Flux decay during thermonuclear X-ray bursts analysed with the dynamic power-law index method

Abstract

The cooling of type-I X-ray bursts can be used to probe the nuclear burning conditions in neutron star envelopes. The flux decay of the bursts has been traditionally modelled with an exponential, even if theoretical considerations predict power-law-like decays. We have analysed a total of 540 type-I X-ray bursts from five low-mass X-ray binaries observed with the Rossi X-ray Timing Explorer. We grouped the bursts according to the source spectral state during which they were observed (hard or soft), flagging those bursts that showed signs of photospheric radius expansion (PRE). The decay phase of all the bursts were then fitted with a dynamic power-law index method. This method provides a new way of probing the chemical composition of the accreted material. Our results show that in the hydrogen-rich sources the power-law decay index is variable during the burst tails and that simple cooling models qualitatively describe the cooling of presumably helium-rich sources 4U 1728-34 and 3A 1820-303. The cooling in the hydrogen-rich sources 4U 1608-52, 4U 1636-536, and GS 1826-24, instead, is clearly different and depends on the spectral states and whether PRE occurred or not. Especially the hard state bursts behave differently than the models predict, exhibiting a peculiar rise in the cooling index at low burst fluxes, which suggests that the cooling in the tail is much faster than expected. Our results indicate that the drivers of the bursting behaviour are not only the accretion rate and chemical composition of the accreted material, but also the cooling that is somehow linked to the spectral states. The latter suggests that the properties of the burning layers deep in the neutron star envelope might be impacted differently depending on the spectral state.

Accretion Disks and Coronae in the X-Ray Flashlight

“…Traditionally, X-ray burst decays have been described with an exponential function, but when high signal to noise ratio data are analyzed, a power-law decay is often a better description and sometimes complex shapes are observed (e.g. in 't Zand et al 2014a;Kuuttila et al 2017). The duration of X-ray bursts commonly ranges from several seconds up to several minutes, and the total radiated output is typically ∼ 10 39 erg (e.g.…”

Neutron star cooling and the rp process in thermonuclear X-ray bursts

“…The exception to this rule is for bursts for which it is more or less certain that there is no H present in the burning layer, like in 4U 1820-30. Kuuttila et al (2017) found perhaps marginal evidence for α evolution in that source, as judged from their Fig. 3, for the phases where dynamical effects (PRE) are negligible.…”

“…11); 2) for Eddington-limited bursts (i.e., bursts with photospheric radius expansion or 'PRE') from 4U 1636-526 and 4U 1608-52, the same applies (although at different α values); 3) there are strong α evolutions for all bursts from GS 1826-24 and non-PRE bursts from 4U 1636-536 and 4U 1608-52; 4) below one-third to onefifth of the peak flux, the α profile is flatter for bursts detected in the hard state than in the soft state. Interestingly, Kuuttila et al (2017) directly compared the α evolution with the theoretical predictions from in 't and found that they are generally inconsistent for fluxes above one-third of the peak flux and marginally consistent below this, with the exception of GS 1826-24, which is strongly inconsistent throughout. In our opinion, both studies (Kuuttila et al 2017 and ours) show that a large portion of the signal in the decay phase of many bursts is due to rp processing and that the difference between PRE and non-PRE bursts is due to a difference in H abundance in the burning layer.…”

“…Their data treatment is different from ours, which makes it interesting as a cross check against our results. Basically, the treatment of Kuuttila et al (2017) follows the f a = 1 method for spectral modeling of the burst spectra, employs the Planck function for the burst emission, and applies a power law to the derived bolometric flux decays with a variable decay index. Complete burst profiles are also included, in contrast to our restriction to fluxes below 55% of the peak.…”

“…Complete burst profiles are also included, in contrast to our restriction to fluxes below 55% of the peak. Kuuttila et al (2017) made no allowance for a second model component next to the power law, like a Gaussian. They found that 1) for the (presumably) H-poor bursts from 4U 1820-30, α has a rather flat profile below about one-third of the peak flux with a value of about 2.0 (cf.…”