A rare composite xenolith from Salt Lake Crater, Oahu: high-pressure fractionation and implications for kimberlitic melts in the Hawaiian mantle

Abstract

A composite xenolith of olivine-bearing garnet clinopyroxenite wall rock intruded by two spinels + garnet veins is described. Vein minerals exhibit textural evidence of a reaction relationship with the mineral phases in the wall rock. Wall rock clinopyroxene contains exsolved blebby garnet and very fine lamellar exsolution of orthopyroxene, indicating that this xenolith had undergone considerable subsolidus cooling. Garnet–clinopyroxene thermometry suggests that the xenolith last equilibrated in the mantle at a temperature of about 1,060 (±25 °C). The spinels in the veins are of two kinds: pleonaste (that occurs with vein garnet) and a high-Mg, high-Al titanomagnetite (MAT spinels). Intriguingly, the MAT spinels are chemically very similar to the spinels found as groundmass in kimberlites, are moderately subhedral to euhedral, have a weakly developed cumulate texture, and, at places, show a reaction relation with the pleonaste + garnet (cumulate?) assemblage in the vein. Based on petrographic, chemical, and phase equilibrium considerations, we propose the following evolutionary history of this composite xenolith. (1) In the first stage the olivine-bearing garnet clinopyroxenite formed as crystal extracts (cumulates) as a result of high pressure fractionation of an alkaline melt in the deepest levels of Hawaiian lithosphere/uppermost asthenosphere (100–110 km). (2) In the second stage, igneous veining (the melt composition of this vein is not precisely known but could be kimberlitic) occurs in the already existing wall rock resulting in the precipitation of pleonaste + garnet. A reaction relation between the igneous veins and the wall rock also characterizes this stage. (3) The last igneous episode in this xenolith is recorded by MAT spinels in the wall rock and their precipitation close to the previous pleonaste + garnet veins. The last igneous stage could well be due again to high pressure fractionation of a kimberlitic melt (the residual melt after precipitation of pleonaste + garnet). The time relationship between exsolution and the later igneous veining stages is not known. The MAT spinels are not a result of sub-solidus solvus processes as partial reaction (melt present) between the pleonaste + garnet (from the second igneous stage) and MAT spinel exists, pointing to the igneous nature of the MAT spinel. Based on striking similarity between the MAT spinels in our xenolith and those found as groundmass in kimberlites, we propose that the veining stages could well have been kimberlitic. Thus, even though kimberlitic melts are not seen on the Koolau shield, this particular xenolith clearly shows the existence of such melts at great depths beneath Hawaii. We also propose that the initial wall rock, which represents crystal extracts (even though it does not exhibit definitive cumulate texture) as a result of high-pressure fractionation of an alkaline melt and subsequent veining episodes, are of pre-Koolau age. This implies that the Koolau shield volcano may have had a pre-shield alkalic stage.