The Origin of r-process Enhanced Metal-Poor Halo Stars In Now-Destroyed Ultra-Faint Dwarf Galaxies

TL;DR

This study uses cosmological simulations to explore whether the observed r-process enhanced metal-poor halo stars originated from now-destroyed ultra-faint dwarf galaxies, suggesting these galaxies contributed significantly to the Milky Way's halo composition.

Contribution

It introduces a simple, empirically motivated model of r-process enrichment in simulations to estimate the origin of r-II stars from ultra-faint dwarf galaxies.

Findings

Simulated r-II fraction is about 1-2%, matching roughly half of the observed 2-4%.

The observed r-II fraction may be overestimated due to sampling biases.

Simulations suggest ultra-faint dwarf galaxies could account for 20-80% of the observed r-II stars.

Abstract

The highly r-process enhanced (r-II) metal-poor halo stars we observe today could play a key role in understanding early ultra-faint dwarf galaxies, the smallest building blocks of the Milky Way. If a significant fraction of metal-poor r-II halo stars originated in the ultra-faint dwarf galaxies that merged to help form the Milky Way, observations of r-II stars could help us study these now-destroyed systems and probe the formation history of our Galaxy. To conduct our initial investigation into this possible connection, we use high-resolution cosmological simulations of Milky-Way-mass galaxies from the Caterpillar suite in combination with a simple, empirically motivated treatment of r-process enrichment. We determine the fraction of metal-poor halo stars that could have formed from highly r-process enhanced gas in now-destroyed low-mass ultra-faint dwarf galaxies, the simulated r-II fraction, and compare it to the "as observed" r-II fraction. We find that the simulated fraction, f_{r-II,sim} ~ 1-2%, can account for around half of the "as observed" fraction, f_{r-II,obs} ~ 2-4%. The "as observed" fraction likely overrepresents the fraction of r-II stars due to incomplete sampling, though, meaning f_{r-II,sim} likely accounts for more than half of the true f_{r-II,obs}. Further considering some parameter variations and scatter between individual simulations, the simulated fraction can account for around 20-80% of the "as observed" fraction.