Untitled Document

The redox state of terrestrial basalts determined by V/Sc olivine-melt partitioning data

Guil Mallmann1, Hugh O'Neill1, Frances Jenner1, Marc Norman1, Steve Eggins1, Richard Arculus1 and Chris Ballhaus2

1 Research School of Earth Sciences, Australian National University, Canberra, ACT 0200, Australia
2 Mineralogisch-Petrologisches Institut, Universität Bonn, Poppelsdorfer Schloss, Bonn 53115, Germany

Figure 1. Partition coefficients obtained empirically for V and Sc between olivine phenocryst and silicate melt (glass or matrix). The positive correlation between is indicative of effects (possibly melt composition) other than oxygen fugacity. The dashed lines, illustrating values of oxygen fugacity relative to the QFM buffer, were calculated based on the experimental partitioning data.



The dependence of the partitioning of V between olivine and silicate melt () on oxygen fugacity was used to estimate directly the redox state of primitive terrestrial basaltic and picritic magmas at that stage in their evolution when they begin to crystallize olivine. The effect of other variables was accounted for by rationing  to , because the partitioning of Sc, a redox insensitive element having approximately the same incompatibility at terrestrial oxygen fugacities, is shown to depend rather similarly on melt composition.

The method was calibrated on basaltic compositions equilibrated in the laboratory (one atmosphere) at QFM and QFM-2.7 between 1300 and 1400°C. We demonstrated that this method can be effective over the entire range of redox conditions observed in geological and cosmochemical materials, and therefore may serve as a universal redox indicator in olivine-phyric mafic volcanic rocks. Our preliminary assessment indicates accuracy in relative oxygen fugacity between 0.2 to 0.5 log units, but precision typically better than ±0.2 log units.

The method was applied to 41 mid-ocean ridge (MORB), 25 ocean island (OIB), and 13 island arc (IAB) recent primitive basalts and picrites. The data indicate that MORBs and OIBs record a very restrict range of redox conditions, between QFM and QFM+1, with no clear distinction between them. However, IABs record consistently more oxidizing conditions, ranging from QFM+0.5 to QFM+3 (average at QFM+1.7). Except for MORBs, for which the data cluster exactly on the maximum redox condition ever reported, the results presented here are in good agreement with previous estimations using various methods in minerals and melts.