The influence of stress history on the grain size and microstructure of experimentally deformed quartzite

Deformation of middle crustal shear zones likely varies with time as a result of the stress build-up and release associated with earthquakes and post-seismic deformation, but the processes involved and their microstructural signature in the rock record are poorly understood. We conducted a series of experiments on quartzite at 900 °C to characterize microstructures associated with changes in stress and strain rate, and to investigate the feasibility of carrying out grain size piezometry in natural rocks that experienced analogous changes. Differential stress (referred to simply as "stress") was varied in two-stage experiments by changing strain rate and by stopping the motor and allowing stress to relax. The two-stage samples preserve a microstructural record that can be interpreted quantitatively in terms of stress history. The microstructure associated with a stress increase is a bimodal distribution of recrystallized grain sizes. The smaller grains associated with the second deformation stage accurately record the stress of the second stage, and the surviving coarse grains remain similar in size to those formed during the earlier stage. The transient microstructure associated with stress decrease is a "partial foam" texture containing a larger concentration of stable 120° triple junctions than occur in samples deformed at a relatively constant strain rate. Our results indicate that microstructures preserved in rocks that experienced relatively simple, two-stage deformation histories can be used to quantitatively assess stress histories. Grain growth rates during deformation are similar to rates observed in previous isostatic growth experiments, supporting theoretical approaches to recrystallized grain size, such as the wattmeter theory (Austin and Evans, 2007), that incorporate static growth rates. From an analysis of the experimental data for quartz recrystallized grain size, we find: 1) Recrystallized grain size quickly reaches a value consistent with ambient deformation conditions. We argue that this explains a good match between average grain sizes predicted by the wattmeter after complete recrystallization and the recrystallized grain sizes of the experiments. 2) The present formulation of the wattmeter overestimates the rates at which porphyroclasts recrystallize by as much as an order of magnitude, and 3) owing to problems with extrapolation of grain growth data for quartz, the wattmeter is not presently applicable to natural samples deformed at low temperatures. We present a simplified flow law for quartz, and suggest that the change in slope of the quartz piezometer at high stress (regime 1) is related to a switch to a linear viscous rheology.