Trace element partitioning between clinopyroxene and silicic alkaline, aluminous melts

Clinopyroxene is considered the most important phase for controlling the partitioning of trace elements during mantle melting. Glasses rich in silica, alumina and alkalis are often found in mantle xenoliths and could coexist with lherzolitic and harzburgitic assemblages, possibly being produced by the diffusive infiltration of alkali process. We have investigated CPX-melt partitioning for two different silica-rich bulk compositions in order to understand how melt polymerization influences element partitioning. Starting materials were synthesized from high purity oxides and corresponding to observed glasses from experiments by Draper and Green (EPSL 170, 255, 1999) and silicic glass found within a harzburgite xenolith studied by Vannucci et al. (EPSL 158, 39, 1998). Experiments were run in a piston cylinder apparatus in unsealed graphite capsules. The Draper composition produces a melt with 62 wt% SiO₂ and 11.8 wt% total alkalis coexisting with olivine (Fo90), orthopyroxene and clinopyroxene at 1.1GPa and 1145° C. The Vannucci composition produces a melt with 67 wt% SiO₂ and 8.1% total alkalis coexisting with diopsidic clinopyroxene at 1 GPa and 1155° C. Selected run products were analyzed for major element composition by EPMA (University of Chicago) and trace element composition by laser ablation ICP-MS (University of Maryland). The most important result is that the heavy rare earth elements (HREE) are compatible in clinopyroxene as previously observed (Blundy et al., EPSL 160, 493. 1998). Four experiments all produce smoothly changing REE patterns with La being slightly incompatible, Sm being slightly compatible and the pattern peaking at Er with DYb always less than DEr. High field strength elements and U and Th have partition coefficients in the low 10^-2 range. These results may imply that element partitioning in the shallow depleted melting column could produce REE patterns similar to those expected for melting in the garnet stability field.