We compare theoretical dust yields for stars with mass 1 Msun < mstar < 8 Msun, and metallicities 0.001 < Z < 0.008 with observed dust production rates (DPR) by carbon- rich and oxygen-rich Asymptotic Giant Branch (C-AGB and O-AGB) stars in the Large
and Small Magellanic Clouds (LMC, SMC). The measured DPR of C-AGB in the LMC are reproduced only if the mass loss from AGB stars is very efficient during the carbon-star stage. The same yields over-predict the observed DPR in the SMC, suggesting a stronger metallicity dependence of the mass-loss rates during the carbon- star stage. DPR of O-AGB stars suggest that rapid silicate dust enrichment occurs due to efficient hot-bottom-burning if mstar > 3 Msun and Z > 0.001. When compared to the most recent observations, our models support a stellar origin for the existing dust mass, if no significant destruction in the ISM occurs, with a contribution from AGB stars of 70% in the LMC and 15% in the SMC.
A large number of spectroscopic studies have provided evidence of the presence of multiple populations in globular clusters by revealing patterns in the stellar chemical abundances. This paper is aimed at studying the origin of these abundance patter
ns. We explore a model in which second generation (SG) stars form out of a mix of pristine gas and ejecta of the first generation of asymptotic giant branch stars. We first study the constraints imposed by the spectroscopic data of SG stars in globular clusters on the chemical properties of the asymptotic and super asymptotic giant branch ejecta. With a simple one-zone chemical model, we then explore the formation of the SG population abundance patterns focussing our attention on the Na-O, Al-Mg anticorrelations and on the helium distribution function. We carry out a survey of models and explore the dependence of the final SG chemical properties on the key parameters affecting the gas dynamics and the SG formation process. Finally, we use our chemical evolution framework to build specific models for NGC 2808 and M4, two Galactic globular clusters which show different patterns in the Na-O and Mg-Al anticorrelation and have different helium distributions. We find that the amount of pristine gas involved in the formation of SG stars is a key parameter to fit the observed O-Na and Mg-Al patterns. The helium distribution function for these models is in general good agreement with the observed one. Our models, by shedding light on the role of different parameters and their interplay in determining the final SG chemical properties, illustrate the basic ingredients, constraints and problems encountered in this self-enrichment scenario which must be addressed by more sophisticated chemical and hydrodynamic simulations.
In the second paper of this series we pursue two objectives. First, in order to make the code more sensitive to small effects, we remove many approximations made in Paper I. Second, we include turbulence and rotation in the two-dimensional framework.
The stellar equilibrium is described by means of a set of five differential equations, with the introduction of a new dependent variable, namely the perturbation to the radial gravity, that is found when the non-radial effects are considered in the solution of the Poisson equation; following the scheme of the first paper, we write the equations in such a way that the two-dimensional effects can be easily disentangled. The key concept introduced in this series is the equipotential surface. We use the underlying cause-effect relation to develop a recurrence relation to calculate the equipotential surface functions for uniform rotation, differential rotation, rotation-like toroidal magnetic fields and turbulence. We also develop a more precise code to numerically solve the two-dimensional stellar structure and evolution equations based on the equipotential surface calculations. We have shown that with this formulation we can achieve the precision required by observations by appropriately selecting the convergence criterion. Several examples are presented to show that the method works well. Since we are interested in modeling the effects of a dynamo-type field on the detailed envelope structure and global properties of the Sun, the code has been optimized for short timescales phenomena (down to 1 yr). The time dependence of the code has so far been tested exclusively to address such problems.
