Unraveling the confusion over the iron oxidation state in MORB glasses. (15th January 2021)

• Record Type: Journal Article
• Authors: Bézos, Antoine; Guivel, Christèle; La, Carole; Fougeroux, Thibault; Humler, Eric
• Is Part Of: Geochimica et cosmochimica acta. Volume 293(2021)
• Page Start: 28
• Page End: 39
• Publication Date: 2021-01-15
• Subjects: MORB -- Iron -- Oxidation state -- Wet chemistry -- XANES spectroscopy; Geochemistry -- Periodicals; Meteorites -- Periodicals; Géochimie -- Périodiques; Météorites -- Périodiques; Geochemie; Astrochemie; Electronic journals; 551.905
• DOI: 10.1016/j.gca.2020.10.004
• Languages: English
• ISSNs: 0016-7037

Abstract

Abstract: The oxidation state of iron in mantle-derived melts is commonly used to determine the oxygen fugacity ( f O2 ) of magmas and their related sources. The accurate and precise determination of the iron oxidation state ratio of Mid-Oceanic Ridge Basalt (MORB) glasses has been a matter of controversy for the last three decades. None of the wet chemical methods used in the literature to measure this ratio converge toward a consensus value. The same difficulties have been observed for the most recent data obtained by XANES spectroscopy, which allows the measurement of the iron oxidation state ratio at high spatial resolution. This study unravels the origin of the analytical biases observed between the colorimetric and direct titration wet chemistries used in the studies of Christie et al. (1986) and Bézos and Humler (2005), respectively. We demonstrate that colorimetric measurements overestimate the ferrous iron contents of sulfide-bearing samples whereas the direct titration method does not suffer from such sulfide interference. Because MORB glasses contain significant amount of sulfur in the +2 valence state, we re-assessed the average MORB FeO/FeOtotal ratio by using the bulk FeOtotal and FeO data from this study and from Bézos and Humler (2005), respectively. We calculated for 49 MORB glasses an average FeO/FeOtotal of 0.90 ± 0.02. This result agrees with colorimetric measurements corrected for sulfide interference and with the results of Bézos and Humler (2005), if corrected for the presence of plagioclase in the sample powders. The average MORB FeO/FeOtotal that we determined in this study agrees also within uncertainty with the most recent XANES spectroscopic data of Berry et al. (2018) (0.90 ± 0.02).