Hot bottom burning in asymptotic giant branch stars and its effect on oxygen isotopic abundances

A self-consistent calculation of asymptotic giant branch (AGB) evolution was carried out, including nucleosynthesis at the base of the convective envelope (hot bottom burning). Hot bottom burning was found to occur for stars between approximately 4.5 and approximately 7 solar mass, producing envelopes with O-18/O-16 less than or equal to 10(exp -6) and 10(exp -3) approximately less than or equal O-17/O-16 approximately less than or equal to 10(exp -1). The O-17 abundance depends sensitively on the nuclear O-17-destruction rate; this rate is only loosely constrained by the requirement that first and second dredge-up models match O-isotope observations of red giant branch (RGB) stars (Boothroyd, Sackmann, & Wasserburg 1994). In some cases, high mass-loss rates can terminate hot bottom burning before further O-17 enrichment takes place or even before all O-18 is destroyed. These predictions are in accord with the very limited stellar observations of J type carbon stars on the AGB and with some of the circumstellar Al2O3 grains from meteorites. In contrast, precise data from a number of grains and data from most low-mass S and C AGB stars (approximately less than 1.7 solar mass) lie in a region of the O-18/O-16 versus O-17/O-16 diagram that is not accessible by first and second dredge-up or by hot bottom burning. We conclude that for AGB stars, the standard models of stellar evolution are not in accord with these observations. We surmise that an additional mixing mechanism must exist that transports material from the cool bottom of the stellar convective envelope to a depth at which O-18 is destroyed. This 'cool bottom processing' mechanism on the AGB is similar to extra mixing mechanisms proposed to explain the excess C-13 (and depleted C-12) observed in the earlier RGB stage of evolution and the large Li-7 depletion observed in low-mass main-sequence stars.