Formation of orange hibonite, as inferred from some Allende inclusions

We studied three fluffy Type A refractory inclusions from Allende that cont ain orange hibonite. The melilite in the present samples is very Al-rich, a veraging Angstromk(6), Angstromk(14), and Angstromk(12) in the three sample s studied. Hibonite in two inclusions, unlike that in Murchison, has low ra re earth element abundances of <10 x Cl; in the other inclusion, the hiboni te, melilite and perovskite have Group II-like patterns. The hibonite and m elilite in all three inclusions studied have excess Mg-26 consistent with ( Al-26/Al-27)(I) = 5 x 10(-5). Much of the hibonite and some of the spinel i n these inclusions is corroded. These phases are found enclosed in melilite , but based on bulk compositions and phase equilibria, hibonite should not be an early-crystallizing phase in these inclusions. We conclude that the h ibonite and probably some of the spinel is relic. Reversely zoned melilite, rounded spinel and isotopically heavy Mg in the inclusions probably reflec t reheating events that involved melting and evaporation. Alteration of the gehlenitic melilite gave rise to some rare phases, including corundum and nearly pure CaTs pyroxene. Studies have shown that blue hibonite contains Ti3+ while orange hibonite d oes not (Ihinger and Stolper, 1986; Beckett et al., 1988). Orange hibonite formed either under oxidizing conditions (such as at oxygen fugacities at l east seven orders of magnitude greater than that of a solar gas at 1700 K), or under conditions reducing enough (e.g., solar) that it contained Ti3+, which was later oxidized in situ. Although V and Ce oxides are volatile at the temperature and range of oxygen fugacities at which orange hibonite is known to be stable, we find that (a) the hibonite is V-rich (<similar to>1 wt% V2O3) and (b) there are no negative Ce anomalies in Allende hibonite. T his indicates that the hibonite did not form by condensation under oxidizin g conditions. In addition, there are slight excesses of Ti + Si cations rel ative to Mg + Fe cations (up to 0.1 of 0.8 cations per 19 oxygen anions), p robably reflecting the original presence of Ti3+. The results of this study strongly support the suggestion (Ihinger and Stolper, 1986) that Allende h ibonite originally formed under reducing conditions and was later oxidized. Oxygen fugacities within similar to2-3 orders of magnitude of that of a so lar gas are implied; otherwise, strong Ce and V depletions would be observe d.