and Piemonte 1998), then peculiar Type Ic (Kay et al. 1998),  

1391: then, at an age of 300 to 400 days, again as Type Ib (Patat et al. 

1392: 1998).} had a peak 6-cm radio luminosity of $\rm \sim 8\times 10^{28}\, erg  

1393: s^{-1} \, Hz^{-1}$, that is 20 to 40 times brighter than other 

1394: radio Type Ib/c SNe, which fall typically in the range $\rm 1.4-2.6\, \times  

1395: 10^{28}\, erg s^{-1} \, Hz^{-1}$. SN1998bw also reached a high radio 

1396: luminosity earlier than any known SN. Simple arguments based on the 

1397: brightness temperature of its radio luminosity (e.g., Readhead 1994) 

1398: required the radiosphere of SN1998bw to have expanded  surprisingly 

1399: fast, at $\rm\geq 200,000\, km\, s^{-1}$, at least during the first few days. 

1400: Its unusually high optical and radio luminosities and  its extraordinarily 

1401: large initial speed of expansion led many authors to conclude that 

1402: SN1999bw was a hypernova (Paczynski 1998) rather than a 

1403: peculiar supernova (e.g., Iwamoto et al. 1998). 

1404: 

1405: \subsection{The pair SN1998bw/GRB 980425 in the CB model}

1406:  

1407: In Dar and De R\'ujula (2000a) we argued that the only 

1408: peculiarity of SN1998bw was that it was viewed very near its axis. The  

1409: peculiarity of GRB 980425 was its nearness, that allowed for 

1410: its detection at an angle, $\theta\sim 8/\gamma(0)$  

1411: unusually large relative to the other GRBs of known redshift,

1412: for which $\theta\sim 1/\gamma_0$. 

1413: These facts conspired to produce a ``normal'' GRB fluence, and 

1414: resulted in an optical AG dominated by the SN. In DDD 2001,  

1415: we  demonstrated 

1416: that the X-ray AG of this GRB was also ``normal'': it has precisely the  

1417: light curve (in shape and normalization) expected in the CB model

1418: if the X rays are produced by the CBs and {\it not}, 

1419: as the observers assume (Pian et al. 2000), by the supernova.  

1420: 

1421: In the CB model (Dar and De R\'ujula 2000a), the 

1422: gamma-ray fluence of GRBs

1423: at large viewing angle ($\gamma_0^2\, \theta^2\gg 1$) is

1424: $\propto \delta_0^3\propto \theta^{-6}$. 

1425: The radio AG spectral energy density is proportional to

1426: $\gamma^{1/2}\, \delta^{7/2}$, as implied by 

1427: Eqs.~(\ref{sync}, \ref{Fnucb}, \ref{Fnuobser}), the 

1428: dependence $\rm\nu_b\propto \gamma^3$ and the

1429: relation $\rm\nu_{obs}\propto\nu_{_{CB}}\,\delta$.

1430: As a function of time, the AG peaks when $\gamma\, \theta\sim 1$, 

1431: so that $\gamma\simeq\delta\simeq 1/\theta$ and the

1432: peak value is proportional to $\theta^{-4}$.

1433: Because its proximity and large viewing angle ``conspired'' to make

1434: GRB 980425 appear ``normal'' in gamma rays, its peak radio intensity

1435: should have been enhanced by a factor $\rm \sim

1436: (\theta/mrad)^2$ relative to that of ordinary GRBs. Thus, for $\theta\sim 8.3$

1437: mrad, as estimated for GRB 980425 in Dar and De R\'ujula 2001, its expected

1438: peak radio intensity is $\sim$ 60 times larger than that of

1439: ordinary GRBs. Observationally, it is 50 to 100 times larger.

1440:  

1441: In Figs.~(\ref{figr042501}) to (\ref{figr042505}) 

1442: we show our CB model fits to the   

1443: temporal and spectral behaviour of the radio afterglow of GRB 980425.

1444: The fit parameters (in particular the large observation angle $\theta$)

1445: are quite close to the ones  

1446: that explain its GRB fluence (Dar and De R\'ujula 2000a), and its  

1447: X-ray afterglow (DDD 2001). These figures show  

1448: how, in the CB model, the radio AG of GRB 980425 also has a ``normal''  

1449: magnitude and shape.  

1450: That is, once more, {\it if} the radio AG is produced by the CBs and  

1451: {\it not} by the SN, unlike, once again, it is generally assumed (e.g., 

1452: Kulkarni et al. 1998; Li and Chevalier; Weiler et al. 2000). 

1453: 

1454: In the case of GRB 980425 the relatively large viewing angle

1455: and the subsequently small Doppler factor imply that, at

1456: late times, even the radio frequencies are above the injection bend.

1457: The large $\nu$ behaviour in Figs.~(\ref{figr042504}) and 

1458: (\ref{figr042504}) is $\rm\nu^{-p/2}\sim\nu^{-1.1}$. Also, the late time

1459: trend of the radio light curves in Figs.~(\ref{figr042501}) and

1460: (\ref{figr042502}) approaches the asymptotic 

1461: $\rm t^{2(p+1)/3}\sim t^{-2.1}$.

1462: 

1463: For GRB 980425 the

1464: radio data are overwhelmingly more abundant and precise than

1465: the X-ray data, and it is interesting to check what the prediction

1466: for the X-ray light curve is, if the input parameters are those

1467: determined in the radio fits. This is done in Fig.~(\ref{X425})

1468: for two values of the electron spectral index $\rm p$. For our

1469: fixed choice, $\rm p=2.2$, the prediction misses the data by a

1470: factor $\sim 20$. There are two excuses for that. First, since GRB

1471: 980425 is seen much more ``sideways'' than other GRBs, and its

1472: Doppler factor $\delta$ is much smaller than usual, the cumulation,

1473: illumination and limb-darkening factors play a bigger role than

1474: usual. These factors involve many simplifying assumptions (such as

1475: spherical symmetry) and significantly affect the normalization of

1476: the radio AG, but not that of the X-rays. Second, the extrapolation

1477: from radio to X-rays is over some 10 orders of magnitude in frequency,

1478: and a small change in the spectral photon's slope, $\rm (p - 1)/2$,

1479: entails a very large change in relative magnitude, as can be seen

1480: in Fig.~(\ref{X425}) by comparing the $\rm p=2.2$ and $\rm p=2$

1481: curves.  

1482: 

1483: The X-ray light-curve of GRB 980245 is essentially 

1484: flat in the time-interval of the first four observed points

1485: (Pian et al. 2000), while the corresponding data

1486: for all other GRBs fall with time much faster. 

1487: The last observational point in Fig.~(\ref{X425}), a preliminary result

1488: from XXM Newton (Pian 2002) and Chandra (Kouveliotou 2002),

1489: falls precisely in the expected  

1490: subsequent fast decline (predicted in DDD 2001) and

1491: definitely not in a naive power-law extrapolation.

1492: The peculiar light curve is a

1493: consequence of the large observing angle (Dar and De R\'ujula 2000a).

1494: For the reasons stated in this paragraph and the preceding one,

1495: we consider the prediction of the X-ray

1496: fluence completely satisfactory.

1497: 

1498: In DDD 2001, on the basis of the very meager X-ray data, we argued

1499: that the last optically-measured point of the SN1998bw/GRB 980425

1500: pair, at day 778 (Fynbo et al., 2000), was due to the CB's AG and

1501: not to the supernova. Redoing the analysis with the input of the

1502: abundant radio data, we must now revise this conclusion. In

1503: Fig.~(\ref{late425}) we show the result, with inclusion of the

1504: late optical measurement. This point lies more than two orders

1505: of magnitude above the predicted CB's AG: it must be due to the

1506: SN. We do not have an explanation ---specific to the CB-model---

1507: of the fact that this point also lies somewhat above the expectation

1508: based on $\rm ^{56}Co$ decay (Sollerman et al. 2000).

1509: 

1510: We are claiming that long duration GRBs are associated with a good

1511: fraction of core-collapse SNe.  Yet, SN1998bw was one of the

1512: brightest in its class. The apparent contradiction may be dispelled

1513: by the increasing evidence that SN explosions are fairly asymmetric.

1514: It is quite conceivable that, viewed very close to their ``CB axis''

1515: SNe  appear to be brighter than when observed from other directions.

1516: 

1517: 

1518: The conclusion is twofold. GRB 980425 is, {\it in every respect},

1519: normal ($\rm z$ and $\theta$ being chance variables). And, deprived

1520: of very abnormal X-ray and radio outputs ---which are not due to

1521: the supernova, but to its ancillary GRB--- SN1998bw loses most of

1522: its ``peculiarity''.

1523:  

1524: 

1525: \subsection{Superluminal motion in SN1998bw/GRB 980425}

1526: 

1527: The transverse projected velocity in the sky of a CB relative

1528: to its parent SN is, for large $\gamma$ and small $\theta$:

1529: \begin{equation}

1530: \rm

1531: V_{_{CB}}(t)\simeq {\gamma(t)\, \delta(t)\, \theta\over (1+z)}\; c\, ,

1532: \label{supervelocity}

1533: \end{equation}

1534: which, for typical parameters, is extremely superluminal.

1535: The resulting angular separation at time $\rm t$ is:

1536: \begin{equation}

1537: \rm

1538: \Delta \alpha(t)={1\over D_A}\int_0^t\,V_{_{CB}}(t')\;dt'\; ,

1539: \label{Deltaalpha}

1540: \end{equation}

1541: where $\rm D_A=(1+z)^2\, D_L$ is the angular distance

1542: to the SN/CB system. In Fig.~(\ref{superluminal}) we show

1543: $\rm \Delta(t)$ for SN1998bw/GRB 980425 with our 

1544: parameters fit to the corresponding radio data (for our

1545: adopted $\rm H_0$, $\rm D_L\simeq D_A\simeq 38$ Mpc). In 

1546: Dar and De R\'ujula (2000a) we argued that this separation

1547: was sufficient to justify a dedicated effort to search for

1548: a ``binary'' source. It is interesting to discuss what the

1549: situation is with the data currently available.

1550: 

1551: The most accurate determination of the position in the sky of the

1552: SN1998bw/GRB 980425 system is based on the radio observations made

1553: with the Australian Telescope Compact Array (ATCA, Wieringa et al.

1554: 1998). Recall that in the CB model the radio coordinates are those

1555: of the CB (GRB 980425 was a single-pulse GRB, that is, it had a

1556: single dominant CB). In days 3, 4 and 10 the source is reported to

1557: be at (RA 19:35:03.31, Dec -52:50:44.7).  In the subsequent 33

1558: observations, ranging from day 12 to day 790, the  position is (RA

1559: 19:35:03.32, Dec -52:50:44.8), some $0.^{\!\!''}18$

1560: away from the original determination,

1561: but not inconsistent with the observational uncertainty of

1562: $0.^{\!\!''}1$. In the penultimate observation at day

1563: 320 the source has faded to the point that it is not observable in

1564: 2 out of 6 frequencies, and in the last observation at date 790

1565: there is no clear sighting at any frequency.  

1566: The predicted values of $\Delta \alpha$

1567: from Eq.~(\ref{supervelocity}) at some relevant dates are 12, 158,

1568: 183 and 292 mas at days 12,  249, 320 and 790, respectively. These

1569: results, the observational error, and the fact that the ATCA

1570: observers were not trying to follow the source's motion imply that

1571: their results are insufficient to claim either that the early change

1572: of position was significant, or that a motion of the CB comparable

1573: to the predicted one is excluded.

1574: 

1575: Observations of the vicinity of the source of GRB 980425 were made

1576: with the Hubble Space Telescope (HST) at day 778 (Fynbo et al.,

1577: 2000), with a tiny astrometric uncertainty of $0.^{\!\!''}018$, and

1578: pointing at ATCA's first reported coordinates. The observations

1579: are compatible with SN1998bw lying at that point, and reveal six

1580: other objects in a ($1.^{\!\!''}0\times 1.^{\!\!''}0$) field centered

1581: there. As a result of our CB model fit to the radio data, as we have

1582: explained, we expect the optical observations to correspond to

1583: SN1998bw, and there it is, at the field's center. We also expect, 

1584: as in Fig.~(\ref{late425}), the CB to be more than two orders of

1585: magnitude fainter:  not observable. It would be nice if this

1586: conclusion was wrong, that is, if the large ``naive'' extrapolation

1587: from radio to optical frequencies in Fig.~(\ref{late425}) was an

1588: underestimate by a considerable factor, which is the case for the

1589: larger extrapolation from optical to X-ray frequencies in

1590: Fig.~(\ref{X425}) (the ``naive'' prediction there is the one labelled

1591: $\rm p=2.2$). In that case, it may be that a subsequent observation

1592: of the same field reveals that one of the closer-by extra sources

1593: has faded away!  Three of these sources are $\sim 0.^{\!\!''}5$

1594: away from the SN, if one of them is the CB, and it is dimming, we

1595: would not excessively mind that this is $\sim 60$\% more distant

1596: than the prediction in Fig.~(\ref{superluminal}), based on a

1597: constant-density approximation for the ISM.

1598: 

1599: \section{The normalization of afterglows}

1600: 

1601: The values of the CB model parameters that were fitted to the broad band 

1602: data on GRBs with known redshift ---and are listed in Table III--- are

1603: narrowly distributed around their reference values, except 

1604: for the overall normalization which is much smaller than originally 

1605: anticipated if  $\eta=1$.  

1606: This normalization ``problem'' may point to inaccuracies in the

1607: various hypothesis that we have made. One example 

1608: is the contention that {\it all} of the incident energy of the ISM electrons 

1609: is radiated in the CB. It may well be 

1610: that the moving CB deflects and scatters the ISM  electrons

1611: before they radiate a large fraction of their acquired energy,

1612: as suggested by the results of the numerical simulations of Achterberg et 

1613: al. (2001).   

1614: But abandoning some of our simplifications would be premature. Indeed,

1615: in concluding that in the CB model the normalization of the AGs is more than

1616: one order of magnitude too large, we have used our reference parameters

1617: to compute the expected values. And there is sufficient elasticity

1618: in these  parameters to obtain a consistent overall picture of all GRB 

1619: properties, as we proceed to review and discuss.

1620:  

1621: \section{The parameters of the CB model}

1622: 

1623: With the current analysis of GRB radio AGs we have completed a 

1624: first round of the study of GRB properties in the CB model, and it

1625:  behooves us

1626: to look back at the various constraints on the relevant parameters.

1627: 

1628: In Dar and De R\'ujula (2000a) we followed Dar (1998) and Dar and

1629: Plaga (1999) in suggesting that the large peculiar velocities of

1630: neutron stars may be due to a ``natal kick'', induced by a momentum

1631: imbalance in the oppositely-directed jets of CBs accompanying

1632: their birth. On this basis we chose as a reference value

1633:  $\rm E_{CB}\sim 10^{52}$ erg, for a jet with a reference

1634: number of CBs (or prominent peaks in the GRB light-curve)

1635: $\rm n_{_{CB}}\sim 10$. Based on a first

1636: analysis of AG properties, and of GRB $\gamma$-ray fluences and 

1637: individual $\gamma$-ray energies, we set $\gamma_0=10^3$

1638: as a reference value. 

1639: 

1640: In Dar and De R\'ujula (2000b) we investigated two extreme

1641: models meant to bracket the behaviour of a CB as it

1642: crosses a SN shell, is heated by the collision with

1643: its constituents, and emits observable $\gamma$-rays as

1644: it reaches the shell's transparent outskirts with a 

1645: radius $\rm R_{_{CB}}^{tr}$, proportional to its early

1646: transverse expansion velocity $\rm \beta_{trans}\,c$,

1647: which we assumed to be close to the sound speed in

1648: a relativistic plasma, $\rm\beta_{trans}=1/\sqrt{3}$. In our

1649: ``surface'' model, which is no doubt closer to a realistic

1650: description, the energy of the GRB in $\gamma$-rays

1651: is proportional to $\rm n_{_{CB}}\, [R_{CB}^{tr}]^2 \,\gamma_0$.

1652: (Eq.~(45) of Dar and De R\'ujula 2000b).

1653: For the chosen reference parameters, in the surface model,

1654: this prediction overestimates the GRB fluences by about

1655: one order of magnitude. Since the individual $\gamma$-ray

1656: energies corroborate the choice $\gamma_0\sim 10^3$, this

1657: means that $\rm [R_{_{CB}}^{tr}]^2$ (and  $\rm\beta_{trans}$)

1658: are overestimated by roughly one order of 

1659: magnitude\footnote{The reduction of thermal

1660: free-electron surface-density discussed in Sections 6.2 and 6.3

1661: is logically independent from the modifications discussed

1662: here (such as a {\it decrease} in radius) which have to do

1663: with the fluence produced by the small fraction of energetic

1664: radiating electrons.}.

1665: 

1666: In Dar and De R\'ujula (2001) we analyzed the X-ray ``Fe'' lines

1667: observed in the AGs of some GRBs, which we attributed to

1668: hydrogen recombination in the CBs, with the corresponding 

1669: Lyman-$\alpha$ lines 

1670: boosted by a large factor $\delta/(1+z)\sim 500$. We equated

1671: the total number of photons in the lines to the baryon number

1672: of the jet of CBs, and found agreement with the baryon

1673: number in the jet, $\rm n_{_{CB}}\,N_{CB}$, to within one

1674: order of magnitude. But in the current investigation, 

1675: we have found that the absorption of radio waves keeps the

1676: CBs hot and ionized (Appendices III and V). This means

1677: that our reference value of $\rm N_{CB}$ is likely

1678: to be an overestimate.

1679: 

1680: In DDD 2001 we proposed a mechanism that would quench the

1681: expansion of a CB in minutes of observer's time, well

1682: after it has exited the SN shell. The CBs reach an asymptotic

1683: radius (Eq.~(16) of DDD 2001):

1684: \begin{equation}

1685: \rm R_{max}^3\simeq 

1686: {3\,N_{CB}\,\beta_{trans}^2\over 2\,\pi\,n_p\,\gamma_0^2}\; .

1687: \label{Rmax}

1688: \end{equation}

1689: On the basis of this calculated radius (for $\rm\beta_{trans}=

1690: 1/(3\sqrt{3})$, we found that the normalization of optical

1691: and X-ray AGs agreed with the reference-value expectations.

1692: On the same basis, we find now that the normalization

1693: is overestimated by an order of magnitude. The reason for

1694: the discrepancy is that, in DDD 2001, we effectively placed

1695: the spectral discontinuity at a ``cooling break'' frequency

1696: corresponding to an electron Lorentz factor $\rm \gamma_e\sim 1$,

1697: while we have now argued that the discontinuity should occur

1698: at a higher value $\rm \gamma_e=\gamma(t)$.

1699: 

1700: Both the GRB fluence and the AG fluence are, in the CB-model,

1701: $\rm F\propto n_{_{CB}}\,R^2$, with $\rm R=R_{_{CB}}^{tr}$ for the

1702: $\gamma$ rays and $\rm R=R_{max}$ for the AG. At a value of

1703: $\rm x_\infty$ fixed by the fit to the AG's temporal behaviour, the AG 

1704: fluence is:

1705: \begin{equation}

1706: \rm

1707: F_{AG}\propto 

1708: n_{_{CB}}\,n_e\,R_{max}^2=

1709: n_{_{CB}}\,N_{CB}\,{1\over \pi\, x_\infty}.

1710: \label{FAG}

1711: \end{equation}

1712: 

1713: All of the above ``problems'' are solved if we reduce our

1714: ``typical'' values of $\rm R_{max}^2$ and $\rm N_{CB}$

1715: by about one order of magnitude, relative to our reference

1716: parameters with, according to Eq.~(\ref{Rmax}), the corresponding 

1717: reduction of the choice of $\rm\beta_{trans}$ by half an order of 

1718: magnitude.

1719: 

1720: The precise location of the injection bend is

1721: not predictable and a modification by up to

1722: one order of magnitude of its position has

1723: a small effect on the quality of the fits

1724: to observations. An increase of the cooling

1725: break frequency $\rm \nu_b$ implies a

1726: corresponding decrease in AG flux, see Eq. (4),

1727: adding to the uncertainty in the prediction

1728: of the precise overall normalization.

1729: 

1730: To summarize, the CB model correctly describes, in terms of a very 

1731: limited set of parameters, the properties of GRBs and their

1732: AGs, including their normalizations. This is the case even if

1733: we adhere to all of the detailed assumptions we have made,

1734: even though they are approximations to a no doubt fairly

1735: convoluted physical problem. 

1736: 

1737: \section{Hyperluminal CBs and radio scintillations} 

1738:  

1739: The radio AGs of GRBs often show temporal variations of a factor of 

1740: two or more on a time scale of hours at early times and on a time scale of  

1741: days at later times; e.g.

1742: GRB 000926 in Figs.~(\ref{figr092602}, \ref{figr092603}), 

1743: GRB 991216 in Fig.~(\ref{figr121602}),

1744: GRB 991208  in Figs.~(\ref{figr120803}, \ref{figr120804}), 

1745: GRB 000418 in Fig.~(\ref{figr041802}),

1746: GRB 990123  in Fig.~(\ref{figr041801}) and

1747: GRB 970508 in Figs.~(\ref{figr050801}, \ref{figr050802}). 

1748: 

1749: Similar variations have not been seen in the optical and X-ray bands. 

1750: The intensity variations of GRB radio AGs are very reminiscent of 

1751: the ones seen in radio signals from pulsars  in 

1752: our galaxy, interpreted as scintillations due to the 

1753: motion of the line of sight through the refractive, diffractive and   

1754: dispersive ISM of the Galaxy (see, e.g., Lyne and Smith 1982).   

1755: Some very compact  active galactic nuclei also show an intraday 

1756: variability that has been the subject of much debate (e.g., Wagner and  

1757: Witzel 1995 and references therein). At least in one case ---the   

1758: variations in 

1759: the radio intensity of the quasar J1819+3845, the most extremely variable  

1760: AGN known at radio wavelengths (Dennett-Thorpe and de Bruyn 2000)--- it  

1761: was shown unambiguously that the variations are scintillations caused by the  

1762: ISM (Dennet-Thorpe and Bruyn 2002). 

1763:  

1764: The (de)coherence properties in time and frequency 

1765: of the radio scintillations have been used to measure the 

1766: transverse speed of pulsars (e.g., Lyne and Smith 1982). 

1767: Gupta (1995) has demonstrated for a sample of 59 pulsars that their 

1768: transverse speed, $\rm V_{iss}$, measured from their inter-stellar 

1769: scintillations, agrees well with their transverse speed, $\rm V_{pm}$,  

1770: measured from their proper motion (see also Nicastro et al. 2001). 

1771:  

1772: The movement of the line of sight to pulsars is in most cases dominated 

1773: by their proper motion at a transverse velocity $\rm V_{pm}$ 

1774: larger than the turbulent speeds in the ISM, or of the 

1775: sun relative to the ISM, or of the Earth around the sun. 

1776: The mean $\rm V_{pm}$ of Gupta's 59 pulsars is 311 $\rm km\, s^{-1}$ 

1777: and their mean distance is estimated to be 1.96 kpc. 

1778: Their angular speeds are within an order of magnitude of a central value:   

1779: \begin{equation} 

1780: \rm \dot \omega_{ps}\sim 

1781: {\langle V_{pm}\rangle \over \langle D\rangle} 

1782: \simeq 5.1\times 10^{-15}\, rad\, s^{-1}\, . 

1783: \label{vpang} 

1784: \end{equation} 

1785:  

1786: Travelling with a characteristic $\gamma\sim 10^3$ and viewed at typical 

1787: angles $\theta$ of milliradians, CBs have apparent superluminal velocities,

1788: $\rm V_{_{CB}}$ of Eq.~(\ref{supervelocity}), that are 

1789: so high (a few hundred times the speed of light) that they deserve 

1790: to be called {\it hyperluminal}. 

1791: The angular speed in the sky is: 

1792: \begin{equation} 

1793: \rm 

1794: \dot\omega_{_{CB}}(t)= 

1795: {V_{_{CB}}(t)\over D_A}\simeq {\gamma(t)\, \delta(t)\, c\, \theta\over   

1796: (1+z)\, D_A}\, , 

1797: \label{vcbang} 

1798: \end{equation} 

1799: For the reference values $\theta\sim 1$ mrad 

1800: and an initial $\gamma_0\sim\delta_0\sim 10^3$, 

1801: the initial angular speed of a CB at redshift $\rm z=1$ is 

1802: $\rm\dot\omega_{_{CB}}(0)\sim 2.7\times 10^{-15}$ rad s$^{-1}$, 

1803: in the very same range as that of 

1804: Galactic pulsars. 

1805: The CBs' angular velocity $\rm \dot\omega_{_{CB}}(t)$ 

1806: and the resulting (inverse) coherence time of the 

1807: scintillations should decline as  $\rm\gamma(t)\,\delta(t)$ does. 

1808: Both pulsars and CBs are pointlike from the point of view 

1809: of their radio scintillations. Thus, all conditions are met to expect 

1810: pulsar-like scintillations in the radio signals from CBs. 

1811: 

1812: 

1813: The deviations from a smooth behaviour of the radio signals in the case of

1814: GRB 980425, as can be seen in Figs. (\ref{figr042501}, \ref{figr042502}),

1815: are chromatic, but correlated in time over a much longer period than for

1816: the other GRBs. Because GRB 980425 is so close ($\rm z=0.0085$) and is

1817: viewed at the unusually large angle of $\sim 8$ mrad (Table III and DDD

1818: 2001), its apparent angular velocity, Eq.(\ref{vcbang}), is much larger

1819: than for other GRBs. The line of sight to this GRB swept a much bigger

1820: region of galactic ISM than for other GRBs or, for that matter, pulsars.

1821: Thus, we have no independent information on the ISM irregularities causing

1822: scintillations on this large scale. 

1823: 

1824: The analysis of CB scintillations could result in a measurement of their  

1825: hyperluminal speeds and a decisive test of the cannonball model 

1826: (fireballs do not have relativistic proper motions, firecones stop moving

1827: close to their progenitors and, unlike CBs, have an increasing size that

1828: should rapidly quench their scintillations with time). 

1829: The presently available information (or the current information 

1830: in its published form) is insufficient for us to attempt 

1831: at the moment to extract conclusions about CB hyperluminal velocities 

1832: from the observed scintillation patterns in the radio AG of GRBs. 

1833: 

1834: 

1835: \section{Conclusions} 

1836: 

1837: The Cannonball model gives an excellent and 

1838: extremely simple description of all 

1839: measured properties of GRB afterglows, 

1840: including their radio afterglows. 

1841:  

1842: 

1843: In the CB model, there is an injection bend in the spectrum,

1844: at the predicted time-dependent frequency $\rm \nu_b$ of Eq.~(\ref{nubend}). 

1845: We have shown that the evidence for the correctness of this

1846: prediction is very strong, see Table II and

1847: Figs.~(\ref{index2}) and (\ref{index}).

1848: It is this spectral bend that governs the relative normalization

1849: of the radio and the optical AG, again in agreement with observation,

1850: as shown in all our figures of AG wide-band spectra.

1851: 

1852: Since we have always set the electron index to its theoretical

1853: value, $\rm p=2.2$, just {\it three intrinsic parameters}

1854: are needed to describe an optical or X-ray AG: 

1855: $\gamma_0$, $\rm x_\infty$ and the normalization; the viewing

1856: angle $\theta$, although it must also be fit, is external to the GRB, like

1857: the redshift and the absorption in the host and in the Galaxy are.

1858: We have shown that, in the CB model, the extension of these results

1859: to the radio domain

1860: requires the introduction of {\it just one extra parameter:} the 

1861: free-free absorption frequency $\rm \nu_a$ of Eq.~(\ref{absfreq}),

1862: and that, in spite of various approximations,

1863: this simplest of descriptions is at the moment entirely satisfactory. 

1864: Notice that what one has to parametrize is a two

1865: dimensional surface: the fluence as a function of frequency and

1866: time. The shape of this surface is that of a relatively simple

1867: ``mountain'', various cuts of which at fixed $\rm t$ or $\nu$

1868: are shown if Fig.~(\ref{figCBpreds}). It would be easy, and it may well

1869: be misleading, to overparametrize this rather

1870: featureless surface with more than a few parameters. 

1871: 

1872: It is instructive to compare, or so Occam would have 

1873: thought, the understanding of wide-band AG spectra in the CB model 

1874: with that in the fireball or firetrumpet models. 

1875: In the latter, the number of intrinsic parameters varies: 

1876: seven [e.g. Berger et al. 2001d], eight [e.g. Yost et al. 2001]

1877: nine [e.g. Yost et al. 2002] and even thirteen [e.g. Galama et al. 2000].

1878: This counting does not include the viewing angle, since

1879: the firetrumpets in these works point precisely at the 

1880: observer\footnote{This serious limitation (DDD 2001) is beginning to

1881: be remedied in the firecone literature, in which the CB-model's

1882: geometry is being ---though with no reference--- ``standardized''

1883: and its consequences (Dar and De R\'ujula 2000a) explained again;

1884: see, e.g.  Rossi, et al. (2001),  

1885: Zhang and Meszaros (2001),

1886: Salmonson and Galama (2001),

1887: Granot J. et al., (2002), Panaitescu and Kumar (2001).}.

1888: Moreover, even before the ``break'' in the time-evolution

1889: ---a period during which it is not inconsistent to use the

1890: quasi-spherical self-similar  approximation of Blandford and McKee

1891: (1976) for the expanding material---  the ordering of the ``breaks''

1892: in frequency implies a multiple choice of spectral shapes

1893: and of their evolution (Granot and Sari 2002).

1894: 

1895: Countrary to established custom, we are not presenting the $\chi^2$

1896: values of our fits, which are generally reasonable and would become

1897: quite good if, again following the consuetudinary path, we artificially

1898: increased the errors to compensate for scintillations in the radio

1899: data and/or uncertainties in attenuation.  The reason is that the

1900: CB model is a very simplified description of a no doubt very

1901: complicated reality (e.g. CBs could be somewhat comet-like, as

1902: opposed to spherical, their inner distributions of density,

1903: ionization, magnetic field and temperature could be non-trivial,

1904: even chaotic, etc, etc).  Even when the physics is much simpler

1905: than in the analysis of radio emissions, and the fits are very good

1906: ---as is the case in our description of optical and X-ray AGs in

1907: DDD 2001--- we do not report their quite impressive $\chi^2$

1908: values\footnote{A ``$\chi$-by-eye'' of the figures reporting here

1909: the optical AG light curves should suffice to prove this

1910: statement.}.  We view our ``fits'' as rough descriptions,

1911: rather than true fits.  Under such circumstances, the overintrepretation

1912: of a $\chi^2$ test has every chance of being  misleading, much more

1913: so in models containing many more parameters than the CB model.

1914: 

1915: For the same reasons, 

1916: and because of the systematic errors in the data, 

1917: the values of the parameters we extract from our fits should not 

1918: be taken entirely at face value, even though the minimization procedure 

1919: ---which attributes to the errors a counterfactual purely statistical origin--- 

1920: results in tiny 1 $\sigma$ spreads for the fitted parameters, and in 

1921: $\chi^2 $ values that are in most cases  satisfactory.

1922: 

1923: In the radio domain, as in every other aspect, the pair SN1998bw/GRB

1924: 980425 is particularly fascinating. On the basis of this GRB's

1925: observed fluence and distance, and given the (totally trivial but

1926: all important) dependence of the fluence on observation angle, we

1927: claimed in Dar and De R\'ujula (2000a) that the only peculiarity

1928: of this pair was that it was observed uncharacteristically far from

1929: its axis (for a GRB) and uncharacteristically close to it (for a

1930: SN). In DDD 2001 we proved that the X-ray AG of GRB

1931: 980245 was also what is expected in the CB model, depriving the

1932: supernova of its X-ray peculiarity: it did not make the observed

1933: X-rays.  In this paper, by understanding the magnitude, time- and

1934: frequency-dependence of the pair's radio signals ---which were not

1935: emitted, either, by the SN--- we have demonstrated that SN1998bw

1936: was also ``radio normal''. Neither this GRB,  nor its SN have ---in

1937: the CB model--- anything in particular, except the chance occurrences

1938: of the distance and observation angle.  Alas, the unique occasion

1939: to make a fundamental discovery by actually resolving the SN and

1940: the CB, as proposed in Dar and De R\'ujula (2000a), may now be

1941: very difficult, but, as we have explained, not entirely out of the 

1942: question.

1943: 

1944: 

1945: By pure coincidence, the apparent angular velocities of galactic

1946: pulsars and cosmological cannonballs are of the same order of

1947: magnitude.  The analysis of radio scintillations, one of the methods

1948: used to measure pulsars velocities, should also be applicable to

1949: the GRB ejecta.  Thus, it ought to be possible to test the CB-model's

1950: prediction of  hyperluminal cannonball velocities.

1951: 

1952: 

1953: 

1954: \section*{Appendix I: The slowdown of a CB} 

1955: 

1956: We review the functional form of 

1957: the time dependent Lorentz factor $\rm \gamma(t)$, which

1958: is explicit and analytical in a fair approximation (DDD 2001).

1959: 

1960: In minutes of observer's time, CBs reach a roughly constant radius 

1961: $\rm R_{max}$ and

1962: are parsecs away from their progenitor star,

1963: a domain where a constant-density ISM may be a reasonable approximation.

1964: Relativistic energy-momentum conservation in the  progenitor's      

1965: rest frame results in the equation governing the deceleration

1966: of a CB in the ISM:

1967: \begin{equation}   

1968: \rm  M_{_{CB}}\, d\gamma =-\, m_p\,c^2\, n_p\, \pi\, R_{max}^2\,

1969:                            \gamma^2\, c\, dt_{SN}\, .

1970: \label{conservation}

1971: \end{equation}

1972: Interestingly, the above expression is correct both if the incoming

1973: ISM protons are isotropically reemitted in the CB rest frame,

1974: or if they are ingurgitated by the CB (in the first case, they are

1975: reemitted with average energy $\rm m_p\,c^2\,\gamma^2$ in the

1976: progenitor's frame, in the second, the change in $\gamma$ per

1977: added proton is $\rm d\gamma=-[m_p/M_{_{CB}}]\gamma$).

1978: 

1979: 

1980: Use the relation $\rm dt_{SN}=\gamma(t)\, dt$ between the

1981: times  measured in the supernova and CB rest frames,

1982: divide both sides of the Eq.~(\ref{conservation}) by 

1983: $\rm M_{_{CB}}\, \gamma^3 $ and 

1984: integrate to obtain the relation:

1985: \begin{equation}

1986: \rm {1\over \gamma^2(t)} - {1\over \gamma_0^2}\simeq {2\, c\, t\over

1987: x_\infty}\, ,

1988: \label{gammacb}

1989: \end{equation} 

1990: where $\rm t$ is CB time, and:

1991: \begin{equation} 

1992: \rm 

1993: x_\infty\equiv{N_{CB}\over\pi\, R_{max}^2\, n_p}  \, , 

1994: \label{range} 

1995: \end{equation} 

1996: with $\rm N_{CB}\approx M_{_{CB}}/m_p$ the CB's baryon number. 

1997: 

1998: It is important to know the number of electrons  accumulated

1999: by a CB as its Lorentz factor decreases from $\gamma_0$ to

2000: $\rm\gamma(t)$ (in the

2001: approximation $\rm n_e=n_p$ of a Hydrogenic ISM

2002: this number equals that of scattered or incorporated protons). 

2003: The number rate of accumulation is related to the energy-loss rate of

2004: Eq.~(\ref{conservation}) so that:

2005: \begin{equation}

2006: \rm dN= - \eta\, {M_{_{CB}}\, d\gamma\over m_p\,c^2\,\gamma^2}=

2007: -\eta\, N_{_{CB}}\,{d\gamma\over \gamma^2}\, .

2008: \label{accrate}

2009: \end{equation}

2010: Assuming constant $\eta$, the total number of ISM electrons 

2011: accumulated at a CB time $\rm t$  is then:

2012: \begin{equation}\rm

2013: N(t)=\eta\, N_{_{CB}}\,\left[{1\over \gamma(t)}-{1\over 

2014: \gamma_0}\right]\, . 

2015: \label{acctot}

2016: \end{equation}

2017:  

2018: The time-dependence of $\rm \gamma(t)$ with $\rm t$ the observer's

2019: time is more complicated than Eq.~(\ref{gammacb}): 

2020:  the relation between the two times

2021: ($\rm dt_{obs}=dt_{_{CB}}(1+z)/\delta$) introduces a $\gamma$ (or

2022: $\rm t$) dependence via $\delta$. The result for $\rm\gamma(t_{obs})$

2023: is (DDD 2001): 

2024: \begin{eqnarray} 

2025: \rm \gamma&=&\rm\gamma(\gamma_0,\theta,x_\infty;t) 

2026: =\rm {1\over B} \,\left[\theta^2+C\,\theta^4+{1\over C}\right]\nonumber\\   

2027: \rm C&\equiv&\rm 

2028: \left[{2\over B^2+2\,\theta^6+B\,\sqrt{B^2+4\,\theta^6}}\right]^{1/3}   

2029: \nonumber\\ 

2030: \rm B&\equiv&\rm 

2031: {1\over \gamma_0^3}+{3\,\theta^2\over\gamma_0}+ 

2032: {6\,c\, t\over  (1+z)\, x_\infty}\, . 

2033: \label{cubic} 

2034: \end{eqnarray} 

2035: The Lorentz factor of the CB decreases from $\gamma_0$ 

2036: to $\gamma_0/2$ as the CB travels a distance 

2037: $\rm x_\infty/\gamma_0$, whose

2038: reference value is 1.3 kpc, as in Table I. 

2039: 

2040: \section*{Appendix II: The synchrotron cooling break}

2041: 

2042: We argue that the conventional synchrotron spectral break occurs at a very

2043: non-relativistic electron energy. The corresponding break in the

2044: radio spectrum is unobservable.

2045: 

2046: In writing the electron energy-loss rate Eq.~(\ref{Eloss}), we have

2047: assumed that synchrotron radiation, which is quadratic in energy,

2048: dominates (inverse Compton scattering has the same energy dependence,

2049: but it is negligible, since

2050: the magnetic energy density within a CB is much higher than 

2051: the radiation energy density). The general result for the energy loss of

2052: high-energy electrons is of the form:

2053: \begin{equation}

2054: \rm -{dE\over dt}\simeq A_C\,(ln {E\over m_e\,c^2}+a)+A_B\,E+A_S\, E^2\, .

2055: \label{b1}

2056: \end{equation}

2057: The term proportional to $\rm A_C$ describes Coulomb

2058: scattering and ionization losses in the CB, which are negligible at high

2059: energies. The second term represents bremsstrahlung and its

2060: coefficient is (e.g. Shu 1991 )

2061: \begin{equation}

2062: \rm A_B={34.35\,\alpha\, \sigma_T\, c\,\bar{n}_b \over 2\, \pi}\, ,

2063: \label{A31}

2064: \end{equation}

2065: where $\alpha=1/137$ and $\rm \bar n_b$ is the baryon number density in the CB.

2066: Adiabatic losses and electron escape would have the same energy dependence

2067: as bremsstrahlung, but in the AG regime we are discussing the CBs are no longer

2068: expanding and the electron's Larmor radii are so small ---relative to the CB's

2069: radius--- that escape losses should also be negligible.

2070: 

2071: 

2072: The spectral index of high energy electrons injected with a power-law 

2073: spectrum steepens by one unit at a ``cooling break'' energy 

2074: $\rm [\beta(E_c)]^2\,E_c=A_B/A_S$.

2075: For  $\gamma \simeq 10^3$ and the reference values of

2076: $\rm \bar n_b=\bar n_e$ and $\rm n_e$,

2077: the synchrotron cooling break is at a subrelativistic energy

2078: ($\beta\sim 0.8$). This is in

2079: contrast with the injection bend at the highly relativistic energy

2080: $\rm E_b\simeq \gamma\, m_e\, c^2\,.$

2081: The synchrotron radiation of electrons below the cooling break

2082: is, for the current data, at unobservably low observer's radio frequencies.

2083: 

2084: \section*{Appendix III: Photoionization and recombination in the CB}

2085: 

2086: We argue that the synchrotron radiation in a CB is intense 

2087: enough to maintain its plasma partially ionized with ion and 

2088: free electron densities proportional to $\rm \gamma(t)$.

2089: 

2090: 

2091: The bound-free cross section for photoionization of atomic hydrogen in its

2092: $\rm n$-th excited state by photons with frequency above the ionization 

2093: threshold,

2094: $\rm \nu_n=3.29\times 10^{15}/n^2\, Hz$, is given by 

2095: $\rm \sigma_\nu(n)=n\,\sigma_1\,\bar{g}_n (\nu/\nu_n)^{-3}$, with

2096: $\rm \sigma_1= {64\, \alpha\, \pi\, a_0^2/(3\, \sqrt{3}})$ $\simeq

2097: 7.91\times 10^{-18}$ cm$^2$ ($\rm a_0=0.53\times 10^{-8}$ cm 

2098: is the Bohr radius

2099: and $\rm \bar{g}_n$ is the Gaunt factor 

2100: for photoabsorption by hydrogen).

2101: For the surface flux of photons of Eq.~(\ref{Fnucb}), we obtain 

2102: an ionization rate of the $\rm n$-th level of atomic hydrogen:

2103: \begin{equation} 

2104: \rm R_i(n)=  {\eta\, n_e\, m_e\, c^3 \gamma(t)^2\, n\,  \sigma_1\, 

2105:              (p-2)\over 8\, (p+6)\,(p-1)\, h\, \nu_n}\, 

2106:               \left({\nu_b\over\nu_n}\right)^{(p-2)/ 2}\, .    

2107: \label{ionizationrate1}

2108: \end{equation} 

2109: For the reference values of $\rm n_e$, $\rm p$ and $\gamma=\gamma_0$,

2110: the ionization rate of  

2111: the ground state is $\rm R_i(1)\sim 1.1\times 10^{-2}\,\eta\, s^{-1}$.

2112: 

2113: The recombination rate per unit volume

2114: of hydrogen in an hydrogenic CB is (Osterbrock

2115: 1989):  

2116: \begin{equation} 

2117: \rm R_{rec}\simeq 1.0\times 10^{-12}\,\bar n_e\,

2118:                   \left[{T\over 10^4\,K}\right]^{-0.7}\,s^{-1}\, . 

2119: \label{Recombination}

2120: \end{equation} 

2121: For the reference value of the CB parameters, and T=10,000

2122: K the recombination rate of hydrogen is $\rm R_{rec}\sim 1.1\times

2123: 10^{-5}\, s^{-1}$. In quasi equilibrium, the ionization rate per unit

2124: volume, which is proportional to the number density of recombined hydrogen

2125: atoms, must be equal to the recombination rate per unit volume, which is

2126: proportional to the product of the ion and free electron densities in the

2127: CB. Thus, initially the CB is highly ionized.  But, for small values of

2128: $\eta$ and later times when $\rm \gamma(t)$ becomes sufficiently small,

2129: equilibrium between the ionization and recombination rates in the partially

2130: ionized hydrogenic plasma results in $\rm \bar n_i=\bar n_e\propto

2131: \gamma(t)\, .$

2132: 

2133: \section*{Appendix IV: Synchrotron and Bremsstrahlung Self Absorption}

2134: 

2135: We argue that the self-attenuation in the CB of the observed radio waves is 

2136: dominated by free-free absorption.

2137: 

2138: 

2139: \subsection*{The density of energetic electrons }

2140: 

2141: In the rest system of a CB, the ISM electrons arrive at a rate

2142: $\rm d\,N_e/dt\simeq n_p\,\gamma\,  \pi\, R_{max}^2\, c\, .$

2143: A fraction $\eta$ of their energy, $\rm E_b=m_e\, c^2\, \gamma \,,$ is 

2144: synchrotron

2145: re-radiated. Let $\rm n_{eff}$ be the density of the emitting

2146: electrons, so that:

2147: \begin{equation}

2148: \rm \eta\, E_b\, {dN_e\over dt}=\int F_{_{CB}}[\nu]\,d\nu

2149: =\int dV\,dE_e\,E_e\,{dn_{eff}\over dE_e}\,{1\over \tau(E_e)}\; ,

2150: \label{neffV}

2151: \end{equation}

2152: with $\rm \tau(E_e)=E_e/(dE_e/dt)$ the  cooling time for electrons

2153: of energy $\rm E_e$,  determined from Eqs.~(\ref{mag}) and

2154: (\ref{Eloss}) to be:

2155: \begin{equation}

2156: \rm

2157: \rm \tau(E_e)=4\,{m_e\over m_p}\,{1\over n_p\,\sigma_{_T}\,c\,\gamma_e^3}\, .

2158: \label{cooltimes}

2159: \end{equation}

2160: For the electron spectrum of Eq.~(\ref{approx1}), the second integral 

2161: in Eq.~(\ref{neffV}) is dominated by energies $\rm E_e\sim E_b$

2162: ($\rm \gamma_e\sim \gamma_b=\gamma$). Thus, for a uniform distribution

2163: of electrons in the CB's volume $\rm V$, we obtain:

2164: \begin{eqnarray}

2165: \rm

2166: n_{eff}&\simeq&\rm -\,{\eta\over V}\,{dN_e\over dt}\,\tau(E_b)

2167: \nonumber\\ 

2168: &=&\rm{m_e\over m_p}\,{3\,\eta\over R_{max}\,\sigma_{_T}\,\gamma^2}

2169: \simeq (11\,\eta\, cm^{-3})\left[10^3\over\gamma\right]^2 ,

2170: \label{neff}

2171: \end{eqnarray}

2172: where we have used our reference $\rm R_{max}$.

2173: This density  could be somewhat 

2174: higher if the emitting electrons are concentrated on the CB's front surface.

2175: Note that this density increases with time like $\rm[\gamma(t)]^{-2}$. 

2176: 

2177: \subsection*{Synchrotron self absorption}

2178: For a power-law distribution, 

2179: $\rm dn_e/d\gamma_e= (p-1)\,n_{eff}\, \gamma_e^{-p}$,  

2180: the correct attenuation coefficient for synchrotron self absorption 

2181: at frequency $\nu$ is

2182: (e.g., Shu 1991, Eqs.~ 19.37,38):

2183: \begin{equation}

2184: \rm \chi_\nu= K_0\, (p-1)\,n_{eff}\,  {B_\perp}^{(p+2)/2}\, \nu^{-(p+4)/2}\, ,

2185: \label{chisyn}

2186: \end{equation}

2187: where: 

2188: \begin{equation}

2189: \rm K_0={c\, r_e\over 4\, \sqrt{3}}\, 

2190: \left ({3\, e\over 2\, \pi\, m_e\, c }\right)^{p+2\over 2}

2191:        \Gamma\left({3\, p+2\over 12}\right) 

2192:        \Gamma\left({3p+22\over 12}\right),

2193: \label{K0}

2194: \end{equation}

2195: with $\rm r_e\, =\, e^2/m_e\, c^2\, .$ 

2196: An  observed frequency $\rm \nu_{obs}$ was emitted at 

2197: $\rm \nu(t)=(1+z)\, \nu_{ob}/\delta(t)$  in 

2198: the CB's rest frame.  As an example, $\rm \nu_{ob}=5$ GHz

2199: from a decelerating CB with

2200: $\rm\delta(t)\simeq \gamma_0/2 \simeq  500$,  if emitted at a typical

2201: $\rm z=1$, corresponds to $\rm \nu=20$ MHz in the CB rest frame. 

2202: For our reference parameters, the synchrotron 

2203: self absorption coefficient in the CB is $\rm\chi_\nu \simeq 2.4\times 

2204: 10^{-10}\,\eta\, cm^{-1}$.  Since $\rm n_{eff}\propto \gamma^{-2}$, 

2205: $\rm B\propto \gamma$, 

2206: and, after a few observer's days, $\delta\sim 2\gamma$ and

2207: $\rm\rm\gamma(t)\sim t^{-1/3}$, $\chi_\nu$ decreases 

2208: with time like $\rm \gamma^{(p+1)}\sim t^{-(p+1)/3}\sim t^{-1.1}$.

2209: 

2210: \subsection*{Bremsstrahlung self absorption} 

2211: The  X-ray AG is dominated first by bremsstrahlung from plasma

2212: electrons and later by synchrotron radiation from the swept up high

2213: energy electrons (DDD 2001). The observed X-ray flux, or the

2214: theoretical UV flux in the CB rest frame, can be used to show that

2215: the CB is partially ionized during the radio AG observations and

2216: that the ionized fraction of the CB plasma is proportional to $\rm

2217: \gamma(t)$ (see Appendix III).  The logarithmic dependence of the

2218: plasma temperature in the Saha equation on the fractional ionization,

2219: keeps the CB's temperature nearly constant during its  AG phase.

2220: Consequently, in the CB rest frame, the free-free attenuation at

2221: a fixed frequency is proportional to $\rm [\gamma(t)]^2$ and the

2222: free-free (bremsstrahlung) absorption coefficient is that of

2223: Eq.~(\ref{chifree}).

2224: 

2225: 

2226: The temperature of the partially ionized CB is of ${\cal{O}}(1)$

2227: eV and almost constant during the observed AG.  For $\sim$ 20 MHz

2228: emission from a thermal plasma at such temperature, $\rm g\sim 10$

2229: and for one tenth of the typical bulk CB density, $\rm \bar n_e\sim

2230: 10^6\ cm^{-3}$, one obtains from Eq.~(\ref{chifree})

2231:  $\rm\chi_\nu\simeq 4\times 10^{-10}\, cm^{-1}$, which is $\sim

2232: 1.7/\eta$ larger than the synchrotron absorption coefficient of

2233: the energetic electrons in the CB (the values of $\eta$ are listed

2234: in Table III).  At a fixed observer frequency, $\rm \nu =(1+z)\,

2235: \nu_{ob}/\delta$, the free-free opacity of the CB decreases roughly

2236: like $\rm \sim\gamma^2\sim t^{-2/3}$ compared with the $\rm \sim

2237: t^{-1.1}$ decline of the synchrotron self opacity.

2238: 

2239: The conclusion is that free-free absorption is dominant for

2240: as long as the ionization of the CB is considerable.

2241: 

2242: 

2243: \section*{Appendix V: Emission of self absorbed radiations}

2244: 

2245: We argue that the energy of the self absorbed radio waves and 

2246: ionizing photons in the CB is radiated mainly by thermal bremsstrahlung

2247: and line emission from the CB contributing significantly to the 

2248: observed X-ray afterglow.

2249:  

2250: The absorbed radio power is roughly equal to the integrated

2251: emissivity of the CB over all frequencies below $\rm \nu_a$,

2252: defined by Eq.~(\ref{absfreq}).  For the spectrum of Eq.~(\ref{sync}), 

2253: and the normalization of Eq.~(\ref{cbemissivity}), the

2254: absorbed  power is:

2255: \begin{equation}

2256: \rm {dE_{ab}\over dt}\simeq {\eta\, \pi\, R_{max}^2\, n_e\, m_e\, c^3\, 

2257:                        [\gamma(t)]^2\, (p-2)\over (p-1)}\, 

2258:                         \left[{\nu_a\over \nu_b}\right]^{1/2}\, .

2259: \end{equation} 

2260: The CB self absorbed radio energy becomes part of the thermal 

2261: energy of the CB plasma. It is radiated by the plasma

2262: as thermal bremsstrahlung at optical wavelengths (in the CB rest frame).

2263: For our reference parameters, $\rm \nu_a\sim$ 100 MHz and  

2264: $\eta<1$, this absorbed power is smaller than the power

2265: absorbed by photoionization.

2266: 

2267: The recombination energy is radiated  at a rate 

2268: $\rm \approx R_{rec}\,  x\, N_{cb} \, I $ in the CB rest frame

2269: where $\rm x=\bar n_e/\bar n_b $ is the fraction of ionized hydrogen in 

2270: the CB and $\rm I=13.6\, eV$ is the binding energy of hydrogen in 

2271: its ground state. In the distant observer frame, the observed  radiation

2272: is boosted and collimated by the highly relativistic motion of the CB  

2273: and redshifted by the cosmological expansion to:

2274: \begin{equation} 

2275: \rm {dE_{rec}\over dt} 

2276: \simeq {R_{rec}\, x\, N_{cb}\, I\, (1+z)\, [\delta(t)]^4\over 4\, \pi\, 

2277: D_L^2}\, . 

2278: \label{Erecombination} 

2279: \end{equation} 

2280: For our reference parameters, hydrogen recombination 

2281: produces X-ray 

2282: lines with a total energy flux of $\rm \sim x^2\, \times 10^{-12}\, erg\, 

2283: s^{-1}\,cm^{-2}\, .$ 

2284: 

2285: Due to their large Doppler shift $\delta$, the 

2286: hydrogen emission lines (and the emission lines from the swept up ISM and

2287: supernova shell material) as well as the CB's  thermal bremsstrahlung, 

2288: are shifted to the observer's X-ray band. They

2289: contribute significantly to the X-ray afterglow and may provide a simple

2290: alternative explanation (Dar and De R\'ujula 2000) to the commonly assumed

2291: Fe-line origin of the X-ray lines observed in the afterglows of GRB

2292: 970508: Piro et al. (1998), GRB 970828: Yoshida et al. (1999; 2001), GRB

2293: 991216: Piro et al. (2000) and GRB 000214: Antonelli et al. (2000). 

2294: 

2295: 

2296: 

2297: 

2298: \begin{thebibliography}{} 

2299: 

2300: \bibitem{}

2301: Achterberg A., et al., 2001, MNRAS 328, 393 

2302: \bibitem{} 

2303: Andersen M.I., et al., 2000, A\&A 364, 54L

2304: \bibitem{} 

2305: Antonelli L.A., et al., 2000, ApJ 545, L39 

2306: \bibitem{}

2307: Axford W.I., Leer E., Skadron G., 1997 {\it Proc. International 

2308: Cosmic Ray Conf.,}  11, 132

2309: \bibitem{}

2310: Ballard K.R. \&  Heavens A.F., 1992, MNRAS 259, 89 

2311: \bibitem{} 

2312: Bell, A.R. 1978, MNRAS 182, 147.

2313: \bibitem{} 

2314: Berger E., et al., 2000, ApJ, 545, 56 

2315: \bibitem{} 

2316: Berger E., et al., 2001a, ApJ 556, 556 

2317: \bibitem{}

2318: Berger E., et al., 2001b, ApJ 549, L7

2319: \bibitem{}

2320: Berger E., et al., 2001c, astro-ph/0112558 

2321: \bibitem{}

2322: Berger E., et al., 2001d, astro-ph/0102278

2323: \bibitem{}

2324: Beuermann K., et al., 1999, A\&A, 352, L26

2325: \bibitem{}

2326: Blandford R.D. \& McKee C.F., 1976, Phys. of Fluids, 19, 1130 

2327: \bibitem{}

2328: Blandford R.D. \& Ostriker J.P. 1978 ApJ 221, L29 

2329: \bibitem{}

2330: Castro-Tirado A.J., et al., 1998, Science 279, 1011

2331: \bibitem{}

2332: Castro-Tirado A.J., et al., 1999, Science 283, 2069

2333: \bibitem{} 

2334: Castro-Tirado A.J., et al., 2001, A\&A 370, 398

2335: \bibitem{}

2336: Chary, R. et al. 1998, ApJ 498, L9

2337: \bibitem{}

2338: Dado, S., Dar, A. \& De R\'ujula, A., 2001, A\&A 

2339: 388 (2002) 1079 (astro-ph/0107367) 

2340: \bibitem{}

2341: Dado, S., Dar, A. \& De R\'ujula, A., 2002,

2342: astro-ph/0207015

2343: \bibitem{} 

2344: Dar A. \& De R\'ujula, A., 2000a, astro-ph/0008474 

2345: \bibitem{} 

2346: Dar A. \& De R\'ujula, A., 2000b, astro-ph/0012227 

2347: \bibitem{} 

2348: Dar A. \& De R\'ujula, A., 2001, astro-ph/0102115 

2349: \bibitem{} 

2350: Dennett-Thorpe J. \& de Bruyn A.G., 2000, ApJ 529, L65 

2351: \bibitem{} 

2352: Dennett-Thorpe J., de Bruyn A.G., 2002, astro-ph/0201061

2353: \bibitem{}

2354: Di Paola A., et al., 2000, GCN Circ. 816

2355: \bibitem{} 

2356: Frail D.A.,  et al., 2000a, ApJ 537, 191  

2357: \bibitem{} 

2358: Frail D.A., et al., 2000b,  ApJ, 538, L129 

2359: \bibitem{} 

2360: Fruchter A.S. et al., 1999, ApJ, 519 L13

2361: \bibitem{} 

2362: Fynbo J.U., et al., 2000, ApJ 542, L89 

2363: \bibitem{}

2364: Fynbo J.U., et al., 2001, A\&A 373, 796

2365: \bibitem{}

2366: Galama T.J., et al., 1998a, ApJ 500, L97

2367: \bibitem{} 

2368: Galama T.J., et al., 1998b, ApJ 497, L13G

2369: \bibitem{} 

2370: Galama T.J., et al., 1998c, Nature 395, 670 

2371: \bibitem{} 

2372: Galama T.J., et al., 1999, Nature 398, 394

2373: \bibitem{} 

2374: Galama, T.J., et al.,  2000, ApJ 541, L45 

2375: \bibitem{}

2376: Garcia M.R., et al., 1998, ApJ 500, L105

2377: \bibitem{} 

2378: Garnavich P.M., et al.,  2000a, ApJ 543, 61

2379: \bibitem{}

2380: Garnavich P.M., et al.,  2000b, ApJ 544, L11

2381: \bibitem{}

2382: Granot J. et al., astro-ph/0201322

2383: \bibitem{} 

2384: Granot J. \& and Sari R., 2002, ApJ 568, 820

2385: \bibitem{} 

2386: Gupta Y., 1995, ApJ 451, 717 

2387: \bibitem{}

2388: Halpern J.P., et al., 2000, ApJ 543, 697

2389: \bibitem{} 

2390: Harrison F.A., et al., 1999 ApJ 523, L121 

2391: \bibitem{} 

2392: Harrison F.A., et al., 2001 ApJ 559, 123 

2393: \bibitem{}

2394: Holland S., et al., 2000, A\&A 364, 467 

2395: \bibitem{} 

2396: Iwamoto K., et al., 1998, Nature 395, 672 

2397: \bibitem{}

2398: Jensen B.L., et al., 2001, A\&A, 370, 909

2399: % \bibitem{} 

2400: % Karzas W.,  Latter R., 1961, ApJ Suppl. 6, 167 

2401: \bibitem{} 

2402: Kay L.E.  et al., 1998, IAU Circ. 6969 

2403: \bibitem{}

2404: Klose S.,  et al.,  2000 ApJ 545, 271

2405: \bibitem{}

2406: Kouveliotou C. 2002, Talk at the GRB 2002 Conference, Rome

2407: \bibitem{}

2408: Krymsky G.F. 1977, Dok. Acad. Nauk. USSR 234, 1306

2409: \bibitem{} 

2410: Kulkarni S.R., et al., 1998, Nature 395, 663 

2411: \bibitem{}

2412: Kulkarni S.R., et al., 1999a, Nature 398, 389

2413: \bibitem{} 

2414: Kulkarni, S.R., et al., 1999b, ApJ 522, L97 

2415: \bibitem{} 

2416: Li Z., Chevalier R.A., 1999, ApJ 526, L716 

2417: \bibitem{} 

2418: Lidman C., et al., 1998, IAU 6895 

2419: \bibitem{} 

2420: Lyne A.G., Smith F.G., 1982, Nature 289, 825 

2421: \bibitem{}

2422: Masetti N., et al., 2000, A\&A 359, 23

2423: \bibitem{}

2424: Masetti N., et al., 2001, A\&A 374, 382

2425: \bibitem{}

2426: Metzger M.R., et al., 1997, Nature 387, 878

2427: \bibitem{} 

2428: Mirabel I.F.,  Rodriguez L.F., 1994, Nature 371, 46 

2429: \bibitem{} 

2430: Mirabel I.F., Rodriguez L.F., 1999, ARA\&A 37, 409 

2431: \bibitem{} 

2432: Nicastro L. et al. 2001, astro-ph/0101232 

2433: \bibitem{} 

2434: Paczynski B., 1998, ApJ 494, L45 

2435: \bibitem{}

2436: Panaitescu, A. \& Kumar, P. astro-ph/0109124, v2

2437: \bibitem{} 

2438: Patat F., Piemonte A., 1998 IAU Circ. 6918 

2439: \bibitem{} 

2440: Patat F., et al., 1998, IAU Circ. 7215 

2441: \bibitem{} 

2442: Pedersen H., et al., 1998, ApJ 496, 311

2443: \bibitem{} 

2444: Pian E. et al., 1998, ApJ 492, L103 

2445: \bibitem{} 

2446: Pian E., et al., 2000, ApJ 536, 778 

2447: \bibitem{}

2448: Pian E., 2002, Talk at the GRB 2002 Conference, Rome

2449: \bibitem{} 

2450: Piro L., et al., 1998, A\&A 331, L41 

2451: \bibitem{} 

2452: Piro L., et al., 2000, Science 290, 955 

2453: \bibitem{}

2454: Price P., et al., 2001, ApJ 549, L7

2455: \bibitem{} 

2456: Readhead A.C.S., 1994, ApJ 426, 51 

2457: \bibitem{}

2458: Rhoads J.E., Fruchter A.S., 2001, ApJ 546, 117

2459: \bibitem{angle1}

2460: Rossi, E. et al. astro-ph/0112083

2461: \bibitem{} 

2462: Rybicki G.B.,  Lightman A.P., 1979,  {\it Radiative Processes in 

2463: Astrophysics} John Wiley \& Sons, Inc. 

2464: \bibitem{}

2465: Sadler E.M., et al., 1998, IAUC Circ. 6901

2466: \bibitem{}

2467: Sagar R., et al., 2000a, BASI 15, 15

2468: \bibitem{}

2469: Sagar R., et al., 2000b, BASI 28, 499

2470: \bibitem{} 

2471: Sagar R., et al., 2001, BASI, 29, 1

2472: \bibitem{} 

2473: Sahu K.C., et al., 1997 ApJ 489, L127

2474: \bibitem{} 

2475: Sahu K.C., et al., 2000, ApJ 540, 74   

2476: \bibitem{}

2477: Salmonson J.D. \& Galama T.J., astro-ph/0112298

2478: \bibitem{}

2479: Schaefer B. et al., 1997, IAU Circ. 6658

2480: \bibitem{}

2481: Schlegel D.J., Finkbeiner D.P., Davis M., 1998, ApJ, 500, 525

2482: \bibitem{}

2483: Shu F.H., et 

2484: al., 1991, {it The Physics of Astrophysics} University Science Books 

2485: \bibitem{}

2486: Smette A., 2001, ApJ 556, 70

2487: \bibitem{}

2488: Soffita P., et al., 1998, IAU Circ. 6884 

2489: \bibitem{}

2490: Sokolov V.V., et al., 1998, A\&A 334, 117

2491: \bibitem{}

2492: Sollerman J., et al., 2000, astro-ph/0006406

2493: \bibitem{}

2494: Stanek K.Z., et al., 1999, ApJ 522, L39

2495: \bibitem{}

2496: Stanek K.Z., et al., 2001; ApJ 563, 592

2497: \bibitem{} 

2498: Tinney C., et al., 1998 IAU Circ. 6896 

2499: \bibitem{} 

2500: Wagner S.J., Witzel A., 1995, ARAA 33, 163 

2501: \bibitem{} 

2502: Weiler K.W., et al., 2001, ApJ 562, 670 

2503: \bibitem{}

2504: Wieringa, M., et al., 1998, GCN Circ. 63 \\

2505: (http://gcn.gsfc.nasa.gov/gcn/cgn3/063.gcn3)

2506: \bibitem{} 

2507: Yoshida A., et al., 1999, A\&A 138S, 433 

2508: \bibitem{}

2509: Yoshida A. et al., 2001, ApJ 557,  L27 

2510: \bibitem{}

2511: Yost, S.A. et al. 2001, astro-ph/0107556

2512: \bibitem{}

2513: Yost, S.A. et al. 2002, astro-ph/0204141

2514: \bibitem{} 

2515: Zhang, B. \& Meszaros, P.L., astro-ph/0112118

2516: \bibitem{}

2517: Zharikov S.V., et al., 1998, A\&A 337, 356

2518: 

2519: 

2520:  

2521: \end{thebibliography} 

2522:  

2523: \clearpage 

2524: \newpage 

2525: 

2526: { \vskip 0.3 true cm

2527: \noindent 

2528: {\bf Table I - Reference parameters} }

2529: \vskip -0.5 true cm 

2530: \begin{table}[h] 

2531: %\vskip 0.1 true cm 

2532: %\huge\bf

2533: \normalsize

2534: \hspace{.0cm} %if you want to center your table act on this argument 

2535: \begin{tabular}{|l|c|c|c|} 

2536: \hline 

2537: %\multicolumn{6}{GRB with redshifts}\\ 

2538: \hline 

2539: Fitted & Value & Definition \\ 

2540: \hline 

2541: $\theta$             & $10^{-3}$          & Observer's viewing angle   \\ 

2542: $\gamma_0$     & $10^3 $             & Lorentz factor at $\rm t=0$  \\ 

2543: $\rm x_\infty$   & 1.3  Mpc            & Deceleration parameter \\ 

2544: %\hline

2545: %$\rm p $            & 2.2                     & Uncooled electron index \\

2546: %\hline

2547: \hline

2548: \hline

2549: Other &  &  \\

2550: \hline

2551: $\delta_0$               & $10^3$              & Doppler factor at $\rm t=0$ \\

2552: $\rm x_\infty/\gamma_0$   & 1.3 kpc & Distance $\rm \gamma_0\to \gamma_0/2$\\ 

2553: $\rm R_{max}$       & $2.2\times10^{14}$  cm         & CB's maximum radius \\ 

2554: $\rm \bar{n}_e$       &$\rm10^{7}\,cm^{-3}$    & CB e number-density \\

2555: $\rm N_{_{CB}}$     & $6\times10^{50}$            & CB's baryon number   \\

2556: \hline 

2557: \hline 

2558: Ambient &  &  \\

2559: \hline

2560: $\rm n_p $      & $\rm 10^{-3}\,cm^{-3}$        & Distant p number-density \\

2561: $\rm n_p^{SN} $      & $\rm1 \,cm^{-3}$       & Close-by p number-density \\

2562: \hline

2563: \hline

2564: \end{tabular} 

2565: \end{table} 

2566: \vskip -0.3 true cm 

2567: \noindent 

2568: {\bf Comments:} The ``Fitted'' parameters are the typical values 

2569: in the fits to optical and X-ray AGs. 

2570: %The spectral index $\rm p$ could be input, rather than fit.

2571: The ``Other'' parameters

2572: are deduced from the fitted ones ($\delta_0$), are calculated 

2573: ($\rm R_{max}$),  or are deduced from the rest

2574: ($\rm N_{_{CB}}$  and $\rm \bar{n}_e$). ``Ambient''

2575: numbers refer to the ISM, not the CBs.

2576:  

2577: 

2578: 

2579: %\pagebreak 

2580: \vspace{.5  cm} 

2581: 

2582: \noindent

2583: {\bf

2584: Table II -  The crossing of the bend frequency through the U to I bands

2585: [($\sim$ 10 to 3)  $\times 10^{15}$ Hz]}

2586: %\vskip 0.2 true cm

2587: \begin{table}[h]

2588: %\vskip 0.1 true cm

2589: \hspace{-.1cm} %if you want to center your table act on this argument

2590: \begin{tabular}{|l|c|c|c|c|c|c|l|}

2591: \hline

2592: \hline

2593: GRB   &$\rm\nu_b^0$  & $\rm t_1$ & $\rm\beta(t_1)$ & $\rm t_2$ & $\rm\beta(t_2)$    

2594: \\

2595: \hline

2596: \hline

2597: 970508& $3.7$ & 0.1-1.5 & $-0.58\!\pm\! 0.40$ & 12.1 & $-1.12\!\pm\!0.04$ \\

2598: 000301c&$5.8$ &1.8 & $-0.90\!\pm\! 0.20$ & 6-8 &$-1.19\!\pm\! 0.15$ \\

2599: 000926& $7.3$ &0.9 & $-0.90\!\pm\! 0.18$ & 3.9 &$-1.00\!\pm\! 0.18$ \\

2600: 990712& $13$ &0.5-1 & $-0.70\!\pm\! 0.10$ &    &$                 $ \\

2601: 991208& $17$ &      & $                 $ & 3.8 &$-1.05\!\pm\! 0.05$ \\

2602: 010222& $18$ & 0.20 & $-0.88\!\pm\! 0.10$ & 1-5 &$-1.10\!\pm\! 0.10$ \\

2603: 991216& $20$ & 1.67 & $-0.58\!\pm\! 0.08$ &     &$                 $ \\

2604: 990510& $27$ & 0.89 & $-0.61\!\pm\! 0.12$ & 3.6 &$-1.29\!\pm\! 0.23$ \\

2605:             &           &          &                                     &        &$-1.11\!\pm\! 0.12$ \\

2606: 990123& $45$ & 0.033& $-0.69\!\pm\! 0.10$ & 1-3 &$-0.90\!\pm\! 0.18$ \\

2607: \hline

2608: \end{tabular}

2609: \end{table}

2610: \vskip -0.3 true cm

2611: \noindent

2612: {\bf Comments:} $\rm\nu_b^0$: bend frequency in units of $10^{14}$ Hz.

2613: $\beta$: spectral index.

2614: $\rm t_i$: times after burst in days. For the second entry on

2615: GRB 990510, see the text.\\  

2616:     {\bf References}:\\

2617: {   GRB 970508}: Galama et al.~1998a\\   

2618: {   GRB 990123}: Andersen et al. 1999; Holland et al.~2000\\

2619: {   GRB 990510}: Stanek et al.~1999; Holland et al.~2000; 

2620: Beuermann et al. 1999 \\

2621: {   GRB 990712}: Sahu, et al.~2000\\

2622: {   GRB 991208}: Castro-Tirado et al.~2001\\

2623: {   GRB 991216}: Garnavich et al.~2000a\\

2624: {   GRB 000301c}: Jensen et al. 2001; Rhoads \& Fruchter~2000 \\

2625: {   GRB 000926}: Fynbo et al. 2001; Harrison et al 2001\\

2626: {   GRB 010222}: Stanek et al. 2001; Masetti et al.~2001 \\

2627: 

2628: \newpage

2629: 

2630: \noindent

2631: {\bf

2632: Table III -  The Afterglow  Parameters}

2633: %\vskip 0.2 true cm

2634: \begin{table}[h]

2635: %\vskip 0.1 true cm

2636: \hspace{+1.1cm} %if you want to center your table act on this argument

2637: \begin{tabular}{|l|c|c|c|c|c|c|l|}

2638: \hline

2639: \hline

2640: GRB   &$\gamma_0 $ &  $\theta $ &$\rm x_\infty $& $\rm \nu_a$& $\eta$ \\

2641: \hline

2642: 000301c       &1061      &2.321    & 0.128    & 552   &  0.025  \\

2643: 000926        & 787      &0.235    & 0.083    & 722   &  0.027  \\

2644: 991216        & 906      &0.403    & 0.462    &  46   &  0.029  \\

2645: 991208        &1034      &0.111    & 1.014    & 103   &  0.011  \\

2646: 000418        &1241      &2.061    & 0.332    & 298   &  0.024  \\

2647: 990123        &1208      &0.464    & 0.364    &1604   &  0.009  \\

2648: 990510        &1009      &0.261    & 0.372    & 107   &  0.015  \\

2649: 970508        & 769      & 2.51    & 0.516    & 559   &  0.035  \\

2650: 980425        & 495      &7.831    & 0.425    & 102   &  0.007  \\  

2651: \hline

2652: \end{tabular}

2653: \end{table}

2654: \vskip -0.3 true cm

2655: \noindent

2656: {\bf Comments:}

2657: $\gamma_0$: Initial Lorentz factor.

2658: $\theta$: Viewing angle relative to the CB line of motion,

2659: in milliradians.

2660: $\rm x_\infty$: Deceleration parameter in Mpc

2661: ($\gamma=\gamma_0/2$ at $\rm x= x_\infty/\gamma_0$).

2662: $\rm \nu_a $: absorption frequency in MHz in the CB rest frame at t=0.

2663: $\eta$: Our best fit normalization divided by the expected normalization  

2664: for the reference parameters that we had chosen in previous works.

2665:  

2666: \vspace{1cm}

2667: {\bf 

2668: \noindent 

2669: Table IV - Frequencies, in GHz, at which the radio AGs  of GRBs  

2670: of known redshift were measured }

2671: \vspace{-.3cm} 

2672: \begin{table}[h] 

2673: \hspace{1cm}  

2674: \begin{tabular}{|l|c|c|c|c|c|l|} 

2675: \hline 

2676: \hline 

2677:  991208 &  991216 & 000301c & 000418 & 000926 \\ 

2678: \hline 

2679:         & 350     & 350    &        &       \\  

2680:         &         & 250    &        &       \\ 

2681:   100   &  100    & 100    &        & 98.48 \\ 

2682:   86.24 &         &        &        &       \\  

2683:   30    &         &        &        &       \\ 

2684:   22.5  &         & 22.5   & 22.46  & 22.5  \\ 

2685:   14.97 & 15      & 15     & 15     & 15    \\ 

2686:   8.46  & 8.46    & 8.46   & 8.46   & 8.46  \\  

2687:   4.86  & 4.86    & 4.86   & 4.86   & 4.86  \\ 

2688:   1.43  & 1.43    & 1.43   &        & 1.43  \\ 

2689: \hline 

2690:  970508 & 980425 & 990123 & 990510 &   \\          

2691: \hline 

2692:         &        &  351   &       &    \\    

2693:         &        &  222   &       &    \\  

2694:         &        &  15    &  13.68&    \\  

2695:  8.46   & 8.64   &  8.46  &   8.65&    \\    

2696:  4.86   & 4.8    &  4.88  &   4.8 &    \\ 

2697:  1.43   & 2.49   &        &       &    \\  

2698:         & 1.38   &  1.38  &       &    \\ 

2699: \hline 

2700: \hline 

2701: \end{tabular} 

2702: \end{table} 

2703: \vskip -0.3 true cm 

2704: \noindent 

2705: {\bf References}:\\ 

2706: GRB 970508: Frail et al. 2000a\\ 

2707: GRB 980425: Kulkarni et al. 1998;\\ 

2708: GRB 990123: Kulkarni et al. 1999b; Galama et al. 1999\\ 

2709: GRB 990510: Harrison et al. 1999\\  

2710: GRB 991208: Galama et al. 2000 \\ 

2711: GRB 991216: Frail et al. 2000b \\  

2712: GRB 000301c: Berger et al 2000\\   

2713: GRB 000418: Berger  et al. 2001a\\ 

2714: GRB 000926: Harrison et al. 2001\\ 

2715: 

2716: 

2717: \clearpage

2718: 

2719: \begin{figure}[t]  

2720: \begin{tabular}{cc}  

2721: \hskip 1.0truecm  

2722: \vspace*{2cm} 

2723: \hspace*{-2.5cm}  

2724: \epsfig{file=SPECTRA2.eps, width=8cm}

2725: \vspace*{-1.5cm}

2726: \\ 

2727: %\hskip 1truecm 

2728: \hspace*{-.2cm}  

2729: \epsfig{file=ag32nr0301A.eps, width=9.5cm} 

2730: \end{tabular}  

2731: \caption{Typical predictions for the CB model's radio afterglow. 

2732: Upper panel: spectra at different times, from 1 to 300 days.

2733: The peak frequencies correspond to CB self-opacities of $\cal{O}$(1).

2734: The black dots are the location of the synchrotron frequency

2735: corresponding to the injection bend.

2736: Lower panel: Light curves at different radio frequencies,

2737: from 350 to 1.43 GHz. The asymptotic curve is $\rm t^{-2\,(p+1)/3}$

2738: (for $\rm \nu\gg\nu_b$, as is the case at all frequencies shown

2739: in this figure).}

2740: \label{figCBpreds}  

2741: \end{figure}

2742: 

2743: \begin{figure}[t]  

2744: \begin{tabular}{cc}  

2745: \hskip 2truecm  

2746: \vspace*{2cm} 

2747: \hspace*{-2cm}  

2748: \epsfig{file=INJECFREQ.eps, width=8cm}

2749: \vspace*{-1.5cm}\\ 

2750: %\hskip 1truecm 

2751: \hspace*{.2cm}  

2752: \epsfig{file=INJECFREQ2.eps, width=8cm} 

2753: \end{tabular}  

2754: \caption{Typical predictions for  the bend frequency in the AG spectrum

2755: as a function of time, for $\theta=0,\,3/\gamma_0$

2756: and $10/\gamma_0$. The ``optical'' U to I band is shown as a horizontal

2757: band. Upper panel: for $\gamma_0=1250$. Lower panel: for $\gamma_0=750$.}  

2758: \label{figinjection}  

2759: \end{figure} 

2760: 

2761: \clearpage

2762: 

2763: \begin{figure}[t]  

2764: \hskip 0truecm   

2765: \epsfig{file=ABBB.eps, width=8.5cm}    

2766: \vspace*{-0.1cm}

2767: \caption{A comparison between the predicted evolution in time

2768: of the effective spectral slope through the optical/NIR band and

2769: the data collected by Galama et al. (1998a) for

2770: the U, B, V, $\rm R_c$ $\rm I_c$ band of the AG of GRB 970508

2771: (upper panel), for the K and $\rm R_c$ band  (full squares,

2772: lower panel, Chary et al. 1998) and for the H and $\rm R_c$ band  (triangle,

2773: lower panel, Pian et al. 1998) The three coloured lines, in the same order,

2774: are the (parameter-less) predictions.}  

2775: \label{index2}  

2776: \end{figure} 

2777: 

2778: 

2779: \begin{figure}[t]  

2780: %\begin{tabular}{cc}  

2781: \hskip 0truecm   

2782: \epsfig{file=INJECTION3.eps, width=8.5cm}  

2783: %\end{tabular}  

2784: \caption{Comparison between the observations and the

2785: (parameter-less) prediction

2786: for the spectral {\it shape} of the optical AG of GRB 000301c, at 

2787: $\sim 3$ days after burst. Data from Jensen et al. (2001).}  

2788: \label{index}  

2789: \end{figure} 

2790: 

2791: 

2792: \begin{figure}[t]  

2793: %\begin{tabular}{cc}  

2794: \hskip 0truecm   

2795: \epsfig{file=ag33nro0301.eps, width=8.5cm}  

2796: %\end{tabular}  

2797: \caption{Results of a fit to radio and optical observations of 

2798: the light curves of GRB 000301c. The narrowly spaced lines refer

2799: ---from top to bottom--- to the K, J, I, R, V, B and U bands.

2800: The more widely spaced lines refer

2801: ---from top to bottom at the figure's left side--- to 

2802: frequencies of 1.43, 4.86, 8.46, 15, 22.5, 100, 250 and 350 GHz.

2803: The comparison with data is shown in Figs.~(\ref{K301}) to (\ref{rad-opt301b}).

2804: }  

2805: \label{all301}  

2806: \end{figure} 

2807: 

2808: 

2809: 

2810: 

2811: 

2812: \begin{figure}[t]  

2813: %\begin{tabular}{cc}  

2814: \hskip 0truecm   

2815: \epsfig{file=ag42nro0301.eps, width=8.5cm}  

2816: %\end{tabular}  

2817: \caption{Comparisons between our fitted CB model AG of GRB 000301c, 

2818: at $\rm z=2.033$, 

2819: Eq.~(\ref{Fnuobser}) and Fig.~(\ref{all301}), with the observed optical

2820: data. 

2821: The figure shows (from top to bottom) 1000 times the K-band results,

2822: 100 times the J-band, 10 times the I-band, the R-band, 1/10 of the V-band,

2823: 1/100 of the B-band and 1/1000 of the U-band.

2824: The contributions of the underlying galaxy and 

2825: an expected (but,  in this case, unobservable) SN1998bw-like 

2826: SN have been subtracted. }  

2827: \label{K301}  

2828: \end{figure} 

2829:  

2830: %\clearpage

2831: 

2832: \begin{figure}[t]  

2833: \begin{tabular}{cc}  

2834: \hskip 2truecm  

2835: \vspace*{2cm} 

2836: \hspace*{-1.7cm}  

2837: \epsfig{file=ag11nr0301.eps, width=8cm} \\ 

2838: %\hskip 1truecm  

2839: \hspace*{.5cm}  

2840: \epsfig{file=ag12nr0301.eps, width=8cm} 

2841: \end{tabular}  

2842: \caption{Comparisons between our fitted CB model afterglow, 

2843: Eq.~(\ref{Fnuobser}), and the observed radio afterglow of GRB 000301c. 

2844: Upper panel: the light curve at 350 

2845: GHz. Lower panel: the light curve at 250 GHz.}  

2846: \label{figr030101}  

2847: \end{figure} 

2848:  

2849: \clearpage

2850: 

2851: \begin{figure}[t]  

2852: \begin{tabular}{cc}  

2853: \hskip 2truecm  

2854: \vspace*{2cm} 

2855: \hspace*{-1.7cm}  

2856: \epsfig{file=ag13nr0301.eps, width=8cm} \\ 

2857: %\hskip 1truecm  

2858: \hspace*{.5cm}  

2859: \epsfig{file=ag14nr0301.eps, width=8cm} 

2860: \end{tabular}  

2861: \caption{Comparisons between our fitted CB model afterglow, 

2862: Eq.~(\ref{Fnuobser}), and the observed radio afterglow of GRB 000301c.

2863: Upper panel: the light curve at 100 

2864: GHz. Lower panel: the light curve at 22.5 GHz.}  

2865: \label{figr030102}  

2866: \end{figure} 

2867:  

2868: 

2869:  

2870: \begin{figure}[t]  

2871: \begin{tabular}{cc}  

2872: \hskip 2truecm  

2873: \vspace*{2cm} 

2874: \hspace*{-1.7cm}  

2875: \epsfig{file=ag15nr0301.eps, width=8cm} \\ 

2876: %\hskip 1truecm  

2877: \hspace*{.5cm}  

2878: \epsfig{file=ag16nr0301.eps, width=8cm} 

2879: \end{tabular}  

2880: \caption{Comparisons between our fitted CB model afterglow, 

2881: Eq.~(\ref{Fnuobser}), and the observed radio afterglow of GRB 000301c.

2882: Upper panel: the light curve at 15 

2883: GHz. Lower panel: the light curve at 8.46 GHz.}  

2884: \label{figr030103}  

2885: \end{figure} 

2886:  

2887: \clearpage

2888:  

2889: \begin{figure}[t]  

2890: \begin{tabular}{cc}  

2891: \hskip 2truecm  

2892: \vspace*{2cm} 

2893: \hspace*{-1.7cm}  

2894: \epsfig{file=ag17nr0301.eps, width=8cm} \\ 

2895: %\hskip 1truecm  

2896: \hspace*{.5cm}  

2897: \epsfig{file=ag18nr0301.eps, width=8cm} 

2898: \end{tabular}  

2899: \caption{Comparisons between our fitted CB model afterglow, 

2900: Eq.~(\ref{Fnuobser}), and the observed radio afterglow of GRB 000301c.

2901: Upper panel: the light curve at 4.86 

2902: GHz. Lower panel: the light curve at 1.43 GHz.}  

2903: \label{figr030104}  

2904: \end{figure} 

2905: 

2906: %\clearpage

2907: 

2908: \begin{figure}[t]  

2909: \begin{tabular}{cc}  

2910: \hskip 2.5truecm  

2911: \vspace*{2cm} 

2912: \hspace*{-2.7cm}  

2913: \epsfig{file=ag21nro0301.eps, width=8cm}

2914: \vspace*{-1.5cm}

2915: \\ 

2916: %\hskip 1truecm 

2917: \hspace*{-.2cm}  

2918: \epsfig{file=ag22nro0301.eps, width=8cm} 

2919: \end{tabular}  

2920: \caption{The spectrum of the AG of GRB 000301c from radio to optical

2921: frequencies.

2922: Upper panel: in the time interval between 1 and 5 days after burst.

2923: Lower panel:  in the time interval between 5 and 10 days after burst. 

2924: The highest peaking curve in the upper pannel corresponds to the later 

2925: time and in the lower panel to the earlier time.} 

2926: \label{rad-opt301}  

2927: \end{figure} 

2928: 

2929: \clearpage

2930: 

2931: \begin{figure}[t]  

2932: \begin{tabular}{cc}  

2933: \hskip 2.5truecm  

2934: \vspace*{2cm} 

2935: \hspace*{-2.7cm}  

2936: \epsfig{file=ag23nro0301.eps, width=8cm}

2937: \vspace*{-1.5cm}

2938: \\ 

2939: %\hskip 1truecm 

2940: \hspace*{-.2cm}  

2941: \epsfig{file=ag24nro0301.eps, width=8cm} 

2942: \end{tabular}  

2943: \caption{The spectrum of the AG of GRB 000301c from radio to optical

2944: frequencies.

2945: Upper panel: in the time interval between 10 and 20 days after burst.

2946: Lower panel:  in the time interval between 20 and 30 days after burst. 

2947: In both cases the highest peaking curve

2948: corresponds to the earlier time.}  

2949: \label{rad-opt301b}  

2950: \end{figure} 

2951: 

2952: %\clearpage

2953: 

2954: %desde 

2955: 

2956: 

2957: \begin{figure}[t]  

2958: %\begin{tabular}{cc}  

2959: \hskip 0truecm   

2960: \epsfig{file=ag42nro0926.eps, width=8.5cm}  

2961: %\end{tabular}  

2962: \caption{Comparisons between our fitted CB model afterglow, 

2963: Eq.~(\ref{Fnuobser}), 

2964: and the observed optical afterglow of GRB 000926 

2965: at $\rm z=2.037$. 

2966: The figure shows (from top to bottom) 1000 times the K-band results,

2967: 100 times the J-band, 10 times the I-band, the R-band, 1/10 of the V-band,

2968: 1/100 of the B-band and 1/1000 of the U-band.

2969: The contributions of the underlying galaxy and 

2970: an expected (but,  in this case, unobservable) SN1998bw-like 

2971: SN have been subtracted. }  

2972: \label{opt926}  

2973: \end{figure}

2974: 

2975: \clearpage

2976: 

2977: \begin{figure}[t] 

2978: \begin{tabular}{cc} 

2979: \hskip 2truecm 

2980: \vspace*{2cm} 

2981: \hspace*{-1.7cm} 

2982: \epsfig{file=ag11nr0926.eps, width=8cm} \\ 

2983: %\hskip 1truecm 

2984: \hspace*{.5cm} 

2985: \epsfig{file=ag12nr0926.eps, width=8cm} 

2986: \end{tabular} 

2987: \caption{Comparisons between our fitted CB model afterglow, 

2988: Eq.~(\ref{Fnuobser}), 

2989: and the observed radio afterglow of GRB 000926. 

2990: Upper panel: the light curve at 98.48 GHz. 

2991: Lower panel: the light curve at 22.5 GHz.} 

2992: \label{figr092601} 

2993: \end{figure} 

2994:  

2995: 

2996: 

2997: \begin{figure}[t] 

2998: \begin{tabular}{cc} 

2999: \hskip 2truecm 

3000: \vspace*{2cm} 

3001: \hspace*{-1.7cm} 

3002: \epsfig{file=ag13nr0926.eps, width=8cm} \\ 

3003: %\hskip 1truecm 

3004: \hspace*{.5cm} 

3005: \epsfig{file=ag14nr0926.eps, width=8cm} 

3006: \end{tabular} 

3007: \caption{Comparisons between our fitted CB model afterglow, 

3008: Eq.~(\ref{Fnuobser}), 

3009: and the observed radio afterglow of GRB 000926. 

3010: Upper panel: the light curve at 15 GHz. 

3011: Lower panel: the light curve at 8.46 GHz.} 

3012: \label{figr092602} 

3013: \end{figure} 

3014: 

3015: \clearpage 

3016: 

3017: \begin{figure}[t] 

3018: \begin{tabular}{cc} 

3019: \hskip 2truecm 

3020: \vspace*{2cm} 

3021: \hspace*{-1.7cm} 

3022: \epsfig{file=ag15nr0926.eps, width=8cm} \\ 

3023: %\hskip 1truecm 

3024: \hspace*{.5cm} 

3025: \epsfig{file=ag16nr0926.eps, width=8cm} 

3026: \end{tabular} 

3027: \caption{Comparisons between our fitted CB model afterglow, 

3028: Eq.~(\ref{Fnuobser}), 

3029: and the observed radio afterglow of GRB 000926. 

3030: Upper panel: the light curve at 4.86 GHz. 

3031: Lower panel: the light curve at 1.43 GHz.} 

3032: \label{figr092603} 

3033: \end{figure} 

3034: 

3035: %\clearpage

3036: 

3037: \begin{figure}[t]  

3038: \begin{tabular}{cc}  

3039: \hskip 2.5truecm  

3040: \vspace*{2cm} 

3041: \hspace*{-2.7cm}  

3042: \epsfig{file=ag21nro0926.eps, width=8cm}

3043: \vspace*{-1.5cm}

3044: \\ 

3045: %\hskip 1truecm 

3046: \hspace*{-.2cm}  

3047: \epsfig{file=ag22nro0926.eps, width=8cm} 

3048: \end{tabular}  

3049: \caption{The spectrum of the AG of GRB 000926 

3050: from radio to optical frequencies.

3051: Upper panel: in the time interval between 0.8 and 1.4 days after burst.

3052: Lower panel:  in the time interval between 1.4 and 2.5 days after burst.

3053: In both cases the highest peaking curve

3054: corresponds to the earlier time. }  

3055: \label{rad-opt926}  

3056: \end{figure} 

3057: 

3058: \clearpage

3059: 

3060: \begin{figure}[t]  

3061: \begin{tabular}{cc}  

3062: \hskip 2.5truecm  

3063: \vspace*{2cm} 

3064: \hspace*{-2.7cm}  

3065: \epsfig{file=ag23nro0926.eps, width=8cm}

3066: \vspace*{-1.5cm}

3067: \\ 

3068: %\hskip 1truecm 

3069: \hspace*{-.2cm}  

3070: \epsfig{file=ag24nro0926.eps, width=8cm} 

3071: \end{tabular}  

3072: \caption{The spectrum of the AG of GRB 000926 

3073: from radio to optical frequencies.

3074: Upper panel: in the time interval between 2.5 and 4 days after burst.

3075: Lower panel:  in the time interval between 4 and 8 days after burst.

3076: In both cases the highest peaking curve

3077: corresponds to the earlier time.}  

3078: \label{rad-opt926b}  

3079: \end{figure} 

3080: 

3081: %\clearpage

3082: 

3083: \begin{figure}[t]  

3084: \begin{tabular}{cc}  

3085: \hskip 2.5truecm  

3086: \vspace*{2cm} 

3087: \hspace*{-2.7cm}  

3088: \epsfig{file=ag25nro0926.eps, width=8cm}

3089: \vspace*{-1.5cm}

3090: \\ 

3091: %\hskip 1truecm 

3092: \hspace*{-.2cm}  

3093: \epsfig{file=ag26nro0926.eps, width=8cm} 

3094: \end{tabular}  

3095: \caption{The spectrum of the AG of GRB 000926 

3096: from radio to optical frequencies.

3097: Upper panel: in the time interval between 8 and 20 days after burst.

3098: Lower panel:  in the time interval between 20 and 100 days after burst.

3099: In both cases the highest peaking curve

3100: corresponds to the earlier time.}  

3101: \label{rad-opt926c}  

3102: \end{figure} 

3103: 

3104: \clearpage

3105: 

3106: \begin{figure}[t]  

3107: %\begin{tabular}{cc}  

3108: \hskip 0truecm   

3109: \epsfig{file=ag42nro1216.eps, width=8.5cm}  

3110: %\end{tabular}  

3111: \caption{Comparisons between our fitted CB model afterglow, 

3112: Eq.~(\ref{Fnuobser}), 

3113: and the observed optical afterglow of GRB 991216 

3114: at $\rm z=1.02\, .$ 

3115: The figure shows (from top to bottom) 1000 times the K-band results,

3116: 100 times the J-band, 10 times the I-band, the R-band, 1/10 of the V-band,

3117: 1/100 of the B-band and 1/1000 of the U-band.

3118: The contributions of the underlying galaxy and the

3119: expected SN1998bw-like 

3120: SN have been subtracted. In a CB-model fit, there is in this case

3121: some evidence for such a SN (DDD 2001).}

3122: \label{opt216}  

3123: \end{figure}

3124: 

3125: \clearpage

3126: 

3127: \begin{figure}[t]  

3128: \begin{tabular}{cc}  

3129: \hskip 2truecm  

3130: \vspace*{2cm} 

3131: \hspace*{-1.7cm}  

3132: \epsfig{file=ag13nr1216.eps, width=8cm} \\ 

3133: %\hskip 1truecm  

3134: \hspace*{.5cm}  

3135: \epsfig{file=ag14nr1216.eps, width=8cm} 

3136: \end{tabular}  

3137: \caption{Comparisons between our fitted CB model afterglow, 

3138: Eq.~(\ref{Fnuobser}), and the observed radio afterglow of GRB 991216.

3139: Upper panel: the light curve at 15 

3140: GHz. Lower panel: the light curve at 8.46 GHz.}  

3141: \label{figr121602}  

3142: \end{figure} 

3143: 

3144: 

3145: 

3146: \begin{figure}[t]  

3147: \begin{tabular}{cc}  

3148: \hskip 2truecm  

3149: \vspace*{2cm} 

3150: \hspace*{-1.7cm}  

3151: \epsfig{file=ag15nr1216.eps, width=8cm} \\ 

3152: %\hskip 1truecm  

3153: \hspace*{.5cm}  

3154: \epsfig{file=ag16nr1216.eps, width=8cm} 

3155: \end{tabular}  

3156: \caption{Comparisons between our fitted CB model afterglow, 

3157: Eq.~(\ref{Fnuobser}), and the observed radio afterglow of GRB 991216.

3158: Upper panel: the light curve at 4.86 

3159: GHz. Lower panel: the light curve at 1.43 GHz.}  

3160: \label{figr121603}  

3161: \end{figure} 

3162: 

3163: 

3164: \clearpage

3165: 

3166: \begin{figure}[t]  

3167: \begin{tabular}{cc}  

3168: \hskip 2.5truecm  

3169: \vspace*{2cm} 

3170: \hspace*{-2.7cm}  

3171: \epsfig{file=ag21nro1216.eps, width=8cm}

3172: \vspace*{-1.5cm}

3173: \\ 

3174: %\hskip 1truecm 

3175: \hspace*{-.2cm}  

3176: \epsfig{file=ag22nro1216.eps, width=8cm} 

3177: \end{tabular}  

3178: \caption{The spectrum of the AG of GRB 991216 from radio to optical

3179: frequencies. 

3180: Upper panel: in the time interval between 0.44 and 2 days after burst.

3181: Lower panel:  in the time interval between 2 and 4 days after burst. 

3182: In both cases the highest peaking curve

3183: corresponds to the earlier time.}  

3184: \label{rad-opt216}  

3185: \end{figure} 

3186: 

3187: 

3188: 

3189: \begin{figure}[t]  

3190: \begin{tabular}{cc}  

3191: \hskip 2.5truecm  

3192: \vspace*{2cm} 

3193: \hspace*{-2.7cm}  

3194: \epsfig{file=ag23nro1216.eps, width=8cm}

3195: \vspace*{-1.5cm}

3196: \\ 

3197: %\hskip 1truecm 

3198: \hspace*{-.2cm}  

3199: \epsfig{file=ag24nro1216.eps, width=8cm} 

3200: \end{tabular}  

3201: \caption{The spectrum of the AG of GRB 991216 from radio to optical

3202: frequencies. 

3203: Upper panel: in the time interval between 4 and 7 days after burst.

3204: Lower panel:  in the time interval between 7 and 13 days after burst. 

3205: In both cases the highest peaking curve

3206: corresponds to the earlier time.}  

3207: \label{rad-opt216b}  

3208: \end{figure} 

3209: 

3210: \clearpage

3211: 

3212: \begin{figure}[t]  

3213: \begin{tabular}{cc}  

3214: \hskip 2.5truecm  

3215: \vspace*{2cm} 

3216: \hspace*{-2.7cm}  

3217: \epsfig{file=ag25nro1216.eps, width=8cm}

3218: \vspace*{-1.5cm}

3219: \\ 

3220: %\hskip 1truecm 

3221: \hspace*{-.2cm}  

3222: \epsfig{file=ag26nro1216.eps, width=8cm} 

3223: \end{tabular}  

3224: \caption{The spectrum of the AG of GRB 991216 from radio to optical

3225: frequencies. 

3226: Upper panel: in the time interval between 13 and 30 days after burst.

3227: Lower panel:  in the time interval between  30 and 80 days after burst. 

3228: In both cases the highest peaking curve

3229: corresponds to the earlier time.}  

3230: \label{rad-opt216c}  

3231: \end{figure} 

3232: 

3233: %end216

3234: 

3235: \begin{figure}[t]  

3236: %\begin{tabular}{cc}  

3237: \hskip 0truecm   

3238: \epsfig{file=ag42nro1208.eps, width=8.5cm}  

3239: %\end{tabular}  

3240: \caption{Comparisons between our fitted CB model AG of GRB 991208, 

3241: at $\rm z=0.706$,

3242: Eq.~(\ref{Fnuobser}), with the observed optical data. 

3243: The figure shows (from top to bottom) 1000 times the K-band results,

3244: 100 times the J-band, 10 times the I-band, the R-band, 1/10 of the V-band,

3245: 1/100 of the B-band and 1/1000 of the U-band.

3246: The contribution of the underlying galaxy and associated

3247: supernova has been subtracted. 

3248: The contributions of the underlying galaxy and the

3249: expected SN1998bw-like 

3250: SN have been subtracted. In a CB-model fit, there is in this case

3251: strong evidence for such a SN (DDD 2001).

3252: }  

3253: \label{opt208}  

3254: \end{figure} 

3255: 

3256: \clearpage

3257: 

3258: \begin{figure}[t] 

3259: \begin{tabular}{cc} 

3260: \hskip 2truecm 

3261: \vspace*{2cm} 

3262: \hspace*{-1.7cm} 

3263: \epsfig{file=ag11nr1208.eps, width=8cm} \\ 

3264: %\hskip 1truecm 

3265: \hspace*{.5cm} 

3266: \epsfig{file=ag12nr1208.eps, width=8cm} 

3267: \end{tabular} 

3268: \caption{Comparisons between our fitted CB model afterglow, 

3269: Eq.~(\ref{Fnuobser}), 

3270: and the observed radio afterglow of GRB 991208. 

3271: Upper panel: the light curve at 100 GHz. 

3272: Lower panel: the light curve at 86.2 GHz.} 

3273: \label{figr120801} 

3274: \end{figure} 

3275:  

3276: %\clearpage

3277: 

3278: \begin{figure}[t] 

3279: \begin{tabular}{cc} 

3280: \hskip 2truecm 

3281: \vspace*{2cm} 

3282: \hspace*{-1.7cm} 

3283: \epsfig{file=ag13nr1208.eps, width=8cm} \\ 

3284: %\hskip 1truecm 

3285: \hspace*{.5cm} 

3286: \epsfig{file=ag14nr1208.eps, width=8cm} 

3287: \end{tabular} 

3288: \caption{Comparisons between our fitted CB model afterglow, 

3289: Eq.~(\ref{Fnuobser}), 

3290: and the observed radio afterglow of GRB 991208. 

3291: Upper panel: the light curve at 30 GHz. 

3292: Lower panel: the light curve at 22.5 GHz.} 

3293: \label{figr120802} 

3294: \end{figure} 

3295:  

3296: \clearpage

3297: 

3298: \begin{figure}[t] 

3299: \begin{tabular}{cc} 

3300: \hskip 2truecm 

3301: \vspace*{2cm} 

3302: \hspace*{-1.7cm} 

3303: \epsfig{file=ag15nr1208.eps, width=8cm} \\ 

3304: %\hskip 1truecm 

3305: \hspace*{.5cm} 

3306: \epsfig{file=ag16nr1208.eps, width=8cm} 

3307: \end{tabular} 

3308: \caption{Comparisons between our fitted CB model afterglow, 

3309: Eq.~(\ref{Fnuobser}), 

3310: and the observed radio afterglow of GRB 991208. 

3311: Upper panel: the light curve at 15 GHz. 

3312: Lower panel: the light curve at 8.46 GHz.} 

3313: \label{figr120803} 

3314: \end{figure} 

3315:  

3316: %\clearpage

3317:  

3318: \begin{figure}[t] 

3319: \begin{tabular}{cc} 

3320: \hskip 2truecm 

3321: \vspace*{2cm} 

3322: \hspace*{-1.7cm} 

3323: \epsfig{file=ag17nr1208.eps, width=8cm} \\ 

3324: %\hskip 1truecm 

3325: \hspace*{.5cm} 

3326: \epsfig{file=ag18nr1208.eps, width=8cm} 

3327: \end{tabular} 

3328: \caption{Comparisons between our fitted CB model afterglow, 

3329: Eq.~(\ref{Fnuobser}), 

3330: and the observed radio afterglow of GRB 991208. 

3331: Upper panel: the light curve at 4.86 GHz. 

3332: Lower panel: the light curve at 1.43 GHz.} 

3333: \label{figr120804}

3334: \end{figure} 

3335:  

3336: \clearpage

3337:  

3338: \begin{figure}[t] 

3339: \begin{tabular}{cc} 

3340: \hskip 2truecm 

3341: \vspace*{2cm} 

3342: \hspace*{-1.7cm} 

3343: \epsfig{file=ag21nro1208.eps, width=8cm} \\ 

3344: %\hskip 1truecm 

3345: \hspace*{.5cm} 

3346: \epsfig{file=ag22nro1208.eps, width=8cm} 

3347: \end{tabular} 

3348: \caption{The spectrum of the AG of GRB 991208 from radio to optical

3349: frequencies. 

3350: Upper panel: in the time interval between 2 and 5 days after burst.

3351: Lower panel: in the time interval between 5 and 10 days after burst.

3352: In both cases the highest peaking curve

3353: corresponds to the earlier time.} 

3354: \label{rad-opt208} 

3355: \end{figure} 

3356: 

3357:  

3358: %\clearpage

3359:  

3360: \begin{figure}[t] 

3361: \begin{tabular}{cc} 

3362: \hskip 2truecm 

3363: \vspace*{2cm} 

3364: \hspace*{-1.7cm} 

3365: \epsfig{file=ag23nro1208.eps, width=8cm} \\ 

3366: %\hskip 1truecm 

3367: \hspace*{.5cm} 

3368: \end{tabular} 

3369: \caption{The spectrum of the AG of GRB 991208 from radio to optical

3370: frequencies in the time interval between 10 and 14.3 days.

3371: The highest peaking curve corresponds to the earlier time.} 

3372: \label{rad-opt208b} 

3373: \end{figure} 

3374: 

3375: \clearpage

3376: 

3377: %418starts

3378: 

3379:  \begin{figure}[t]  

3380: %\begin{tabular}{cc}  

3381: \hskip 0truecm   

3382: \epsfig{file=ag42nro0418.eps, width=8.5cm}  

3383: %\end{tabular}  

3384: \caption{Comparisons between our fitted CB model AG of GRB 000418, 

3385: at $\rm z=1.118$,

3386: Eq.~(\ref{Fnuobser}), with the observed optical data. 

3387: The figure shows (from top to bottom) 1000 times the K-band results,

3388: 100 times the J-band, 10 times the I-band, the R-band, 1/10 of the V-band,

3389: 1/100 of the B-band and 1/1000 of the U-band.

3390: The contributions of the underlying galaxy and the

3391: expected SN1998bw-like 

3392: SN have been subtracted. In a CB-model fit, there is in this case

3393: strong evidence for such a SN (DDD 2001).}

3394: 

3395: \label{opt418}  

3396: \end{figure} 

3397:  

3398: \begin{figure}[t] 

3399: \begin{tabular}{cc} 

3400: \hskip 2truecm 

3401: \vspace*{2cm} 

3402: \hspace*{-1.7cm} 

3403: \epsfig{file=ag11nr0418.eps, width=8cm} \\ 

3404: %\hskip 1truecm 

3405: \hspace*{.5cm} 

3406: \epsfig{file=ag12nr0418.eps, width=8cm} 

3407: \end{tabular} 

3408: \caption{Comparisons between our fitted CB model afterglow, 

3409: Eq.~(\ref{Fnuobser}), 

3410: and the observed radio afterglow of GRB 000418. 

3411: Upper panel: the light curve at 22.46 GHz. 

3412: Lower panel: the light curve at 15 GHz.} 

3413: \label{figr041801} 

3414: \end{figure} 

3415: 

3416: \clearpage

3417:  

3418: \begin{figure}[t] 

3419: \begin{tabular}{cc} 

3420: \hskip 2truecm 

3421: \vspace*{2cm} 

3422: \hspace*{-1.7cm} 

3423: \epsfig{file=ag13nr0418.eps, width=8cm} \\ 

3424: %\hskip 1truecm 

3425: \hspace*{.5cm} 

3426: \epsfig{file=ag15nr0418.eps, width=8cm} 

3427: \end{tabular} 

3428: \caption{Comparisons between our fitted CB model afterglow, 

3429: Eq.~(\ref{Fnuobser}), 

3430: and the observed radio afterglow of GRB 000418. 

3431: Upper panel: the light curve at 8.46 GHz. 

3432: Lower panel: the light curve at 4.86 GHz.

3433: } 

3434: \label{figr041802} 

3435: \end{figure} 

3436:  

3437: 

3438: 

3439: \begin{figure}[t] 

3440: \begin{tabular}{cc} 

3441: \hskip 2truecm 

3442: \vspace*{2cm} 

3443: \hspace*{-1.7cm} 

3444: \epsfig{file=ag21nro0418.eps, width=8cm} \\ 

3445: %\hskip 1truecm 

3446: \hspace*{.5cm} 

3447: \epsfig{file=ag22nro0418.eps, width=8cm} 

3448: \end{tabular} 

3449: \caption{The spectrum of the AG of GRB 000418 from radio to optical

3450: frequencies.

3451: Upper panel: in the time interval

3452: between 9.5 and 30 days after burst.

3453: Lower panel: in the time interval

3454: between 30 and 100 days after burst.

3455: In both cases the highest peaking curve

3456: corresponds to the earlier time.} 

3457: \label{rad-opt418} 

3458: \end{figure} 

3459: 

3460: %418ends

3461: 

3462: \clearpage

3463: 

3464: %510starts

3465: 

3466:  \begin{figure}[t]  

3467: %\begin{tabular}{cc}  

3468: \hskip 0truecm   

3469: \epsfig{file=ag42nro0510.eps, width=8.5cm}  

3470: %\end{tabular}  

3471: \caption{

3472: Comparisons between our fitted CB model AG of GRB 990510,

3473: at $\rm z=1.619$,

3474: Eq.~(\ref{Fnuobser}), with the observed optical data.

3475: The figure shows (from top to bottom) 1000 times the K-band results,

3476: 100 times the J-band, 10 times the I-band, the R-band, 1/10 of the V-band,

3477: 1/100 of the B-band and 1/1000 of the U-band.

3478: The contribution of the underlying galaxy and the

3479: (in this case unobservable) associated

3480: supernova  has been subtracted. }  

3481: \label{opt510}  

3482: \end{figure} 

3483: 

3484:  

3485: \begin{figure}[t] 

3486: \begin{tabular}{cc} 

3487: \hskip 2truecm 

3488: \vspace*{2cm} 

3489: \hspace*{-1.7cm} 

3490: \epsfig{file=ag11nr0510.eps, width=8cm} \\ 

3491: %\hskip 1truecm 

3492: \hspace*{.5cm} 

3493: \epsfig{file=ag12nr0510.eps, width=8cm} 

3494: \end{tabular} 

3495: \caption{Comparisons between our fitted CB model afterglow, 

3496: Eq.~(\ref{Fnuobser}), 

3497: and the observed radio afterglow of GRB 990510. 

3498: Upper panel: the light curve at 13.68 GHz. 

3499: Lower  panel: the light curve at 8.6-8.7 GHz.} 

3500: \label{figr051001} 

3501: \end{figure} 

3502:  

3503: \clearpage

3504:  

3505: \begin{figure}[t] 

3506: \begin{tabular}{cc} 

3507: \hskip 2truecm 

3508: \vspace*{2cm} 

3509: \hspace*{-1.7cm} 

3510: \epsfig{file=ag13nr0510.eps, width=8cm} \\ 

3511: %\hskip 1truecm 

3512: \hspace*{.5cm} 

3513: \epsfig{file=ag21nro0510.eps, width=8cm} 

3514: \end{tabular} 

3515: \caption{Comparisons between our fitted CB model afterglow, 

3516: Eq.~(\ref{Fnuobser}), 

3517: and the observed radio afterglow of GRB 990510. 

3518: Upper panel: the light curve at 4.86 GHz. 

3519: Lower panel: the spectrum from radio to optical frequencies in the 

3520: time interval between 1 and 6 days after burst. 

3521: The highest peaking curve corresponds to the earlier time.}

3522: \label{figr051002} 

3523: \end{figure} 

3524: 

3525: 

3526: \begin{figure}[t] 

3527: \begin{tabular}{cc} 

3528: \hskip 2truecm 

3529: \vspace*{2cm} 

3530: \hspace*{-1.7cm} 

3531: \epsfig{file=ag22nro0510.eps, width=8cm} \\ 

3532: %\hskip 1truecm 

3533: \hspace*{.5cm} 

3534: \epsfig{file=ag23nro0510.eps, width=8cm} 

3535: \end{tabular} 

3536: \caption{The spectrum of the AG of GRB 990510 from radio to optical

3537: frequencies.

3538: Upper panel: in the time interval

3539: between 6 and 20 days after burst. 

3540: Lower panel: in the time interval between 20 and 40  days after burst.

3541: In both cases the highest peaking curve

3542: corresponds to the earlier time.}

3543: \label{rad-opt510} 

3544: \end{figure} 

3545:  

3546: %510ends

3547: 

3548: %123starts

3549: 

3550:   \begin{figure}[t]  

3551: %\begin{tabular}{cc}  

3552: \hskip 0truecm   

3553: \epsfig{file=ag42nro0123.eps, width=8.5cm}  

3554: %\end{tabular}  

3555: \caption{

3556: Comparisons between our fitted CB model AG of GRB 990123, 

3557: at $\rm z=1.600$,

3558: Eq.~(\ref{Fnuobser}), and the observed optical data. 

3559: The figure shows (from top to bottom) 1000 times the K-band results,

3560: 100 times the J-band, 10 times the I-band, the R-band, 1/10 of the V-band,

3561: 1/100 of the B-band and 1/1000 of the U-band.

3562: The contributions of the underlying galaxy and 

3563: an expected (but,  in this case, unobservable) SN1998bw-like 

3564: SN have been subtracted. }

3565: \label{opt123}  

3566: \end{figure} 

3567:  

3568: 

3569: \begin{figure}[t] 

3570: \begin{tabular}{cc} 

3571: \hskip 2truecm 

3572: \vspace*{2cm} 

3573: \hspace*{-1.7cm} 

3574: \epsfig{file=ag13nr0123.eps, width=8cm} \\ 

3575: %\hskip 1truecm 

3576: \hspace*{.5cm} 

3577: \epsfig{file=ag14nr0123.eps, width=8cm} 

3578: \end{tabular} 

3579: \caption{Comparisons between our fitted CB model afterglow, 

3580: Eq.~(\ref{Fnuobser}), 

3581: and the observed radio afterglow of GRB 990123. 

3582: Upper panel: the light curve at 15 GHz. 

3583: Lower  panel: the light curve at 8.46 GHz.} 

3584: \label{figr012301} 

3585: \end{figure} 

3586:  

3587: %\clearpage

3588:  

3589: \begin{figure}[t] 

3590: \begin{tabular}{cc} 

3591: \hskip 2truecm 

3592: \vspace*{2cm} 

3593: \hspace*{-1.7cm} 

3594: \epsfig{file=ag21nro0123.eps, width=8cm} \\ 

3595: %\hskip 1truecm 

3596: \hspace*{.5cm} 

3597: \epsfig{file=ag22nro0123.eps, width=8cm} 

3598: \end{tabular} 

3599: \caption{The spectrum of the AG of GRB 990123 

3600: from radio to optical frequencies.

3601: Upper panel: in the time interval between 0.1 and 1 day after burst.

3602: Lower panel: in the time interval between 1 and 3 days after burst.

3603: In both cases the highest peaking curve

3604: corresponds to the earlier time.} 

3605: \label{rad-opt123a} 

3606: \end{figure} 

3607: 

3608: \begin{figure}[t] 

3609: \begin{tabular}{cc} 

3610: \hskip 2truecm 

3611: \vspace*{2cm} 

3612: \hspace*{-1.7cm} 

3613: \epsfig{file=ag23nro0123.eps, width=8cm} \\ 

3614: %\hskip 1truecm 

3615: \hspace*{.5cm} 

3616: \epsfig{file=ag24nro0123.eps, width=8cm} 

3617: \end{tabular} 

3618: \caption{The spectrum of the AG of GRB 990123

3619: from radio to optical frequencies.

3620: Upper panel: in the time interval between 3 and 10 days after burst.

3621: Lower panel: in the time interval between 10 and 20 days after burst.

3622: In both cases the highest peaking curve

3623: corresponds to the earlier time.} 

3624: \label{rad-opt123b} 

3625: \end{figure} 

3626: 

3627: 

3628: 

3629: %123ends

3630:  

3631: 

3632:  \begin{figure}[t]  

3633: %\begin{tabular}{cc}  

3634: \hskip 0truecm   

3635: \epsfig{file=ag42nro0508bb.eps, width=8.5cm}  

3636: %\end{tabular}  

3637: \caption{Comparisons between our fitted CB model AG of GRB 970508, 

3638: at $\rm z=0.835$,

3639: Eq.~(\ref{Fnuobser}), with the observed optical data. 

3640: The figure shows (from top to bottom), 

3641: 10 times the I-band, the R-band, 1/10 of the V-band and

3642: 1/100 of the B-band.

3643: The contributions of the underlying galaxy and the

3644: expected SN1998bw-like 

3645: SN have been subtracted. In a CB-model fit, there is in this case

3646: strong evidence for such a SN (DDD 2001).}

3647: \label{opt508}  

3648: \end{figure} 

3649:  

3650:  

3651: \begin{figure}[t]  

3652: \begin{tabular}{cc}  

3653: \hskip 2truecm  

3654: \vspace*{2cm} 

3655: \hspace*{-1.7cm}  

3656: \epsfig{file=ag11nr0508.eps, width=8cm} \\ 

3657: %\hskip 1truecm  

3658: \hspace*{.5cm}  

3659: \epsfig{file=ag12nr0508.eps, width=8cm} 

3660: \end{tabular}  

3661: \caption{Comparisons between our fitted CB model afterglow, 

3662: Eq.~(\ref{Fnuobser}), and the observed radio afterglow of GRB 970508.

3663: Upper panel: the light curve at 8.46 

3664: GHz. Lower panel: the light curve at 4.86 GHz.}  

3665: \label{figr050801}  

3666: \end{figure} 

3667: 

3668: \clearpage 

3669: 

3670: \begin{figure}[t]  

3671: \begin{tabular}{cc}  

3672: \hskip 2truecm  

3673: \vspace*{2cm} 

3674: \hspace*{-1.7cm}  

3675: \epsfig{file=ag13nr0508.eps, width=8cm} \\ 

3676: %\hskip 1truecm  

3677: \hspace*{.5cm}  

3678: \epsfig{file=ag21nro0508.eps, width=8cm} 

3679: \end{tabular}  

3680: \caption{Comparisons between our fitted CB model afterglow, 

3681: Eq.~(\ref{Fnuobser}), and the observed radio afterglow of GRB 970508.

3682: Upper panel: the light curve at 1.43 GHz. 

3683: Lower panel: the spectral behaviour in the time

3684: interval  between 0.12 and 6 

3685: days after burst. The highest peaking curve

3686: corresponds to the earlier time.}

3687: \label{figr050802} 

3688: \end{figure} 

3689:  

3690: 

3691:  

3692: \begin{figure}[t]  

3693: \begin{tabular}{cc}  

3694: \hskip 2truecm  

3695: \vspace*{2cm} 

3696: \hspace*{-1.7cm}  

3697: \epsfig{file=ag22nro0508.eps, width=8cm} \\ 

3698: %\hskip 1truecm  

3699: \hspace*{.5cm}  

3700: \epsfig{file=ag23nro0508.eps, width=8cm} 

3701: \end{tabular}  

3702: \caption{The spectrum of the AG of GRB 970508

3703: from radio to optical frequencies.

3704: Upper panel: in the time interval between 6 and 20 days after burst.

3705: Lower panel: in the time interval between 20 and 40 days after burst.

3706: In both cases the highest peaking curve

3707: corresponds to the earlier time.} 

3708: \label{rad-opt508a} 

3709: \end{figure} 

3710: 

3711: \clearpage

3712:  

3713: \begin{figure}[t]  

3714: \begin{tabular}{cc}  

3715: \hskip 2truecm  

3716: \vspace*{2cm} 

3717: \hspace*{-1.7cm}  

3718: \epsfig{file=ag24nro0508.eps, width=8cm} \\ 

3719: %\hskip 1truecm  

3720: \hspace*{.5cm}  

3721: \epsfig{file=ag25nro0508.eps, width=8cm} 

3722: \end{tabular}  

3723: \caption{The spectrum of the AG of GRB 970508

3724: from radio to optical frequencies.

3725: Upper panel: in the time interval  between 40 and 100 days after burst.

3726: Lower panel: in the time interval between 100 and 470 days after burst. 

3727: In both cases the highest peaking curve

3728: corresponds to the earlier time.}

3729: \label{rad-opt508b} 

3730: \end{figure} 

3731:  

3732: 

3733: %\clearpage

3734:  

3735:  

3736: \begin{figure}[t] 

3737: \begin{tabular}{cc} 

3738: \hskip 2truecm 

3739: \vspace*{2cm} 

3740: \hspace*{-1.7cm} 

3741: \epsfig{file=ag11nr0425.eps, width=8cm} \\ 

3742: %\hskip 1truecm 

3743: \hspace*{.5cm} 

3744: \epsfig{file=ag12nr0425.eps, width=8cm} 

3745: \end{tabular} 

3746: \caption{Comparisons between our fitted CB model afterglow of

3747: GRB 980425 at z=0.0085, 

3748: Eq.~(\ref{Fnuobser}), 

3749: and its observed radio afterglow. 

3750: Upper panel: the light curve at 8.64 GHz. 

3751: Lower panel: the light curve at 4.80 GHz.} 

3752: \label{figr042501} 

3753: \end{figure} 

3754:  

3755: \clearpage

3756: 

3757: \begin{figure}[t] 

3758: \begin{tabular}{cc} 

3759: \hskip 2truecm 

3760: \vspace*{2cm} 

3761: \hspace*{-1.7cm} 

3762: \epsfig{file=ag13nr0425.eps, width=8cm} \\ 

3763: %\hskip 1truecm 

3764: \hspace*{.5cm} 

3765: \epsfig{file=ag14nr0425.eps, width=8cm} 

3766: \end{tabular} 

3767: \caption{Comparisons between our fitted CB model afterglow, 

3768: Eq.~(\ref{Fnuobser}), 

3769: and the observed radio afterglow of GRB 980425. 

3770: Upper panel: the light curve at 2.49 GHz. 

3771: Lower panel: the light curve at 1.38 GHz.} 

3772: \label{figr042502} 

3773: \end{figure} 

3774:  

3775: %\clearpage

3776: 

3777: \begin{figure}[t] 

3778: \begin{tabular}{cc} 

3779: \hskip 2truecm 

3780: \vspace*{2cm} 

3781: \hspace*{-1.7cm} 

3782: \epsfig{file=ag21nr0425.eps, width=8cm} \\ 

3783: %\hskip 1truecm 

3784: \hspace*{.5cm} 

3785: \epsfig{file=ag22nr0425.eps, width=8cm} 

3786: \end{tabular} 

3787: \caption{Comparison between the observed spectrum of the AG of GRB 980425 in 

3788: the radio band and the fitted CB model spectrum. 

3789: Upper panel: in the time interval  between 2.5 and 20 days after burst.

3790: Lower panel: in the time interval between 20 and 40 days after burst. 

3791: The highest peaking curve in the upper pannel corresponds to the later 

3792: time and in the lower panel to the earlier time.}

3793: \label{figr042503} 

3794: \end{figure} 

3795: 

3796: 

3797: 

3798:  

3799: \begin{figure}[t] 

3800: \begin{tabular}{cc} 

3801: \hskip 2truecm 

3802: \vspace*{2cm} 

3803: \hspace*{-1.7cm} 

3804: \epsfig{file=ag23nr0425.eps, width=8cm} \\ 

3805: %\hskip 1truecm 

3806: \hspace*{.5cm} 

3807: \epsfig{file=ag24nr0425.eps, width=8cm} 

3808: \end{tabular} 

3809: \caption{Comparison between the observed spectrum of the AG of GRB 980425 in 

3810: the radio band and the fitted CB model spectrum. 

3811: Upper panel: in the time interval  between 40 and 70 days after burst.

3812: Lower panel: in the time interval between 70 and 150 days after burst. 

3813: In both cases the highest peaking curve

3814: corresponds to the earlier time.}

3815: \label{figr042504} 

3816: \end{figure} 

3817: 

3818: %\clearpage

3819:  

3820: \begin{figure}[t]  

3821: \hskip 0truecm   

3822: \epsfig{file=ag27nr0425.eps, width=8cm}  

3823: \caption{The late spectrum of GRB 980425 in the time interval 

3824: between 150 and 759 days after burst. The highest 

3825: peaking line corresponds to the earlier time. The late isolated

3826: point is the last optical observation, which must correspond

3827: to SN1998bw and not to the AG of the CB of GRB 980425.

3828: }  

3829: \label{late425}  

3830: \end{figure}

3831: 

3832: \begin{figure}[t]  

3833: \hskip 0truecm   

3834: \vskip 1cm

3835: \epsfig{file=X425new.eps, width=8cm}  

3836: \caption{The X-ray afterglow of GRB 980425.}  

3837: \label{X425}  

3838: \end{figure}

3839: 

3840: \begin{figure}[t]  

3841: \hskip 0truecm   

3842: \epsfig{file=425displace.eps, width=8cm}  

3843: \caption{The predicted angular separation of SN1998bw

3844: and GRB 980425, in milliarcseconds, as a function of time.

3845: }  

3846: \label{superluminal}  

3847: \end{figure}

3848: 

3849: 

3850: \end{document}

3851: 

3852:  

3853: \begin{figure}[t] 

3854: \begin{tabular}{cc} 

3855: \hskip 2truecm 

3856: \vspace*{-0cm} 

3857: \hspace*{-1.7cm} 

3858: \epsfig{file=ag25nr0425.eps, width=8cm} \\ 

3859: %\hskip 1truecm 

3860: \hspace*{.5cm} 

3861: \end{tabular} 

3862: \caption{Comparison between the observed spectrum of the AG of GRB 980425 in 

3863: the radio band and the fitted CB model spectrum 

3864: in the time interval between 150 and 500 days after burst. The highest 

3865: peaking line corresponds to the earlier time.}

3866: \label{figr042505} 

3867: \end{figure}  

3868: 

3869: 

3870: 

3871: 

3872: 

3873:  

3874: \begin{figure}[t] 

3875: \begin{tabular}{cc} 

3876: \hskip 2truecm 

3877: \vspace*{2cm} 

3878: \hspace*{-1.7cm} 

3879: \epsfig{file=ag25nr0425.eps, width=8cm} \\ 

3880: %\hskip 1truecm 

3881: \hspace*{.5cm} 

3882: \epsfig{file=ag26nr0425.eps, width=8cm} 

3883: \end{tabular} 

3884: \caption{Comparison between the observed spectrum of the AG of GRB 980425 in 

3885: the radio band and the fitted CB model spectrum. 

3886: Upper panel: in the time interval between 150 and 500 days after burst.

3887: Lower panel: in the time interval between 500 and 800 days after burst. 

3888: In both cases the highest peaking curve

3889: corresponds to the earlier time.}

3890: \label{figr042505} 

3891: \end{figure}  

3892: 

3893: 

3894: 

3895: \begin{figure}[t]  

3896: \begin{tabular}{cc}  

3897: \hskip 2truecm  

3898: \vspace*{2cm} 

3899: \hspace*{-1.7cm}  

3900: \epsfig{file=ag11nr1216.eps, width=8cm} \\ 

3901: %\hskip 1truecm  

3902: \hspace*{.5cm}  

3903: \epsfig{file=ag12nr1216.eps, width=8cm} 

3904: \end{tabular}  

3905: \caption{Comparisons between our fitted CB model afterglow, 

3906: Eq.~(\ref{Fnuobser}), and the observed radio afterglow of GRB 991216.

3907: Upper panel: the light curve at 350 

3908: GHz. Lower panel: the light curve at 100 GHz.}  

3909: \label{figr121601}  

3910: \end{figure} 

3911: 

3912: