Using 3D dynamic models to reproduce X-ray properties of colliding wind binaries

Colliding wind binaries (CWBs) are unique laboratories for X-ray astrophysics. The two massive stars contained in these systems have powerful radiatively driven stellar winds, and the conversion of their kinetic energy to heat (up to 108 K) at the wind-wind collision region generates hard thermal X-rays (up to 10 keV). Rich data sets exist of several multi-year-period systems, as well as key observations of shorter period systems, and detailed models are required to disentangle the phase-locked emission and absorption processes in these systems. To interpret these X-ray light curves and spectra, this dissertation models the wind-wind interaction of CWBs using 3D smoothed particle hydrodynamics (SPH), and solves the 3D formal solution of radiative transfer to synthesize the model X-ray properties, allowing direct comparison with the colliding-wind X-ray spectra observed by, e.g., RXTE and XMM. The multi-year-period, highly eccentric CWBs we examine are eta Carinae and WR140. For the commonly inferred primary mass loss rate of ∼10 -3 Msun/yr, eta Carinae's 3D model reproduces quite well the 2-10 keV RXTE light curve, hardness ratio, and dynamic spectra in absolute units. This agreement includes the ∼3 month X-ray minimum associated with the 1998.0 and 2003.5 periastron passages, which we find to occur as the primary wind encroaches into the secondary wind's acceleration region. This modeling provides further evidence that the observer is mainly viewing the system through the secondary's shock cone, and suggests that periastron occurs ~1 month after the onset of the X-ray minimum. The model RXTE observables of WR140 match the data well in absolute units, although the decrease in model X-rays around periastron is less than observed. There is very good agreement between the observed XMM spectrum taken on the rise before periastron and the model. We also model two short-period CWBs, HD150136, which has a wind-star collision, and delta Orionis A, the closest eclipsing CWB. The asymmetry predicted in the unobserved portion of HD150136's orbit, and the line profile variations due to the cavity carved into the primary wind by the secondary in delta Orionis A, helped provide a basis for newly approved Chandra observations of both systems.