We study the origins of 122 ultra-diffuse galaxies (UDGs) in the RomulusC zoom-in cosmological simulation of a galaxy cluster (M200 = 1.15 × 1014 M⊙), one of the only such simulations capable of resolving the evolution and structure of dwarf galaxies (M⋆ < 109 M⊙). We find broad agreement with observed cluster UDGs and predict that they are not separate from the overall cluster dwarf population. UDGs in cluster environments form primarily from dwarf galaxies that experienced early cluster in-fall and subsequent quenching due to ram pressure. The ensuing dimming of these dwarf galaxies due to passive stellar evolution results in a population of very low surface brightness galaxies that are otherwise typical dwarfs. UDGs and non-UDGs alike are affected by tidal interactions with the cluster potential. Tidal stripping of dark matter, as well as mass loss from stellar evolution, results in the adiabatic expansion of stars, particularly in the lowest mass dwarfs. High mass dwarf galaxies show signatures of tidal heating while low mass dwarfs that survive until z = 0 typically have not experienced such impulsive interactions. There is little difference between UDGs and non-UDGs in terms of their dark matter halos, stellar morphology, colors, and location within the cluster. In most respects cluster UDG and non-UDGs alike are similar to isolated dwarf galaxies, except for the fact that they are typically quenched.

We use the Romulus25 cosmological simulation volume to identify the largest-ever simulated sample of field ultra-diffuse galaxies (UDGs). At z = 0, we find that isolated UDGs have average star formation rates, colors, and virial masses for their stellar masses and environment. UDGs have moderately elevated HI masses, being 70% (300%) more HI-rich than typical isolated dwarf galaxies at luminosities brighter (fainter) than MB=-14. However, UDGs are consistent with the general isolated dwarf galaxy population and make up ∼20% of all field galaxies with 107<M⋆/M⊙<109. The HI masses, effective radii, and overall appearances of our UDGs are consistent with existing observations of field UDGs, but we predict that many isolated UDGs have been missed by current surveys. Despite their isolation at z = 0, the UDGs in our sample are the products of major mergers. Mergers are no more common in UDG than non-UDG progenitors, but mergers that create UDGs tend to happen earlier – almost never occurring after z = 1, produce a temporary boost in spin, and cause star formation to be redistributed to the outskirts of galaxies, resulting in lower central star formation rates. The centers of the galaxies fade as their central stellar populations age, but their global star formation rates are maintained through bursts of star formation at larger radii, producing steeper negative g-r color gradients. This formation channel is unique relative to other proposals for UDG formation in isolated galaxies, demonstrating that UDGs can potentially be formed through multiple mechanisms.

The Local Group hosts a number of dwarf galaxies that show evidence of periods of little to no star formation. We use a suite of cosmological simulations to study how star formation is reignited in such galaxies. We focus on isolated galaxies at z = 0 with halo masses between 9.2×10 8 M and 8.4×10 9 M , where star formation is typically shut off by reionization or by supernova feedback. Nearly 20% of these simulated galaxies later restart star formation, due to interactions with streams of gas in the intergalactic medium, indicating that this mechanism is relatively common in this mass range and that many isolated dwarfs at z = 0 may not have been isolated throughout their histories. While high ram pressure interactions lead to stripping, the encounters that reignite star formation are low density and/or low velocity and thus low ram pressure, resulting in compression of the hot gas in the halos of our dwarfs. The gas mass bound up in hot halos can be substantial -at least an order of magnitude greater than the mass contained in HI. Consequently, we find that dwarfs that have experienced reignition tend to be more HI-rich and have a higher M H I /M * ratio at z = 0 than galaxies with continuous star formation. Using this fact, we identify galaxies in the Local Volume that might have "gappy" star formation histories, and can be studied by the Hubble Space Telescope or the James Webb Space Telescope.

We investigate the effects of massive black hole growth on the structural evolution of dwarf galaxies within the Romulus25 cosmological hydrodynamical simulation. We study a sample of 205 central, isolated dwarf galaxies with stellar masses and a central BH. We find that the local M BH–M star relation exhibits a high degree of scatter below M star < 1010 M ⊙, which we use to classify BHs as overmassive or undermassive relative to their host M star. Within isolated dwarf galaxies, only 8% of undermassive BHs ever undergo a BH merger, while 95% of overmassive BHs grow through a mixture of BH mergers and accretion. We find that isolated dwarf galaxies that host overmassive BHs also follow different evolutionary tracks relative to their undermassive BH counterparts, building up their stars and dark matter earlier and experiencing star formation suppression starting around z = 2. By z = 0.05, overmassive BH hosts above M star > 109 M ⊙ are more likely to exhibit lower central stellar mass density, lower H i gas content, and lower star formation rates than their undermassive BH counterparts. Our results suggest that overmassive BHs in isolated galaxies above M star > 109 M ⊙ are capable of driving feedback, in many cases suppressing and even quenching star formation by late times.

To investigate the origin of elevated globular cluster (GC) abundances observed around Ultra-Diffuse Galaxies (UDGs), we simulate GC populations hosted by UDGs formed through tidal heating. Specifically, GC formation is modelled as occurring in regions of dense star formation. Because star formation-rate densities are higher at high redshift, dwarf galaxies in massive galaxy clusters, which formed most of their stars at high redshift, form a large fraction of their stars in GCs. Given that UDGs formed through environmental processes are more likely to be accreted at high redshift, these systems have more GCs than non-UDGs. In particular, our model predicts that massive UDGs have twice the GC mass of non-UDGs of similar stellar mass, in rough agreement with observations. Although this effect is somewhat diminished by GC disruption, we find that the relationship between GC mass fraction and cluster-centric distance, and the relationship between GC mass fraction and galaxy half-light radius are remarkably similar to observations. Among our model objects, both UDGs and non-UDGs present a correlation between halo mass and GC mass, although UDGs have lower dynamical masses at a given GC mass. Furthermore, because of the effectiveness of GC disruption, we predict that GCs around UDGs should have a more top heavy mass function than GCs around non-UDGs. This analysis suggests that dwarfs with older stellar populations, such as UDGs, should have higher GC mass fractions than objects with young stellar populations, such as isolated dwarfs.