Carbon and nitrogen abundance variations on the main sequence of 47 Tucanae

The FOCAP multi-object fibre optic system on the AAT has been used to obtain spectra for a sample of more than a hundred faint stars in the globular cluster 47 Tuc (NGC 104), from the base of the giant branch to about a magnitude below the main-sequence turn-off. The spectra cover the ultraviolet and blue bands of CN and the G band of CH. Quantitative abundances have been derived by comparing the observations with synthetic spectra. The main-sequence stars can be divided into two approximately equal groups, one having N enhanced by a factor of ∼ 7, and C depleted by 40 per cent, relative to the other, similar to what is inferred for the red giants. The mean heavy-element abundance, as measured by Ca or Fe, differs between the two groups by no more than 0.01 ± 0.05 dex. All the available abundance data indicate that the atmospheres of 47 Tuc stars contain different amounts of material which has been processed through the CNO cycle, but not through more advanced stages of nucleosynthesis. In the red giants, this could be a result of the convective dredge-up of processed material from deep within the stars themselves, as occurs in some more metal-poor clusters. However, stars on the upper main sequence of globular clusters do not have convective envelopes, while any deep mixing there would have consequences incompatible with other observational data. Thus it seems that the main-sequence C and N variations must be either primordial or caused by a pollution mechanism. The similarity of the bimodal abundance patterns among the dwarfs and giants then suggests that the dredge up of CNO-cycled material is not the dominant process in the red giants either, at least not in 47 Tuc. The CNO-processed material seems most likely to have come from intermediate-mass stars on the asymptotic giant branch; various mechanisms by which such material might appear in the main-sequence stars are discussed. Explanations involving a truly primordial origin, or having successive generations of stars, seem to require rather special ‘tuning’. A somewhat speculative alternative self-enrichment process is proposed, in which stellar winds from intermediate-mass stars were captured by existing low-mass stars, early in the life of the cluster. Whatever the mechanism, the abundance data presumably contain important clues about the formation and early evolution of globular clusters.