Formation of garnet polycrystals during metamorphic crystallization D. L. WHITNEY, 1 E. T. GOERGEN, 1 R. A. KETCHAM 2 AND K. KUNZE 3 1 Department of Geology and Geophysics, University of Minnesota, Minneapolis, MN 55455, USA (dwhitney@umn.edu) 2 Department of Geological Sciences, University of Texas, Austin, TX 78712, USA 3 Geologisches Institut, ETH-Zentrum, CH-8092 Zu ¨ rich, Switzerland ABSTRACT Garnet polycrystals may form throughout the metamorphic history of a rock, starting at the earliest stages of garnet growth when closely spaced nuclei coalesce. In mica schist from Townshend Dam, VT, electron back-scattered diffraction (EBSD) analysis shows that garnet polycrystals possess two or more distinct lattice orientations separated by high-angle boundaries (28–61°). The minimum rotational displacements required to bring these lattice orientations into concordance with each other are commonly normal to the same low-energy planes that occur as crystal faces of euhedral garnet. There is no evidence for intracrystalline deformation, and the polycrystals therefore probably represent individual garnet crystals that coalesced during growth. The boundaries cross-cut growth zoning and inclusion trails of the polycrystals, indicating that early-formed polycrystals, once coalesced, behave chemically and physically as single crystals. Statistical analysis of a 3D, high-resolution X-ray computed tomographic data set of a large sample (912 cm 3 ) of a Townshend Dam schist, combined with microprobe and EBSD analyses of garnet, are consistent with a high degree of clustering at all stages of garnet growth. The formation and prevalence of polycrystals implies that garnet nuclei impinged on each other and coalesced, and that coalescence was a common feature throughout garnet growth in the rock. Key words: electron back-scattered diffraction (EBSD); garnet; metamorphic crystallization; Townshend Dam; Vermont; X-ray computed tomography. GARNET TEXTURES AND GROWTH MECHANISMS Garnet compositions and textures are extremely useful for determining metamorphic conditions of tectonic processes, e.g. pressure, temperature, deformation and fluid history, as well as timing of metamorphism. In addition, the distribution of garnet in a rock provides information about metamorphic crystallization mech- anisms (Jones & Galwey, 1964, 1966; Kretz, 1966a, 1993; Carlson, 1989, 1991; Carlson et al., 1995; Hirsch et al., 2000). Because of the central role of garnet for evaluating tectonometamorphic processes, it is impor- tant to understand how and where garnet nucleates and grows during metamorphism. Evidence for garnet growth histories and processes may be preserved in garnet microstructures and chemical zoning. Although optical examination of garnet may not show any structural complexity, electron back- scattered diffraction (EBSD) studies of garnet micro- structures have revealed subgrain structures with low-angle boundaries (Kleinschrodt & McGrew, 2000; Kleinschrodt & Duyster, 2002; Prior et al., 2000, 2002; Mainprice et al., 2004; Storey & Prior, 2005), granular and polygonal textures (Spiess et al., 2001; Storey & Prior, 2005) and high-angle boundaries (Hirsch et al., 2003). These microstructures are important for understanding garnet growth mechanisms, and the interaction of metamorphism and deformation during and after garnet growth. Additional information about garnet growth histo- ries comes from textural analysis of garnet crystal distribution in a rock volume, allowing analysis of whether garnet growth was diffusion limited (e.g. Carlson, 1989; Spear & Daniel, 2001) (Fig. 1a) or interface controlled (Daniel & Spear, 1999; Zeh & Holness, 2003) (Fig. 1b). Although both processes may operate in a rock at different times, each process will in theory result in a distinct spatial distribution of garnet porphyroblasts. Although deformation and other fac- tors may influence the distribution of porphyroblasts in a rock, statistical analysis of the three-dimensional spacing (location and distances between neighbouring crystals) may be used as a first-order approach in assessing crystallization mechanisms (Kretz, 1966a). In diffusion-limited growth, the likelihood of impingement of crystals is reduced because each garnet will deplete the surrounding region in elements involved in garnet growth, thereby inhibiting new nucleation nearby (Fig. 1a). However, a high degree of clustering of preferred nucleation sites can counteract this effect, potentially occluding any tendency toward reduced impingement or spatial ordering. Daniel & Spear (1998) proposed that garnet porphyroblasts may J. metamorphic Geol., 2008, 26, 365–383 doi:10.1111/j.1525-1314.2008.00763.x Ó 2008 Blackwell Publishing Ltd 365