The Na-O anticorrelation in horizontal branch stars. III. 47 Tuc and M 5

TL;DR

This study investigates the chemical composition of horizontal branch stars in 47 Tuc and M5, revealing correlations between element abundances, stellar colors, and helium content, supporting multiple population scenarios in globular clusters.

Contribution

It provides new observational evidence linking p-capture element abundances to HB star colors and helium content, and discusses implications for multiple population models.

Findings

He content correlates with star color in 47 Tuc

Small He abundance variation (ΔY<0.03) explains observations

Na-O and N abundances correlate with color in M5

Abstract

To check the impact of the multiple population scenario for globular clusters on their HB, we present an analysis of the composition of 110 red HB (RHB) stars in 47 Tucanae and of 61 blue HB (BHB) and 30 RHB stars in M5. In 47 Tuc we found tight relations between the colours of the stars and their abundances of p-capture elements. This strongly supports the idea that the He content - which is expected to be closely correlated with the abundances of p-capture elements - is the third parameter (after overall metallicity and age) that determines the colour of HB stars. However, the range in He abundance must be small (Delta Y<0.03) in 47 Tuc to reproduce our observations; this agrees with previous analyses. There is possibly a correlation between the abundances of p- and n-capture elements in 47 Tuc. If confirmed, this might suggest that AGB stars of moderate mass contributed to the gas from which second-generation stars formed. Considering the selection effects in our sample (which does not include stars warmer than 11000 K and RR Lyrae variables) is important to understand our results for M5. In this case, we find that, as expected, RHB stars are Na-poor and O-rich, and likely belong to the primordial population. There is a clear correlation of the [Na/O] ratio and N abundance with colour along the BHB. A derivation of the He abundance for these stars yields a low value of Y=0.22\pm 0.03. This is expected because HB stars of a putative He-rich population in this cluster should be warmer than 11000 K, and would accordingly not have been sampled by our analysis. However, we need some additional source of scatter in the total mass loss of stars climbing up the RGB to reproduce our results for M5. Finally, we found a C-star on the HB of 47 Tuc and a Ba-rich, fast-rotating, likely binary star on the HB of M5. These stars are among the brightest and coolest HB stars.