AN EXPERIMENTAL-STUDY OF THE INFLUENCE OF OXYGEN FUGACITY ON FE-TI OXIDE STABILITY, PHASE-RELATIONS, AND MINERAL-MELT EQUILIBRIA IN FERRO-BASALTIC SYSTEMS

Equilibrium crystallization experiments at atmospheric pressure and over a range of oxygen fugacity (f(O2)) have been carried out on a ferro-basaltic composition similar to liquids proposed to have been parental to much of the exposed portion of the Skaergaard intrusion. Before Fe-Ti oxide saturation the liquid line of descent is little affected by f(O2). However, the appearance temperatures of the magnetite-ulvospinel solid solution (Mt) and the ilmenite-haematite solid solution (Ilm) depend strongly on f(O2). Above the fayalite-magnetite-quartz (FMQ) buffer Mt is the first oxide phase to appear on the liquidus, but below the FMQ buffer Ilm is the first oxide to crystallize. The appearance temperature of Mt is similar to 1100 degrees C at FMQ and the Mt liquidus slope is similar to 30 degrees C/log f(O2) unit between FMQ - 2 and FMQ + 1. The Ilm liquidus is at similar to 1100 degrees C between FMQ and FMQ - 2, but moves to lower temperature at higher f(O2) where Mt is the first oxide phase. The results indicate that the ferric iron content of Mt-saturated melts varies linearly with inverse temperature, and that nm saturation is closely related to melt TiO2 content. Mt saturation Produces an immediate enrichment of SiO2 and depletion in FeO* in the melt phase, whereas Ilm saturation Produces similar enrichment in SiO2, but iron enrichment may continue for similar to 10 degrees C below the ilmenite liquidus. The experimental liquids reach a maximum of similar to 18 wt % FeO*, at similar to 48 wt % SiO2 for ilmenite-saturated melts at lore f(O2), more differentiated melts having lower iron and higher silica. Cotectic proportions, derived from mass balance calculations, are in good agreement with data from natural samples and other experimental studies. Olivine resorption is inferred at all f(O2), with the onset of resorption occurring similar to 10 degrees C higher than the appearance of magnetite. The effect of f(O2) on silicate mineral compositions, and partitioning of elements between coexisting mineral-melt pairs, is small. Thermodynamic considerations suggest that variations of Fe-Mg partitioning between the iron-rich olivines, pyroxenes and melts produced in this study may be explained by known non-idealities of Fe-Mg mixing in the crystalline phases, rather than non-idealities in the coexisting melts. These experiments also provide insights into many features common to natural tholeiitic series of volcanic and plutonic rocks, and provide experimental data required for modelling of fractional crystallization and crystallization closed to oxygen, processes which are not easily investigated experimentally.