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

2: %\documentclass[manuscript]{aastex}

3: %\documentclass{aastex}

4: %\usepackage{emulateapj5}

5: %\documentclass[preprint2]{aastex}

6: %\documentstyle[12pt,aps,epsfig]

7: 

8: %definitions go here

9: \newcommand\degree{$^{\circ}$}

10: \newcommand\cR{{\cal R}}

11: \newcommand\km{{\rm \ km}}

12: \newcommand\simlt{\lower.5ex\hbox{$\; \buildrel < \over \sim \;$}}

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

14: 

15: 

16: 

17: %\slugcomment{Submitted to {\it Astrophysical Journal Letters},DATE}

18: \begin{document}

19: 

20: 

21: \title{Delayed Afterglow Onset Interpreted as Baryon-Poor Viewing Angle}

22: \author {David Eichler\altaffilmark{1}}

23: %\altaffiltext{1}{School of Physics \& Astronomy, Tel Aviv University,

24: %Tel Aviv 69978, Israel; Levinson@wise.tau.ac.il}

25: \altaffiltext{1}{Physics Department, Ben-Gurion University,

26: Beer-Sheva 84105, Israel; eichler@bgu.ac.il}

27: %\author{??}

28: 

29: \begin{abstract}

30:     We have suggested previously that  baryons in GRB fireballs

31:  infiltrate from the surrounding walls that

32: collimate the fireball.  The efficiency $\epsilon_b$ for

33: generating blast energy

34:  can then be angle dependent.  Delayed onset of

35: afterglow can be interpreted as being due to a baryon-poor viewing

36: angle.

37: \end{abstract}

38: 

39: 

40: \keywords{black hole physics --- gamma-rays: bursts and theory }

41: 

42: %\mbox{}\\

43: %\newpage

44: \section{Introduction}

45: 

46: GRB are suspected to come from black holes.  Nearly all GRB

47: produce X-ray afterglow. This is significant because, if GRB were

48: driven by  the pure energy that can be extracted from a black

49: hole, it is unlikely that there would be afterglow. Pairs could be

50: produced in a variety of ways, but they would probably annihilate

51: within $10^{11}$ cm of the central engine before making afterglow.

52: This is not an embarrassment either for black hole models of GRB

53: or for models of afterglow, because baryons can be picked up after

54: the pair fireball is launched - either from a surrounding baryonic

55: wind that emanates

56:  from the accretion disk, from

57: the walls of a host star, or from ambient material, such as a

58: presupernova wind emitted by the  star that hosts the GRB.

59: Moreover, the possibility exists that the GRB is  driven by the

60: accretion disk surrounding the black hole,

61:  and that

62: the latter plays no active role other than to provide

63: gravitational energy to  the accreting matter. Nevertheless,

64:  the  event horizon of the underlying black hole would be more convincingly

65: revealed

66:   by {\it failure} to produce afterglow,

67: especially now that afterglow is regarded as routine. A line of

68: sight in which $\gamma$-rays, but not  baryons, were emitted could

69: be the signature of an event horizon, which permits energy, but

70: not baryons,

71:  to escape along the field lines that thread it.  The recent giant

72:  flare of 27 December 2004 from SGR 1806-20  displayed both strong

73:  mass ejection and bright gamma-ray emission, implying that the

74:  latter can take place in the presence of the former without an

75:  event horizon. This is probably to large variations in the baryon

76:  loading that naturally result from magnetic reconnection in a highly

77:  stratified medium. However, the duration of the mass ejection during this event

78:  was smaller than the rotation period of the central object, which is unlikely

79:  to be the case for GRB.

80: 

81: 

82: Very recently,  Piro et al (2004)  have tracked what appears to be

83: the beginning of the afterglow phase from two relatively bright

84: GRB using  data from the Wide Field Camera (WFC) of BeppoSax. The

85: WFC was able to track  the  GRB's 011121 and 011211, which lasted

86: tens of seconds.  They then show a lull for about 200 seconds

87: (after correcting for redshift) and, after this lull, there is a

88: soft but non-thermal "revival" detected in the X-ray band.

89: Previous timings of afterglow onset (e.g. Pian et al 2001) are

90: consistent with delays of order 70 seconds, but are hard to

91: separate from the end of the prompt phase of the GRB.  The revival

92: decreases as a power law and fits the backwards extrapolation of

93: the afterglow observed several hours later with the Narrow field

94: Instrument (NFI). Although such lulls are not seen for all GRB,

95: they are not atypical and many are shorter, lasting perhaps 70

96: seconds on average. For GRB that are not so bright, a revival

97: following a long lull would not have been picked up by the WFC, so

98: short lulls and/or  revival-free prompt emission may be positively

99: selected in all but the brightest GRB's.

100: 

101: Noting that the revived X-ray emission appears to be the beginning

102: of the X-ray afterglow,   Piro et al. suggest that GRB's 011121

103: and 011211`were particularly  prolonged  - enough that the

104: afterglow it produces should peak late, i.e. that the reverse

105: shock and onset of self-similar expansion should begin after

106: several minutes. They tentatively attribute  lulls occurring from

107: $\sim 70 - 200$s to a decrease in the efficiency of converting

108: outflow energy to gamma rays, e.g. by internal shocks. In this

109: letter I suggest an alternative explanation - that the late onset

110: of the X-ray afterglow is due to a viewing angle effect. I suggest

111: that our line of sight is not in the primary direction of the

112: blast of kinetic energy and that the beaming angle of the

113: afterglow has become wide enough to include our line of sight only

114: after $\sim 200$s. This was predicted in Eichler and Levinson

115: (2004), who noted that the Amati et al relation (Amati et al,

116: 2002, Atteia et al, 2004) between spectral peak frequency

117: $\nu_{peak}$ can be attributed to off-beam viewing.

118: 

119: The present suggestion is at odds with the model of GRB in which

120: $\gamma$-rays are produced by internal shocks in a baryonic

121: outflow downstream of the photosphere (Mezsaros and Rees, 1994).

122: In this popular model, the presence of $\gamma$-rays implies the

123: existence of a somewhat larger energy reservoir of baryonic

124: kinetic energy, which then produces the afterglow along any line

125: of sight on which there were observed $\gamma$-rays. (The reverse

126: need not be true if there happen to be no internal shocks.) The

127: present suggestion, on the other hand, allows for the possibility

128: that the beam of $\gamma$-ray emission need not lie entirely

129: within  the beam of afterglow-producing baryons.

130: 

131: It has already been suggested (Eichler and Levinson, 2004 ) that

132: X-ray flashes and their adherence to the Amati et al relation

133: between spectral peak $E_{peak}$ and isotropic equivalent fluence

134: $E_{iso}$ can be attributed to an observer's view that is not in

135: the direct line of fire of the $\gamma$-ray fireball. It was also

136: noted (Eichler and Levinson, 2004) that such X-ray flashes may not

137: always be accompanied by early afterglow, and that a late

138: afterglow onset would confirm the whole picture of delayed onset

139:   due to  viewing angle effects (Granot et al., 2002, 2005)

140: as applied to these X-ray flashes. [Off angle viewing as an

141: explanation for X-ray bursts has also been discussed by Yamazaki

142: et al. (2002, 2004), who derive a different viewing angle-induced

143: relation because of different assumptions. They attribute the

144: Amati et al. relation at higher energy to some other unexplained

145: effect.]

146: 

147: Below, we use parameters guided by the observed Amati et al

148: relation, and show that even within 200 seconds of observer time,

149: the blast can decelerate enough that  the afterglow beam

150: encompasses baryon-poor viewing angles.

151: 

152: 

153: \section{The Basic Picture}

154: 

155: Consider the following topology for a GRB fireball: The fireball

156: is  emitted or collimated into a cone of opening angle $\theta_o$.

157: The baryons infiltrate into the fireball from the side and

158: penetrate an angle $\Delta$ so that they  flow out along the

159: annulus $\theta_o - \Delta \le \theta \le \theta_o$.  Assume that

160: the $\gamma$-rays within this annulus are isotropic at their point

161: of emission or last scattering, in a frame that moves at Lorentz

162: factor $\Gamma_e$ relative to the observer. Thus the $\gamma$-rays

163: ultimately fill a cone of opening angle $\theta_o + 1/\Gamma_e$.

164: At first, only observers in the line of sight  $\theta_o - \Delta

165: -1/\Gamma_e \le \theta \le \theta_o + 1/\Gamma_e$ see early

166: afterglow. The Lorentz factor of the blast decreases as it sweeps

167: up ambient matter, so that at observer time t, afterglow is seen

168: in an enhanced annulus $\theta_o - \Delta-1/\Gamma(t) \le \theta

169: \le \theta_o + 1/\Gamma(t)$

170: %Assume the blast sweeps up a wind that has an $r^{-2}$ density

171: %dependence, so that $\Gamma(t) \propto t^{-1/4}$, or $\Gamma(t) =

172: %(t/t_5)^{-1/4}\Gamma(t_5)$.

173: (Here and throughout  $Q_n \equiv Q/10^n$ in cgs units.)

174: 

175:  Given that softened GRB spectra

176: are attributed to off-beam viewing angles, the Amati et al

177: relation can be best understood  if a) $\theta_o \gg 1/\Gamma_e$

178: and b) $\Delta \ge 3/\Gamma_e$ (Levinson and Eichler 2004). These

179: requirements are based on the fact  that the off-beam viewing

180: angle that is offset by $\delta$ from the $\gamma$-ray beam

181: receives contributions at comparable Doppler factors from a patch

182: that is $\sim \delta$ in width and in length, and is thus

183: proportional to $\delta^2$. This somewhat compensates the large

184: decrease in $E_{iso}$ due to the decrease in the Doppler factor

185: $1/\Gamma_s(1-\beta_e cos\delta)$.

186: 

187: Now divide the set of possible observer viewing angles into the

188: following zones:

189: 

190: The inner zone is $\theta \le \theta_o -\Delta $

191: 

192: The baryon-rich zone is $\theta_o -\Delta \le \theta \le \theta_o$

193: 

194: The outer zone is $\theta \ge \theta_o $

195: 

196: Baryons are directed only at  observers in the baryon rich zone.

197: Hence, an observer sees afterglow only if he is within

198: $1/\Gamma(t)$ of this zone. Observers in the outer zone see X-ray

199: flashes peaking at $\nu$ if they are at a viewing angle $\delta

200: \theta \sim (\nu^*/\nu)^{1/2}/\Gamma_e$ from the baryon rich zone.

201: Here $\nu^*$ is the spectral peak seen by a head-on observer,

202: apparently about 1 to 2 Mev in the assumptions of the off-beam

203: viewing angle of the Amati et al relation.  Observers  in the

204: inner zone would see a $\gamma$-ray burst with  weak or delayed

205: afterglow, or possibly no afterglow at all. Observers  within

206: $\delta  \sim (\nu^*/\nu)^{1/2}/\Gamma_e$ of the baryon rich zone

207: see an X-ray rich GRB, where the soft contribution comes from the

208: baryon rich zone viewed obliquely.

209: 

210: Note that the assumption that $\gamma$-rays fill the inner zone is

211: not quite the same  as  in (Eichler and Levinson 2004), where it

212: was assumed that {\it all} the $\gamma$- ray emission was from the

213: baryon rich zone. This zone, slightly expanded by $1/\Gamma_e$,

214: may include the collection  all of the photons that hit the

215: baryonic lining of the corridor that the GRB fireball bores

216: through the host star, and then get scattered forward into the

217: annulus defined, more or less, by the baryonic outflow. Because

218: the original solid angle  of this luminosity component can be

219: rather large, the luminosity from the baryonic annulus may be

220: considerably larger than that coming directly from the inner zone.

221: An observer in the inner zone might see both direct photons and

222: off-beam photons last scattered in the baryon-rich zone, and the

223: relative strength of each would be strongly viewing-angle

224: dependent.

225: 

226: In   principle an observer near the axis could see a GRB with no

227: afterglow. How often would this occur? Observationally, it must be

228: a small fraction of the total, as all but one GRB localized with

229: BeppoSax displayed X-ray afterglow observed by the NFI, which

230: makes its observations $10^4t_4$ seconds after the burst trigger,

231: $t_4 \gtrsim 1$. The expected fraction of GRB that would display

232: no afterglow by time t is then

233: 

234: \begin{equation}

235: f(t) \simeq  [\theta_o -\Delta-1/\Gamma(t)]^2 / [\theta_o

236: +1/\Gamma_e]^2.

237: \end{equation}

238: 

239: As an example, we consider the parameter choice $\theta_o =

240: 9/\Gamma_e$, $\Delta = 4/\Gamma_e $, and  $\Gamma_e=\Gamma(100s) =

241: 100$. [The choice $3/\Gamma_e \la \Delta  \la 5/\Gamma_e $ has

242: both empirical motivation (Eichler and Levinson, 2004) as well as

243: {\it a priori} motivation in the model of Levinson and Eichler

244: (2003), where it is estimated that neutrons freely stream at a

245: surface with $\Gamma \sim 30$, which is somewhat greater than

246: $1/\theta_o \sim 10$ for typical GRB but by less than an order of

247: magnitude.] The fraction of observers seeing hard $\gamma$-rays

248: along baryon poor lines of sight  is  $\sim 0.25$, a not

249: insignificant fraction. If the blast decelerates as $\Gamma

250: \propto t^{-3/8}$, the case of a constant density ambient medium,

251: then within 3 hours ($10^4$s, say), $\Gamma = 10^{5/4}$, and even

252: an observer exactly at the axis would detect afterglow. If $\Gamma

253: \propto t^{-1/4}$, the case for a wind, then after $10^4$s

254: ($10^{4.5}$s), $\Gamma(t)= 10^{3/2}$ ($10^{11/8}$), and a small

255: fraction $f= 0.04$ ($\le 0.0001$) fail to see afterglow. This

256: illustrates the point that {\it afterglow may be nearly guaranteed

257: for an observer of a GRB after several hours, even if the original

258: line of sight is baryon-poor}. At a time t of only 200s, on the

259: other hand, still assuming the above GRB parameters and that

260: $\Gamma(t) \propto t^{1/4}$, it follows that $\Gamma(200s) \sim

261: 80$, and the probability for a viewer of the prompt emission of a

262: GRB ( defined here to be within 0.1 [radians] of the axis) to not

263: observe its afterglow by this time (i.e. to be within 0.04 of the

264: axis), would be about 0.16.

265: 

266:  It is thus reasonable that some modest fraction of all X-ray afterglows

267: begin, from the observer's point of view, at $t \ge 200$s. It goes

268: without saying that the numbers here are somewhat uncertain, and

269: the sharp zone boundaries invoked  here are an oversimplification.

270: Where we have formulated the results in terms of  afterglows being

271: seen or not seen, with no middle ground, it might actually be the

272: case that they would be bright or dim relative to the prompt

273: emission. Observations with Swift will allow better measurements

274: of afterglow onset, but the full BeppoSax data set already allows

275: modest statistical analysis.

276: 

277: I thank Dr. E. Pian   for useful discussions. This research was

278: supported by the Arnow Chair of Astrophysics at Ben Gurion

279: University, by a Center of Excellence grant awarded by the Israel

280: Science Foundation, and by a grant from the Israel-U.S. Binational

281: Science Foundation.

282: 

283: \begin{thebibliography}{99}

284: \bibitem [] {1} Amati, L. et al., 2002, Astron. and Astroph. 390,

285: 81

286: \bibitem [] {21} J-L. Atteia, G.R. Ricker, D.Q. Lamb, T. Sakamoto, C.

287: Graziani, T. Donaghy, C. Barraud, and the HETE-2 Science Team,

288: astro-ph 0312371

289: \bibitem[]{} Eichler, D. \& Levinson, A. 1999, ApJ, 521, L117

290: \bibitem[]{} Eichler, D. \& Levinson, A. 2004, ApJ, 614, L13

291: \bibitem[]{} Frontera, F. et al. 2000, ApJS, 127, 59

292: %\bibitem[]{}  Frail, D., Kulkarni, S. \& Bloom, J.S. 1999, Nature, 398, 127

293: %\bibitem[]{}  Goodman, J. 1986, ApJ, 308, L47

294: \bibitem[]{} Granot, J., Panaitescu, A., Kumar, P. and Woosley, S.

295: (2202), ApJ, 570, L61

296: \bibitem[]{} Granot, J., Ramirez-Ruiz, E., Perna R., 2005

297: astro-ph/0502300 v2

298: \bibitem[]{} Meszaros, P.  \& Rees, M. J. 1994, ApJ Lett., 430,

299: L93

300: \bibitem[]{} Pian, e. Soffitta, P, Alessi, A. Amati, L., Costa, E.

301: Frontera, F. Fruchter, A., Massetti, N., Palazzi, E., Panataiscu,

302: A. and Kumar, P.  (2001) Astron. and Ap., 372, 456

303: \bibitem[]{} Piro. L. De Pasquale, M. Soffitta, Lazzati, D.

304: Amati, L., Costa, E., Feroci, M., Frontera, F., Fuidorzi, C.,

305: Zand,, J. M. J., Montanari, E., and Nicastro, L., (2004)

306: astro-ph/0412589

307: 

308: \bibitem[]{} Yamazaki, R., Ioka, K., \& Nakamura, T. 2004, ApJ, 606, L33

309: \bibitem[]{} Yamazaki, R., Ioka, K., \& Nakamura, T. 2002, ApJ, 571, L31

310: %\bibitem[]{} Hurley, K. et al 2002, ApJ, 567, 447

311: %\bibitem[]{} Levinson, A. and Eichler, D. 1993, ApJ, 418, 386

312: %\bibitem[]{} Levinson, A. and Eichler, D. 2003, ApJ, 594, L19

313: %\bibitem[]{} Vlahakis, N., Peng, F., Konigl, A., 2003, ApJ, 594,  L23

314: %\bibitem[]{} Rossi, E. Beloborodov, A. and Rees, M.J. astro-ph/0401355

315: %\bibitem[]{}  Paczynski, B. 1986, ApJ, 308, L43

316: %\bibitem[]{}  Paczynski, B. and Xu, G. 1994, Ap. J. 427, 708

317: %\bibitem[]{}  Tasevsky, M. (2001) hep-exp 0110084 and references therein

318: %\bibitem[]{} Tan, J., Matzner, C.D. \& McKee, C.F. 2001, ApJ, 551, 946

319: %   \bibitem[]{} Sikora, M. and Wilson, D.B. 1981, MNRAS, 197,529S%

320: %\bibitem[]{}  van Putten, M.V.P. \& Levinson, A. 2003, ApJ, 584, 937

321: %\bibitem[]{}  Waxman, E. \& Meszaros, P. 2003, ApJ, 584, 390

322: %\bibitem[]{}  Waxman, E. 2001, Nucl. Phys. B (Proc. Suppl.)91,494

323: %\bibitem[]{}  Waxman, E. and Bahcall. J. 1997, Phys. Rev. Lett.,

324: %78, 2292

325: \end{thebibliography}

326: %\newpage

327: %\section{Figure captions}

328: 

329: \newpage

330: \plotone{delayfig2.eps} \figcaption{ The three zones described in

331: the text are separated by the black lines.  The purple-shaded

332: region denotes the set of viewing angles that observe X-ray

333: afterglow after a given early time. The shaded blue area denotes

334: the viewing angles that see afterglow at a somewhat later time.

335: The white dot in the inner zone denotes viewing angles in which

336: prompt emission might be seen with no X-ray  afterglow. }

337: \end{document}

338: