Modelling multi-wavelength evolution of AGN across cosmic time

The evolution of Active Galactic Nuclei (AGNs) is crucial to galaxy evolution, given that AGNs affect their host galaxies via AGN feedback. In this thesis, I present predictions for the evolution of supermassive black holes (SMBHs) and AGNs from the semi-analytic model of galaxy formation galform, over a range of redshift (0 < z < 15) and wavelength (from radio to X-ray). First, I compare SMBH masses and AGN optical to X-ray luminosities from the model for z < 6 to observations, and explore the evolution of typical SMBHs within the model. I find that the median SMBH spin evolves very little over this redshift range. Secondly, I present predictions for z ≥ 7 for future surveys by JWST, EUCLID, ATHENA, and Lynx. I find that Lynx will detect the smallest SMBHs in the smallest host galaxies and host haloes, and that the predictions are generally insensitive to the SMBH seed mass. Thirdly, I predict the evolution of jet powers and (core-dominated) radio luminosities from the model for z < 6, and compare the evolution of these to observations. I predict the jet powers, halo masses, and fuelling mechanisms that dominate the model predictions. Finally, I present predictions of radio luminosities, lobe sizes and Fanaroff-Riley types of radio sources by combining a radio lobe evolution model appropriate for extended sources with the galaxy formation model. I find that this model generally is in good agreement with observed radio properties at z = 0, except for the fractions of Fanaroff-Riley sources, the number of low luminosity radio sources in high stellar mass galaxies, and the number of large sources. I explore the effect of varying different free parameters of this radio model, and suggest potential improvements.