Development of Fluid Veins during Deformation of Fluid-rich Rocks close to the Brittle-Ductile Transition: Comparison between Experimental and Physical Models

International audienceTo understand the factors affecting the orientation and spacing of fluid veins in deforming silicate matrices, experimental results on the deformation of sub-micron flint have been examined. New results and data from the literature show that deformation of flint at high-temperature, fluid-rich conditions (∼6 vol. % fluid) and strains, γ, less than 0*1-0*2 leads to the development of fluid banding orthogonal to , consistent with theoretical expectations. On the other hand, above these strains and with fluid-poor conditions (∼1 vol. %), conjugate R1 and R2 Riedel bands are observed, related to brittle fracture. These latter bands cross the first generation of 45° bands formed at lower γ. It is of note that for strains up to 0*5 the strain rate remains uniform throughout the entire sample, but that above this value deformation is concentrated in a single prominent Riedel band. To understand and rationalize these observations, one-dimensional numerical modelling of fluid-rock separation during shear has been performed. The model assumes a constant strain rate and uses the interstitial fluid dependence of the pressure-solution viscosity of quartz. At the beginning of the numerical simulation fluid and solid pressures are equal (pf = ps), but shearing leads to the development of zones of compaction from which fluid is expelled, pf drops and the solid viscosity rises sharply. Because strain rate is uniform across the bulk sample, the model implies that local stress rises sharply in the compaction bands but remains low in zones where fluid accumulates. Indeed, the model shows that, in the zones of fluid extraction, both the deviatoric stress and the excess pressure (pf -ps) have the same amplitude. Their value exceeds the bulk shear stress by a factor of about five. In light of these numerical results, we tentatively suggest that in certain experimental studies low-angle fluid-rich veins may be initiated by transient embrittlement of the rock in zones of fluid loss. Subsequently, a drop in permeability orthogonal to cracking may lead to compaction in that direction, filling the cracks with fluid. Consideration of the strain rate conditions of the mantle calls into question whether low-angle melt veins will occur or not in natural olivine-basalt systems, although such veins may be relevant for fluid percolation in quartz fault gouges