Smectite dehydration: a key to fault mechanics? – SMEC

The recent geophysical data acquired to describe subduction zones revealed the very variable mechanical behaviour of the plate interface fault zone, from aseismic slip to megaearthquakes. So far, there is no model to describe the processes controlling these deformation modes.
The fault cores contain in many cases a large proportion of smectite. This clay mineral incorporates a variable proportion of water in interlayer space. Dehydration reactions are a potential trigger of mechanical instabilities, and the aim of SMEC is to unravel possible connections between dehydration reactions and slip instabilities.
Understanding smectite dehydration in fault rocks requires first a new conceptual framework to be applied. All available experimental data, as well as thermodynamic models, consider a system where the same pressure is applied to the solid and the free fluid. On the contrary, fault rocks are best described as a two-phase system composed of a solid skeleton that includes fluid-filled cavities, with two independent pressures.
Based on this fundamental assumption, we propose to reassess smectite dehydration by combining experiments and modelling (WP1). Dehydration reactions in the space (Pfluid-Psolid-T conditions) will be analysed using X-ray diffraction experiments in a semi-transparent high-pressure vessel at ISTO, along with Synchrotron experiments. These experimental results will then constitute inputs to a model of dehydration, which includes molecular simulations at the clay interlayer scale, and upscaling to a macroscopic thermodynamic model.
This analysis of dehydration will form the ground of an experimental study of smectite deformation (WP2), to unravel potential connections between slip instabilities and dehydration. Frictional properties of smectite revealed by the experiments will be modelled using a combination of molecular simulations at the interlayer scale and thermo-poro-mechanics to retrieve macroscopic properties and to establish instability criteria.
Last, in WP3, the same approach of WP1 and WP2 will be applied to natural fault rocks, to validate and extend the hydration and frictional behaviour characterized on synthetic, pure smectite samples to natural material with a larger complexity, in particular in terms of mineralogy. The hydration and frictional properties obtained from the analysis of synthetic and natural samples will finally be used as inputs of numerical models of the slip behaviour of large-scale fault zones, able to reproduce the seismic cycle and to catch a large scope of slip behaviour, from slow slip events to regular earthquakes.