In high mass X-ray binaries (HMXBs), the black hole or neutron star accretes matter from the wind of a massive supergiant companion: the stellar wind drives changes in the accretion and thus the system’s X-ray emission. But the interaction of this emission with the wind material can also be used to study the wind itself, in particular its geometry, porosity (or clumpiness), mass-loss rate and interaction with the compact object. Such winds are strong and fast and can significantly influence the evolution of massive stars; their observational studies are, however, hampered by a lack of direct diagnostics of the stellar structure. HMXBs are our unique chance to probe these wind structure, in particular through the variable absorption imprinted by the wind onto the continuum emission.
In this talk, I will address some of our recent work on understanding wind accretion in the HMXBs, focusing on different variability timescales. High resolution spectroscopy reveals a complex, multi-phase medium such as would be expected in the case of cold clumps embedded in a hotter wind. X-ray color-color diagrams can be used to assess absorption dips on timescales as short as a few tens seconds even in faint HMXBs; the resulting measure of the stochastic absorption variability constrains the size of the wind clumps. Low cadence observations allow us to constrain the orbital variability of absorption and through it the large-scale accretion geometry onto the compact object and the porosity of wind of the massive stellar companion. I will in particularly highlight how population studies that will be enabled by eROSITA will help us to understand the evolution of the wind properties in HMXBs.