Multifrequency and Machine Learning methods to Search for Early Super Massive Black Holes

Early Super Massive Black Holes (SMBH) continuously push our understanding of the formation of galaxies and structures in the Universe. SMBH accreting matter under the radio/jet mode produces powerful relativistic jets and emit beamed non-thermal radiation from radio up to very high energy gamma-rays. Those jets pointing directly to Earth create the so-called Blazar phenomena, where the source appears exceptionally bright due to relativistic magnification (beaming) effects. We can spot Blazars up to high redshifts, but they are rare (given the geometrical alignment constraints involved). To date, only a few distant blazars are known (e.g. QJ0906+6930 z=5.57 and PSO J030947+271757 z=6.1), and a direct search for new ones is impactful because each source at z > 5 implies the existence of thousands of similar misaligned objects. A systematic investigation at z > 5-6 will provide a robust lower limit for the density of Jetted SMBH close and within the Epoch of Reionization (EoR). This research proposal aims to apply Machine Learning (ML) techniques coupled with Multifrequency data to search for high redshift blazars candidates. We plan to select promising z~7 candidates based on the Damping Wing Pattern (DWP). The DWP manifests as the absorption of the observed wavelength < 970nm ( 7 and is very sensitive to the neutral fraction of the IGM. The DWP allows to probe well within the EoR phase and provide a remarkable view into the early Universe. This proposal will leverage fresh survey releases (as the CatWISE2020 in IR and eROSAT Q4-2022 in X-rays) and benefit from the leading role of Instituto de Astrofsica (IA) within ASKAP and MOONS projects (which will provide deep radio data and support for optical observations). This plan will apply ML to a complex Multifrequency data frame in search of high-redshift sources and contribute to establishing the fast-emerging branch of Astroinformatics.