Peridotite melting at 1·0 and 1·5 GPa: An experimental evaluation of techniques using diamond aggregates and mineral mixes for determination of near-solidus melts

The experimental determination of liquid compositions in lherzolite as functions of pressure and temperature provides constraints on mantle dynamics and magma genesis. In this paper, we present a detailed evaluation of the use of natural mineral mixes as starting material in peridotite melting studies at 1·0 GPa. As an example we have chosen to test the data obtained by Baker and Stolper (1994, Geochimica et Cosmochimica Acta 58, 2811-2827) on a lherzolite composition (MM-3) presented as a potential source for mid-ocean ridge basalts (MORB). That study is the most fully documented published melting study using natural mineral mixes. We have tested the Baker and Stolper data in three ways: (1) we have defined the liquidus phases and conditions of the partial melt compositions obtained by Baker and Stolper; (2) we have reacted these partial melt compositions with a fine-grained synthetic starting mix of MM-3 composition; (3) we have performed additional melting experiments at 1·0 and 1·5 GPa using the synthetic mix of peridotite MM-3. Our results demonstrate that only the highest temperature experiment of Baker and Stolper, performed at 1390°C, approached an equilibrium melt of peridotite MM-3 composition and that lower temperature experiments have not reached equilibrium, retaining residual unreacted minerals and metastable melt compositions. The degree of disequilibrium increases progressively with lower temperature. Disequilibrium is attributed to the lack of reaction of the natural mineral mix and to disequilibrium melting reactions of the metastable, relatively coarse-grained mineral mix. Other contributing factors include disequilibrium caused by the use of a diamond aggregate trap. We also present peridotite melting experiments using the mineral mix KLB-1 at 1·0 GPa. Our results demonstrate that the mineral mix KLB-1 fails to equilibrate even after ~340 h at temperatures of 1280-1300°C. We present reversals of the 1·0 GPa peridotite melting experiments of Hirose and Kushiro (1993, Earth and Planetary Science Letters 114, 477-489). Our reversals demonstrate that the mineral mix-diamond aggregate trap technique used by Hirose and Kushiro has also failed to produce equilibrium melts of a mantle peridotite composition. It is recommended that data from peridotite melting studies utilizing natural mineral mixes be used with reservation and that natural mineral mixes are not a suitable starting material for such studies. The use of diamond aggregate for separation and trapping of the melt phase compounds rather than solves the problems inherent in the use of natural mineral mixes.