Microphysical models for dry and wet phyllosilicate friction
MetadataShow full item record
Despite extensive research and experimental studies on the frictional behaviour of phyllosilicates little is known about the physical mechanisms controlling (pure) phyllosilicate friction. The aim of this thesis is to formulate a microphysical model for the steady state frictional behaviour of phyllosilicate gouges at relatively low temperatures (room temperature to ~400 0C), where dislocation and diffusional processes are of limited importance as grain scale deformation mechanisms. Based on a literature review it is proposed that phyllosilicate friction can potentially be controlled at one of the following scales of interaction; i) atomic scale, ii) asperity (nm) scale, iii) grain (1-10 µm) scale and iv) clast scale (10-30 µm). In order to determine the interaction scales occurring during phyllosilicates friction axial loading experiments were conducted on muscovite gouge using an applied normal load of 1.9, 2.8, 6.5 or 12.1 MPa for 24 hours. In addition, shear experiments at 6.5 MPa normal stress, strain rate velocities of 1-2 mm/s and reaching shear strains between 10-50 were performed. The deformed samples were impregnated with epoxy resin while still under load in order to eliminate unloading artefacts and preserve the actual deformed microstructure under load, thus providing the microstructural record of the active deformation mechanisms. Backscattered Scanning Electron Microscope micrographs show that this approach was indeed successful in eliminating unloading artefacts. From SEM-micrographs it is observed that during frictional sliding grains interact on level of the grain scale, i.e. the edges of interacting grains are in contact and fail by both ductile and brittle deformation during sliding. Microphysical models were subsequently derived for friction controlled by atomic and grain scale interactions for both wet and dry phyllosilicate friction, incorporating characteristics of the microstructures observed in the literature and experiments. No models were derived for asperity controlled friction, since previous authors have already studied this process extensively Thermodynamic considerations concerning the frictional processes were employed to predict the corresponding frictional parameters, such as the coefficient of friction and its dependence on velocity. By correlating the predicted frictional parameters with observations on phyllosilicate friction in the literature it is concluded that i) dry and wet phyllosilicate friction is not controlled by atomic scale interactions, ii) the leading hypothesis regarding wet phyllosilicate friction by Moore and Lockner (2004), namely viscous slip on an thin adsorbed film, is not in agreement with the observations on wet phyllosilicate friction and iii) based on the derived grain scale controlled microphysical models cleavage failure of edge-contacts do occur during frictional sliding, but do not control friction.