1 Introduction

Gamma-ray bursts are observed with a large variety in duration, ranging from seconds to minutes (Norris et al. 1996), intensity and variability. The shortest temporal structures are unresolved by detectors and reflect the activity of a highly variable inner engine (Fenimore et al. 1996). On the other hand some bursts last for several minutes which indicates that the energy generation within the burst region has a rather long time scale.

  

Figure 1: The upper panel gives the flux (Y-axis) as a function of time (X-axis) for the gamma-ray burst with BATSE trigger numbers 1609. The lower panel gives the result of the fitted burst. Time is in units of 64 ms, the time resolution of BATSE. The upper right corner gives a schematic representation of the central locus of the black hole (central $\circ$) and the trajectory of our line of sight (solid line) starting at the $\bullet$, moving clockwise. The inner dotted line identifies the angle at which the luminosity distribution within the luminosity cone is maximum, it drops to zero at the outer dotted line (see PZLL98) The simulated burst is binned in 64 ms bins to make a comparison with the observation more easy

In the proposed model a neutron star transfers mass to a black hole with a mass of 2.2 to 5.5 ${M}_\odot$. A strong magnetic field is anchored in the disc, threads the black hole and taps its rotation energy via the Blandford-Znajek (1977) mechanism. Gamma-rays are emitted in a narrow beam. The luminosity distribution within the beam is given by the details of the Blandford-Znajek process. Precession of the inner part of the accretion disc causes the bean to sweep through space. This results in repeated pulses or flashes for an observer at a distant planet.

This model was proposed by Portegies Zwart et al. (1999, hereafter PZLL) to explain the complex temporal structure of gamma-ray bursts.