Mémoire

Abstract

This master thesis focuses on the analysis of two O-star colliding wind binaries (CWBs). Massive O-type stars are crucial objects that have a huge impact on their surroundings, notably via their stellar winds. Nowadays, we estimate that 71% of massive stars are in binary systems. This, in turn, gives rise to the study of CWBs. The plasma heated by the shock between the winds will then cool down following either an adiabatic or a radiative regime. The goal of this work is to describe the X-ray behaviour of two O-type binaries, HD168112, and HD167971. These two systems have a long period, in addition of a short one for HD167971 which is a triple system. We made the analysis using six XMM-Newton’s observations collected over a period of two decades with the EPIC and RGS instruments. We selected the best model to fit them. It assumed three temperatures for the emitting plasma. From this model we retrieved the fluxes in three energy bands: from 0.5 to 1 keV, from 1 to 10 keV, and from 0.5 to 10 keV, along with fluxes corrected for the ISM absorption. Thanks to the cooling parameter, we conclude the wind interaction in HD168112 is in an adiabatic regime throughout its orbit, as is the the wind interaction in the wide outer orbit of HD167971. Otherwise, the wind-wind collision of the inner eclipsing binary is in a radiative regime. The analysis of the corrected fluxes Fcorr along the orbit of HD168112 allowed us to adjust a relation: Fcorr = A × a/d + B. This fit confirms the 1/d behaviour expected for an adiabatic wind interaction. We processed two ROSAT observations in the soft band that confirmed this result. Using the parameter B, we get the intrinsic X-ray luminosity, LX,intrin = 2.97 × 1032 erg s−1 and LX,intrin/Lbol = 9.97 × 10−8. This finding is in perfect agreement with the canonical value near 10−7 obtained by many previous studies. The analysis of HD168112 fluxes permitted to posit a partial answer to one of the main question in the context of CWB: the lack of over-luminosity of certain CWBs. In fact, HD168112 displays an over-luminosity for only 12% of its orbital cycle. The observations are not well distributed along the orbital phases of HD16797. Neither for the long nor the short period. Nevertheless, by plotting the flux as a function of the long and short period, we concludes that the variation is rather shaped by the eclipsing binary than by the long orbital cycle. A comparison of our fluxes with the optical light curve reveals a phase shift between the two that can possibly be caused by Coriolis deflection. We concluded that new observations are needed to undertake a more thorough study of the X-ray behaviour of HD167971, especially for the eclipsing binary.