1: %\documentclass[preprint,flushrt]{aastex}

2: %\documentclass[12pt]{article}\usepackage{natbib,amssymb,psfig,epsfig}

3: \documentclass[12pt,preprint]{aastex}\usepackage{amssymb,epsfig}

4: %\documentclass{article}\usepackage{natbib,psfig,epsfig,emulateapj}

5: \newcommand{\beq}{\begin{equation}}

6: \newcommand{\eeq}{\end{equation}}

7: \newcommand{\bea}{\begin{eqnarray}}

8: \newcommand{\eea}{\end{eqnarray}}

9: \newcommand{\rem}[1]{ }

10: \bibpunct[,]{(}{)}{;}{a}{}{,}

11: \begin{document}

12: %\date{\today}

13: 

14: \title{Jitter radiation as a possible mechanism for Gamma-Ray Burst

15: afterglows. Spectra and lightcurves} 

16: \author{Mikhail V. Medvedev\altaffilmark{1},}

17: \affil{Department of Physics and Astronomy, University of Kansas, KS 66045}

18: \altaffiltext{1}{

19: Also at the Institute for Nuclear Fusion, RRC ``Kurchatov

20: Institute'', Moscow 123182, Russia}

21: \author{Davide Lazzati, Brian C. Morsony and Jared C. Workman}

22: \affil{JILA, University of Colorado, 440 UCB, Boulder, CO 80309-0440}

23: 

24: 

25: \begin{abstract}

26: The standard model of GRB afterglows assumes that the radiation

27: observed as a delayed emission is of synchrotron origin, which

28: requires the shock magnetic field to be relatively homogeneous on

29: small scales.  An alternative mechanism -- jitter radiation, which

30: traditionally has been applied to the prompt emission -- substitutes

31: synchrotron when the magnetic field is tangled on a microscopic

32: scale. Such fields are produced at relativistic shocks by the

33: Weibel instability. Here we explore the possibility that small-scale

34: fields populate afterglow shocks. We derive the spectrum of jitter

35: radiation under the afterglow conditions. We also derive the afterglow

36: lightcurves for the ISM and Wind profiles of the ambient

37: density. Jitter self-absorption is calculated here for the first

38: time. We find that jitter radiation can produce afterglows similar to

39: synchrotron-generated ones, but with some important differences. We

40: compare the predictions of the two emission mechanisms. By fitting 

41: observational data to the synchrotron and jitter afterglow lightcurves, 

42: it can be possible to discriminate between the small-scale vs large-scale 

43: magnetic field models in afterglow shocks.  

44: \end{abstract}

45: 

46: \keywords{ gamma rays: bursts --- radiation processes --- shock waves 

47: --- magnetic fields }

48: 

49: %\maketitle

50: 

51: \section{Introduction}

52: 

53: The general framework for the interpretation of the long-wavelength

54: radiation of gamma-ray burst (GRB) afterglows is the external shock

55: synchrotron model \citep{MR97,Wax97,P99}. In that scenario, a

56: blastwave is generated by the interaction of the GRB ejecta with the

57: interstellar medium. At the shock front, electrons are accelerated in

58: a power-law distribution with energy (or with Lorentz factor $\gamma$)

59: and a strong magnetic field is generated by some mechanism. The model

60: assumes that the magnetic field is coherent on the Larmor scale of the

61: emitting electron, hence allowing for synchrotron emission.  Both the

62: relativistic electron population and the magnetic field are originally

63: thought to share a sizable fraction of $\sim10\%$ of the internal

64: energy of the blastwave; those fractions are being called $\epsilon_e$

65: and $\epsilon_B$, respectively. Afterglow spectral fits yield typical

66: values, $\epsilon_e\sim0.1-0.01$ and $\epsilon_B\sim0.01-0.0001$, with

67: relatively large scatter \citep{PK01,P06}.

68: 

69: No mechanism or instability capable of generating a sub-equipartition

70: magnetic field in GRBs has been identified for a while, until

71: \citet{ML99} suggested that the field can be generated through the

72: Weibel instability. This prediction has been extended to

73: non-relativistic shocks, e.g., in supernovae and galaxy clusters

74: \citep{MSK05}, and confirmed via numerical PIC simulations

75: \citep{Silva+03,Nish+03,fred04,M+05,Spit05}.  The volume-averaged value of

76: $\epsilon_B$ deduced from the simulation is indeed $\sim0.01-0.0001$

77: (depending on location with respect to the main shock compression). 

78: An intriguing relation: 

79: \beq

80: \epsilon_e\simeq\epsilon_B^{1/2}, 

81: \eeq 

82: recenly found by \citet{M06b},

83: can allow one to reduce the number of fit parametes in afterglow studies.

84: 

85: Weibel-generated fields have a very short coherence length-scale

86: (smaller than $1/\gamma^2$ times the electron Larmor radius) and

87: standard synchrotron equations cannot be adopted. The theory of jitter

88: radiation has been proposed by \citet{M00} and further developed in

89: subsequent works \citep{M06,Fleish06}. (Note, however, that

90: \citealp{Fleish06} incorrectly predicts the absence of steep spectral

91: slopes, such as $F_\nu\propto\nu^1$, just below $E_{\rm peak}$ in

92: GRBs).

93: 

94: Unlike synchrotron, jitter radiation is sensitive to the statistical

95: properties of the magnetic field in the shock, that is to the spectrum

96: of magnetic fluctuations, and not just to its ``global property'' --

97: the strength \citep{M00}. In addition, the spectrum of jitter radiation

98: depends on the shock viewing angle, i.e., the angle, $\theta'$ between

99: the shock velocity (= propagation direction) and the line of sight

100: ($\theta'$ is measured in the shock comoving frame).  The two extreme

101: cases are characterized by the emissivity function being a power-law

102: at frequencies below the spectral peak:

103: \beq

104: P(\omega)\propto\omega^\alpha, 

105: \eeq 

106: with $\alpha=1$ for $\theta'=0$ --

107: a shock viewed face-on (this is also the case of the ``effective'' 1D

108: magnetic turbulence, considered in \citealp{M00}), and $\alpha=0$ for

109: $\theta'=\pi/2$ -- an edge-on shock (this is also the case in

110: isotropic 2D and 3D turbulence).  These asymptotes, along with a

111: general case of $0\le\theta'\le\pi/2$, are considered elsewhere

112: (Medvedev 2006, Workman et al. in preparation).  For reference,

113: synchrotron radiation has $\alpha=1/3$ at low energies.

114: 

115: Should afterglows be resolved, one would observe a limb-brightened

116: object \citep{GPS99}. For a spherical outflow, the emission spectrum

117: will be entirely dominated by the limb emission. Because of

118: relativistic aberration, the photons emitted within an angle of $\sim

119: 1/\Gamma_{\rm sh}$ with respect to the line of sight in the lab

120: (observer's) frame, have been emitted at angle $\theta'$ in the shock

121: comoving frame: 

122: %

123: \beq \cos\theta'=\frac{\cos(1/\Gamma_{\rm

124: sh})-\beta_{\rm sh}} {1-\beta_{\rm sh}\cos(1/\Gamma_{\rm sh})}= 0,

125: \eeq 

126: %

127: where $\beta_{\rm sh}$ is the dimensionless shock speed.  That is,

128: these photons are emitted at the angle $\theta'= 90^\circ$ away from

129: the shock normal. This consideration is also true for a jet outflow at

130: times when the jet opening angle is larger than $1/\Gamma_{\rm sh}$.

131: 

132: In general, after the jet break, the emission will still be dominated

133: by the limb. However, the shock there will not be seen edge-on, but

134: rather at some angle $\theta'<90^\circ$.  The jitter spectrum is, in

135: general, anisotropic, i.e., it depends on $\theta'$. There is a break,

136: $\nu_b(\theta')$ in the jitter spectrum \citep{M06} below the jitter

137: peak, $\nu_b<\nu_{\rm jitter~peak}$, such that $\alpha=1$ above the

138: break (at $\nu_b<\nu<\nu_{\rm jitter~peak}$) and $\alpha=0$ below it

139: (at $\nu<\nu_b$).  The dependence of the break frequency on $\theta'$

140: is not easy to parameterize. Since the break is quite smooth is also

141: not easy to establish its precise position.  One characteristic point

142: is: $\nu_{b} \sim 0.01 \nu_{\rm jitter~peak}$ at $\theta'=\pi/10 =

143: 18$~degrees.

144: 

145: Thus, for spherical outflows and for jets seen before the ``jet

146: break'', the reasonable approximation will be the ``edge-on'' shock,

147: with $\alpha=0$. However, at late times, especially when the jet

148: becomes weakly relativistic, the ``face-on'' case with $\alpha=1$

149: should be a better approximation.

150: 

151: One cautious remark. The above consideration assumes strong anisotropy 

152: of the magnetic turbulence in the shock \citep{ML99}, which is likely 

153: true in the internal shocks. We do not understand well the properties

154: of magnetic turbulence far downstream the external shock. If the magnetic 

155: turbulence will become nearly isotropic, then a 3D jitter regime

156: gives $\alpha=0$, independent of the viewing angle.

157: 

158: Below, we consider both the $\alpha=0$ and the $\alpha=1$ cases in

159: calculating the jitter self-absorption frequency. This paper is

160: organized as follows: in \S~2 we summarize the shock kinematics, in

161: \S~3 we derive the jitter self absorption frequency in various regimes

162: while in \S~4 we compute all the observable properties of the

163: spectrum. We summarize and discuss our results in \S~5.

164:  

165: 

166: \section{Shock kinematics}

167: 

168: Here we sumarize, for future use, the kinematic relations of several

169: blast wave parameters. For a highly relativistic blast wave with

170: Lorentz factor $\Gamma_{\rm sh}$, the density jump condition relates

171: the pre-shock ISM density to the density downstream, $n'$, measured in

172: the shock frame: 

173: %

174: \beq 

175: n'\simeq4\Gamma_{\rm sh}n_{\rm ISM}.  

176: \eeq

177: %

178: The magnetic field strength downstream 

179: %

180: \beq 

181: B'=\left(32\pi\Gamma_{\rm sh}^2 n_{\rm ISM} m_p c^2\epsilon_B\right)^{1/2}, 

182: \eeq 

183: %

184: where $\epsilon_B$ is the magnetic field equipartition parameter.  The

185: minimum Lorentz factor of the accelerated electrons, which share a

186: fraction $\epsilon_e<1$ of the total energy of the shock (outflow) is:

187: %

188: \beq 

189: \gamma_m=\left(\frac{s-2}{s-1}\right)\frac{m_p}{m_e}\

190: \epsilon_e\Gamma_{\rm sh}\approx6.12\times10^2\epsilon_e

191: \Gamma_{\rm sh}.  

192: \eeq 

193: %

194: Here the last expression is calculated for a typical electron

195: power-law index $s=2.5$, see definition below, Eq (\ref{N}).

196: 

197: \subsection{Constant density ISM}

198: 

199: The radius and Lorentz factor of a blast wave propagating in a

200: constant density ISM depend on the observed time as \citep{GPS99}

201: %

202: \begin{eqnarray}

203: \Gamma_{\rm sh}&\approx& 3.65\left(\frac{E_{52}}{n_{\rm

204: ISM,0}}\right)^{1/8} \left(\frac{t_{\rm days}}{1+z}\right)^{-3/8}, \\

205: R&\approx& 5.53\times10^{17} \left[\frac{E_{52} t_{\rm days}}{n_{\rm

206: ISM,0}(1+z)} \right]^{1/4}\textrm{ cm}.

207: \label{ISM}

208: \end{eqnarray}

209: %

210: Here $E_{52}=E_{\rm explosion}/10^{52}$~erg, and $n_{\rm ISM,0}=n_{\rm

211: ISM}/1$~cm$^{-3}$.

212: 

213: \subsection{Wind model}

214: 

215: If the blast wave is propagating in the wind environment with the

216: density dicreasing with distance, $n\propto r^{-2}$, the Lorentz

217: factor and the radius of the blast wave are \citep{CL00}

218: %

219: \begin{eqnarray}

220: \Gamma_{\rm sh}&\approx& 4.96\,

221: E_{52}^{1/2}\,\left(\frac{A_* t_{\rm days}}{1+z}\right)^{-1/4}, \\

222: R&\approx& 1.56\times10^{17}

223: \left[\frac{ E_{52} t_{\rm days}}{A_* (1+z)}

224: \right]^{1/2}\textrm{ cm}.

225: \label{Wind}

226: \end{eqnarray}

227: %

228: Here the wind parameter $A_*=[\dot M_w/(10^{-5}M_{\sun}\textrm{

229: yr}^{-1})]/ [V_w/(10^3\textrm{ km s}^{-1})]$, where $\dot M_w$ is the

230: mass loss rate, $V_w$ is the wind velocity. Since the ambient density

231: is no longer constant, one need to substitute $n_{\rm ISM}$ with the

232: wind density 

233: %

234: \beq 

235: n_{\rm wind}=AR^{-2}\approx(3.00\times10^{35}\,A_*\textrm{ cm}^{-1})\,R^{-2}.

236: \eeq

237: 

238: \section{Theory}

239: 

240: Here we work in the shock comoving frame. Thus, all the frequencies

241: are expressed in this frame. Also, all the shock parameters, such as

242: the particle number density $n'$ and magnetic field $B'$, are those in

243: the shocked region and are measured in the comoving frame as well,

244: unless stated otherwise. We use ``prime'' to denote quantities

245: measured in the shock frame. Sometimes we omit ``prime'' when this

246: does not cause any confusion.

247: 

248: We introduce the accelerated electron distribution function:

249: %

250: \beq

251: N(\gamma)=(s-1)N_e\gamma_m^{s-1}\gamma^{-s}, \qquad \gamma\ge\gamma_m,

252: \label{N}

253: \eeq

254: %

255: where $s$ is the power-law index, $\gamma_m$ is the minimum Lorenz

256: factor (low-energy cut-off), $N_e=4\pi R^2\Delta' n_{\rm ISM}$ is the

257: total number of the nonthermal (emitting) electrons, $R$ is the radius

258: of the blast wave, $\Delta'$ is its thickness, $n_{\rm ISM}$ is the

259: number density of the electrons in the ambient medium.

260: 

261: The absorption coefficient at comovong frequency $\nu'$ is

262: \citep{DBC00}:

263: %

264: \beq

265: \kappa_{\nu'}=\frac{1}{8\pi m_e V_{\rm bw} \nu'^2}

266: \int_1^{\infty} d\gamma\; P(\nu',\gamma)~

267: \gamma^2\frac{\partial}{\partial\gamma}\left[\frac{N(\gamma)}{\gamma^2}\right],

268: \eeq

269: %

270: where $V_{\rm bw}=4\pi r^2\Delta'$ is the volume of the blast wave of

271: the comoving thickness $\Delta'$ and $ P(\nu',\gamma)$ is the

272: emissivity function. Straightforwardly, 

273: %

274: \beq

275: \gamma^2\frac{\partial}{\partial\gamma}\left[\frac{N(\gamma)}{\gamma^2}\right]

276: =-(s+2)(s-1)N_e\gamma_m^{s-1}\gamma^{-s-1}.

277: \label{kappa}

278: \eeq

279: 

280: \subsection{Emissivity functions}

281: 

282: We first discuss synchrotron, for reference.  The synchrotron comoving

283: peak frequency is at 

284: %

285: \beq \nu'_s=(3/2)\nu_B\gamma^2, 

286: \eeq 

287: %

288: where $\nu_B=eB'/2\pi m_ec$ is the cyclotron (Larmor) frequency in a

289: homogeneous magnetic field of strength $B'$, the pitch-angle is

290: assumed to be $\pi/2$, for simplicity.  At $\nu'\ll\nu'_s$, the

291: synchrotron emissivity is 

292: %

293: \beq P_{\rm synch}(\nu',\gamma)\simeq

294: \frac{e^2}{c}\sqrt{3}\nu_B\left[\frac{4\pi}

295: {\sqrt{3}\Gamma(1/3)}\left(\frac{\nu'}{3\nu_B\gamma^2}\right)\right],

296: \eeq 

297: %

298: where $\Gamma(1/3)\simeq2.68$ is a gamma function.

299: 

300: The jitter peak frequency is determined by the magnetic field spectrum

301: in the post-shock medium.  Recent numerical simulations

302: \citep{fred04,MSK05} demonstrate that the field generation by both the

303: electrons and the protons, occurs well before the main shock

304: compression. The wave-vector of the fastest growing mode of the Weibel

305: instability in the linear regime is \citep{ML99}: 

306: %

307: \beq 

308: k_{\rm Weibel}\approx 2^{-1/4}(\omega_{p,e}/c){\bar\gamma_e}^{-1/2}, 

309: \eeq

310: %

311: where $\omega_{p,e}=\sqrt{4\pi e^2 n/m_e}\approx

312: 5.64\times10^4n^{1/2}$~s$^{-1}$ is the plasma frequency. Because

313: $\omega_{p,e}$ is a Lorentz invariant, $k_{\rm Weibel}$ is determined

314: by the parameters of the ambient, unshocked medium alone. Thus, the

315: density $n=n_{\rm ISM}$ is the ambient medium density, and the mean

316: Lorentz factor of the electrons in the ambient medium, $\bar\gamma_e$,

317: is close unity (the ISM is cold, non-relativistic). The shock

318: compression and the proton thermalization occur far downstream,

319: significantly behind the region where the field has been

320: generated. Therefore, the correlation length of the field decreases

321: (primarily in the parallel direction) because of the

322: compression. Hence, the characteristic wave-vector of the downstream

323: random fields is

324: %

325: \beq 

326: k_{\rm rand}\simeq (4\Gamma_{\rm sh})k_{\rm  Weibel}.  

327: \eeq 

328: %

329: Numerical simulations also indicate that the correlation scale of the

330: field $\lambda_B=2\pi/k_{\rm rand}$ varies with distance from the

331: shock front because of the highly nonlinear dynamics of the

332: Weibel-generated currents and fields \citep{fred04,M+05}. We therefore

333: parameterize this as 

334: %

335: \beq 

336: k_{\rm rand}=\eta(\omega_{p,e}/c), 

337: \eeq

338: %

339: where the parameter $\eta$ incorporates all relativistic effects, the

340: shock compression and the nonlinear evolution of the Weibel

341: turbulence.  Thereafter, we use 

342: %

343: \beq 

344: \eta\simeq 2^{-1/4}(4\Gamma_{\rm sh})\bar\gamma_e^{-1/2}

345: \simeq2^{7/4}\Gamma_{\rm sh}, 

346: \eeq 

347: %

348: where we assumed that $\bar\gamma_e\sim1$.

349: 

350: The characteristic frequency of the electron's jitter while it moves

351: at roughly the speed of light through these magentic fields is

352: $\nu\sim c/\lambda_B$. More precisely, 

353: %

354: \beq 

355: \nu_r=k_{\rm rand}c/2\pi=\eta\omega_{p,e}/2\pi=\eta\nu_{p,{\rm ISM}}, 

356: \eeq 

357: %

358: where we introduced the plasma frequency of the ISM, $\nu_{p,{\rm

359: ISM}}=(4\pi e^2 n_{\rm ISM}/m_e)^{1/2}/2\pi =8.98\times10^3 n_{\rm

360: ISM}^{1/2}$~Hz.

361: 

362: The peak of the emitted jitter radiation is $\nu'_j\sim\nu_r\gamma^2$,

363: but its exact position slightly depends on the magnetic field

364: spectrum, defined as $(B'^2)_k\propto k^{2\mu}$. For steep spectra,

365: $\mu\gg1$, the peak is at, roughly, twice the jitter frequency: 

366: \beq

367: \nu'_{j}\simeq 2\nu_r\gamma^2. 

368: \eeq 

369: %

370: The jitter and synchrotron peak frequencies in the magnetic fields of

371: identical strengths are related to each other via the identities \beq

372: \nu_r=\nu_B/\delta, \textrm{ or } \nu'_{j}=\nu'_s/(3\delta/4), \eeq

373: These identities define the parameter $\delta\lesssim1$, which is the

374: ratio of the deflection angle of the particle path in chaotic fields

375: to the beaming angle $1/\gamma$ \citep{M00}. It is expressed via the

376: magnetic field equipartition parameter in the shock, $\epsilon_B$, as

377: \beq \delta=\left(\frac{m_p}{m_e}\frac{8\Gamma_{\rm sh}}{\eta^2}\,

378: {\epsilon_B}\right)^{1/2} \simeq

379: \left(\frac{m_p}{m_e}\,2^{-1/2}{\epsilon_B}\right)^{1/2} \approx

380: 36.0\sqrt{\epsilon_B}.  \eeq The jitter emissivity function below the

381: peak is \beq P_{\rm jitter}(\nu',\gamma)\simeq\frac{e^2}{c}\pi

382: f(\mu)\delta^2\nu_r \left(\frac{\nu'}{2\nu_r\gamma^2}\right)^\alpha.

383: \eeq Later on, we neglected a factor $f(\mu)=(2\mu+1)/(2\mu-1)$

384: (calculated for $\alpha=1$) because it is of order unity when

385: $\mu\gg1$.

386: 

387: Thus, for both emission mechanisms,

388: \beq

389: P(\nu',\gamma)\simeq a \left(\frac{\nu'}{b\gamma^2}\right)^\alpha.

390: \label{P}

391: \eeq

392: Here, for synchrotron

393: \beq

394: a=\frac{e^2}{c}\frac{4\pi}{\Gamma(1/3)}\;\nu_B, \qquad

395: b=3\nu_B, \qquad

396: \alpha=1/3,

397: \eeq

398: and for jitter 

399: \beq

400: a=\frac{e^2}{2c}\pi\delta^2\nu_r,\qquad

401: b=\nu_r,\qquad

402: \alpha=0,\textrm{  and  } \alpha=1.

403: \eeq

404: 

405: 

406: \subsection{Self-absorption frequencies for $\nu_a<\nu_{m}$}

407: 

408: Here we assume that the self-absorption frequency is below the jitter

409: peak in the spectrum from the ensemble of the electrons,

410: $\nu'_a\la\nu'_{m}=2\nu_{j,m}\equiv 2\nu_r\gamma_m^2$.

411: 

412: The opacity is easily calculated from Eqs. (\ref{kappa}), (\ref{P}) to

413: yield 

414: %

415: \beq 

416: \kappa_{\nu'}=\frac{(s+2)(s-1)N_e\; a\;

417: b^{-\alpha}\;(\nu')^{\alpha-2}} {8\pi m_e V_{\rm bw} (s+2\alpha)

418: \gamma_m^{2\alpha+1}}.  

419: \eeq

420: 

421: The self-absorption frequency is that, at which the optical thickness

422: of the blast wave shell is unity:

423: %

424: \beq

425: \kappa_{\nu'_a}\Delta'\sim1.

426: \eeq

427: %

428: This condition gives:

429: %

430: \beq

431: \nu'^{2-\alpha}_a\simeq\frac{(s+2)(s-1)}{(s+2\alpha)}

432: \frac{R\; n_{\rm ISM}\; a\; b^{-\alpha}}{24\pi m_e\gamma_m^{2\alpha+1}}. 

433: \eeq

434: %

435: This equation works for both synchrotron and jitter radiation.  One

436: just need to replace the general parameters $a,\ b,\ \alpha$ with

437: those above, for a process of interest.

438: 

439: For jitter radiation, we have

440: \beq

441: a\;b^{-\alpha}=\frac{e^2}{c}\frac{\pi}{2^\alpha}\delta^2\nu_r^{1-\alpha}.

442: \eeq

443: Therefore,

444: \beq

445: \nu'^{2-\alpha}_a\simeq\frac{(s+2)(s-1)}{(s+2\alpha)}

446: \frac{\pi}{24}

447: \frac{(R/c) \nu_{p,{\rm ISM}}^2\nu_r^{1-\alpha}\delta^2}

448: {\gamma_m^{2\alpha+1}}. 

449: \eeq

450: 

451: 

452: \subsubsection{Case $\alpha=0$}

453: 

454: In this case,

455: \begin{eqnarray}

456: \nu'_a&\simeq&\left(\frac{\pi}{24}\frac{(s+2)(s-1)}{s}\right)^{1/2}

457: \frac{(R/c)^{1/2}\nu_{p,{\rm ISM}}\nu_r^{1/2}\delta}{\gamma_m^{1/2}} 

458: \nonumber\\

459: &\approx&2.92\textrm{ Hz } 

460: R^{1/2}n_{\rm ISM}^{3/4}\gamma_m^{-1/2}\eta^{1/2}\delta,

461: \nonumber\\

462: &\approx&5.36\textrm{ Hz } 

463: R^{1/2}n_{\rm ISM}^{3/4}\gamma_m^{-1/2}\Gamma_{\rm sh}^{1/2}\delta.

464: \end{eqnarray}

465: Hereafter, we use a typical value, $s=2.5$,  in numerical estimates.

466: 

467: 

468: \subsubsection{Case $\alpha=1$}

469: 

470: This case may be relevant to the late afterglow from a jet. 

471: \begin{eqnarray}

472: \nu'_a&\simeq&(s-1)\frac{\pi}{24}

473: \frac{(R/c)\nu_{p,{\rm ISM}}^{2}\delta^2}{\gamma_m^{3}} 

474: \nonumber\\

475: &\approx&5.28\times10^{-4}\textrm{ Hz } R\,n_{\rm ISM}\,\gamma_m^{-3}\delta^2.

476: \end{eqnarray}

477: 

478: 

479: \subsection{Self-absorption frequency for $\nu_a>\nu_m$}

480: 

481: In order to calculate the self-absorption frequency in this regime, 

482: we need the full expression for the emissivity function:

483: %\begin{eqnarray}

484: \beq

485: P(\nu',\gamma)%&=&\frac{e^2}{2c}\,\delta\,\frac{2\pi\nu_j}{\gamma^2}\,

486: %J\left(\frac{\nu'}{\nu_j}\right) 

487: %\nonumber\\&=&

488: =\frac{e^2}{2c}\,\delta\,2\pi\nu_r\,

489: J\left(\frac{\nu'}{\nu_r\gamma^2}\right), 

490: \eeq

491: %\end{eqnarray}

492: where the function $J(\xi)$ for $\mu\gg1$ and $\delta\ll1$ is

493: \begin{eqnarray}

494: J(\xi)&=&(2\mu+1)\xi^{2\mu}\left[I(2)-I(\xi)\right], \\

495: I(\xi)&=&-\left(\frac{\xi^{-2\mu+1}}{2\mu-1}-

496: \frac{\xi^{-2\mu+2}}{2\mu-2}+\frac{1}{2}\frac{\xi^{-2\mu+3}}{2\mu-3}\right).

497: \end{eqnarray}

498: This function corresponds to the $\alpha=1$ case. The expression for

499: the $\alpha=0$ is more complicated. Our analysis indicates, however,

500: that the jitter emissivity function is approximated by a sharply

501: broken power-law very well (much better than the synchrotron one does).

502: Therefore, we will use the approximate expression for $J(\xi)$,

503: which describes well the $\alpha=1$ and $\alpha=0$ spectra.

504: We use

505: \beq

506: J(\xi)=\left\{

507: \begin{array}{ll}

508: \xi^\alpha, & \textrm{  if } \xi<2,\\

509: 0, & \textrm{  if } \xi\ge2.

510: \end{array}\right.

511: \eeq

512: This function mimics a power-law spectrum up to the

513: peak jitter frequency $\nu_{j}\simeq 2\nu_r\gamma^2$,

514: with the sharp cutoff above it.

515: 

516: The opacity is:

517: \begin{eqnarray}

518: \kappa_{\nu'}=

519: \frac{(s+2)(s-1)N_e\gamma_m^{s-1}}{8\pi m_e V_{\rm bw}\nu'^2}

520: \frac{e^2}{2c}\delta 2\pi\nu_r \int_{\gamma_m}^\infty 

521: J\left(\frac{\nu'}{\nu_r\gamma^2}\right)\gamma^{-s-1}\, d\gamma.

522: \end{eqnarray}

523: Upon substitution of $J(\xi)$ the integral becomes

524: \beq

525: \frac{1}{2}\left(\frac{\nu}{\nu_r}\right)^{-s/2}

526: \int_0^{\textrm{min}\!\left({\nu'}/({\nu_r\gamma_m^2});~ 2\right)}

527: z^{\alpha+{s}/{2}-1}\; dz.

528: \eeq

529: We are interested in the high-energy part of the spectrum, above

530: $\nu'\gg\nu'_m\simeq2\nu_r\gamma_m^2$. Therefore, the upper limit in

531: the integral is equal to 2. Thus, we have for the opacity at

532: frequencies above the spectral peak:

533: \beq

534: \kappa_{\nu'}=\frac{(s+2)(s-1)}{(s+2\alpha)}\frac{2^{s/2+\alpha-3}}{12\pi}

535: \frac{(R/c)\omega_{p,e,{\rm ISM}}^2}{\Delta'\gamma_m^3}\frac{\delta^2}{\nu'_m}

536: \left(\frac{\nu'}{\nu'_m}\right)^{-s/2-2}.

537: \eeq

538: Since $F_\nu\propto\nu^{5/2}$ for $\nu_m<\nu<\nu_a$, rather than

539: $\propto\nu^2$, the absorption frequency should be defined as the

540: frequency at which $dF_\nu/d\nu=0$, where

541: $F_\nu\propto\nu^{5/2}(1-e^{-\kappa_{\nu'}\Delta'})$.  For a simple

542: estimate, we can still use the condition:

543: $\kappa_{\nu'_a}\Delta'\simeq1$. We have 

544: \beq 

545: \nu'_a\simeq\left(

546: 2^{s/2+\alpha-3}\frac{\pi}{3}\frac{(s+2)(s-1)}{(s+2\alpha)}\right)^{2/(s+4)}

547: \left[\frac{(R/c)\nu_{p,{\rm

548: ISM}}^2\nu'^{s/2+1}_m\delta^2}{\gamma_m^3} \right]^{2/(s+4)}.

549: \label{nu-a39}

550: \eeq

551: The peak frequency, $\nu'_m$, is calculated in the next section, 

552: Eq. (\ref{nu-m}). To proceed furter, we again use the typical value of 

553: the electron power-law exponent, $s=2.5$. For such $s$, the 

554: values of the numerical factor (the first term) in Eq. (\ref{nu-a39}) is

555: equal to 0.95 and 0.99 for $\alpha=0$ and $\alpha=1$, respectively. 

556: We, therefore

557: use the representative value of 0.97 for both cases, which gives

558: an error in the $\nu'_a$ value to within a couple of percents. Thus,

559: \beq

560: \nu'_a\approx320\textrm{ Hz } R^{0.31}n_{\rm ISM}^{0.65}

561: \Gamma_{\rm sh}^{0.69}\gamma_m^{0.46}\delta^{0.62}.

562: \eeq

563: 

564: \subsection{Peak frequency}

565: 

566: The peak frequency of the shock spectrum (from an ensemble of electrons)

567: is determined by the jitter frequency at $\gamma_m$:

568: \begin{eqnarray}

569: \nu'_m&\simeq&2\nu_r\gamma_m^2 \nonumber\\

570: &\approx&1.80\times10^4\textrm{ Hz } n_{\rm ISM}^{1/2}\gamma_m^2\eta

571: \nonumber\\ 

572: &\approx&6.04\times10^4\textrm{ Hz } n_{\rm

573: ISM}^{1/2}\Gamma_{\rm sh}\gamma_m^2.

574: \label{nu-m}

575: \end{eqnarray}

576: 

577: 

578: \subsection{Cooling break frequency}

579: 

580: The total (integrated over frequencies) emitted power by an electron

581: is identical in jitter and synchrotron regimes.  Thus, the cooling

582: break is unchanged. We quote results from \citet{SPN98} 

583: \beq

584: \nu'_c=(3/2)\nu_B\gamma_c^2, 

585: \eeq 

586: where the cooling Lorentz factor (neglecting Compton losses) is 

587: \beq

588: \gamma_c=\frac{3m_e}{16m_pc\sigma_T\epsilon_B} \frac{1}{n_{\rm

589: ISM}\Gamma_{\rm sh}t_{\rm loc}} 

590: \eeq 

591: where $t_{\rm loc}=t/(1+z)$ is the cosmologically local time for a

592: GRB. Thus, 

593: \beq 

594: \nu'_c\approx4.29\times10^{24}\textrm{ Hz } n_{\rm

595: ISM}^{-3/2} \Gamma_{\rm sh}^{-5}\epsilon_B^{-3/2} t_{\rm loc}^{-2}.

596: \eeq

597: 

598: \section{Observables}

599: 

600: The power-law segments in $\nu_a<\nu_m$ and $\nu_a>\nu_m$ regimes are, 

601: respectively:

602: \beq

603: F_\nu^{(\nu_a<\nu_m)}\propto\left\{

604: \begin{array}{ll}

605: \nu^2, & \textrm{ if   } \nu<\nu_a; \\

606: \nu^\alpha, & \textrm{ if   } \nu_a<\nu<\nu_m; \\

607: \nu^{-(s-1)/2}, & \textrm{ if   } \nu_m<\nu<\nu_{c}; \\

608: \nu^{-s/2}, & \textrm{ if   } \nu_{c}<\nu.

609: \end{array}

610: \right.

611: \eeq

612: \beq

613: F_\nu^{(\nu_a>\nu_m)}\propto\left\{

614: \begin{array}{ll}

615: \nu^2, & \textrm{ if   } \nu<\nu_m; \\

616: \nu^{5/2}, & \textrm{ if   } \nu_m<\nu<\nu_a; \\

617: \nu^{-(s-1)/2}, & \textrm{ if   } \nu_a<\nu<\nu_{c}; \\

618: \nu^{-s/2}, & \textrm{ if   } \nu_{c}<\nu.

619: \end{array}

620: \right.

621: \eeq

622: 

623: 

624: \subsection{Frequencies: ISM}

625: 

626: We now calculate the frequencies in the observer's frame 

627: (thus, all frequencies are boosted by $\Gamma_{\rm sh}/(1+z)$) and their

628: dependencies on the burst parameters for $s=2.5$.

629: \begin{eqnarray}

630: \nu_a^{(\alpha=0)}&\approx& 5.88\times10^{8}\textrm{ Hz }

631: (1+z)^{-3/4}E_{52}^{1/4}\epsilon_e^{-1/2}\delta\,

632: n_{\rm ISM,0}^{1/2} t_{\rm days}^{-1/4}, 

633: \\

634: \nu_a^{(\alpha=1)}&\approx& 9.56\times10^{4}\textrm{ Hz }

635: (1+z)^{-2}E_{52}^{0}\epsilon_e^{-3}\delta^2

636: n_{\rm ISM,0}^{} t_{\rm days}^{},

637: \\

638: \nu_a^{(>\nu_m)}&\approx& 3.13\times10^{10}\textrm{ Hz }

639: (1+z)^{-0.27} E_{52}^{0.35}\epsilon_e^{0.46}\delta^{0.62}

640: n_{\rm ISM,0}^{0.30}t_{\rm days}^{-0.73},

641: \\

642: \nu_m&\approx& 4.02\times10^{12}\textrm{ Hz }

643: (1+z)^{1/2}E_{52}^{1/2}\epsilon_e^2 t_{\rm days}^{-3/2},

644: \\

645: \nu_c&\approx& 8.54\times10^{16}\textrm{ Hz } (1+z)^{-1/2}

646: E_{52}^{-1/2}\epsilon_{B,-3}^{-3/2}n_{\rm ISM,0}^{-1}t_{\rm days}^{-1/2}.

647: \end{eqnarray}

648: Here we also quoted the result from \citet{SPN98} for the

649: cooling frequency in the adiabatic regime of blast wave evolution

650: in the constant ISM density environment,

651: corrected for the redshift.

652: 

653: Finally, we compute the peak flux of jitter afterglows by using the

654: relation $F_{\nu,\max}=\delta^2 F_{\nu,\max,{\rm synch}}

655: \nu_m/\nu_{m,{\rm synch}}$ and the synchrotron peak flux from Sari et

656: al. (1998):

657: %

658: \begin{equation}

659: F_{\nu,\max} = 10^3 \,E_{52}\,\epsilon_{B,-3}\,n_{\rm{ISM},0}\,

660: D_{28}^{-2}\,\mu Jy

661: \end{equation}

662: 

663: We remind that the jitter deflection parameter is related 

664: to the magnetic field equipartition parameter as follows:

665: \beq

666: \delta\approx36.0\epsilon_B^{1/2}\approx1.12\epsilon_{B,-3}^{1/2}.

667: \eeq

668: 

669: Figure~\ref{fig:spex} shows the jitter and synchrotron spectra for an

670: afterglow with $E=10^{53}$~erg, $\epsilon_e=0.1$, $\epsilon_B=0.0001$,

671: $n_{ISM}=1$ and $s=2.5$, computed at $t=$0.1, 1 and 10 days.

672: 

673: %

674: \begin{figure}

675: \psfig{file=f1.ps}

676: \caption{{Jitter and synchrotron spectra for a typical afterglow

677: running into a uniform medium. The parameter set is: $E=10^{53}$~erg,

678: $\epsilon_e=0.1$, $\epsilon_B=0.0001$, $n_{\rm ISM}=1$ and the electron

679: energy distribution power-law index $s=2.5$,

680: computed at $t=$0.1, 1 and 10 days.}

681: \label{fig:spex}}

682: \end{figure}

683: %

684: 

685: \subsection{Frequencies: Wind}

686: 

687: We now calculate the frequencies in the observer's frame 

688: in the wind environment.

689: \begin{eqnarray}

690: \nu_a^{(\alpha=0)}&\approx& 2.79\times10^{9}\textrm{ Hz }

691: (1+z)^{-1/4}E_{52}^{0}\epsilon_e^{-1/2}\delta\,

692: A_*^{} t_{\rm days}^{-3/4}, 

693: \\

694: \nu_a^{(\alpha=1)}&\approx& 1.80\times10^{5}\textrm{ Hz }

695: (1+z)^{-1}E_{52}^{-3/2}\epsilon_e^{-3}\delta^2

696: A_*^{2} t_{\rm days}^{0},

697: \\

698: \nu_a^{(>\nu_m)}&\approx& 1.41\times10^{11}\textrm{ Hz }

699: (1+z)^{0.04} E_{52}^{0.58}\epsilon_e^{0.46}\delta^{0.62}

700: A_*^{0.62}t_{\rm days}^{-1.04},

701: \\

702: \nu_m&\approx& 2.40\times10^{13}\textrm{ Hz }

703: (1+z)^{1/2}E_{52}^{3/2}\epsilon_e^2 t_{\rm days}^{-3/2},

704: \\

705: \nu_c&\approx& 5.66\times10^{15}\textrm{ Hz } (1+z)^{-3/2}

706: E_{52}^{1/2}\epsilon_{B,-3}^{3/2}A_*^{-2}t_{\rm days}^{1/2}.

707: \end{eqnarray}

708: Here we also quoted the result from \citet{CL00} for the

709: cooling frequency in the adiabatic regime of blast wave evolution

710: in the wind environment. 

711: 

712: \section{Discussion}

713: 

714: We considered in this work the propertiies of GRB afterglows with

715: radiation produced by jitter radiation instead of synchrotron. For the

716: first time we evaluate the self absorption frequency in various

717: regimes and for blastwaves rpopagating in different ambient media.

718: 

719: Within the present framework, we analyzed two possible regimes of the

720: jitter mechanism. If the post-shock magnetic turbulence is isotropic

721: (which is very likely in the far downstream region), then

722: $\alpha=0$. If the turbulence remains anisotropic throughout the shell

723: than (i) if we observe a shock before the jet break or if the outflow

724: is spherical, then still $\alpha=0$. However, (ii) for anisotropic

725: turbulence and late times (long after the jet break), one can have

726: $\alpha=1$. It is likely that one can have an intermediate regime for

727: most of the time, as the properties of turbulence likely vary

728: downstream as a function of the distance from the shock front.

729: 

730: Note that in the jitter regime, the peak frequency is independent of

731: the magnetic field strength. In general, it depends on the ambient

732: density. However, for the assumed parameter $\eta\propto\Gamma_{\rm

733: sh}$, which incorporates all details (not so well known) of the

734: magnetic field evolution far downstream, the jitter peak is

735: independent of the density at all, either $n_{\rm ISM}$ or $A_*$.

736: Note also that the jitter self-absorption frequencies and the peak

737: frequency strongly depend on the electron equipartition parameter.  It

738: may be helpful to remember that the jitter peak frequency is higher

739: than the synchrotron peak frequency in the field of the same strength

740: (same $\epsilon_B$) and the same electron energy distribution (same

741: $\epsilon_e$ and the electron index $s$) by a factor $\sim\delta^{-1}$, 

742: which, in turn, is $\propto\epsilon_B^{1/2}$.

743: 

744: Finally, we estimate the times when the emission is becomes optically 

745: thick, $\nu_a\sim\nu_m$ for both $\alpha$'s, for the ISM case.

746: \begin{eqnarray}

747: t_a^{(\alpha=0)}&\sim&1170\textrm{ days }(1+z)E_{52}^{1/5}

748: \epsilon_e^{2}\delta^{-4/5}n_{\rm ISM,0}^{-2/5}, \\

749: t_a^{(\alpha=1)}&\sim& 1120\textrm{ days }(1+z)E_{52}^{1/5}

750: \epsilon_e^{2}\delta^{-4/5}n_{\rm ISM,0}^{-2/5},\\

751: t_a^{(\nu_m<\nu_a)}&\sim& 548\textrm{ days }(1+z)E_{52}^{0.20}

752: \epsilon_e^{2.00}\delta^{-0.80}n_{\rm ISM,0}^{0.40}

753: \end{eqnarray}

754: Ideally, these times should coincide. The discrepancies are due to the

755: approximations made in our analysis. In particular, the

756: self-absorption frequency for $\nu_m<\nu_a$ is overestimated by a

757: factor of two (as is explained above), hence the thin-to-thick

758: transition time is earlier. A more detailed treatment of the

759: self-absorption frequencies will be presented elsewere (Jared et

760: al. in preparation; Morsony et al. in preparation).  Also, the time when

761: the cooling break is equal to the peak, $\nu_c\sim\nu_c$ is 

762: \beq

763: t_c\sim4.71\times10^{-5} \textrm{ days }

764: (1+z)E_{52}\epsilon_e^2\epsilon_{B,-3}^{3/2}n_{\rm ISM,0}.  

765: \eeq 

766: The numerical factor in the above equation is equal to 4.07 seconds.

767: 

768: Similartly, we calculate $t_a$'s and $t_c$ for the wind case. 

769: \begin{eqnarray}

770: t_a^{(\alpha=0)}&\sim&1.76\times10^{5}\textrm{ days }(1+z)E_{52}^{2}

771: \epsilon_e^{10/3}\delta^{-4/3}A_*^{-4/3}, \\

772: t_a^{(\alpha=1)}&\sim& 2.61\times10^{5}\textrm{ days }(1+z)E_{52}^{2}

773: \epsilon_e^{10/3}\delta^{-4/3}A_*^{-4/3}, \\

774: t_a^{(\nu_m<\nu_a)}&\sim& 7.08\times10^{4}\textrm{ days }(1+z)E_{52}^{2.00}

775: \epsilon_e^{-3.35}\delta^{-1.34}A_*^{-1.34}, \\

776: %(1+z)E_{52}^{0.20}

777: %\epsilon_e^{2.00}\delta^{-0.80}n_{\rm ISM,0}^{0.40}\\

778: t_c&\sim&6.51\times10^{-2} \textrm{ days }

779: (1+z)E_{52}^{1/2}\epsilon_e^{}\epsilon_{B,-3}^{-3/4}A_*.

780: \end{eqnarray}

781: The discrepancies in $t_a$'s are, again, due to the approximations

782: made in calculating $\nu_a$.

783: 

784: Figure~\ref{fig:spex} allows us to comment on the differences between

785: synchrotron and jitter afterglows. The high energy part of the

786: spectrum (mainly the optical and X-ray regimes) are hardly

787: distinguishable between the two mechanisms, especially if, as expected

788: from simulation, jitter fields are created such that

789: $\delta\lesssim1$.

790: 

791: The two main differences are in the low energy branches, around when

792: most radio observations are performed. First, the spectral slope at

793: the left of the peak frequency is flat, $\propto\nu^0$, rather than

794: the canonical $\nu^{1/3}$; second, the location of the self-absorption

795: break is different and evolves in time differently than in synchrotron

796: afterglows. Since one important observation to nail down the density

797: of the ambient medium is the radio regime, modelling afterglows with

798: jitter radiation may lead to different results compared to those of

799: synchtron modelling (Morsony et al. and Workman et al., in

800: preparation).

801: 

802: \acknowledgements

803: 

804: MM gratefully acknowledges support from 

805: the Institute for Advanced Study.

806: This work was supported by NASA grants NNG-04GM41G (MM) and

807: NNG-06GI06G (BM, DL), Swift Guest Investigator grant 06-SWIFT306-0001

808: (MM) and NNX06AB69G (BM, DL), DoE grant DE-FG02-04ER54790 (MM), and

809: NSF grant AST-0307502 (BM, DL).

810: 

811: %Work by MM has been supported by NASA grant NNG-04GM41G, Swift Guest

812: %Investigator grant 06-SWIFT306-0001 and DoE grant DE-FG02-04ER54790.

813: 

814: 

815: 

816: \begin{thebibliography}{dummy}

817: %

818: \bibitem[Chevalier \& Li(2000)]{CL00} Chevalier, R. A., \& Li,

819: Z.-Y. 2000, \apj, 536, 195

820: %

821: \bibitem[Dermer et al.(2000)]{DBC00} Dermer, C. D., \& B\"ottcher, M.,

822: \& Chiang, J. 2000, \apj, 537, 225

823: %

824: \bibitem[Fleishman(2006)]{Fleish06}

825: Fleishman, G. 2006, \apj, 638, 348

826: %

827: \bibitem[Frederiksen et al.(2004)]{fred04} Frederiksen, J. T.,

828: Hededal, C. B., Haugb\o lle, T., Nordlund, \AA. 2004 \apjl, 608, L13

829: %

830: \bibitem[Granot et al.(1999)]{GPS99} Granot, J., Piran, T., \& Sari,

831: R. 1999, \apj, 527, 236

832: %

833: \bibitem[Meszaros \& Rees(1997)]{MR97} Meszaros, P., \& Rees, M.~J.\

834: 1997, \apj, 476, 232

835: %

836: \bibitem[Medvedev \& Loeb(1999)]{ML99} Medvedev, M. V., \& Loeb,

837: A. 1999, \apj, 526, 697

838: %

839: \bibitem[Medvedev(2000)]{M00} Medvedev, M. V. 2000, \apj, 540, 704

840: %

841: \bibitem[Medvedev(2006a)]{M06} Medvedev, M.~V.\ 2006a, \apj, 637, 869

842: %

843: \bibitem[Medvedev(2006b)]{M06b} Medvedev, M.~V.\ 2006b, \apjl, 651, L9

844: %

845: \bibitem[Medvedev, et al.(2005)]{M+05} Medvedev, M. V., Fiore, M.,

846: Fonseca, R. A., Silva, L O., Mori, W. B. 2005, \apjl, 618, L75

847: %

848: \bibitem[Medvedev, Silva, Kamionkowski(2006)]{MSK05} Medvedev, M. V.,

849: Silva, L. O., \& Kamionkowski, M. 2006, \apjl, 642, L1

850: %

851: \bibitem[Nishivawa, et al.(2003)]{Nish+03}

852: Nishikawa, K.-I., Hardee, P., Richardson, G., Preece, R., Sol, H., \&

853: Fishman, G. J. 2003, \apj, 595, 555 

854: %

855: \bibitem[Panaitescu(2006)]{P06}

856: Panaitescu, A. 2006, \mnras, 366, 1357

857: %

858: \bibitem[Panaitescu \& Kumar(2001)]{PK01}

859: Panaitescu, A., \& Kumar, P. 2001, \apjl, 560, L49

860: %

861: \bibitem[Piran(1999)]{P99} Piran, T.\ 1999, \physrep, 314, 575

862: %

863: \bibitem[Sari et al.(1998)]{SPN98} Sari, R., Piran, T., \& Narayan,

864: R. 1998, \apjl, 497, L17

865: %

866: \bibitem[Silva, et al.(2003)]{Silva+03} Silva, L.~O., Fonseca, R.~A.,

867: Tonge, J.~W., Dawson, J.~M., Mori, W.~B., \& Medvedev, M.~V.\ 2003,

868: \apjl, 596, L121

869: %

870: \bibitem[Spitkovski(2005)]{Spit05}

871: Spitkovsky, A. 2005, in AIP Conf. Proc. 801, Astrophysical Sources of High

872: Energy Particles and Radiation, ed. T. Bulik \& B. Rudak (Melville: AIP), 345

873: %

874: \bibitem[Waxman(1997)]{Wax97}

875: Waxman, E. 1997, \apjl, 491, L19

876: %

877: \end{thebibliography}

878: 

879: \rem{

880: %\clearpage

881: %

882: \begin{figure}

883: %\psfig{file=f1.eps,width=3.5in}

884: \caption{ 

885: \label{f:1} }

886: \end{figure} 

887: %

888: }

889: 

890: \end{document}

891: 

892: 

893: