Engineering dwarfs at galaxy formation's edge

Dwarf galaxies provide unique insights into the cosmology of our Universe. They areat the scale most sensitive to the properties of dark matter and the physical processesgoverning their evolution. They are, therefore, an ideal laboratory for testing dark matterand galaxy formation models. Furthermore, due to new surveys like SDSS and DES, thenumber of known dwarf galaxies in the Local Group and beyond continues to grow, withupwards of 50 detected within the Milky Way to date. Due to their close proximity, we canresolve these into their individual stars, providing stellar velocities, metallicities, and starformation histories for hundreds to thousands of members. In this work I present a suiteof cosmological simulations designed to probe the low-mass dwarf regime in a LambdaCold Dark Matter (ΛCDM) cosmology at a mass and spatial resolution of 100 M sun and3 pc respectively.First, I present a selection of dwarf galaxies produced as part of the Engineering Dwarfsat Galaxy formation’s Edge (EDGE) simulation suite and show they conform to all knownobservations of dwarf galaxies to date.Then, I investigate the cusp-core problem in the very smallest dwarf galaxies. I findthat even galaxies with M∗ ∼ 10^5 M sun can have their inner dark matter density loweredby baryonic processes. This proceeds via two heating mechanisms, the typical ‘gas flow’mechanism and a subdominant ‘minor merger’ mechanism. I also show that cusps can beregrown as a result of dense mergers.Next, I follow up with more detailed analyses to test whether my simulations canexplain the lone star cluster discovered in the ultra faint dwarf galaxy Eridanus II. Thisincludes the development of a large suite of star cluster simulations. I find that none ofthe 1×10^9 < M200/M sun < 2×10^9 EDGE dwarf simulations run to date produce a coresubstantial enough to explain the properties of the star cluster in Eridanus II. I discusspossible explanations for this discrepancy and how these can be tested.Finally, I explore whether the aforementioned dark matter heating has any measurableimpact on the halo shape in EDGE simulation. Above M200 ∼ 3×10^9 Msun , halo shapesbecome more oblate and aligned with the angular momentum vector of the gas out to 5-10times the stellar half light radius. This effect owes to gas dissipation rather than cusp-coretransformations. Below M200 ∼ 2×10^9 Msun , haloes retain their typically prolate shapesas in pure dark matter ΛCDM simulations. I predict, therefore, that reionisation fossilsshould be triaxial, providing a novel means to test cosmological models and the nature ofdark matter