The DARWIN Project: Cosmological Simulations of Galaxy Formation down to Dwarf Scales in Diverse Environments

Abstract

We present the DARWIN project, a suite of cosmological simulations with RAMSES to study dwarf-galaxy formation and evolution across diverse environments. DARWIN-1 is a full-box simulation in a cubic volume of $(65.5\,{\rm cMpc})^3$ with a minimum cell size of 500pc, while DARWIN-2 comprises zoom-in simulations of 24 field regions and a Milky Way analog connected to a filament, reaching 62.5pc resolution. The simulations employ baryonic physics tailored to their aims. For DARWIN-2, we include radiative transfer in three photon energy bands to trace hydrogen and helium ionization. DARWIN-1 reproduces the cosmic star formation rate density and galaxy stellar mass functions reasonably well at $z>1$. Both simulations broadly reproduce observed stellar mass-halo mass relations, chemical abundances, star-formation surface densities, and black hole-galaxy scaling relations. At lower redshift, DARWIN-1 shows an excess in the number density of massive galaxies, although the inferred stellar content of massive halos depends on the definition of galaxy stellar mass. Other systematic differences arise from spatial resolution, feedback efficiency, and environmental selection. To facilitate comparison with observations, we generate mock images using the radiative transfer codes SKIRT and Powderday. We also explore a machine-learning approach to mitigate resolution-driven biases in the dwarf galaxy properties of DARWIN-1 using the higher-resolution DARWIN-2 simulations. The DARWIN simulation suite therefore provides a comprehensive numerical framework for studying the formation, evolution, and internal physics of dwarf galaxies in a cosmological context.