Experimental constraints on the low-pressure evolution of transitional and mildly alkalic basalts: the effect of Fe-Ti oxide minerals and the origin of basaltic andesites

Abstract

One-atmosphere, melting experiments, controlled at the fayalite-magnetite-quartz oxygen buffer, on mildly alkalic and transitional basalts from Iceland show that these begin to crystallize Fe-Ti oxide minerals (magnetite and/or ilmenite) at 1105±5°C, apparently independently of bulk composition and the order of silicate and oxide mineral crystallization. Most samples crystalline plagioclase and olivine as the first two crystalline phases, augite as the third phase, and an Fe-Ti oxide mineral as the fourth phase. The main effects of Fe-Ti oxide crystallization are a marked decrease in FeO and TiO₂ in the liquid, and a notable increase in SiO₂ and Al₂O₃, and the minor oxides K₂O and P₂O₅, with decreasing temperature. The most silicic glasses are compositionally mugearitic and shoshonitic basaltic andesites. Because the smallest amount of glass that could be analyzed with the microprobe represents 20–55 percent liquid remaining, it can be expected that more silicic liquids will occur at lower temperatures. On normative, pseudoternary projections, the general effect of Fe-Ti oxide crystallization for mildly alkalic and transitional basalts is a marked increase in normative quartz. This is caused by a strong systematic convergence, with the appearance of Fe-Ti oxides, of the bulk solid precipitates toward the liquid compositions, as projected on the triangle plagioclase-diopside-olivine. For alkalic basalts, the bulk solid precipitate shows an increase in normative diopside with falling temperature and Fe-Ti oxide crystallization. This causes the liquids to move toward decreasing normative diopside and relatively little variation in nepheline. The experimental observations imply that mildly alkalic and transitional magmas, without stabilizing a Fe-Ti oxide mineral, will not evolve toward early silica saturation.