MSA ROEBLING MEDAL LECTURE: THE EFFECTS OF SOLID-SOLID PHASE EQUILIBRIA AND PARTIAL MELTING ON THE OXYGEN FUGACITY OF THE UPPER MANTLE

The redox variability of basalts and their mantle sources is of growing interest due to the increasing number of measurements in glasses and minerals of ƒO₂–sensitive ratios such as Fe³⁺/Fe²⁺. A simple interpretation of such variability is that if ƒO₂ and Fe³⁺/Fe²⁺ are higher in one basalt than in another, then its source also had a higher ƒO₂ and Fe³⁺/Fe²⁺. We show, however, that phase equilibria of peridotite can lead to significant variations in the ƒO₂ of the system and Fe³⁺/Fe²⁺ of individual phases (including melt) even when the peridotite’s bulk composition (i.e., total O content and Fe³⁺/Fe²⁺) is constant.

The effects of phase changes on the ƒO₂ of lherzolite can be largely understood based on a simple example: Assume subsolidus plagioclase (plag) lherzolite with a small modal abundance of spinel (sp); its ƒO₂ will vary positively with the magnetite content (mt) of the sp. As P increases, the peridotite will transform into a sp lherzolite, and the mt content of the sp will decrease due to dilution by the aluminous sp components produced by this transformation; as a result, the ƒO₂ of the system drops despite its unchanged bulk Fe³⁺/Fe²⁺. This ƒO₂ decrease as peridotite transforms from plag to sp lherzolite is ~1.5 log units based on thermodynamic modeling. The ƒO₂ then increases by ~1 log unit as P increases further and the peridotite transforms to garnet (gt) lherzolite, due to the reaction of aluminous sp components into pyroxenes and gt, largely reversing the effects of the reaction to sp lherzolite. Quantitative modeling of peridotite ƒO₂ at 0–4 GPa confirms these and other effects of continuous and discontinuous phase changes.

Although isobaric melting generally leads to decreasing ƒO₂, isentropic decompression melting can result in the ∆FMQ of peridotite increasing by ~1 log unit. This also results from continuous and discontinuous solid–solid phase transitions, with melting itself only introducing a small perturbation on melt-absent trends. Shallower melts on a single isentrope are thus expected to have generally higher ƒO₂ than deeper melts.

We conclude that although observed variations in ƒO₂ of basalts may reflect subduction of oxidizing fluids, sediments, and altered oceanic crust and/or the opposing effects of subducted organic material, the effects of peridotite phase equilibria must also be considered.