1:
2: \documentclass[12pt,preprint]{aastex}
3: %\documentclass[12pt,preprint]{aastex}
4: \usepackage{epsfig}
5: % for a referee version
6: %\usepackage{epsfig,deluxe}
7: % \documentclass[manuscript]{aastex}
8:
9: % %documentclass[article]{aa} \usepackage{epsfig} %
10:
11: \begin{document}
12:
13: \newcommand{\gsim}{\hbox{\rlap{$^>$}$_\sim$}}
14: \voffset-.6cm
15:
16: \title{Fireballs and cannonballs confront the afterglow of GRB
17: 991208}
18:
19: \author{Shlomo Dado\altaffilmark{1}, Arnon Dar\altaffilmark{1,2}
20: and A. De R\'ujula\altaffilmark{2}}
21:
22: \altaffiltext{1}{Physics Department and Space Research Institute,
23: Technion, Haifa 32000, Israel} \altaffiltext{2}{Theory Division,
24: CERN,CH-1211 Geneva 23, Switzerland}
25:
26:
27: % \maketitle
28:
29:
30: \begin{abstract}
31:
32: Galama et al. have recently reported their follow-up measurements
33: of the radio afterglow (AG) of the Gamma Ray Burst (GRB) 991208,
34: up to 293 days after burst, and their reanalysis of the broad-band
35: AG, in the framework of standard fireball models. They advocate a
36: serious revision of their prior analysis and conclusions, based on
37: optical data and on their earlier observations during the first
38: two weeks of the AG. We comment on their work and fill a lacuna:
39: these authors have overlooked the possibility of comparing their
40: new data to the available predictions of the cannonball (CB) model,
41: based ---like their incorrect predictions--- on the first round of
42: data. The new data are in good agreement with these CB-model
43: predictions. This is in spite of the fact that, in comparison to
44: the fireball models, the CB model is much simpler, much more
45: predictive, has many fewer parameters, practically no free choices...
46: and it describes well ---on a universal basis--- all the
47: measured AGs of GRBs of known redshift.
48:
49: \end{abstract}
50:
51:
52: \keywords{gamma rays: bursts}
53:
54: \section*{Introduction}
55:
56: We discuss the afterglow of GRB 991208 as a good and simple example
57: of a three-sided confrontation: the data, the generally accepted
58: theory (the fireball model in its various incarnations, thereafter
59: the ``standard model'': SM), and the cannonball (CB) model.
60:
61: Gamma Ray Burst (GRB) 991208 was detected with the Interplanetary
62: Network (IPN) on December 8.192 UT, 1999 and its afterglow (AG)
63: was first detected on December 10.92 UT in the radio band (Hurley et
64: al. 2000). The optical AG of GRB 991208 was first detected 2.1
65: days after burst and has been measured by Castro-Tirado et al.
66: (2001) and Sagar et al. (2000). ``Early'' radio measurements, from
67: 11.77 UT to 21.96 UT, December 1999 between 1.4 and 250 GHz were
68: reported by Galama et al. (2000) (thereafter G1) and were analyzed
69: in the framework of the SM along with the optical data.
70: Follow up observations, up to day $\rm 293$ after burst have been
71: recently reported and reanalyzed by Galama et al. (2002), (thereafter
72: G2), requiring a severe revision of their prior analysis and
73: conclusions.
74:
75:
76:
77: We have analized the AG of GRB 991208 in the realm of the CB model
78: twice before: in Dado et al. (2002a), thereafter DDD1, we fitted
79: the available R-band data on all GRBs of known redshift, $\rm z$,
80: including this one. In Dado et al. (2002b), hereinafter DDD2, we
81: extended the analysis to wide-band fits of all the available optical
82: and radio data. We present here a comparison between these CB-model
83: predictions and the new data of G2. We also refit all the data
84: including the new observations in G2. Consistently, the fit
85: parameters change
86: little as the new data are added. More importantly, the CB model,
87: in spite of its simplicity and its scarcity of parameters and
88: choices, is found to be successful: the predictions of DDD2 agree
89: with the new data of G2, which is not the case for the SM
90: predictions in G1.
91:
92:
93: \section*{GRB 991208 secundum Galama et al. (2000) [G1]}
94:
95: In G1 the authors first make a fit to the radio-to-optical spectral
96: behaviour of data modified to pertain to a fixed ``unified'' date:
97: December 1999, 15.5 UT. The ``spectral parameters'' are the
98: self-absorption and peak frequencies, $\rm \nu_a$ and $\rm \nu_m$,
99: the peak flux density, $\rm F_m$, and the power-law index of the
100: electron distribution, $\rm p$. With the resulting $\rm p=2.52$
101: fixed, they fit the radio-only spectrum at 3 other unified
102: dates within the first 2 weeks, by extracting values of $\rm
103: \nu_a$, $\rm \nu_m$ and $\rm F_m$ separately for each of these
104: dates. This procedure is useful as a test of the time evolution
105: of these quantities but, {\it in totto}, theirs is a 13-parameter fit.
106:
107: On the basis of their measured parameter-evolution, the authors of
108: G1 discard spherical-explosion models, either with a constant-density
109: interstellar medium (ISM), or a $\rm 1/r^2$ ``wind'' circumburst
110: profile (the ``ISM'' and ``WIND'' models). They also note that a
111: ``JET'' model disagrees by $>\! 4\sigma$ with the predicted power
112: decline of $\rm F_m$. They advocate a model with a transition from
113: a quasi-spherical to a jet evolution (Kumar and Panaitescu, 2000),
114: but they stop short of an explicit analysis on these grounds. Yet,
115: they conclude, with no proof, that ``the jet model can account for
116: the [observations...] provided that the jet transition has not been
117: fully completed in the first two weeks after the event''.
118:
119: The closing predictions in G1 are that $\rm \nu_a\propto t^{-14/13}$
120: at $\rm t\!>\! 10$ days, and the flux density $\rm F_\nu\propto
121: t^{-(2.2\; to \; 2.5)}$ for $\nu=8.46$ GHz at $\rm t\!>\! 12$ days,
122: as well as for $\nu=4.86$ GHz at $\rm t\!>\! 17$ days.
123:
124: \section*{GRB 991208 secundum Galama et al. (2002) [G2]}
125:
126: Two of the results of the analysis in G2 are that $\rm \nu_a\propto
127: t^{-0.29{-0.21\atop +0.17}}$ for a ``FREE fit'' to the ensemble of
128: data, and $\rm F_\nu\propto t^{-1.07\pm 0.09}$ at $\nu=8.46$ GHz,
129: for $\rm t=53$ to 293 days after burst. These results are in stark
130: contrast with the quoted predictions of G1.
131:
132: In the FREE model of G2 one extra cooling frequency, $\rm \nu_c$,
133: is introduced, and $\rm F_m$ as well as $\rm\nu_{a}$, $\rm \nu_m$
134: and $\rm \nu_c$ are assumed to behave as $\rm C_i\, t^{-\alpha_i}$,
135: for a total of nine parameters (including $\rm p$). Three SM
136: variations are also discussed: ISM, WIND and JET, each of which
137: predicts the values of $\rm \alpha_{_{F}}$, $\rm\nu_a$, $\rm \nu_m$
138: and $\rm\nu_c$. The ISM and WIND models turn out to be inadequate,
139: as in G1. The JET model (with its two new parameters) is an
140: improvement over that of G1, but fails to reconcile the late-time
141: decay $\rm F_\nu\sim t^{-1.1}$ at 8.46 GHz with the much steeper
142: optical decay $\rm F_R\propto t^{-2.2\pm 0.1}$ ---observed by Sagar
143: et al. (2000) at $\rm t=2$ to 10 days after burst--- which should
144: similarly decline. The FREE model provides a satisfactory fit to
145: the data, but implies that the combination $\rm\gamma\,B^3$ of the
146: bulk Lorentz factor of the flow and the post-shock magnetic field
147: ought to be roughly constant, while both are expected to decline
148: with time. In G2, the predictions by Li and Chevalier (2001) on
149: the late-time behaviour of this AG (in a model with two electron
150: energy distributions) are also found to fail.
151:
152: Faced with so much unsuccess, the authors of G2 conclude that, in
153: analogy to work by Frail et al. (2000) on GRB 970508, ``the simplest
154: explanation which is consistent with the data and requires no
155: significant modifications is that the blast wave of GRB 991208
156: entered a non-relativistic expansion phase several months after
157: the burst''. As in G1, no explicit support is given to the conclusion.
158:
159: The authors of G2 do not explain why they eliminate from their
160: analysis the optical measurements at $\rm t\!>\! 7$ days, a total
161: of 10 points in the R, V, B and I bands (Fig. 2 of Castro-Tirado
162: et al. 2001). Just the two R-band points at days $\sim\! 24$ and
163: 30 in our Fig. \ref{fone} are each $7 \sigma$ above the extrapolation
164: of the R-band fit in G2, after subtraction of the host galaxy
165: contribution ($\rm R=24.27\pm 0.13$, Castro-Tirado et al. 2001).
166: It is not obvious that the advocated late non-relativistic blast-wave
167: would remedy this discrepancy.
168:
169: \section*{The parameters of the CB model}
170:
171: In the CB model four parameters suffice to describe the {\it optical}
172: AGs in their various frequency bands. Three of them are ``intrinsic''
173: to the model: $\gamma_0$, the initial Lorentz factor of the CBs;
174: $\rm x_\infty$, the single parameter governing the deceleration of
175: a CB in the approximation of a constant-density interstellar medium
176: ($\rm x_\infty/\gamma_0$ is the distance required to half the
177: original Lorentz factor); and an overall normalization. A fourth
178: parameter $\theta$, the angle between the line of sight to the
179: observer and the direction of the CBs, must be extracted from the
180: AG fits, but it is, in the same sense as the redshift, not a
181: parameter describing the model per se.
182:
183: In extending the description of AGs from the optical to the radio
184: domain only one extra time-independent parameter is necessary: a
185: characteristic frequency for self-absorption within the CBs, $\rm
186: \nu_a$, for a total of 4 intrinsic parameters (DDD2). In DDD1 we
187: fit yet another parameter to the observations: the index $\rm p$
188: of the electron spectrum, prior to radiation losses. Having found
189: that, in the CB model, it was always compatible with the theoretical
190: expectation $\rm p\sim 2.2$, we no longer use it here, or elsewhere,
191: as a free parameter.
192:
193:
194: \section*{GRB 991208 in the CB model}
195:
196: In Fig. \ref{fone} we show the fit of DDD1 to the R-band AG. The
197: upper panel contains three contributions: the AG proper, the host
198: galaxy and a ``standard candle'' supernova akin to SN1998bw,
199: transported to the GRB's redshift\footnote{In the CB model, GRB
200: 980425 ---associated to SN1998bw--- is in no way exceptional (Dar
201: and De R\'ujula 2000, DDD1, DDD2). Unlike in the SM, it makes sense
202: to use this SN as a putative standard candle.}. In the lower panel
203: the galaxy is subtracted, demonstrating the presence of the SN: in
204: a CB-model analysis, in all instances wherein such a SN could be
205: seen, it was seen. This is so for all AGs with $\rm z\!<\!1.2$
206: (DDD1), including the cases where the presence of a 1998bw-like SN
207: was a prediction, based on the optical data {\it preceding} the
208: observable SN contribution (GRBs 011121 and 020405; Dado et al.
209: 2002c,d).
210:
211: In Figs. \ref{ftwo} and \ref{fthree} we show two fits to the radio
212: data of GRB 991208, for $\nu=1.43$, 4.86, 8.46 GHz (for which there
213: is abundant new data in G2) and for 15 GHz, at which an earlier
214: measurement had escaped our attention in DDD2. One of these wide-band
215: fits (WB1) is the one published in DDD2, the other is a new fit
216: (WB2), along identical lines, including the new radio data of G2.
217: The figures show that the predictions of DDD2 were very satisfactory:
218: the WB1 and WB2 curves are very similar and they both provide a
219: good description of the data, their difference not being larger
220: than the scintillating ups and downs of the data.
221:
222: In the Table we give the parameters for the R-band fit of DDD1,
223: the WB1 fit of DDD2 and the current WB2 fit. They appear to be
224: quite stable. Even the fit to only the R-band data determines
225: $\gamma_0$, $\theta$ and $\rm x_\infty$ to within a few percent of
226: the results of the WB2 fit (117 data points in total), even though
227: it is based on the mere dozen of early data points that are not
228: dominated by the SN. The value of $\rm x_\infty$ in the WB1 fit is
229: a bit smaller than in the others, the reason being that ---as can
230: be seen by inspection of Figs. \ref{ftwo}, \ref{fthree}--- this parameter is
231: sensitive to the late observations, and the early radio data of G1
232: dominated the WB1 fit. For fits that are so similar, their single
233: parameters describing the overall normalization are also necessarily
234: similar: we have not reported them in the Table.
235:
236: The CB model could be tested further by comparing the sky-projected
237: superluminal velocity of the CBs (that may be extracted from the AG's
238: radio scintillations, as for Galactic pulsars) with the predicted
239: $\rm v_{_{T}}(t)\simeq c\,\gamma(t)\,\delta(t)\, \theta/(1+z)$ (DDD2).
240:
241: \section*{Asymptotic behaviours in the CB model}
242:
243: According to G2 {\it ``one of the main challenges in modelling the AG
244: of GRB 991028''} is to reconcile the late radio decline
245: at 8.46 GHz with the
246: optical decay, which is ``twice'' as fast. In the CB model this is not a
247: challenge, both behaviours are correctly predicted:
248:
249: Let SEF refer to the ``proper'' CB wide-band
250: spectral energy flux (after subtraction
251: of the host galaxy and the associated SN).
252: Let $\rm\gamma=\gamma(t)$ be the explicit function describing the decreasing
253: Lorentz factor of the CBs (DDD1) and
254: $\delta\simeq 2\,\gamma/(1+\theta^2\gamma^2)$ the varying
255: Doppler factor of the radiation. The CB-model SEF has only two ``bends''.
256: Self-absorption within the CBs results, in their rest system,
257: in an opacity $\rm\tau =(\nu_a/\nu)^2\,(\gamma/\gamma_0)^2$,
258: parametrized by the single parameter $\rm \nu_a$, and responsible
259: for the turn down of the SEF towards low $\nu$. At higher
260: $\nu$ the spectral index steepens from $\sim\! -1/2$ to $\rm -p/2$
261: at an ``injection bend'' frequency:
262: \begin{equation}
263: \rm \nu_b(t) \simeq \rm {1.87\times 10^{15}\over 1+z}\,
264: \left[{\gamma^3\, \delta\over 10^{12}}\right ]\,
265: \left[{n_p\over 10^{-3}cm^{-3}}\right]^{1/2}\;\; Hz,
266: \label{nubend}
267: \end{equation}
268: in the observer's frame, with $\rm n_p$ the ISM number density (DDD2).
269:
270: After a couple of (observer) days, $\rm \gamma(t)\sim \delta(t)\sim t^{-1/3}$
271: and for frequencies above the opacity bend, or ``peak'', the SEF behaves
272: as:
273: \begin{eqnarray}
274: \rm F_{\nu(t)\ll \nu_b(t)}&\!\sim\!& \rm
275: \gamma^4\,
276: \nu^{-0.5}\sim
277: t^{-1.33}\, \nu^{-0.5}\, , \\
278: \rm F_{\nu(t)\gg \nu_b(t)}&\!\sim\!& \rm
279: \gamma^{2p+2}\,\nu^{-p/2}\sim t^{-2.13}\,
280: \nu^{-1.1}\, ,
281: \label{nuall}
282: \end{eqnarray}
283: where there may be a $\sim\!\pm 0.1 $ indetermination in the
284: the exponents of $\rm t$ and $\nu$,
285: due to the uncertainty around our adopted value, $\rm p=2.2$.
286: These predictions (or, rather, the explicit formula in DDD2 interpolating
287: them) are in agreement with the observations of all GRBs of known
288: $\rm z$. For the parameters that we fit to GRB 991208,
289: $\rm \nu(t)\!<\! \nu_b(t)$ in the radio at the late observed times,
290: and $\rm \nu(t)\!>\! \nu_b(t)$ in the optical, even during the
291: early optical observations.
292: In spite of their dependence on the arbitrarily chosen time intervals,
293: the observed late radio-behaviour at 8.46 GHz ($\rm \sim\! t^{-1.07\pm
294: 0.09}$, G2) and the optical result ($\rm\sim\! t^{-2.3 \pm 0.07}\,
295: \nu^{-1.05\pm 0.09}$, Castro-Tirado et al. 2001; $\rm\sim\! t^{-2.2
296: \pm 0.1}$, Sagar et al. 2000) are compatible with the
297: above CB-model expectations, q.e.d.
298:
299: \section*{Conclusions}
300:
301: We have reviewed a particular example of the failure of the standard
302: model of GRBs in describing an afterglow (G1, G2), and the existence
303: of a much simpler, predictive and successful alternative. This is
304: not an exception, to date there is no satisfactory and comprehensive
305: SM explanation of the AGs of all GRBs of known redshift. In DDD1
306: we have commented on some of the most complete studies (Frail et
307: al. 2001; Kumar and Panaitescu 2001), which have, among others,
308: the limitation of being ``anthropoaxial'' (all jets point to the
309: observer, an unlikely circumstance).
310:
311:
312: Most researches in most areas of science are motivated by challenging
313: their respective ``standard models''. This is the case even for
314: models that are currently flawless, such as the SM of particle
315: physics, or even for ``sacred'' pillars of science, such as quantum
316: mechanics and general relativity. In studying G1 and G2, as well
317: as most of the current GRB literature, it is difficult to suppress
318: the impression that, for observers and theorists alike, the opposite
319: motivation prevails. True enough, some of the profound inadequacies
320: of the ``fireball'' or ``firecone'' SMs models are occasionally
321: aired (e.g. Ghisellini 2001, Lazzati 2002), but the final verdict
322: is always benevolent. The fact that new epicycles must be added
323: with every ``novel'' observation is not unwelcome, even if the
324: additions are ponderous and totally ad-hoc, as is the case in the
325: interpretation of the $\rm Fe$ or other ``metal'' X-ray lines in
326: GRB AGs (reviewed in Lazzati 2002) or of the wiggly AG of GRB
327: 021004 (Lazzati et al. 2002; Nakar et al. 2002; Heyl and Perna
328: 2002). In everyone of these cases the CB model offers an incredibly
329: simpler alternative (Dado et al. 2002e,f).
330:
331: We are not saying that the CB model is entirely correct, it is a
332: simplification of what is no doubt a very complicated phenomenon;
333: it will either require modifications or turn out to be completely
334: wrong. But its assumptions and predictions should be tested against
335: observations, or challenged for consistency. In other realms of
336: science the existence of a sensible model challenging the standard
337: lore would be very welcome, as opposed to ignored.
338:
339:
340:
341:
342: {\bf Acknowledgment:} This research was supported in part by the Helen
343: Asher Space Research Fund and by the VPR fund for research at the
344: Technion. One of us, Arnon Dar, is grateful for hospitality
345: at the CERN Theory Division.
346:
347: \newpage
348: \begin{thebibliography}{}
349: \bibitem{}
350: Castro-Tirado, A.J., et al. 2001, A\&A, 370, 398
351: \bibitem{}
352: Dado S., Dar A., De R\'ujula A., 2002a, A\&A 388, 1079
353: \bibitem{}
354: Dado S., Dar A., De R\'ujula A., 2002b, A\&A in press (astro-ph/0204474)
355: \bibitem{}
356: Dado S., Dar A., De R\'ujula A., 2002c, ApJ, 572, L143
357: \bibitem{}
358: Dado S., Dar A., De R\'ujula A., 2002d, A\&A, 393, L25
359: \bibitem{}
360: Dado S., Dar A., De R\'ujula A., 2002e, ApJ, in press (astro-ph/0207015)
361: \bibitem{}
362: Dado S., Dar A., De R\'ujula A., 2002f, astro-ph/0211224
363: \bibitem{}
364: Dar A., De R\'ujula A., 2000, astro-ph/0008474
365: \bibitem{}
366: Frail, D.A., Waxman, E., Kulkarni, S.R., 2000, ApJ, 537, 191
367: \bibitem{}
368: Frail, D.A., et al. 2001, ApJ, 562, L55
369: \bibitem{}
370: Galama, T.J., et al. 2000, ApJ, 541, L45
371: \bibitem{}
372: Galama, T.J., et al. 2002, astro-ph/0211221
373: \bibitem{}
374: Ghisellini, G., 2001, astro-ph/0111584
375: \bibitem{}
376: Heyl, J.S., Perna, R., 2002, astro-ph/0211256
377: \bibitem{}
378: Hurley, K., et al. 2000, ApJ, 534, L23
379: \bibitem{}
380: Lazzati, D., 2002, astro-ph/0211174
381: \bibitem{}
382: Lazzati, D., et al. 2002, astro-ph/0210333
383: \bibitem{}
384: Li, Z., Chevalier, R.A., 2001, ApJ, 551, 940
385: \bibitem{}
386: Nakar, E., Piran, T., Granot, J., 2002, astro-ph/0210631
387: \bibitem{}
388: Panaitescu, A., Kumar, P., 2001, ApJ, 554, 667
389: \bibitem{}
390: Sagar, R., et al. 2000, BASI, 28, 15
391:
392:
393: \end{thebibliography}{}
394:
395:
396: \newpage
397:
398:
399: { \vskip 0.3 true cm
400: \noindent
401: {\bf Table:} Successive CB-model fits to the AG of GRB 991208.
402: R-band is a fit to only that optical frequency (DDD1). WB1 is the
403: wide-band fit in DDD2, with only the early radio-data. WB2 is
404: the current fit to all data.}
405: \vskip 0.5 true cm
406: \begin{table}[h]
407: %\huge\bf
408: \normalsize
409: \hspace{.0cm} %if you want to center your table act on this argument
410: \begin{tabular}{|l|c|c|c|c|}
411: \hline
412:
413: Parameter & R-band & WB1 & WB2 \\
414: \hline
415: $\theta$ [mrad] & 0.100 & 0.111 & 0.103 \\
416: $\gamma_0$ & 1034 & 1034 & 1089 \\
417: $\rm x_{\infty} $ [Mpc] & 1.357 & 1.014 & 1.382 \\
418: $\rm\nu_a $ [MHz] & **** & 103 & 89 \\
419: \hline
420: %
421:
422: \end{tabular}
423: \end{table}
424:
425:
426:
427: \begin{figure}[t]
428: \begin{tabular}{cc}
429: \hskip 2truecm
430: %\vspace*{-1cm}
431: \hspace*{-1.7cm}
432: \epsfig{file=f1.eps, width=9cm} \\
433: %\hskip 1truecm
434: \hspace*{.5cm}
435: \epsfig{file=f2.eps, width=9cm}
436: \end{tabular}
437: \caption{Comparisons between the R-band AG
438: (upper curves) and the observations,
439: not corrected for extinction,
440: for GRB 991208, at $\rm z=0.706$ (DDD1).
441: Upper panel: without subtraction of the host
442: galaxy's contribution (the straight line).
443: Lower panel: with the host galaxy subtracted. The contribution
444: from a 1998bw-like supernova placed at the GRB's
445: redshift, corrected for extinction,
446: is indicated in both panels by a line of crosses.
447: The SN contribution is clearly discernible.}
448: \label{fone}
449: \end{figure}
450:
451: \begin{figure}[t]
452: \begin{tabular}{cc}
453: \hskip 2truecm
454: %\vspace*{2cm}
455: \hspace*{-1.7cm}
456: \epsfig{file=f3.eps, width=10cm} \\
457: %\hskip 1truecm
458: \hspace*{.5cm}
459: \epsfig{file=f4.eps, width=10cm}
460: \end{tabular}
461: \caption{Comparisons between the CB-model radio light-curves and
462: the observations at 1.43 and 4.86 GHz. The (red) continuous curve
463: in the upper panel and the (red) higher-up dotted line in the lower
464: panel are the results of DDD1, obtained without the new data,
465: represented by stars. The other dotted lines in both panels depict
466: the current fit to all data.}
467:
468: \label{ftwo}
469: \end{figure}
470:
471: \begin{figure}[t]
472: \begin{tabular}{cc}
473: \hskip 2truecm
474: %\vspace*{2cm}
475: \hspace*{-1.7cm}
476: \epsfig{file=f5.eps, width=10cm} \\
477: %\hskip 1truecm
478: \hspace*{.5cm}
479: \epsfig{file=f6.eps, width=10cm}
480: \end{tabular}
481: \caption{Comparisons between the CB-model radio light-curves
482: and the observations at 8.46 and 15 GHz. The (red)
483: higher-up dotted lines are the results of DDD1, obtained
484: without the new data, represented by stars. The other dotted lines
485: in both panels depict the current fit to all data.}
486: \label{fthree}
487: \end{figure}
488:
489:
490: \end{document}
491:
492:
493:
494:
495:
496:
497:
498: