A radiogenic Os component in the oceanic lithosphere? Constraints from Hawaiian pyroxenite xenoliths

Platinum Group Element (PGE) concentrations in garnet pyroxenite xenoliths from Oahu, Hawaii, are significantly lower than those in mantle peridotites and show fractionated patterns (e.g. PdN/OsN = 2–10, PdN/IrN = 4–24; N = chondrite normalized) and very high ReN/OsN ratios ( 9–248). Mass balance calculations show that the bulk rock pyroxenite PGE inventory is controlled by the presence of sulfide phases. The Os/Os ratios of these pyroxenites vary from subchondritic to suprachondritic (0.123–0.164); and the Os/Os ratios show good correlations with bulk rock and clinopyroxene major and trace element compositions, and bulk rock PGE and sulfur abundances. These observations suggest that the Os isotope compositions in these pyroxenites largely reflect primary processes in the oceanic mantle and Pacific lithosphere. In contrast, bulk rock Os/Os ratios do not correlate with other lithophile isotopic tracers (e.g. Rb–Sr, Sm–Nd, Lu–Hf) which show limited isotopic variability (Bizimis et al., 2005). This and the lack of Os/Os vs. Re/Os correlations suggest that the range in Os isotope ratios is not likely the result of mixing between long-lived depleted and enriched components or aging of these pyroxenites within the Pacific lithosphere after its formation at a mid-oceanic ridge setting some 80– 100 million years ago. We interpret the Os isotopes, PGE and lithophile element systematics as the result of melt–lithosphere interaction at the base of the Pacific lithosphere. The major and trace element systematics of the clinopyroxenes and bulk rock pyroxenites and the relatively constant lithophile element isotope systematics are best explained by fractional crystallization of a rather homogenous parental magma. We suggest that during melt crystallization and percolation within the lithosphere, the parental pyroxenite melt assimilated radiogenic Os from the grain boundaries of the peridotitic lithosphere. This radiogenic Os component may reside in the grain boundary sulfides or other trace phases, and may be due to fluids or melts that had previously percolated through the basal part of the lithosphere during its transit from a mid-oceanic ridge to its present position above the Hawaiian plume. As the solidus of the parental pyroxenite melt is lower than the solidus of the lithospheric peridotite, we envision that the pyroxenite–parent melt selectively assimilated the grain boundary sulfide phases with lower melting temperature as it percolated through the lithosphere, without significantly reacting with the silicate minerals. Thus while the parental melt of these pyroxenites originate within the Hawaiian plume, melt–lithosphere interaction during progressive crystallization may have selectively enriched the resulting melts with radiogenic Os, thereby decoupling Os from the lithophile element isotopes, but retaining a link between Os, PGE and fractional crystallization systematics. In this model, Oahu pyroxenites essentially represent melts from different stages of this melt–mantle reaction process at the base of the lithosphere, and we suggest that this process may also explain the similar Os vs. lithophile element decoupling seen in the rejuvenated volcanism in Oahu and Kauai. We further show that the pyroxenites do not posses the requisite Pt/Re ratios, where upon, recycling and aging would generate the coupled enrichments of Os–Os isotope ratios observed in Hawaiian and other lavas. 2011 Elsevier Ltd. All rights reserved. 0016-7037/$ see front matter 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.gca.2011.06.008 ⇑ Corresponding author. Present address: Department of Marine Chemistry and Geochemistry, Woods Hole Oceanographic Institution, 360 Woods Hole Rd., Mailstop #25, Woods Hole, MA 02543-1541, USA. Tel.: +1 508 289 3339; fax: +1 508 457 2193. E-mail address: [email protected] (I.S. Sen). 4900 I.S. Sen et al. /Geochimica et Cosmochimica Acta 75 (2011) 4899–4916