Abstract English

The matter injected to the interstellar medium as a consequence of stellar nucleosynthesis and evolution is a key ingredient to our understanding of the chemical evolution of galaxies. This project investigates the mixing events (dredge up) in low and intermediate mass stars, which bring material processed by nuclear reactions from the stellar interior to the surface of the star during the late stages of stellar evolution (asymptotic giant branch / AGB phase). We propose to obtain abundances for the key elements C and O and some of their isotopes for a sample of single AGB stars in galactic and extragalactic globular clusters covering a wide range in mass (age) and metallicity. The innovative approach of this study lies in the use of pulsational properties for mass determination of the observed stars and of new model atmospheres for deriving abundances from high and medium resolution near infrared spectra. The first aspect will allow to pin down stellar mass more accurately than in previous studies of single stars. The second aspect will include the self consistent use of spherical atmospheric structure and spherical radiative transfer (Aringer 2004) as well as a first time study on the influence of pulsation on the abundance determination by using most recent dynamical model atmospheres (Höfner et al. 2003). Both aspects, knowing the mass of the studied object and dealing with pulsational effects on the investigated atmospheres are crucial for comparing models of stellar evolution and mixing with observations.

Our investigation shall set tight observational constraints on the lower mass limit for dredge up during the AGB phase at various metallicities. We will compare our findings with current nucleosynthesis and mixing models testing also new approaches for extramixing (Busso et al. 2004). This will help to tune the evolutionary models (Straniero et al. 2003) and will allow for improved predicitions of stellar yields for stellar and galactic evolution models. Our study of the influence of dynamical effects on abundance determination will show for which stars hydrostatic models can be used and which spectral lines are least affected by dynamical effects. As a side product we will derive accurate period-luminosity relations for AGB variables in different stellar environments and give a new calibration for the mass loss by comparison of observations with current pulsation models (Wood et al. 1999).

Abstract German

german abstract