The cosmic evolution of star-forming galaxies

Galaxies separate into passive and star-forming galaxies, where star-forming galaxies show a tight correlation between stellar mass and star formation rate (SFR), referred to as the star forming main sequence. In this thesis, we have measured the average galaxy star formation rate – stellar mass relation out to z~5 in the COSMOS (Cosmic Evolution Survey) field. This survey includes deep radio observations taken as part of the VLA-COSMOS 3 GHz large program that provide a dust-unbiased view of star-formation. To measure SFRs over a wide range of galaxy masses, including galaxies too faint to be detected individually, we employed a stacking analysis on the 3 GHz data. We found a flattening of the star-forming main sequence at high masses that can be explained by the increasing fraction of bulge-dominated galaxies which follow a shallower SFR – stellar mass relation than disk-dominated galaxies. As bulges grow more prominent in the low-redshift galaxy population, the flattening of the main sequence becomes more significant. We found that galaxy environment, probed by X-ray-groups and local galaxy number density, has no significant effect on the shape of the star-forming main sequence at z>0.3. We have compared SFRs derived from publicly available mid-infrared (MIR), far-infrared (FIR), radio, and ultraviolet (UV) photometry for massive star-forming galaxies selected consistently at z~0 and z~0.7. We probed the dust properties of these massive disk galaxies by analysing how their UV luminosity depends on galaxy inclination. By comparing our observed trends with radiative transfer model predictions we constrained the average opacity and clumpiness of the dust. We found that UV attenuation has increased between z~0 and z~0.7 by a factor of 3.5. A higher fraction of clumpy dust around nascent star-forming regions can explain the substantial UV attenuation at z~0.7. If the gas and dust geometry at high-redshift are significantly different than inferred from our current models, this would have significant implications for our SFR calibrations relying on the UV and IR emission from galaxies. Reproducing the spatial distribution of galaxy components like stellar mass, newly formed stars, dust, metals, and gas will be a key objective for future theories of galaxy evolution