| Type I X-ray bursts are thermonuclear explosions that occur on accreting neutron stars. Observations and theoretical burst models are discrepant; nearly all models predict bursts occur frequently for all M˙/M˙ Edd ≲ 1, whereas observations suggest bursts are infrequent when 0.1 ≲ M˙/M˙Edd ≲ 0.3 and cease when M˙/M˙Edd ≳ 0.3. In this thesis, I address this discrepancy by investigating the physics of ignition and its ramifications. The triple-alpha reaction produces 12C, which feeds the hot CNO cycle, and the hot CNO cycle produces 4He, which feeds the triple-alpha reaction. This interplay amplifies hot CNO cycle burning and helps thermally stabilize the matter. Ignition requires a breakout from the hot CNO cycle via 15O(alpha,gamma)19Ne, the rate of which is experimentally poorly-constrained. Models in which the rate is lower by a factor ∼ 3 agree with observations, whereas models with higher trial rates, including the fiducial rate, do not. This suggests the true 15O(alpha,gamma) 19Ne rate is lower than usually assumed.;Superbursts are energetic bursts triggered by 12C burning. The triple-alpha reaction generates the 12C. If present, H immediately fuses with 12C and destroys the fuel. Stable burning of a solar abundance of accreted matter produces an insufficient amount of 12C for ignition. If the accreted matter's metallicity is sufficiently above solar, H depletes before 4He burns and leaves enough 12C for ignition. 4He-triggered bursts from systems that exhibit superbursts differ from those that do not: they have shorter durations and are less frequent. Increasing the accreted matter's CNO abundance reproduces both of these observations as well.;Astronomers have detected oscillations in burst lightcurves from 20 sources. The rotational modulation of a growing hot spot generates oscillations during the rise, whereas the epsilon-mechanism drives oscillations during the decay. A long-lived hot spot requires nonequatorial ignition. I show that nonequatorial ignition occurs preferentially at high M˙, which explains why oscillations occur during the burst rise only when M˙ is high. Furthermore, I show that convection dampens oscillations; this may explain the rarity of oscillation during the burst peak. |
