2000A&A...358..535I

The origin of abundance spreads observed in ω Centauri is studied in the context of the self-enrichment scenario. Five chemical evolution models are constructed and are compared with empirical metallicity distribution of ω Cen. After a series of simulations, it is found that neither of closed-box, outflow, nor infall models can reproduce the empirical metallicity distribution of ω Cen, while a modified outflow model with a bimodal initial mass function (IMF) gives a metallicity distribution that fits closely to the empirical ones. In the modified outflow model, long-lived stars are assumed to form after the first explosion of type II supernovae (SNII) in a proto-cloud. The modified outflow model involves gas infall at the very first chemical evolution. Thus we conclude that self-enrichment causes the abundance dispersion in ω Cen. A success of the outflow model with the bimodal IMF implies that low mass stars in a globular cluster (GC) should have formed in the gas already enriched by the first generation of SNII. This scenario, originally proposed by Cayrel (1986A&A...168...81C), can explain a lack of globular clusters with [Fe/H]≲-2.2 in the Milky Way Galaxy.