Extra-Mixing During the Red Giant Branch Evolution of Low-Mass Stars

Abstract

It is thought that rotation leads to extra-mixing between the hydrogenburning shell and the convective envelope in low-mass red giants. The atmospheric abundances of Li, C, and N are therefore expected to change with luminosity on the upper RGB. This theoretical prediction is in excellent agreement with the observed RGB abundance variations in both field and globular-cluster red giants. The latest observational data on the evolutionary variations of the surface chemical composition in low-mass metal-poor stars are used to constrain the basic properties of extra-mixing in the majority of upper RGB stars. A possible mechanism of this ”canonical” extramixing is turbulent diffusion driven by rotation. The progress in modeling this process is reviewed. In addition, observational evidence for primordial composition modifications and ideas about their origin in the context of extra-mixing are discussed. 1 Canonical Low-Mass Star Evolution The evolution of low-mass stars (M ∼< 2− 2.5M ) starts with central hydrogen burning on the main-sequence. Due to the rather low core temperatures, protonnucleosynthesis affects only the nuclei involved in the pp-chains and CNO-cycles. Even for the lower mass range of, say 0.7−0.9M , the initial C → N conversion takes place, although the whole cycle (involving changes in O) is completed only rarely. Nuclei of higher cycles, such as those of the NeNa and MgAl cycles, are not affected. At the end of the main-sequence, the hydrogen-burning shell establishes itself within the former burning core, such that above the shell we find a region of composition modified by H-burning: a C/C value close to the equilibrium value of ≈ 3, almost vanishing carbon, and strongly enhanced nitrogen abundance. In addition, He is increased by huge factors in those regions of the former core where temperature was low enough for a high He equilibrium abundance within the ppI-chain, but high enough for nuclear reactions establishing the equilibrium quickly [8]. Finally, Li is strongly depleted over a large fraction of the stellar interior, due to its low p-capture temperature of ≈ 2.5 · 10 K. As the star becomes cooler on the subgiant and early red giant branch, the convective envelope deepens until it touches these regions and mixes the modified material to the surface. This is the first dredge-up (1 du, [18]), which is the only mixing event predicted by canonical stellar evolution theory for such stars. The resulting changes in element and isotope abundances are shown in Fig. 1. After the 1 du the convective envelope retreats ahead of the advancing hydrogen-shell and no further changes in the surface abundances are expected. Extra Mixing in Low-Mass Red Giants 299 When the shell encounters the point of deepest convective penetration, the evolution is slowed down, resulting in the so-called bump, and, more importantly, no molecular weight barrier between shell and convective envelope remains. Theory and observation nicely agree about the size and location (around logL/L = 2) of the bump [28]. Fig. 1. Changing surface composition of a metal-poor 0.9M star during its RGB evolution (courtesy C. Charbonnel) This picture is supported by observations, one important set of data coming from [17] for field stars with accurate Hipparcos parallaxes (see Fig. 2). The strength of the 1 du increases with decreasing metallicity, with mass up to ≈ 1M , decreasing again for higher masses, and with decreasing initial helium content. It also leads to a slight enrichment of the convective envelope with helium. For more massive stars, the energy generation of which is dominated by the CNO-cycle, a slight decrease in oxygen would be expected, too. 2 Observational Evidence for Extra-Mixing [17] not only demonstrated the reality of the 1 du as predicted consistently by all stellar evolution calculations, but – more importantly – showed that in field stars a further decrease of Li, C, C/C, and an accompanying increase in N takes place along the upper RGB. This additional evolutionary effect can not be explained by canonical evolution, but can be fully accounted for by requiring an additional extra-mixing between the bottom of the convective envelope and the outermost layers of the advancing hydrogen-shell as demonstrated in Fig. 2, taken from [12]. The shell is always hot enough for CNO-burning being the main energy source, and thus the CN-conversion takes place in regions outside the major energy producing layers. This explanation of the abundance