Frictional anisotropy in casted seismic faults: or how to 3D print a fault to better characterize it

Anisotropic phenomena have long been studied in the vicinity of seismic faults. It has for instance been shown that both in situ pore fluids and seismic mechanical waves travel at different velocities along various directions of a fault zone. Yet, while more and more complexity and disorder in seismic models are introduced to better understand earthquakes, frictional anisotropy is only rarely regarded. In many other domains than geophysics, however, such anisotropy in solid friction is believed to be crucial. For instance, the tribology of rubber tires, skis or advanced adhesives is improved when those are designed to have a preferential frictional direction. But numerous natural systems also benefit from such anisotropy: is is notably essential in the motion of numerous animal skins and in the efficient hydration of some plants. In most cases, these frictional anisotropies derive from the existence of preferential topographic orientations on, at least, one of the contact surfaces, with scales for such structural directivity that can be multiple and various. Seismic faults also exhibit such preferential directions in their topography: unique rock crystals, such as antigorite, can already display some frictional anisotropy, fault zones are  initiated by early fractures that often propagates through layered sediments, generating ramp-flat morphology in their surfaces and, finally, mature faults are marked by grooves of various wavelengths due to the slip induced erosion.

 

In this work, we study how the morphology of faults affects their stability, as understood by their Coulomb static coefficient of friction. In particular we study its anisotropy with the slip direction. To do so, we make use of the 3D-printing technology and print actual fault surfaces, that were measured in the field. We perform friction experiments with gypsum casts of these 3D-printed faults, as mineral-like materials might deform differently under shear than plastic materials. With these experiments, we show that the friction coefficient along seismic faults is highly anisotropic, with slip motions that are easier in, but not limited to, the direction of the main grooves. This anisotropy could for instance be paramount to better predict the next direction of rupture along some faults under a varying stress state. In some cases, it could indeed not only be related to the orientation of the main regional stress, but also to the structural anisotropy, and  depending on stress and friction anisotropy, along which orientation a rupture criterion will first be exceeded.