The origin of garnet and clinopyroxene in "depleted" Kaapvaal peridotites

A detailed petrographic, major and trace element and isotope (Re-Os) study is presented on 18 xenoliths from Northern Lesotho kimberlites. The samples represent typical coarse, low-temperature garnet and spinel peridotites and span a P-T range from similar to 60 to 150 km depth. With the exception of one sample (that belongs to the ilmenite-rutile-phlogopite-sulphide suite (IRPS) suite first described by [B. Harte, P.A. Winterburn, J.J. Gurney, Metasomatic and enrichment phenomena in garnet peridotite facies mantle xenoliths from the Matsoku kimberlite pipe, Lesotho. In: Menzies, M. (Ed.), Mantle metsasomatism. Academic Press, London 1987, 145-220.]), all samples considered here have high Mg# and show strong depletion in CaO and Al2O3. They have bulk rock Re depletion ages (T-RD) >2.5 Ga and are therefore interpreted as residua from large volume melting in the Archaean. A characteristic of Kaapvaal xenoliths, however, is their high SiO2 concentrations, and hence, modal orthopyroxene contents that are inconsistent with a simple residual origin of these samples. Moreover, trace element signatures show strong overall incompatible element enrichment and REE disequilibrium between garnet and clinopyroxene. Textural and subtle major element disequilibria were also observed. We therefore conclude that garnet and clinopyroxene are not co-genetic and suggest that (most) clinopyroxene in the Archaean Kaapvaal peridotite xenoliths is of metasomatic origin and crystallized relatively recently, possibly from a melt precursory to the kimberlite. Possible explanations for the origin of garnet are exsolution from a high-temperature, Al- and Ca-rich orthopyroxene (indicating primary melt extraction at shallow levels) or a majorite phase (primary melting at >6 GPa). Mass balance calculations, however, show that not all garnet observed in the samples today is of a simple exsolution origin. The extreme LREE enrichment (sigmoidal REE pattern in all garnet cores) is also inconsistent with exsolution from a residual orthopyroxene. Therefore, extensive metasomatism and probably re-crystallization of the lithosphere after melt-depletion and garnet exsolution is required to obtain the present textural and compositional features of the xenoliths. The metasomatic agent that modified or perhaps even precipitated garnet was a highly fractionated melt or fluid that might have been derived from the asthenosphere or from recycled oceanic crust. Since, to date, partitioning of trace elements between orthopyroxene and garnet/clinopyroxene is poorly constrained, it was impossible to assess if orthopyroxene is in chemical equilibrium with garnet or clinopyroxene. Therefore, further trace element and isotopic studies are required to constrain the timing of garnet introduction/modification and its possible link with the SiO2 enrichment of the Kaapvaal lithosphere. (C) 2003 Elsevier B.V All rights reserved.