2021MNRAS.507..834M

Recent observations of globular clusters (GCs) suggest that elemental abundance variations may exist between first-generation (1G) stars. We propose that metal abundance ('metallicity') spreads within GC forming giant molecular clouds (GMCs) can influence the iron abundances of future cluster members. To investigate this, we use original hydrodynamical simulations to model GMC formation in a high redshift dwarf galaxy. Our simulations self-consistently model physical processes such as stellar feedback, dust formation and destruction, and molecular gas formation on dust grains, making them well suited to the study of GMC formation. We conclude that iron abundance variations in GMCs are due to the merging of gas clumps and self-enrichment processes. The metallicity dispersions of GC forming clumps is ∼0.1 dex, reflecting a growing number of studies that claim a non-zero dispersion within GCs. The galactic gas fraction is a key parameter for the formation of clumps and the metallicity 'floor' observed for both Galactic and extragalactic GCs are associated with the parent galaxy's capacity to form massive GMCs. Finally, we argue that GMCs have the potential to trap surrounding metal-poor galactic disc stars, which we interpret as a precursor population (0G). These low metallicity stars are representative of the [Fe/H] value of the host dwarf and thus the chemistry of this 0G may be a fossilized record of the parent galaxy. These results depend on the initial metallicity and radial gradient of the galaxy, the threshold gas density for star formation, and the star formation prescription.