1: %%

2: %% Beginning of file 'sample.tex'

3: %%

4: %% Modified 2004 January 9

5: %%

6: %% This is a sample manuscript marked up using the

7: %% AASTeX v5.x LaTeX 2e macros.

8: 

9: 

10: 

11: %% The command below calls the preprint style

12: %% which will produce a one-column, single-spaced document.

13: %% Examples of commands for other substyles follow. Use

14: %% whichever is most appropriate for your purposes.

15: %%

16: \documentclass[12pt,preprint]{aastex}

17: 

18: %% manuscript produces a one-column, double-spaced document:

19: 

20: %% \documentclass[manuscript]{aastex}

21: 

22: %% preprint2 produces a double-column, single-spaced document:

23: % \documentclass[preprint2]{aastex}

24: 

25: %% \documentclass[preprint2,longabstract]{aastex}

26: 

27: 

28: \newcommand{\nablaa}{\nabla_{\alpha}}

29: \newcommand{\nablab}{\nabla_{\beta}}

30: \newcommand{\pad}{\partial}

31: \newcommand{\md}{\mbox{d}}

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

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

34: \newcommand{\beqn}{\begin{eqnarray}}

35: \newcommand{\eeqn}{\end{eqnarray}}

36: \newcommand{\lppr}{\stackrel{<}{\scriptstyle \sim}}

37: \newcommand{\gppr}{\stackrel{>}{\scriptstyle \sim}}

38: 

39: 

40: \slugcomment{ApJ 632 (2005), L21}

41: \shorttitle{Particle acceleration in GRB jets}

42: \shortauthors{}

43: 

44: 

45: \begin{document}

46: 

47: \title{Particle Acceleration in Gamma-Ray Burst Jets}

48: \author{Frank M. Rieger and Peter Duffy}

49: \affil{UCD School of Mathematical Sciences, University College Dublin,

50:        Dublin 4, Ireland}

51: \email{frank.rieger@ucd.ie; peter.duffy@ucd.ie}

52: 

53: 

54: \begin{abstract}

55: Gradual shear acceleration of energetic particles in gamma-ray burst 

56: (GRB) jets is considered. Special emphasis is given to the analysis 

57: of universal structured jets, and characteristic acceleration timescales 

58: are determined for a power-law and a Gaussian evolution of the bulk 

59: flow Lorentz factor $\gamma_b$ with angle $\phi$ from the jet axis. 

60: The results suggest that local power-law particle distributions may be 

61: generated and that higher energy particles are generally concentrated 

62: closer to the jet axis. Taking several constraints into account we show 

63: that efficient electron acceleration in gradual shear flows, with maximum 

64: particle energy successively decreasing with time, may be possible on 

65: scales larger than $r \sim 10^{15}$ cm, provided the jet magnetic field 

66: becomes sufficiently weak and/or decreases rapidly enough with distance, 

67: while efficient acceleration of protons to ultra-high energies $> 10^{20}$ 

68: eV may be possible under a wide range of conditions.

69: \end{abstract}

70: \keywords{gamma rays: bursts --- acceleration of particles}

71: 

72: 

73: \section{Introduction}

74: There is mounting evidence today that gamma-ray bursts (GRBs) are associated 

75: with collimated relativistic outflows or jets \citep[][]{rho99,kul99,gre03}. 

76: It is still a matter of ongoing research, however, what types of internal 

77: structures are actualized within these jets. Within the fireball framework, 

78: for example, two different kinds of jet models seem to be compatible with the 

79: observations, i.e., (i) the uniform "top-hat" jet model \citep[][]{rho99,

80: fra01,lam04}, in which all the hydrodynamical quantities (e.g., Lorentz factor, 

81: energy density) are essentially the same within some well-defined opening angle 

82: $\phi_j$ around the jet axis but sharply drop outside of $\phi_j$, and (ii) 

83: the universal structured ("power-law" or "Gaussian") jet model \citep[][]{ros02,

84: zha02,kum03,zha04}, in which the hydrodynamical quantities are rather smooth 

85: functions of the angle $\phi$ from the jet axis.\\

86: Here we analyze the implications of such jet structures for the acceleration 

87: of energetic particles. We focus on shear acceleration as a promising mechanism 

88: for converting the kinetic energy of the flow into nonthermal particles and 

89: radiation. Such a mechanism has previously been successfully applied to the 

90: relativistic jets in Active Galactic Nuclei \citep[][]{ost98,ost00,rie04,rie05a,

91: rie05b}.

92: Shear acceleration is based on the simple fact that particles may gain energy 

93: by scattering off (small-scale) magnetic field irregularities with different 

94: local velocities due to being systematically embedded in a collisionless shear 

95: flow \citep[cf.][for a recent review]{rie05b}. In the case of nonrelativistic 

96: shear acceleration it is straightforward to show that local power-law particle 

97: momentum distributions $f(p) \propto p^{-\,(3 + \alpha)}$ can be generated, 

98: assuming a momentum-dependent mean scattering time of the form $\tau \propto 

99: p^{\alpha}$ with $\alpha>0$ \citep[cf.][]{berk81,rie05a}. Hence, for $\tau$ 

100: scaling with the gyro-radius, i.e., $\tau \propto p$, this results in a 

101: power-law particle number density $n(p)\propto p^2\,f(p) \propto p^{-2}$, and 

102: thus a synchrotron emissivity $j_{\nu}\propto \nu^{-1/2}$.

103: 

104: 

105: \section{Shear acceleration in structured GRB-type jets}

106: Let us consider an idealized radial, relativistic flow profile, which in four 

107: vector notation is given by \citep[cf.][for the more general case]{rie05c}

108: \beq

109:    u^{\alpha}= \gamma_b\,\left(1,v_r(\phi)/c,0,0\right)\,,

110: \eeq where $\alpha=0,1,2,3$, and $\phi$ denotes the polar angle in spherical

111: coordinates, and $\gamma_b \equiv \gamma_b(\phi)=[1-v_r(\phi)^2/c^2]^{-1/2}$ 

112: is the bulk Lorentz factor of the flow. In the comoving frame, the related 

113: gradual shear acceleration coefficient can be cast into the form \citep[e.g.,]

114: [eq.~3.27]{web89}

115: \beq

116:    < \dot{p}\,'\,> = \frac{1}{p\,'^{\,2}}\frac{\pad }{\pad p\,'}

117:                 \left(p\,'^{\,4}\,\tau'\,\Gamma\right)

118: \eeq where $p'$ denotes the comoving particle momentum, $\tau' \simeq \lambda'

119: /c$ is the mean scattering time and $\Gamma$ is the relativistic shear coefficient. 

120: We are interested in the strong scattering limit (corresponding to $\omega_g\,'

121: \,\tau' \ll 1$, with $\omega_g'$ being the relativistic gyro-frequency measured in 

122: the comoving jet frame, i.e., assuming a sufficiently weak longitudinal mean 

123: magnetic field and the presence of strong turbulence, so that collisions are 

124: efficient enough to restore isotropy), where the shear coefficient is given by 

125: $\Gamma =(c^2/30)\,\sigma_{\alpha \beta}\,\sigma^{\alpha \beta}$\citep[see][eq.~3.34]

126: {web89} and $\sigma_{\alpha \beta}$, with $\alpha,\beta=0,1,2,3$, is the usual 

127: covariant fluid shear tensor \citep[cf.][for more details]{web89,rie04,rie05c}. 

128: Using spherical coordinates and the velocity profile above, it can be shown 

129: that the relativistic shear coefficient becomes \citep[cf.][]{rie05c} 

130: \beq\label{shear-coeff}

131:   \Gamma = \frac{4}{45}\,\gamma_b^2 \left[\frac{v_r^2}{r^2}+\frac{3}{4\,r^2}\,

132:            \gamma_b^2\,\left(\frac{\pad v_r}{\pad \phi}\right)^2\right]\,,

133: \eeq which in the nonrelativistic limit ($\gamma_b \rightarrow 1$) reduces 

134: to the (nonrelativistic) viscous transfer coefficient derived by \citet{ear88} 

135: (their eq.~7) when the latter is expressed in spherical coordinates and the

136: corresponding velocity profile $\vec{v} = v_r(\phi)\,\vec{e}_r$ is applied. 

137: The (comoving) timescale $t_{\rm acc} \simeq < p' >/< \dot{p}' >$ for the 

138: shear flow acceleration of particles then becomes 

139: \beq\label{tshear}

140:     t_{\rm acc}(r,\phi) \simeq \frac{45}{4 (4 + \alpha)}\,

141:                \frac{c}{\lambda'}\,

142:                 \frac{r^2}{\gamma_b^2\,\left[v_r^2 + 0.75\,\gamma_b^2\,   

143:                 (\pad v_r/\pad \phi)^2\right]}\,,

144: \eeq where $r$ is the radial coordinate measured in the cosmological rest 

145: frame (i.e., the supernova, collapsar or merger rest frame), and where a power-law

146: dependence $\tau'=\lambda'/c \propto p'^{\alpha}$ has been assumed. As the jet 

147: flow is diverging, $t_{\rm acc}$ obviously increases with $r$ to the square.\\

148: In order to investigate the acceleration potential of structured GRB jets, two 

149: applications appear particularly interesting \citep[e.g.,][]{zha02,kum03,zha04}: 

150: a power-law model, in which $\gamma_b$ is a power-law function of $\phi$ 

151: outside a core of opening angle $\phi_c$, i.e., $\gamma_b(\phi) = 1 + (\gamma_{b0} 

152: -1)(1+[\phi/\phi_c]^2)^{-b/2}$, with $1.5 < b \lppr 2$ \citep[cf.][]{zha02}, 

153: and a Gaussian model with $\gamma_b(\phi) = 1 + (\gamma_{b0} -1) \exp[-\phi^2/

154: 2\phi_c^2]$, where $\gamma_{b0}$ denotes the Lorentz factor at the jet axis, and

155: typically $\phi_c = 0.1$ rad \citep[][]{zha04}. The shear acceleration timescale 

156: (with the time and space coordinates measured in the cosmological rest frame and 

157: particle Lorentz factor measured in the comoving jet frame) then becomes 

158: \beq\label{tshear1}

159:   t_{\rm acc}(r,\phi) = \frac{45}{4 (4 + \alpha)}\,

160:       \frac{r^2}{c\,\lambda'\,\gamma_b(\phi)^2}\,\times 

161:       \left\{ \begin{array}{rl}

162:       \left(\frac{v_r^2}{c^2}+\frac{3}{4}\,\frac{\left(\gamma_b(\phi)-1\right)}

163:       {\gamma_b(\phi)^2 \left(\gamma_b(\phi)+1\right)}\frac{b^2}{(1+\phi^2/\phi_c^2)^2}

164:       \,\frac{\phi^2}{\phi_c^4}\right)^{-1} & {\rm power-law} \\

165:       \left(\frac{v_r^2}{c^2}+\frac{3}{4}\,\frac{\left(\gamma_b(\phi)-1\right)}

166:       {\gamma_b^2(\phi)\left(\gamma_b(\phi)+1\right)}\frac{\phi^2}{\phi_c^4}\right)^{-1} 

167:       & {\rm Gaussian\,model}

168:       \end{array}\,. \right. 

169: \eeq Note that in contrast to nonrelativistic parallel shock acceleration, 

170: $t_{\rm acc} \propto 1/\lambda$ as the probability of a particle sampling a higher

171: shear, and thus a more energetic scattering event increases with $\lambda$.

172: The general evolution of $t_{\rm acc}/t_0$ as a function of $\phi$, with $t_0 = 45\,r^2/[4

173: (4+\alpha)\,\lambda'\, c]$, is illustrated in Fig.~\ref{fig1}, indicating that a power-law 

174: model becomes somewhat more favourable than a Gaussian model. Obviously, the shear 

175: acceleration timescale in structured jets generally increases with $\phi$ due to the 

176: decrease in $\gamma_b$, suggesting different maximum attainable energies in the local 

177: comoving frame. Hence, at constant $r$ the higher energy particles and thus the higher 

178: energy emission are expected to be naturally concentrated closer to the jet axis, i.e., 

179: toward smaller $\phi$ (provided the magnetic field is not very inhomogeneous across 

180: the jet), enforcing the effect that observers looking at the same GRB from different 

181: directions will see different gamma-ray light curves.

182: 

183: In general several constraints need to be satisfied in order to allow for 

184: efficient particle acceleration in GRB jets, e.g., the particle acceleration process 

185: may be limited by (i) radiative synchrotron losses, (ii) the transversal confinement 

186: size, (iii) particle escape via cross-field diffusion or (iv) the duration of the jet 

187: expansion. The first constraint implies that the acceleration timescale has to be smaller 

188: than the comoving synchrotron cooling timescale, which, for particles with Lorentz 

189: factors $\gamma'$ and isotropic pitch angle distribution, is given by

190: \beq\label{tcool}

191:   t_{\rm cool} =\frac{9\,m^3\,c^5}{4\,e^4}\frac{1}{\gamma'\,B'^2}

192:                  \nonumber

193:                = 4.8\cdot 10^{12} \,\left(\frac{1}{\gamma'}\right)\,

194:                   \left(\frac{m}{m_p}\right)^3

195:                    \left(\frac{10^3\;{\rm G}}{B'}\right)^{2}\;[{\rm s}]\,,

196: \eeq with $m_p$ being the proton mass. According to eq.~(\ref{tshear1}) the minimum  

197: shear acceleration timescale is 

198: \beq\label{tmin}

199:  t_{\rm acc} \sim \frac{5}{2}\,\frac{r^2}{c\,\lambda'\,\gamma_{b}^2}\,,

200: \eeq so that for a gyro-dependent particle mean free path $\lambda' = \xi\, r_g'$, 

201: with $\xi < 1$, (as motivated by the observationally implied particle spectra with 

202: momentum index $\sim -2$) the ratio of shear to cooling timescale becomes independent 

203: of the particle Lorentz factor. Note that eq.~(\ref{tmin}) is also the timescale for a 

204: uniform flow with $v_r$ independent of $\phi$ (cf. eq.~[\ref{tshear}])! For a given 

205: magnetic field strength the acceleration process may thus work efficiently as long as 

206: the velocity shear remains sufficiently high, i.e., as long as the radial coordinate 

207: satisfies

208: \beq

209:  r < 4 \cdot 10^{15}\, \xi^{1/2}\left(\frac{m}{m_p}\right)^2 

210:          \left(\frac{ 10^3\;{\rm G}}{B'}\right)^{3/2} 

211:           \left(\frac{\gamma_{b0}}{300}\right)\;\,{\rm [cm]}\,,

212: \eeq suggesting that, in contrast to the acceleration of protons, efficient electron 

213: acceleration is suppressed in the presence of high magnetic fields. It seems very likely, 

214: however, that due to the expansion of the wind, the comoving magnetic field $B'$ will 

215: depend inversely on $r$, i.e., $B' \propto 1/r^{\beta}$ with $\beta > 0$. In order to 

216: study possible implications in more details, we may thus consider the following simple 

217: parametrization for the commoving jet magnetic field strength, $B'= 1000\,b_0\,(10^{13}

218: \,{\rm cm}/r)^{\beta}$ G. For $\beta=1$ and $b_0=30$ this expression corresponds 

219: to the lower limit required by \citet{wax95} in order to allow for efficient (second

220: order) Fermi-type acceleration of protons to ultra-high energies during expansion of 

221: the wind (cf. his eq. [4a]). While it can be shown then that proton acceleration is 

222: nearly unconstrained by radiative synchrotron losses, electron acceleration can only 

223: work efficiently if the field decays rapidly enough and becomes comparatively weak, 

224: e.g., for $\beta=2$ efficient electron acceleration may occur for $r \geq 10^{15} 

225: \xi^{-1/4}\,b_0^{3/4}(300/\gamma_{b0})^{1/2}$ cm.\\

226: The second constraint requires the particle mean free path $\lambda'$ to be 

227: smaller than the transversal width $R \sim r\,\phi_j$ of the jet (with $\phi_j > 

228: \phi_c$ being the jet opening angle in the cosmological rest frame). For $\lambda' = 

229: \xi\, r_g'$, $\xi < 1$ and $r_g'=\gamma' m\, c^2/(e\,B')$, this implies an upper 

230: limit for the maximum possible (comoving) particle Lorentz factor given by

231: \beq

232: \gamma_{\rm max}' \sim 10^{9} \left(\frac{1}{\xi}\right) \left(\frac{m_p}

233:             {m}\right) \left(\frac{B'}{10^3\;{\rm G}}\right) \left(\frac{r}

234:             {10^{13}{\rm cm}}\right)\left(\frac{\phi_j}{0.3\,{\rm rad}}\right)\,.

235: \eeq Note again, that for $B'=1000\,b_0\,(10^{13}\,{\rm cm}/r)^{\beta}$ G, this 

236: translates into maximum particle energies of $E \sim 2.8\cdot 10^{20}\,\xi^{-1}\,

237: b_0\,(10^{13}\,{\rm cm}/r)^{\beta-1}\,(\phi_j/0.3\,{\rm rad})\,(\gamma_{b}/300)$ 

238: eV in the observer frame. For a simple linear scaling, i.e., $\beta=1$, the maximum 

239: energy would become independent of $r$, while for $\beta>1$ the higher energy 

240: particles would originate on smaller scales.\\

241: The third constraint requires that the timescale for (cross-field) diffusion in the 

242: comoving frame $t_{\rm cf} \sim R^2/\kappa'$, with $\kappa' \sim c\,\lambda'/3$, 

243: is larger than $t_{\rm acc}$, i.e., $t_{\rm acc} \leq 3 \,(r^2\phi_j^2)

244: /(c\,\lambda')$. Comparison with eq.~(\ref{tshear1}) reveals that $t_{\rm acc}$ 

245: scales with $r$ and $\lambda'$ in the same way as $t_{\rm cf}$, indicating that 

246: the third constraint may be easily satisfied as long as $\gamma_b(\phi)$ is larger 

247: than a few, cf. Fig.~(\ref{fig1}).\\

248: Finally, the fourth constraint requires $t_{\rm acc}$ to be smaller than 

249: the time $t_w' =\Delta'/c$ needed to transverse the overall radial (comoving) 

250: width $\Delta'$ of the flow, which translates into a lower limit for the required 

251: particle Lorentz factors. This width is essentially determined by the activity 

252: of the inner engine driving the GRB and, when specified in the 

253: observer frame, should be at least of order of the observed GRB duration $t_{\Delta}$ 

254: \citep[cf.][]{pir05}, i.e., $\Delta = t_{\Delta}\,c=\Delta'/\gamma_b$. Allowing

255: for expansion effects on larger scales with $\Delta' \gppr r/\gamma_b$ \citep[cf.]

256: []{mes93}, we have $t_w' = {\rm max}\{\gamma_b\,t_{\Delta}, r/[\gamma_b\,c]\}$.

257: Using Eq.~(\ref{tmin}) one thus finds

258: \beq\label{constraint4}

259: \gamma_{\rm min}' \simeq 10^4 \,\xi^{-1} \left(\frac{m_p}

260:                   {m}\right)\left(\frac{B'}{10^3\;{\rm G}}\right)

261:                   \left(\frac{300}{\gamma_b}\right)^3\left(\frac{r}{10^{13}\,

262:                   {\rm cm}}\right)^2 \times {\rm min}\left\{\left( \frac{10\,{\rm sec}}

263:                   {t_{\Delta}}\right),\chi(r)\right\} \,,

264: \eeq where $\chi(r)= 2.7\cdot 10^3\,(\gamma_b/300)^4\,(10^{13}\,{\rm cm}/r)$. For the 

265: chosen magnetic field parametrization $B'=1000\,b_0\,(10^{13}\,{\rm cm}/r)^{\beta}$ G, 

266: this results in $\gamma_{\rm min}' \simeq 10^4\, \xi^{-1}\,(m_p/m)\,\,b_0

267: \,(r/10^{13}\,{\rm cm})^{2-\beta}$ $(300/\gamma_{b})^3 \times{\rm min}\{(10\,{\rm sec}/

268: t_{\Delta}), \chi(r)\}$, indicating that for $\beta=2$ the leading term becomes 

269: independent of $r$, which may be favourable for efficient electron acceleration. 

270: A more restrictive condition, roughly corresponding to $\gamma_{\rm min}'$

271: as implied by $\chi(r)$ in eq.~(\ref{constraint4}), is usually associated with 

272: the requirement to overcome adiabatic losses.

273: Accordingly, shear acceleration can only act efficiently on particles that were 

274: sufficiently accelerated through other processes. In the GRB context it is possible 

275: that such particles can be provided, for example, by the mechanism responsible for 

276: the prompt burst of emission (e.g., shock acceleration).

277: 

278: In summary, if the jet magnetic field is sufficiently weak (e.g., $b_0 \sim 1$) 

279: and/or decays rapidly enough (say, e.g., with $\beta = 2$), both the first and

280: the fourth constraint may be satisfied even for electrons, so that the maximum 

281: energy is essentially determined by the third constraint. Particles may then, for 

282: example, be accelerated efficiently at distances larger than $r \sim 10^{12}\,

283: (m_p/m)$ cm (cf. constraint [i]), with possible maximum energies, measured in the 

284: observer frame, of less than $\sim 10^{18}$ eV for electrons and $\sim 10^{21}$ eV 

285: for protons, and successively (i.e., linearly for $\beta=2$) decreasing with time. 

286: For a high (e.g., $b_0 \sim 10$) and slowly decaying (say, e.g., $\beta \leq 1$) 

287: jet magnetic field, on the other hand, efficient shear acceleration of electrons is 

288: virtually excluded, while protons again may well be accelerated up to energies 

289: $>10^{20}$ eV. 

290:  

291: 

292: \section{Comparison with shock acceleration} 

293: It is widely believed that shock-accelerated electrons are responsible for the 

294: observed prompt GRB and afterglow emission via synchrotron radiation processes 

295: \citep[e.g.,][]{pir05}. In particular, diffusive electron acceleration at midly 

296: relativistic internal shocks (with Lorentz factor $\Gamma_s \sim$ a few), caused 

297: by velocity variations in the relativistic outflow is usually thought to be 

298: behind the powerful burst of $\gamma$-rays \citep[][]{ree94}. In the case of

299: nonrelativistic shocks Fermi acceleration leads to power-law particle spectra 

300: $N(\gamma) \propto \gamma^{-s}$, which are only dependent on the shock compression 

301: ratio $\rho=u_1/u_2$ (where $1 < \rho \leq 4$), i.e., $s=(\rho+2)/(\rho-1)$, so 

302: that for strong shocks ($\rho=4$) the famous $s=2$ result is obtained \citep[e.g.,]

303: []{dru83,kir99}. In general, the acceleration timescale for diffusive shock 

304: acceleration depends on both the upstream and downstream residence times. For 

305: an unmodified nonrelativistic shock, one thus obtains \citep[e.g.,][]{dru83,jok87}

306: \beq 

307:   t_{\rm acc}=\frac{3}{u_1-u_2}\left(\frac{\kappa_1}{u_1}+\frac{\kappa_2}{u_2}\right)

308:              =\frac{3\,\rho}{(\rho-1)}\frac{\left(\kappa_1+\rho\,\kappa_2\right)}

309:                 {u_s^2}\,,

310: \eeq with $u_1$ and $u_2$ the upstream and downstream flow velocities measured in 

311: the shock frame, respectively, $\kappa_1$ and $\kappa_2$ the upstream and downstream 

312: (spatial) diffusion coefficients, and $u_s$ the shock speed as measured in the upstream 

313: frame. If the acceleration process operates at nearly the Bohm limit [i.e., $\kappa 

314: \simeq \lambda\,c/3 = \eta\, r_g\,c/3$, with $\eta \simeq O(1)$], one finds

315: \beq\label{tacc2}

316:  t_{\rm acc} \gppr 6\, \eta\,r_g\,\frac{c}{u_s^2}\,,

317: \eeq assuming $\kappa_1 \simeq \kappa_2$. When $u_s\sim c$ this is comparable to 

318: the Larmor time which is a result also predicted by the simulations of \citet{lem03}.

319: It has been often suggested that the nonrelativistic limit can also be approximately

320: applied to midly relativistic shocks in GRBs, proposing, for example, that efficient 

321: proton acceleration to ultra-high energies might be possible \citep[e.g.,][]{wax04}. 

322: Equating the acceleration timescale (eq.~[\ref{tacc2}]), replacing $r_g$ by $r_g'$ and 

323: $u_s$ by $u_s'$ as the (comoving) internal shock speed, with the cooling timescale 

324: eq.~(\ref{tcool}), using $B'=1000\,b_0\,(10^{13}\,{\rm cm}/r)^{\beta}$ G and $u_s'

325: =\beta_s\,c$ with $\beta_s < 1$, gives

326: \beq\label{c_shock1}

327:  \gamma_{\rm max}' \simeq 2.5\cdot 10^9\,\eta^{-1/2}\,b_0^{-1/2}

328:                     \left(\frac{\beta_s}{0.9}\right)\,\left(\frac{m}{m_p}\right)\,

329:                     \left(\frac{r}{10^{13}\,{\rm cm}}\right)^{\beta/2}\,.

330: \eeq Note that magnetic field amplification in the vicinity of a shock may yield a 

331: field value $b_0$ well above the usual (background) flow magnetic field, thus leading 

332: to a somewhat more restrictive condition. In addition, the acceleration timescale 

333: also has to be smaller than the comoving time $\sim r/{c\,\gamma_b}$ needed to 

334: transverse the width of a shell, which gives

335: \beq\label{c_shock2}

336: \gamma_{\rm max}' \lppr 1.5 \cdot 10^6\,\eta^{-1}\,b_0\, \left(\frac{\beta_s}{0.9}

337:                   \right)^2\,\left(\frac{m_p}{m}\right)\,\left(\frac{300}{\gamma_b}

338:                   \right)\,\left(\frac{10^{13}\,{\rm cm}}{r}\right)^{\beta-1}\,.

339: \eeq 

340: A more detailed comparison of shear with shock acceleration may perhaps be reached 

341: by analyzing a critical particle Lorentz factor $\gamma_{\rm eq}'\propto b_0\,

342: r^{1-\beta}\,(m_p/m)$, defined by equating eq.~(\ref{tmin}) with eq.~(\ref{tacc2}), 

343: above which shear acceleration may become more efficient than shock acceleration.

344: Several applications with respect to the acceleration of protons are illustrated in 

345: Fig.~(\ref{fig2}), suggesting the possibility that under a reasonable range of 

346: conditions shear acceleration in GRB jets may become more relevant for the production 

347: of UHE cosmic rays than shock-type acceleration processes. Concerning electron 

348: acceleration in weak magnetic fields (e.g., $b_0 \leq 1$, $\beta=2$, $r \gppr 

349: 10^{15}$ cm), on the other hand, shock-type processes would allow for a much 

350: quicker particle energization (but would also be more severely limited by 

351: eq.~[\ref{c_shock1}] and eq.~[\ref{c_shock2}]) and shear effects would only 

352: become dominant above $\gamma' \sim 10^8\,(10^{15}\,{\rm cm}/r)$.

353: 

354: \section{Conclusions}

355: Using an idealized model we have analyzed the possible role of shear acceleration

356: for the energization of particles in relativistic GRB-type jets. Our results suggest

357: that efficient electron acceleration on scales $r\gppr 10^{15}$ cm, with maximum 

358: energy decreasing with distance, may be possible in the presence of weak magnetic 

359: fields, assuming that high energy seed particles are provided by the mechanism 

360: responsible for the prompt burst of emission. This may result in a weak and long 

361: duration component in the GRB emission. Protons, on the other hand, may reach UHE 

362: energies $>10^{20}$ eV under a broad range of conditions.

363:  

364: 

365: \acknowledgments

366: We are grateful to R.~Blandford for discussions and the referee for very useful 

367: comments. FMR acknowledges support through a Marie-Curie Individual and a 

368: Cosmogrid Fellowship. 

369: 

370: 

371: 

372: \begin{thebibliography}{}

373: \bibitem[Berezhko \& Krymskii(1981)]{berk81}

374:  Berezhko, E. G., \& Krymskii, G. F. 1981, Sov. Astr. Lett., 7, 352

375: \bibitem[Drury(1983)]{dru83}

376:  Drury, L.O'C 1983, Rep. Prog. Phys., 46, 973

377: \bibitem[Earl et al.(1988)]{ear88} 

378:  Earl, J.~A., Jokipii, J.~R., \& Morfill, G.\ 1988, \apjl, 331, L91 

379:  \bibitem[Frail et al.(2001)]{fra01} 

380:  Frail, D.~A., et al.\ 2001, \apjl, 562, L55 

381: \bibitem[Greiner et al.(2003)]{gre03}

382:  Greiner, J., Klose, S., Reinsch, K. et al.~2003, Nature 426, 157  

383: \bibitem[Jokipii(1987)]{jok87}

384:  Jokipii, J.R. 1987, ApJ, 313, 842

385: \bibitem[Kirk \& Duffy(1999)]{kir99} 

386:  Kirk, J.~G.~\& Duffy, P.\ 1999, JPhG, 25, 163 

387: \bibitem[Kulkarni et al.(1999)]{kul99} 

388:  Kulkarni, S.~R., et al.\ 1999, \nat, 398, 389 

389: \bibitem[Kumar \& Granot(2003)]{kum03} 

390:  Kumar, P.~\& Granot, J.\ 2003, \apj, 591, 1075 

391: \bibitem[Lamb et al(2004)]{lam04} 

392:  Lamb, D.~Q., Donaghy, T.~Q., \& Graziani, C.\ 2004, New Astronomy Review, 48, 459 

393: \bibitem[Lemoine \& Pelletier(2003)]{lem03}

394:  Lemoine, M., \& Pelletier, G. 2003, ApJ, 589, L73

395: \bibitem[ M{\' e}sz{\' a}ros \& Rees(1993)]{mes93} 

396:   M{\' e}sz{\' a}ros, P., \& Rees, M.~J.\ 1993, \apj, 405, 278 

397: \bibitem[Ostrowski(1998)]{ost98}

398:  Ostrowski, M. 1998, \aap, 335, 134

399: \bibitem[Ostrowski(2000)]{ost00}

400:  Ostrowski, M. 2000, \mnras, 312, 579

401: \bibitem[Piran(2005)]{pir05}

402:  Piran, T. 2005, Rev. of Mod. Phys., 76, 1143 

403: \bibitem[Rees \& M{\' e}sz{\' a}ros(1994)]{ree94} 

404:  Rees, M.~J.~\& M{\' e}sz{\' a}ros, P.\ 1994, \apjl, 430, L93 

405: \bibitem[Rhoads(1999)]{rho99} 

406:  Rhoads, J.~E.\ 1999, \apj, 525, 737 

407: \bibitem[Rieger \& Duffy(2004)]{rie04}

408:  Rieger, F.M., \& Duffy, P. 2004, \apj, 617, 155

409: \bibitem[Rieger \& Duffy(2005a)]{rie05a}

410:  Rieger, F.M., \& Duffy, P. 2005a, ChJA\&A 5S, 195

411: \bibitem[Rieger \& Duffy(2005b)]{rie05b}

412:  Rieger, F.M., \& Duffy, P. 2005b, Proc. of 22nd Texas Symposium 

413:  on Relativistic Astrophysics (Stanford 2004), eds. P. Chen et al.,

414:  eConf:C041213, 2521

415: \bibitem[Rieger \& Duffy(2005c)]{rie05c}

416:  Rieger, F.M., \& Duffy, P. 2005c, ApJ, to be submitted

417: \bibitem[Rossi et al.(2002)]{ros02} 

418:  Rossi, E., Lazzati, D., \& Rees, M.~J.\ 2002, \mnras, 332, 945 

419: \bibitem[Waxman(1995)]{wax95}

420:  Waxman, E. 1995, Phys. Rev. Lett., 75, 386

421: \bibitem[Waxman(2004)]{wax04}

422:  Waxman, E. 2004, ApJ, 606, 988 

423: \bibitem[Webb(1989)]{web89}

424:  Webb, G. M. 1989, \apj, 340, 1112

425: \bibitem[Zhang \& M{\' e}sz{\' a}ros(2002)]{zha02} 

426:  Zhang, B.~\& M{\' e}sz{\' a}ros, P.\ 2002, \apj, 571, 876 

427: \bibitem[Zhang et al.(2004)]{zha04} 

428:  Zhang, B., Dai, X., Lloyd-Ronning, N.~M., \& M{\' e}sz{\' a}ros, P.\ 2004, 

429:  \apjl, 601, L119 

430: \end{thebibliography}

431: 

432: \clearpage

433: 

434: 

435: \begin{figure}

436: \plottwo{f1a.eps}{f1b.eps}

437: \caption{Left: Illustration of the evolution of the bulk Lorentz factor $\gamma_b$

438:  with angle $\phi$ for a power-law ($b=1.8$) and Gaussian model, respectively, 

439:  using $\gamma_{b0}=300$. Right: Associated timescales for shear acceleration,

440:  calculated for $\phi_c=0.1$ [rad].\label{fig1}}

441: \end{figure}

442: 

443: \clearpage

444: 

445: \begin{figure}

446: \epsscale{.60}

447: \plotone{f2.eps}

448: \caption{Evolution of the critical comoving proton Lorentz factor $\gamma_{\rm 

449: eq}'$ as a function of the radial coordinate, shown for $\gamma_b =300$, $\xi\,\eta 

450: =1$, $\beta_s =0.9$, and different jet magnetic field configurations. Above 

451: $\gamma_{\rm eq}'$ the shear acceleration timescale becomes smaller than the shock 

452: acceleration timescale and proton acceleration by shear becomes more efficient 

453: than shock acceleration.\label{fig2}}

454: \end{figure}

455: 

456: 

457: \end{document}

458: 

459: 

460: 

461: