X-ray spectral and timing properties of the High Mass X-ray Binaries GX 304-1 and Vela X-1

Abstract

Accreting X-ray pulsars are binary systems hosting a rotating, highly magnetized neutron star, and an optical companion star whose expelled plasma is accreted onto the compact object, a process that eventually leads to pulsed X-ray emission. Discovered more than forty years ago, substantial progress has been reached in the comprehension of these objects, in terms of both their individual binary components and of the X-ray emission properties. However, further efforts (both observational and theoretical) are needed to constrain the many unsolved aspects. Among the key open issues there are the details of the emission mechanism, the geometry and the radiation beaming pattern of the accretion structure, as well as the influence of the optical companion on the observed X-ray properties. This thesis is focused on the spectroscopical and timing analysis of two accreting X-ray pulsars: the Be/X-ray Binary GX 304-1, and the wind-fed binary Vela X-1. INTEGRAL data (5 − 100 keV) have been used for the analysis of GX 304-1, which was observed during a bright outburst in 2012. MAXI data (2 − 20 keV), on the other hand, provide continuous observations of Vela X-1 since 2009 and therefore offer a valuable opportunity for studying its orbital variability. The INTEGRAL analysis of GX 304-1 allowed to characterize the spectral and timing behaviour of the source as a function of the luminosity. A timing solution valid throughout the outburst has been found, which allowed to construct pulse profiles at different luminosities and for different energy bands. The pulse profiles appear strongly luminosity-dependent, thus suggesting a change of the radiation beam pattern with luminosity. Pulse profiles also appear strongly energy-dependent, possibly due to a geometrical effect arising from the rotation of the neutron star or to a distorted magnetic dipole field. Also, contrary to other accreting pulsars, GX 304-1 pulse profiles show only a small pulsed fraction that does not correlate with energy nor luminosity. The pulse period has negative derivative during the accretion phase, implying a spin up episode, and likely indicating the presence of an accretion disk. The pulse-phase averaged spectroscopy confirms the correlation between the cyclotron line energy and the luminosity, with a more precise detector photon energy calibration, and therefore a better estimation of the neutron star magnetic field and of its critical luminosity. The spectral photon index and folding energy have been found to be negatively correlated with luminosity (as expected for a Comptonization spectrum), thus suggesting that Compton cooling becomes more efficient at higher luminosities. The timing solution has also been used to identify pulse-phases of the rotating neutron star, which allowed to perform pulse-phase resolved spectroscopy. Pulse-phase resolved spectra of individual INTEGRAL observations show a variation of the cyclotron line energy up to ~16% with pulse phase. Furthermore, pulse-phase resolved spectra of stacked observations have been analyzed, again favoring a small variation of the cyclotron line energy with pulse phase. These results can be interpreted in terms of a geometrical configuration such that the observer is exploring only a small part of the radiation beam pattern emerging from the accretion structure, thus causing the spectral parameters to result mostly insensitive to the pulsar rotation, a scenario that is also supported by the observed small pulsed fraction. Concerning Vela X-1, spectral analysis on the sub-orbital timescale has been performed, to explore the light curve features and the effects of large scale (i.e. of the order of the optical companion size) structures that are known to affect the binary system. Such studies are important to constrain the stellar wind properties and to study the mechanisms behind the high X-ray variability typical of such sources. First, a sample of double-peaked orbital light curves has been extracted from the entire population of orbital profiles. The double-peaked sample shows a dip around the inferior conjunction that is difficult to explain invoking absorption by neutral matter alone. Instead, Thomson scattering from an extended and ionized accretion wake is shown to be a possible reason for the observed dip. Orbital-phase resolved spectra show that a phenomenological cutoff power-law model, commonly used in literature for accreting pulsars, leads to orbital variation of the photon index, thus it is likely inadequate to describe the Vela X-1 spectral properties. The addition of a partial covering component to model certain orbital phase-bins spectra avoids the photon index modulation and offers a physical interpretation: a highly structured ambient wind that affects only a part of the original X-ray emission, leaving the other part unaffected. The partial covering component is usually attributed to the clumpy nature of the stellar wind, but the observed orbital modulation of such a component does not favor this interpretation. The necessity of such a component in Vela X-1 seems to be more likely due to either a wobbling, or to an intrinsically structured accretion wake. Due to the ability of the neutron star traveling in the ambient wind to develop density inhomogeneities (i.e. the structured accretion wake), the compact object itself can be considered responsible for the formation of clumps which are then accreted, thus feeding the observed high X-ray variability typical of this system.