Discovery of absorption features of the accretion disc corona and systematic acceleration of the X-ray burst rate in XB 1323-619

Abstract

We present results from analysis of the observation of the dipping, quasi-periodic bursting low mass X-ray binary XB 1323-619 made with XMM-Newton in 2003 January. In spectral analysis of the non-dip, non-burst EPIC PN spectrum, a number of absorption lines were discovered, notably at 6.70 and 6.98 keV, which we identify with scattering by high ionization state ions Fe xxv and Fe xxvi. Such features have been seen in other dipping sources, but their origin was not understood. Curve of growth analysis provided a consistent solution in which the line ratio was reproduced assuming collisional ionization with kT = 31 keV, close to the electron temperature we previously determined for the accretion disc corona (ADC) in this source. We thus propose that the absorption lines in the dipping low mass X-ray binaries (LMXB) are produced in the ADC. Spectral evolution in dipping was well described by the progressive covering model, which we have previously shown to give very good explanations of many dipping sources. We discuss the proposal of Boirin et al. (2004b), based on analysis of the present observation, that spectral evolution in all of the dipping LMXB may be explained by subjecting the continuum to a highly ionized absorber. This would require a decrease in X-ray intensity by a factor of ∼3 in dipping at energies where photoelectric absorption is not effective (40-100 keV), which previous BeppoSAX analysis of several sources firmly rules out. In XB 1323-619, any decrease was less than 10 ± 10 per cent and so the ionized absorber proposal can be ruled out. We find a remarkable linear increase in the rate of X-ray bursts with time over the 14-yr period since 1989 and a systematic non-linear increase in source luminosity (L). The linear variation of burst rate with L shows that the burst rate is proportional to mass accretion rate and if continued implies that the gap between bursts will become zero on 2008 January 11. In reality, we expect the source to undergo a transition from X-ray bursting to X-ray flaring, which would confirm the suggestion we previously made that flaring is unstable nuclear burning effectively consisting of a superposition of X-ray bursts.