The accretion process in neutron-star low-mass X-ray binaries

Abstract

There had been long-standing fundamental problems in the spectral studies of accreting neutron stars (NSs) in low-mass X-ray binaries involving the X-ray spectral decomposition, the relations between subtypes (mainly atoll and Z sources), and the origins of different X- ray states. Atoll sources are less luminous and have both hard and soft spectral states, while Z sources have three distinct branches (horizontal(HB)/normal(NB)/flaring(FB)) whose spectra are mostly soft. I analyzed more than twelve-year RXTE observations (~ 2500 in total) of four atoll sources Aql X-1, 4U 1608-522, 4U 1705-44, and 4U 1636-536. I developed a hybrid spec- tral model for accreting NSs. In this model, atoll hard-state spectra are described by a single-temperature blackbody (BB), presumed to model emission from the boundary layer where the accreted material impacts the NS surface, and a strong Comptonized compo- nent, modeled by a cutoffpl power law (CPL). Atoll soft-state spectra are described by two thermal components, i.e., a multicolor disk (MCD) and a BB, with additional weak Comp- tonized component, modeled by a single power law. I found that the accretion disk in most of the soft state is truncated at a constant value, most probably at the innermost stable circular orbit (ISCO), predicted by general relativity. This allows us to derive upper limits of magnetic fields on the NS surface of the above four atoll sources. The apparent emission area of the boundary layer is small, ~1/16 of the whole NS surface, but is fairly constant, spanning the hard and soft states. All this was not seen if the classical models for thermal emission plus high Comptonization were used instead. By tracking the accretion rate onto the NS surface, I inferred a strong mass jet in the hard state. My study of 4U 1705-44 using broadband spectra from Suzaku and BeppoSAX supported the above results. From my spectral study of the above four atoll sources, I also found that in a part of the soft state with frequent occurrences of kilohertz quasi-periodic oscillations (kHz QPOs), the accretion disk appears to be truncated at larger radii than in other parts of the soft state where the disk is presumably truncated at the ISCO. Thus the production of kHz QPOs in accreting NSs should be closely related to the behavior of the accretion disk. It is well known that the kHz QPO amplitude spectrum tracks the BB, even though we see no changes in the BB spectral evolution track when kHz QPOs are present. The simplest interpretation is that accretion oscillations are imparted in the inner disk and then seen as the waves impact the NS surface in the boundary layer. The transient XTE J1701-462 (2006-2007) is the only source known to exhibit properties of both the Z and atoll types. I carried out the state/branch classifications of all the ~900 RXTE observations. The Z-source branches evolved substantially in the X-ray color-color diagram during this outburst. In the decay, the HB, NB and FB disappeared successively, with the NB/FB transition evolving to the atoll-source soft state. Spectral analyses using my new spectral model show that the inner disk radius maintains at a nearly constant value, presumably at ISCO, when the source behaves as an atoll source in the soft state, but increases with accretion rates when the source behaves as a Z source at high luminosity. We interpreted this as local Eddington limit effects and advection domination in the accretion disk. The disks in the two Z vertices probably represent two stable accretion configurations, and we speculate that the lower (NB/FB) vertex represents a standard thin disk and the upper (HB/NB) vertex a slim disk. The changes in the accretion rate are responsible for movement of Z-source branches and the evolution from one source type to another. However, the three Z-source branches are caused by three mechanisms that operate at a roughly constant accretion rate. The FB is an instability tied to the Eddington limit. It is formed as the inner disk radius temporarily decreases toward the ISCO. The NB is traced out mostly due to changes in the boundary layer emission area, as a result of the system transiting from a standard thin disk to a slim disk. The HB is formed with the increase in Comptonization, consistent with strong radio emission detected from this branch.