Star Formation and Quenching in Galaxies: Groups, Clusters, and Mergers

Abstract

This thesis studies the impact of galaxy environment on star formation and ‘quenching’, by using simple physically-motivated models that can be fit using available observed quantities. Quenching refers to the close to total suppression, whether gradual or abrupt, of star formation in a galaxy, and remains a challenging process to understand due to the many tangled non-linear physical processes involved in galaxy formation. By examining the effects of specific environments on star formation, we are effectively given naturally controlled experiments. In particular, this work addresses gaps in our understanding of quenching during and prior to (‘pre-processing’) infall into galaxy groups and clusters, as well as the poorly studied star formation burst that occurs when one galaxy merges with another.

The first section of this thesis presents observed properties of galaxies of the GOGREEN 1 < z < 1.5 galaxy groups. Using publicly available COSMOS and SXDF, with supporting GOGREEN spectroscopic data for confirming group properties, I use background subtraction to determine stellar mass functions of groups at z > 1 for the first time. I see enhanced quenching in higher mass galaxies in these groups compared to galaxies in the average field population. Using this result and previously published work measuring the quiescent fraction of galaxies for GOGREEN 1 < z < 1.5 clusters, as well as similar measurements at lower redshifts, I find a halo mass dependence of quiescent fraction excess when controlled for stellar mass, with logarithmic slope, d(QFE)/dlog(Mhalo) ∼ 0.24 ± 0.04 at all redshifts. I find this trend is qualitatively reproduced in the BAHAMAS hydrodynamical simulation at z ∼ 1, but not the increasing quenched fraction with stellar mass trend. I then interpret my observational results using two toy accretion-quenching models. From this analysis, time until quenching in a group/cluster appears to be shorter for larger halos, with a particularly intense dependence required if there is no pre-processing. Our results strongly support a scenario where environmental quenching begins in low-mass < 10^14M⊙ at z > 1.

The second section turns to infall quenching and preprocessing in z ∼ 0 galaxy clusters using SDSS data. Numerous works have looked at low redshift clusters using quiescent fractions and star formation rates, but struggle to break degeneracies in infall quenching timescales or come to agreement. To address this, I build on a string of works by Kyle Oman and Michael Hudson, which employ statistical infall time in projected phase space information from N-body simulations, by adding an additional observable: spectroscopically derived mass-weighted stellar ages (MWAs). I then forward model the MWAs and quiescent fractions in projected phase space using star-formation histories from the stochastic UniverseMachine model, finding overall infall quenching times of ∼ 4 Gyr after first pericentre. The use of MWAs enables breaking degeneracy in a two-parameter model, yielding both time of quenching onset and SFR suppression timescale for our stellar mass bins 9 < log(M⋆/M⊙) < 10 and 10 < log(M⋆/M⊙) < 10.5. The results of this modeling suggest quenching begins close to, or just after first pericentre, but the suppression timescale is relatively long (∼ 2.3 Gyr versus τ < 1 Gyr) for the higher stellar mass bin, indicating ram-pressure stripping is not complete on first pericentric passage. Prior works required short suppression timescales to maintain the SFR bimodality, but we show that the use of stochastic star formation histories removes the need for this constraint.

The third section determines the mass of stars created when two galaxies merge, a long-standing unknown due to not having large and pure samples of post-merger galaxies until this past year (2022). In particular, I forward model the difference in stellar age between post-coalescence mergers and a control sample, controlling for stellar mass, environmental density, and redshift. We find an age difference of up to 3 Gyr, best fit by a stellar mass burst fraction of 0.18 ± 0.02, consistent with some previously published measurements, but much higher than found in hydrodynamical simulations. Our model is robust to choice of analytic star formation history as well as differences in burst duration. Using published SFRs of Luminous InfraRed Galaxies (LIRGs), we estimate a burst duration of 120–250 Myr, which is consistent with simulations and longer than is estimated for post-starbursts in the literature. We find our stellar mass burst fraction is consistent with the amount of molecular gas reported for very close pairs (pre-coalescence) in the literature. Additionally, we find that the difference between published cold gas measurements for pre- and post-coalescence is consistent with our estimated stellar mass burst fraction, lending credence to our approach.