1: \documentclass [12pt,preprint]{aastex}
2: % for a referee version
3: %\documentclass [manuscript]{aastex}
4: %documentclass [article]{aa}
5: \usepackage{epsfig}
6: %
7: \begin{document}
8: \voffset-1cm
9: \newcommand{\gsim}{\hbox{\rlap{$^>$}$_\sim$}}
10: \newcommand{\lsim}{\hbox{\rlap{$^<$}$_\sim$}}
11:
12:
13: \title{On the origin of the correlations between \\
14: Gamma-Ray Burst observables}
15:
16:
17: \author{Shlomo Dado\altaffilmark{1}, Arnon Dar\altaffilmark{1}
18: and A. De
19: R\'ujula\altaffilmark{2}}
20:
21: \altaffiltext{1}{dado@phep3.technion.ac.il, arnon@physics.technion.ac.il,
22: dar@cern.ch.\\
23: Physics Department and Space Research Institute, Technion, Haifa 32000,
24: Israel}
25: \altaffiltext{2}{alvaro.derujula@cern.ch; Theory Unit, CERN,
26: 1211 Geneva 23, Switzerland \\
27: Physics Department, Boston University, USA}
28:
29:
30:
31: \begin{abstract}
32: Several pairs of observable properties of Gamma Ray Bursts (GRBs)
33: are known to be correlated.
34: Many such correlations are straightforward predictions of the
35: `cannonball' model of GRBs.
36: We extend our previous discussions of the subject to a wealth of new data,
37: and to correlations between `lag-time', `variability' and `minimum rise-time',
38: with other observables. Schaefer's recent systematic analysis of the
39: observations of many GRBs of known red-shift gives us a good and
40: updated data-basis for our study.
41: \end{abstract}
42:
43: \keywords{Gamma Ray Burst}
44:
45: \section{Introduction}
46:
47: Quite a few independent observable quantities can be measured
48: in a long-duration gamma-ray burst (GRB). These include its
49: spherical equivalent energy,
50: its peak isotropic luminosity,
51: the `peak energy' of its spectrum
52: and the red-shift of its host galaxy.
53: In order of increasing effort on the data analysis, one can also define
54: and determine the number of pulses in the GRB's light-curve and
55: their widths, rise-times, and `lag-times'.
56: Finally, with considerable toil and embarrassment of choices,
57: one can define and measure `variability'
58: (Fenimore \& Ramirez-Ruiz, 2000; Plaga, 2001;
59: Reichart et al.~2001;
60: Guidorzi et al.~2006; Schaefer 2006).
61: Pairs of the above quantities are known to be correlated
62: as approximate power laws, sometimes fairly
63: tightly and over spans of several orders of magnitude.
64: A model of GRBs ought to be able to predict these power laws
65: and to pin-point the choices of red-shift corrections that should
66: make them tightest.
67:
68:
69: In the `cannonball' (CB) model of GRBs the correlations
70: between the cited observables, as we shall discuss, are
71: predictable. They are based on very simple physics:
72: the production of $\gamma$ rays by inverse Compton scattering
73: of much softer photons by a relativistically-moving,
74: quasi-point-like object (Shaviv \& Dar 1995, Dar \& De R\'ujula 2004).
75: The correlations very satisfactorily test
76: these individual theoretical ingredients, or combinations thereof.
77: A reader desiring to start by evaluating these claims may
78: choose to read Section \ref{basis} first.
79:
80: The data, particularly on
81: GRBs of known red-shift, has become much more extensive
82: in the time elapsed since the CB-model correlations were
83: predicted (Dar \& De R\'ujula 2001) and several of them
84: were tested (Dar \& De R\'ujula 2004). It is time
85: to restudy the subject, which we do for many correlations,
86: relying mainly on the data analysis by Schaefer (2006).
87:
88:
89: \section{The CB model}
90:
91: In the CB model (Dar \& De R\'ujula~2000, 2004; Dado et al.~2002, 2003),
92: {\it long-duration} GRBs
93: and their AGs are produced by bipolar jets of CBs, ejected in
94: core-collapse SN explosions (Dar \& Plaga~1999). An
95: accretion disk is hypothesized to be produced around the newly
96: formed compact object, either by stellar material originally close to the
97: surface of the imploding core and left behind by the explosion-generating
98: outgoing shock, or by more distant stellar matter falling back after its
99: passage (De R\'ujula~1987). As observed in microquasars, each
100: time part of the disk falls abruptly onto the compact object, a
101: pair of CBs made of {\it ordinary plasma} are emitted with high
102: bulk-motion Lorentz factors, $\gamma$, in opposite directions along the
103: rotation axis, wherefrom matter has already fallen onto the compact
104: object, due to lack of rotational support.
105: The $\gamma$-rays of a single
106: pulse in a GRB are produced as a CB coasts through the SN {\it glory}
107: --the SN light scattered away from the radial direction
108: by the SN and pre-SN ejecta. The electrons enclosed in the
109: CB Compton up-scatter glory's photons to GRB energies.
110:
111: Each pulse of a GRB
112: corresponds to one CB. The emission times of the individual CBs
113: reflect the chaotic accretion process and are not predictable.
114: At the moment,
115: neither are the characteristic baryon number and Lorentz factor
116: of CBs, which can be inferred from the analysis of GRB afterglows
117: (Dado et al.~2002, 2003a). Given this information,
118: two other `priors' (the typical early
119: luminosity of a supernova and the typical density
120: distribution of the parent star's
121: wind-fed circumburst material), and a single extra hypothesis
122: (that the wind's column density in the `polar' directions
123: is significantly smaller than average) all observed properties of
124: the GRB pulses can be derived (Dar \& De R\'ujula~2004).
125: All that is required are explicit simple calculations involving
126: Compton scattering.
127:
128: Strictly speaking, our results refer to single pulses in a GRB, whose
129: properties reflect those of {\it one} CB. The statistics on single-pulse
130: GRBs of known redshift are too meager to be significant. Thus, we
131: apply our results to entire GRBs, irrespective of their number of pulses,
132: which ranges from 1 to $\sim\!12$. This implies averaging the properties
133: of a GRB over its distinct pulses,
134: and is no doubt a source of dispersion in the
135: correlations we study.
136:
137: %The rapid expansion of the CBs stops shortly after ejection
138: %(Dado et al.~2002, Dar \& De R\'ujula 2006) by their
139: %interaction with the inter-stellar medium (ISM). During this initial
140: %rapid expansion and cooling phase, their AG is
141: %dominated by thermal bremstrahlung and line emission. Later,
142: %their AG becomes dominated by synchrotron radiation from
143: %swept-in ISM electrons spiraling in the CBs' inner magnetic fields
144: %(Dado et al.~2002, 2006) and from ISM electrons scattered to higher
145: %energies by the moving CBs (Dado \& Dar 2005).
146:
147: \section{The basis of the correlations, and a summary of results}
148: \label{basis}
149:
150: Cannonballs
151: are highly relativistic, their typical Lorentz factors are
152: $\gamma\!=\!{\cal{O}}(10^3)$. They
153: are quasi-point-like: the angle a CB subtends from its point
154: of emission is comparable or
155: smaller than the characteristic opening angle,
156: $1/\gamma$, of its relativistically beamed radiation.
157: Let the typical viewing angle of an
158: observer of a CB, relative to its direction of motion,
159: be $\theta\!=\!{\cal{O}}$(1 mrad), and let
160: $\delta\!=\!{\cal{O}}(10^3)$ be the corresponding Doppler factor:
161: \begin{equation}
162: \delta \equiv {1\over\gamma\,(1-\beta\, cos\theta)}
163: \simeq {2\, \gamma
164: \over 1+\gamma^2\, \theta^2}\; ,
165: \label{delta}
166: \end{equation}
167: where the approximation is excellent
168: for $\theta\ll 1$ and $\gamma \gg 1$.
169:
170: To correlate two GRB observables, all one needs to know is
171: their functional dependence on $\delta$ and $\gamma$. The reason
172: is that, in Eq.~(\ref{delta}),
173: the $\theta$ dependence of $\delta(\gamma,\theta)$ is so pronounced, that
174: it may be expected to be the
175: largest source of the case-to-case spread in the measured quantities
176: (for GRBs of known redshift $z$, the correlations are sharpened
177: by use of the explicit $z$-dependences). In the CB model,
178: the $(\gamma,\delta,z)$ dependences of the
179: spherical equivalent energy of a GRB, $E_\gamma^{\rm iso}$;
180: its peak isotropic luminosity $L_p^{\rm iso}$;
181: its peak energy, $E_p$ (Dar \& De R\'ujula 2001);
182: and its pulse rise-time $t_{\rm rise}$ (Dar \& De R\'ujula 2004), are:
183: \begin{equation}
184: E_\gamma^{\rm iso}\propto\delta^3,\;\;\;
185: (1+z)^2\,L_p^{\rm iso}\propto \delta^4,\;\;\;
186: (1+z)\,E_p\propto \gamma\,\delta,\;\;\;
187: t_{\rm rise}/(1+z)\propto 1/(\gamma\,\delta),
188: \label{brief}
189: \end{equation}
190: The first two of these results are simple consequences of relativity
191: and the quasi-point-like character of the CB-model's sources
192: (they would be different for an assumed GRB-generating
193: jet with an opening angle much greater than $1/\theta$). The expression for
194: $E_p$ reflects the inverse Compton scattering by the CB's electrons
195: (comoving with it with a Lorentz factor $\gamma$) of the glory's photons,
196: that are approximately isotropic in the supernova rest system, and
197: are Doppler-shifted by the CB's motion by a factor $\delta$
198: (the result would be different, for instance, for synchrotron radiation
199: from the GRB's source, or self-Compton scattering of photons
200: comoving with it). The expression for the pulses's rise-time has the
201: same physical basis as that for $E_p$, but we shall see in more
202: detail in Section \ref{variability}, Eq.~(\ref{twtrans22}), that it also reflects
203: the production of $\gamma$-rays in an illuminated, previously
204: wind-fed medium\footnote{The coefficients of
205: proportionality in Eqs.~(\ref{brief}) have explicit dependences on
206: the number of CBs in a GRB, their
207: initial expansion velocity and baryon number.
208: With typical values fixed by the analysis of GRB afterglows,
209: the predictions agree with the
210: observations (Dar \& De R\'ujula 2004, Dado et al.~2006).}.
211:
212: In Sections \ref{variability} and \ref{lag}
213: we derive the CB-model's expectation for the
214: the variability of a GRB, $V$, and the prediction for the
215: `lag-time' of its pulses, $t_{\rm lag}$, to wit:
216: \begin{equation}
217: V\propto \gamma\,\delta/(1+z),\;\;\;
218: t_{\rm lag}\propto (1+z)^2/(\delta^2\,\gamma^2) .
219: \label{tlagV}
220: \end{equation}
221: The physics of the first of these relations is essentially the same as
222: that of a pulse's rise-time. The behaviour of $t_{\rm lag}$ reflects
223: the CB-model's specific prediction for how, as a pulse evolves in time,
224: the photon's energy spectrum softens, due to the increasingly non-isotropic
225: character of the glory's photons which the CBs encounter as they travel,
226: and to the softening of the energy spectrum
227: of the CB's electron population
228: (Dar \& De R\'ujula 2004).
229:
230: Given the 6 relations in Eqs.~(\ref{brief},\ref{tlagV}), it is straightforward
231: to derive the 15 ensuing two-observable correlations, of which subgroups
232: of 5 are non-redundant (if $A$ is correlated to $B$ and $B$ to $C$...).
233: We consider first one of these subgroups: the 5 correlations
234: most often phenomenologically
235: discussed to date.
236: All these correlations are derived in the same manner. Consequently,
237: we proceed by way of example and outline only the derivation of the
238: $[E_p\,,E_\gamma^{\rm iso}]$ correlation
239: (Dar \& De R\'ujula 2001, Dado et al.~2006). The full derivation is
240: in the Appendix.
241:
242: An $[E_p\,,E_\gamma^{\rm iso}]$
243: correlation was predicted [and tested]
244: in Dar \& De R\'ujula (2001, [2004]).
245: According to Eqs.~(\ref{brief}),
246: $(1\!+\! z)\,E_p\!\propto\!\gamma\delta$
247: and $E_\gamma^{\rm iso}\propto\delta^3$. If most of the variability
248: is attributed to the very fast-varying $\theta$-dependence of $\delta$ in
249: Eq.~(\ref{delta}),
250: $(1\!+\! z) E_p\!\propto\![E_\gamma^{\rm iso}]^{1/3}$.
251: (Dar \& De R\'ujula 2001).
252: This original expectation can be refined
253: by exploiting another prediction (Dado et al.~2006).
254: A typical observer's
255: angle is $\theta\!\sim\!1/\gamma$. A relatively large $E_p$
256: implies a relatively large $\delta$, and a relatively small viewing angle,
257: $\theta<1/\gamma$. For $\theta^2 \ll 1/\gamma^2$,
258: $\delta\propto\gamma$, implying that $(1\!+\! z)E_p\propto
259: [E_\gamma^{\rm iso}]^{2/3}$ {\it for the largest observed values} of
260: $E_\gamma^{\rm iso}$. On the other hand, for $\theta^2 \gg 1/\gamma^2$,
261: the Dar \& De R\'ujula (2001) correlation is unchanged:
262: it should be increasingly accurate
263: {\it for smaller values} of $E_\gamma^{\rm iso}$.
264: We may interpolate between these extremes by positing:
265: \begin{equation}
266: (1\!+\! z)\,E_p=E_p^0\,\left\{
267: [E_\gamma^{\rm iso}/E_0^{\rm iso}]^{1/3}
268: +[E_\gamma^{\rm iso}/E_0^{\rm iso}]^{2/3}\right\} \; ,
269: \label{epi}
270: \end{equation}
271: an expression with two parameters ($E_p^0$, $E_0^{\rm iso}$);
272: like the correlation $E_p=a\,[E_\gamma^{\rm iso}]^b$
273: (see, e.g., Amati~2006a,b,c), whose power behaviour
274: is arbitrary. A fit to Eq.~(\ref{epi}) is shown in Fig.~\ref{f1}a.
275: The variances around the
276: mean trends of all the correlations we study
277: have roughly log-normal distributions. Thus our fits are to
278: the logarithms of the observed quantities.
279:
280: The number fluence of a GRB is proportional
281: to $\delta^2$, its energy fluence to $\delta^3$. The individual-photon
282: energies are $\propto\delta\,\gamma$. All these facts imply that one
283: expects most observed events to correspond to small $\theta$ (though
284: obviously not to $\theta\simeq 0$, a set of null solid angle). Small
285: $\theta$ means large $E_p$ and $E_\gamma^{\rm iso}$, approximately
286: related by $(1\!+\! z)E_p\propto [E_\gamma^{\rm iso}]^{2/3}$.
287: The expectation is supported by the data in Fig.~\ref{f1}a.
288: These comments
289: on the $[E_p\,,E_\gamma^{\rm iso}]$
290: correlation and the way it is derived are extensible to
291: the other correlations we shall discuss, e.g., to the
292: $[E_p\,,L_p^{\rm iso}]$ correlation predicted and tested
293: in Dar \& De R\'ujula (2001, 2004). Given Eq.~(\ref{brief})
294: and in analogy with Eq.~(\ref{epi}),
295: we expect (Dado et al.~2006):
296: \begin{equation}
297: (1+z)\,E_p \simeq E_p^0\,
298: \left\{[(1+z)^2\,L_p^{\rm iso}/L_p^0]^{1/4} +
299: [(1+z)^2\,L_p^{\rm iso}/L_p^0]^{1/2}\right\} .
300: \label{lpep}
301: \end{equation}
302: This correlation is akin to the one proposed by Yonetoku et al.~(2004),
303: but its power behaviour is not arbitrary.
304:
305: To derive the other correlations we shall confront with data, we must
306: recall the CB-model
307: expectations for pulse rise-times and event variabilities,
308: and derive the one for the lag-time. But Eqs.~(\ref{brief},\ref{tlagV})
309: allow us to anticipate the results:
310: \begin{equation}
311: {t_{\rm min}^{\rm rise}\, (1+z)^{-1}}=t_0 \left\{
312: \left[(1+z)^2\,L_p^{\rm iso}/L_p^0 \right]^{1\over 4}
313: +\left[(1+z)^2\,L_p^{\rm iso}/L_p^0\right]^{1\over 2} \right\}^{-1}
314: \label{tmineq}
315: \end{equation}
316: \begin{equation}
317: {(1+z)\,V}=V_0\;\left\{
318: \left[(1+z)^2\,L_p^{\rm iso}/L_p^0\right]^{1\over 4}
319: +\left[(1+z)^2\,L_p^{\rm iso}/L_p^0\right]^{1\over 2}\right\}
320: \label{Veq}
321: \end{equation}
322: \begin{equation}
323: {t_{\rm lag}\, (1+z)^{-2}}=t_0 \left\{
324: \left[(1+z)^2\,L_p^{\rm iso}/L_p^0 \right]^{1\over 2}
325: +\left[(1+z)^2\,L_p^{\rm iso}/L_p^0\right]^{1} \right\}^{-1}
326: \label{tlageq}
327: \end{equation}
328:
329: The predictions of Eqs.~(\ref{epi}) to (\ref{tlageq}) are
330: tested in Figs.~\ref{f1} and \ref{f2}a. The data are
331: from Schaefer (2006); values of $E_0^{\rm iso}$
332: and data for some extra GRBs at low $E_0^{\rm iso}$ and $E_p$
333: (which could be classified as X-ray flashes) are from Amati
334: (2006a,b,c). All results are satisfactory.
335:
336: Two observables ($L_p^{\rm iso}$ and $t_{\rm min}^{\rm rise}$)
337: reflect fixed values of $\gamma$ and $\delta$, since they
338: refer to a particular pulse in a GRB light curve, not necessarily
339: the same one.
340: For multi-pulse GRBs, the other observables reflect averages over the
341: various pulses. Single-peak correlations should
342: be tighter than the ones we have discussed,
343: but properly analized data are not available.
344: We also expect
345: correlations between two multi-pulse-averaged observables to be
346: tighter than those between a multi-pulse observable and a
347: single-pulse one.
348: The only correlation of the former type in Figs.~\ref{f1} and \ref{f2} is
349: that between $E_p$ and $E_\gamma^{\rm iso}$ in Fig.~\ref{f1}a;
350: it is indeed the tightest.
351:
352: The correlations we have discussed are not the simplest ones the
353: CB model suggests. Indeed, it follows from Eqs.~(\ref{brief},\ref{tlagV})
354: that:
355: \begin{equation}
356: t_{\rm lag} \propto E_p^{-2},\;\;\;\;\;
357: E_\gamma^{\rm iso}\propto
358: [(1+z)^2\,L_p^{\rm iso}]^{3/4},\;\;\;\;\;
359: V\propto E_p,\;\;\;\;\;
360: t_{\rm rise} \propto E_p^{-1}.
361: \label{simple}
362: \end{equation}
363: These relations involve just one
364: parameter: the proportionality factor, and deserve to be studied,
365: even if they add no significance to the results, for they are redundant
366: with the correlations we have already discussed. The first two
367: predictions in Eqs.~(\ref{simple}) are shown in Figs.~\ref{f2}b,c.
368: The correlations of $V$ and $t^{\rm rise}_{\rm min}$
369: with $E_p$ are less informative,
370: not only because the first two of these observables are of
371: a somewhat debatable significance, but because
372: the dynamical ranges of $V$, $t^{\rm rise}_{\rm min}$
373: and $E_p$ span $\sim 2$ orders of
374: magnitude, while the data on $L_p^{\rm iso}$,
375: $E_\gamma^{\rm iso}$ and $t_{\rm lag}$ span $\sim 3$ times
376: as many.
377:
378:
379:
380: \section{Variability and minimum rise-time}
381: \label{variability}
382:
383: In the CB model there are two a priori time scales determining
384: the rise-time and duration of a pulse: the time it takes
385: a CB to expand to the point at which it becomes transparent
386: to radiation and the time it takes it to travel to a distance
387: from which the remaining of its path is transparent
388: to $\gamma$ rays (Dar \& De R\'ujula 2004). These two times
389: are, for typical parameters, of the same order of magnitude.
390: We discuss the second time scale here, for it is the one
391: naturally leading to
392: larger variabilities and differences in rise-time.
393:
394: The $\gamma$ rays of a GRB's pulse must traverse the
395: pre-SN wind material remaining upstream of their production
396: point, at a typical distance of $r={\cal{O}}(10^{16}\,\rm cm)$
397: from the parent SN. At these `short' distances, the observed
398: circumburst material is located in layers whose density
399: decreases roughly
400: as $1/r^2$ and whose typical $\rho\,r^2$ is large:
401: $\sim 10^{16}\,{\rm g \,cm}^{-1}$
402: (Chugai et al.~2003; Chugai \& Danzinger 2003).
403: Compton absorption in such a wind implies that a pulse of
404: a GRB initially rises with time as ${\rm Exp}[-(t^w_{tr}/t)^2]$, where
405: (Dar \& De R\'ujula 2004):
406: \begin{equation}
407: {t^w_{tr}\over 1+z}=(0.13\,{\rm s})\;
408: {\rho\,r^2\over 10^{16}\,{\rm g \,cm}^{-1}}\;
409: {10^6\over \gamma\,\delta}.
410: \label{twtrans22}
411: \end{equation}
412: The values of $\delta\,\gamma$
413: may differ for the different CBs (pulses) of a GRB
414: even if they are emitted in precisely the same direction, which
415: need not be the case, e.g. if the emission axis precesses.
416: The minimum rise-time, $t_{\rm rise}$,
417: used as a variability measure in Schaefer (2006), satisfies
418: Eq.~(\ref{twtrans22}), and was used in Fig.~\ref{f1}c.
419:
420: The result in Eq.~(\ref{twtrans22}) is for an ideal spherically-symmetric
421: wind. Actual wind distributions are layered and patchy,
422: implying an in-homogeneous distribution of the glory's light density.
423: Since the number of photons Compton up-scattered by the CB
424: is proportional to this density, the
425: inhomogeneities would directly translate into a variability on top
426: of a smooth pulse shape, which reflects the average density distribution
427: of the wind-fed medium.
428: This corresponds to a source of variability
429: that, as a function of $\delta$, $\gamma$ and $z$,
430: behaves as the inverse of $t^w_{tr}$, the form used for $V$
431: in Eq.~(\ref{brief}) and Fig.~\ref{f1}d
432: (in the CB model, the deviations from a smooth
433: behaviour observed in some optical and X-ray AGs also trace
434: the inhomogeneities in the density of the interstellar
435: medium; see Dado et al.~2003a, 2006). There are many ways to define the
436: variability of a GRB; variations
437: of the sort we have described are the ones
438: studied by Schaefer (2006). The data in
439: Figs.~\ref{f1}c,d are from his analysis and definitions.
440:
441:
442: \section{The lag-time}
443: \label{lag}
444:
445: In the CB model a pulse's $\gamma$-ray number flux
446: as a function of energy and time is of the form (Dar \& De R\'ujula, 2004):
447: \begin{equation}
448: N(E,t)\equiv{d^2N\over dE\,dt}={dN_1(E,t)\over dE}\,{dN_2(t)\over dt}\, .
449: \label{NEt}
450: \end{equation}
451: The function $dN_2/dt$
452: is well approximated by
453: ${\rm Exp}[-(t^w_{tr}/t)^2]\,\{[1-{\rm Exp}[-(t^w_{tr}/t)^2]\}$,
454: with $t^w_{tr}$ as in Eq.~(\ref{twtrans22}).
455: The predicted shape of $dN_1/dE$ is amazingly similar to that of
456: `Band's' phenomenological spectrum and has a weak time-dependence that makes the
457: spectrum within a pulse soften with time, in a time of
458: order $t^w_{tr}$. The energies $E$ in $dN_1/dE$ scale with $T$:
459: \begin{equation}
460: T\equiv {4\over 3} \;T_i\;{\gamma\;\delta\over 1+z}\;\,
461: \langle 1+\cos\theta_i\rangle,
462: \label{Teff}
463: \end{equation}
464: where $\theta_i$ is the angle of incidence of a glory's photon onto
465: the CB (in the SN rest system) and $T_i$ is the pseudo-temperature
466: in the thin thermal-bresstrahlung spectrum $[{\rm Exp}(-E_i/T_i)]/E_i$
467: of the glory's light. We conclude that
468: $N(E,t)=F\left({E/T}\; , {t/ t^w_{tr}}\right)$
469: with $F$ a predicted function. This implies that
470: \begin{equation}
471: t_{\rm rise}(E)=t^w_{tr}\; G(E/T),
472: \label{tpformula}
473: \end{equation}
474: with $G(x)$ a slowly-varying function of $x$
475: (a fact that can be traced back to the time dependence
476: of $dN_1/dE$ being much slower than that of $dN_2/dt$).
477:
478: Let $t_{\rm rise}(E_i)$ be the rise-time of a pulse at a given $\gamma$-ray energy $E_i$. The lagtime is defined and approximated as:
479: \begin{equation}
480: t_{\rm lag}\simeq t_{\rm rise}(E_2)-t_{\rm rise}(E_1) \approx \Delta\,E\;{dt_{\rm rise}\over dE}\; ,
481: \label{tlagdef}
482: \end{equation}
483: where $\Delta E$ is generally taken to be a fixed
484: energy interval between two `channels'
485: in a given detector.
486: Use Eqs.~(\ref{tpformula}, \ref{tlagdef}) to deduce that:
487: \begin{equation}
488: t_{\rm lag}\approx t^w_{tr}\,{dG\over dE}\,\Delta\,E\propto
489: {t^w_{tr}\over T}\,\Delta\,E
490: \label{tlagr1}
491: \end{equation}
492: where, on dimensional grounds, we used $dG/dE\propto 1/T$.
493: It follows from Eqs.~(\ref{Teff},\ref{twtrans22}) that:
494: \begin{equation}
495: t_{\rm lag}\propto
496: {t^w_{tr}\over T} \propto {(1+z)^2\over \delta^2 \, \gamma^2}\,,
497: \label{tlagr2}
498: \end{equation}
499: the result announced in Eq.~(\ref{tlagV}) and used in Fig.~\ref{f2}a.
500:
501:
502: \section{The duration of a GRB pulse}
503: \label{fwhm}
504:
505: Some correlations do not follow from comparisons of $\gamma$
506: and $\delta$ dependences. One of them (Dar \& De R\'ujula 2004)
507: is the following. As the CB reaches the more transparent
508: outskirts of the wind, its ambient light becomes
509: increasingly radially directed, so that the average $1+\cos\theta_i$ in
510: Eq.~(\ref{Teff}) will tend to $0$
511: as $1/r^2\propto 1/t^2$.
512: Since the (exponential) rise of a
513: typical pulse is much faster than its
514: (power) decay, the width of a peak is dominated by its late
515: behaviour at $t>t_{tr}$. At such times, $T\propto 1/t^2$ in Eq.~(\ref{Teff}),
516: so that $dN/dE$ is, approximately, a function of the combination $E\,t^2$.
517: Consequently, the width of a GRB pulse in different energy bands is:
518: $ \Delta t\propto E^{-1/2}$,
519: in agreement with the observation,
520: $t_{_{\rm FWHM}}\propto E^{-0.43\pm 0.10}$ ,
521: for the average FWHM of peaks as a function of the energies
522: of the four BATSE channels (Fenimore et al.~1995, Norris et al.~1996).
523: This correlation is shown in Fig.~\ref{f2}d.
524:
525:
526:
527:
528: \section{Conclusions}
529:
530: The CB model is very successful in its description of all properties
531: of the pulses of long-duration GRBs (Dar \& De R\'ujula 2004).
532: We have extended our previous discussions
533: of one of these properties: the correlations between pairs of observables.
534: We have analyzed a wealth of newly available data and derived
535: predictions for some observables which we had not studied before.
536: Although our predictions are expected to be better satisfied for individual pulses, the results are very satisfactory
537: even when applied to entire GRBs: all of the
538: predicted trends agree with the observations. The
539: correlations we have discussed have a common and simple physical basis:
540: relativistic kinematics and Compton scattering. The viewing
541: angle $\theta$ is the most crucial parameter underlying the
542: correlations, and determining the properties of GRBs and
543: their larger-$\theta$ counterparts, X-ray flashes
544: (Dar \& De R\'ujula 2004, Dado et al.~2004).
545:
546:
547:
548: \noindent
549: {\bf Acknowledgements.}
550: { This research was supported in part by the
551: Helen Asher Space Research Fund and the Institute for
552: Theoretical Physics at the Technion Institute.
553: ADR thanks the Institute for its hospitality.}
554:
555: \section{Appendix}
556:
557: In this Appendix, and in the example of the $[E_p,\,E_\gamma^{\rm iso}]$
558: correlation, we prove in detail the ``double-power" nature of many of
559: the correlations predicted by the CB model.
560:
561:
562: For a typical angle of incidence (Dar \& De R\'ujula 2004), the
563: energy of a Compton up-scattered photon from the SN glory
564: is Lorentz and Doppler boosted by a factor $\sim\!\gamma\,\delta/2$
565: and redshifted by $1\!+\! z$. The peak energy $E_p$ of the GRB's
566: $\gamma$-rays
567: is related to the peak energy, $\epsilon_p\!\sim\! 1$ eV, of the
568: glory's light by:
569: \begin{equation}
570: (1+z)\,E_p\simeq {\gamma\,\delta\, \epsilon_p\over 2}\simeq
571: (500\;{\rm keV})\; {\gamma\,\delta\over 10^6}\,
572: {\epsilon_p\over 1\;\rm eV}\; .
573: \label{eobs}
574: \end{equation}
575: %The upscattered radiation, emitted nearly isotropically
576: %in the CB's rest frame, is boosted by its highly relativistic motion
577: %to a narrow angular distribution whose number density is:
578: %\begin{equation}
579: %{dn_\gamma \over d\Omega}\simeq {n_\gamma \over 4\, \pi}\, \delta^2
580: % \simeq {n_\gamma \over 4\, \pi}\, {4\, \gamma^2
581: % \over (1+\gamma^2\, \theta^2)^2}\, ,
582: %\label{beaming}
583: %\end{equation}
584: The spherical equivalent
585: energy, $E_\gamma^{\rm iso}$, is (Dar \& De R\'ujula 2004, Dado et al.~2006b):
586: \begin{equation}
587: E_\gamma^{\rm iso} \simeq
588: {\delta^3\, L_{_{\rm SN}}\,N_{_{\rm CB}}\,\beta_s\over 6\, c}\,
589: \sqrt{\sigma_{_{\rm T}}\, N_b\over 4\, \pi}\sim
590: (3.8\! \times\! 10^{53}\,{\rm erg})\,{\delta^3\over 10^9}\,
591: {L_{_{\rm SN}}\over L_{_{\rm SN}}^{\rm bw}}\,{N_{_{\rm CB}}\over 6}\,
592: \beta_s\sqrt{ N_b\over 10^{50}}\; ,
593: \label{eiso}
594: \end{equation}
595: where $L_{_{\rm SN}}$ is the mean SN optical luminosity just
596: prior to the ejection of CBs, $N_{_{\rm CB}}$ is the number of CBs in
597: the jet, $N_b$ is their mean baryon number, $\beta_s$ is the comoving early
598: expansion velocity of a CB (in units of $c/\sqrt{3}$),
599: and $\sigma_{_{\rm T}}$ is the Thomson cross section. The early SN luminosity required to
600: produce the mean isotropic energy, $E_\gamma^{\rm iso}\!\sim\! 4\!\times\! 10^{53}$
601: erg, of ordinary long GRBs is
602: $L_{_{\rm SN}}^{\rm bw}\!\simeq\! 5\!\times\! 10^{42}\, {\rm erg\,
603: s^{-1}}$, the estimated early luminosity of SN1998bw.
604:
605: The explicit proportionality factors in the relations
606: $E_p\propto\gamma\,\delta$
607: and $E_\gamma^{\rm iso}\propto\delta^3$ are given by
608: Eqs.~(\ref{eobs},\ref{eiso}). Let us first consider them fixed
609: at their typical values.
610: The typical $[\gamma,\,\delta]$ domain of observable GRBs is
611: then the one shown in
612: Fig.~\ref{f3}a. The observed values of $\gamma$ are fairly
613: narrowly distributed around $\gamma\!\sim\!10^3$ (Dado et al.~2003a,
614: Dar \& De R\'ujula, 2004),
615: as in the blue strip of the figure. The $[\gamma,\,\delta]$ domain
616: is also limited by a minimum
617: observable isotropic energy or fluence (both $\propto\,\delta^3$), by
618: a minimum observable peak energy, and by the line
619: $\theta=0$ or, if one takes into account that phase space for
620: observability diminishes as $\theta\to 0$, by a line corresponding to
621: a minimum fixed $\theta$. The elliptical ``sweet spot" in Fig.~\ref{f3}a
622: is the region wherein GRBs are most easily detectable, particularly
623: in pre-Swift times. In the CM model X-ray Flashes are GRBs
624: seen at a relatively large $\theta\gamma$ (Dar \& De R\'ujula, 2004,
625: Dado et al.~2004) they populate the region labeled
626: XRF in the figure, above the fixed $\gamma\theta$ line or to the
627: left of the fixed $E_p$ line.
628:
629:
630: The blue line in Fig.~\ref{f3}d is the contour of the blue domain of
631: Fig.~\ref{f3}a,
632: shown in the $[\gamma\delta,\,\delta^3]$ plane of the
633: $[E_p,\,E_\gamma^{\rm iso}]$ correlation. At low values of these quantities,
634: the correlation is $E_p\,\propto\!(E_\gamma^{\rm iso})^{1/3}$. At the opposite
635: extreme, the expected power is half-way
636: between 1/3 and 2/3. But there is another effect
637: increasing this expectation to $\sim\!2/3$.
638:
639: A CB that is expanding, in its rest system, at a speed of relativistic sound
640: ($\beta_{\rm exp}\!=\!\beta_s$)
641: --or at the speed of light ($\beta_{\rm exp}\!=\!1$)--
642: subtends a non-vanishing angle from its
643: point of emission. In the SN rest system, this (half-)angle is
644: $\theta_{_{\rm CB}}\!=\!\beta_{\rm exp}/\gamma$. At a fixed observer's angle,
645: $\theta$,
646: the value of $\delta$ and $\delta^3$ entering Eqs.~(\ref{eobs},\ref{eiso})
647: are not the ``naive" ones of Eq.~(\ref{delta}), but are averages,
648: $\langle\delta\rangle$ and $\langle\delta^3\rangle$, over the
649: CB's non-vanishing surface. In Fig.~\ref{f3}b we show function
650: $\langle\delta(\theta\gamma)\rangle$, for fixed $\gamma$ and
651: $\theta_{_{\rm CB}}\!=\!1/\gamma$ (the result is very similar for
652: $\beta_{\rm exp}=1/\sqrt{3}$). This figure is easy to interpret: for
653: $\theta\gg\theta_{_{\rm CB}}$, the CB is effectively point-like and
654: $\langle\delta(\theta\gamma)\rangle\to \delta$. At the opposite
655: extreme, $\theta\ll\theta_{_{\rm CB}}$, the observer's angle is immaterial
656: and $\langle\delta(\theta\gamma)\propto\gamma$. The consequences
657: of this fact on the $[E_p,\,E_\gamma^{\rm iso}]$ correlation
658: can be seen in Fig.~\ref{f3}c, where we have plotted
659: $E_p\!\propto\!\gamma\langle\delta\rangle$ versus
660: $E_\gamma^{\rm iso}\!\propto\!\langle\delta^3\rangle$, at fixed $\gamma$.
661: The $E_p(E_\gamma^{\rm iso})$ functional
662: dependence smoothly evolves from a 1/3 to a 2/3 power.
663: We plot in Fig.~\ref{f3}d, as the banana-like
664: dashed line, the border of the blue
665: domain of Fig.~\ref{f3}a, taking into account the geometrical effect we just
666: described. The result is an $[E_p,\,E_\gamma^{\rm iso}]$ correlation
667: with an index varying from 1/3 to $\sim\!2/3$.
668:
669: Finally, we may consider the effect of varying the proportionality factors
670: in Eqs.~(\ref{eobs},\ref{eiso}) around their reference values. This results
671: in a superposition of banana-like domains, the general behaviour of
672: which we have approximated by the $[E_p,\,E_\gamma^{\rm iso}]$ correlation
673: of Eq.~(\ref{epi}). Similar considerations apply to all the other two-power
674: correlations that we have studied.
675:
676:
677:
678: \begin{thebibliography}
679:
680:
681:
682: \bibitem[2006]{AM2006a}
683: Amati, L., 2006a, MMRAS, 372, 233
684: %\bibitem[2006]{AM2006b}
685: %Amati, L.,??? Pian 2006b, astro-ph/0607148
686: \bibitem[2006]{AM2006b}
687: Amati, L. 2006b, astro-ph/0601553
688: \bibitem[2006]{AM2006c}
689: Amati, L., 2006c, astro-ph/0611189
690:
691: \bibitem[2003]{ChugaiD}
692: Chugai, N.N. \& Danziger, I.J.~2003, Astron. Lett. 29 649
693:
694: \bibitem[2003]{Chugaietal}
695: Chugai, N.N. et al.
696: Proceedings of IAU Colloquium 192, {\ it Supernovae},
697: eds. J. M. Marcaide and K. W. Weiler
698:
699: \bibitem[2002]{DDD2002}
700: Dado, S., Dar, A. \& De R\'ujula, A.~2002, A\&A, 388, 1079
701:
702: \bibitem[2003]{DDD2003a}
703: Dado, S., Dar, A. \& De R\'ujula, A.~2003a, A\&A, 401, 243
704:
705: \bibitem[2003]{DDD2003b}
706: Dado, S., Dar, A. \& De R\'ujula, A.~2003b, ApJ, 594, L89
707:
708: \bibitem[2004]{DDD2004}
709: Dado, S., Dar, A. \& De R\'ujula, A.~2004, A\&A, 422, 381
710:
711: %\bibitem[2004]{DDD2004b}
712: %Dado, S., Dar, A. \& De R\'ujula, A.~2004b,
713:
714: %\bibitem[2006]{DDD2006}
715: %Dado, S., Dar, A. \& De R\'ujula, A.~2006, ApJ. 646, L2106325
716:
717: %\bibitem[2006]{DDD2006}
718: %Dado, S., Dar, A. \& De R\'ujula, A.~2006, ApJ. 646, L21
719:
720: \bibitem[2003]{DDD2006a}
721: Dado, S., Dar, A. \& De R\'ujula, A.~2006a, ApJ, 646, L21
722:
723: \bibitem[2006b]{DDDP2006b}
724: Dado, S., Dar, A., De R\'ujula and Plaga, R., astro-ph/0611161
725:
726: %\bibitem[2005]{DD2005}
727:
728: %Dado, S. \& Dar, A. 2005 ApJ. 627 L109
729:
730: %\bibitem[2005]{DD2005a}
731: %Dado, S. \& Dar, A., 2005a, Nuovo Cimento 120, 731
732:
733: %\bibitem[2004]{DAR2004}
734: %Dar, A. 2004, Proc. 2004 Vulcano Workshop (eds. F. Giovannelli
735: %\& G. Mannocchi) p. 287 (astro-ph/0405386)
736:
737: \bibitem[2000]{DP1999}
738: Dar, A. \& Plaga, R. ~1999, A\&A, 349, 259
739:
740: \bibitem[2000]{DD2000}
741: Dar, A. \& De R\'ujula, A.~2000, astro-ph/0008474
742:
743: \bibitem[2000]{DD2001}
744: Dar, A. \& De R\'ujula, A.~2001, astro-ph/0012227
745:
746: \bibitem[2004]{DD2004}
747: Dar, A. \& De R\'ujula, A.~2004, Physics Reports, 405, 203
748:
749: %\bibitem[2004]{DD2006}
750: %Dar, A. \& De R\'ujula, A.~2006, hep-ph/0606199
751:
752: %\bibitem[1992]{DKNR1992}
753: %Dar, A., Kozlovsky, B.; Nussinov, S. \& Ramaty, R. 1992, ApJ, 388, 164
754:
755: \bibitem[1987]{DR1987}
756: De R\'ujula, A, 1987, Phys. Lett. 193, 514
757:
758: \bibitem{}
759: Fenimore, E. E., in 't Zand, J. J. M., Norris, J. P., Bonnell, J. T.
760: \& Nemiroff, R. J. 1995, ApJ, 448, L101
761:
762: \bibitem[2000]{FenRR2000}
763: Fenimore, E.E. \& Ramirez-Ruiz, E. 2000, astro-ph/0004176
764:
765: \bibitem[2006]{Guid}
766: Guidorzi, C., et al. 2006, MNRAS, 371, 843
767:
768:
769: \bibitem{}
770: Norris, J. P., Nemiroff, R. J., Bonnell, J. T., Scargle, J. D.,
771: Kouveliotou, C., Paciesas, W. S., Meegan, C. A. \& Fishman, G. J. 1996,
772: ApJ, 459, 393
773:
774: \bibitem[2001]{Plaga}
775: Plaga, R. 2001, A\&A, 370, 351
776:
777: \bibitem[2001]{Reichart}
778: Reichart, D. E., et al. 2001, ApJ, 552, 57
779:
780: \bibitem[2006]{WB2006}
781: Schaefer, B. E., astro-ph/0612285
782:
783: \bibitem[1995]{SD1995}
784: Shaviv, N. J., Dar, A. 1995, ApJ, 447, 863
785:
786:
787: \bibitem[2004]{YO2004}
788: Yonetoku, D., et al. 2004, ApJ. 609, 935
789:
790:
791: \end{thebibliography}
792:
793:
794:
795: \begin{figure}[]
796: \centering
797: \vbox{
798: \hbox{
799: \epsfig{file=f1a.eps,width=8cm}
800: \epsfig{file=f1b.eps,width=8cm}
801: }}
802: \vbox{
803: \hbox{
804: \epsfig{file=f1c.eps,width=8cm}
805: \epsfig{file=f1d.eps,width=8cm}
806: }}
807: \vspace*{8pt}
808: \caption{
809: Left to right and top to bottom (a to d):
810: a) The $[E_p,E_\gamma^{\rm iso}]$ correlation of Eq.~(\ref{epi}).
811: b) The $[E_p,L_p^{\rm iso}]$ correlation of Eq.~(\ref{lpep}).
812: c) The $[t_{\rm min}^{\rm rise},L_p^{\rm iso}]$ correlation of
813: Eq.~(\ref{tmineq}).
814: d) The $[V,L_p^{\rm iso}]$ correlation of Eq.~(\ref{Veq}).
815: The dotted `variance lines' are to guide the eye, they are not
816: always symmetric about the best fit.
817: }
818: \label{f1}
819: \end{figure}
820:
821: \begin{figure}[]
822: \centering
823: \vbox{
824: \hbox{
825: \epsfig{file=f2a.eps,width=8cm}
826: \epsfig{file=f2b.eps,width=8cm }
827: }}
828: \vspace{-3cm}
829: %\hskip -.8cm
830: \vbox{
831: \hbox{
832: \epsfig{file=f2c.eps,width=8.cm}
833: \hskip -.3cm
834: \vspace*{30pt}
835: \epsfig{file=f2d.ps,width=8.2cm}
836: \vspace{-1cm}
837: }}
838: \vspace*{8pt}
839: \caption{ Left to right and top to bottom (a to d):
840: a) The $[t_{\rm lag},L_p^{\rm iso}]$ correlation of Eq.~(\ref{tlageq}).
841: b) The $[t_{\rm lag},E_p]$ correlation of Eq.~(\ref{simple}).
842: c) The $[E_\gamma^{\rm iso},L_p^{\rm iso}]$ correlation of
843: Eq.~(\ref{simple}).
844: d) The correlation between $t_{\rm FHHM}$ and BATSE
845: $E$-channel of Section \ref{fwhm}.
846: The dotted `variance lines' are to guide the eye, they are not
847: always symmetric about the best fit.
848: }
849: \label{f2}
850: \end{figure}
851: \begin{figure}[]
852: \centering
853: \vbox{
854: \hbox{
855: %\vspace{2cm}
856: \epsfig{file=f3a.eps,width=8cm}
857: \hspace{1.cm}
858: \epsfig{file=f3b.eps,width=7.5cm}
859: }
860: \vspace{2cm}
861: \hbox{
862: \epsfig{file=f3c.eps,width=16cm}
863: %\vspace*{-50pt}
864: %\epsfig{file=EpEisoCurved2.eps,width=7.5cm}
865: }}
866: \vspace*{8pt}
867: \caption{
868: Left to right and top to bottom: a) The $[\delta,\,\gamma]$ domain.
869: b) The ratio $\langle\delta\rangle/\delta$ of average to naive Doppler
870: shifts for a (Lorentz-contracted)
871: disk-like CB of angular size $1/\gamma$ (in the SN rest
872: system), as a function of $\theta\,\gamma$. c) $E_p$ versus
873: $E_\gamma^{\rm iso}$ for the same CB, at fixed $\gamma$.
874: d) Contours of $E_p$ versus $E_\gamma^{\rm iso}$ for an ensemble of CBs
875: whose values of $[\delta,\,\gamma]$ are in the blue domain of
876: Fig.~\ref{f3}a. The continuous (blue) contour does not take into account
877: the non-point-like character of CBs. The dashed contour does.
878: }
879: \label{f3}
880: \end{figure}
881:
882: \end{document}
883:
884: