1: %\documentclass[preprint2,epsf]{aastex}

2: %\documentstyle[12pt,aasms4,epsf,rotate]{article}

3: \documentstyle[11pt,aaspp4,epsf,rotate]{article}

4: %\documentstyle[aas2pp4]{article} 

5: %\tighten 

6: %\eqsecnum 

7: %\received{}

8: %\accepted{} 

9: %\journalid{}{} 

10: %\articleid{}{} 

11: \newcommand\D{{\cal D}}

12: \def\psim{\lower.5ex\hbox{$\; \buildrel \propto \over\sim \;$}}

13: \def\gtrsim{\lower.5ex\hbox{$\; \buildrel > \over\sim \;$}}

14: \def\lesssim{\lower.5ex\hbox{$\; \buildrel < \over\sim \;$}}

15: \def\gm{\gamma_m} 

16: \def\g2{\gamma_2} 

17: \def\tT{\tau_T}

18: \def\elnumax{L_{\nu,{\rm max}}} 

19: \def\e{{\epsilon}}

20: \def\ag{\alpha_{\gamma}}

21: %\slugcomment{}

22: \begin{document}

23: 

24: \title{Flash-Heating of Circumstellar Clouds by Gamma Ray Bursts}

25: 

26: \author{Charles D.  Dermer\altaffilmark{1} \& Markus

27: B\"ottcher\altaffilmark{2,3}}

28: 

29: \altaffiltext{1}{E.  O.  Hulburt Center for Space Research, Code 7653,

30: Naval Research Laboratory, Washington, DC 20375-5352}

31: \altaffiltext{2}{Department of Space Physics and Astronomy, Rice

32: University, Houston, TX 77005-1892} \altaffiltext{3}{Chandra Fellow}

33: 

34: 

35: \begin{abstract}

36: 

37: The blast-wave model for gamma-ray bursts (GRBs) has been called into

38: question by observations of spectra from GRBs that are harder than can

39: be produced through optically thin synchrotron emission.  If GRBs originate 

40: from the collapse of massive stars, then circumstellar clouds near burst

41: sources will be illuminated by intense $\gamma$ radiation, and the

42: electrons in these clouds will be rapidly scattered to energies as

43: large as several hundred keV.  Low-energy photons that subsequently

44: pass through the hot plasma will be scattered to higher energies, thus

45: hardening the intrisic spectrum.  This effect resolves the

46: ``line-of-death'' objection to the synchrotron shock model. Illuminated

47: clouds near GRBs will form relativistic plasmas containing large

48: numbers of electron-positron pairs that can be detected within $\sim$

49: 1-2 days of the explosion before expanding and dissipating.  Localized

50: regions of pair annihilation radiation in the Galaxy would reveal past

51: GRB explosions.  \end{abstract}

52: 

53: \keywords{gamma rays:  bursts -- massive stars -- nonthermal radiation

54: processes }

55: 

56: \section{Introduction}

57: 

58: The identification of flaring and fading X-ray, optical and radio

59: counterparts to gamma-ray burst (GRB) sources (e.g., Costa et al.\

60: \markcite{cea97}1997; van Paradijs et al.\ \markcite{vpea97}1997;

61: Djorgovski et al.\ \markcite{dea97}1997; Frail et al.\

62: \markcite{fea97}1997), and the large energy releases implied by

63: redshift measurements, find a consistent explanation in an expanding

64: relativistic blast-wave model (Paczy\'nski \& Rhoads

65: \markcite{pr93}1993; M\'esz\'aros \& Rees \markcite{mr97}1997).  As a

66: result of Beppo-SAX and optical follow-on observations, the redshifts

67: of about one dozen GRBs with durations greater than $\sim 1$ s have

68: been measured.  The distribution of redshifts is broad and centered

69: near $z\sim 1$, corresponding to the cosmological epoch of active star

70: formation (Hogg \& Fruchter \markcite{hf99}1999).  GRBs are extremely

71: luminous and energetic at hard X-ray and $\gamma$-ray energies.  The

72: degree of GRB collimation is unknown, but peak directional

73: $\gamma$-ray luminosities and energy releases as large as $\partial

74: L/\partial \Omega\simeq 3\times 10^{51}$ ergs (s-sr)$^{-1}$ and $\partial

75: E/\partial \Omega \simeq 3\times 10^{53}$ ergs sr$^{-1}$, respectively,

76: have been measured (Kulkarni et al.\ \markcite{kea99}1999).  Less

77: powerful GRBs and less luminous episodes during the GRB produce

78: smaller $\gamma$-ray powers, but the apparent isotropic $\gamma$-ray

79: luminosities from typical GRBs could regularly reach values exceeding

80: $10^{50}L_{50}$ ergs s$^{-1}$ with $L_{50} \sim 1$, with some GRBs

81: reaching $L_{50} > 10^2$.  Because the energy radiated in $\gamma$

82: rays is less than the total energy released by a GRB, the apparent

83: isotropic energy release of GRB sources could often reach values of

84: $10^{54}E_{54}$ ergs, with $E_{54} \sim$1.

85: 

86: Cosmological gamma-ray burst and afterglow observations are best

87: explained through the fireball/blast-wave model, where the deposition

88: of large quantities of energy into a small region yields a fireball

89: that expands until it reaches a relativistic speed determined by the

90: amount of baryons mixed into the fireball (see, e.g., Piran

91: \markcite{tp99}1999 for a review).  Nonthermal synchrotron radiation

92: from energetic electrons in the relativistic blast wave is thought to

93: account for the origin of the prompt $\gamma$-ray emission and

94: afterglow radiation.  This paradigm has been called into question,

95: however, by observations of very hard X-ray emission during the prompt

96: $\gamma$-ray luminous phase of a significant number of GRBs (Crider et al.\

97: \markcite{crider97}1997; Preece et al.\ \markcite{pea98}1998).  Photon

98: fluxes $\phi(\e ) \propto \e^{-\alpha_X}$ with $\alpha_X \sim$ 0,

99: where $\e = h\nu/m_ec^2$ is the dimensionless photon energy, have been

100: observed in 5-10\% of GRBs that are bright enough to permit spectral

101: analysis.  This strongly contradicts the optically-thin synchrotron 

102: shock model, which predicts that only radiation spectra with 

103: $\alpha_X \geq 2/3$ can emerge from the blast wave. In view 

104: of the severity of this challenge to the model, these observations 

105: have been termed the ``line-of-death" to the synchrotron shock model.  

106: Possible explanations for this phenomeonon involve photoelectric 

107: absorption by optically thick cold matter (Liang \& Kargatis

108: \markcite{lk94}1994; Brainerd \markcite{jb94}1994; B\"ottcher 

109: et al.\ \markcite{bea99}1999), synchrotron self-absorption 

110: (Crider \& Liang \markcite{cl99}1999, Granot, Piran, \& Sari

111: \markcite{granot00}2000, Lloyd \& Petrosian \markcite{lp99}1999), 

112: Compton scattering (Liang \markcite{liang97}1997; Liang et al.\ 

113: \markcite{lcbs99}1999), or the existence of a

114: pair-photosphere (M\'esz\'aros \& Rees \markcite{mr00}2000) within the

115: blast wave.  Except for the last model cited, these explanations

116: are inconsistent with the standard synchrotron shock model.  Here we offer a

117: solution to this problem that is consistent with the standard model

118: and recent observations pointing to a massive star origin of GRBs.

119: 

120: \section{Massive Star Origin of GRBs}

121: 

122: Considerable evidence linking the sources of GRBs with star-forming

123: regions has recently been obtained (e.g., Lamb \markcite{lamb99}1999).

124: For example, the associated host galaxies have blue colors, consistent

125: with galaxy types that are undergoing active star formation.  GRB

126: counterparts are found within the optical radii and central regions of

127: the host galaxies (e.g.  Bloom et al.\ \markcite{bea99a}1999a), rather

128: than far outside the galaxies' disks, as might be expected in a

129: scenario of merging neutron stars and black holes (Narayan, Paczy\'nski,

130: \& Piran \markcite{npp92}1992).  Lack of optical counterparts in some

131: GRBs could be due to extreme reddening from large quantities of gas

132: and dust in the host galaxy.  This, together with the appearance of

133: supernova-like emissions in the late time optical decay curves of a

134: few GRBs (e.g., Bloom et al.\ \markcite{bea99b}1999b) and weak X-ray

135: evidence for Fe K$_\alpha$-line signatures (Piro et al.\

136: \markcite{pea99}1999), supports a massive star hypernova/collapsar

137: (Woosley \markcite{sw93}1993; Paczy\'nski \markcite{bp98}1998)

138: origin for the long duration gamma-ray bursters.

139: 

140: The observations thus favor a model for GRBs involving the collapse of

141: the core of a $\gtrsim 30 M_\odot$ star to a black hole, with the

142: collapse events producing fireballs and relativistic outflows with

143: large directed energy releases.  Earlier treatments of the blast-wave

144: model considered systems where the density of the surrounding medium

145: is either uniform or monotonically decreasing as a result of a

146: circumstellar medium formed by a hot stellar wind (M\'esz\'aros, Rees,

147: \& Wijers \markcite{mrw98}1998).  Until recently (Chevalier \&

148: Li \markcite{cl99}1999; Li \& Chevalier \markcite{lc99}1999),

149: less attention has been paid to the actual 

150: environment found in the vicinity of massive stars.  For this

151: we consider $\eta$ Carinae (Davidson \& Humphreys

152: \markcite{dh97}1997), the best-studied massive star that might

153: correspond to a GRB progenitor.  It is an evolved star with

154: present-day mass $\geq 90 M_\odot$, distance of $2300\pm 200$ pc, and

155: lifetime of about 3 million years.  It anisotropically ejects mass at

156: a current rate of $\lesssim 0.003 M_\odot$ yr$^{-1}$ to form its

157: unusual ``homonculus nebula."  Several Solar masses of material

158: surround $\eta$ Carinae.  In the immediate vicinity of the central

159: star, dense clouds of slow-moving gas with radii $r \sim 10^{15}$ cm

160: and densities between $10^7$ and $10^{10}$ cm$^{-3}$ were discovered

161: with speckle techniques (Hofman \& Weigelt \markcite{hw88}1988) and

162: confirmed with high-resolution HST observations (Davidson et al.\

163: \markcite{dea95}1995).  This material, moving with speeds of $\sim 50$

164: km s$^{-1}$, is apparently ejected nonuniformly from the equatorial

165: zone, but may remain trapped by the gravitational field of the star.

166: Inferences (Davidson \& Humphreys \markcite{dh97}1997) from [FeII]

167: observations suggest that $\gtrsim 0.02 M_\odot$ of gas are contained

168: within $\sim 2\times 10^{16}$ cm, implying a volume-averaged gas

169: density $\gtrsim 7\times 10^5$ cm$^{-3}$.  Model results (B\"ottcher

170: \markcite{mb99}1999) imply that a dense ($n \sim 10^{12}$ cm$^{-3}$)

171: torus of gas at mean distance $d\sim 2\times 10^{15}$ cm and with a

172: 10-fold enhancement of Fe relative to Solar abundance is required to

173: explain the Fe K$_\alpha$ emission weakly detected from GRB 970508

174: with Beppo-SAX (Piro et al.\ \markcite{pea99}1999).

175: 

176: Guided by the optical observations of $\eta$ Carinae, we assume that

177: the volume-averaged density of gas at $d \lesssim 10^{16}$ cm of a GRB

178: source is $\langle n \rangle = 10^6 n_6$ cm$^{-3}$.  Dense clouds of

179: radius $10^{15}r_{15}$ cm and radial Thomson depths $\tT = n_c\sigma_T

180: r$ are assumed to be embedded within this region, so that the mean

181: density of particles in a cloud is $n_c = 1.5\times 10^9 \tT /r_{15}$

182: cm$^{-3}$.  Thus $\tT \sim 1$ clouds located very close to a GRB

183: source are consistent with the observations of dense blobs near $\eta$

184: Carinae.  The deceleration length scale of a blast wave with initial

185: Lorentz factor $\Gamma_0 = 100\Gamma_2$ in a uniform medium is $x_d =

186: (3E_0/4\pi \Gamma_0^2 \langle n \rangle m_pc^2)^{1/3} = 2.5\times

187: 10^{15}(E_{54}/\Gamma_2^2 n_6)^{1/3}$ cm; hence the blast wave would

188: emit a significant fraction of its energy before reaching distances of

189: $\sim 10^{16}$ cm.  The deceleration time, which corresponds to the

190: duration of the prompt $\gamma$-ray luminous phase of a GRB in the

191: external shock model (Rees \& M\'esz\'aros \markcite{rm92}1992), is

192: $t_d = (1+z) x_d/(c \Gamma_0^2) \cong 8(1+z) (E_{54}/\Gamma_2^8

193: n_6)^{1/3}$ s.  These parameters are not unique, and we expect that

194: GRBs display a wide range of energies, Lorentz factors and surrounding

195: mean densities that could accommodate the diverse range of GRB

196: observations.

197: 

198: \section{Blast-Wave/Cloud Interaction}

199: 

200: A wave of photons impinging on a cloud located $10^{16} d_{16}$ cm

201: from the explosion center will photoionize and Compton-scatter the

202: ambient electrons to energies characteristic of the incident $\gamma$

203: rays (Madau \& Thompson \markcite{mt99}1999).  The $\gamma$-ray photon

204: front has a width of $\sim$10-100 lt-s, corresponding to the duration

205: of the GRB, whereas the plasma cloud has a width of $\sim 3\times 10^4

206: r_{15}$ lt-s, so that the radiation effects must be treated locally.

207: The radiation force driving the electrons outward is balanced by

208: strong electrostatic forces from the more massive protons and ions

209: that anchor the system until the net impulse is sufficient to drive

210: the entire plasma cloud outward.  The Compton back-scattered photons

211: provide targets for successive waves of incident GRB photons through

212: $\gamma\gamma$ pair-production interactions (Thompson \& Madau

213: \markcite{tm99}1999).  Higher-energy photons are preferentially

214: attenuated, forming an additional injection source of $\gtrsim 1$ MeV

215: electron-positron pairs.  The nonthermal electrons and pairs will

216: Compton scatter successive waves of photons, thereby modifying the

217: incident spectrum.  The pairs, no longer bound by electrostatic

218: attraction with the ions, will be driven outward by both radiation

219: forces and restoring electrostatic fields to form a mildly

220: relativistic pair wind passing through the more slowly moving normal

221: plasma.  Shortly after the $\gamma$-ray photon front has passed, the

222: decelerating blast wave from the GRB will plow into the cloud,

223: shock-heating the relativistic plasma.

224: 

225: Nonthermal synchrotron photons with energy $\e$ impinge on the atoms

226: in the cloud with a flux which can be parametrized as

227: \begin{equation} 

228: \Phi(\e) = (4\pi d^2)^{-1}\; {L \over m_ec^2 \e_0^2 \zeta_1}

229: \;\big[{1\over (\e/\e_0)^{2/3}+(\e/\e_0)^{\alpha_\gamma}}\bigr]\; 

230: \label{Phi}

231: \end{equation}

232: (Dermer, Chiang, \& B\"ottcher \markcite{dcb99}1999), where $\e_0 \sim

233: 1$ is the photon energy of the peak of the $\nu F_\nu$ spectrum,

234: $\alpha_\gamma \sim 2$-3 is the photon spectral index at energies

235: $\e\gg \e_0$, and $\zeta_1 \cong [3/4 + (\ag-2)^{-1}]$.  Hydrogen, the most

236: abundant species in the cloud, will be ionized on a time scale of

237: $5\times 10^{-5} (1+z) d_{16}^2 \zeta_1 \e_0^{4/3}/L_{50}$~s.  Fe

238: features might persist briefly during the early periods of very weak

239: GRBs on a time scale of $4\times 10^{-3} (1+z) d_{16}^2 \zeta_1

240: \e_0^{4/3}/L_{50}$~s, and would be identified by a rapidly evolving Fe

241: absorption feature at $9.1/(1+z)$ keV.  After the H and Fe are

242: ionized, the coupling between the GRB photons and gas is dominated by

243: Compton scattering interactions.  Pair production through

244: photon-particle processes are negligible by comparison with Compton

245: interactions except for photons with $\e \gtrsim 200$.  A lower limit

246: to the time scale for an electron to be scattered by a photon is

247: $t_T(s) \approx 15 (1+z) d_{16}^2 \zeta_1\e_0/\zeta_2 L_{50}$, where

248: $\zeta_2 = [3 + (\alpha_\gamma -1)^{-1}]$, assuming that all Compton

249: scattering events occur in the Thomson limit.  The Klein-Nishina 

250: decline in the Compton cross section will increase this

251: estimate by a factor of $\sim 1$-3, depending on the incident

252: spectrum.  Most of the electrons in the cloud will therefore be

253: scattered to high energies during a very luminous ($L_{50}\gg 1$) GRB,

254: or when the cloud is located at $d_{16} \ll 1$.

255: 

256: The average energy transferred to an electron at rest when

257: Compton-scattered by a photon with energy $\e$ is $\Delta\e \cong

258: \e^2/(1+1.5\e )$ (this expression is accurate to better than 18\% for

259: $\e < 10^3$).  Defining $\eta = \gamma -1$ as the dimensionless

260: electron kinetic energy, we can easily estimate the production rate

261: $f(\eta )$ of electrons scattered to energy $\eta$ in the

262: nonrelativistic ($\eta \ll 1$) and extreme relativistic ($\eta \gg 1$)

263: limits, noting that $f(\eta)d\eta \propto \Phi(\e)\sigma_C(\e)d\e$ and

264: letting $\eta\cong \Delta\e$.  In the former limit, the Compton cross

265: section $\sigma_C(\e)\rightarrow \sigma_T$ and $\Phi(\e) \propto

266: \e^{-2/3}$, so that $f(\eta)\propto \eta^{-5/6}$ when $\eta\ll$ min(1,

267: $\e_0^2$).  In the high energy limit, $\sigma_C(\e)\propto

268: \ln(3.3\e)/\e$ and $\Phi(\e) \propto \e^{-\alpha_\gamma}$, so that

269: $f(\eta )\propto \eta^{-(\alpha_\gamma+1)} \ln(2.2\eta)$ when

270: $\eta\gg$ max$(1, \e_0^2$).  Thus electrons are Compton-scattered on

271: the time scale derived above to form a hard spectrum that turns over

272: at kinetic energies of $\gtrsim 500\times$min($1,\e_0^2$) keV.  For a

273: GRB with $\e_0 \sim 1$, most of the kinetic energy is therefore

274: carried by nonthermal electrons with energies of $\sim 500$ keV.

275: Successive waves of photons that pass through this plasma will

276: continue to Compton-scatter the nonthermal electrons.  Only the lowest

277: energy photons, however, will be strongly affected by the radiative

278: transfer because both the Compton scattering cross section and energy

279: change per scattering is largest for the lowest energy photons.

280: 

281: Following the initial wave of photons, successive photon fronts also

282: encounter the back-scattered radiation (Madau \& Thompson

283: \markcite{mt99}1999; Thompson \& Madau \markcite{tm99}1999).  The

284: kinematics of the Compton process dictate that the energy $\e_s$ of a

285: photon back-scattered through $180^\circ$ by an electron at rest is

286: $\e_s = \e/(1+ 2\e)$; thus $\e_s$ cannot exceed 1/2 the electron

287: rest-mass energy.  Head-on collisions of the back-scattered photons by

288: primary GRB photons with $\e_1 > 2/\e_s = 2+2\sqrt{3}$ can thus produce

289: nonthermal e$^+$-$e^-$ pairs.  The cross section for $\gamma\gamma$

290: pair production peaks near threshold with a value of $\sim \sigma_T/3$.

291: The $\gamma\gamma$ pair-production optical depth $\tau_{\gamma\gamma}

292: (\e_1)$ of a photon that trails the onset of the GRB by $\Delta t$

293: seconds can be estimated by noting that the photon traverses a

294: distance $\sim r$ through a backscattered radiation field of spectral

295: density $n_s(\e_s)\approx n_e \sigma_T \cdot \Delta t \cdot \Phi [\e_s

296: /(1-2\e_s )]$ -- a more accurate calculation would replace the term

297: $\Delta t \cdot \Phi [\e_s/(1-2\e_s)]$ by an integral over the

298: time-varying flux.  Approximating $\tau_{\gamma\gamma} (\e_1)\approx r

299: (\sigma_T/3) \Delta\e_s n_s(2/\e_1)$, where $\Delta \e_s \simeq 2/\e_1$ is

300: the bandwidth that is effective for producing pairs, we obtain

301: \begin{equation} 

302: \tau_{\gamma\gamma} (\e_1) \approx 0.02 \; {\tau_T \Delta t [s] \, 

303: L_{50} k(\e_1) \over d_{16}^2 \e_1\e_0^2\zeta_1}\;\big[{1\over 

304: (\e^\prime/\e_0)^{2/3} + (\e^\prime/\e_0)^{\alpha_\gamma}}\bigr]\; , 

305: \label{tau_gg}

306: \end{equation}

307: where $\e^\prime = 2/(\e_1-4)$.  The coefficient results from a more

308: detailed derivation, and the term $k(\e_1) = 1-4\e_1^{-1} + \e_1

309: /(\e_1 -4)$ is a Klein-Nishina correction.  Eq.\ (\ref{tau_gg}) shows 

310: that photons with energies above several MeV will be severely attenuated 

311: in Thomson thick clouds if $L_{50}\gg 1$ or $d_{16}\ll 1$.  Photons with

312: MeV energies are most severely attenuated, and $\tau_{\gamma\gamma}

313: (\e_1)\propto \e_1^{-1/3}$ at energies $\e_1 \gg \max (1, \e_0)$.

314: The $\gamma\gamma$ pair injection process provides another source of

315: nonthermal leptons with $\eta \sim 1$.  The pairs will not, however,

316: be electrostatically bound but will be accelerated by the photon

317: pressure and electrostatic field.  

318: 

319: Fig.\ 1 shows Monte Carlo simulations of radiation spectra 

320: described by eq.\ (\ref{Phi}) that pass through a hot electron 

321: scattering medium.  For simplicity, we approximate the 

322: hard nonthermal electron spectrum by a thermal

323: distribution with temperatures of 100 and 300~keV.  These

324: calculations show that the lowest energy photons of the primary

325: synchrotron spectrum are most strongly scattered, and that the 

326: ``line-of-death" problem of the synchrotron shock model of 

327: GRBs (Preece et al.\ \markcite{pea98}1998) can be solved by 

328: radiation transfer effects through a hot scattering cloud 

329: with $\tT \gtrsim$ 1-2.  

330: 

331: According to this interpretation, GRBs 

332: displaying very hard spectra could display one break from the 

333: intrinsic synchrotron shock spectrum and a second break

334: from the scattering process. In the examples shown in Fig. 1,

335: these two breaks are so close to each other that they appear

336: as one smooth turnover. Two breaks are observed from GRB

337: 970111 (Crider \& Liang \markcite{cl99}1999), a GRB that strongly

338: violates the ``line of death." A prediction of this model is

339: that GRBs showing such flat X-ray spectra should also display softer

340: MeV spectra than typical GRBs due to $\gamma\gamma$ attenuation

341: processes in the hot scattering cloud.

342: 

343: \section{Observational Signatures of the Flash-Heated Cloud}

344: 

345: The Compton-scattered electrons transfer momentum to the $N_p =

346: 4\times 10^{54}r_{15}^3n_9$ protons of the cloud through their

347: electrostatic coupling.  If the radiation efficiency is $\xi_r$, then

348: the Compton impulse gives each proton in the cloud $\approx \xi_r

349: [1-\exp(-\tT )]E \pi r^2/(4\pi d^2N_p) \approx 40 [1-\exp(-\tT )]

350: (\xi_r/0.1) E_{54}/(d_{16}^2 r_{15}n_9)$ MeV of directed energy.

351: Pairs, by contrast, will be accelerated to mildly relativistic speeds

352: until Compton drag or streaming instabilities limit further

353: acceleration.  In the simplification that the medium interior to the

354: cloud is uniform, and neglecting pair-loading of the swept-up material

355: (Thompson \& Madau \markcite{tm99}1999), the decelerating blast wave

356: follows the dynamical equation $\Gamma(x) = \Gamma_0(x/x_d)^{-g}$,

357: where $g = 3/2$ and 3 for adiabatic and radiative blast waves,

358: respectively.  Using the standard parameters adopted here, the blast

359: wave slows to between $\xi= 0.01 - 0.1$ of its initial Lorentz factor

360: before reaching a cloud at $d = 10^{16}$ cm.  Even considering the

361: radiative acceleration of the cloud, the blast wave reaches the cloud

362: at time $t_{bw} = t_d(d/x_d)^{(2g+1)}/(2g+1) \lesssim t_{\rm dyn}$,

363: where the dynamical time scale of the cloud is $t_{\rm dyn} = r/c =

364: 3\times 10^4 r_{15}$ s.  Because the cloud is so dense, a large

365: fraction of the residual energy of the blast wave is deposited into

366: the $N_p$ particles of the cloud.  Thus each proton in the cloud

367: receives an additional $m_p\beta_p^2 c^2 \approx \xi E \pi r^2/(4\pi

368: d^2N_p) \approx 40 (\xi/0.1) E_{54}/(d_{16}^2 r_{15}n_9$) MeV of

369: kinetic energy, divided roughly equally into directed outflow and

370: random thermal energy.

371: 

372: If the circumstellar medium at $d_{16}\gg 1$ is much more dilute than

373: the interior region, as suggested by observations of $\eta$ Carinae,

374: then we can neglect further interactions of the cloud/blast wave

375: system with their surroundings.  The observational signatures and fate

376: of the cloud at late times can be outlined by comparing time scales.

377: The cloud expands on a time scale $t_{\rm ex} = t_{\rm dyn}/\beta =

378: 1.5\times 10^5 r_{15}/(\beta/0.2)$ s.  The basic time scale

379: governing radiative processes in the cloud is the Thomson time $t_{\rm

380: T} = (n_e\sigma_T c)^{-1} = 5\times 10^4/n_9$~s $= t_{\rm dyn} /\tau_{\rm

381: T}$.  The electrons thermalize on a time scale $t_{\rm T}/\ln\Lambda

382: \ll t_{\rm dyn}$, where the Coulomb logarithm $\ln\Lambda \approx 20$.

383: The protons transfer their energy to the electrons on the time

384: scale $t_{ep}\cong \theta^{3/2}(m_p/m_e)t_{\rm T}/\ln\Lambda$, 

385: where $\theta = kT/m_ec^2$ is an effective dimensionless

386: electron temperature, and we assume collective plasma processes for

387: energy exchange to be negligible.

388: 

389: The flash-heated cloud can evolve in two limiting regimes.  When the

390: external soft-photon energy density is small, the system emits by a

391: hard bremsstrahlung spectrum with $\theta \sim$ 1 and luminosity

392: $L_{ff}\cong N_e\alpha_f m_ec^2\theta^{1/2}/t_{\rm T} \approx 5\times

393: 10^{41} r_{15}^3 n_9^2\theta^{1/2}$ ergs s$^{-1}$.  In the more likely

394: case when abundant soft photons are present, for example, from the

395: Compton echo (Madau, Blandford, \& Rees \markcite{mbr99}1999), then

396: Compton cooling will balance ion heating to produce a luminous

397: Comptonized soft-photon spectrum with effective temperature $\theta

398: \sim 0.1$ and luminosity $L_{\rm C}\cong \xi E \pi r^2/(4\pi d^2

399: t_{ep}) \approx 2\times 10^{45} (\xi/0.1) E_{54} r_{15}^2 n_9

400: d_{16}^{-2}(\theta/0.1)^{-3/2}$ ergs s$^{-1}$.  In either case, the

401: spectra persist until the plasma expands and adiabatically cools, that

402: is, for a period $\sim t_{ex} \sim$ day.  The hot bremsstrahlung

403: plasma will be too dim to be detectable with current instrumentation,

404: but a 50-100 keV Comptonized plasma at redshift $z \sim 1$ and luminosity 

405: distance of $10^{28}D_{28}$ cm would have

406: a flux of $\sim 1.6\times 10^{-12} (\xi/0.1) E_{54} r_{15}^2 n_9

407: d_{16}^{-2}(\theta/0.1)^{-3/2} D_{28}^{-2}$ ergs cm$^{-2}$ s$^{-1}$.

408: The hot plasma formed by a nearby GRB at $z \sim 0.1$ would be

409:  easily detectable with the

410: INTEGRAL and Swift missions at hard X-ray and soft $\gamma$-ray

411: energies.  In either case, e$^+$-e$^-$

412: pairs would be formed with moderate efficiency, 

413: and the cooling, expanding plasma

414: would produce a broad pair annihilation feature (Guilbert \& Stepney

415: \markcite{gs85}1985).  The residual pairs formed in the relativistic

416: plasma and the pair wind would diffuse into the dilute interstellar

417: medium with density $n_{\rm ISM}$ to annihilate on a time scale

418: $(n_{\rm ISM} \sigma_{\rm T} c)^{-1} \sim 2 \times 10^6/n_{\rm ISM}$

419: yr.  If the energy intercepted by a single cloud is converted to pairs

420: with a conservative 1\% pair yield, past GRBs in the Milky Way would

421: be revealed by localized regions of annihilation radiation with flux

422: $\sim 2\times 10^{-5} E_{54} n_{\rm ISM} ~(d/10 {\rm kpc})^{-2}$ 0.511

423: MeV ph cm$^{-2}$ s$^{-1}$.  The high-latitude annihilation feature

424: discovered with OSSE on the Compton Gamma Ray Observatory (Purcell et

425: al.\ \markcite{pea97}1997), or other localized hot spots of

426: annihilation radiation that will be mapped in detail with INTEGRAL,

427: could reveal sites of past GRB explosions. 

428: 

429: \acknowledgments{CD thanks B.  Paczy\'nski for stressing the importance

430: of massive star observations in developing blast-wave models of GRBs.

431: The work of CD is supported by the Office of Naval Research and NASA

432: Astrophysical Theory Program (DPR S-13756G). The work of MB is

433: supported by NASA through Chandra Postdoctoral Fellowship grant

434: PF~9-10007, awarded by the Chandra X-ray Center, which is operated

435: by the Smithsonian Astrophysical Observatory for NASA under

436: contract NAS~8-39073.}

437: 

438: \begin{references}

439: 

440: \reference{bea99a} Bloom, J.  S.  et al.\ 1999a, ApJ, 518, L1

441: 

442: \reference{bea99b} Bloom, J.  S.  et al.\ 1999b, Nature 401, 453

443: 

444: \reference{bdcl} B\"ottcher, M., Dermer, C.  D., Crider, A.  W.  \&

445: Liang, E.  P.  1999, A\&A 343, 111 

446: 

447: \reference{mb99} B\"ottcher, M. 1999, ApJ, submitted (astro-ph/9912030) 

448: 

449: \reference{jb94} Brainerd, J. J.  1994, ApJ, 428, 21 

450: 

451: \reference{cl99} Chevalier, R. A., \& Li, Z.-Y., 1999, ApJ, 520, L29

452: 

453: \reference{crider97} Crider, A.  et al.\ 1997, ApJ 479, L39 

454: 

455: \reference{cl99} Crider, A.  W.,  \& Liang, E.  P.  1999, A\&AS, 138, 405 

456: 

457: \reference{cea97} Costa, E. et al.\ 1997, Nature, 387, 783 

458: 

459: \reference{dh97} Davidson, K.,  \& Humphreys, R.  M.  1997, ARAA, 35, 1 

460: 

461: \reference{dea95} Davidson, K.  et al.\ 1995, AJ, 109, 1784 

462: 

463: \reference{dcb99} Dermer, C.  D., Chiang, J.,  \& B\"ottcher, M. 1999, 

464: ApJ 513, 656 

465: 

466: \reference{dea97} Djorgovski, S.  G.  et al.\ 1997, Nature, 387, 876 

467: 

468: \reference{fea97} Frail, D.  A., Kulkarni, S.  R., Nicastro, L., 

469: Feroci, M.  \& Taylor, G.  B.  1997, Nature 389, 261

470: 

471: \reference{granot00} Granot, J., Piran, T., \& Sari, R., 2000, ApJ,

472: submitted (astro-ph/0001160)

473: 

474: \reference{gs85} Guilbert, P.  W.,  \& Stepney, S., 1985, MNRAS, 212, 523 

475: 

476: \reference{hw88} Hofmann, K.  H., \& Weigelt, G.  1988, A\&A, 203, L21 

477: 

478: \reference{hf99} Hogg, D.  W., \& Fruchter, A.  S.  1999, ApJ 520, 54 

479: 

480: \reference{kea99} Kulkarni, S. R. et al.\ 1999, Nature 398, 389 

481: 

482: \reference{lamb99} Lamb, D.  Q. 1999, A\&AS, 138, 607 

483: 

484: \reference{lc99} Li, Z.-Y., \& Chevalier, R. A., 1999, ApJ, 526, 716

485: 

486: \reference{liang97} Liang, E. P., 1997, ApJ, 491, L15

487: 

488: \reference{lcbs99} Liang, E.  P., Crider, A. W., B\"ottcher, M., \& 

489: Smith, I.  1999, ApJ 519, L21 

490: 

491: \reference{lk94} Liang, E. P., \& Kargatis, V. E., 1994, ApJ, 432, L111

492: 

493: \reference{lp99} Lloyd, N. M., \& Petrosian, V., 1999, in proc. of 5$^{th}$

494: Huntsville Symposium on Gamma-Ray Bursts, AIP proc. in press (astro-ph/9912205)

495: 

496: \reference{mt99} Madau, P.  \& Thompson, C.  1999, ApJ, in press 

497: (astro-ph/9909060)

498: 

499: \reference{mbr99} Madau, P., Blandford, R.  D., \& Rees, M.  J.  1999,

500: ApJ, submitted (astro-ph/9912276) 

501: 

502: \reference{mr00} M\'esz\'aros, P. \& Rees, M.  J.  2000, ApJL, in press 

503: 

504: \reference{mrw98} M\'esz\'aros, P., Rees, M.  J.  \& Wijers, R.  A.  M.  J.  

505: 1998, ApJ 499, 301

506: 

507: \reference{mr97}M\'esz\'aros, P., \& Rees, M.  J.  1997, ApJ, 476, 232

508: 

509: \reference{npp92} Narayan, R., Paczy\'nski, B., \& Piran, T.  1992, ApJ

510: 395, L83 

511: 

512: \reference{bp98} Paczy\'nski, B.  1998, ApJ, 494, L4

513: 

514: \reference{pr93}Paczy\'nski, B., \& Rhoads, J.  1993, ApJ, 418, L5

515: 

516: \reference{tp99} Piran, T.  1999, Phys Rpts, 314, 575

517: 

518: \reference{pea99} Piro, L.  et al.\ 1999, ApJ, 514, L73

519: 

520: \reference{pea98} Preece, R.  D.  et al.\ 1998, ApJ, 506, L23

521: 

522: \reference{pea97} Purcell, W.  R., et al.\ 1997, ApJ 491, 725

523: 

524: \reference{rm92} Rees, M.  J., \& M\'esz\'aros, P.  1992, MNRAS, 258, 41P 

525: 

526: \reference{tm99} Thompson, C., \& Madau, P.  1999, ApJ, submitted

527: (astro-ph/9909111) 

528: 

529: \reference{vpea97} van Paradijs, J.  et al.\ 1997, Nature, 386, 686 

530: 

531: \reference{sw93} Woosley, S.  E.  1993, ApJ, 405, 273

532: 

533: \end{references} 

534: 

535: \eject

536: 

537: \setcounter{figure}{0}

538: \begin{figure} 

539: \rotate[r]{ \epsfysize=12cm

540: \epsffile[150 0 550 500]{fig1.ps} }

541: 

542: \caption[]{Radiation transfer

543: effects on GRB emission that passes through electrons energized by an

544: earlier portion of the photon front.  The intrinsic spectrum eq.(1),

545: with $\alpha_X = 2/3$, $\alpha_\gamma = 2.5$ and $\e_0 = 0.5$, is

546: shown by the thick dashed curve.  The nonthermal electrons are

547: approximated by a thermal distribution with temperatures of 100 keV

548: (thin curves) and 300 keV (thick curves), and Thomson depths $\tT = 1$

549: (solid curves) and $\tT = 2$ (dotted curves).  Spectral indices

550: $\alpha_X$ calculated at 50 keV are 0.5 ($T = 100$ keV, $\tT = 1$),

551: 0.44 ($T = 300$ keV, $\tT = 1$), 0.14 ($T = 100$ keV, $\tT = 2$), and

552: 0.05 ($T = 300$ keV, $\tT = 2$).  } 

553:  

554: \label{figure1} 

555: \end{figure}

556: \end{document}

557: