New Astronomy Reviews 48 (2004) 445–451

www.elsevier.com/locate/newastrev

GeV–TeV emission from c-ray bursts

P. M�esz�aros a,b,*, S. Razzaque a,b, B. Zhang a

a Department of Astronomy and Astrophysics, Pennsylvania State University, University Park, PA 16802, USAb Department of Physics, Pennsylvania State University, University Park, PA 16802, USA

Abstract

Most of the current knowledge about c-ray bursts (GRB) is based on electromagnetic observations at MeV and

lower energies. Here, we focus on some recent theoretical work on GRB, in particular the higher energy (GeV–TeV)

photon emission, and the expected TeV and higher energy neutrino emission.

� 2003 Elsevier B.V. All rights reserved.

Keywords: c-ray bursts; c-rays; Neutrinos; TeV

1. Introduction

A new era in c-ray Burst (GRB) research

opened in 1991 with the launch of the Compton

Gamma-Ray Observatory (CGRO), whose

ground-breaking results have been summarized in

Fishman and Meegan (1995). The most significant

results came from the all-sky survey by the Burst

and Transient Experiment (BATSE) on CGRO,which recorded over 2700 bursts, complemented

by data from the OSSE, Comptel and EGRET

experiments, in the 20 keV–1 MeV, 0.1–10 MeV,

1–30 MeV and 20 MeV–30 GeV c-ray bands, re-

spectively. The next major advance was ushered in

by the Italian–Dutch satellite Beppo-SAX in 1997,

which obtained the first high resolution 0.2–10 keV

X-ray images (Costa et al., 1997) of the fadingafterglow of a burst, GRB 970228, a phenomenon

* Corresponding author.

E-mail address: nnp@astro.psu.edu (P. M�esz�aros).

1387-6473/$ - see front matter � 2003 Elsevier B.V. All rights reserv

doi:10.1016/j.newar.2003.12.022

which had been previously theoretically predicted.This was followed by an increasing list of burst

afterglow detections with Beppo-SAX. These X-

ray detections, after a 4–6 h delay needed for data

processing, led to arc-minute accuracy positions,

which finally made possible the optical detection

and the follow-up of the GRB afterglows at longer

wavelengths (e.g., Frail et al., 1997; van Paradijs

et al., 1997). This paved the way for the mea-surement of redshift distances, the identification of

candidate host galaxies, and the confirmation that

they were at cosmological distances (Metzger

et al., 1997; Kulkarni et al., 2000), etc. In the last

year, the HETE-2 spacecraft 1 has been discover-

ing bursts at the rate of �2/month, in the 0.5–400

keV range. Besides classical bursts, a number of

these have been in the sub-class of X-ray flashes,first identified with Beppo-SAX (Heise et al.,

2001), or X-ray rich bursts (Kippen et al., 2002).

1 http://space.mit.edu/HETE/.

ed.

446 P. M�esz�aros et al. / New Astronomy Reviews 48 (2004) 445–451

Another remarkable observation (Akerlof et al.,

1999) was that of a prompt and extremely bright

(mv � 9) optical flash in the burst GRB 990123, 15

s after the GRB started (and while it was still going

on). A prompt multi-wavelength flash, contem-

poraneous with the c-ray emission and reachingsuch optical magnitude levels is an expected con-

sequence of the reverse component of external

shocks (M�esz�aros and Rees, 1993, 1997). The

prompt optical flash of 990123 is generally inter-

preted (M�esz�aros and Rees, 1999; Sari and Piran,

1999) as the radiation from a reverse external

shock (although prompt optical flashes can be

expected from both an external reverse shock andfrom an internal shock (M�esz�aros and Rees,

1997)). The behavior of the reverse shock is more

sensitive than the forward shock on the details of

the shock physics (Kobayashi, 2000). This has

become apparent with the recent prompt flashes

seen from ground-based optical follow-up of two

other bursts, GRB 021004 and GRB 021211 (Fox

et al., 2003a,b; Li et al., 2003), made possible byHETE-2 spacecraft position alerts. In the near

future, the prospects look bright for detecting 100–

150 bursts per year with a new multi-wavelength

GRB follow-up spacecraft called Swift, 2 which is

due for launch in December 2003 and is equipped

with c-ray, X-ray and optical detectors. Predic-

tions of the dependence of spectra and lightcurves

on the initial bulk Lorentz factor, magnetic energycontent, etc. (Zhang et al., 2003) opens up the in-

teresting possibility of charting in much greater

detail the shock physics and the dynamics of the

relativistic outflow in GRB.

The large bulk of the information on GRB has

come from electromagnetic signals at MeV and

lower energies, except for a handful of bursts de-

tected at higher photon energies with SMM(Harris and Share, 1998), with the EGRET, OSSE

and Comptel experiments on CGRO (Fishman

and Meegan, 1995), and more recently INTE-

GRAL (von Kienlin et al., 2003). However, new

missions are planned for ultra-high energies

(UHE). One of these, aimed at the 20 MeV–300

GeV as well as 10 keV–30 MeV range, is GLAST

2 http://swift.gsfc.nasa.gov/.

(Gehrels and Michelson, 1999), due for launch in

2006. In addition, several new ground-based air

and water Cherenkov telescopes as well as solar

array facilities are coming or will shortly come on-

line, sensitive in the 0.2–10 TeV range, with GRB

as one possible class of targets (Weekes, 2000).Furthermore, there are several other, as yet un-

confirmed, but potentially interesting observing

channels for GRBs, relating to the baryonic com-

ponent of the outflow, the shock physics and the

progenitor collapse dynamics. There is the pros-

pect of measuring ultra-high energy cosmic rays

from GRB with the Auger array (Letessier-Selvon,

2003); TeV and higher energy neutrinos with large,cubic kilometer ice or water Cherenkov telescopes

such as ICECUBE or ANTARES (Halzen, 2000);

and measuring gravitational wave signals accom-

panying the GRB formation with LIGO and

similar arrays (Finn et al., 1999). We discuss below

some of the theoretical expectations for the GRB

ultra-high energy JGeV electromagnetic signals,

and the UHE neutrino signals.

2. UHE photons from GRB

Ultra-high energy emission, in the range of GeV

and harder, is expected from electron inverse

Compton in external shocks (M�esz�aros et al.,

1994) as well as from internal shocks (Papath-anassiou and M�esz�aros, 1996) in the prompt

phase. The combination of prompt MeV radiation

from internal shocks and a more prolonged GeV

IC component for external shocks (M�esz�aros andRees, 1994) is a likely explanation for the delayed

GeV emission seen in some GRB (Hurley et al.,

1994). (For an alternative invoking photomeson

processes from ejecta protons impacting a nearbybinary stellar companion, see Katz, 1994.) The

GeV photon emission from the long-term IC

component in external afterglow shocks has been

considered by (Derishev et al., 2001; Dermer et al.,

2000; Zhang and M�esz�aros, 2001). The IC GeV

photon component is likely to be significantly

more important (Zhang and M�esz�aros, 2001) thana possible proton synchrotron or electron syn-chrotron component at these energies. Another

possible contributor at these energies may be p0

P. M�esz�aros et al. / New Astronomy Reviews 48 (2004) 445–451 447

decay from pc interactions between shock-accel-

erated protons and MeV or other photons in the

GRB shock region (B€ottcher and Dermer, 1998;

Fragile et al., 2002; Totani, 1998). However, under

the conservative assumption that the relativistic

proton energy does not exceed the energy in rela-tivistic electrons or in c-rays, and that the proton

spectral index is )2.2 instead of )2, both the

proton synchrotron and the pc components can be

shown to be substantially less important at GeV–

TeV than the IC component (Zhang and

M�esz�aros, 2001). Another GeV photon compo-

nent is expected from the fact that in a baryonic

GRB outflow neutrons are likely to be present,and when these decouple from the protons, before

any shocks occur, pn inelastic collisions will lead to

pions, including p0, resulting in UHE photons

which cascade down to the GeV range (Derishev

et al., 1999b, 2000). The final GeV spectrum results

from a complex cascade, but a rough estimate in-

dicates that 1–10 GeV flux should be detectable

with GLAST for bursts at zK 0:1 (Bahcall andM�esz�aros, 2000).

In these models, due to the high photon densi-

ties implied by GRB models, cc absorption within

the GRB emission region must be taken into ac-

count (see also Baring, 2000, 2001). So far, these

calculations have been largely analytical. One in-

teresting result is that the observation of photons

up to a certain energy, say 10–20 GeV withEGRET, puts a lower limit on the bulk Lorentz

factor of the outflow, from the fact that the com-

pactness parameter (optical depth to cc) is directlyproportional to the comoving photon density, and

both this as well as the energy of the photons de-

pend on the bulk Lorentz factor. This has been

used by Lithwick and Sari (2001) to estimate lower

limits on C in the range 300–600 for a number ofspecific bursts observed with EGRET.

At higher energies, a tentative J 0.1 TeV de-

tection at the 3r level of GRB970417a has been

reported with the water Cherenkov detector Mil-

agrito (Atkins et al., 2000). Another possible TeV

detection (Poirier et al., 2003) of GRB971110 has

been reported with the GRAND array, at the 2:7rlevel. Stacking of data from the TIBET array for alarge number of GRB time windows has led to an

estimate of a � 6r composite detection significance

(Amenomori et al., 2003). However, later analyses

by the same group do not find statistically signif-

icant individual detections (Amenomori et al.,

2001). Better sensitivity is expected from the up-

graded larger version of MILAGRO, as well as

from atmospheric Cherenkov telescopes underconstruction such as VERITAS, HESS, MAGIC

and CANGAROO-III (Weekes, 2000). However,

GRB detections in the TeV range are expected

only for rare nearby events, since at this energy the

mean free path against cc absorption on the dif-

fuse IR photon background is � few hundred Mpc

(Coppi and Aharonian, 1997, 2002). The mean free

path is much larger at GeV energies, and based onthe handful of GRB reported in this range with

EGRET, several hundred should be detectable

with large area space-based detectors such as

GLAST (Gehrels and Michelson, 1999, 2001). See

Fig. 1.

3. UHE neutrinos from GRB

GRBs, like supernovae, also generate large

amount of thermal (�10–30 MeV) neutrinos from

its stellar environment. However, these neutrinos,

typically from redshift z � 1, are hard to detect

because of tiny neutrino–nucleon cross section at

these energies. Increasing neutrino–nucleon cross

section with energy makes high energy (>TeV)neutrino detection possible in kilometer scale ice

or water Cherenkov detectors, such as ICECUBE

or ANTARES.

A mechanism leading to lower (GeV) energy

neutrinos in GRB is inelastic nuclear collisions.

Proton–neutron inelastic collisions are expected, at

much lower radii than where shocks occur, due to

the decoupling of neutrons and protons in thefireball or jet (Derishev et al., 1999a). If the fireball

has a substantial neutron/proton ratio, as expected

in most GRB progenitors, the collisions become

inelastic and their rate peaks at when the nuclear

scattering time becomes comparable to the ex-

pansion time. This occurs when the n and p fluids

decouple, their relative drift velocity becoming

comparable to c, which is easier due to the lack ofcharge of the neutrons. Inelastic n; p collisions thenlead to charged pions and GeV muon and electron

Fig. 1. Left panel: Parameter regimes for GRB GeV emission dominated by proton synchrotron (I), electron synchrotron (III) and

electron inverse Compton (II). Right panel: GeV lightcurves for the three types of bursts in the range 400 MeV–200 GeV. The solid

curves indicate bursts in regimes I (p-sy), II (e-IC) and III (e-sy), at a distance z ¼ 1. The three dotted unmarked curves are the same

but located at z ¼ 0:1. Also shown are the sensitivity curves for EGRET and GLAST.

448 P. M�esz�aros et al. / New Astronomy Reviews 48 (2004) 445–451

neutrinos (Bahcall and M�esz�aros, 2000). The earlydecoupling and saturation of the n also leads to a

somewhat higher final p Lorentz factor (Bahcall

and M�esz�aros, 2000; Derishev et al., 1999a; Fuller

et al., 2000; Pruet, 2003), implying a possible re-

lation between the n=p ratio and the observablefireball dynamics, relevant for the progenitor

question. Fusion and photodissociation processes

also play a role in such effects (Lemoine, 2002,

2003). Inelastic p; n collisions leading to neutrinos

can also occur in fireball outflows with transverse

inhomogeneities in the bulk Lorentz factor, where

the n can drift sideways into regions of different

bulk velocity flow, or in situations where internalshocks involving n and p occur close to the satu-

ration radius or below the photon photosphere

(M�esz�aros and Rees, 2000). The typical n; p neu-

trino energies are in the 1–10 GeV range, which

could be detectable in coincidence with observed

GRBs for a sufficiently close photo-tube spacing in

km3 detectors such as ICECUBE (Bahcall and

M�esz�aros, 2000).High energy (J100 TeV) neutrinos are pro-

duced mainly from photo–meson (pc) interactions.The most straightforward is the interaction be-

tween observed MeV photons produced by syn-

chrotron radiation from electrons accelerated in

internal shocks and relativistic protons accelerated

by the same shocks (Waxman and Bahcall, 1997),

leading to charged pions, muons and neutrinos.

The condition for pc interaction at D resonance

threshold in the fluid frame is �p�c J 0:2 GeV2c2,where c is the relativistic Lorentz factor of the fluidmotion. For 1 MeV photons this implies J 1016

eV protons, and neutrinos with �5% of that en-

ergy, �m J 1014 eV in the observer frame. The

typical spectrum of neutrino energy per decade is

�2mUm � constant for ��2p injected proton spectrum.

Synchrotron and adiabatic losses limit the muon

lifetime (Rachen and M�esz�aros, 1998), leading to a

suppression of the neutrino flux above �m � 1016

eV. In external shocks another copious source of

targets are the O/UV photons in the afterglow

reverse shock (e.g., as deduced from the GRB

990123 prompt flash of Akerlof et al., 1999). In

this case higher energy protons undergo pc inter-

actions, leading to neutrinos of 1017–1019 eV

(Waxman and Bahcall, 1999a; Vietri, 1998). These

neutrinos are expected to be detectable in ICE-CUBE (Halzen, 2000). Useful limits on their total

contribution to the diffuse ultra-high energy neu-

trino flux can be derived from observed cosmic ray

and diffuse c-ray fluxes (Waxman and Bahcall,

1999b; Bahcall and Waxman, 2001; Mannheim,

P. M�esz�aros et al. / New Astronomy Reviews 48 (2004) 445–451 449

2001). While the pc interactions leading to J100

TeV energy neutrinos provide a direct probe of the

internal shock acceleration process, the J10 PeV

neutrinos would probe the reverse external shocks.

The most intense neutrino signals may be due to

interactions occurring inside collapsars while thejet is still burrowing its way out of the star

(M�esz�aros and Waxman, 2001), before it has had a

chance to break through (or not) the stellar enve-

lope to produce a GRB outside of it. While still

inside the star, the buried jet produces thermal X-

rays at �1 keV which interact with J105 GeV

protons which could be accelerated in internal

shocks occurring well inside the jet/stellar envelopeterminal shock, producing J few TeV energy

neutrinos for tens of seconds, which penetrate the

envelope. The most conspicuous contribution of

TeV ml is due to pp; pn collisions involving the

buried jet relativistic protons and thermal nucleons

in the jet and the stellar envelope (Razzaque et al.,

2003b), while pc is important at higher energies

J100 TeV (see Fig. 2). At TeV energies thenumber of neutrinos is larger for the same total

energy output, which improves the detection sta-

tistics. Based on a total rate of 103/year the rare

bright, nearby or high c collapsars could occur at

the rate of �1/year, including both c-ray bright

3 4 5 6 7 8 9–13

–12

–11

–10

–9

–8

–7

WB Limit

ICECUBE

burst

pp

p (head)γ

γ (shocks)p

p (head)γ

νµ flux (10 bursts/yr)3

atmospheric

ν

[log1

0(G

eV/c

m s

sr)

]2

E [log10(GeV)]

opε

E2 νΦ

ν-1

(stellar)

pp (shocks)

HHe

Fig. 2. Left panel: Diffuse ml fluxes (E2mUm�

�1op ) from sub-stellar jet sh

shown as solid lines. These neutrinos arrive as precursors (10–100 s) of

is the diffuse flux of ml�s arriving simultaneously with the c-rays from(WB) diffuse cosmic ray bound (light long-dashed curves); and the a

pothetical 100:1 ratio of c-ray dark (in which the jets do not emerge) to

be 100 times higher. Right panel: Diffuse ml flux (E2mUm) from post-sup

that (top curve) all observed GRBs have an SNR shell, or (bottom) 10

correspond to the WB cosmic-ray limit, short dashed curves are the d

GRBs (where the jet broke through the envelope)

and c-ray dark events where the jet is choked

(failed to break through), and both such c-brightand dark events could have a TeV neutrino fluence

of � few neutrinos/burst, detectable by ICECUBE

in individual bursts. A specific example (Razzaqueet al., 2003c) has been calculated based on the

recent nearby z � 0:17 burst GRB 030329.

Recent reports of detection of X-ray lines from

several GRB afterglows have been interpreted

(Piro et al., 2000), although not unambiguously, as

providing support for a version of the collapsar

model in which a SN explosion occurs weeks be-

fore the GRB (the ‘‘supranova’’ model, Vietri andStella, 1998). In the supranova scenario the MHD

wind of a pre-GRB pulsar (Granot and Guetta,

2003a,b) and/or protons escaping pc interactions

in the GRB prompt phase (Razzaque et al., 2003a)

may interact with nucleons (pp) of the supernova

remnant (SNR) shell as well as photons (pc)trapped within (other nucleon targets from a stel-

lar companion disruption leading to p0 decay GeVc-rays were discussed by Katz, 1994). We have

calculated the expected m-fluxes and l-event ratesfrom individual bursts as well as the diffuse con-

tribution from all bursts (Razzaque et al., 2003a).

The neutrinos produced by pp and pc interactions

3 4 5 6 7 8 9 10–11

–10

–9

–8

−7

[log1

0(G

eV/c

m s

sr)

]2

WB Limit

νΦ

E2 ν

Burst

Afterglow

Supranova

E [log10(GeV)]ν

ocks in two GRB progenitor models r12:5 (H) and r11 (He) are

c-ray bright (electromagnetically detectable) bursts. Also shown

observed GRB (dark short-dashed curve); the Waxman–Bahcall

tmospheric neutrino flux (light short-dashed curves). For a hy-

c-ray bright collapses, the solid neutrino spectral curves would

ernova (supra-nova) models of GRBs (solid curves), assuming

% of all observed GRBs have an SNR shell. Long dashed lines

iffuse m flux from GRB internal shocks and afterglows.

450 P. M�esz�aros et al. / New Astronomy Reviews 48 (2004) 445–451

between GRB relativistic protons and SNR shell

target protons and photons will be contempora-

nious with the GRB electromagnetic event. The

high pp optical depth of the shell also implies a

moderately high average Thomson optical depth

sT / t�2d of the shell, dropping below unity after

�100 days. Large scale anisotropy as well as

clumpiness of the shell will result in a mixture of

higher and lower optical depth regions being ob-

servable simultaneously, as required in the supra-

nova interpretation of X-ray lines and photon

continua in some GRB afterglows (Piro et al.,

2000). Depending on the fraction of GRB with

SNR shells, the contributions of these to the GRBdiffuse neutrino flux has a pp component which is

relatively stronger at TeV–PeV energies than the

internal shock pc component of Waxman and

Bahcall (1997), and a shell pc component which is

a factor 1 (0.1) of the internal shock pc component

(Fig. 2) for a fraction 1 (0.1) of GRB with SNR

shells. Due to a higher synchrotron cooling break

in the shell, at Em J 1017 eV the shell componentcould compete with the internal (Waxman and

Bahcall, 1997) and afterglow (Waxman and Bah-

call, 1999a) components. These neutrino fluxes

provide a test for the supranova hypothesis, the

predicted event rates being detectable with kilo-

meter scale planned Cherenkov detectors.

Acknowledgements

This research has been partially supported

through NASA NAG5-9192 and NSF AST-

0098416.

References

Akerlof, C.W. et al., 1999. Nature 398, 400.

Amenomori, M. et al., 2001. In: AIP Conf. Proceedings, vol.

558, p. 844.

Amenomori, M. et al., 2003. A&A 311, 919.

Atkins, R. et al., 2000. ApJ 533, L119.

Bahcall, J.N., M�esz�aros, P., 2000. Phys. Rev. Lett. 85, 1362.

Bahcall, J.N., M�esz�aros, P., 2000. Phys. Rev. Lett. 85, 1362.

Bahcall, J.N., Waxman, E., 2001. Phys. Rev. D 64, 023002.

Baring, M.G., GeV–TeV Gamma-Ray Astrophysics Work-

shop. In: Dingus, B.L. et al. (Ed.), AIP Conf. Proceedings,

vol. 515, p. 238.

Beloborodov, A.M., 2003. ApJ 588, 931.

B€ottcher, M., Dermer, C.D., 1998. ApJ 499, L131.

Coppi, P., Aharonian, F., 1997. ApJ 487, L9.

Costa, E. et al., 1997. Nature 387, 783.

de Jager, O.C., Stecker, F.W., 2002. ApJ 566, 738.

Derishev, E.V., Kocharovsky, V.V., Kocharovsky, Vl.V.,

1999a. ApJ 521, 640.

Derishev, E.V. et al., 1999b. ApJ 521, 640.

Derishev, E.V., Kocharovsky, V.V., Kocharovsky, Vl.V., 2001.

A&A 372, 1071.

Dermer, C.D., Chiang, J., Mitman, K.E., 2000. ApJ 537, 785.

Finn, L.S., Mohanty, S.D., Romano, J.D., 1999. Phys. Rev. D

60, 121101.

Fishman, G.J., Meegan, C.A., 1995. ARA&A 33, 415.

Fox, D.W. et al., 2003a. Nature 422, 284.

Fox, D.W. et al., 2003b. ApJ 586, L5.

Fragile, P.C., et al., 2002. ApJ, submitted. Available from

<astro-ph/0206383>.

Frail, D. et al., 1997. Nature 389, 261.

Fuller, G., Pruet, J., Kevork, A., 2000. Phys. Rev. Lett. 85,

2673.

Gehrels, N., Michelson, P., 1999. Astropart. Phys. 11, 277.

Granot, J., Guetta, D., 2003a. Phys. Rev. Lett. 90, 1102.

Guetta, D., Granot, J., 2003b. Phys. Rev. Lett. 90, 1103.

Halzen, F., 2000. In: Weak Interactions and Neutrinos,

Proceedings of the 17th International Workshop. World

Scientific, Singapore, p. 123.

Harris, M.J., Share, G.H., 1998. ApJ 494, 724.

Heise, J. et al., 2001. In: Costa, E., Frontera, F. (Eds.), Gamma-

ray Bursts in the Afterglow Era, p. 16.

Hurley, K. et al., 1994. Nature 372, 652.

Katz, J., 1994. ApJ 432, L27.

Kippen, R.M., Woods, P.M., Heise, J., in�tZand, J.J.M., Briggs,

M.S., Preece, R.D., 2002. preprint. Available from <astro-

ph/0203114>.

Kobayashi, S., 2000. ApJ 545, 807.

Kulkarni, S. et al., 2000. In: Kippen, M., Mallozzi, R., Fishman,

G. (Eds.), Gamma-ray Bursts. AIP Conf. Proc., vol. 526.

AIP, New York, p. 277.

Lemoine, M., 2002. A&A 390, L31.

Letessier-Selvon, A., 2003. In: 20th International Conference

on Neutrino Physics and Astrophysics (Neutrino 2002),

Munich, Germany. Available from <astro-ph/0208526>.

Li, W. et al., 2003. ApJ 586, L9.

Lithwick, Y., Sari, R., 2001. ApJ 555, 540.

Mannheim, K., 2001. In: Aharonian, F, V€olk, H (Eds.),

Heidelberg 2000, High Energy Gamma-ray Astronomy.

In: AIP Conf. Proceedings, vol. 558. AIP, New York, p.

417.

M�esz�aros, P., Rees, M.J., 1993. ApJ 405, 278.

M�esz�aros, P., Rees, M.J., 1994. MNRAS 269, L41.

M�esz�aros, P., Rees, M.J., Papathanassiou, H., 1994. ApJ 432,

181.

M�esz�aros, P., Rees, M.J., 1997. ApJ 476, 232.

M�esz�aros, P., Rees, M.J., 1999. MNRAS 306, L39.

M�esz�aros, P., Rees, M.J., 2000. ApJ 541, L5.

M�esz�aros, P., Waxman, E., 2001. Phys. Rev. Lett. 87, 171102.

P. M�esz�aros et al. / New Astronomy Reviews 48 (2004) 445–451 451

Metzger, M. et al., 1997. Nature 387, 878.

Papathanassiou, H., M�esz�aros, P., 1996. ApJ 471, L91.

Piro, L. et al., 2000. Science 290, 955.

Poirier, J. et al., 2003. Phys. Rev. D 67, 2001.

Pruet, J., 2003. ApJ 591, 1104.

Rachen, J., M�esz�aros, P., 1998. Phys. Rev. D 58, 123005.

Razzaque, S., M�esz�aros, P., Waxman, E., 2003a. Phys. Rev.

Lett. 90, 1103.

Razzaque, S., M�esz�aros, P., Waxman, E., 2003b. Phys. Rev. D

68, 3001.

Razzaque, S., M�esz�aros, P., Waxman, E., 2003c. Phys. Rev. D,

in press. Available from <astro-ph/0308239>.

Sari, R., Piran, T., 1999. ApJ 517, L109.

Totani, T., 1998. ApJ 502, L13.

van Paradijs, J. et al., 1997. Nature 386, 686.

Vietri, M., 1998. Phys. Rev. Lett. 80, 3690.

Vietri, M., Stella, L., 1998. ApJ 507, L45.

von Kienlin, A. et al., 2003. In: SPIE Conf. Proc., 4851, X-ray

and Gamma-ray Telescopes and Instruments for Astron-

omy, p. 1336.

Waxman, E., Bahcall, J.N., 1997. Phys. Rev. Lett. 78, 2292.

Waxman, E., Bahcall, J.N., 1999a. ApJ 541, 707.

Waxman, E., Bahcall, J.N., 1999b. Phys. Rev. D. 59, 023002.

Weekes, T., 2000. In: Aharonian, F, V€olk, H (Eds.), Heidelberg

2000, High Energy Gamma-ray Astronomy. In: AIP Conf.

Proc., 558. AIP, NY, p. 15.

Zhang, B., M�esz�aros, P., 2001. ApJ 559, 110.

Zhang, B., Kobayashi, S., M�esz�aros, P., 2003. ApJ 596, 950.