next previous
Up: Highlights of the Rome


2 How many classes of GRBs are there?

The discovery six months ago of an unusual Type Ic supernova, SN 1998bw (Galama et al. 1998; Galama 1999), in the BeppoSAX WFC error circle for GRB 980425 (Soffitta et al. 1998) (see Fig. 1) has focused attention once again on the question: How many distinct classes of GRBs are there?

If GRB 980425 were associated with SN 1998bw, the luminosity of the burst would be $\sim 10^{46}$ erg s-1 and its energy would be $\sim 10^{47}$ erg. These values are five orders of magnitude less than those of the other BeppoSAX bursts, whose luminosities range from 1050 to 1053 ergs s-1 and whose energies range from 1052 to 1055 ergs (see below). Moreover, the behaviors of the X-ray and optical afterglows would be very different from those of the other BeppoSAX bursts, yet the burst itself is indistinguishable from other BeppoSAX and BATSE GRBs with respect to duration, time history, spectral shape, peak flux, and fluence (Galama et al. 1998).

There is another troubling aspect about the proposed association between GRB 980425 and SN 1998bw: Also inside the BeppoSAX WFC error circle was a fading X-ray source (Pian et al. 1998a,b; Pian et al. 1999; Piro et al. 1998) (see Fig. 1). Connecting this fading X-ray source with the burst gives a power-law index of $\sim 1.2$ for the temporal decay rate (Pian et al. 1998b), which is similar to the behavior of the other X-ray afterglows observed using BeppoSAX, ROSAT and ASCA. This fading X-ray source must therefore be viewed as a strong candidate for the X-ray afterglow of GRB 980425. There is also strong statistical evidence that all Type Ib-Ic supernovae (SNe) do not produce observable GRBs (Graziani et al. 1999a,b).

  
\begin{figure}

\includegraphics [width=6cm,clip]{grb_980425.ps}\end{figure} Figure: GRB 980425 field, showing the positional error circle for GRB 980425 determined using the BeppoSAX WFC (large solid circle), the positional error circles for the fading X-ray source detected by the BeppoSAX NFI (small solid circle labeled 1SAX1935.3-5252) and for the host galaxy of SN 1998bw (small solid circle labeled 1SAX1935.0-5248). From Graziani et al. (1999a)

Approaching the possible association between SN 1998bw and GRBs from the opposite direction, one can ask: What fraction $f_{\rm GRB}$ of the GRBs detected by BATSE could have been produced by Type Ib-Ic SNe, assuming that the proposed association between GRB 980425 and SN 1998bw is correct, and that the bursts produced are similar to GRB 980425? Assuming that the association between SN1998bw and GRB 980425 is real, using this association to estimate the BATSE sampling distance for such events under the admittedly dubious assumption that the GRBs produced by Type Ib-Ic SNe are roughly standard candles, and assuming that all Type Ib-Ic SNe produce observable GRBs, Graziani et al. (1999a,b) find that no more than $\sim 90$ such events could have been detected by BATSE during the lifetime of the Compton Gamma-Ray Observatory, indicating that the fraction $f_{\rm GRB}$ of such events in the BATSE catalog can be no more than about 5%. This result suggests that the observation of another burst like GRB 980425 is unlikely to happen any time soon, even assuming that the association is real, and consequently, the question of whether Type Ib-Ic SNe can produce extremely faint GRBs is likely to remain open for a long time.

Earlier studies have shown that gamma-ray bursts can be separated into two classes: short, harder, more variable bursts; and long, softer, smoother bursts (see, e.g., Lamb et al. 1993; Kouveliotou et al. 1993). Recently, Mukherjee et al. (1999) have provided evidence for the possible existence of a third class of bursts, based on these same properties of duration, hardness and smoothness properties of the bursts. Also, the hardest long bursts exhibit a pronounced deviation from the -3/2 power-law required for a homogeneous spatial distribution of sources, whereas the short bursts and the softest long bursts do not (Pizzichini 1995; Kouveliotou 1996; Belli 1997, 1999; Tavani 1998). These results contradict the expectation that the most distant bursts should be the most affected by cosmological energy redshift and time dilation. Some bursts show considerable high-energy (E > 300 keV) emission whereas others do not, but it is doubtful that this difference signifies two separate GRB classes, since a similar difference in behavior is seen for peaks within a burst (Pendleton et al. 1998).

It is not clear whether the short and long classes, and the other differences among various burst properties, reflect distinct burst mechanisms, or whether they are due to beaming-or some other property of the bursts-and different viewing angles. Some theorists say, however, that the "collapsar'' or "hypernova'' model cannot explain the short bursts (see, e.g., Woosley 1999).

Because of observational selection effects, all of the GRBs that have been detected by the BeppoSAX GRBM and observed by the WFC have been long bursts. It may be possible for BeppoSAX to revise its GRB detection algorithm in order to detect short bursts. We also expect that HETE-2 will detect short bursts and determine their positions (Kawai et al. 1999; Ricker et al. 1999). If so, follow-up observations may well lead to a breakthrough in our understanding of the nature of the short bursts similar to that which has occurred for the long bursts.

A nightmare I sometimes have is that HETE-2 provides accurate positions for a number of short bursts, but the positions are not coincident with any host galaxies because the bursts are due to merging compact object binaries that have drifted away from their galaxy of origin (see below). And furthermore, the bursts exhibit no soft X-ray, optical, or radio afterglows because any envelope that the progenitors of the compact objects might have expelled has been left behind, and the intergalactic medium is too tenuous to dissipate efficiently the energy in the relativistic external shock that is widely thought to be the origin of GRB afterglows. The redshifts of such bursts would be difficult, if not impossible, to determine, since they could not be inferred from the redshift of any host galaxy, nor constrained by the observation of absorption-line systems in the spectrum of any optical afterglow.

On a more positive note, future radio, optical, and X-ray observations of GRB afterglows and host galaxies, may well lead to the identification of new subclasses of GRBs.


next previous
Up: Highlights of the Rome

Copyright The European Southern Observatory (ESO)