Type II burst properties of the Rapid Burster (MXB 1730-335; abbreviated as RB) have been studied with the X-ray satellite ASCA. The observation were performed in 1998 August 26-27 and 1999 March 26-28. The data contains segments showing mode-1 bursting pattern, a pattern in which a series of rapidly repetitive bursts is followed by the last biggest burst (in the 1999 observation), as well as segments showing lower frequency of burst occurrence (in the 1998 observation) in which two long flat-topped bursts are detected. Due to the imaging capability of ASCA, we were able to avoid the contaminating radiation from the bursting LMXB 4U 1728-34 (located ∼0°.5 from the RB), which had been an obstacle in the previous observations. Therefore, we did not subtract the persistent spectra from that of the burst throughout our analysis.
Properties of individual bursts, such as duration, fluence, peak luminosity, temperature near the peak (approximated with a blackbody spectrum), and time to the next burst were calculated. Three average luminosities associated with each burst: <Lb> (during the burst), <Lq> (during the persistent until the next burst), and <Lc> (during one cycle) are also calculated. The properties of bursts in the 1998 observation are different from that in 1999 observation, when plotted as a function of the fluence.
We do phase-resolved spectral analysis on a set of 42 spectra extracted from six composite bursts containing bursts with similar fluence (bursts from 1998 and 1999 observation are separated). By studying spectral ratios between the 42 spectra, we find evidence that there are two components in the spectra. One component, whose shape is relatively constant at higher energy (≥7 keV), vary greatly in luminosity, while the other one stays relatively constant (both in shape and luminosity). When fitted with a single blackbody model, the 42 spectra give poor results with similar pattern in the residuals. Based on these results we decide to analyze the spectral set using a two-component model consisting of a blackbody component and a multicolor (blackbody) disk model.
Based on some recent theoretical results on the neutron-star--accretion-disk boundary layer (Inogamov & Sunyaev; Popham & Sunyaev), we interpret the best-fit parameters of the blackbody component under an assumption of a disk-like geometry (rather than the traditionally assumed spherical geometry). Then, we argue that the set of best-fit parameters, obtained with the adopted two-component model, are consistent with the expectations from attributing the type II bursts to disk-instability-related process.
We develop a simple phenomenological model, based on the disk instability theory, to explain the range of bursting behavior observed in our data.