Slip-system and EBSD analysis on compressively deformed fine-grained polycrystalline olivine

A slip-system analysis was performed on two synthetic compressively deformed olivine aggregates, derived from experimental solution–gelation (sol–gel) and natural San Carlos precursors to determine how dislocation density relates to Schmid factor for slip in olivine. Individual grain orientations were measured with electron backscatter diffraction. Using decorated dislocations, grain populations were separated into subsets of high versus low dislocation density. Analysis of preferred orientations and distributions of Schmid factors suggests that there is only weak correlation between Schmid factor and dislocation density, slip on (010)[100] in San Carlos grains but (001)[100] in sol–gel material, with multiple slip or stress heterogeneity in both. Olivine, the main constituent of the Earth’s upper mantle, has been thoroughly studied in the past decades to understand how crystallographic preferred orientations (CPO) develop under upper mantle conditions due to prevailing plate tectonic forces (Ben Ismail & Mainprice 1998; Tommasi et al. 2000; Jung et al. 2006; Karato 2008), but less attention has been given to fine-grained polycrystalline olivine. CPO development commonly occurs in coarse-grained olivine during plastic deformation (i.e. high-temperature dislocation creep, .1000 8C) where individual grains rotate and/or recrystallize, promoting anisotropy in seismic measurements (Karato 1988; Lee et al. 2002; Drury & Pennock 2007). For a fine-grained aggregate, however, stabilization of grain size below that at which extensive subgrain development takes place decreases the probability of recrystallization (White 1979) and is observed in essentially dry, solution–gelation (sol–gel) derived olivine (Faul et al. 2011) or ultra-mylonites in shear zones. Several slip systems potentially operate in olivine during plastic deformation. According to the von Mises criterion, five independent slip systems are needed for maximum strain compatibility (Von Mises 1928). For each slip system (hkl)[uvw], a geometrical Schmid factor S provides a measure of resolved shear stress (0 value 0.5) from the crystal orientation relative to external stress (Hull & Bacon 2001). For uniaxial compression, the Schmid factor is defined S 1⁄4 cos a cos b where a is the angle between compression direction and slip plane normal and b is the angle between compression direction and slip direction. For a polycrystalline aggregate, homogeneous stress must be assumed to calculate S for each grain. Studies on the fabric and seismic anisotropy of olivine have shown that dominant slip systems relate to the critical resolved shear stress on each slip system and temperature. Other dependencies can include chemical potential of components such as oxygen and silica (in the case of olivine) and ‘water’. The type-A CPO typically develops in low-OH natural olivine for high-temperature plastic deformation (Green & Radcliffe 1972; Goetze 1978; Ben Ismail & Mainprice 1998; Jung & Karato 2001). This CPO is characterized in shear deformation by [100] grain directions subparallel to the shear direction and [010] subnormal to the shear plane. Models employing the three most active slip systems in olivine, (010)[100], (001)[100] and (010)[001] (Ribe & Yu 1991; Wenk et al. 1991), can replicate this CPO type. Additional slip systems may also be operating, for example, (0kl)[100] (Passchier & Trouw 1998). The range of observed CPO types is summarized in Jung et al. (2006) for a variety of water concentrations in olivine. Two end-member states describe the interactions between grains during deformation, the uniform strain (Taylor) and uniform stress (Sachs) models (e.g. Winther et al. 1997; Dawson & Wenk 2000). In the Taylor model, some grains require higher From: Prior, D. J., Rutter, E. H. & Tatham, D. J. (eds) Deformation Mechanisms, Rheology and Tectonics: Microstructures, Mechanics and Anisotropy. Geological Society, London, Special Publications, 360, 225–235. DOI: 10.1144/SP360.13 # The Geological Society of London 2011. Publishing disclaimer: www.geolsoc.org.uk/pub_ethics grain-boundary stresses than others to produce uniform strain. In the Sachs model, some grains in favourable orientations deform more than neighbouring grains; some unfavourably oriented grains do not deform at all. In general, a positive correlation between dislocation density (r) and Schmid factor is more likely if grain interactions obey the Sachs model rather than the Taylor model (Karato & Lee 1999). However, it is likely that the deformation involves heterogeneous stress or strain as implemented in the Taylor–Bishop–Hill model (Lister et al. 1978) and the self-consistent model for a limited number of slip systems (Wenk et al. 1991; Tommasi et al. 2000). More recent modelling requires activation of four essential slip systems in olivine plus another degree of freedom to describe the microscopic stress heterogeneities (Castelnau et al. 2008). This study builds upon previous studies by Karato & Lee (1999) and Lee et al. (2002) and investigates to what extent the Taylor and Sachs models are applicable in deformed polycrystalline olivine by comparing fine-grained sol–gel olivine with a coarser-grained San Carlos for high-stress (but low-strain) compressive deformation. A robust analysis of slip systems from large populations of grains with highand low-dislocationdensities has been possible because grain rotation and/or recrystallization is minor and CPO is weak or absent from our aggregates. Our new analyses have advantages over those from previous studies in two ways: firstly, Karato & Lee (1999) analysed only 5–85 grains in each of four aggregates; secondly the CPO/dislocation density analyses conducted by Lee et al. (2002) involved two aggregates which both possessed a strong overall CPO (resulting from high shear strains).