An experimental investigation of brittle failure mechanisms in scratch tests of rock

This paper reports the results of scratch tests on three rocks (Bonne Terre Dolomite, Alabama Marble and Dunnville Sandstone) conducted at depths of cut ranging from 0.2 mm to 2.6 mm. The experiments were performed on slabs of thickness slightly smaller than the cutter width, with a camera recording the entire cutting process. The test results reveal the existence of three cutting regimes: ductile, fragmentation, and brittle with increasing depth of cut. The regimes differ by the failure mode, the dependence of the specific energy ϵ on the depth of cut d, and the fragment shapes. In the ductile regime, ϵ∼d⁰ and the fragments are constitutive grains of the rock or powder as a result of intense shearing. In the fragmentation regime, ϵ∼d^−1/2 and the fragments is a mix of ductile-type particles and flake-like chips also the product of shear failure. Finally, the brittle regime is characterized by ϵ∼d⁻¹ and is essentially associated with chip-like fragments that come in two varieties depending on whether they are created by shear or by tension. The d⁻¹-dependence of the specific energy can only be observed if the rock is sufficiently homogeneous at the scale of the experiment. The distinction between different modes of failure is based on evidence from video recording and fractography with a Scanning Electron Microscope. It is shown that the dominance of tensile over shear fracture is intimately related to the existence of a crushed zone, which acts as a wedge to initiate a tensile macrocrack in a compressive stress field. Finally, an analysis of the energy partitioning in the brittle regime indicates that the energy dissipated in creating new surfaces with a tensile crack is actually negligible compared to the energy expended in the formation of the wedge. This result indicates that the mode I toughness cannot be determined from the specific energy in the brittle regime, on the basis of a model of a crack parallel to the free surface, in contradiction with published claims to that effect. Finally, this experimental investigation clarifies some confusion in the literature, rooted on assuming the existence of only two modes of failure, ductile and brittle.