Introduction

Over 30 years after their discovery, gamma-ray bursts (GRBs) are still among the greatest mysteries of modern astrophysics. Our understanding of the GRB phenomenon has made rapid progress over the past ~ 10 years with the detection of X-ray, optical, and radio afterglows, the identification of host galaxies and more than a dozen redshift determinations, which led to the ultimate establishment of their cosmological distance scale (at least for the subclass of GRBs with durations longer than ~ 1 second, for which optical afterglows have only been localized so far). However, the nature of GRB progenitors and the primary mechanism driving GRB explosions are still largely unknown. This is mainly because the now well-understood afterglow phase of GRBs, in which a relativistic blastwave is believed to plough into a dilute surrounding medium, continuously energizing swept-up particles and thereby slowing down its own bulk velocity, is only weakly dependent on the details of the event that produced the blast wave.

One way of unveiling the nature of GRB progenitors is through diagnostics of the environment of GRBs. In a few cases, there have now been positive detections of iron emission lines in the X-ray afterglow. These allow rather stringent constraints on the location and thus the structure and geometry of the GRB environment. The first such detection was reported for GRB 970508 (see Fig. 1 for an image of the optical afterglow of this burst and Fig. 2 for a model calculation to reproduce the observed iron emission line in its X-ray afterglow). Three more iron emission line detections have been reported since then. In the following section on modeling of the iron lines in GRB afterglows a brief overview of our recent efforts on a theoretical interpretation of these line features will be given.

In one case, namely GRB 990705, the BeppoSAX team has also reported the detection of a time-dependent absorption feature. We had previously modelled expected time-dependent photoelectric absorption as a possible diagnostic of GRB environments, and have done extensive research on the interpretation of the particular feature seen in GRB 990705, which, however, turned out to be rather problematic in the context of the currently most widely discussed GRB models. This work is summarized in the section on X-ray absorption signatures from GRBs.

Important diagnostics about the matter distribution in the vicinity of GRBs can also be deduced from the spectra and spectral evolution in the prompt GRB phase and the early X-ray and broadband afterglows since those depend on the overall density and possible inhomogeneities of the material surrounding the GRB progenitor, on the beaming of the relativistic material ejected during the GRB explosion, and on the energetics of the primary event, all of which may give hints towards the nature of the progenitor. We have done extensive work on gamma-ray burst spectra and spectral evolution predicted in the external-shock model for GRBs and their afterglows as well as on alternative radiation mechanisms to produce the prompt emission from GRBs. Using the specific predictions of the internal shock model, we have done a detailed study of the statistics of cosmological GRBs by fitting simultaneously the statistical distributions of various GRB quantities to the observed ones (see Fig. 3). For details on these projects see the section on GRB spectral models and statistics.

Gamma-ray bursts are also known to emit high-energy gamma radiation since several GRBs have been detected by the EGRET instrument on board the Compton Gamma-Ray Observatory at energies > 100 MeV. The highly relativistic nature of the outflows most likely associated with GRBs also suggests that they may be a source of ultra-high energy cosmic rays (UHECRs). Predictions about the high-energy emission from GRBs are particularly interesting in view of the upcoming GLAST satellite mission which will carry out a comprehensive high-energy gamma-ray monitoring program, providing very promising prospects of detecting a large number of GRBs at GeV energies. In the course of our spectral modeling efforts, described in the section on GRB spectral models and statistics, we have pointed out that Compton scattering of synchrotron photons (the SSC mechanism) in relativistic shocks can very plausibly produce a significant level of very-high energy emission in GRBs. In the section on high-energy emission from GRBs, we will focus on alternative mechanisms for producing MeV and GeV emission in GRBs, which may also be related to the production of UHECRs.

Using part of Ohio University's share of the MDM Observatory, we are contributing to rapid follow-up observations of optical afterglows of GRBs. Recent results from such observations are briefly summarized in the Section ``Observations of Optical Afterglows''.


Fig. 1: The optical afterglow of GRB 970508, the first GRB for which an optical and radio counterpart have been detected, and a redshift could be measured. The X-ray afterglow of this burst also allowed the first (marginal) detection of an iron emission line by the BeppoSAX satellite. Credit: P. Groot et al.


Fig. 2: Modeling result of the X-ray afterglow of GRB 970508, including the time-dependent iron emission line (from Böttcher 2000). Panel a) shows the broadband X-ray afterglow in two continuum energy channels (black curves: solid: 40 - 50 keV; dotted: 6.4 - 6.7 keV) and the modeled isotropic luminosity in the iron K-alpha emission line at different observing angles w.r.t. the symmetry axis of a thick torus around the burst source, in which the line emission originates. Panel b) shows the modeled X-ray spectra at three different times after the onset of the burst. For more details see the section on modeling of the iron lines in GRB afterglows.


Fig. 3: Cosmological model fits to the observed peak flux, Epk, and duration distribution (long bursts only) from the BATSE 3B catalog, and the predicted redshift distribution, using the specific predictions about spectra and spectral evolution of GRBs in the framework of the external synchrotron shock model (from Böttcher & Dermer 2000a).

Modeling the Iron Lines in GRBs

Cosmological GRBs emit short, but extremely energetic flashes of high-energy (X-ray and gamma-ray) radiation. Clearly, such bursts will rapidly photoionize any material in the vicinity of the GRB. With the discovery of X-ray afterglows of GRBs in 1997 by the BeppoSAX satellite, the question arose as to whether the X-ray signatures from such time-dependent photoionization in GRB environments might be observable with current or future X-ray telescopes. This would include time-dependent photoelectric absorption features, but possibly also fluorescence line emission. Our group did some of the pioneering work on model calculations of the time-dependent photoionization and radiation transport associated with gamma-ray bursts (e.g., Böttcher et al. 1999).

As a first step, we investigated a completely isotropic scenario, i.e. an isotropic GRB explosion interacting with an isotropic environment, which could, for example, be provided by a moderately dense cloud of a star-forming region. This was also motivated by the increasing success of GRB models based on supernova-like explosions of very massive stars, which end their lives a relatively short time (a few million years) after their birth, so that they might still be located within - or at least very close to - the star-forming region in which they were formed. We have solved the time-dependent photoionization and radiation transfer problem in such an environment numerically and made interesting predictions about time-dependent photoelectric absorption features expected in this case (see the next section). However, it turned out that in such isotropic scenarios it was not possible to produce an iron K-alpha emission line strong enough to be observable by any current or planned X-ray satellite (see Fig. 4). This was in good part because the amount of material that can be located in the line of sight between the GRB and the observer is limited to be Thomson thin (i.e., the column depth NH has to be less than ~ 1024 cm-2) because otherwise one would observe strong absorption features which have not been seen (except for the peculiar case of GRB 990705, see the next section), and Thomson scattering of the burst radiation may lead to a smearing of the short-term (millisecond) variability, in contradiction to the observed, very rapid variability of many GRBs.

For this reason, modeling efforts for the iron K-alpha line, which had in the meantime been detected in GRB 970508 and GRB 970828 (due to an inconsistency with the redshift of that line detection with the later determined redshift of the host galaxy of this burst, this detection had temporarily been retracted, but recently been re-evaluated, invoking a recombination edge instead of a fluorescence line) soon concentrated on anisotropic scenarios. In Böttcher (2000), I had modelled the time-dependent radiation transport in the case of a dense torus surrounding the GRB progenitor (see Fig. 5 for a sketch of the geometry). Such a torus is an idealized geometry for a configuration that could be provided, e.g., through an anisotropic wind ejected by the GRB progenitor star or through material ejected during a common-envelope phase of two merging stars, preceding the GRB. Those simulations also took into account the hydrodynamic interaction of the relativistic blast wave likely associated with the GRB, with the pre-existing, dense torus. As shown in Fig. 2, this model was capable of reproducing the iron emission line feature seen in GRB 970508 very well.

The next step was now to tie such time-dependent radiation transport calculations to specific GRB progenitor models and investigate inhowfar they would provide the environments required to model the observed iron lines. This was done in Böttcher & Fryer (2001), where we made specific predictions, based on the collapsar / hypernova and the He-merger models for GRBs. Both these GRB models most likely require that the progenitor system goes through a common envelope phase during which most of the hydrogen envelope of the secondary (the supernova explosion of which will ultimately cause the GRB) is ejected into a disk-like structure. In the case of the He-merger scenario, the delay between the ejection of the hydrogen envelope and the GRB may be as short as ~ 10,000 years, and the disk ormed by the ejection of the hydrogen envelope is expected to be located at a typical distance of ~ 1013 cm from the center of the GRB explosion. In this case (as in the collapsar / hypernova scenario), the GRB will be associated with a supernova, which ejects ~ 1 solar mass of material at non-relativistic speeds quasi-isotropically. Only along the symmetry axis of the orbit of the binary system (and the pre-ejected disk) does this event appear as a GRB. A typical result of a simulation of the interaction between the supernova ejecta and the pre-ejected disk is shown in Fig. 6. Such a model setup is very well suited to reproduce the iron emission line features seen in some GRBs. Fig. 7 shows a specific model fit to the iron line seen in GRB 000214. Here, we have assumed that the GRB was caused by the explosion of a 25 M0 secondary, with the inner edge of the pre-ejected disk being located at 6×1013 cm from the secondary. In the course of the supernova explosion, 2 M0 are being ejected isotropically at a velocity of 2×109 cm/s.

In the case of the collapsar / hypernova scenario, one expects that the delay between the common-envelope phase and the GRB explosion is significantly longer, most likely in excess of ~ 100,000 yr. Thus, the pre-ejected disk is expected to be located at least ~ 1015 cm from the secondary. Consequently, there will be a significantly longer delay between the GRB and the hydrodynamic interaction of the (non-relativistic) supernova ejecta with the pre-ejected disk, of the order of a few weeks or longer. Fig. 8 shows a typical simulation result for such a case. An interesting prediction from our calculations is that with the currently operating Chandra and XMM-Newton X-ray telescopes, quasi-thermal X-ray flashes from these interactions between supernova ejecta and pre-ejected disks should be observable out to redshifts of ~ 1.


Fig. 4: Maximum isotropic luminosities in the iron K-alpha fluorescence line from a standard GRB in an isotropic environment with solar-system element abundances. The total column density is held constant as the density of the material in the environment is increased (accordingly, the size scale of the cloud containing this material is decreased). The resulting fluorescence line luminosity reaches a maximum more than 2 orders of magnitude lower than typically observed in the 4 cases of positive iron K-alpha emission line detections so far. For more details, see Böttcher et al. (1999).


Fig. 5: Sketch of the torus geometry assumed for the model calculations of the iron line in GRB 970508 (see Fig. 2) in Böttcher (2000).


Fig. 6: Model simulation of the interaction of a non-relativistic supernova/GRB shock wave interacting with a disk of pre-ejected material from a common-envelope phase in the He-merger GRB model. Panel a) shows the average temperature of the shocked material as a function of time; panel b) shows snap-shot X-ray spectra at several times; panel c) shows the isotropic luminosity in the Fe K-alpha line emitted by the shocked material. From Böttcher & Fryer(2001).


Fig. 7: Model simulation to reproduce the iron K-alpha emission line observed in GRB 000214, using the code of Böttcher & Fryer(2001). The top panel shows the X-ray spectra from the shocked disk only; the middle panel shows those spectra added to the underlying power-law continuum of the afterglow of GRB 000214. The lower level compares the modelled isotropic luminosity of the Fe K-alpha emission line to the observed value (shaded area) and the t-1.2 decay law seen in the continuum X-ray afterglow.


Fig. 8: Model simulation of the interaction of a non-relativistic supernova/GRB shock wave interacting with a disk of pre-ejected material from a common-envelope phase in the collapsar / hypernova GRB model. Panel a) shows the average temperature of the shocked material as a function of time; panel b) shows snap-shot X-ray spectra at several times; panel c) shows the isotropic luminosity in the Fe K-alpha line emitted by the shocked material. From Böttcher & Fryer(2001).

X-ray Absorption Signatures from GRBs

As briefly mentioned above, the first calculations of the time-dependent radiation transport and photoionization in homogeneous GRB environments yielded interesting predictions in terms of absorption features, but did not predict any observable iron lines. In Böttcher et al. (1999), we had found that X-ray absorption edges due to intervening material along the line of sight in the vicinity of the GRB are either basically unaffected by photoionization and thus not time-dependent (if most of the material is located too far away from the burst to be photoionized), or it is being photoionized on sub-minute time scales, i.e. typically within the prompt phase of the GRB. This was an important prediction which called for prompt X-ray observations of GRBs with moderate spectral resolution since time-independent absorption features can not confidently be distinguished from foreground absorption not associated with the host galaxy of the GRB, and can thus only provide a lower limit on the redshift of the GRB. With the advent of the Swift satellite mission, currently scheduled for launch in 2003, such prompt X-ray observations of GRBs might become reality.

However, excitingly, the BeppoSAX team has recently reported the discovery of a time-dependent absorption feature in the prompt X-ray emission of GRB 990705, which was consistent with an iron K edge at the most likely redshift of the burst, z = 0.86, persisting for the first ~&160;13 seconds of the GRB; in the following time segments of the BeppoSAX Wide Field Camera observation, no evidence for excess photoelectric absorption was found. We have thus used the radiation transfer code developed in Böttcher et al. (1999) to model this absorption feature and deduce the distribution of the absorbing material, assuming that it is primarily due to photoelectric absorption in a gradually photoionized GRB environment. Our modeling results are illustrated in Fig. 10. They agree well with analytic estimates of the photoionization time scale, which yields an estimate of the distance of the absorbing material from the burst source, and thus allows an estimate of the amount of iron required to produce the absorption feature. Unfortunately, the deduced amount of iron would be ~ 44 M0/sr within ~ 1.3 pc, which implies substantially more iron than could plausibly be produced in any well-understood astrophysical context, even if the iron is distributed anisotropically around the burst source.

As alternative scenarios, absorption in very dense clouds has been discussed, in which recombination would be sufficiently efficient to balance photoelectric absorption over a limited time. Parameters can be adjusted so that Compton heating by the GRB radiation renders recombination inefficient around the time when the absorption feature is observed to vanish (see, e.g., Böttcher et al. 2001). However, this scenario requires a very large degree of clumping and extremely large density contrasts (of order 105 - 106) in the GRB environment which are unlikely to be realized. Another alternative, invoking resonance scattering out of the line of sight by Fe XXVI in a clumpy, high-velocity outflow from the burst source, faces a similar problem of a large degree of clumping and requires a large amount of energy (~ 1053 ergs) deposited into this outflow (prior to the GRB).

Focusing on scenarios interpreting the absorption feature as being due to photoelectric absorption, in Böttcher, Fryer, & Dermer (2001) we have investigated possible GRB progenitor models in terms of their ability to produce an environment suitable to produce X-ray absorption features as seen in GRB 990705. Primary candidates were the He-merger and the supranova model, both of which involve a supernova explosion several months to several years before the GRB. The results are summarized in Fig. 11. It appears that the supranova model may be the most natural way to produce such a GRB environment, although there is some concern by theorists about its efficiency to actually produce a GRB. The He-merger model, predicting longer delays between the supernova and the GRB, leads to a very low probability of observing significant X-ray absorption features. Thus, we are anxiously awaiting the availability of a larger data base of prompt X-ray observations of GRBs in order to place more stringent constraints on the probability of the occurrence of such absorption features.


Fig. 9: Time evolution of the depth of the Fe K absorption edge through photoionization of a homogeneous GRB environment by the prompt and early afterglow GRB radiation. Standard solar-system element abundances are assumed. From Böttcher et al. (1999).


Fig. 10: Model fit to the transient Fe K absorption edge observed in GRB 990705. The observed depth of the edge is indicated by the shaded regions; the red solid line is the best fit assuming a moderately dense shell of absorbing material around the GRB source. The numbers in the legend are the hydrogen column density NH in units of 1022 cm-2 (assuming an iron overabundance of 75× the standard solar-system value), the radius of the absorbing shell in units of 1018 cm, and the matter density in the shell in units of cm-3. Although the red curve seems to be a good fit, the parameters deduced from this model setup require an implausibly large amount of iron. From Böttcher et al. (2001).


Fig. 11: Parameter space of supernova ejecta mass Mcl concentrated in clumps, and the time delay between the primary's supernova explosion and the GRB. The solid lines indicate the condition that an iron K absorption edge of the depth observed in GRB 990705 is produced, for various values of the ratio of the probability Pcll.o.s. of an absorbing cloud being located in the line of sight, to the iron enhancement XFe w.r.t. standard solar-system values. Constellations which would give a consistent physical scenario, must either be located close to the vertical line corresponding to tion (z) = 6 s (if recombination is inefficient), or within the shaded regions in the upper left corner of the plot, which indicates the condition trec < tion for volume filling factors of the SN ejecta of 1 % and 0.1 %, respectively, 1 year after the SN (from Böttcher, Fryer, & Dermer 2001).

GRB spectral models and statistics

While the afterglow phase of GRBs is now rather well understood in terms of the relativistic blast wave model. In this model, a highly relativistic outflow is expanding into a surrounding medium, continuously accelerating particles at the relativistic (external) shock wave which it forms when interacting with the environment, and thereby gradually decelerating its bulk motion. Energetic electrons behind the shock front are emitting synchrotron radiation, the time dependent spectral properties of which very well match observed GRB afterglow observations. However, the prompt phase of GRBs is currently much less well understood, and it is still not clear what the dominant radiation mechanism is and how particles are primarily energized to produce the observed radiation.

Most researchers now believe that during the prompt GRB phase, internal shocks are primarily responsible for the energy dissipation and particle acceleration, leading to the observable prompt GRB emission. In such a scenario, subsequent shells of material ejected from a central source collide with each other rather than with the external medium, as in the external shock scenario. However, an external shock scenario can not be ruled out for the prompt GRB phase, and it has the appealing feature that it would smoothly connect to the afterglow phase, in which external shocks are generally believed to dominate the overall dynamics and radiation processes. In Dermer, Chiang & Böttcher (1999) and Dermer, Böttcher & Chiang (1999), we had made detailed predictions about the spectra and spectral evolution expected in the external synchrotron shock scenario for GRBs.

The currently most successful models of GRBs are based on explosions of very massive stars, which are exploding in supernova (or GRB) events shortly (i.e. a few million years) after their formation. If those models are correct, one would expect that the cosmological distribution of GRBs would closely follow the star formation history of the Universe. Under this assumption, we have modelled the expected peak flux, peak energy, and duration distribution of GRBs (using the detailed external-shock model predictions mentioned above), in order to constrain GRB model parameters (in particular, the total explosion energy and the initial bulk Lorentz factor or baryon loading factor) required to reproduce the observed statistical distributions of GRBs (Böttcher & Dermer 2000a). The main result of those modeling calculations are shown in Fig. 3. The ability of the external shock model to reproduce all three of those statistical distributions simultaneously may be seen as support for the external shock model, although some issues with respect to the short-term variability of GRB pulses remain to be resolved in this model.

As mentioned above, the afterglow emission from GRBs is fairly well understood today, and its time-dependent broadband emission can be rather easily characterized by simple standard formulae if the relativistic blast wave is evolving in one of the two extreme limits of blast wave evolution: the non-radiative (or adiabatic) limit or the fully radiative regime. However, in the early afterglow phase, it is very likely that the blast wave is in an intermediate radiative regime, in which a non-negligible fraction of the energy going into swept-up particles is radiated away almost instantaneously. In Böttcher & Dermer (2000b), we have shown that in certain idealized cases, the blast wave dynamics can be solved analytically even in the case of such intermediate radiative regimes. We have presented analytical expressions for the spectra and light curves of GRB afterglows in arbitrary radiative regimes, which are relevant primarily to the early afterglow phase of GRBs. Figs. 11 and 12 illustrate some of the key results of that work, which show how sensitive the predictions of the blast wave model are to the value of the radiative efficiency. These predictions will become very important when sensitive early X-ray afterglow observations by the Swift mission become available after its launch which is scheduled for 2003.

While most GRB afterglow modeling work has concentrated on synchrotron emission as the dominant radiation process, we were among the first to point out the potential importance of Compton upscattering of synchrotron photons (the synchrotron self-Compton or SSC mechanism) during the afterglow phase of GRBs (Dermer, Böttcher & Chiang 2000). This process might not only significantly alter the broadband spectral shape of the afterglow emission and the light curves (due to the secondary bump of the SSC emission sweeping through a fixed observing frequency), but it may also provide a source of high-energy gamma-ray emission at energies beyond ~ 100 MeV. We have developed simple analytical fitting formulae and critically investigated their range of validity. Figs. 14 - 16 show some illustrative examples of broadband GRB afterglow spectra and light curves, including the SSC component. They demonstrate that, in particular in the case of a low magnetic-field equipartition factor (the case shown in Fig. 14) the SSC component can become quite dominant in the broadband afterglow spectrum and lead to a significant level of delayed high-energy emission.

A significant fraction of GRBs exhibits time-resolved X-ray and soft gamma-ray spectra at early times, which are harder than possible through optically thin synchrotron emission. For this reason, we are also investigating alternative emission mechanisms for GRBs, such as saturated Comptonization (e.g., Liang et al. 1999), optically thin Comptonization of synchrotron or external soft radiation etc. which may be able to produce harder low-energy spectra than the limiting +1/3 spectral slope (energy index) of optically thin synchrotron emission of a relativistic particle population with a low-energy cut-off.


Fig. 12: Early afterglow light curves at three different observing frequencies for three different values of the radiative efficiency of the blastwave, for standard GRB parameters given in Böttcher & Dermer (2000b).


Fig. 13: Temporal indices (F ~ tw) as a function of the radiative regime of the blastwave, for standard GRB parameters given in Böttcher & Dermer (2000b). The labels on the curves describe the spectral regime in which the respective temporal slopes are valid: 1/3 for the low-frequency regime (typically radio or IR for the early afterglow phase), -1/2 for intermediate frequencies (typically optical/UV), -p/2 (where p is the spectral index with which relativistic particles are injected [accelerated] into the post-shock region) for high frequencies (X-rays).


Fig. 14: Time-dependent GRB and afterglow spectra from a standard external-shock scenario, including synchrotron and SSC emission. The smooth curves are the results of detailed numerical simulations, which are being compared to simple analytic approximations. The labels on the curves denote the logarithm of the elapsed time in the observer's frame in seconds. From Dermer, Böttcher & Chiang (2000). One can see that the analytic approximations become more accurate at later times.


Fig. 15: Time-dependent GRB and afterglow spectra from a standard external-shock scenario, including synchrotron and SSC emission. The smooth curves are the results of detailed numerical simulations, which are being compared to simple analytic approximations. The labels on the curves denote the logarithm of the elapsed time in the observer's frame in seconds. From Dermer, Böttcher & Chiang (2000).


Fig. 16: GRB and afterglow light curves at 4 different observing frequencies, predicted by an external synchrotron shock scenario, including synchrotron and SSC emission. As in Figs. 13 and 14, the smooth curves result from detailed numerical simulations, which are being compared to simple analytic fit formulae. From Dermer, Böttcher & Chiang (2000).

High-Energy Emission from GRBs

It had been suggested that the highly relativistic blast waves most likely associated with GRBs are capable of accelerating ions up to very high energies, and that GRBs may thus be the source of ultra-high energy cosmic rays (UHECRs) up to energies beyond ~ 1019 eV. If this is true and a large number of protons with such energies is co-existing with the dense radiation field of the GRB, then one would expect that those UHECRs are undergoing frequent proton-photon interactions, resulting in the production of neutral and charged pions as well as electron-positron pairs. The subsequent pion decay as well as synchrotron emission from secondary particles and from the protons themselves should result in characteristic high energy emission (beyond ~ 100 MeV, out to TeV energies).

In Böttcher & Dermer (1998), we have calculated the expected spectra and light curves of the high-energy emission from these hadronic processes. We found that the overall luminosity of the emission initiated by UHECRs might be dominated by proton synchrotron emission. Interestingly, this high-energy emission is expected to decay more slowly than the prompt GRB and afterglow emission due to electron synchrotron radiation. Thus, such signatures may be observable through rather long pointed follow-up observations by high-energy instruments, such as GLAST or the new generation of air Cherenkov telescopes. Fig. 16 shows a typical time sequence of spectra resulting from hadronic interactions of UHECRs in the radiation field of a GRB, normalized to the observed spectral evolution of GRB 970508. The figure demonstrates that it might well be possible for GLAST in its pointing mode to detect such signatures.

As briefly mentioned in the previous section, a persisting problem for GRB emission models is that some bursts show time-resolved X-ray and soft gamma-ray spectra which are too hard to be consistent with optically thin synchrotron radiation, which is the favored mechanism in the framework of most current GRB models. A conceivable way out of this dilemma is the hypothesis that intervening material along the line of sight could preferentially scatter low-energy emission out of the line of sight and thus lead to a hardening of the observable X-ray and soft gamma-ray spectrum. In particular, when one accounts for the fact that intervening clouds will be very rapidly ionized and heated to electron temperatures of several 100  keV, the hot electrons in intervening clouds will additionally Compton upscatter soft photons and thus enhance the spectral hardening effect, as we have investigated in Dermer & Böttcher (2000).

However, if this is true, then one would also expect that some of the sideways-scattered GRB photons will interact with high-energy photons from the GRB emission via gamma-gamma absorption and pair production, thus enhancing the density of material along the line of sight (Dermer, Böttcher, & Liang 2001). This effect is illustrated in Fig. 18.

We have recently discovered another potential source of high-energy emission from GRBs, which may arise in the very early, possibly collision-dominated phase of GRB evolution. GRBs are generally believed to be initiated by relativistic pair fireballs or pair plasmoids which are interacting with surrounding medium to form a relativistic blast wave. While the standard view is that 2nd order Fermi acceleration at the shock front will produce relativistic electrons (and protons) behind the shock, an important aspect to keep in mind is that it takes a time comparable to the Alfvén crossing time to build up hydromagnetic turbulence behind the shock front in order for the 2nd order Fermi acceleration mechanism to be efficient. Within the first few tens to hundreds of milliseconds of a GRB, the expected densities of pairs and electron/proton plasma are so high that the fate of any incoming particle which is swept up from the surrounding medium, is strongly dominated by collisional processes rather than 2nd order Fermi acceleration.

In Böttcher, Schlickeiser & Marra (2001), we have investigated the evolution of the particle spectra and the expected radiation signatures such collision-dominated pair fireballs or plasmoids during the very early phase of GRBs. The peak apparent luminosities are expected to be reached after a few tens to hundreds of milliseconds, and the spectra are dominated by marginally optically thick thermal Comptonization spectra, peaking at several hundreds MeV to a few GeV. Figs. 19 and 20 show the typical range of peak luminosities and peak photon energies obtained in our calculations. We find that those early high-energy flashes can be very luminous so that, in spite of their short duration, serendipitous detections by the GLAST satellite during its continuous scanning observations of the high-energy gamma-ray sky should be possible.


Fig. 17: Simulated time history of the high-energy emission expected through hadronic processes by ultra-high energy cosmic rays accelerated in the GRB blast wave. Model parameters and the estimate of the time-dependent synchrotron (set of broken power-law curves at the top left of the figure) and expected SSC emission (broken power-law curves in the range ~ 10-2 - 105 MeV) were chosen to approximate the observables of GRB 970508. The inset shows the time- and energy-integrated fluxes, compared to the EGRET, Whipple, and the anticipated GLAST sensitivity levels. From Böttcher & Dermer (1998).


Fig. 18: Simulated time history of generic GRB spectra accounting for the reprocessing of GRB radiation off an intervening cloud of Thomson depth 1, located at 1016 cm from the GRB source. At low energies, the spectra are attenuated by Compton scattering out of the line of sight, while at high energies, they are attenuated by gamma-gamma pair production with photons that have been backscattered in the cloud. For more details see Dermer & Böttcher (2000) and Dermer, Böttcher, & Liang (2001)


Fig. 19: Apparent isotropic peak luminosities from a collision-dominated pair fireball during the very early evolution of a GRB, as a function of total injected energy per solid angle and bulk Lorentz factor. In the color-coded contour plot in the top panel, red corresponds to ~ 1054 erg/s, blue corresponds to ~ 1051 erg/s. From Böttcher, Schlickeiser & Marra (2001).


Fig. 20: Peak photon energy of the spectrum from a collision-dominated pair fireball at the time of maximum apparent isotropic luminosity during the very early evolution of a GRB, as a function of total injected energy per solid angle and bulk Lorentz factor. In the color-coded contour plot in the top panel, red corresponds to ~ 10 GeV, blue corresponds to ~ 100 MeV. From Böttcher, Schlickeiser & Marra (2001).

Observations of Optical Afterglows

Using part of Ohio University's 1/12 share of the MDM Observatory, we are contributing to rapid follow-up observations of optical afterglows of GRBs. In particular at the 1.3 m telescope, we have so far contributed to follow-up observations of to afterglows, namely of GRB 050502A and of GRB 050922C.

We observed the afterglow of GRB 050502A in 21 R-band exposures, following its light curve from ~ 1.8  to ~ 8.2 hours after the GRB. An animation of the fading afterglow as observed by us, can be found in Fig. 21. Combined with results from other optical observatories, our results indicated a break in the optical light curve ~ 1.6 hr after the burst. We found that the spectral and lightcurve properties of GRB 050502A slightly favor a scenario in which a relativistic outflow from the GRB source expands into a homogeneous interstellar environment.


Fig. 21: The optical afterglow of GRB 050502A, observed with the 1.3 m telescope of the MDM Observatory. The link will show a time lapse animation of our sequence of observations, from ~ 1.8  to ~ 8.2 hours after the GRB. For more details on the scientific results of these follow-up observations, see Yost et al. (2005).