Selenium isotope and S-Se-Te elemental systematics along the Pacific-Antarctic ridge: Role of mantle processes

Abstract

The selenium stable isotope system emerges as a new potential tracer of volatile origin and evolution of the terrestrial planets. Accurate determination of the mantle Se isotope composition requires an assessment of Se isotopic behavior in magmatic processes and potential variations across all mantle reservoirs. Here we report the first high-precision Se isotope and Se–Te abundance data for a suite of basaltic glasses from the Pacific–Antarctic ridge. These MORBs display a narrow range in δ82/76Se values (deviation of 82Se/76Se relative to NIST SRM 3149) between −0.30 ± 0.09‰ and −0.05 ± 0.09‰, with an average of −0.16 ± 0.13‰ (2 s.d., n = 27). We quantify the main processes relevant to MORB petrogenesis in order to better understand the Se–Te elemental behavior in the mantle and investigate if these are systematically related to Se isotope variations. We show that both Se isotopes and S–Se–Te abundances of MORB melts remain unaffected by assimilation of high-temperature hydrothermal fluids and sulfides, whereas the latter has been shown to overprint the 34S/32S ratios. MORB differentiation involving sulfide segregation (sulfide liquid and monosulfide solid solution) significantly fractionates Se and Te (Se/Te ratio ∼ 45–190), with no systematic Se isotope variation. The Se–Te contents of the primary MORB melt corrected for magmatic differentiation can be successfully reproduced by near-fractional decompression melting of a mantle with 170–200 μg g−1 S (as sulfide liquid), which has either (1) “fertile lherzolite-like” Se and Te contents (80 ± 17 and 11 ± 1.7 ng g−1, respectively; 1 s.d.) or (2) distinctly lower Se (49 ± 11 ng g−1) and Te (3.5 ± 1.3 ng g−1) contents depending on the choice of experimental partition coefficients published by different studies. Regardless, our model shows that Se-Te systematics of “fertile” lithospheric peridotites preserve little primary melt depletion signatures and reflect significant, if not complete, metasomatic overprinting. Finally, based on the observed negligible Se isotopic fractionation between sulfide phase and silicate melt, we suggest that MORBs preserve their mantle source isotopic signature (δ82/76Se = −0.16 ± 0.13‰). Our MORB average is similar within uncertainty to chondritic values but significantly lighter than previously published δ82/76Se data for basalts from a variety of geodynamic settings. The subtle but significant Se isotope variation observed within the investigated MORB suite (up to ∼0.25‰) and between other mantle samples analyzed so far may reflect intrinsic source heterogeneity and potential isotopic differences across various mantle reservoirs.