GRB970228: the GRB-afterglow connection

The specific work on the spectral properties of GRB970228 ([Frontera et al. 1998a]) indicates the existence of a connection between GRB event and the associated X-ray afterglow. showed a strong evolution during the main peak (Fig. 5.14).
The maximum value of $\alpha_\Gamma$ ($-0.9\pm 0.2$) achieved at the onset of the 1^st pulse as well as the corresponding $\alpha_X$ during the first 5 s are consistent with the asymptotic spectral index below the peak energy E_p (-0.67), as expected in synchrotron emission models (e.g. the Synchrotron Shock Model (SSM) ([Tavani 1996]). The spectral index of the 1^st pulse rapidly evolves to $\alpha_\Gamma = -2.3\pm 0.3$ at the end of the $\gamma$-ray pulse. In the SSM framework the final value of $\alpha_\Gamma$ during the 1^st pulse corresponds to an index of the non-thermal electron energy distribution function $\delta = 3.7$.During the first 3 s of the 2^nd pulse the spectrum is significantly harder than during the last part of the 1^st pulse. This indicates significant re-energization or relaxation of the particle energy distribution function within the $\sim 20$ s between the 1^st and the 2^nd pulse. It appears that the last three GRB pulses and the X-ray afterglow have a similar non-thermal spectrum and that this spectrum does not appear to change from the first to the second TOO. These results are not consistent with simple cooling models of excited compact objects. An analysis of Ginga data suggested the existence of black-body spectral components in the `precursor' or `delayed' GRB emission in the 1-10 keV band. In this burst we find no evidence for a black-body spectral component of temperature $\sim 1$-2 keV or for a prominent soft X-ray component. Furthermore, there is no evidence of upturn in the soft X-ray intensity with respect to the higher energy spectrum.

An evolving non-thermal spectrum for both the burst and afterglow emission is generally expected in relativistic expanding fireball models ([Mészáros and Rees 1997,Tavani 1997]). In these models, the physics and locations of the shocks associated with the fireballs heavily influence the photon emission mechanisms. Simple fireball models, in which only the forward blast wave radiates efficiently, predict an evolution of the peak energy as a function of the time t from the burst onset, $E_p \propto t^{-3/2}$ ([Tavani 1997,Wijers et al. 1997]). By extrapolating our data for the 1^st pulse of GRB970228 we obtain an initial $E_p \simeq 1$ MeV. In this model, the overall evolution of the X-ray intensity is expected to evolve as $\propto t^\delta$with $\delta = (3/2)(\alpha + 1)$ and $\alpha$ an appropriate photon index ([Wijers et al. 1997]). If we use our best value of $\alpha_\Gamma$ for the decay part of the 2^nd pulse ($-1.94\pm 0.13$), we obtain $\delta = (-1.4\pm 0.2)$, a value that is consistent with the observed decay of the GRB970228 afterglow . However, if we use the value of $\alpha_\Gamma$ determined at the end of the 1^st pulse ($-2.3\pm 0.3$), the resulting value of $\delta$($-1.95\pm 0.45$) would not agree with our observations. The discontinuity in the $\gamma$-ray spectral index observed from the end of the 1^st pulse to the beginning of the 2^nd pulse requires an interpretation. Given the continuity of the spectral index, starting from the 2^nd pulse, any physical relation between the X-ray component of the GRB and the afterglow emission most likely holds with the last set of hard GRB pulses, not with the 1^st one. This suggests that the emission mechanism producing the X-ray afterglow might be already taking place after the 1^st pulse.