Melt specific parameters

Different parameters influence the behavior of melt during experiments and simulations: surface energies, melt fraction and permeability of the rock. In our simulations the permeability is set to infinity so that different melt pockets can freely interact and diffusion is not rate-limiting. We effectively assume that there is a three-dimensional connectivity between melt pockets on the thin-section scale that we model.

In the Elle_melt module, the melt fraction can be easily adjusted. The starting grain fabric for a simulation includes information about the grains and melt pockets, the ratio of the area of melt and area of grains is the melt fraction. Using parameters at program start-up, this melt fraction can be fixed to the starting value or it can be decreased or increased during the simulations. During a simulation, the current melt fraction is constantly compared to the desired melt fraction. If it is the same or the melt fraction is allowed to increase without bounds, the melt fraction does not influence the calculations at all. If the melt fraction is outside the desired range, node-movements that adjust the current melt fraction towards the desired one are preferred. The melt experiments shown below all use the starting melt fraction as a fixed value.

In the Elle_melt module, the surface energies of solid-solid boundaries (Δtss), solid-liquid boundaries (Δtsl) and liquid-liquid boundaries (Δtll) can be adjusted. Liquid-liquid boundaries (Δtll) have no physical meaning, since melt composition is always the same. They may however occur in the models for topological reasons, whereby their surface energy is set to zero. Solid-solid surface energies are set to unity and solid-liquid surface energies range between 0.50191 (simulation 1a, 1b, 4, 5, 6, 8), 0.57735 (simulation 2) and 1 (simulation 3). The different relative solid-solid to solid-liquid surface energies lead to different wetting angles and hence test the predictions of Bulau et al. (1979) or Laporte et al. (1997) concerning the shape of grains and melt pockets and their behavior at different wetting angles (simulations 1-3).