Liquid unmixing kinetics and the extent of immiscibility in the system K 2O–CaO–FeO–Al 2O 3–SiO 2

Abstract

This experimental study examines the kinetics of silicate liquid immiscibility, and the effects of melt chemistry on phase separation. We used three synthetic compositions in the systems K 2O–FeO–Al 2O 3–SiO 2 and K 2O–CaO–FeO–Al 2O 3–SiO 2. Two of them lie inside the known miscibility gaps on the joins fayalite–orthoclase–silica and fayalite–hedenbergite–orthoclase–silica; the third one lies on the anorthite–fayalite cotectic, well outside any known region of stable or metastable unmixing. In order to distinguish between sub-micron exsolutions due to quenching and those that were formed by immiscibility above liquidus, we performed high temperature in situ centrifugation experiments in addition to conventional static runs. Both types of experiments were performed at the atmospheric pressure and super-liquidus temperatures. Run durations varied from a few hours to a few days, and melts were quenched to glasses in air at the end of each run. Exsolution textures were observed in all the glasses. Droplet sizes varied widely in the micron to sub-micron range, with the smallest exsolutions being visible only with the aid of a transmission electron microscope (TEM). Those immiscible droplets that settled over macroscopic distances during high temperature centrifugation were interpreted, regardless of their size, as products of stable unmixing. Of particular interest is the settling of very fine, 20–30 nm emulsion droplets in the composition on the anorthite–fayalite cotectic, in which stable unmixing was previously unknown. Our results show that unmixing of melts with high contents of Ca aluminosilicate components may develop slowly, even at high temperatures above liquidus, and remain in a latent, sub-micron form for a long time. Textures of quenched charges imply that interfacial energies between immiscible liquids are low, and this fact is likely to significantly impede the coarsening of the emulsions. Slow growth of immiscible droplets and the uncertainty about interpretation of very fine, sub-micron emulsions raise the possibility that regions of silicate immiscibility, including those in natural basaltic melts, may be broader than presently thought, with the implication that the petrogenetic role of immiscibility should be revisited.