What controls the preservation of hydration interfaces in high grade metamorphic rocks?

The hydration of initially dry, lower crustal metamorphic rocks during orogenesis is a commonly observed phenomenon and hydration/reaction interfaces are also often preserved, providing a unique insight into the evolution of the lithosphere. Often the interfaces between unreacted and reacted rock are very sharp, even on a thin section scale, and various explanations have been proposed to account for the abrupt changes in mineral assemblage on such a small spatial scale. Common to a wide range of specific examples is the role of an infiltrating aqueous fluid that is generally assumed to be required for the reaction to take place, although other features of such reaction fronts can differ widely in terms of density changes and the apparent difference in metamorphic grade across a sharp interface. 

The examples discussed here all involve the hydration of basement granulite rocks formed during the Caledonian Orogeny and now exposed in the Bergen Arcs in Norway. All stages of hydration can be observed from totally unreacted dry granulites with a wide range of composition to either eclogite facies or amphibolite facies overprints. In these cases the density changes across the interface can either be positive (in the case of eclogite formation from anorthosite granulites), can be negative (in the amphibolitisation of basic rocks) or virtually zero (during the amphibolitisation of garnet bearing anorthosites). The preservation of volume across such interfaces has led to investigations of the coupling between the consequent stress generation and mass transfer, which in turn focusses on the evolution of porosity/permeability in the parent dry rock. The extent of hydration in the Bergen Arcs as a whole ( 90% hydration of ~10⁵ km³ of granulite) suggests a plentiful supply of aqueous solution introduced seismically by fracturing and the consequent generation of shear zones from which hydration fronts spread. The hydration to either eclogite or amphibolite, often observed at the same structural level (i.e. depth in the crust) continues to be an enigma.

Although the details of the reactions and density changes are different, a common feature is the need for an infiltrating aqueous solution and hence the question of what drives the fluid and the hydration reaction and finally why the reaction stops at the sharp interfaces observed in the field. Terminated reactions can be studied by the extent of alteration around fracture planes by modelling the likely fluid pressure gradients that drive Darcy flow from the fluid source towards the reaction interface. Fracture planes represent zones of localised high permeability that facilitate the infiltration of fluid. The difference in fluid saturation between the fracture plane and the alteration halo is thought to be responsible for both the degree of reaction and the difference in assemblage. Additionally, the width of the initial fracture plane is thought to be proportional to the extent of the alteration halo. Examples will be given of hydration fronts associated with amphibolite facies shear zones as well as the observation of eclogite fingers within a partly hydrated granulite host. Combined reaction-fluid flow models attempt to explain these phenomena.