Experimental Determination of Composition of Melt Formed by Equilibrium Partial Melting of Peridotite at High Pressures Using Aggregates of Diamond Grains.

Abstract

The composition of melt formed by equilibrium partial melting of peridotite at high pressures using a new experimental method is described. A thin layer of diamond powder or a small loosely sintered diamond chip is sandwiched between finely ground peridotite and heated to partially melt the host peridotite at high pressures. Melt formed by partial melting of peridotite segregates and migrates into the pore space between diamond grains. The melt equilibrates with the host peridotite only after the pore space is completely filled with melt. Microprobe analysis of segregated quenched melt trapped between diamond grains can be made without problems involving quenching or overgrowth of coexisting crystals. Melt which fills the pore space of the diamond grains is equilibrated with residual solids within 24 hours at 1400°C in the synthetic system Mg2Si04-NaA1Sio4-Si02 at 10, 15 and 20 kbar. This method can be applied to determine both compatible and incompatible elements in melt formed by partial melting of peridotite or any other rocks at high pressures. Using this method the compositions of melts formed by partial melting of a spinel lherzolite were determined at pressures between 10 and 20 kbar. Introduction. Determination of the composition of melt formed by partial melting of peridotites at high pressures is important for understanding genesis of magmas formed in the upper mantle. Many experimental studies have been made to determine such compositions in both synthetic and natural peridotite systems using several different methods. Estimation of the melt composition was first made on the basis of the liquidus phase relations at high pressures. 11) This approach gives constraints on the melt compositions over a wide pressure range, but it cannot give specific melt compositions. Direct microprobe analysis of quenched melts formed by partial melting of peridotites gives more specific melt compositions5~-7~; however, the compositions of melts are significantly modified due to overgrowth of minerals during quenching and the direct analyses do not represent the compositions of equilibrium melts in most cases. Partial melt compositions, based on the modal analysis of the run products and microprobe analyses of residual minerals, have also been calculated.8 This method does not involve the above-mentioned quench problem, but the calculation involves significant uncertainties originating from small errors in modal analysis and presence of relic minerals. Melting experiments with peridotite capsules9~ and those with the sandwich method1o>-12) were successful in the determination of major elements or "compatible elements" of partial melts; however, the compositions, particularly the abundances of incompatible elements, of partial melts and host peridotites are affected by the materials added to peridotite and do not represent the true partial melt compositions. If analysable amounts of partial melts can be separated from host peridotite, the true partial melt compositions can be determined. In the present experiments, analysable