Lewis, Alexander John Robert
Dusty star-forming galaxies (DSFGs) are highly luminous (LFIR ≳ 10¹² L⊙), very distant (z ≳ 2.5) and ≳ 10× rarer then ‘normal’, NIR-selected (BzK), star-forming galaxies. Capable of producing a Mstars ≳ 10¹¹-M⊙ galaxy after a tburst ∼ 100-Myr burst of late-stage, merger-induced star formation, DSFGs naturally provide extreme laboratories with which to study the formation and evolution of massive structures within the Universe. Thus far, however, theoretical models have struggled to reconcile the observed abundance and redshift distribution of these massive, dust-enshrouded galaxies that occupy the high-end tail of the galaxy stellar mass function. Therefore, it is of paramount importance, from an observational point of view, to both increase the number of known distant (z ≳ 4) DSFGs and to thoroughly explore their extreme environments in order to provide further constraints on such models. Thus, in Chapter 2 of this thesis, I report on efforts to substantially increase the number of distant DSFGs using the uniquely wide H-ATLAS imaging survey. I analysed a sample of 109, so-called ‘ultra-red galaxies’ selected via their red Herschel -SPIRE flux densities (σ500 > 3.5 and S500 < 100mJy) and flux-density ratios (S500/S250 > 1.5 and S500/S350 > 0.85). Ground-Based continuum imaging at ∼ 850 μm with the JCMT and APEX telescopes allowed me to locate the dust peaks of these S500 ≳ 30-mJy ultra-red galaxies and derive a median photometric redshift of zphot = 3.66 (3.30–4.27, IQR) for them (assuming that they can satisfactorily be represented by a Tdust ∼ 30-K template SED). Using 25 spectroscopically confirmed DSFGs with SPIRE flux densities matching this ultra-red criteria, I determined that these FIR photometric redshift estimates have a minimum intrinsic scatter of σ = 0.14(1 + zspec) and systematically under-estimate the spectroscopic redshifts below zspec ≲ 5. With over a third of these ultra-red galaxies lying above z > 4, I derived a space density of at least p ≈ 6 × 10−⁷Mpc−³ for this sample (assuming a tburst = 100-Myr burst of star formation), which is only a factor of 7× less numerous than that of the most massive (Mstars ≳ 10¹¹M⊙), compact, quiescent galaxies selected in the NIR at z ∼ 3. Finally, although the space density of z > 4 ultra-red galaxies aligns very well with that of massive (MBH ≳ 10⁸M⊙) AGN at z > 6, none have yet to be uncovered within this sample to date. In the following chapter, I present wide images obtained with LABOCA of a sub-sample of 22, representative ultra-red galaxies to see if these galaxies are signposting over-dense regions in the early Universe, as might be expected if they were to evolve into the most massive, compact, quiescent galaxies at z ∼ 0. This LABOCA ultra-red galaxy survey covers an area of ≈ 0.8 deg² down to an average r.m.s. of 3.9mJy beam−¹, with the five deepest images going ≈ 2× deeper still. I catalogue 86 galaxies detected above > 3.5σ870 surrounding these 22 ultra-red galaxies, which implies a δ ≈ 100 ± 30% over-density of S870 > 8.5-mJy (LFIR ≈ (7–30) × 10¹² L⊙) DSFGs when compared against LESS. Thus, I am 99.93% confident that these ultra-red galaxies are pinpointing over-dense regions in the Universe, and ≈ 95% confident that these regions are over-dense by a factor of at least ≥ 1.5×. Using the same template SEDs as in the previous chapter, I derived a consistent median photometric redshift of z = 3.2 ± 0.2 with an IQR of z = 2.8–3.6 for these ultra-red galaxies. I constrained the surrounding galaxies likely responsible for this over-density to within |∆z| ≤ 0.65 of their respective ultra-red galaxies. However, on average, I was only able to associate one surrounding galaxy to within |∆z| ≤ 0.5 of its respective ultra-red galaxy. These ‘associated’ galaxies are radially distributed within (physical) distances of 1.6 ± 0.5Mpc from their ultra-red galaxies, have median SFRs of vI≈ (1.0±0.2)×10³M⊙ yr−¹ (assuming a Salpeter stellar IMF) and median gas reservoirs of Mgas ∼ 1.7 × 10¹¹M⊙. These candidate proto-clusters have average total SFRs of at least ψ ≈ (2.3 ± 0.5) × 10³M⊙ yr−¹ and a space density of ∼ 9 × 10−⁷Mpc−³, consistent with the idea that their constituents may occupy the centres of rich galaxy clusters seen today. Finally, in Chapter 4 of this thesis, I extracted Herschel -SPIRE photometry at the 850-μm positions of DSFGs detected within in the S2CLS and S2COSMOS imaging surveys. I then analysed the multi-wavelength environmental properties around a robust sample of 64 ultra-red galaxies selected via their ‘ultra-red probability’. Similar to the findings in Chapter 3, I found that these ultra-red galaxies are preferentially located in over-dense regions extending over scales of ∼ 5′ (or ∼ 2Mpc at z ∼ 3). Furthermore, I found that these candidate, high-redshift proto-clusters have FIR total dust masses and total SFRs of Mdust ∼ 10⁹M⊙ and ψ ∼ 10³M⊙ yr−¹, respectively. Ground-Based, optical/NIR imaging around a subset of 42 ultra-red galaxies shows a factor of ∼ 5× increase in both the stellar mass and the (MB −MI )-colour of associated LBGs as the radial distance decreases from . 500 kpc – consistent with the emergence of a galaxy red sequence at z ∼ 3. Furthermore, these data show a 1-σ increase in the fraction of ‘green-valley’ galaxies within scales of ∼ 5′ compared to the field – supporting the concept that red-sequence galaxies are appearing at a faster rate around ultra-red galaxies compared to the field. There is a sizeable contribution of Mstars ∼ 10¹²M⊙ from these high-redshift LBGs within the environments of ultra-red galaxies. On average, I was able to associate ≈ 28 LBGs to a given ultra-red galaxy (∼ 30× the number of associated DSFGs). Although these systems have average optical/NIR/FIR properties that are consistent with their evolution into present-day galaxy clusters with DM halos of mass Mhalo ∼ 10¹⁴–10¹⁵M⊙, I am still likely missing a sizeable contribution from unassociated LBGs and DSFGs due to the large photometric redshift uncertainties for the ultra-red galaxies. Therefore, the results presented in this thesis should be regarded as firm lower limits on these environmental properties around ultra-red galaxies, which can now only be improved upon when spectroscopic data increases the accuracy of the photometric redshift estimates presented here.