Radiation drag on the accretion disc in type-I X-ray bursts

Type I X-ray bursts on accreting neutron stars offer a means to probe the interior of these stars by providing information about the conditions at its surface, via the measurement of radius and surface gravitational redshift. Our understanding has been hampered by difficulties in finding spectral models that adequately describe the spectra, in part because we do not properly understand how the burst luminosity influences the structure of the disc and the properties of the accretion flow. I have used archival Rossi X-ray Timing Explorer data to perform spectral analyses of 1,759 bursts from 56 sources. I investigated the effect of allowing the pre-burst persistent emission to vary with time, while holding its spectral shape fixed. I found that an increase in the intensity, to several times its pre-burst level, significantly improves the quality of the spectral fits. This increase cannot be attributed to changes in the shape of either the burst or the persistent components. One possible interpretation of these results is a temporary increase in accretion rate, possibly as a consequence of radiation drag on the disc. I have shown that the magnitude of the persistent flux increase is consistent with the results of earlier computer simulations. I further test this hypothesis using smoothed particle hydrodynamics (SPH) computer simulations of an irradiated accretion disc. The simulations evolve the Navier-Stokes equations with an extra velocity-dependent force representing radiation drag. I test the effect of partial self-shading of the disc by devising four methods for propagating radiation within it. The results suggest that large accretion enhancements are possible provided that angular momentum loss can be efficiently communicated to the interior of the disc.