The effect of grain boundary sliding on the rheology of polymineralic rocks: Nature and numerical experiments

Geodynamic models of deformation in the crust and mantle require constitutive equations to provide a physical description of the rheology. A particularly challenging task is to derive constitutive equations that account for a variety of different micromechanical processes. In this project, microstructural maps of naturally deformed mylonites motivate numerical simulations in Elle, an open-source modeling platform for simulating the two-dimensional evolution of microstructures. The simulations presented here investigate the effect of a combination of different deformation mechanisms on rheological behavior and microstructural features. As natural examples, we use two samples from New Zealand's Alpine Faults Zone that contain microstructures interpreted to be representative of aggregates deforming by grain size insensitive (GSI) creep, grain size sensitive (GSS) creep, and their combination. The first sample contains a well-mixed assemblage consisting primarily of quartz, plagioclase, biotite, and muscovite. The foliation is defined by the alignment of individual mineral long axes and is deflected around garnet porphyroclasts and along C'-type shear bands. In addition, several thin (250 μm) but continuous monophase quartz bands occur parallel to the foliation. Electron backscatter diffraction (EBSD) analysis indicates that quartz grains located within a monophase layer (15% of the area mapped) have an average grain size of 20 μm and contain a moderately strong crystallographic preferred orientation (CPO), interpreted to signify the activity of GSI creep. In contrast, quartz grains within the mixed phase regions have an average grain size of 14 μm and exhibit statistically random orientations. The random orientations are thought to originate from GSS creep. The second sample consists of a fine-grained assemblage of quartz, plagioclase, calcite and opaque material. The foliation in this sample is defined by planar intervals of alternating quartz-rich and opaque-rich aggregates. EBSD indicates that the quartz grains have an average grain size of 7 μm and possess a weak CPO. The prevalence of very fine grained material within the second sample suggests that grain boundary sliding (GBS) played an important role in the deformation. In this case, the weak CPO of quartz could have developed if one particular crystallographic plane (e.g. rhomb) is also a preferred grain boundary orientation and commonly aligns with the shear plane. This is indicated by prevalence of rhomb-slip <c> axes in other Alpine Fault mylonites (Toy et al., 2008). GBS could be accommodated both by grain boundary diffusion and dislocation creep. Based on these typical microstructures, we introduce a new method for modeling GBS in Elle, allowing accommodation of GBS by either diffusion only or by diffusion and dislocation creep. We apply this model to simplified microstructures with different grain size distributions and evaluate the effect of the resultant dominance of GBS and GSS on the developing microstructure and bulk constitutive behavior.