Abstract

We use cosmological, magnetohydrodynamical simulations of Milky Way-mass galaxies from the Auriga project to study their enrichment with rapid neutron capture (r-process) elements. We implement a variety of enrichment models from both binary neutron star mergers and rare core-collapse supernovae. We focus on the abundances of (extremely) metal-poor stars, most of which were formed during the first similar to Gyr of the Universe in external galaxies and later accreted on to the main galaxy. We find that the majority of metal-poor stars are r-process enriched in all our enrichment models. Neutron star merger models result in a median rprocess abundance ratio, which increases with metallicity, whereas the median trend in rare core-collapse supernova models is approximately flat. The scatter in r-process abundance increases for models with longer delay times or lower rates of r-process-producing events. Our results are nearly perfectly converged, in part due to themixing of gas between mesh cells in the simulations. Additionally, different MilkyWay-mass galaxies showonly small variation in their respective r-process abundance ratios. Current (sparse and potentially biased) observations of metal-poor stars in the Milky Way seem to prefer rare core-collapse supernovae over neutron star mergers as the dominant source of r-process elements at low metallicity, but we discuss possible caveats to our models. Dwarf galaxies that experience a single r-process event early in their history show highly enhanced r-process abundances at low metallicity, which is seen both in observations and in our simulations. We also find that the elements produced in a single event are mixed with approximate to 10(8) M-circle dot of gas relatively quickly, distributing the r-process elements over a large region.