Phase relationships at 30 kbar for quartz eclogite composition in CaO-MgO-Al_2O_3-SiO_2-H_2O with implications for subduction zone magmas

Abstract

Synthetic starting material, corresponding to oceanic tholeiite (analysis wt.%: SiO_2, 52.5; Al_2O_3, 15.6; MgO, 14.6; CaO, 17.3), was run in sealed capsules at 30 kbar with 3.5% to 32.5% H_2O for comparison with the phase diagram previously reported for a natural gabbro-H_2O. The subsolidus assemblage is zoisite-coesite-eclogite, with the solidus at 810°C. Coesite/ quartz disappears just above the solidus, zoisite between 900° and 950°C, eclogite + liquid persists through a wider temperature interval, and clinopyroxene + liquid exists below the liquidus. The liquidus is at 1110°C with excess H_2O, and 1360°C with 7.5% H_2O. Compositions of glass, clinopyroxene, garnet and zoisite were measured by electron microprobe for runs with 7.5% H_2O. Glass analyses could not be obtained for runs below 1050°C. Glass analyses for runs at 1100°C and 1200°C included values that could not represent original liquid compositions, suggesting that the liquid composition changes during the quench, for reasons not fully understood. Equilibrium compositions of liquids were calculated from estimated modes and mass balance. The equilibrium liquid paths for both model and natural-rock compositions diverge from the average chemical variation of the calc-alkaline rock series, with divergence greatest for about 20-25% liquid where SiO_2 content corresponds to that of andesite. Partial melting of quartz eclogite in subducted oceanic crust at 100 km depth would produce liquids with Ca/(Mg+Fe) higher than andesites. Therefore, andesite is not a primary magma from subducted oceanic crust deeper than amphibole breakdown.

Additional details