Metal Abundances in Nearby Stars and the Chemical History of the Solar Neighbourhood

Open Access
Publisher Website
Google Scholar
Abstract
A simple theory of galactic enrichment in heavy elements is described in some detail, for the purpose of indicating clearly how the assumptions in numerical models of galactic evolution are related to the observational constraints imposed by the dependence of stellar chemical composition on time of formation and by the statistical distribution of metal abundances in long-lived stars such as the G dwarfs. Observational data bearing on these two questions are reviewed and. the cumulative distribution of metal abundances in G dwarfs is derived from an analysis of photometric data in catalogues of nearby stars. The observed distribution is approximately lognormal, but the correction for both intrinsic and observational scatter that has to be applied in order to derive the distribution function of [Fe/H] due to progressive enrichment is rather uncertain. Within the limits imposed by this uncertainty, the distribution function resembles a (truncated) gaussian with a mean [Fe/H] = – 0·3 relative to the Hyades or – 0·2 relative to the mean of young stars, and a dispersion between 0·075 and 0·18. As is well known, simple models of galactic evolution do not fit the distribution function of metal abundance because they predict too many metal-deficient stars. Modifications of the simple model that give a satisfactory fit include prompt initial enrichment (PIE) and the early version of metal-enhanced star formation (MESF(1)); a particularly successful version of PIE is that one of the series of two-component disk–halo models of Ostriker & Thuan which assumes that 10 per cent of the mass of the disk is secondary material from the halo. The more sophisticated MESF(2) model of Talbot does not fit so well as PIE or MESF(1). Inhomogeneous collapse and in fall models only fit the data if they are combined with PIE or MESF(1) and if the present rate of in fall is not too large; such processes acting mainly in a short initial phase could, however, lead to a successful model effectively incorporating PIE.

This publication has 0 references indexed in Scilit: