Micro-XANES Determination of Fe Oxidation State in MORB and its Relationship to Water Content

We report the oxidation state of iron and H2O content for a global suite of mid-ocean ridge basalt (MORB) glasses in order to assess the relationship between water and oxidation state in mantle-derived basalts unaffected by subduction zone processes. Fe3+/ΣFe ratios were determined with Fe K-edge μX-ray absorption near-edge structure spectroscopy (μXANES). Fe oxidation state was previously determined with microcolorimetry and/or direct titration on a subset of the samples. Fe3+/ΣFe ratios determined using the Mossbauer-based μXANES basalt calibration averaged 0.16±0.01 (n=46) and negatively correlate with MgO suggesting the influence of olivine fractionation. The subset of samples also analyzed by colorimetry ranged from 0.04-0.15 compared to 0.13-0.17 with XANES (n=27) and those also analyzed by direct titration ranged from 0.09-0.14 compared to 0.14-0.17 with XANES (n=9). To investigate the cause of these offsets, Fe3+/ΣFe ratios were determined directly by Mossbauer spectroscopy on two hand-picked, aphyric MORB glass separates and were found to be identical within error to values determined by μXANES. Experimentally equilibrated basaltic glasses from the original study of Kress and Carmichael (CMP, 1991), for which Fe3+/ΣFe ratios were determined by microcolorimetry, were also found to yield Fe3+/ΣFe ratios identical to those determined by μXANES within error. Acid digestion of olivine microphenocrysts in natural MORB samples may lead to an offset between XANES, which samples only glass, and wet-chemical determinations of Fe3+/ΣFe ratios. For example, acid digestion of 20% Fo85 olivine is sufficient to cause the observed offset between XANES and direct titration. Total H2O content, determined with FTIR on the same suite of glasses, shows no correlation with iron oxidation state globally (n=46) or locally within a cogenetic suite (n=10, CHEPR, EPR 6°-13° N, 6-9wt% MgO). In contrast, subduction-influenced back-arc basin (BABB) settings show clear correlations between H2O and Fe oxidation state (Kelley and Cottrell, Science, 2009) both globally and locally. This indicates that, while fluids or melts derived from subducted altered oceanic crust may oxidize the mantle locally beneath subduction zones, water itself does not. Moreover, shallow or magma-chamber scale processes common to both MORB and BABB settings are not likely to be responsible for the global correlation observed between water and Fe oxidation state in basaltic melts.