EFFECT OF GLASS COMPOSITION ON FE-XANES: IMPLICATIONS FOR THE REDOX STATE OF MORBS

Since iron is the most abundant naturally-occurring element that has multiple oxidation states, the relative concentrations of Fe³⁺ and Fe²⁺ in minerals and melts provides insight into the oxygen fugacities (fO₂s) under which they were formed. Given the extent that the oxidation state controls some geological processes (e.g., planetary core formation and magmatic degassing), numerous efforts have been made to measure the Fe³⁺ and Fe²⁺ concentrations of natural samples in order to estimate the fO₂ inside the Earth and the other terrestrial planets.

The small sizes and heterogeneities in many natural samples make it difficult to measure their Fe³⁺/∑Fe ratios. Cottrell et al., (2009) developed a XANES-based micro-analytical approach applicable to basaltic glasses. They showed that for two basaltic compositions (6.1 and 7.4 wt. % MgO), the centroid energy of the iron 1s→3d doublet (measured by XANES) displays a positive correlation with the Fe³⁺/∑Fe, as measured by Mössbauer spectroscopy. However, in addition to being affected by the Fe³⁺/∑Fe, the XANES spectra should also depend on the coordination environment of the iron atoms in silicate glass (cf., Wilke et al. 2005). Here, we report the results of experiments aimed to test the magnitude of the effect of glass composition on XANES-based Fe redox determination. We performed one-atmosphere experiments on three bulk compositions, a primitive, an intermediate and an evolved average MORB (16.1, 8.6, and 4.7 wt. % MgO, respectively) at log fO₂s from −10 to −0.68. Fe³⁺/∑Fe in the quenched liquids were measured by Mössbauer spectroscopy, and Fe-XANES spectra were collected at Diamond light source using a Si(333) monochromator and acquisition parameters similar to those of Cottrell et al. (2009). Our results suggest that glass composition does have an effect: for example, at an fO₂ one log unit above the QFM buffer, our evolved basalt has a centroid energy that is 0.05 eV lower than that of our primitive basalt despite having a higher Fe³⁺/∑Fe. Applying a calibration based on a primitive or intermediate composition to an evolved natural sample could result in the underestimation of the fO₂ of the evolved basalt by up to a log unit.

References:

Cottrell et al. (2009) Chemical Geology 268, 167-179.

Wilke et al. (2005) Chemical Geology 220, 143-161.