It has been widely accepted that type I X-ray bursts from low mass X-ray binaries (LMXBs) are due to thermonuclear runaways in accreted materials on the surface of neutron stars (e.g. Taam 1985, Lewin et al. 1993, Bildsten 1998). Detailed evolutionary calculations have been performed taking into account the nuclear process during the flash (e.g. Fujimoto et al. 1987, Taam et al. 1996). Recent observations of LMXBs by the Rossi X-Ray Timing Explorer have revealed several important new features related to X-ray bursts. For example, a burst from 4U 1728-34 would be produced by the spin modulation of a localized thermonuclear hot spot on the surface of a rotating neutron star with a millisecond period (Strohmayer et al. 1998) where kilohertz quasi-periodic oscillations (QPOs) are discussed related to the models to constrain the mass and radius of the neutron star. Furthermore, analyzing the burst observation of Cyg X-2, Smale (1998) suggested super-Eddington bursts which resulted from the photospheric expansion. Also he inferred the source distance using an assumed neutron star mass as high as 2 which seems to be consistent with an estimate of Kaaret et al. (1997) for different observations of QPOs. However, the supernova 1987A could have produced a black hole of (Brown & Bethe 1994) which could not be compatible with the above estimates. Therefore, we can say that recent observations of X-ray burst phenomena provide new challenges to both the model of bursts and the theory of neutron star structure.
On the other hand, many nuclear data have been revised and accumulated in these years, some of which may affect the modeling of X-ray bursts. The nuclear process in the proton rich environments was investigated in detail by Wallace & Woosley (1981) where the rapid proton capture process (rp-process) was first proposed. Recently the rp-process has been investigated extensively from a point of the fundamental nuclear process (see e.g. Wormer et al. 1994, Rembges et al. 1997, Schatz et al. 1998). In those series of the papers, they investigated nuclear flows under the condition of constant temperature and density or the assumption of "adiabatic expansion" to see the effects of uncertainties of nuclear physics. Among all, Schatz et al. (1998) analyzed the relation between the nuclear data and the rp-process which would occur at extreme temperature and density conditions with the use of a large nuclear reaction network. However, more plausible models which simulate the thermonuclear flash would be very necessary to examine how the revised nuclear data affect actually physical conditions during the flash. Unfortunately, at present it is difficult to perform multi-dimensional hydrodynamical calculations which include both general relativity in the strong field and the complete nuclear reaction network. Therefore, a simple but crucial model which represents a thermonuclear flash phase is very useful to extract the effects of physical inputs on the flash.
In the spirit of the one zone model of constant pressure, explosive nucleosyntesis (Hashimoto et al. 1983) and the rp-process (Hanawa et al. 1983, hereafter HSH) were investigated in detail with the use of large networks. In the calculations of the rp-process, they have shown clearly that not only the nucleosynthesis proceeds appreciably beyond 56Ni but also appreciable amounts of the nuclear fuel of hydrogen and helium are left if the peak temperature exceeds K for the pressure of ; it was suggested that the unburnt fuel may be responsible for the X-ray bursts at 10 minute interval. Using their approximate network, Fujimoto et al. (1987) investigated X-ray bursts in detail with the evolutionary calculations of an accreting neutron star. Unfortunately, HSH did not discuss how the uncertainties of the initial abundances and the nuclear data affect the thermonuclear history during the flash; since then there have been many revisions of the nuclear data related to the rp-process, it should be worth while investigating the effects of the new data on the flash. For example, uncertainties of Q-values of the proton capture must influence the path in the nuclear chart and therefore the time scale of the burst (see e.g. Schatz et al. 1998), because the path of the rp-process is almost along the proton drip line after the leakage from the hot CNO (HCNO) cycle.
In the present paper, we will investigate the thermonuclear flash with the use of an extended network up to 94 Kr based on the network constructed by Hashimoto & Arai (1985) which will be coupled to the thermodynamical equation for the flash as was done by HSH. In our calculation, we will use the up-to-date nuclear data and other physical inputs like screening factors. Then, by considering that the nuclear process for the rp-process has been examined by Wormer et al. (1994) and Schatz et al. (1998) in detail, special attention will be paid rather to see the effects of the uncertainty of nuclear data on thermal histories of the shell flash related to the fuel left unburnt.
In Sect. 2, we describe a type I X-ray burst model which presents characteristic features of the thermonuclear flash in accreting neutron stars. Physical data incorporated in our network are explained in Sect. 3. Computational results for the rp-process during the flash are presented in Sect. 4 in connection with uncertainties of the nuclear data. Discussion and conclusions are given in Sect. 5.
© European Southern Observatory (ESO) 1999
Online publication: February 22, 1999