Studies of crustal evolution require a knowledge not only of potential source regions of granitic magmas, but also of their transport paths and emplacement mechanisms. One approach to understanding segregation and transport mechanisms of silicate melts following partial melting has been through experimental investigations. In recent years, the equilibrium melt distribution in partially molten crustal and mantle rocks has been studied experimentally in various laboratories. A critical microstructural parameter in partially molten texturally equilibrated aggregates is the wetting (or dihedral) angle - that is the angle between two adjacent solid-liquid interfaces which bound a melt inclusion. When wetting angles are low (< 60°) the melt pockets in a partially-molten aggregate are interconnected and the permeability is sufficiently high for the melt to segregate. On the other hand, when wetting angles are high (> 60°) the melt pockets are isolated and there will be no melt connectivity except at high melt fractions. Accordingly, considerable effort has been devoted toward determining melt connectivity and matrix permeability in texturally equilibrated systems. This has led many investigators to conclude that efficient melt segregation requires high degrees of partial melting. However, disequilibrium phenomena, particularly those associated with deformation, may result in more important mechanisms for enhancing permeability.
We are currently investigating the influence of microfracturing on permeability during partial melting. The volume increase of many melting reactions results in a melt overpressure which induces brittle failure and crack propagation during the reaction. Propagating fractures enhance permeability by providing pathways for rapid melt transport and this effect can be large provided the fractures interconnect.
Preliminary experiments have been performed on a muscovite-bearing quartzite, under both fluid-absent and H2O-saturated conditions, at 1 and 3 kbar in the temperature range 760 - 850 °C. The aim of the experiments is to study the propagation of fractures and the enhancement of permeability as a function of parameters such as temperature, time and the rate of the melting reaction. The melting reaction involved, muscovite + quartz ± H2O -> melt + mullite + biotite, results in muscovite grains becoming pseudomorphed by the product phases with a local volume increase of ~ 14 %. As a result of the volume increase, melt-filled fractures propagate away from the melting sites as the reaction progresses. The crack length (and therefore the length scale of melt migration) as well as the width of the melt-filled fractures increase with the extent of the melting reaction. Short term experiments (< 12 hours) contain melt-filled fractures which are typically 1 micron wide and can extend up to 150 microns from the melting sites. In longer duration experiments (> 12 hours and at least 50 % reaction) the fractures have propagated further, have a length of several 100 microns, a width of 1-5 microns and form a network which connects many of the melting sites. In a preliminary fluid-absent experiment, crack lengths are shorter (e.g. < 100 microns long after 286 hours as compared to reaching > 200 microns in the water-saturated experiment performed for 214 hours) and are more closely spaced, possibly as a consequence partly of the higher melt viscosity. Estimated permeabilities have been calculated from the density and dimensions of the melt-filled fractures and range between 10-14 to 10-13 m2. Such high permeability values suggest that at relatively fast melting rates, such as during magmatic underplating of the continental crust, transient periods of increased permeability may develop in previously impermeable rock. This process may be the most important first step in the development of a large-scale fracture network necessary for efficient melt segregation.