SCYON Abstract

Received on October 9 2007

Dust, pulsation, chromospheres and their role in driving mass loss from red giants in Galactic globular clusters

AuthorsIain McDonald and Jacco Th. van Loon
AffiliationAstrophysics Group, School of Physical & Geographical Sciences, Keele University, Staffordshire, ST5 5BG, UK
Accepted byAstronomy & Astrophysics
Contactiain@astro.keele.ac.uk
URL
Links

Abstract

Context: Mass loss from red giants in old globular clusters affects the horizontal branch (HB) morphology and post-HB stellar evolution including the production of ultraviolet-bright stars, dredge up of nucleosynthesis products and replenishment of the intra-cluster medium. Studies of mass loss in globular clusters also allows one to investigate the metallicity dependence of the mass loss from cool, low-mass stars down to very low metallicities. Aims: We present an analysis of new VLT/UVES spectra of 47 red giants in the Galactic globular clusters 47 Tuc (NGC 104), NGC 362, omega Cen (NGC 5139), NGC 6388, M54 (NGC 6715) and M15 (NGC 7078). The spectra cover the wavelength region 6100-9900A at a resolving power of R = 110,000. Some of these stars are known to exhibit mid-infrared excess emission indicative of circumstellar dust. Our aim is to detect signatures of mass loss, identify the mechanism(s) responsible for such outflows, and measure the mass-loss rates. Methods: We determine for each star its effective temperature, luminosity, radius and escape velocity. We analyse the H-alpha and near-infrared calcium triplet lines for evidence of outflows, pulsation and chromospheric activity, and present a simple model for estimating mass-loss rates from the H-alpha line profile. We compare our results with a variety of other, independent methods. Results: We argue that a chromosphere persists in Galactic globular cluster giants and controls the mass-loss rate to late-K/early-M spectral types, where pulsation becomes strong enough to drive shock waves at luminosities above the RGB tip. This transition may be metallicity-dependent. We find mass-loss rates of ~10-7 to 10-5 solar masses per year, largely independent of metallicity.