The Hidden Universe: Investigating the Evolution of Dusty Star Formation and Gas Consumption Across Cosmic Time

Half a century has now passed since the first observations of the Universe at infrared wavelengths, revealing a population of dust-enshrouded galaxies, many of which were too faint to be detected in the optical surveys of the time. Along with observations at ultra-violet and optical wavelengths, these observations support a picture of a star-formation rate density, which rises rapidly to a peak at z ≈ 2 and then declines to the present day. However, a key outstanding question remains: what drives this evolution of the star-formation rate density? Is the peak of the star-formation rate density driven by a larger supply of molecular gas in galaxies or because galaxies are able to form stars more efficiency, or both? Observations of the infrared sky also laid the ground work for the discovery of a population of distant, highly infrared luminous galaxies detected at submillimetre wavelengths (Submillimetre Galaxies). Whilst we now have a good understanding of the physical properties of the ‘canonical’ z ≈ 2 submillimetre population, the study of the intrinsically rare, high-redshift, bright-end tail of this population is far from complete. With only a few detections of bright (S850μm > 15mJy) submillimetre sources the single-dish number counts and the redshift distribution of this population are poorly constrained. It is also commonly accepted that bright sources detected in single-dish submillimetre surveys are either gravitationally lensed, intrinsically bright or blends of multiple galaxies, but the relative contribution of each of these subpopulations to the bright end of the single-dish submillimetre counts is currently unknown. Current models struggle to reproduce the abundance and redshift distribution of the bright-end of the submillimetre population, hampered by a lack of robust observational data. To better inform these models we not only need accurate submillimetre number counts and a robust redshift distribution for the bright-end of the submillimetre population (single-dish sources with fluxes S850μm > 15mJy), but also a clear picture of the contribution of lensed and blended galaxies to the over-abundance of bright submillimetre galaxies observed. In this thesis we use data from the two largest extragalactic surveys at 850μm with JCMT to date; the SCUBA-2 Cosmology Legacy Survey and the SCUBA-2 Large eXtragalactic Survey, as well as recent data from an ALMA follow-up survey of the SCUBA-2 Large eXtragalactic Survey XMM-LSS field. We employ a statistical approach to explore the cosmological evolution of the molecular gas mass density (ρH2 ) measuring the average observed 850μmflux density of near-infrared selected galaxies as a function of redshift. The redshift range considered corresponds to a span where the 850μm band probes the Rayleigh-Jeans tail of thermal dust emission in the rest-frame, and can therefore be used as an estimate of the mass of the interstellar medium (ISM). With a sample approximately 2 orders of magnitude larger than in previous works we significantly reduce statistical uncertainties on ρH2 to z ≈ 2.5. Our measurements are in broad agreement with recent direct estimates from blank field molecular gas surveys, finding that the epoch of molecular gas coincides with the peak epoch of star formation with ρH2 ≈ 2×107M⊙ Mpc−3 at z ≈ 2. We demonstrate that ρH2 can be broadly modelled by inverting the star-formation rate density with a fixed or weakly evolving star-formation efficiency. This “constant efficiency” model shows a similar evolution to our statistically derived ρH2 , indicating that the dominant factor driving the peak star formation history at z ≈ 2 is a larger supply of molecular gas in galaxies rather than a significant evolution of the star-formation rate efficiency within individual galaxies. We use data from the SCUBA-2 Large eXtragalactic survey of the XMM-LSS field to investigate the abundance of sources at the bright-end of the 850μm number counts. The S2LXS XMMLSS survey maps an area of 9deg2, reaching a moderate depth of 1σ ≃ 4mJybeam−1. This is the largest contiguous area of extragalactic sky mapped by JCMT at 850μm to date. The wide area of the S2LXS XMM-LSS survey allows us to probe the ultra-bright (S850μm ≳ 15mJy), yet rare submillimetre population. We present the S2LXS XMM-LSS catalogue, which comprises 40 sources detected at >5σ significance, with deboosted flux densities in the range of 7mJy to 48mJy. We robustly measure the bright-end of the 850μm number counts at flux densities >7mJy, reducing the Poisson errors by a factor ≈ 2 compared to existing measurements. The S2LXS XMM-LSS observed number counts show the characteristic upturn at bright fluxes, expected to be motivated by local sources of submillimetre emission and high-redshift strongly lensed galaxies. We find that the observed 850μm number counts are best reproduced by model predictions that include either strong lensing or source blending from a 15 arcsec beam, indicating that both may make an important contribution to the observed over-abundance of bright single-dish 850μm selected sources. We explore the multiplicity fraction of bright single-dish 850μm selected sources using data from a follow up ALMA survey of 17 single-dish detected submillimetre sources from the S2LXS XMM-LSS field. Our ALMA maps reach a median sensitivity of 1σ = 0.11mJy, with a median synthesised beam size of 0.59′′×0.50′′. In our deep ALMA maps we detect 22 sources at a significance of >5σ, finding a multiplicity fraction of 54% at S850μm > 12mJy. Our initial results suggest that source blending does not significantly contribute to the abundance of bright sources observed in single-dish 850μm surveys. This is an unexpected result given that the S2LXS XMM-LSS number counts are broadly reproduced by models that incorporate source blending, and further work is required to confirm this