Imperfect accretion: ejecting matter in X-ray binaries

X-ray binaries are binary stars composed of a compact object (a black hole, a neutron star) accreting matter from a companion star. These sources can be considered perfect astrophysical laboratories to test our knowledge of, e.g., General Relativity and Magneto-Hydrodynamics. Accretion is the key phenomenon characterizing these systems, but it is not always completely efficient. In many systems, ejections of matter are also observed, e.g. in the form of jets and winds, or also suggested, e.g. to explain the observed strong orbital expansion of a number of systems. Furthermore accretion and ejection seems to be somehow interconnected but the nature of this correlation is not completely clear. The purpose of this thesis is the study of a number of cases where the accretion is imperfect and mass losses have to be taken into account to correctly model the physical properties of the binaries. In the first of the featured projects, I focus on the spectral study of the accretion flow in the Neutron Star (NS) Low Mass X-ray Binary (LMXB) 1RXS J180408.9-342058, an intriguing system which in the past exhibited “very faint” phases of activity. I performed a spectral analysis of data collected by different X-ray telescopes, i.e. INTEGRAL, Swift and NuSTAR, The study led to several interesting results, in particular the observation of the intermediate spectral state, hard to catch in NS LMXBs because very short-lived, and new constraints on the nature of the companion star, which exclude the hypothesis of a helium dwarf companion as suggested in the past. The second project presents a systematic study of (almost) all known Accreting Millisecond X-ray Pulsars (AMXPs), i.e. LMXBs hosting an X-ray pulsar spinning at millisecond periods, with the aim of looking for indications of non-conservative mass-transfer in this class. Comparing this observed luminosity averaged over twenty years with the one expected from the theory in a conservative scenario, I found that over a sample of 19 sources, around one half of it shows indications for mass losses. The third project in this thesis is dedicated to jets, the most known form of mass ejection in X-ray binaries. Jets are characterized by flat radio-to-mid-IR spectra, which have been modelled in the last few decades using the Internal Shocks model ISHEM. The basic idea of this model consists in using the observed X-ray variability as a proxy for the fluctuations of the Lorentz factor in the ejected shells along the jet. I applied the model on the multi-wavelength data set of the NS LMXB 4U 0614+091. I found that ISHEM describes satisfactorily the data only in two cases: using the X-ray variability but in non-conical geometry or either in conical geometry but using flicker noise instead of the X-ray variability. The final project of my thesis aims at testing a unified accretion-ejection model to the Black Hole LMXB MAXI J1820+070. The model considers the accretion flow in X-ray binaries as two-fold, comprising a truncated geometrically thin disk far from the Black Hole and a so-called jet emitting disk serving as the base of the jet close to the Black Hole. Interestingly, the model allows not only to describe the X-rays data, but also to predict the radio power emitted by the jet. In order to test the model, I used X-rays data from Swift and NuSTAR. The preliminary results of the spectral fitting suggest that the model is indeed effective in describing the observed X-ray spectra. Furthermore, the analysis reveals the need for describing the reflection spectrum with two reflection components instead of one: the origin of such intriguing component, if confirmed, will be object of future investigations.