Document Type


Date of Original Version





To improve the accuracy of X-ray absorption near-edge structure (XANES) calibrations for the Fe3 +/ΣFe ratio in basaltic glasses, we reevaluated the Fe3 +/ΣFe ratios of glasses used as standards by Cottrell et al. (2009), and available to the community (NMNH catalog #117393). Here we take into account the effect of recoilless fraction on the apparent Fe3 +/ΣFe ratio measured from room temperature Mössbauer spectra in that original study. Recoilless fractions were determined from Mössbauer spectra collected from 40 to 320 K for one basaltic glass, AII_25, and from spectra acquired at 10 K for the 13 basaltic glass standards from the study of Cottrell et al. (2009). The recoilless fractions, f, of Fe2 + and Fe3 + in glass AII_25 were calculated from variable-temperature Mössbauer spectra by a relative method (RM), based on the temperature dependence of the absorption area ratios of Fe3 + and Fe2 + paramagnetic doublets. The resulting correction factor applicable to room temperature determinations (C293, the ratio of recoilless fractions for Fe3 + and Fe2 +) is 1.125 ± 0.068 (2σ). Comparison of the spectra at 10 K for the 13 basaltic glasses with those from 293 K suggests C293 equal to 1.105 ± 0.015 (2σ). Although the 10 K estimate is more precise, the relative method determination is believed to be more accurate, as it does not depend on the assumption that recoilless fractions are equal at 10 K. Applying the effects of recoilless fraction to the relationship between Mössbauer-determined Fe3 +/ΣFe ratios and revised average XANES pre-edge centroids for the 13 standard glasses allows regression of a new calibration of the relationship between the Fe XANES pre-edge centroid energy and the Fe3 +/ΣFe ratio of silicate glass. We also update the calibration of Zhang et al. (2016) for andesites and present a more general calibration for mafic glasses including both basaltic and andesitic compositions. Recalculation of Fe3 +/ΣFe ratios for the mid-ocean ridge basalt (MORB) glasses analyzed previously by XANES by Cottrell and Kelley (2011) results in an average Fe3 +/ΣFe ratio for MORB of 0.143 ± 0.008 (1σ), taking into account only analytical precision, and 0.14 ± 0.01(1σ), taking into account uncertainty on the value of C293. This revised average is lower than the average of 0.16 ± 0.01 given by Cottrell and Kelley (2011). The revised average oxygen fugacity for MORB based on the database of Cottrell and Kelley (2011) is − 0.18 ± 0.16 log units less than the quartz-fayalite-magnetite buffer of Frost (1991) at 100 kPa (∆ QFM = − 0.18 ± 0.16).

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.


A corrigendum was later published in Chemical Geology (Vol. 485, May 2018, pp. 113) listing changes to the publication. It can be found listed in the Additional Files of this page.

The changes can also be found at this link: