Role of AGN feedback in galaxy evolution

Over the past two decades, detailed studies in the nearby Universe have shown that accreting supermassive black holes (SMBHs; or active galactic nuclei; AGN) can have a significant effect on their host galaxies, suppressing star formation and regulating their growth (known as AGN feedback). One of the most striking examples of AGN feedback in action comes from detailed studies of local AGN which exhibit powerful bi-polar jet outflows that can deposit significant energy into the galaxy halo, heating the surrounding gas and thereby regulating their own growth and suppressing star-formation activity. Therefore, studying the properties of AGN and the galaxies in which they reside is crucial in understanding and further developing our current models of galaxy formation and evolution. AGN can be split into two distinct categories, based on the accretion efficiency of the SMBH: radiative-mode AGN, and jet-mode AGN. Radiative-mode AGN are associated with efficient accretion, typically from cold gas, resulting in the formation of a geometrically thin, optically thick accretion disk that is typically surrounded by a dusty obscuring structure; these AGN are known to drive powerful outflows. Jet-mode AGN are associated with inefficient accretion, typically from hot gas, and display powerful bi-polar synchrotron radio jets that emit the bulk of their energetic output into the surrounding medium in the form of mechanical energy; these AGN are identified as such based on radio observations, showing no signs of AGN activity (e.g. accretion disk or torus) at other wavelengths. Based on the nature of the excitation lines, the jet-mode and radio-loud radiative-mode populations are also known as low-excitation radio galaxies (LERGs) and high-excitation radio galaxies (HERGs), respectively. However, our understanding of these AGN and their feedback effect is built primarily from detailed local Universe observations. Determining the physical mechanisms underpinning triggering and fuelling of AGN and how this affects AGN feedback activity across cosmic time is crucial but lacking. In this thesis, I address this shortcoming using deep observations carried out by the LOw Frequency ARray (LOFAR) telescope: the LOFAR Deep Fields; this forms the deepest radio continuum survey to date at low frequencies. I generated key science-enhanced datasets using this survey and studied the cosmic evolution of AGN feedback from low-luminosity radio-AGN within the past 10 Gyrs and how this feedback affects the growth and evolution of galaxies. In the first science chapter of the thesis, I detail the pipeline I developed to generate new, more robust multi-wavelength catalogues in the LOFAR Deep Fields. The existing catalogues in the literature either did not include the deepest available datasets in each survey field, or were created using different methodologies for detecting sources and measuring their fluxes; all this meant the catalogues were not sufficiently robust for the scientific aims of the thesis. To overcome these issues, I generated new catalogues in two of the Deep Fields by combining information from the ultraviolet to the mid-infrared wavelengths to detect sources and extract their properties in a clean and homogeneous manner. These are some of the best-studied regions of the sky and therefore these catalogues are also expected to provide a legacy value beyond the aims of the LOFAR surveys. Then, in the next chapter, using the multi-wavelength catalogues generated, I identify the host-galaxy counterparts of the radio-detected sources in the LOFAR Deep Fields. Host-galaxy identification and characterisation is crucial, in particular for radio surveys, in determining the photometric redshifts and physical properties (e.g. stellar masses, luminosities, star-formation rates) of the radio-source host galaxies, greatly expanding the scientific scope of the survey. I identified the host-galaxy counterparts of the LOFAR sources using a combination of the statistical Likelihood Ratio method and a visual classification scheme, using a workflow to decide the most appropriate method of identification for each source. This process results in a value-added catalogue of over 80 000 radio sources with multi-wavelength counterparts identified for > 97% of them. In this chapter, I then also investigate the properties of host galaxies of the faint radio population in the LOFAR Deep Fields. In the fourth chapter I focus on studying the evolution of the radio-AGN population and their properties in the LOFAR Deep Fields and how feedback from these AGN evolves across cosmic time. Of particular interest, and the focus of this chapter are the LERGs, which dominate at low radio luminosities and are thought to play a key role in the formation of massive galaxies in the local Universe; however, the evolution of this population beyond z ��� 1 is not well known. In this chapter, I present the first robust measurement of the LERG luminosity functions out to z ��� 2.5 and characterise the evolution of their host galaxy properties. This population shows relatively mild evolution across the redshifts examined; this is explained by the different evolution of the LERGs hosted in star-forming galaxies and those hosted in quiescent galaxies. The evolution of the quiescent LERGs show a strong decline in their space densities with increasing redshift, in accordance with the available host galaxies, while there is also an increase in the characteristic luminosity. I also find that unlike in the local Universe, the bulk of the LERGs are hosted by star-forming galaxies at higher redshifts and that the AGN in these galaxies appear to be fuelled by a different mechanism, likely associated with the cold gas, as compared to the LERGs in quiescent galaxies. In the final chapter I present the conclusions and look towards further exploration of this dataset. In particular, I discuss the characterisation of 3% of the radio-sources that were found to be completely invisible at optical and near-infrared wavelengths; an investigation of the far-infrared and radio properties of this subset found that the vast majority of these sources are likely high-redshift star-forming galaxies hosting a radio-AGN. To understand the nature of these extreme sources at early epochs, I present preliminary analysis from recent sub-millimetre follow-up via the sub-millimetre array (SMA) and the James Clerk Maxwell Telescope (JCMT). In addition, I also discuss the spectroscopic follow-up of the radio-AGN population found in the previous chapter with the upcoming multi-object WEAVE spectrograph on the William Herschel Telescope to study the prevalence of AGN activity as a function of different galaxy properties out to high redshifts. Finally, I also describe plans to compare the observational results found in the previous chapters with predictions from the latest cosmological simulations.