A predictive model for divalent element partitioning between clinopyroxene and basaltic melt and a europium-in-plagioclase-clinopyroxene oxybarometer for cumulate rocks

Abstract

Under geologically relevant conditions, Eu is a multivalent element, exhibiting divalent character in reduced systems and trivalent character in oxidized systems. Its mineral-melt and mineral-mineral partitioning behavior is sensitive to oxygen fugacity (fO₂) and can be leveraged in oxybarometers that recover fO₂s from natural samples if the partitioning behaviors of the divalent and trivalent species are known. Here, we parameterize a lattice strain-based predictive model for partitioning of divalent elements between basaltic silicate melts and the clinopyroxene M2 site. The model accurately and precisely recovers experimentally determined partition coefficients (Sr, Ca, Mn and Zn) as a function of pyroxene composition and temperature. The new divalent element partitioning model is coupled with published trivalent element partitioning models to develop an fO₂-, temperature- and composition-dependent clinopyroxene-melt Eu partitioning model. The clinopyroxene-melt Eu partitioning model is coupled with a plagioclase-melt Eu partitioning model to develop a Eu-in-plagioclase-clinopyroxene oxybarometer for cumulate rocks. Applied to lower crustal gabbros from oceanic lithosphere exposed at the southern Samail ophiolite and the east Pacific rise (Hess Deep), the oxybarometer recovers fO₂s indistinguishable from the fayalite-magnetite-quartz buffer within error (∼±0.5 log units), in agreement with basaltic glasses from mid-ocean ridges. Kinetic analysis is applied to evaluate the time-dependent response of the oxybarometer to a change in temperature. Eu³⁺ diffusion is much slower than Eu²⁺ diffusion; nonetheless, under moderately to highly reducing conditions, fO₂s recovered using the oxybarometer approach equilibrium conditions over timescales of 1000s of years, even while Eu³⁺ exchange is still active.