If you could look out into the night sky with X-ray eyes, you would see the sky littered with points. See the image above from the eROSITA/SRG survey telescope. These are not stars, but growing, super-massive black holes in the centers of galaxies.
How common are super-massive black holes? How are they created? How frequently do they grow and ignite as luminous Active Galactic Nuclei (AGN)? Is it a random phenomenon or strongly inflenced by the life of the host galaxy? Does the energy released impact the host galaxy directly or indirectly?
To answer these questions, I have worked on X-ray surveys with the Chandra, XMM-Newton and eROSITA telescopes, and received better answers by advancing the state-of-the-art in multi-wavelength analysis, X-ray spectroscopy.
Multi-wavelength association by catalog matching
To know the host galaxy and its distance, the X-ray detections need to be cross-referenced with optical, near-infrared and infrared imaging. This is a match of N catalogs, and has many possible combinations if the positions are uncertain. We can take advantage of knowing the typical colors of AGN host galaxies (as opposed to non-X-ray sources).
The state-of-the art code for self-consistent catalog matching is the Bayesian code NWAY. I am the lead code developer. In this era of large astronomical surveys, NWAY is being used more and more, including in LSST and eROSITA collaborations.
Demographics of AGN
Distant, faint and heavily obscured AGN are under-detected in X-ray survey. Carefully treating the detection limitations allows powerful inference on the entire population.
I have pioneered non-parametric AGN space density estimation. For undetectable AGN, this method reports large uncertainties, as it should be. This contrasts ad-hoc models which extrapolate unreliably.
With eROSITA, I am working on the most rapidly growing black holes, and in which host galaxies they live.
Black holes seeds
Most super-massive black holes were probably created in the first 750 million years, before reionisation. We do not know whether stars or large gas clouds are the progenitors of these black holes. Because we do not know the chemistry of star-formation at these early epochs, and we do not know how black holes impact their environments, this is virtually impossible to test robustly with simulations. A new generation of instruments, including LISA, Athena and Lynx may open the door.
Seeding galaxies with black holes cannot be too inefficient, otherwise we would not see as many AGN as we do. In 2019, I presented this argument by analysing cosmological dark-matter N-body simulations with minimal assumptions. Predictions for LISA are very encouraging!