Experimental Rock-on-Rock Abrasive Wear Under Aqueous Conditions: its Role in Subglacial Abrasion

We have determined experimentally the rate of abrasive wear of rock on rock for a range of rock types as a function of normal stress and shear displacement. Unlike abrasive wear in fault zones, where wear products accumulate as a thickening gouge zone, in our experiments wear particles were removed by flowing water. The experiments are thus directly pertinent to one of the most important processes in subglacial erosion, and to some extent in river incision. Wear was produced between rotating discs machined from rock samples and measured from the progressive approach of the disc axes towards each other under various levels of normal load. Shear displacements of several km were produced. Optical and scanning electron microscopy were used to study the worn rock surfaces, and particle size distributions in wear products were characterized using a laser particle size analyzer. Rock types studied were sandstones of various porosities and cement characteristics, schists and a granite. In all cases abrasion rate decreased logarithmically with displacement by up to 2 orders of magnitude until a steady state was approached, but only after at least 1 km displacement. The more porous, less-well cemented rocks wore fastest. Amount of abrasion could be characterized quantitatively using an exponentially decaying plus a steady-state term. Wear rate increased non-linearly with normal contact stress, apparently to an asymptote defined by the unconfined compressive strength. Microstructural study showed that the well-cemented and/or lowest porosity rocks wore by progressive abrasion of grains without plucking, whereas whole grains were plucked out of weakly-cemented and/or more porous rocks. This difference in behavior was reflected in wear-product particle size distributions. Where whole-grain plucking was possible, wear products were dominated by particles of the original grain size rather than finer rock flour. Comparison of our results to glacier basal abrasive wear estimated from suspended sediment load (Findeln Glacier, Switzerland) showed the steady-state experimental data seriously to underestimate the natural wear rate. This suggests continuous resetting of the subglacial surface occurs, so that wear is continuously in the 'running-in' stage.