Comments on: Liquid Immiscibility and the Evolution of Basaltic Magma

Abstract

Veksler et al. (2007) present the results of interesting experiments that purport to demonstrate that fractionating basaltic magma can split into immiscible liquid fractions at temperatures above 11008C, which is considerably higher than has been found in all previous experimental studies for melts of similar composition. Previous studies showed that immiscibility occurred just above 10008C and could therefore play a role during only the late stages of fractionation of basaltic magma. Veksler et al., however, claim that the higher temperature immiscibility demonstrated in their experiments shows that immiscibility could play a significant role at much earlier stages of differentiation. If correct, this is an important finding. Having worked on silicate liquid immiscibility since the 1960s, I would be only too pleased if their conclusion were correct. Unfortunately, I believe their results do not indicate stable high-temperature immiscibility but rather metastable phase separation developed during quenching of their experimental charges that were run in a centrifuge. All binary systems involving SiO2 with FeO, MnO, MgO, CaO, and TiO2 exhibit liquid immiscibility at temperatures near 17008C (e.g. the binary SiO2^Fayalite in Fig. 1). With addition of small amounts of alumina and alkalis to these binary systems, the two-liquid field disappears in ternary or more complex systems (Fig. 1a). However, Roedder (1951) showed that the two-liquid field reappears at lower temperatures in the system K2O^FeO^ Al2O3^SiO2 (Fig. 1a). It is important to recognize that, although the low-temperature two-liquid field is separate from the high-temperature field, they are both intersections of the same two-liquid solvus with the liquidus surface in this system (Visser & Koster van Groos, 1979). To illustrate this phase relation, Fig. 1c shows a schematic perspective view of the system, where the two-liquid solvus can be visualized as analogous to an anticline plunging from high temperatures on the SiO2^Fayalite binary to low temperatures in the ternary system. This solvus plunges below the liquidus on leaving the SiO2^Fayalite binary and hence becomes metastable. However, at lower temperatures along the Fayalite^Tridymite cotectic, the solvus reappears above the liquidus and again becomes stable. The two-liquid solvus then plunges beneath the liquidus again and becomes metastable as it approaches the Fayalite^Leucite binary. It is important to recognize that even where the two-liquid field is not stable, this field is lurking just beneath the stable liquidus. In their experiments,Visser & Koster van Groos (1979) were able to trace the metastable two-liquid field to compositions beyond the Fayalite^Orthoclase join (Fig. 1c). If a single-phase liquid is quenched rapidly enough to prevent crystallization, it can easily intersect the metastable two-liquid field, or its spinode, and develop fine-scale phase separation. Indeed, this metastable phase separation is taken advantage of commercially by Corning to create some of the desirable physical properties of their PyroCeram , which is used, for example, in the nose-cones of rockets and Corningware for cooking. Given the similarity of the supposed immiscibility textures described by Veksler et al. (2007) to the metastable immiscibility textures found by Visser & Koster van Groos (1979) in the system K2O^ FeO^Al2O3^SiO2, it is surprising that they make no reference to this paper. We can draw three important conclusions from the nature of immiscibility in the system K2O^FeO^Al2O3^ SiO2 that are relevant to the discussion of theVeksler et al. paper. First, the top of the two-liquid solvus and the