1: \documentclass[preprint,12pt]{aastex}
2: %\documentclass[preprint2]{aastex}
3: %\documentclass{emulateapj}
4:
5:
6: \usepackage{graphicx}
7: \usepackage{amsmath}
8: \usepackage{amssymb}
9: %\usepackage{mathrsfs}
10: %\usepackage{aastex}
11:
12: %\linespread{0.8}
13:
14: %\newcommand{\Tfl}{$\theta_{fl}$}
15: \newcommand{\EQ}[1]{Eq.(\ref{EQ #1})} % Equation referencing
16: \newcommand{\FIG}[1]{Fig.(\ref{FIG #1})} % figure referencing
17:
18: \begin{document}
19:
20: \title{GRB 990123 revisited: Further Evidence for a Reverse Shock}
21:
22: \author{Ehud Nakar \altaffilmark{1,2} and Tsvi Piran \altaffilmark{1}}
23:
24: \altaffiltext{1}{Racah Institute for Physics, The Hebrew
25: University, Jerusalem 91904, Israel}
26:
27: \altaffiltext{2}{ Institut d`Astrophysique de Paris, 75014 Paris,
28: France}
29:
30: \begin{abstract}
31: Recently we have presented a new theoretical analysis of the
32: reverse shock emission. We use this analysis here to revisit GRB
33: 990123. We find new and compelling evidences that the optical
34: flash and the radio flare of GRB 990123 resulted from a reverse
35: shock. This suggests that a significant fraction of the energy of
36: the relativistic ejecta must have been carried by baryons. It also
37: suggests that the external medium is an ISM and that in this burst
38: the reverse shock emission dominated at early time over other
39: possible processes. We use the early optical emission to constrain
40: the physical parameters of the original ejecta and the microscopic
41: parameters in the emitting reverse shocked region.
42: \end{abstract}
43: \maketitle
44:
45: \section{Introduction}
46:
47: The 9th magnitude optical flash of GRB 990123 was one of the most
48: exciting discoveries associated with Gamma-Ray Bursts (GRBs). This
49: was followed by a detection of a strong unusual radio flare from
50: the same source (Kulkarni et al. 1999, Galama et al. 1999). One of
51: the most impressive facts concerning the strong optical flash of
52: GRB 990123 was that such a strong prompt optical emission was
53: predicted, just a few months earlier to arise from the reverse
54: shock during the early afterglow (Sari, \& Piran 1999a). Later
55: this interpretation was questioned and other models for prompt
56: optical emission were proposed (e.g. Beloborodov, 2002). We return
57: to this issue here.
58:
59: In the popular fireball model (Piran 1999) for Gamma-Ray Bursts
60: (GRBs), the afterglow results from a blast-wave that propagates
61: into the circum-burst medium. This blast-wave is originated by the
62: energy dissipation of the relativistic ejecta to the circum-burst
63: medium. In this model the early afterglow is produced during this
64: dissipation, and thus can be used to study the nature of the
65: ejecta. In case that the ejecta is Baryonic, a reverse shock is
66: produced during this dissipation. This reverse shock is predicted
67: to produce strong optical flash and radio flare (Sari \& Piran
68: 1999b,c).
69:
70: Recently we have presented new theoretical analysis of the reverse
71: shock emission (Nakar \& Piran 2004; hereafter NP04). This
72: analysis includes a new test of reverse shock emission and
73: diagnostic tools for the physical parameters of the original
74: ejecta. We use these new tools, here, to re-investigate the early
75: afterglow of GRB 990123. Note that numerous authors analyzed the
76: early afterglow of GRB 990123 (Sari \& Piran 1999b; M\'esz\'aros
77: \& Rees 1999; Kulkarni et al. 1999; Galama et al. 1999; Wang, Dai
78: \& Lu 2000; Panaitescu \& Kumar 2001; Fan et al. 2002; Soderberg
79: \& Ramirez-Ruiz 2002). We don't attempt to review here these
80: results and compare them as the scope of such a review is too long
81: for a short letter.
82:
83: The reverse shock optical emission has a typical and
84: characteristic $t^{-2}$ decay that was observed in the light
85: curve of GRB 990123 (Fenimore et al. 1999; Sari \& Piran
86: 1999b,c). We find additional conclusive evidence that the optical
87: flash and the radio flare of GRB 990123 indeed resulted from the
88: emission of a reverse shock produced by an interaction of the
89: ejecta with and ISM. These results suggest first that at least in
90: this burst the reverse shock emission, and not another mechanism
91: such as a pair enriched forward shock (due to interaction of the
92: prompt $\gamma$-rays with the circum burst medium; Thompson \&
93: Madau 2000, Beloborodov, 2002), dominated the early afterglow. As
94: the reverse shock requires, a baryonic component in the ejecta it
95: also implies that a significant fraction of the energy of the
96: relativistic ejecta was carried by baryons.
97:
98: The sparse observations before the peak of the optical flash (only
99: one measurement before the peak at each wavelength) poses
100: difficulties in analyzing the properties of the ejecta, since this
101: analysis depends on the rising power-law index of the optical
102: light curve\footnote{Since there is only a single observation
103: before the peak, the rising phase may also be very irregular with
104: no power-law behavior. In this case the analysis of the ejecta
105: properties presented in NP04 does not apply, as it imply that the
106: relativistic ejecta hydrodynamical profile is highly irregular.
107: However, having no detailed observations of this phase we use here
108: the simplest light curve that fits the observations - a broken
109: power-law. Note that the tests, which show that the emission is a
110: reverse shock emission, are independent of the shape of the rising
111: light curve, and can be carried also if it is irregular.}. Still,
112: we are able to determine that the reverse shock was mildly
113: relativistic and that the initial Lorentz factor is most likely
114: smaller than $\sim 600 (E_{54}/n)^{1/8}$, where $E_{54}$ is the
115: (isotropic equivalent) kinetic energy in the ejecta after the
116: prompt emission phase in units of $10^{54}$ergs and $n$ is the
117: protons density of the circum-burst interstellar medium in c.g.s.
118: The width of the ejecta is found to be, as expected, roughly equal
119: to the burst duration (multiplied by the light speed). Because of
120: the sparse data available we obtain only upper limits to the
121: initial Lorentz factor and the shell's width and only lower limits
122: to the microscopic equipartition parameters and the external
123: density. We expect that in the future detailed optical light
124: curves, provided by Swift and rapid follow up observations, would
125: enable an accurate determination of the initial physical
126: parameters of the relativistic wind by using the method we apply
127: here.
128:
129: For completeness we begin (\S \ref{sec Theory}) with a short
130: summary of the theoretical model of the reverse shock emission
131: presented in NP04. We summarize, in \S \ref{sec Observations}, the
132: early afterglow observations of GRB 990123. The analysis of the
133: optical light curve and the resulting constraints on the physical
134: parameters of the relativistic wind is presented in \S \ref{sec
135: analysis}. In \S \ref{sec test} we test whether the observations
136: fit with a reverse shock emission. The results, their
137: implications to GRB modelling and future prospects are summarized
138: in \S \ref{sec Conclusions}.
139:
140: \section{The reverse shock emission - Theory}\label{sec Theory}
141:
142: We begin with a brief summary of the main results of NP04. These
143: results give the theoretical predictions of the optical and the
144: radio emission of a baryonic reverse shock in an ISM environment.
145: NP04 suggest to model the early optical light curve (similarity
146: to late afterglow parametrization; Beuermann et al. 1999) as a
147: broken power-law with five parameters:
148: \begin{equation}\label{EQ FnurParam}
149: F_{\nu,opt}^r(t)=F_0^r \left( \frac {1}{2} \left( \frac{t}{t_0} \right) ^{-s\alpha_1}
150: + \frac {1}{2} \left( \frac{t}{t_0} \right) ^{-s\alpha_2} \right)
151: ^{-\frac{1}{s}} ,
152: \end{equation}
153: where $\alpha_1>0$ and $\alpha_2<0$ are the asymptotic power-law
154: indices, $F_{\nu,opt}^r(t_0)=F_0^r$ is the peak flux and the
155: parameter $s$ determines the sharpness of the peak (a large $s$
156: corresponds to a sharp peak). The physical parameters which
157: determine the emission of the reverse shock are: (1) The
158: (isotropic equivalent) energy $E$, the width $\Delta$ and the
159: initial Lorentz factor $\Gamma_o$ of the ejected wind. (2) The
160: circum burst density, $n$, that we consider as a constant
161: interstellar medium (ISM). (3) The microscopic parameters in the
162: emitting region, namely the energy equipartition parameters
163: $\epsilon_e$ and $\epsilon_B$ that describe the ratio of the
164: electrons and Magnetic field energy to the total internal energy,
165: and $p$, the power-law index of the electrons' energy
166: distribution.
167:
168: The strength of the reverse shock is determined by the
169: dimensionless parameter $\xi$ (Sari \& Piran 1995):
170: \begin{equation}\label{EQ xi}
171: \xi\equiv\frac{l^{1/2}}{\Delta^{1/2}\Gamma_o^{4/3}} ,
172: \end{equation}
173: where $l \equiv (3E/(4\pi nm_pc^2))^{1/3}$ ($m_pc^2$ is the proton
174: rest mass energy). According to this definition:
175: \begin{equation}\label{EQ gamma0}
176: \Gamma_o=188\xi^{-3/4}\Delta_{12}^{-3/8}(E_{52}/n)^{1/8} .
177: \end{equation}
178: The peak of the optical emission is observed at
179: \begin{equation}\label{EQ t0}
180: t_0=\frac{\Delta}{c}(1+0.7\xi^{3/2})(1+z) ,
181: \end{equation}
182: when the reverse shock finishes crossing the ejecta. $c$ is the
183: speed of light and $z$ is the redshift. The structure of the
184: optical and the radio reverse shock light curves depend on the
185: relative values of the three reverse shock break frequencies at
186: $t_0$: $\nu_a^r$, the self-absorbtion frequency, $\nu_m^r$, the
187: synchrotron frequency and $\nu_c^r$ the cooling frequency. In
188: principal numerous patterns are possible. However, over a wide
189: range of the parameter space (and as we show below, in
190: particularly for GRB 990123) these frequencies satisfy
191: $\nu_{radio}<\nu_m^r(t_0)<\nu_a^r(t_0)<\nu_{opt}<\nu_c^r(t_0)$,
192: where $\nu_{radio}$ and $\nu_{opt}$ are the observed radio and
193: optical frequencies respectively. We consider this frequencies
194: sequence as the {\it generic case}. In this case the peak optical
195: flux is:
196: \begin{equation}\label{EQ F0r}
197: F_0^r=0.1{\rm mJy}(1+z)^{-\frac{4+p}{8}}1.5^{2.5-p}
198: \left(\frac{3(p-2)}{p-1}\right)^{p-1}\epsilon_{e-1}^{p-1}
199: \epsilon_{B-2}^{\frac{p+1}{4}}n^{\frac{p+2}{8}}E_{52}^{1+\frac{p}{8}}
200: t_{0,2}^{-\frac{3p}{8}}D_{28}^{-2}A_{F,0}^r(\xi),
201: \end{equation}
202: where throughout the paper we denote by $Q_{x}$ the value of the
203: quantity $Q$ in units of $10^{x}$ (c.g.s), and $D$ is the proper
204: distance. The function $A_{F,0}^r(\xi)$ is approximated in the
205: range of $0.1<\xi<2.5$ by:
206: \begin{equation}\label{EQ AF0rApprox}
207: A_{F,0}^r(\xi) \approx 180 \xi^{0.65} \left(6 \cdot 10^{-4}
208: \xi^{-2.6}\right)^\frac{p-1}{2}.
209: \end{equation}
210: $A_{F,0}^r(\xi)$ values out of this range can be found in NP04.
211:
212: The flux before $t_0$ depends strongly on $\xi$, and thus the
213: rising slope, $\alpha_1$ can be used to determine $\xi$ using (in
214: the generic case):
215: \begin{equation}\label{EQ alpha1}
216: \alpha_1 \approx 1.2(0.5+\frac{p}{2}(\xi-0.07\xi^2)).
217: \end{equation}
218: The reserve shock decay index is, however, a constant:
219: \begin{equation}\label{EQ alpha1}
220: \alpha_2 \approx -2,
221: \end{equation}
222: and its main use is to identify the reverse shock emission.
223: Finally, the sharpness parameter, $s$, depends mainly on the
224: profile of the ejecta. A sharp break with $s \gtrsim 3$ is
225: expected when the ejecta is rather homogenous, while a gradual
226: break with a lower value of $s \sim 1$ is expected if the spread
227: of the initial Lorentz factor within the shell is large.
228:
229: In contrast to the optical emission, the radio continues to rise
230: at $t>t_0$ and it peaks at a later time, $t_*$, when
231: $\nu_{radio}=\nu_a^r$. Over a wide range of $\xi$ values (when the
232: shock is not ultra relativistic) $\nu_{radio}<\nu_m^r(t_0) <
233: \nu_a^r(t_0) \approx 10^{12-13}$Hz. In this case the radio
234: emission is expected to rise first as $\sim t^{0.5}$ until
235: $\nu_{radio}=\nu_m^r$ and then as $\sim t^{1.25}$ until $t_*$. At
236: $t>t_*$ it is expected to decay, similarly to the optical
237: emission, as $\sim t^{-2}$. This behavior provides a second and
238: independent test for the reverse shock emission:
239: \begin{equation}\label{EQ radioTest}
240: \frac{F_*}{F_0}\left(\frac{t_*}{t_0}\right)^{\frac{p-1}{2}+1.3}=\left(\frac{\nu_{opt}}{\nu_{radio}}\right)^\frac{p-1}{2}
241: \sim 1000,
242: \end{equation}
243: where this value can be larger or smaller by a factor of $\sim 3$
244: (for a given $p$) due to uncertainty in the hydrodynamics (see
245: Kobayashi \& Sari 2000). It provides also an independent
246: measurement of $\nu_a^r(t_0)$:
247: \begin{equation}\label{EQ nuat0}
248: \nu_a^r(t_0)\approx \frac{t_*}{t_0} \nu_{radio}
249: \end{equation}
250: Detailed radio observations at $t<t_*$ that identify the break
251: during the rising phase, when $\nu_m^r=\nu_{radio}$, would enable
252: determination of $\nu_m^r(t_0)$ as well.
253:
254:
255: \section{Observations}\label{sec Observations}
256: At $z=1.6$ GRB 990123 is one of the brightest $\gamma$-rays bursts
257: observed so far (Briggs et al. 1999). The main event of this burst
258: is composed of two very energetic and hard pulses which last
259: together $\sim 25$sec. These two pulses contain almost all the
260: emitted energy during the burst. A much softer and less energetic
261: emission ($<120$keV) is observed $\sim 20$sec before and up to
262: $\sim 50$sec after this event. The total isotropic equivalent
263: energy emitted in $\gamma$-rays during the burst is $\sim
264: 10^{54}$ergs. The timing of the main $\gamma$-rays event differ
265: between different detectors. While BATSE triggered on the first
266: soft emission, $\sim 20$sec before the main event, BeppoSAX
267: triggered $18$sec later than BATSE on the beginning of the main
268: event.
269:
270: GRB 990123 is also the only burst where optical emission was
271: observed during the prompt emission (Akerlof et al. 1999). The
272: first detection of $11.7$mag is recorded $\sim 7 \rm sec$ after
273: BeppoSAX trigger. The optical emission peaks at the second
274: snapshot that took place $\sim 32$sec after BeppoSAX trigger with
275: $8.9$mag ($\approx 0.8$Jy)! After the peak, a fast decay is
276: observed with a power law slope of $\approx -2$ which becomes
277: shallower at later times.
278:
279: The early radio observations of GRB 990123 show also an unusual
280: flare (Kulkarni et al. 1999). The first observation after
281: $0.25$days shows a 8.46GHz intensity of $62\pm 32\mu$Jy. The flux
282: in this band rises to a peak flux of $260\pm 32\mu$Jy after
283: $1.25$days and then decline rapidly. Observations at other
284: wavelengths from this epoch ($\sim 1$day) (Galama et al. 1999)
285: show a flux of $118 \pm 40\mu$Jy at $4.88$GHz and upper limits of
286: several hundreds $\mu$Jy in $15$, $86$ and $222$GHz.
287:
288:
289:
290: \section{Constraining the burst parameters} \label{sec analysis}
291:
292: Already in 1999 Fenimore et al. (1999) pointed out that the
293: initial decay index of the optical flash $\alpha_2 \approx -2$
294: agrees with the predictions of a reverse shock emission (Sari \&
295: Piran 1999a,b,c). At later times the decay becomes more moderate
296: as expected from the forward shock contribution. Motivated by
297: this observation, we first use the optical observations to
298: constrain the physical parameters of the burst assuming that the
299: flash arises from a reverse shock emission. Later, we carry the
300: additional tests of reverse shock emission, described in NP04.
301:
302: When modelling the early afterglow, it is important to find the
303: time when the main part of the relativistic ejecta is emitted from
304: the source. It is not necessarily the trigger time. In GRB 990123
305: the radiated energy, and therefore most likely the energy ejected
306: from the source, is clearly dominated by the main two pulses.
307: Kobayashi, Sari \& Piran 1997 (see also Nakar \& Piran 2002), have
308: shown that in internal shocks the observed time of the
309: $\gamma$-rays pulses reflects the emission time of the wind from
310: the source. Therefore we estimate that the main part of the
311: relativistic wind is beginning to be ejected by the source at the
312: time that the first dominant pulse starts rising - the BeppoSax
313: triggering time ($18$sec after the BATSE trigger). This is the
314: point where we set the observer clock to $t=0$. According to this
315: setting the duration of the burst is $\sim 25$sec and the peak of
316: the optical emission is observed at $t=32$sec.
317:
318: The observations of the optical flash of GRB 990123 (Akerlof et
319: al. 1999) include only one data point before the peak. Since there
320: are two observations after the peak which show similar decay
321: ($t=57$sec and $t=142$sec), $\alpha_2 \approx -2$, a broken power
322: law fit is possible only if the peak is between the first and the
323: second observations. Thus, we have only lower limits to
324: $F_0>0.8$Jy, $\alpha_1>2$, an upper limit to $t_0<32$sec and $s$
325: can take any value (see Fig. 1). As we have only a lower limit on
326: $\alpha_1$, Eq. (\ref{EQ alpha1}) yields only a lower limit on
327: $\xi$ (we use $p=2.3$ thorough out.):
328: \begin{equation}\label{EQ xi990123}
329: \xi \gtrsim 1.
330: \end{equation}
331: Eqs. (\ref{EQ t0} \& \ref{EQ gamma0}) result in the upper
332: limits:
333: \begin{equation}\label{EQ D990123}
334: \Delta \lesssim 2.5 \cdot 10^{11} \rm cm,
335: \end{equation}
336: and
337: \begin{equation}\label{EQ Gamma990123}
338: \Gamma_o \lesssim 600 \left(\frac{E_{54}}{n}\right)^{1/8} .
339: \end{equation}
340: It is reassuring to find that the width of the shell obtained for
341: $\xi \approx 1$ divided by the speed of light is similar to the
342: duration of the GRB ($\sim 10$sec in the burst's frame).
343:
344:
345:
346: Next we use Eq. (\ref{EQ F0r}) to constrain the microscopic
347: equipartition parameters, $\epsilon_e$ and $\epsilon_B$, and the
348: external density, $n$. We consider the total energy emitted in
349: $\gamma$-rays, $\sim 10^{54}$ergs, as a reasonable estimate of
350: the remaining energy in the ejecta, $E$. Therefore, we do not
351: attempt to estimate it from the optical flash observations.
352: Instead we express the dependance of the resulting constraints on
353: the value of $E$. Taking $\xi>1$ and requiring $F_0>0.8$Jy leads,
354: using Eq. (\ref{EQ F0r}), to:
355: \begin{equation}\label{EQ constraint}
356: \epsilon_{e-1}^{p-1}
357: \epsilon_{B-2}^{\frac{p+1}{4}}n^{\frac{p+2}{8}} \gtrsim 15E_{54}^{-(1+p/8)}
358: \end{equation}
359: Another constraint follows the requirement that
360: $\nu_{opt}<\nu_c^r(t_0)$ (otherwise the optical light curve would
361: not show the generic behavior of $\alpha_2 \approx 2$) \footnote{
362: The requirement that $\nu_{radio}<\nu_m^r(t_0)<\nu_{opt}$ is
363: satisfied for any reasonable value of the parameters and therefore
364: cannot be used as a constraint. The values of $\nu_m^r(t_0)$ and
365: $\nu_c^r(t_0)$ can be found in NP04.}:
366: \begin{equation}\label{EQ nuc_constraint}
367: n \epsilon_{B-2}^{3/2} \lesssim 20E_{54}^{-1/2}.
368: \end{equation}
369: These two constraints (together with the trivial one, $\epsilon_e+
370: \epsilon_B<1$) result in the following limits:
371: \begin{equation}\label{EQ ee constraint}
372: \epsilon_e \gtrsim 0.1 E_{54}^{-0.75},
373: \end{equation}
374: \begin{equation}\label{EQ eb constraint}
375: 5\cdot 10^{-3} E_{54}^{-1.6} n^{-2/3} \lesssim \epsilon_B \lesssim 0.1 E_{54}^{-1/3} n^{-2/3},
376: \end{equation}
377: \begin{equation}\label{EQ ee constraint}
378: n \gtrsim 5 \cdot 10^{-3} E_{54}^{-2.4} \rm cm^{-3}
379: \end{equation}
380:
381:
382:
383:
384: \section{Was the optical flash a result of a reverse shock
385: emission?}\label{sec test}
386:
387:
388: We have already mentioned that the optical flash of GRB 990123
389: passes the first test of a reverse shock emission: $\alpha_2
390: \approx -2$ (Fenimore et al. 1999, Sari \& Piran 1999a,b,c). NP04
391: have shown that in the generic case
392: ($\nu_{radio}<\nu_m^r<\nu_a^r<\nu_{opt}<\nu_c$ at $t_0$) the
393: reverse shock emission results also in a tight relation between
394: the radio and the optical emission, Eq. (\ref{EQ radioTest}). For
395: the observed frequencies of GRB 990123 ($\nu_{opt}=5\cdot
396: 10^{14}$Hz and $\nu_{radio}=8.6$GHz) and $p=2.3$:
397: \begin{equation}\label{EQ radioobs1}
398: \left(\frac{\nu_{opt}}{\nu_{radio}}\right)^\frac{p-1}{2} =1300,
399: \end{equation}
400: The observations of GRB 990123 show a remarkable agreement with
401: this prediction (taking the times and the fluxes of the flash and
402: the flare as the times and the fluxes of the peak observation,
403: $t_0=32$sec, $F_0=0.8$Jy, $t_*=1.25$day and $F_*=260\mu$Jy):
404: \begin{equation}\label{EQ radioObs2}
405: \frac{F_*}{F_0}\left(\frac{t_*}{t_0}\right)^{\frac{p-1}{2}+1.3}
406: \approx 800-4000,
407: \end{equation}
408: where the result include the uncertainty in the hydrodynamics.
409: NP04 show also that if the above test is passed, then the self
410: absorption frequency at $t_0$, $\nu_a^r(t0)$, can be estimated
411: from the the ratio between $t_*$ and $t_0$:
412: \begin{equation}\label{EQ nuaObs}
413: \nu_a^r(t_0) \approx \frac{t_*}{t_0}\nu_{radio}=3 \cdot
414: 10^{13}Hz.
415: \end{equation}
416: This frequency is determined by the physical conditions at $t_0$.
417: As a consistency check we compare this result to the possible
418: values of $\nu_a^r(t_0)$ obtained by the limits on the physical
419: parameters determined from the analysis of the optical emission.
420: We consider the case of $\xi=1$ and p=2.3 for which:
421: \begin{equation}
422: \nu_a^r(t_0)=6 \cdot 10^{12}Hz \left[(1+z)^{-\frac{p+6}{8}}
423: \epsilon_{e,-1}^{\;p-1}\epsilon_{B,-2}^\frac{p+2}{4} (n
424: E_{54})^\frac{p+6}{8}t_{0,2}^{-\frac{3p+10}{8}}
425: \right]^\frac{2}{p+4}.
426: \end{equation}
427: Within the parameters space of $\epsilon_e$,$\epsilon_B$ and $n$
428: which satisfy Eqs. (\ref{EQ constraint} \& \ref{EQ
429: nuc_constraint}) and taking $E_{54}=1$ we obtain $1.5 \cdot
430: 10^{13}Hz<\nu_a^r(t_0)<3.5 \cdot 10^{13} n^{1/6}Hz$! Although the
431: dependance of this range on the value of $E$ is not trivial, it is
432: weak and the value $\nu_a^r(t_0)=3 \cdot 10^{13}Hz$ is consistent
433: with the allowed range for any reasonable value of $E$,
434: $0.5<E_{54}<10$. This constraint on the value of $\nu_a^r(t_0)$ is
435: independent of the value found in Eq. (\ref{EQ nuaObs}). Once more
436: the agreement between the two estimates is remarkable.
437:
438: So far we focused on the radio observations at 8.46Ghz. The lower
439: radio fluxes both above and below 8.46GHz, at $t_*$, are
440: consistent with reverse shock emission, which predicts that the
441: peak frequency at $t_*$ is the observed $\nu_{radio}$ (8.46Ghz in
442: this case). Finally we note that although radio observations from
443: $t<t_*$ can theoretically be used to farther constrain the
444: conditions at $t_0$, the large errors of the single radio
445: observation from this epoch is insufficient to do so.
446:
447: \section{Conclusions}
448: \label{sec Conclusions}
449:
450: We have analyzed the early afterglow of GRB
451: 990123 according to the theoretical predictions of the reverse
452: shock emission presented in NP04. We find that GRB 990123 shows
453: the clear signature of a reverse shock emission in two independent
454: and robust tests. Apart from these two tests we carry a
455: consistency check between the radio emission and our analysis of
456: the optical emission which is passed successfully. This
457: consistency check gives us further confidence both in the analysis
458: and in the fact that the optical flash and the radio flare of GRB
459: 990123 are a generic case of a reverse shock emission.
460:
461: The consistency with the reverse shock model further suggests, but
462: does not proof of course, that the optical flash of GRB 990123 did
463: not arise from a pair loaded forward shock (Beloborodov, 2002). In
464: turn this indicates that the ejecta of GRB 990123 had (at least at
465: this stage) a significant baryonic component, as otherwise, for
466: example in a pure Poynting flux flow, a reverse shock is not
467: expected.
468:
469: Our analysis of the optical emission shows that the reverse shock
470: was mildly relativistic and that the initial Lorentz factor is
471: less than $600(E_{54}/n)^{1/8}$. We find that the width of the
472: initial shell is similar to the duration of the burst or smaller.
473: We constrain also the equipartition parameters and find that
474: $\epsilon_e \gtrsim 0.1 E_{54}^{-0.75}$ while $5\cdot 10^{-3}
475: E_{54}^{-1.6} n^{-2/3} \lesssim \epsilon_B \lesssim 0.1
476: E_{54}^{-1/3} n^{-2/3}$. The external density is limited by $n
477: \gtrsim 5 \cdot 10^{-3} E_{54}^{-2.4}\rm cm^{-3}$.
478:
479:
480:
481: The determination of the initial conditions of GRB 990123 and the
482: constraints of the microscopic parameters and the external density
483: is very limited due to the sparse optical and radio observations.
484: Swifts and its follow up observations would hopefully provide
485: during the next few years detailed light curves. These would
486: enable a more detailed analysis and a much better determination of
487: the parameters of the burst's relativistic outflow.
488:
489: We thank Robert Mochkovitch, Frederic Daigne and Elena Rossi for
490: helpful discussions. The research was supported by the US-Israel
491: BSF and by EU-RTN: GRBs - Enigma and a Tool. EN is supported by
492: the Horowitz foundation and by a Dan David Prize Scholarship 2003.
493:
494: \begin{thebibliography}{99}
495: \bibitem[]{}Akerlof, C. et al., 1999, Nature, 398, 400
496:
497: \bibitem[]{}Beloborodov, A.~M.\ 2002, \apj, 565, 808
498:
499: \bibitem[]{}Beuermann et al., 1999, A\&A, 352, L26
500:
501: \bibitem[]{}Briggs, M. S. et al., 1999, ApJ, 524, 82
502:
503: \bibitem[]{}Fan, Y. Z., Dai, Z. G., Huang, Y. F. \& Lu T., 2002, ChJAA, 2,
504: 449
505:
506: \bibitem[]{}Fenimore, E. E., Ramirez-Ruiz, E. \& Wu, B., 1999. ApJ, 518, L73
507:
508: \bibitem[]{}Galama, T, J., 1999, Nature, 398, 394
509:
510: \bibitem[]{}Kulkarni, S. et al, 1999, ApJ, 522, L97
511:
512: \bibitem[]{}Kobayashi, S., Piran, T. \& Sari, R, 1997, ApJ, 490, 92
513:
514: \bibitem[]{}Kobayashi, S. \& Sari, R., 2000, ApJ, 542, 819
515:
516: \bibitem[]{}M\'esz\'aros, P. \& M.J. Rees, 1997,ApJ, 476, 232
517:
518: \bibitem[]{}M\'esz\'aros, P. \& M.J. Rees, 1999, MNRAS, 306, L39
519:
520: \bibitem[]{}Nakar, E. \& Piran, T., 2002, ApJ, 572, L139
521:
522: \bibitem[]{}Nakar, E. \& Piran, T., 2004, astro-ph/0403461 (NP04)
523:
524: \bibitem[]{}Panaitescu, A. \& Kumar, P., 2001, ApJ, 554, 11
525:
526: \bibitem[]{}Piran, T., 1999, Physics Reports, 314, 575
527:
528: \bibitem[]{}Sari, R. \& Piran, T., 1995, ApJ, 455, L143
529:
530: \bibitem[]{} Sari, R.~\& Piran, T. 1999a, Astron. \& Astrophys. Supp., 138, 537
531:
532: \bibitem[]{}Sari, R. \& Piran, T., 1999b, ApJ, 517, L109
533:
534: \bibitem[]{}Sari, R. \& Piran, T., 1999c, ApJ, 520, 641
535:
536: \bibitem[]{} Thompson, C. \& Madau, P., 2000, Apj 538, 105
537:
538: \bibitem[]{}Soderberg, A. M. \& Ramirez-Ruiz, E., 2002, MNRAS, 330, L24
539:
540: \bibitem[]{}Wang, X. Y., Dai, Z. G. \& Lu, T. 2000, MNRAS, 319, 1159
541: \end{thebibliography}
542:
543: \begin{figure}[h]
544: \begin{center}
545: \includegraphics[width=14cm,height=8cm]{f1.eps}
546: \caption{Possible light curves (with the value of $s$ beside each
547: light curve) fitted according to Eq. (\ref{EQ FnurParam}) to the
548: optical observations ({\it full dots}) of GRB 990123.}
549: \label{plotone}
550: \end{center}
551: \end{figure}
552:
553: \end{document}
554: