SIMBAD references

We demonstrate that models of resonant cyclotron radiation transfer in a strong field (i.e., cyclotron scattering) can account for spectral lines seen at two epochs, denoted S1 and S2, in the Ginga data for GRB 870303. S1, which extends 4 s, exhibits one line at ~20 keV, while S2, which extends 9 s, exhibits harmonically spaced lines at ~20 and 40 keV. The midpoints of S1 and S2 are separated by 22.5 s. Using a generalized version of the Monte Carlo code of Wang et al., we model line formation by injecting continuum photons into a static plane-parallel slab of electrons threaded by a strong neutron star magnetic field (∼10¹² G) that may be oriented at an arbitrary angle relative to the slab normal. We examine two source geometries, which we denote ``1-0'' and ``1-1'', with the numbers representing the relative electron column densities above and below the continuum photon source plane. The 1-0 geometry may represent, e.g., a line formation region levitating above the surface of the neutron star, or possibly a plasma-filled flux tube illuminated from below. The 1-1 geometry, on the other hand, corresponds to line formation in a semi-infinite atmosphere at the surface of a neutron star. We apply rigorous statistical inference to compare azimuthally symmetric models, i.e., models in which the magnetic field is parallel to the slab normal, with models having more general magnetic field orientations. If the bursting source has a simple dipole field, these two model classes represent line formation at the magnetic pole, or elsewhere on the stellar surface. We find that the data of S1 and S2, considered individually, are consistent with both geometries, and with all magnetic field orientations, with the exception that the S1 data clearly favors line formation away from a polar cap in the 1-1 geometry, with the best-fit model placing the line-forming region at the magnetic equator. Within both geometries, fits to the combined (S1+S2) data marginally favor models that feature equatorial line formation, and in which the observer's orientation with respect to the slab changes between the two epochs. We interpret this change as being due to neutron star rotation, and we place limits on the rotation period.