2008MNRAS.390..693D

The majority of globular clusters show chemical inhomogeneities in the composition of their stars, apparently due to a second stellar generation in which the forming gas is enriched by hot-CNO cycled material processed in stars belonging to a first stellar generation. Clearly this evidence prompts questions on the modalities of formation of globular clusters. An important preliminary input to any model for the formation of multiple generations is to determine which is today the relative number fraction of `normal' and anomalous stars in each cluster. As it is very difficult to gather very large spectroscopic samples of globular cluster stars to achieve this result with good statistical significance, we propose to use the horizontal branch (HB). We assume that, whichever the progenitors of the second generation, the anomalies also include enhanced helium abundance. In fact, helium variations have been recently recognized to be able to explain several puzzling peculiarities (gaps, RRLyr periods and period distribution, ratio of blue to red stars, blue tails) in HBs. We summarize previous results and extend the analysis in order to infer the percentage in number of the first and second generation in as many clusters as possible. We show that, with few exceptions, approximately 50 per cent or more of the stars belong to the second generation. In other cases, in which at first sight one would think of a simple stellar population, we give arguments and suggest that the stars might all belong to the second generation. We provide in the appendix a detailed discussion and new fits of the optical and ultraviolet data of NGC2808, the classic example of a multiple helium populations cluster, consistently including a reproduction of the main-sequence splittings and an examination of the problem of `blue hook' stars. We also show a detailed fit of the totally blue HB of M13, one among the clusters that are possibly fully made up by second generation stars. We conclude that the formation of the second generation is a crucial event in the life of globular clusters. The problem of the initial mass function required to achieve the observed high fraction of second generation stars can be solved only if the initial cluster was much more massive than the present one and most of the first generation low-mass stars have been preferentially lost. As shown by D'Ercole et al., by modelling the formation and dynamical evolution of the second generation, the mass loss due to the explosions of the Type II supernovae of the first generation may be the process responsible for triggering the expansion of the cluster, the stripping of its outer layers and the loss of most of the first generation low-mass stars.