▪ Abstract We present a review of nucleosynthesis in AGB stars outlining the development of theoretical models and their relationship to observations. We focus on the new high resolution codes with improved opacities, which recently succeeded in accounting for the third dredge-up. This opens the possibility of understanding low luminosity C stars (enriched in s-elements) as the normal outcome of AGB evolution, characterized by production of¹²C and neutron-rich nuclei in the He intershell and by mass loss from strong stellar winds. Neutron captures in AGB stars are driven by two reactions:¹³C(α,n)¹⁶O, which provides the bulk of the neutron flux at low neutron densities (N_n≤ 10⁷n/cm³), and²²Ne(α,n)²⁵Mg, which is mildly activated at higher temperatures and mainly affects the production of s-nuclei depending on reaction branchings. The first reaction is now known to occur in the radiative interpulse phase, immediately below the region previously homogenized by third dredge-up. The second reaction occurs during the convective thermal pulses. The resulting nucleosynthesis phenomena are rather complex and rule out any analytical approximation (exponential distribution of neutron fluences). Nucleosynthesis in AGB stars, modeled at different metallicities, account for several observational constraints, coming from a wide spectrum of sources: evolved red giants rich in s-elements, unevolved stars at different metallicities, presolar grains recovered from meteorites, and the abundances of s-process isotopes in the solar system. In particular, a good reproduction of the solar system main component is obtained as a result of Galactic chemical evolution that mixes the outputs of AGB stars of different stellar generations, born with different metallicities and producing different patterns of s-process nuclei. The main solar s-process pattern is thus not considered to be the result of a standard archetypal s-process occurring in all stars. Concerning the¹³C neutron source, its synthesis requires penetration of small amounts of protons below the convective envelope, where they are captured by the abundant¹²C forming a¹³C-rich pocket. This penetration cannot be modeled in current evolutionary codes, but is treated as a free parameter. Future hydrodynamical studies of time dependent mixing will be required to attack this problem. Evidence of other insufficiencies in the current mixing algorithms is common throughout the evolution of low and intermediate mass stars, as is shown by the inadequacy of stellar models in reproducing the observations of CNO isotopes in red giants and in circumstellar dust grains. These observations require some circulation of matter between the bottom of convective envelopes and regions close to the H-burning shell (cool bottom processing). AGB stars are also discussed in the light of their possible contribution to the inventory of short-lived radioactivities that were found to be alive in the early solar system. We show that the pollution of the protosolar nebula by a close-by AGB star may account for concordant abundances of²⁶Al,⁴¹Ca,⁶⁰Fe, and¹⁰⁷Pd. The AGB star must have undergone a very small neutron exposure, and be of small initial mass ([ ]). There is a shortage of²⁶Al in such models, that however remains within the large uncertainties of crucial reaction rates. The net²⁶Al production problem requires further investigation.

The recently improved information on the stellar (n,gamma) cross sections of neutron-magic nuclei at N = 82, and in particular of 142Nd, turned out to represent a sensitive test for models of s-process nucleosynthesis. While these data were found to be incompatible with the classical approach based on an exponential distribution of neutron exposures, they provide significantly better agreement between the solar abundance distribution of s nuclei and the predictions of models for low mass AGB stars. Particular attention is paid to a consistent description of s-process branchings in the region of the rare earth elements. It is shown that - in certain cases - the nuclear data are sufficiently accurate that the resulting abundance uncertainties can be completely attributed to stellar modelling. Thus, the s process becomes important for testing the role of different stellar masses and metallicities as well as for constraining the assumptions for describing the low neutron density provided by the 13C source.Comment: 49 pagg., 5 figg., accepted for publication by Ap

We present the results of s-process nucleosynthesis calculations for AGB stars of different metallicities and initial masses. The computations were based on previously published stellar evolutionary models that account for the III dredge up phenomenon occurring late on the AGB. Neutron production is driven by the 13C(alpha,n)16O reaction during the interpulse periods in a tiny layer in radiative equilibrium at the top of the He- and C-rich shell. The s-enriched material is subsequently mixed with the envelope by the III dredge up, and the envelope composition is computed after each thermal pulse. We follow the changes in the photospheric abundance of the Ba-peak elements (heavy s, or `hs') and that of the Zr-peak ones (light s, or `ls'), whose logarithmic ratio [hs/ls] has often been adopted as an indicator of the s-process efficiency. The theoretical predictions are compared with published abundances of s elements for Galactic AGB giants of classes MS, S, SC, post-AGB supergiants, and for various classes of binary stars. The observations in general confirm the complex dependence of n captures on metallicity. They suggest that a moderate spread exists in the abundance of 13C that is burnt in different stars. Although additional observations are needed, a good understanding has been achieved of s-process operation in AGB. The detailed abundance distribution including the light elements (CNO) of a few s-enriched stars at different metallicity are examined.Comment: Accepted for ApJ, 59 pages, 19 figures, 5 table