Dr Rohan Varadaraj

Abstract:

In this thesis, I present results on the statistical properties of luminous star-forming galaxies in the early Universe, spanning the first two billion years of cosmic time. I combine deep, degree-scale optical and near-infrared imaging from the latest ground-based surveys with next-generation observatories - Euclid, which provides deep, degree-scale space-based near-infrared imaging for the first time, and JWST, which delivers unrivalled depth and resolution in the near-infrared and infrared. With this unique combination, I place strong constraints on the number densities and size-scaling relations of the first galaxies.

First, by combining imaging from the VISTA telescope with deep ground-based optical surveys, infrared imaging from Spitzer/IRAC, and early data from Euclid, I construct a sample of galaxy candidates at redshift 6.5 < z < 7.5 spanning a rest-UV absolute magnitude range of −23.5 ≤ M_uv ≤ −20.2. These sources represent some of the most luminous and massive galaxies at this epoch. After accounting for brown dwarf contamination through a careful SED-fitting analysis, I find that the rest-frame UV luminosity function at z ≃ 7 is best described by double-power law, showing an excess relative to a Schechter function at absolute rest-UV magnitudes M_uv ≲ −22.5 and evolving slowly from z ≃ 8. This suggests that luminous galaxies at this epoch are not yet significantly affected by dust obscuration or mass quenching, and that active galactic nuclei do not contribute significantly to the luminosity function until very bright magnitudes (M_uv < −24).

I then measure the size-scaling relations of 1,668 luminous galaxies at z ≃ 3 − 5 using the JWST PRIMER survey. These sources were selected from ground-based, seeing-dominated imaging, presenting an unbiased sampling of the morphology and size distributions of luminous sources. I find a build-up of large (R_e > 2 kpc) galaxies at z = 3 relative to z = 4 − 5, a redshift-dependent size evolution leading to larger mean sizes at z = 3, and an increase in the intrinsic scatter of the size-mass relations towards lower redshift. These results suggest that by z = 3, some galaxies have undergone dissipative processes such as mergers and gas accretion, allowing for the formation of rare, larger galaxies. However, the majority of galaxies remain compact over this redshift range, with a typical (modal) size of R_e = 0.7 − 0.9 kpc. Finally, I find that the size-mass and size-luminosity relations are consistent with predictions from simulations such as Illustris and FLARES, providing evidence for centrally concentrated star formation in the most massive galaxies at high redshift.