1:
2: \documentclass[12pt,preprint]{aastex}
3: \usepackage{epsfig}
4: % for a referee version
5: %\usepackage{epsfig,deluxe}
6: % \documentclass[manuscript]{aastex}
7:
8:
9: %\received{2002 November 27}
10: \begin{document}
11:
12: \newcommand{\gsim}{\hbox{\rlap{$^>$}$_\sim$}}
13: %\voffset-.6cm
14:
15: \title{What we learn from the afterglow of GRB 021211}
16:
17: \author{Shlomo Dado\altaffilmark{1}, Arnon Dar\altaffilmark{1,2}
18: and A. De R\'ujula\altaffilmark{2}}
19:
20: \altaffiltext{1}{Physics Department and Space Research Institute,
21: Technion, Haifa 32000, Israel} \altaffiltext{2}{Theory Division,
22: CERN, CH-1211 Geneva 23, Switzerland}
23:
24:
25: % \maketitle
26:
27:
28: \begin{abstract}
29:
30: The behaviour of the afterglow (AG) of gamma-ray bursts (GRBs)
31: directly provides, in the cannonball (CB) model, information about
32: the environment of their progenitor stars. The well observed early
33: temporal decline of the AG of GRB 021211 is precisely the one
34: predicted in the presence of a progenitor's ``wind'' which resulted
35: in a density profile $\propto 1/r^2$ around the star. The subsequent
36: fast fading ---which makes this GRB ``quasi-dark''---
37: is the one anticipated if, further away, the interstellar density is
38: roughly constant and relatively high. The CB-model fit to the AG clearly
39: shows the presence of an associated supernova akin to SN1998bw, and allows
40: even for the determination of the broad-band spectrum of the host galaxy.
41: GRB 990123 and GRB 021004, whose AGs were also measured very early, are
42: also discussed.
43:
44: \end{abstract}
45:
46:
47: \keywords{gamma rays: bursts}
48:
49: \section{Introduction}
50:
51: Massive stars are observed to emit a dense stellar wind, continuously
52: or in a sequence of eruptions, which presumably intensifies before
53: their death as core-collapse supernovae (SNe). Indirect evidence
54: for a large mass loss from the progenitor stars during their
55: late-time evolution is provided by the absence of H lines (and He
56: lines) in the spectra of SNe of Type Ib (Ic). In Type II SNe,
57: direct evidence for a wind emitted shortly before the explosion is
58: provided by the subsequent observations of emission by the circumburst
59: material of narrow optical lines (e.g. Salamanca et al. 1998;
60: 2002; Fassia et al. 2001), radio waves and X rays (for a recent
61: review, see Chevalier 2003). All these observations indicate that
62: the circumstellar density profile is $\rho_w \propto r^{-2}$, as
63: expected from a quasi-steady stellar wind.
64:
65: Gamma ray bursts (GRBs), in the cannonball (CB) model, are associated
66: with the death of massive stars in core collapse SN explosions (Dar
67: \& De R\'ujula 2000a,b; Dado, Dar \& De R\'ujula 2002a,b). Clear
68: indications for this association are the effect of the progenitor's
69: wind on their early-time afterglow (AG) and the presence of a
70: supernova in their late-time AG. In a CB-model analysis, there is
71: very convincing evidence of a GRB--SN association
72: (Dado et al. 20002a--e; 2003). The CB-model fingerprints of the massive
73: wind of the progenitor star have also been spotted in the AGs,
74: but never before with unquestionable certainty.
75:
76: In this paper we fit the broad-band optical afterglow of GRB 021211
77: in the CB model, and extract information about the circumburst
78: density profile, the associated SN and the host galaxy. We also
79: rediscuss GRB 990123 and GRB 021004. In these three cases the AG
80: was observed in the optical band early enough after the onset of
81: the burst for the CBs to be still moving through the progenitor's
82: wind. The observed early AG decline, roughly $t^{-1.6}$, is the
83: CB-model expectation\footnote{We have claimed before that this
84: behaviour ought to be $\sim t^{-2}$. In the case of GRB 990123,
85: discussed in Dado et al. 2002a, that was due to our then-incomplete
86: understanding of the AGs' broad-band spectra, which we improved in
87: Dado et al. 2002b. For GRB 021004 it was an error, which we
88: correct here, although it is not relevant in practice (the early
89: data on this AG are not sufficient to distinguish the two declines).}
90: for a wind density profile $\propto r^{-2}$. In the well measured early
91: AG of GRB 021211, the agreement between the observations and the
92: CB-model prediction is extremely satisfactory.
93:
94:
95: In the CB model we assumed GRBs to be associated with SNe similar
96: to SN1998bw (Dar and De R\'ujula, 2000a). In subsequent work (Dado
97: et al. 2002a--e; 2003) we found that, indeed, in all cases of GRBs of
98: known $z$ in which such a SN was in practice visible (all cases
99: with $z<1.12$), the CB-model fit to the corresponding AG disclosed
100: its presence. GRB 021211, at $z=1.006$, is one more example: the
101: fit clearly requires such a SN contribution, tentatively reported
102: by Fruchter et al. (2002) and Testa et al. (2003). The CB model
103: has so few parameters that we could determine via our fits the
104: unknown contribution of the host galaxy of GRB 021211 to the
105: different optical bands. The resulting broad-band spectrum snugly
106: resembles that of other GRB host galaxies (e.g. Gorosabel et
107: al. 2003) and star-forming galaxies at a similar $z$ (Fruchter et
108: al. 2002).
109:
110:
111: In the CB model ``dark'' GRBs without AGs are not expected. But
112: the AGs' decline is expected to be often fairly fast, in particular
113: if the interstellar medium (ISM)
114: density is relatively high, making their
115: detection difficult or, in the bygone days of slow GRB localization,
116: nearly impossible; see Fig. 6 of Dado et al. (2002a) and its
117: discussion. With faster detectors, more of these faint, fast-declining
118: ``quasi-dark'' GRBs ought to be found, and GRB 021211 is an example.
119: The detailed shape of the fast AG decline observed from 20 to 150
120: minutes, roughly as $1/t$, is well fit with an ISM
121: whose density, beyond the reaches of the wind, is relatively
122: high and approximately constant.
123:
124:
125:
126: \section{GRB 021211}
127: The gamma ray burst 021211 was detected on December 11.471227 UT,
128: 2002, with the HETE FREGATE, WXM, and SXC instruments. It was a
129: relatively long-duration ($\geq$ 5.7 s in the 8--40 keV band),
130: single-pulse GRB with a peak flux of $8\times 10^{-7}$ erg cm$^{-2}$
131: s$^{-1}$ and a fluence of $\sim 10^{-6}$ erg cm$^{-2}$ (Crew et
132: al. 2002). Its early-time optical AG was detected by the robotic
133: telescopes RAPTOR (Wozniak et al. 2002), KAIT (Li et al. 2003) and
134: S-LOTIS (Park et al. 2002), 90, 108 and 143 s after the burst,
135: respectively. A fit to the KAIT first 9 data points, in the interval
136: of 2.2--6.5 minutes after burst, showed a temporal decay with a
137: power-law index of $-1.60 \pm 0.02$ and a fit to later data, taken
138: 20 to 150 minutes after the GRB, gave a slope of $-0.96 \pm 0.04$
139: (Chornock et al. 2002), consistent with the decay seen by Price
140: and Fox (2002a,b) who were first to report the detection and the
141: precise localization of the optical AG. Follow-up observations in
142: the optical, NIR and radio bands were reported by Fox et al. (2003)
143: and by other groups in the GCN. The redshift of the host Galaxy,
144: $z=1.006\pm 0.002$, was measured with VLT by Vreeswijk et al. (2002)
145: and confirmed by Della Valle et al. (2003). A flattening of the
146: light curve due to the host galaxy and a possible contribution from
147: an associated SN were observed with HST 7 and 14 days after burst
148: in the BVIH bands (Fruchter et al. 2002) and with VLT-UT4 between
149: days 30 and 35 in the R band (Testa et al. 2003). Well-sampled
150: optical AG observations have been recently reported by Li et
151: al. (2003). The published observational data on the broad-band
152: optical AG of GRB 021211 are shown in Figs.~\ref{f1} and \ref{f2}.
153:
154:
155: \section{The CB model}
156:
157: In the CB model bipolar jets of CBs are launched axially in
158: core-collapse SNe, with initial Lorentz factors $\gamma_0={\cal{O}}(10^3)$.
159: The CBs are assumed to be produced in an unstable accretion process,
160: as in quasars and micro-quasars and, as in SS 433, to be made of
161: ordinary matter\footnote{Balmer H and He lines (e.g., Eikenberry,
162: et al. 2001) and the K$\alpha$ line of Fe (Migliari et al. 2002)
163: were detected in the mildly relativistic CBs of this $\mu$-quasar.}.
164: Crossing the SN shell and the progenitor's wind with a large
165: $\gamma$, the front surface of a CB is collisionally heated to keV
166: temperatures. The quasi-thermal radiation it emits, when no longer
167: absorbed by the intervening matter, and boosted and collimated by
168: its relativistic motion, is a single $\gamma$-ray pulse in a GRB.
169: The cadence of pulses reflects the chaotic accretion and is not
170: predictable, but the individual-pulse temporal and spectral properties
171: are (Dar and De R\'ujula, 2000b; for recent reviews, see De R\'ujula
172: 2002; Dar 2003). The ejected CBs, as observed in $\mu$-quasars,
173: are assumed to contain a tangled magnetic field. As they plough
174: through matter, they gather and magnetically scatter its constituent
175: protons. The re-emitted protons exert an inward pressure on the
176: CBs, which counters their expansion and makes them reach an asymptotic
177: radius $R_{_{CB}}$, in minutes of observer's time (Dado et al. 2002a).
178: The electrons swept in by the CB in its voyage through the wind
179: and ISM are Fermi-accelerated in the enclosed magnetic maze and
180: cooled by synchrotron radiation. Shortly after the GRB, the optical
181: radiation from the CB is dominated by synchrotron emission from
182: these electrons. So far, the CB model was very successful in
183: fitting the observed broad-band AGs of all GRBs of known redshift
184: (Dado et al. 2002a--e, 2003).
185:
186:
187:
188: \section{The fate of a cannonball}
189:
190: Let $n_p$ be the baryonic number density of the circumburst
191: material or the ISM, both dominated by protons.
192: A spherical CB of radius $R_{_{CB}}$ flying through this material
193: sees, in its rest system, an incoming flux of protons
194: entering it with a total momentum per unit time
195: $\Pi\simeq n_p\,m_p\,c^2\,\gamma^2\,\pi\,R_{_{CB}}^2$,
196: with $\gamma=\gamma(t)$ the diminishing CB's Lorentz factor.
197: If the CB's enclosed magnetic field randomizes these protons
198: so that they are re-emitted isotropically, the CB's surface is subject
199: to an inward pressure $P=\Pi/(4\,\pi\,R_{_{CB}}^2)$, which is independent
200: of $R_{_{CB}}$. In Dado et al. 2002a we have assumed that
201: this pressure is sustained by the equal and opposite pressure
202: $B^2/(8\,\pi)$ of the CB's enclosed magnetic field (which is thereby
203: determined) resulting in an equilibrium situation for a CB of
204: approximately constant $R_{_{CB}}$. We also assumed that $R_{_{CB}}$,
205: for a CB originally expanding at a transverse velocity
206: $\beta_0\,c$, can be estimated as the maximum expansion radius
207: attained as the pressure $P$ opposes the expansion, that is
208: $R_{_{CB}}\!\sim\! [3\,N_{_{CB}}\beta_0^2/(2\,\pi\,n_p\,\gamma_0^2)]^{1/3}$,
209: with $N_{_{CB}}$ the CB's baryon number.
210:
211: To test the above bold assumptions, we have fitted the model to
212: all broad-band AGs of GRBs of known $z$, with satisfactory results.
213: In Dado et al. (2002a) we have also tried a model with continuously
214: expanding CBs, and found it to be completely inadequate, supporting
215: the ansatz of CBs with constant radius. The estimated value of
216: $B\sim 10$ Gauss (for $n_p=10^{-2}$ cm$^{-3}$ and $\gamma_0=10^3$)
217: is adequate for the description of the data, but our original guess
218: $R_{_{CB}}={\cal{O}}(10^{14})$ cm turned out to be an overestimate
219: by at least one order of magnitude (Dado et al. 2002b). In this
220: paper we subject the CB's defining property ---that they reach an
221: approximately constant radius--- to a severe test, by studying in
222: detail its consequences for very early AGs.
223:
224: Let $\theta$ be the angle between the direction of a jet
225: of CBs and the observer.
226: The Doppler factor of the light the CB emits
227: %$\delta(t)=1/\gamma(t)\,(1-\beta(t)\, \cos\theta)$,
228: is well approximated by
229: $ \delta(t)\approx 2\,\gamma(t)/ (1+\theta^2\, \gamma(t)^2)$
230: in the domain of interest for GRBs: large $\gamma$ and small $\theta$
231: (typical values ensuing from our fits are $\gamma_0\sim 1/\theta\sim 10^3$).
232: The relation between time ---as measured by the observer--- and distance
233: ---as travelled by the CB--- is:
234: \begin{equation}
235: dx = {\gamma\,\delta\over(1 + z)}\;c\, dt .
236: \label{dxdt}
237: \end{equation}
238: The ambient protons that a CB scatters in the interval $dx$ slow it
239: down by an amount\footnote{We call $x$ and not $r$
240: the distance from the progenitor, not to insinuate a hypothesis
241: of spherical symmetry.}:
242: \begin{equation}
243: d\gamma=-{\pi\,R_{_{CB}}^2 n_p\gamma^2\over N_{_{CB}}}\,dx,
244: \label{dgamma}
245: \end{equation}
246: where $N_{_{CB}}$ is the CB's baryon number, for which our reference
247: value is $6\times 10^{50}$. The function $\gamma(t)$
248: can be explicitly found by quadrature in the two cases of interest here:
249: a density profile $n(x)=n_w\,(x_w/x)^2$ (for a ``windy''
250: neighbourhood) and a constant density (an adequate approximation
251: as the CBs get further away into the ISM).
252:
253: For an ISM of constant density, $\gamma(t)$ depends on $\theta$,
254: on the initial
255: $\gamma=\gamma_0$ as a CB exits the denser wind domain, and
256: on $x_\infty=N_{_{CB}}/(\pi\, n_p\, R_{_{CB}}^2)$, a deceleration
257: parameter. The value of $\gamma(t)$ is the real root of the cubic:
258: \begin{equation}
259: {1\over\gamma^3}-{1\over\gamma_0^3}
260: +3\,\theta^2\,\left[{1\over\gamma}-{1\over\gamma_0}\right]=
261: {6\,c\, t\over (1+z)\, x_\infty}\, .
262: \label{cubic}
263: \end{equation}
264: It takes a distance $x_{1/2}=x_\infty/\gamma_0$ for $\gamma(t)$
265: to descend to $\gamma_0/2$.
266: Our fitted values of $x_{1/2}$ are in the range 0.1 to 1 kpc, corresponding
267: to AGs that fade in the observed characteristic times of order
268: $(1+z)\,x_{1/2}/(c\, \delta_0)$ $\sim$ days. This range may be
269: affected by an observational bias: AGs with a smaller $x_\infty$ (e.g. those
270: with CBs moving in a relatively high density ISM) decay faster in time
271: and are ``quasi-dark'': harder to detect. GRB 021211 is one such case.
272:
273: We do not report the function $\gamma(t)$ for the small
274: values of $t$ corresponding to the CBs crossing the parent
275: stellar wind, since, for typical parameters, the
276: fractional energy loss $\Delta\gamma/\gamma$ is negligible
277: in that interval. The ``canonical'' stellar wind of a very
278: massive star has a rate $\dot M_w\!\sim\! 10^{-4}\,M_\odot$ y$^{-1}$
279: and a velocity $v_w\!\sim\!100$ km/s, so that
280: $\rho(x)\!\approx\!\rho_w\,(x_w/x)^2$ with
281: $\rho_w\,x_w^2\!\approx\! \dot M_w/(4\pi\,v_w)\!\sim\! 5.1\times 10^{13}$
282: g cm$^{-1}$,
283: or $n_w\,x_w^2\!\sim\! 3\times 10^{37}$ cm$^{-1}$.
284: The actual observations of $M_w$ and $v_w$
285: span an order of magnitude around
286: these canonical values.
287: It follows from Eq.~(\ref{dgamma}) that the
288: condition $\Delta\gamma/\gamma\!\ll\! 1$, for a wind profile extending
289: from $x_i$ to $x_f\!\gg\! x_i$, is
290: $R_{_{CB}}^2\!\ll\! N_{_{CB}}x_i/(\pi\,n_w\,x_w^2\,\gamma)$.
291: For observations starting at $\Delta t=100$ s after burst, $z=1$
292: and typical $\gamma=\delta=10^3$,
293: $x_i=c\gamma\delta\,\Delta t/(1+z)=0.5$ pc, and
294: the constraint is $R_{_{CB}}\!<\! 10^{14}$ cm, which we
295: have found to be amply satisfied (Dado et al. 2002b).
296:
297:
298: The fact that $\Delta\gamma/\gamma$ is typically small as the wind is
299: crossed has, as we shall see anon, an important consequence:
300: {\it the shape of the early optical AG locally and directly reflects the
301: shape of the circumburst density profile}.
302:
303: \section{The GRB afterglow in the CB model}
304:
305: The AG ---the persistent radiation in
306: the direction of an observed GRB--- has three origins: the ejected CBs, the
307: concomitant SN explosion, and the host galaxy (HG). These components are
308: usually unresolved in the observed ``GRB afterglows'', so that the
309: light curves and spectra are measures of
310: the corresponding cumulative energy flux density:
311: $
312: F_{_{AG}}=F_{_{CBs}}+F_{SN}+F_{_{HG}}.
313: $
314: %The contribution of the host galaxy depends on the
315: %angular aperture of the observations and it is usually determined
316: %by late time observations when the CB and SN contributions become
317: %negligible.
318: In all observed cases, but one (GRB 021004, discussed in Dado et al.
319: 2003), it is sufficient to approximate the ensemble of CBs in the AG
320: phase as a single (or dominant) CB.
321:
322: The contribution $F_{SN}$ of the SNe is approximated by that of
323: SN1998bw (Galama et al. 1998),
324: displaced to the GRB's redshift\footnote{The cosmological
325: parameters we use are: $\rm H_0=65$ km/(s Mpc), ${\rm \Omega_M}=0.3$
326: and ${\rm \Omega_\Lambda}=0.7$.}. In the CB model
327: the pair GRB 980425 \& SN1998bw is in no way
328: exceptional (but for the accidentally small $z$ and large $\theta$
329: of the GRB). Thus, the use of SN1998bw as a candidate GRB-associated
330: standard candle makes sense.
331:
332: %Let the energy flux
333: %density of SN1998bw at redshift $\rm z_{bw}=0.0085$ (Galama et al. 1998)
334: %be $\rm F_{bw}[\nu,t]$. For a similar SN placed at a redshift $\rm z$:
335: %\begin{equation}
336: %F_{SN}[\nu,t] = { 1+z \over 1+z_{bw}}\; {D_L^2(z_{bw})\over
337: %D_L^2(z)}\, A(\nu)\,F_{bw}[\nu',t']\, ,
338: %\label{bw}
339: %\end{equation}
340: %where $\rm D_L(z)$ is the
341: %luminosity distance, $\rm A(\nu,z)$ is the extinction along
342: %the line of sight, $\nu'=(1+z)\, \nu /(1+z_{bw})$ and
343: %$ t'= (1+z_{bw})\, t/(1+z)$.
344:
345:
346: The optical AG of a CB is dominated by synchrotron radiation from
347: the ISM electrons that penetrate in it. We argued in Dado et al.
348: (2002b) that these electrons are Fermi-accelerated in the CB-enclosed
349: magnetic maze and cooled by synchrotron radiation to a distribution
350: (in the CB's rest system) $dn_e/dE\propto E^{-2}$ below the {\it
351: injection bend} energy at which they enter the CB
352: ($E_b=\gamma(t)\,m_e\,c^2$); and to a distribution
353: $dn_e/dE\propto E^{-(p+1)}$ above
354: $E_b$, with $p=2$ in analytical approximations and $p=2.2$ in
355: numerical simulations. The emitted synchrotron radiation has a
356: corresponding double power-law form $\nu\,dn_\gamma/d\nu \propto
357: \nu^{-\alpha}$, with a power-law index $\alpha_l\approx 0.5$ well
358: below a {\it bend frequency} $\nu_b$, and $\alpha_h=p/2\approx 1.1$
359: well above it. Broad-band fits to the data on the AGs of all GRBs
360: of known $z$ result in $\alpha_l\approx 0.6\pm 0.1$, $\alpha_h\approx
361: 1.1\pm 0.1$. In practice $\alpha_l$ is less well determined than
362: $\alpha_h$, the CB-model fits being good even for a fixed $\alpha_h=
363: 1.1$. We attribute this to the fact that the extracted $\alpha_h$
364: is sensitive to the AG light-curve at relatively late times, when
365: CBs are typically hundreds of parsecs away from the progenitor and
366: the absorption in the host, which is hard to ascertain, may be
367: minimal.
368:
369: The value of $E_b$ and the estimate of the CB's magnetic field
370: described in the previous section allow us to estimate
371: the bend frequency of the synchrotron-radiation spectrum:
372: \begin{equation}
373: \nu_b(t) \simeq {1.87\times 10^3\, [\gamma(t)]^3\,\delta(t)\over 1+z}\,
374: \left[{n_p(x)\over 10^{-3}\;{\rm cm}^{-3}}\right]^{1/2}\, \rm Hz.
375: \label{nubend}
376: \end{equation}
377: In the observer frame, this radiation is Doppler-boosted and
378: collimated by the relativistic motion of the CB, and redshifted
379: by the cosmic expansion. Its explicit form, extending from
380: radio to X-ray frequencies, is given in Eqs.~(4,6,8)
381: of Dado et al. 2002b; here we only report its behaviour near the
382: optical domain, which is simple above and below the injection bend.
383: Let $\alpha=(\alpha_l,\,\alpha_h)=(0.5,1.1)$ be the predicted power
384: indices at $\nu\!\ll\! \nu_b$ and $\nu\!\gg\! \nu_b$, respectively. The
385: time and frequency dependence of the energy fluence in these
386: limits is:
387: \begin{equation}
388: F_\nu \propto { A'(\nu,t)\;n_e^{(1+\alpha)/2}\,
389: [\gamma(t)]^{3\alpha-1}\, [\delta(t)]^{3+\alpha}} \,
390: \nu^{-\alpha}\, ,
391: \label{sync}
392: \end{equation}
393: where $n_e=n_e(x)\simeq n_p(x)$ is the electron density along the
394: CBs' trajectory and $A'(\nu,t)$ corrects for absorption in the
395: host galaxy and in ours, its possible time dependence originating
396: in the kiloparsec length of the CBs' trajectory in the host galaxy.
397: Self-absorption in the CB is irrelevant
398: at the optical and NIR wavelengths relevant here.
399: In the CB model the early AG is dominated by the last significant
400: pulse (or CB) of the GRB, so that $t$ in Eq.~(\ref{sync}) is the time
401: after that pulse. The nuance may be significant at very small $t$.
402:
403: In deriving Eq.~(\ref{sync}) we have assumed that a
404: fixed fraction of the energy-deposition rate by ISM electrons in a CB
405: ($\pi\,R_{_{CB}}^2\,n_e\,m_e\,c^3\,\gamma^2$, in the CB's
406: system) is re-emitted as the AG. The AG spectrum is a
407: function of $\nu$ and $\nu_b$, so that at fixed $\nu$ it depends on
408: $n_e=n_p$ via $\nu_b$, as in Eq.~(\ref{nubend}). Thus a result which is
409: not linear in $n_e$, and peculiar at first sight.
410:
411: In our fits we assume as the circumburst density profile
412: a constant plus a ``windy'' term:
413: \begin{equation}
414: n(x)=n_0+n_w(x_w/x)^2\equiv n_0\,[1+(\bar x/x)^2].
415: \label{density}
416: \end{equation}
417:
418:
419: \section{The very early optical AG}
420:
421:
422: For typical parameters,
423: the first few parsecs of a CB's voyage ---as it crosses the ``wind''
424: material emitted by the progenitor star--- are seen by an observer
425: in the first few minutes of the AG. During that ``early'' time, the
426: ambient density is high enough for $\nu_b$, as in Eq.~(\ref{nubend}),
427: to be comfortably above the optical frequencies, so that
428: $\alpha=\alpha_l=0.5$ in Eq.~(\ref{sync}).
429: We have seen that, at ``early'' times, a CB
430: has its Lorentz factor insignificantly changed
431: by collisions with the wind material, so that the observer's time and
432: the CB's travelled distance, as in Eq.~(\ref{dxdt}),
433: are strictly proportional. Thus Eqs.~(\ref{sync},\ref{density}) collapse to:
434: \begin{equation}
435: F_\nu\simeq n_e^{(1+\alpha)/2}\propto
436: \left[1+{\left(\bar t/ t\right)^2}\right]^{3/4}.
437: \label{early}
438: \end{equation}
439: Had we used the observed $\alpha_l=0.6\pm 0.1$, as opposed to the
440: naive theoretical $\alpha_l=0.5$, we would have obtained
441: $F_\nu\simeq t^{-1.6}$ at veary early times, in close agreement
442: with the fit by Chornock et al. (2002) for GRB 021211,
443: and the observations of Akerlof et al. (1999) for
444: GRB 990123 and Fox et al. (2002) for GRB 021004. But we shall
445: see that Eq.~(\ref{early}) provides an excellent fit to the data.
446: The observed deviations relative to the smooth curve
447: of Eq.~(\ref{early}) reflect the variations of the wind
448: that generated the profile. In the case of GRB 021211,
449: these variations are at the $\sim 10$\% level (Li et al. 2003).
450:
451:
452: \section{The overall AG of GRB 021211 in the CB model}
453:
454: The copious early data on GRB 021211 obtained by Li et al. (2003),
455: along with all other data communicated in the GCN notices
456: quoted in section 2 and in
457: Fox et al. 2003 are shown in Fig.~\ref{f1} for the R-band,
458: and in Fig.~\ref{f2} for the broad-band optical (IRVB) and NIR
459: (HKJ) passbands. In fitting these results we neglect the
460: unknown extinction in the host, due to the lack of relevant spectral
461: information\footnote{An unabsorbed SN1998bw-like contribution fits the
462: late-time broad-band data, indicating that host extinction is not large.}.
463: For this GRB the R-band data are relatively so copious that it is useful
464: to perform separate R-band and broad-band fits.
465:
466: In our fit, the density is that of Eq.~(\ref{density}) and the
467: limiting exponents in the synchrotron spectrum
468: $\propto\nu^{-\alpha}$ are set to their theoretical values
469: $(\alpha_l,\alpha_h)=(0.5,1.1)$. The fit is made with
470: the complete analytical broad-band predictions Eqs.~(4,6,8) of Dado et
471: al. 2002b, of which Eq.~(\ref{sync}) are the limiting behaviours and
472: Eq.~(\ref{cubic}) is the deceleration law.
473: The best fitted parameters for the R-band fit shown in Fig.~\ref{f1}
474: are $\gamma_0=262$, $\theta= 1.76$ mrad (i.e.
475: $\delta_0=431$) and $x_{\infty}=4.7$ kpc, with 51 d.o.f.~and
476: $\chi^2/{\rm (d.o.f.)}=0.97$ if the out-lying point at $t\!\sim\! 2$ days is
477: eliminated. The constant contribution
478: of the host galaxy was left as a free parameter, for which the
479: fit returned $F_{_{HG}}=0.34\, \mu$Jy.
480: The contribution of a standard-candle SN1998bw
481: at the GRB position is discernible in Fig.~\ref{f1}, in which $F_{_{HG}}$
482: has been subtracted.
483:
484: We have also tried a
485: fit with an arbitrary power $a$ in the windy term: $n_w(x_w/x)^a$,
486: resulting in a best fitted $a=1.92\pm 0.03$, close to the
487: ``canonical'' expectation $a=2$. Thus, the agreement of the fits
488: with the canonical ``windy'' power is not a consequence of the
489: scarcity of data. Also, at early times, the fit is totally
490: insensitive to the other parameters ($\gamma_0$, $\theta$ and $x_\infty$).
491: Thus, the result for $a$ is robust: the signature of a wind is there.
492:
493: The overall normalization of the AG is also a free parameter in our
494: fits; it is proportional to the ambient density $n(x)$, to the CBs' cross
495: sections, to their number, and to the unknown fraction of the
496: energy of the gathered ISM electrons that is re-emitted as the AG.
497: Thus, extracting a value of $n(x)$ from the fit normalization requires
498: extra assumptions. Yet, the value of $n(x)$ at some reference distance
499: (e.g. $n_0$, its limiting value at large distance)
500: can be extracted from the fact that we use it as a free
501: parameter in the expression of Eq.~(\ref{nubend}) for $\nu_b$.
502: The fitted result is $n_0=(2.97\pm 0.22)$ cm$^{-3}$.
503: Combined with the fitted value
504: $\bar x=1.2$ pc in Eq.(\ref{density}), this
505: yields $\rho_w\,x_w^2=m_p\,n_w\,x_w^2=
506: (6.8\pm 0.5)\times 10^{13}$ g cm$^{-1}$,
507: compatible with the expectation, quoted in section 4, for a
508: ``canonical wind'': $5.1\times 10^{13}$ g cm$^{-1}$.
509:
510: The wide-band fit is shown in Fig.~\ref{f2}; it has essentially
511: the same fit parameters as the R-band fit. The contributions
512: $F_{_{HG}}$ of the host galaxy in the different bands, resulting
513: from the fit, are 0.50, 0.34, 0.23 and 0.23 $\mu$Jy for the IRV
514: and B bands, respectively. This extracted spectrum, and in particular
515: its flattening above the V frequencies, is compatible with that of
516: other GRB hosts (e.g. Gorosabel et al. 2003) and star-forming
517: galaxies at $z\!\sim\! 1$ (e.g. Fruchter et al. 2002).
518:
519: The comparison we have described between the CB-model expectations
520: and the data for GRB 021211 is simply spectacular. The fits are
521: good, particularly that to the shape of the R-band light curves.
522: The agreement between the expected and extracted ambient density
523: profiles is eery. A word of caution, however, is necessary. The
524: number of parameters in our fits is high: the ``ambient'' parameters
525: describing the host galaxy ($F_{_{HG}}$) and the density profile
526: ($n_0$ and $\bar x$), the observer's viewing angle $\theta$, and
527: the 3 quantities that are specific to the CBs: the overall AG
528: normalization, $\gamma_0$ and $x_\infty$. Even if the fitted ambient
529: parameters turn out to have the expected values, a model with this
530: many parameters has to be very wrong not to describe a very simple
531: surface: the fluence as a function of frequency and time\footnote{It
532: is possible to construct models that disagree with observation or
533: are unphysical, but have even more parameters. For the case of
534: GRB 991208, see section 2 of Dado et al. 2002e.}. Moreover, the
535: data are not cross-calibrated, their systematic errors are unknown,
536: the model is no doubt a simplification of a very complicated
537: phenomenon, and the parameters are not uncorrelated (except at the
538: early times of particular interest here). As a consequence of all
539: this, the ``formal errors'' of the parameters (as given by the
540: standard program MINUIT) are no doubt underestimated. Whence our
541: habit of not always making the errors explicit.
542:
543: \section{The radio AG of GRB 021211}
544:
545: In the CB model the radio AG at early time is suppressed because of limb
546: darkening and the finite time it takes the electrons gathered by a CB to
547: cool radiatively and reach an energy at which they emit synchrotron
548: radiation of a given typical frequency (Dado et al. 2002b). At later times
549: the radio emission, like the optical and X-ray emission, becomes
550: proportional to a high power of the Lorentz factor.
551: In the case of the radio AG of GRB 021211, only upper limits (or
552: perhaps marginal detections) have been reported, by Hoge et al. (2002)
553: and Fox et al. (2002). The later authors, who analyze the AG in detail,
554: consider the absence of a stronger radio signal to be a problem.
555: In the CB model, it is not.
556:
557: The wide-band spectrum of AGs is, in the CB model, very simple.
558: Its only feature ---besides the bend frequency that we have discussed---
559: is due to self-absorption in the CB, and it is described by a single
560: parameter $\nu_a$ (Dado et al. 2002b). In the case of GRB 021211
561: we cannot fix this parameter, having no single secure radio signal.
562: The best we can do is to show, by way of example, what the radio
563: signal would be if the value of $\nu_a$ was similar to the average
564: for other radio AGs ($\bar \nu_a\!\sim\! 1$ GHz) or to the value for
565: GRB 021004 (0.98 GHz), which is otherwise
566: akin to GRB 021211, and is very well described by the CB model
567: (Dado et al. 2003). The result of this exercise is shown in
568: Fig.~\ref{f4}, and it is compatible with the observational limits or
569: marginal detections: the largest ``signals''
570: reported in Fox et al. 2002 are $45\pm 23$ and $60\pm 38$ $\mu$Jy,
571: at $t\!\sim\! 5,\,10$ days, both at 8.46 GHz, while Hoge et al. (2002)
572: obtain a 3$\sigma$ upper limit of 7.5 mJy at 347 GHz, on day $t\!\sim\! 1$.
573:
574:
575: \section{The GRB proper}
576:
577: In the CB model, the fluence $F$
578: from a GRB viewed at a small $\theta$, is amplified by
579: a huge factor $\delta_0^3$, due to Doppler boosting and
580: relativistic collimation (Dar \& De R\'ujula 2000):
581: \begin{equation}
582: F={(1+z) \,\delta_0^3\over 4\,\pi\, D_L^2}\, E_\gamma\, ,
583: \label{fluence}
584: \end{equation}
585: where $E_\gamma$ is the total energy in photons emitted by CBs in their
586: rest system. The total ``equivalent spherical'', or would-be
587: isotropic energy, $E^{iso}$, inferred from the observed fluence,
588: is a factor $(\delta_0)^3$ larger than $E_\gamma$.
589: In Dado et al. 2002a we deduced that the $E_\gamma$ values
590: of the GRBs of known $z$ span the surprisingly
591: narrow\footnote{GRBs in the CB model are much better standard candles
592: than in the standard model (Frail et al. 2001).} interval
593: $10^{44\pm 0.3}$ erg, the spread in $E^{iso}$
594: being mainly due to the spread in their values of $\delta_0$
595: (deduced from the fits to their AGs). For GRB 021211,
596: the CB-model expectation is $E^{iso} \approx \delta_0^3\,
597: E_\gamma\approx 0.8\times 10^{52\pm 0.3}$ erg,
598: in agreement with the observed $E^{iso}\approx 1\times 10^{52}$
599: erg, deduced from its measured redshift ($z=1.006$;
600: Vreeswijk et al. 2002) and fluence
601: in the 8--40 keV band ($\sim 10^{-6}$ erg cm$^{-2}$; Crew et
602: al. 2002).
603:
604:
605:
606: \section {GRBs 990123 and 021004}
607:
608: The AG of these two GRBs was also observed particularly early, by
609: Akerlof et al. (1999) in the case of GRB 990123. We have recalibrated
610: their data assuming $V-R=0.28$ mag (our predicted $\alpha_l=0.5$),
611: rather than a constant colour. We have also recalibrated the
612: later-time data compiled in Castro-Tirado et al. 1999; Kulkarni et
613: al. 1999; Galama et al. 1999 and Fruchter et al. 1999, assuming
614: $R-r=0.015$ mag ($\alpha_l=0.5$). The R-band results of our broad-band
615: fit are shown in Fig.~\ref{f4}, in which $t=0$ corresponds to the
616: start of the last (second) prominent
617: pulse in the GRB, 37s after trigger (Briggs et al. 1999). There is an
618: excellent agreement between the early
619: data tracing the circumburst windy density and
620: the prediction of Eq.~(\ref{density}). A fit with
621: an arbitrary power $a$ in the windy term, $n_w(x_w/x)^a$,
622: results in a best-fitted $a=1.98\pm 0.06$, again in agreement
623: with the ``canonical'' expectation: $a=2$.
624:
625: Though the data have been recalibrated, the parameters we fit to
626: the AG of GRB
627: 990123 are close to those in Dado et al. 2002a,b. In particular
628: $\gamma_0=1204$ and $\delta_0=1630$ ---determined from the {\it
629: shape} of the AG--- are the highest for any GRB of known $z$,
630: implying that the {\it magnitude} of the GRB and AG fluences should
631: also be record-breaking, as they are (even for this GRB, the value of
632: $E_\gamma$ in Eq.~(\ref{fluence}) is not exceptional). Since
633: $\gamma$ and $\delta$ are so large, the bend frequency $\nu_b$ of
634: Eq.~(\ref{nubend}) ``crosses'' the optical band very late in time,
635: where the data are quite imprecise. As a consequence, the R-band
636: and wide-band fits are very insensitive to $\nu_b$, and we cannot
637: reliably use its value to determine the absolute magnitude of the
638: ambient and wind densities. If we assume that the density
639: profile $n_w\propto x_w^{-2}$ prevails until $n_w$
640: declines to the typical
641: superbubble or halo value, $n_p\sim 10^{-3}$, we find that
642: $\rho_w\, x_w^2=m_p\, n_p\, x_w^2\approx 4.2\times 10^{13}$ g
643: cm$^{-1}$, again in agreement with the
644: expectation, quoted in section 4, for a
645: ``canonical wind'': $5.1\times 10^{13}$ g cm$^{-1}$.
646: The windy profile would in this case
647: extend all the way to $x_w\sim 50$ pc, yielding a
648: total mass loss of $\sim 40\, M_\odot$ during the life of the very
649: massive progenitor star. Such a large total mass loss is
650: not unusual for the progenitors of SNe of Type Ib and Ic.
651:
652:
653:
654: The CB-model fit to the R-band AG of GRB 021004 is shown in Fig.~\ref{f5}.
655: The data are those measured and compiled by Fox et al. 2002; Holland et
656: al. 2002; Pandey et al. 2002 and Bersier et al. 2003. Once more, the
657: agreement between the prediction of Eq.~(\ref{density}) and the early data
658: is good, though these data are neither as abundant nor as close to the GRB
659: trigger as for the other two cases we discussed, implying that we cannot
660: claim to have clearly spotted, for this GRB, a typical windy profile.
661: The fitted result for the asymptotic density is $n_0=(4.08 \pm 0.17)
662: \times 10^{-3}$ cm$^{-3}$. Combined with the fitted value $\bar x\sim 5.1$
663: pc in Eq.~(\ref{density}), this yields $\rho_w\,x_w^2=m_p\,n_w\,x_w^2\sim
664: 1.7\times 10^{12}$ g cm$^{-1}$, a bit over an order of magnitude below
665: the central ``canonical'' expectation.
666:
667: \section{Discussion and conclusions}
668:
669: It is, as usual, interesting to compare the CB model with the standard
670: fireball or blast-wave models of GRBs. For GRB 021211 a detailed SM
671: analysis is given in Fox et al. 2003. The relevant issues at hand concern
672: the signatures for ``windy'' neighbourhoods, and the predictive power for
673: early AGs\footnote{For Fox et al. (2003) the non-observation of a radio AG
674: is also significant, suggesting that ``the burst may have suffered
675: substantial radiative corrections''. No such specific cure is necessary in
676: the CB model.}.
677:
678: In the standard model (SM) the scarcity of AGs indicating a windy
679: circumburst density distribution is a problem, admitted even by its
680: staunchest defenders (Piran 2001; Price et al. 2002). In the CB model all
681: the observed AGs that should show a windy signature do. Indeed, the
682: relation Eq.~(\ref{dxdt}) between observer's time and CBs travelled
683: distance implies that, for the typical $\gamma_0\sim \delta_0\sim 10^3$,
684: the AG of a GRB at a cosmological distance must be ``caught'' within the
685: first few minutes for the CBs to be still travelling close enough to the
686: parent star, so that the contribution of its wind to the ambient density
687: is still significant and observable. Only the three GRBs that we have
688: discussed pertain to this category.
689:
690: In the SM the fast-decreasing ``early'' optical AG, as first observed in
691: GRB 990123, is generally attributed to a {\it reverse} shock (M\'esz\'aros
692: and Rees 1999). For GRB 021211, this is the interpretation spoused by the
693: observers (Fox et al. 2003; Li et al. 2003). In this view the index of the
694: approximate power law of the decay of the early AG is not fixed, and its
695: normalization is unrelated to that of the late AG. In the CB model,
696: contrariwise, the early optical AG does not have its own separate origin:
697: it is made by the very same mechanism as the late AG. The temporal shape
698: of the AG is not a succession of power laws with breaks between different
699: mechanisms or regimes. This temporal shape is not arbitrary: at early
700: times it is directly related to the shape of the circumburst density
701: profile, as in Eq.~(\ref{early}), and this relation is fully vindicated by
702: the data for the expected windy profile. The AGs shown in our figures
703: depend, in the ``early'' domain, on only one parameter: the absolute
704: normalization. The ambient density that defines the early temporal shape
705: can also be extracted from the CB-model fit, via the bend frequency of
706: Eq.~(\ref{nubend}), and its magnitude turns out to coincide with the
707: expected density for the canonical wind of a heavy star.
708:
709: Fast-declining optical light curves ---masquerading as ``dark'' AGs--- are
710: expected in the CB model (Dado et al. 2002a), and GRB 021211 is an
711: example. The CB-model fit to its AG has allowed us to infer the presence
712: of an associated SN akin to SN1998bw, as for all other GRBs where the SN
713: could in practice be detected. The fit also yields the ``colour'' of the
714: host galaxy, which is the expected one.
715:
716: We have seen how, concerning the early optical AGs, the CB model
717: is very predictive and successful. In previous works we have shown
718: this to be the case also for all properties of the AGs of all GRBs
719: of known redshift: they can {\bf all} be described in a simple,
720: analytical and parameter-thrifty manner, which does not involve
721: multiple choices, multiple mechanisms and multiple exceptions. To
722: put it in a way with which even the most allegiant apologists of
723: the SM might agree: the contrast between the CB model and the
724: standard model is striking.
725:
726:
727:
728: {\bf Acknowledgement:} This research was supported in part by the Helen
729: Asher Space Research Fund and by the VPR fund for research at the
730: Technion. One of us, Arnon Dar, is grateful for hospitality
731: at the CERN Theory Division.
732:
733: \begin{thebibliography}{}
734:
735: \bibitem[]{Ak}
736: Akerlof, C., et al. 1999, Nature, 398, 400
737: \bibitem[]{1}
738: Bersier, D., et al. 2003, ApJ, 584, L43
739: \bibitem[]{12}
740: Briggs, M. S., et al. 1999, ApJ, 524, 82
741: \bibitem[]{13}
742: Castro--Tirado, A., et al. 1999, Science, 283, 2069
743: \bibitem[]{349}
744: Chevalier, R. A. 2003, astro-ph/0301368
745: \bibitem[]{3491}
746: Chornock, R., Li, W., Filippenko, A. V. \& Jha, S. 2002, GCN 1754
747: \bibitem[]{3492}
748: Crew, G., et al. 2002, GCN 1734
749: \bibitem[]{350}
750: Dado, S., Dar, A. \& De R\'ujula, A. 2002a, A\&A 388, 1079
751: \bibitem[]{352}
752: Dado, S., Dar, A. \& De R\'ujula, A. 2002b, A\&A in press (astro-ph/0204474)
753: \bibitem[]{354}
754: Dado, S., Dar, A. \& De R\'ujula, A. 2002c, ApJ, 572, L143
755: \bibitem[]{356}
756: Dado, S., Dar, A. \& De R\'ujula, A. 2002d, A\&A, 393, L25
757: \bibitem[]{3582}
758: Dado, S., Dar, A. \& De R\'ujula, A. 2002e, astro-ph/0211596
759: \bibitem[]{3583}
760: Dado, S., Dar, A. \& De R\'ujula, A. 2003, ApJ, 585, L15D
761: \bibitem[]{18085}
762: Dar, A. 2003, astro-ph/0301389
763: \bibitem[]{360}
764: Dar, A. \& De R\'ujula, A. 2000a, astro-ph/0008474
765: \bibitem[]{1808}
766: Dar, A. \& De R\'ujula, A. 2000b, astro-ph/0012227
767: \bibitem[]{1809}
768: Della Valle, M., et al. 2003, GCN 1809
769: \bibitem[]{2000}
770: De R\'ujula, A. 2002, astro-ph/0207033
771: \bibitem[]{3601}
772: Eikenberry, S. S., Cameron, P. B., Fierce, B. W.,
773: Kull, D. M., Dror, D. H., Houck, J. R. \& Margon, B. 2001, ApJ, 561, 1027
774: \bibitem[]{3660}
775: Fassia, A., et al. 2001, MNRAS, 325, 907
776: \bibitem[]{3661}
777: Fox, D. W., et al. 2002, GCN 1564
778: \bibitem[]{3662}
779: Fox, D. W.,, et al. 2003, astro-ph/0301377
780: \bibitem[]{3663}
781: Frail, D. A., et al. 2001, ApJ, 562, L55
782: \bibitem[]{1313}
783: Fruchter, A. S., et al. 1999, ApJ, 519, L13
784: \bibitem[]{1791}
785: Fruchter, A., et al. 2002, GCN 1781
786: \bibitem[]{368}
787: Galama T.J., et al., 1998, Nature, 395, 670
788: \bibitem[]{369}
789: Galama, T. J., et al. 1999, Nature, 398, 394
790: \bibitem[]{370}
791: Gorosabel, J., et al. 2003, astro-ph/0301564
792: \bibitem[]{375}
793: Hoge, J., et al. 2002, GCN 1742
794: \bibitem[]{1111}
795: Holland, S. T. 2002, astro-ph/0211094
796: \bibitem[]{378}
797: Kulkarni, S. R., et al. 1999, Nature, 398, 389
798: \bibitem[]{382}
799: Li, W., Filippenko, A.V., Chornock, R. \& Jha, S. 2002, GCN 1737
800: \bibitem[]{38205}
801: Li, W., Filippenko, A.V., Chornock, R. \& Jha, S. 2003, astro-ph/0302136
802: \bibitem[]{38208}
803: M\'esz\'aros P. \& Rees, M. J. 1999, MNRAS, 306, L39
804: \bibitem[]{3821}
805: Migliari, S., Fender, R. \& Mendez, M. 2002, Science, 297, 1673
806: \bibitem[]{333}
807: Pandey, S.B., et al. 2002, astro-ph/0211108
808: \bibitem[]{1783}
809: Park, H. S., et al. 2002, GCN 1736
810: \bibitem[]{17834}
811: Piran, T. 2001, astro-ph/0104134
812: \bibitem[]{17835}
813: Price, P. A., et al. 2002, astro-ph/0203467
814: \bibitem[]{1782}
815: Price, P. A. \& Fox, D. W. 2002a, GCN 1731
816: \bibitem[]{1784}
817: Price, P. A. \& Fox, D. W. 2002b, GCN 1731
818: \bibitem{1792}
819: Schlegel, D. J., Finkbeiner, D. P. \& Davis, M. 1998, ApJ, 500, 525
820: \bibitem[]{1793}
821: Salamanca, I., Cid-Fernandes, R., Tenorio-Tagle, G., Telles, E.,
822: Terlevich, R. J. \& Munoz-Tunon, C. 1999, MNRAS, 300, L17
823: \bibitem[]{1785}
824: Salamanca, I., Terlevich, R. J. \& Tenorio-Tagle, G. 2002, MNRAS,
825: 330, 844
826: \bibitem[]{1786}
827: Testa, V., et al. 2003, GCN 1821
828: \bibitem[]{1787}
829: Vreeswijk, P., et al. 2002, GCN 1785
830: \bibitem[]{1788}
831: Wozniak, P., et al. 2002, GCN 1757
832:
833:
834:
835: \end{thebibliography}{}
836:
837: \clearpage
838:
839: \begin{figure}[t]
840: %\plotone{ag346nro021211.eps}
841: \plotone{f1.eps}
842: \figcaption{The optical
843: observations of the R-band
844: AG of GRB 021211, and their CB-model fit. The ISM density
845: is a constant plus a ``wind'' contribution decreasing as the inverse square
846: of the distance. The two contributions are equal at $\bar x\simeq 1.2$ pc, a
847: distance reached by the CBs in an observer's time $\bar t\simeq 0.025$ days
848: after burst. The data are those
849: reported to date, in the GCN notices quoted in section 2, in Fox et al.
850: 2003 and in Li et al. 2003. The
851: contribution of a SN1998bw-like SN at the GRB position
852: is discernible at late times.
853: The host galaxy's contribution, which was fitted, is subtracted in this plot.
854: \label{f1}}
855: \end{figure}
856:
857: \begin{figure}[t]
858: %\plotone{ag448nro021211.eps}
859: \plotone{f2.eps}
860: \figcaption{The wide-band CB-model fit to the optical and NIR observations
861: for GRB 021211.
862: For clarity, only the better-sampled I, R, V and B bands are shown,
863: scaled by 10, 1, 1/10 and 1/100, respectively. The
864: contribution of a SN1998bw-like SN at the GRB position
865: is discernible at late times.
866: The host galaxy's contributions, which were fitted, are subtracted in
867: this plot.
868: \label{f2}}
869: \end{figure}
870:
871:
872: \begin{figure}[t]
873: %\plotone{ag502nro021211.eps}
874: \plotone{f3.eps}
875: \figcaption{The optical R-band
876: AG of GRB 021211, superimposed on the predictions for the radio
877: AG at the various labelled frequencies, for an assumed $\nu_a=1$ GHz.
878: \label{f3}}
879: \end{figure}
880:
881: \begin{figure}[t]
882: %\plotone{ag001nro990123.eps}
883: \plotone{f4.eps}
884: \figcaption{The optical
885: observations of the R-band
886: AG of GRB 990123, and their CB-model fit. The ISM density
887: is a constant plus a ``wind'' contribution decreasing as the inverse
888: square of the distance. The contribution of the host galaxy has been
889: subtracted. The small extinction in the Galaxy (Schlegel et al. 1998)
890: was neglected.
891: \label{f4}}
892: \end{figure}
893:
894:
895: \begin{figure}[t]
896: %\plotone{ag607nro1004.eps}
897: \plotone{f5.eps}
898: \figcaption{The optical observations of the R-band AG of GRB 021004,
899: and their CB-model fit with two CBs, whose individual contributions
900: are depicted along with the total (Dado et al. 2002e). The ISM
901: density is a constant plus a ``wind'' contribution decreasing as
902: the inverse square of the distance. These two contributions are
903: equal at $\bar x\simeq 3$ pc, a distance reached by the CBs in an
904: observer's time $\bar t\simeq 0.01$ day after burst. The data are
905: the same as the ones quoted in Dado et al. 2002e. The contribution
906: of a SN1998bw-like SN at the GRB position and the host galaxy's
907: contribution were subtracted in this plot.
908: \label{f5}}
909: \end{figure}
910:
911:
912:
913:
914: \end{document}
915:
916:
917:
918:
919:
920:
921:
922:
923:
924:
925: