Fusion by earthquake fault friction: Stick or slip?

Abstract

[1] Field observations of pseudotachylites and experimental studies of high-speed friction indicate that melting on a slipping interface may significantly affect the magnitude of shear stresses resisting slip. We investigate the effects of rock melting on the dynamic friction using theoretical models of shear heating that couple heat transfer, thermodynamics of phase transitions, and fluid mechanics. Results of laboratory experiments conducted at high (order of m/s) slip velocities but low (order of MPa) normal stresses suggest that the onset of frictional melting may give rise to substantial increases in the effective fault strength, presumably due to viscous effects. However, extrapolation of the modeling results to in situ conditions suggests that the efficiency of viscous braking is significantly reduced under high normal and shear stresses. When transient increases in the dynamic fault strength due to fusion are not sufficient to inhibit slip, decreases in the effective melt viscosity due to shear heating and melting of clasts drastically decrease the dynamic friction, resulting in a nearly complete stress drop (''thermal runaway''). The amount of energy dissipation associated with the formation of pseudotachylites is governed by the temperature dependence of melt viscosity and the average clast size in the fault gouge prior to melting. Clasts from a coarse-grained gouge have lower chances of survival in a pseudotachylite due to a higher likelihood of nonequilibrium overheating. The maximum temperature and energy dissipation attainable on the fault surface are ultimately limited by either the rock solidus (via viscous braking, and slip arrest) or liquidus (via thermal runaway and vanishing resistance to sliding). Our modeling results indicate that the thermally activated fault strengthening and rupture arrest are unlikely to occur in most mafic protoliths but might be relevant for quartz-rich rocks, especially at shallow (<5-7 km) depths where the driving shear stress is relatively low.

Experimental investigation of frictional melting of argillite at high slip rates: Implications for seismic slip in subduction‐accretion complexes

“…However, the ratio of shear to normal stress during the strengthening stage decreased with an increase in s n : for s n of 2.67, 8.0, and 13.33 MPa, the ratio of shear to normal stress was found to be 1.2 -1.4, 0.8 -1.0, and 0.4 -0.6, respectively, at the end of the experiment (Figure 3). Since the development of a melt layer was independently confirmed during the strengthening stage, our data are suggestive of a decrease in the apparent coefficient of friction with normal stress in the postmelting regime, as predicted by theoretical models [Fialko and Khazan, 2005]. Unfortunately, the experiments were too short to confirm the postmelting steady state.…”

“…The common feature of HVRFE on crystalline rocks is that slip weakening occurs once a melt layer formed and separated host rocks completely. The HVREE on crystalline rocks shows a weak dependence of steady state shear stress on melt composition [Di Toro et al, 2006b], and numerical modeling incorporating the HVRFE results also indicates that shear stress drop during frictional melting is weakly dependent on melt viscosity likely due to a feedback between temperature and melt viscosity [Fialko and Khazan, 2005;Sirono et al, 2006]. On the basis of HVRFE on gabbro, Hirose and Shimamoto [2005] pointed out that the physical process responsible for the slip weakening after the formation of melt layer is a decrease in shear strain rate due to the growth of melt layer.…”

“…These experimental data show a complex evolution of shear stresses prior to and after the onset of melting, and explained by coupled thermal-fluidmechanical models [Fialko and Khazan, 2005;Sirono et al, 2006;Nielsen et al, 2008]. Tsutsumi and Shimamoto [1997] and Hirose and Shimamoto [2005] reported a typical evolution of apparent friction of gabbro during HVRFE: the initial slip weakening followed by the slip strengthening up to a peak shear stress (related to the formation and coalescence of discontinuous melt patches) and finally by the secondary slip weakening toward the steady state shear stress (related to temperature-dependent melt rheology).…”

“…On the basis of HVRFE on gabbro, Hirose and Shimamoto [2005] pointed out that the physical process responsible for the slip weakening after the formation of melt layer is a decrease in shear strain rate due to the growth of melt layer. Fialko and Khazan [2005] argued that the growth of melt layer is limited by squeezing of melt from the shear zone due to normal load, and attributed the observed weakening primarily to temperature dependence of melt viscosity. The rotary friction experiments on granite at high-velocities (2.0 -4.0 m/s) by Spray [2005] suggested that melt viscosity plays an important role on melt lubrication of faults: when selective melting of minerals constituting $50% of granite (i.e., feldspar) occurred at temperatures higher than the liquidus for granite ($600°C at low pressure), low-viscosity melt can be generated owing to the high temperature ($1150°C) and the low content of SiO 2 in melt zone relative to the bulk composition of granite.…”

Rock‐and‐soil avalanches: Theory and simulation

“…However, high-velocity friction experiments on dry samples of gabbro suggest that the formation of a thin molten layer along a shear zone may induce a substantial increase in the shear stress up to a local peak value, presumably due to viscous effects [Fialko and Khazan, 2005;Hirose and Shimamoto, 2005]. The peak strength is followed by an exponential decrease down to a steady state value well above zero (e.g., the peak and steady state apparent friction coefficients are m p $ 1 and m ss $ 0.5).…”