Speaker
Description
High Mass X-Ray binaries are fascinating systems. They consist of an O/B-type donor and a compact object (neutron star or black hole) embedded in the donor’s dense stellar wind, often in a close orbit, where orbital geometry and variability can be exploited to map the circumstellar environment.
During the years, we analysed interesting individual systems and, at the same time, developed some tools that we published to share with the community. In the eclipsing, wind and disc-fed system Cen X-3, the high inclination enables compact object eclipses and direct measurements of NS spin orbital Doppler modulations. During state transitions, from low-hard to high-soft, Doppler shifts of the Fe XXV emission line allow us to reconstruct the ballistic trajectory of emitting plasma as it is captured and accreted by the neutron star. These analyses motivated \texttt{xraybinaryorbit}, a Python package for modelling and fitting orbital modulations of different observables, including absorption and Doppler shifts.
We also explore rapid variability in the magnetar-candidate system 4U~0114+65, one of the slowest known X-ray pulsars ($P_{\rm spin}\sim 9000$ s). Its light-curves show peaks compatible to Rayleigh--Taylor instabilities at the magnetosphere. Short-lived dips and/or spikes in light-curves have the power to probe both accretion dynamics and wind structure (e.g. partial occultations by overdense clumps). To systematically characterise these features, we developed dipspeaks, an automatic dip and peak search based on autoencoders that returns a catalogue of events with noise probabilities. Finally, anticipating the diagnostic power of upcoming high-resolution spectroscopy with many time-binned, phase-resolved spectra, we introduce BLISS, a continuum independent blind line-search algorithm that streamlines the automatic detection and incorporation of emission lines in spectral modelling.
