Garnet and pyroxenes in the mantle: A test of the majorite fractionation hypothesis

Abstract

Experimental data are reported for the compositions of garnets and pyroxenes that coexist with liquid in the system CaO‐MgO‐Al2O3‐SiO2 (CMAS) at 80–160 kbar. All experiments were run using the uniaxial split sphere apparatus at Stony Brook. Our data have permitted a refinement of Gasparik's (1990a) thermodynamic analysis of garnets and pyroxenes that occur at arid below the solidus; our results should provide precise estimates of the temperature‐pressure conditions at which peridotite and garnet xenoliths equilibrated in the mantle. An examination is also made of the composition of garnet that crystallizes as a liquidus phase; it is demonstrated to be a function of both composition and pressure. Pressure has the effect of changing the composition of garnet from pyropic at around 80 kilobars to pyroxene‐like (majorite) at pressures in excess of 140 kbar, changes that mirror those seen for garnets in the subsolidus. Failure to consider the effects of pressure on the composition of liquidus garnet can result in an erroneous test of the hypothesis that garnet fractionation occurred in early Earth history. It is shown that primitive mantle peridotite could have formed as a liquid by 35% majorite garnet fractionation from a CV chondrite like Allende at 200–260 kbar. These results support the hypothesis that extensive or possibly total melting occurred during the growth of the Earth by accretion of planetesimals (e.g., Davies, 1985; Abe and Matsui, 1985; Stevenson, 1987; Wetherill, 1990) and that majorite fractionation could have occurred during this time (Ohtani and Sawamoto, 1987). Removal of majorite to the lower mantle during the crystallization of a magma ocean is a successful way of explaining differences in MgO/SiO2 and CaO/Al2O3 between mantle peridotite and chondrite.