A fertile harzburgite–garnet lherzolite transition: possible inferences for the roles of strain and metasomatism in upper mantle peridotites

Abstract

Porphyroclastic enstatite in a garnet lherzolite xenolith from the Monastery Mine kimberlite, South Africa, has exsolved pyrope garnet, Cr-diopside and Al-chromite, and the specimen is interpreted as representing a transition from fertile harzburgite, (containing high Ca-Al-Cr enstatite) to granular garnet lherzolite. Although the exsolved phases occur in morphologically different forms (fine and coarse lamellae; equant, ripened grains), indicating textural disequilibrium, the exsolved grains are very constant in composition, indicating chemical equilibrium. Theoretically, the exsolution could have been due to a fall in temperature, but the close association of exsolution and deformation of the host enstatite suggests that exsolution was also aided by straining of the enstatite lattice. The phase compositions can be broadly matched with those in other mantle peridotites, except that all phases are characterised by a virtual absence of Ti. In the garnet and diopside Ti, Co, Zr and most of the REE are lower than in published analyses of garnet and diopside in both granular and sheared garnet lherzolites from Southern African kimberlites, and diopside/garnet partitioning for Sr and the REE is higher. Comparison with the trace element chemistry of an enstatite from a fertile harzburgite indicates that, except for Nb, the trace element content and distribution found in the Monastery phases could arise by isochemical exsolution from such an enstatite. On the assumption that (a) the Monastery specimen represents a transition from harzburgite to garnet lherzolite, and (b) many garnet lherzolites are of exsolution origin (as suggested by their modal compositions), the inference is that most garnet lherzolites, and not just the sheared variety, have been subject to varying degrees of Ti, Zr, Sr and REE metasomatism.