D/H and Water Concentrations of Submarine MORB Glass Around the World: Analytical Aspects, Standardization, and (re)defining Mantle D/H Ranges

Abstract:

The advent and calibration of the Thermal Combustion Element Analyzer (TCEA) continuous flow system coupled with the large-radius mass spectrometer MAT253 permits precise (±0.02 wt.% H₂O, ±1-3‰ D/H) measurements in 1-10 mg of volcanic glass (0.1 wt.% H₂O requires ~10 mg glass), which permits the targeting of small amounts of the freshest concentrate. This is a >100 factor reduction in sample size over conventional methods, four times over more common Delta series instruments. We investigated in triplicate 115 samples of submarine MORB glasses ranging from water-poor (0.1-0.2wt%) to water-rich (1.2-1.5wt%). These samples were previously investigated for major and trace elements, radiogenic isotopes; a large subset of these samples coming from the FAZAR expedition were studied previously by FTIR for water concentration. We also ran samples previously studied by the conventional off-line technique: MORB glass including those from the Easter Platform and the Alvin 526-1 standard (0.2wt% H2O). We observe excellent 1:1 correspondence (1.02x+0.02, R2=0.94) of wt% water by FTIR and TCEA suggesting complete extraction of water and no dependence on water concentration. We measure 51‰ total range in D/H that correlates with all other chemical and isotopic indicators of mantle enrichment, with the heaviest values occurring in the most enriched samples. When used uncorrected values of H2 gas run against H2 gas of known composition, this range agrees nicely with previous D/H range for MORB (-30 to -90‰), measured for samples run conventionally. Uncorrected analyses of Alvin glass 526-1 gives -66‰. When run against SMOW, SLAP and -41‰ water sealed in silver cups, the range is shifted by -15‰; when standardization is done by with three commonly used mica standards as is done most commonly in different labs, the range is shifted downward by -30-32‰. There are no isotopic offsets related to total water or D/H range requiring different slope or non-linear correction. The NBS30 mica standard has been recently shown to be heavier and more heterogeneous than previously thought, and older conventional methods that relied on Pt reduction unreliable. Based on these new TCEA results, the D/H values of MORB and mantle samples may need to be revised to lighter values by 15‰.