Title: Can spinodal decomposition occur during decompression-induced vesiculation of magma?

Abstract: Volcanic eruptions are driven by decompression-induced vesiculation of volatiles in magma. Its initial phase has long been described as nucleation. Recently, it was proposed that spinodal decomposition (SD; an energetically spontaneous phase separation without forming a distinct interface) may occur during magma vesiculation. This suggestion is currently based only on qualitative textural observations of the products of decompression experiments. In this study, I used a simple thermodynamic approach to quantitatively investigate whether SD can occur during magma vesiculation. Using the previous water solubility data, I plotted the binodal and spinodal curves on the chemical composition-pressure plane for several hydrous magmas, treating them as two-component symmetric regular solutions of silicate and water. The spinodal curves were much lower than the binodal curves at pressures sufficiently below the second critical endpoints. In addition, the final pressure of all the decompression experiments performed to date fell between these two curves. This suggests that SD is unlikely to occur in the pressure range of magmatic processes in the continental crust or at realistic decompression rates i.e. magma vesiculation results from nucleation, as previously suggested. To test this thermodynamic approach, I estimated the microscopic surface tension between the melt and bubble nucleus in previous decompression experiments by substituting the spinodal pressure into the nonclassical nucleation theory (non-CNT) equation for the dependence of the surface tension on the degree of supersaturation. The resulting values were significantly more scattered than those obtained by the conventional method: inversion of the experimental bubble number density using the classical nucleation theory (CNT) formula. Thus, it is difficult to judge whether the non-CNT equation is appropriate for magma systems.