On the way to the superburst - a numerical simulation study

Regular X-ray bursts are the most prevalent thermonuclear stellar explosions observed in the Galaxy. They occur in accreting binary systems and provide important constraints about the physics of the involved neutron star – a highly compact object. The thermonuclear runaway is periodically triggered in the accreted shell in the atmosphere of the neutron star and can be observed for a few minutes as an intense increase of the luminosity. Observation have shown that, after thousands of X-ray bursts, a rare superburst event of day-long duration, releasing thousand times more energy, may take place. These superbursts are not yet fully understood though thought to be triggered by unstable carbon-burning in the ashes of the previous X-ray bursts. Given that superbursts have a recurrence time of a few years, performing a self-consistent numerical simulation of the gradual build-up of 12C up to the moment when it is ignited is extremely expensive in terms of computational resources. Therefore, most simulations of superbursts start from artificial initial conditions, without simulating the thousands of Type I X-ray bursts that lead to a superburst. Simulations are not yet capable of self-consistently reproducing the event of a superburst providing all its observable features.

In this thesis we present a one-dimensional model which is capable of simulating thousands of Type I X-ray bursts in the surface layer of an accreting neutron star. Our code couples general relativistic hydrodynamics with a detailed nuclear reaction network to investigate the scenario of Type I X-ray bursts. Consequently, we are able to make predictions for the evolution of the composition of the ashes of Type I X-ray bursts. Various parameters influence the ignition of an X-ray burst and the resulting layer of ashes. Therefore, we perform a huge parameter study, focussing mainly on accretion rates and crustal heating, to find fitting sets for a superburst simulation. Investigating the change of these parameters we find three different burning regimes, restricting the range of usable parameter sets for self-consistently simulating a superburst.