Clues from Garnet-Spinel-Bearing Peridotite and Pyroxenite Xenoliths about the Formation of the Subcratonic Lithospheric Mantle—a Case Study from the Orapa Kimberlite Cluster

Abstract

Abstract A series of 18 ultramafic mantle xenoliths from the Orapa and Letlhakane mines containing both garnet and spinel were investigated petrographically and by means of classical and multi-equilibrium geothermobarometry in order to understand their P–T evolution and infer tectonic processes consistent with that evolution, assuming that samples from the shallow subcontinental lithospheric mantle (SCLM), may have remained cool enough to preserve memory of its formation. Samples broadly classify into peridotites, mainly harzburgites, and pyroxenites, mainly (olivine-) websterites with less than 15 vol% olivine. The websterites all show thin garnet coronas around xenomorphic spinel clusters and often coarse garnet and very fine spinel exsolving from pyroxenes. Classical geothermobarometry results vary consistently between the 40 mW/m2 geotherm at high pressures and the 50 mW/m2 geotherm at lower pressures but partly fail the reliability test indicating some degree of disequilibrium, particularly, as may be expected, toward lower P and T. Eleven samples were analyzed in detail for mineral modes, and pseudosections were calculated for the four compositionally most extreme samples. All four samples show systematic divergence of the mode and composition isopleths indicating overall chemical disequilibrium, but partial isopleth convergence in two distinct P–T regions that were interpreted as the main equilibration stage and a stage of late ‘freeze-in’ of a few composition isopleths at lower temperature and higher pressure. Specifically, the P–T changes between stage 1 and stage 2 were from 900°C and 13 kbar to 750°C and 23 kbar for a websterite and from 850°C and 24 kbar to 730°C and 33 kbar for a harzburgite. For an olivine-orthopyroxenite and -clinopyroxenite the change was from 910°C, 17 kbar and 770°C, 17 kbar for stage 1, respectively, to less well-constrained conditions for stage 2 of 650°C, 36 kbar and 630°C, 20 kbar, respectively. This interpretation of a protracted or stepwise freeze-in along a P–T path holds even after considering thermodynamic, analytical and, as far as possible, geological uncertainties. The result is consistent with a ‘subcretion’ or ‘buoyant subduction’ model for the formation of the SCLM, specifically with the down-drag of an already subcreted lithospheric platelet by a subsequently arriving platelet. In the framework of the crust–mantle evolution in the Orapa region this most likely occurred in the Neoarchean, between 2.8 and 2.5 Ga.