Slow neutron captures are responsible for the production of about 50% of elements heavier than iron, mainly, occurring during the asymptotic giant branch phase of low-mass stars (1 M/M ⊙ 3), where the main neutron source is the 13 C(α,n) 16 O reaction. This last is activated from locally-produced 13 C, formed by partial mixing of hydrogen into the He-rich layers. We present here the first attempt at describing a physical mechanism for the formation of the 13 C reservoir, studying the mass circulation induced by magnetic buoyancy and without adding new free parameters to those already involved in stellar modelling.Our approach represents the application, to the stellar layers relevant for s-processing, of recent exact, analytical 2D and 3D models for magneto-hydrodynamic processes at the base of convective envelopes in evolved stars in order to promote downflows of envelope material for mass conservation, during the occurrence of a dredge-up phenomenon. We find that the proton penetration is characterized by small concentrations, but extended over a large fractional mass of the He-layers, thus producing 13 C reservoirs of several 10 −3 M ⊙ .The ensuing 13 C-enriched zone has an almost flat profile, while only a limited production of 14 N occurs. In order to verify the effects of our new findings we show how the abundances of the main s-component nuclei can be accounted for in solar proportions and how our large 13 C-reservoir allows us to solve a few so far unexplained features in the abundance distribution of post-AGB objects.

Low-mass asymptotic giant branch stars are among the most important polluters of the interstellar medium. In their interiors, the main component (A ≳ 90) of the slow neutron capture process (the s-process) is synthesized, the most important neutron source being the 13C(α,n)16O reaction. In this paper, we review its current experimental status, discussing possible future synergies between some experiments currently focused on the determination of its rate. Moreover, in order to determine the level of precision needed to fully characterize this reaction, we present a theoretical sensitivity study, carried out with the FUNS evolutionary stellar code and the NEWTON post-process code. We modify the rate up to a factor of 2 with respect to a reference case. We find that variations of the 13C(α,n)16O rate do not appreciably affect s-process distributions for masses above 3 M ⊙ at any metallicity. Apart from a few isotopes, in fact, the differences are always below 5%. The situation is completely different if some 13C burns in a convective environment: this occurs in FUNS models with M < 3 M ⊙ at solar-like metallicities. In this case, a change of the 13C(α,n)16O reaction rate leads to nonnegligible variations of the element surface distribution (10% on average), with larger peaks for some elements (such as rubidium) and neutron-rich isotopes (such as 86Kr and 96Zr). Larger variations are found in low-mass, low-metallicity models if protons are mixed and burned at very high temperatures. In this case, the surface abundances of the heavier elements may vary by more than a factor of 50.

We present here the application of a model for a mass circulation mechanism in between the H-burning shell and the base of the convective envelope of low mass AGB stars, aimed at studying the isotopic composition of those presolar