Taming Giants: Studies on the Growth, Regulation, and Evolution of Dusty, Star-Forming Galaxies in the Early Universe

Half of all cosmic starlight is obscured by dust and re-radiated at cooler, infrared wavelengths. The majority of stellar mass is built in these dust-obscured regions at z > 0.5 -- with the most extreme manifestation taking place in a rare population of galaxies: dusty, star-forming galaxies (DSFGs). DSFGs form stars at extreme rates (~10^2-3 Msun/yr), becoming extremely massive (M* >= 10^10 Msun) in just a few hundred million years. Prodigious star formation generates abundances of dust that obscures starlight, making some DSFGs nearly invisible to even the deepest rest-frame ultraviolet and optical surveys. Though rare in the local Universe, DSFGs are a thousand times more populous at z ~ 1-3. What is their evolutionary fate and what might their descendants look like today? In my doctoral research, I led detailed case studies, large statistical analyses, and I developed an empirically-based numerical model to uncover critical insights into the answers of these questions. I focus on constraining the stellar growth and evolution of DSFGs discovered with the Herschel Space Observatory, complemented by data from the Chandra X-ray Observatory, Hubble Space Telescope (HST), Spitzer Space Telescope, and other ground-based telescopes. First, I present a detailed case study on a protocluster of DSFGs found at z = 4, the Distant Red Core (DRC). In this work, I presented the first measurement of both the stellar and} cold gas content in a massive, z > 3 protocluster, and determined that this protocluster occupies an exceptionally massive dark matter halo (>10^14 Msun), potentially in tension with a simple Lambda CDM cosmological model. I forward evolved the protocluster members to show that these galaxies will likely become the massive quiescent ellipticals dominating cluster cores by z ~ 2-3. Then, I present my first lead-author work where I showed that there was no statistically significant evidence of star formation suppression in dusty galaxies with actively growing black holes when compared to those without. This implies that feedback and heating from actively growing supermassive black holes may not be the primary mechanism that shuts down of star formation in massive galaxies at z > 1. I also derived the first statistically significant quantification of black hole versus star-forming emission as a function of wavelength, which can be used to argue for/against certain photometric filters as "pure'' star-formation indicators in distant, dusty galaxies. Finally, I present my latest in-progress work where I use empirical data on dusty star-forming galaxies to create a novel, numerical model that re-shapes the primary function describing stellar mass assembly in the Universe: the stellar mass function. Using the infrared luminosity function as a nearly-complete census of dust-obscured galaxies, I built a Markov Chain Monte Carlo model that infers the stellar masses of mock populations of dusty galaxies throughout cosmic time. Current results show that the massive end (M* >= 10^11 Msun) of the most robust galaxy stellar mass functions in the literature are deficient by up to an order of magnitude; this is of major concern for galaxy evolution models, which often invoke extreme feedback prescriptions (e.g. AGN) to prevent galaxies from growing much beyond this pivotal mass. Using simple assumptions to forward evolve these mock DSFGs, I demonstrate that massive DSFGs at early times can evolve to match the observed population densities of massive quiescent galaxies at later times, and are therefore the likely dominant ancestral population. Many of the results and predictions presented in this thesis are immediately testable with with uniformly-selected galaxy samples from Cycle 1 JWST GTO / ERS programs