Oxygen isotope variations in Cr-poor megacrysts from kimberlite

Abstract

As radiogenic isotope compositions of Cr-poor megacrysts from kimberlite demonstrate that the megacrysts and kimberlite hosts are genetically related (megacrysts are likely deep-seated liquidus phases), the oxygen isotope compositions of Cr-poor megacrysts are indicative of the δ18O of the host kimberlites and their mantle source regions. Oxygen isotope ratios of garnet megacrysts from Group I kimberlites (from slightly depleted mantle sources) worldwide (North America, southern Africa, Australia) are restricted (δ18OVSMOW = 5.24‰, SD = 0.15, SE = 0.01, n = 121) and typical of phenocrysts in magmas derived from upper mantle with “normal” oxygen isotope compositions. No significant involvement of subducted oceanic crust is indicated in the genesis of Group I kimberlites. Garnet megacrysts from Group II kimberlites (from “enriched” mantle, and restricted to southern Africa) have anomalously high 18O/16O ratios (δ18OVSMOW = 5.59‰, SD = 0.18, SE = 0.02, n = 55). Similarities in equilibration temperature and major element composition between garnet megacrysts from Group I and II kimberlites rule out roles for these parameters in producing the marked difference in 18O/16O between these two suites. Instead, the high 18O/16O values of the Group II garnets suggests incorporation of anomalously heavy oxygen from subducted ocean floor material in the source region of Group II kimberlites (as much as 30% if eclogite [with δ18O approximately 6.5‰] with an ocean floor protolith is the contaminant, less if the source of anomalous oxygen is pelagic or terrestrial sediment with higher δ18O). Actual samples of potential Group II kimberlite source rocks (e.g., mica/amphibole veined peridotites) or plutonic crystallization products of such kimberlites (e.g., mica-amphibole-rutile-ilmenite-diopside rocks) should have elevated δ18O values, but no such material has been previously described in southern African xenolith suites.