The complex stellar system M 22: constraining the chemical enrichment from AGB stars using magnesium isotope ratios
arXiv:2607.14066 · doi:10.1093/mnras/stad2999 10.1093/mnras/stad2999
The paper measures magnesium isotope ratios in the globular‑cluster-like system M 22 and finds that low‑mass (1–3 M☉) AGB stars can explain its s‑process enrichment, while higher‑mass AGB stars are ruled out, implying an age gap of a few hundred Myr between its two stellar populations.
Abstract
The complex star cluster M 22 (NGC 6656) provides a unique opportunity for studying slow neutron-capture (s-process) nucleosynthesis at low metallicity due to its two stellar groups with distinct iron-peak and neutron-capture element abundances. Previous studies attribute these abundance differences to pollution from 3-6 solar-mass asymptotic giant branch (AGB) stars, which produce significant quantities of the neutron-rich Mg isotopes 25Mg and 26Mg. We report the first-ever measurements of Mg isotopic abundance ratios at [Fe/H] approximately -2 in a globular-cluster-like system using very high-resolution and high signal-to-noise spectra (R = 110000, S/N = 300 per pixel at 514 nm) from the VLT/UVES spectrograph for six stars, three in each s-process group. Despite the presence of star-to-star variations in 24Mg, 25Mg, and 26Mg, we find no correlation with heavy-element abundances, implying that the nucleosynthetic source of s-process enrichment must not influence Mg isotope ratios. Instead, a key result of this work is that we identify correlations between 26Mg/24Mg and some light elements. Using a custom suite of AGB nucleosynthesis yields tailored to the metallicity of M 22, we find that low-mass AGB stars of approximately 1-3 solar masses are capable of reproducing the observed s-process abundances of M 22 and that the absence of any difference in Mg isotope ratios between the two s-process groups precludes AGBs with masses above approximately 3 solar masses. This places tighter constraints on possible formation scenarios and suggests an age difference of at least approximately 280-480 Myr between the two populations that is independent of isochrone fitting.
16 pages, 13 figures, 5 tables. Published in MNRAS