Magnesium and titanium partitioning between anorthite and Type B CAI liquid: Dependence on oxygen fugacity and liquid composition

Abstract

Experiments were conducted in air and at low oxygen fugacity (f_O_2)) to evaluate Mg and Ti partitioning between anorthite and liquid (D_(Mg) and D_(Ti) in a synthetic composition similar to that of a Type B Ca, AI-rich inclusion (CAI). The starting material showed a range of compositions, which allowed assessment of the composition dependence of D_(Mg) and D_(Ti) in this system. Additional experiments using a homogeneous split of the same material investigated the effect of oxygen fugacity on the partitioning of Ti^(3+) and Ti^(4+) between anorthite and liquid. The low foe charges were purple, consistent with the presence of significant amounts of Ti^(3+).This was verified by electron spin resonance (ESR) spectra, and quantitative estimates of Ti^(3+) contents were obtained using ESR. The Ti and Mg partition coefficients in the air run using the homogeneous starting material are similar (0.034 and 0.036, respectively) and consistent with those determined in other studies. However, D_(Ti) at low f_(O_2) is slightly greater than D_(Ti) in the air experiments. Using Ti^(3+)/total Ti from the ESR measurements, D_(Ti^(3+)) is calculated to be about 0.040. The range of compositions reveal a clustering of D_(Mg) and D_(Ti) within charges, but a wide range of D_s between charges of different composition. A well-defined inverse correlation exists between D_(Mg) and D_(Ti). This variation is not due to temperature-dependence, but is instead due to the dependence of D_(Mg) and D_(Ti) on liquid composition (Si and Al in particular). D_(Mg) correlates positively with Si content and negatively with Al content, while D_(Ti) shows the opposite correlations. The results of these experiments have interesting implications for the petrogenesis of Type B CAIs and for substitution mechanisms of Mg, Ti^(4+), and Ti^(3+) into anorthite. Crystallization models for Type B CAIs permit certain predictions concerning trace element systematics in plagioclase. The Mg and Ti systematics are best explained by a fractional crystallization model where plagioclase crystallizes very late (>95% crystallization), and D_(Ti^(3+)). is equal to D_(Ti^(4+)). The results from our experiments support this model for the relative partitioning of Ti^(4+) and Ti^(3+) between plagioclase and liquid. In addition, the dependence of D_(Mg), and D_(Ti) on the Si content of a Type B CAI liquid helps explain systematics expected during late-stage crystallization of plagioclase. The composition dependence of D_(Mg) and D_(Ti) also allows assessment of substitution mechanisms in anorthite using a crystallization reaction approach. Using these methods, a plausible mechanism for Mg involves substitution for tetrahedral A1 by the reaction Mg^(2+) + Si^(4+) = 2AI^(3+), consistent with that proposed by previous workers. The systematics are also consistent with Ti^(4+) and Ti^(3+) substitution for tetrahedral Si^(4+) by the reactions 2Al^(3+) + Ti^(4+) = Ca^(2+) + 2Si^(4+) and Al^(3+) + Ti^(3+) = Ca^(2+) + Si^(4+), respectively.