Electron backscatter diffraction analysis to determine the mechanisms that operated during dynamic recrystallisation of quartz-rich rocks

Abstract

Determination of the controlling nucleation and recrystallisation mechanisms from a samples microstructure are essential for understanding how the microstructure formed and evolved through time. The aim of our research was to apply a quantified analytical approach to the identification of the controlling nucleation, recrystallisation and microstructural modification mechanisms. We used electron backscatter diffraction to quantify the microstructures of naturally deformed quartz-rich rocks which were deformed at various temperature and pressure conditions. Our results show that ratios of the recrystallised grain size to the subgrain size with values less than 1 (0.5–0.7 in the data presented here) suggest bulge nucleation, whereas ratios of ∼1 suggest subgrain rotation nucleation. Other supporting evidence for subgrain rotation nucleation is an increase in misorientation from the centre of an original protolith ‘parent’ grain to the edge. All samples show evidence for modification of the microstructure due to grain boundary sliding including increased misorientation angles between grains and movement of recrystallised grains between parent grains. By systematically analysing sample microstructures it is possible to separate out evidence to determine the controlling nucleation and recrystallisation mechanisms, as well as being able to identify microstructure modification mechanisms. Using microstructural quantification via EBSD allows a systematic methodology to analyse samples from any location from an objective viewpoint.

Introduction

The microstructures of quartz rocks deformed by creep mechanisms are important as they can be used to interpret deformation processes and conditions (Bell and Etheridge, 1976, Lloyd and Freeman, 1994, Poirier and Guillope, 1979, Urai et al., 1986) in the Earth’s crust. The pattern of quartz crystallographic preferred orientations (CPOs) (Schmid et al., 1986) can be used to constrain the conditions of deformation and the kinematic strain path. Recrystallised grain size can be used to estimate stress magnitudes (Ord and Christie, 1984, Stipp and Tullis, 2003, Twiss, 1986). An understanding of dynamic recrystallisation is crucial to understanding both CPO evolution and quantitative use of microstructures.

Dynamic recrystallisation processes are embedded in the concept of the three regimes of dislocation creep that have been identified from the microstructures of experimentally deformed quartz samples (Hirth and Tullis, 1992). Regime 1 represents lower temperatures, faster strain rates and the microstructure is characterised by inhomogenously deformed parent grains with patchy undulatory extinction and very fine recrystallised grains. Regime 2 represents increased temperature or decreased strain rate, the microstructure is characterised by flattened parent grains with sweeping undulatory extinction. The microstructure can show a core and mantle structure, and the recrystallised grain size is similar to the subgrain size of the parent grains. Regime 3 represents high temperature or low strain rate, and the microstructure is characterised by elongate parent grains with highly lobate boundaries and a high proportion of recrystallised grains.

The dominant recrystallisation mechanisms in regimes 1–3 have become a particular focus of attention (Stipp et al., 2002a). Strain-induced grain boundary migration (SIGBM) (Urai et al., 1986) and subgrain rotation recrystallisation (SGR) (Guillope and Poirier, 1979) are the two key processes identified as important in the dynamic recrystallisation of minerals. There is general consensus that regimes 1 and 3 are dominated by SIGBM, facilitated by a high driving force (dislocation density) or a high mobility (related to higher temperatures), respectively, and that regime 2 is dominated by SGR recrystallisation (G. Hirth personal communication). The nucleation mechanisms will control the initial recrystallised grain size and will influence the evolution of both rheology and CPOs. The SGR nucleation mechanism dominates in regime 2 and may also be an important nucleation mechanism in regime 3, creating new grains that can then grow by SIGBM (Urai et al., 1986). Since dislocation climb is difficult in regime 1 due to the low temperatures, SGR should not be an effective nucleation mechanism therefore nucleation is by grain boundary bulging. The relative roles of SGR and bulging for nucleation of recrystallised grains in each of the dislocation creep regimes have never been fully established.

Bestmann and Prior (2003) and Wheeler et al. (2003) suggested using the orientation of the recrystallised grains in relation to their parent grains as a tool to indicate nucleation mechanism. Nucleation via SGR or by bulging should both give rise to similar daughter and parent grain orientations and rational parent–daughter misorientation axes, which form the subgrain walls (Bestmann and Prior, 2003, Mariani et al., 2009). However many observations suggest that recrystallised grains tend to have high-angle misorientations to parent grains with non-rational, random misorientation axes (Bestmann et al., 2008, Bestmann and Prior, 2003, Halfpenny et al., 2006, Halfpenny et al., 2004, Prior et al., 2004, Skemer and Karato, 2008, Storey and Prior, 2005). Such observations may relate to nucleation mechanisms that are not fully understood (Hobbs, 1968, Vernooij et al., 2006a, Wheeler et al., 2003) or to modification by twinning. However, a more common explanation is grain boundary sliding (GBS) which allows recrystallised grains to become dispersed throughout the microstructure with a weakening of the CPO strength (Bestmann et al., 2008, Bestmann and Prior, 2003, Halfpenny et al., 2006, Skemer and Karato, 2008, Storey and Prior, 2005) and randomisation of grain boundary misorientation axes (Jiang et al., 2000). It has been suggested that SGR leads to recrystallised grains which are (within error) the same size as the internal subgrains of the parent grain (Halfpenny et al., 2006, Stipp et al., 2010, Urai et al., 1986). In contrast bulging produces recrystallised grains which are smaller than the internal subgrain size (Halfpenny et al., 2006).

Automated electron backscatter diffraction EBSD (Prior et al., 1999) is critical to this work, because it allows quantification of the microstructure in terms of grain size, subgrain size, full crystallographic orientation and misorientation (Bestmann and Prior, 2003, Halfpenny et al., 2006, Pennock et al., 2006, Stipp and Kunze, 2008, Trimby et al., 1998, Valcke et al., 2006). Naturally deformed quartz aggregates representative of regimes 1–3 (Hirth and Tullis, 1992) have been chosen for this study. The aim of this research is to understand the nucleation mechanisms during dynamic recrystallisation of quartz. We reason that if microstructures can be quantified then they can be systematically compared regardless of location or deformation conditions. This will allow the determination of the controlling nucleation and recrystallisation mechanisms as well as defining microstructural modification mechanisms after nucleation and recrystallisation have taken place.