SILICA LUBRICATION IN FAULTS

Silica-rich rocks are common, so silica lubrication may be important for causing fault weakening during earthquakes. In laboratory friction experiments on chert, dramatic shear weakening has been attributed to amorphization and attraction of water from atmospheric humidity to form a “silica gel”, but the details of weakening mechanism(s) remain enigmatic. Few observations of the slip surfaces have been reported, so no criteria exist on which to make comparisons of experimental to natural materials. We performed friction experiments, characterized the sliding surface, and compared these to a geological fault in the same rock type.

Experiments were performed in room humidity at 2.5 MPa normal stress with 3 and 30 m displacement for slip rates from 10^-4 – 10^-1 m/s. The friction coefficient fell from >0.6 to ~0.2 at 10^-1 m/s, but only to ~0.4 at 10^-2 – 10^-4 m/s. The slip surfaces and wear material were observed using laser confocal Raman microscopy, electron microprobe, X-ray diffraction, and transmission electron microscopy. Experiments at 10^-1 m/s formed ≤1 µm powder that is aggregated into irregular 5-20 µm clumps. Some material disaggregated during analysis with electron beams and lasers, suggesting hydrous and unstable components. Compressed powder forms smooth pavements on the surface in which grains are not visible (if present, they are <100 nm). Powder contains amorphous material and as yet unidentified crystalline and non-crystalline forms of silica (not quartz), while the worn chert surface underneath shows Raman spectra consistent with a mixture of quartz and amorphous material.

If silica amorphization facilitates shear weakening in natural faults, similar wear materials may be formed, and we may be able to identify them. However, the sub-micron particles of unstable materials are unlikely to survive in the crust over geologic time, so a direct comparison of fresh material and ancient fault rock needs to account for alteration and crystallization. The Corona fault is coated by a translucent shiny layer consisting of ~100 nm interlocking groundmass of dislocation-free quartz, 10 nm ellipsoidal particles, and interstitial patches of amorphous silica. We interpret this layer as the equivalent of the experimentally produced amorphous material after crystallizing to more stable forms over geological time.