Ductile faulting, dynamic recrystallization and grain-size-sensitive flow of olivine

Twiss (1976) has suggested that the “ductile faulting” events observed by Post (1973) during high temperature creep of dunite are due to a transition from creep by dislocation movement to a diffusion accommodated, grain-boundary sliding mechanism following a reduction in grain size by dynamic recrystallization. Similarly, Goetze (1978) has explained both ductile faulting and water weakening of dunite by transition to a “nonlinear Coble” creep mechanism. However, the fundamental assumption made by Twiss (1976) that the stress exponent, n, reduces to unity during ductile faulting events is questionable. If the stress exponent remains high, (n≥3), then a diffusion-accomodated grain-boundary sliding mechanism is excluded. “Nonlinear Coble” creep would remain a viable alternative; however, this model fails to adequately explain the water weakening phenomenon, and the available data do not constrain us to this model. Assuming that the water-weakening phenomenon can be explained by other models (e.g., Blacic, 1972), it will be shown (by analogy with the behavior of metals) that a third model, also consistent with the available data, also qualitatively explains the observations associated with ductile faulting without appeal to a transition in creep mechanisms. The model is similar to one for metals undergoing deformation by dislocation movement and recovery by dynamic recrystallization, which commonly exhibit behavior virtually identical to that observed in dunite during ductile faulting events without transition to grain-size-sensitive creep mechanisms.