Tectonics of Metamorphic Crystallization A DISSERTATION SUBMITTED TO THE FACULTY OF THE GRADUATE SCHOOL OF THE UNIVERSITY OF MINNESOTA BY Eric Thomas Goergen IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY IN GEOLOGY May 2009 Professor Donna L. Whitney (advisor) © Eric Thomas Goergen, May, 2009 i Acknowledgements This dissertation would not have been possible without the academic, emotional, and monetary support of many people and institutions. I would first like to thank my advisor, Donna Whitney, for her unflinching support in my work despite many roadblocks that were encountered during my years at Minnesota. Her humor and patience during my presence here was instrumental to my successful completion of my PhD. I also appreciate her constantly encouraging me to try new techniques, attend and present at meetings, and always working to make sure I was financially supported in some capacity. All of this opened up communication with other researchers as well as exposing me to directions of research I would not have otherwise sought out. I would also like to thank my thesis committee: Christian Teyssier, Marc Hirschmann, and David Kohlstedt. Christian has been a constant cheerleader for my research as well as always knowing when I needed a pep talk. I cannot thank him enough for all of his support and encouragement. Marc has been an incredible resource for me during the development of this project. He is responsible for identifying some philosophical errors to my approach in studying reaction textures early on, and this work would not have been as successful without his help. David has been instrumental to my developing an understanding of the kinetic processes active in rocks, both during deformation as well as during diffusion and reaction. David was always open for questions and his amazingly humble personality always made me feel comfortable to ask anything, regardless of how simplistic the question. ii The entire faculty of the Department of Geology and Geophysics is acknowledged for providing outstanding education and helpful and interesting classes. I would like to thank Jim Stout for our many conversations regarding metamorphic processes. These conversations ultimately led to fruitful lines of research and I thank Jim for his time and interest in my work. David Fox is thanked for being a great friend as well as an understanding ear over the years. David’s humor and encouragement was always well timed and I owe a lot of my persistence to his advice. I would also like to thank the departmental administrative and technical staff for their help and support. In particular I would like Sharon Kressler for always being there with answers to question about paperwork as well as with words of encouragement. Greg Gambeski is thanked for being a willing ear when I needed to vent, as well as the provider of great conversation and reimbursments! Mark Griffith is thanked for allowing me to borrow tools as well as use some equipment in his shop as well as for being a great guy. All of the graduate students in the department are also thanked. There are too many fellow graduate students to mention by name here, but there are several that had a huge impact on my life here. Seth Kruckenberg was the only one I knew coming into the department and his friendship and support over the years has been a constant and I could always rely on him in dark times. We kept each other sane over the past few months as we both pushed to finish our dissertations. Rory McFadden has been an incredible friend. His support and advice, while usually straight to the point, made me a better scientist as well as a better friend and person. Peter Davis is thanked for his friendship and iii unwavering support. I could not have finished this degree without his advice and friendship. Former officemates Nick Pester and Fawna Korhonen are thanked for many conversations, related and unrelated to research, as well as being great friends. Dylan Blumentritt, Aydin Aycenk, Adam Nagle, Kristina Brady, Megan Kelly, Jess Till and many many other students are thanked for their friendship and support. I want to thank my family for being unbelievable throughout the course of my education. I especially want to thank my parents for their belief in me and counseling when I doubted myself. I also want to thank my Father for being a great field assistant during a field season in Thor-Odin dome. I cannot express into words what my family has meant to and done for me over the years. Simply put, I would not have accomplished this without their unending faith and love. Finally, I want to thank Alissa Hodges for being my partner in crime, my strong shoulder, and constant cheerleader. I would have given up long ago without her in my life. She put up with a lot of long nights alone, but was always there when I needed her most. This work was supported financially by The National Science Foundation in a grant to Donna Whitney as well as a graduate student research grant from the Geological Society of America awarded to ETG. iv Dedication This dissertation is dedicated to my grandparents George and Helen Goergen and Dale and Mary Len Ockerman. They provided me an amazing example of how to live a rich and fruitful life as well as teaching me that the most important things in life are the relationships with those you surround yourself with. v Table of Contents Acknowledgements ............................................................................................................ i Dedication ......................................................................................................................... iv Table of Contents .............................................................................................................. v List of Tables .................................................................................................................... ix List of Figures.................................................................................................................... x Chapter 1 Introduction ................................................................................................... 1 GENERAL BACKROUND AND MOTIVATION................................................................. 1 Interaction of deformation and reaction processes .................................................................. 2 P-T evolution and application of phase-equilibria to texturally complex rocks ..................... 4 Changing scales of mass transport and disequilibria during evolving P-T conditions............ 6 Chapter 2 Deformation-induced polymorphic transformation: experimental deformation of kyanite, andalusite, and sillimanite..................................................... 15 Introduction.............................................................................................................................. 16 Al2SiO5 polymorphs: crystal chemistry, structure, and phase stability ................................. 19 Experimental Methods ............................................................................................................ 20 Sample preparation................................................................................................................ 20 Experimental conditions and methods................................................................................... 21 Microstructural Characterization........................................................................................... 23 EBSD analyses ...................................................................................................................... 24 vi STEM imaging ...................................................................................................................... 25 Results ....................................................................................................................................... 26 Pre-deformed (hot-pressed) sample characterization ............................................................ 26 Torsion results ....................................................................................................................... 27 Discussion ................................................................................................................................. 32 Evaluation of experimental stress-strain relationships .......................................................... 33 The effect of ductile deformation on sillimanite ................................................................... 35 Deformation mechanisms, microstructural development and phase transformation: interaction of metamorphism and deformation ..................................................................... 36 Summary................................................................................................................................... 40 Chapter 3 Application of phase equilibria to compositionally and texturally complex rocks: orthoamphibole-cordierite gneiss of Thor-Odin Dome .................... 67 INTRODUCTION ................................................................................................................... 68 GEOLOGIC SETTING .......................................................................................................... 71 ORTHOAMPHIBOLE-CORDIERITE GNEISS OF THE THOR-ODIN DOME........... 72 PETROLOGY AND MINERAL CHEMISTRY .................................................................. 74 Southern Thor-Odin dome: samples 06ET-2, 02-3.03, 05-2.54 01-2.07 .............................. 75 PETROGENESIS OF ORTHOAMPHIBOLE-CORDIERITE ROCKS........................... 80 BULK COMPOSITION AND PHASE DIAGRAM CALCULATIONS............................ 82 Determination of bulk composition for use in pseudosections ............................................. 83 Pseudosection calculations .................................................................................................... 84 DISCUSSION ........................................................................................................................... 89 Timing of bulk chemical modification .................................................................................. 91 P-T conditions ....................................................................................................................... 92 vii Timing of reaction texture initiation...................................................................................... 94 Chapter 4 Tectonic evolution and evolving scales of open system behavior: controls on textural heterogeneity in symplectitic coronal reaction textures ........................ 139 OVERVIEW OF REACTION TEXTURES ....................................................................... 140 GEOLOGIC SETTING ........................................................................................................ 146 MINERAL CHEMISTRY AND DESCRIPTION OF TEXTURES................................. 148 Coronal textures centered on Al2SiO5 ................................................................................. 149 Coronal textures centered on garnet .................................................................................... 151 2D Morphology and quantitative textural assessment of symplectitic coronas .................. 152 3D Morphology (HRXCT) .................................................................................................. 154 CRYSTALLOGRAPHIC REALTIONSHIPS IN SYMPLECTITIC CORONAS ......... 156 Background and motivation ................................................................................................ 157 EBSD analysis of symplectitic coronal textures ................................................................. 158 Grain-scale structure and general crystallographic trends................................................... 159 Crystallographic relationships between host and vermicular phases .................................. 161 Crystallographic relationships between corona layers and Al2SiO5.................................... 163 ESTIMATES OF MATERIAL TRANSPORT................................................................... 164 Estimates of material transport in symplectitic coronas associated with Al2SiO5 .............. 166 Estimates of material transport in corona textures associated with garnet.......................... 170 DISCUSSION ......................................................................................................................... 171 Controls on textural heterogeneity ...................................................................................... 173 Controls on morphology and grain/interface-scale processes in symplectitic reaction textures ............................................................................................................................................. 179 CONCLUSIONS .................................................................................................................... 183 viii Chapter 5 Conclusion to the thesis ............................................................................. 236 The effect of deformation on reactions ............................................................................... 237 Controls on reaction texture development and heterogeneity ............................................. 238 References...................................................................................................................... 242 Appendix........................................................................................................................ 270
 ix List of Tables Table 3.1 Mineral assemblages for orthoamphibole-cordierite gneisses from Thor-Odin Dome. ..................................................................................................................113 Table 3.2 Representative microprobe analyses of orthoamphibole. ...............................114 Table 3.3 Representative microprobe analyses of biotite. .............................................115 Table 3.4 Representative microprobe analyses of garnet. ............................................. 116 Table 3.5 Representative microprobe analyses of cordierite, spinel, sapphirine, plagioclase........................................................................................................... 117 Table 3.6 Bulk compositions of orthoamphibole-cordierite rocks from Thor-Odin Gneiss Dome................................................................................................................... 118 Table 3.7 Bulk composition correction for biotite replacement. ................................... 119 Table 3.8 Bulk compositions used in pseudosection calculations. ................................ 120 Table 4.1 Representative microprobe analyses of orthoamphibole, cordierite, spinel, sapphirine, plagioclase, and biotite. ....................................................................207 Table 4.2 Measurements of symplectitic layer assemblages. ........................................208 Table 4.3 Estimated volume change calculations for symplectitic reaction coronas. ...209 Table 4.4 Survey of crystallography of symplectitic phase assemblages. .....................210 x List of Figures Figure 2.1 Phase diagram for the Al2SiO5 system. .......................................................... 54 Figure 2.2 Schematic of sample charge displaying different sections cut for analysis. ..55 Figure 2.3 Pole figures for kyanite, andalusite and sillimanite from the pre-deformed (hot-press) state. ....................................................................................................56 Figure 2.4 STEM dark-field images of pre-deformed kyanite and andalusite. ................57 Figure 2.5 The assembly of the Al2SiO5 triple stack experiment. ....................................58 Figure 2.6 Pole-figures of kyanite, andalusite and sillimanite from the triple stack experiment. ............................................................................................................59 Figure 2.7 Pole-figures from the andalusite and sillimanite individual torsion experiments. .......................................................................................................... 60 Figure 2.8 M-index calculations for the pre-deformed, andalusite axial, andalusite tangential and sillimanite tangential EBSD analyses. ...........................................61 Figure 2.9 Photomicrographs comparing the Andalusite torsion experiment to torsion experiments on Olivine+MORB. ..........................................................................62 Figure 2.10 SEI images of fibrolite mantles around relict andalusite. .............................63 Figure 2.11 Photomicrographs and STEM dark field images illustrating representative microstructures in Andalusite torsion experiments. ..............................................64 Figure 2.12 Photomicrographs and STEM dark field images illustrating representative microstructures in Kyanite torsion experiments. ..................................................65 Figure 2.13 Stress-strain dependence and the effect of stress-exponent. ........................66 Figure 3.1 Regional map of the Shuswap metamorphic core complex. ........................121 xi Figure 3.2 Simplified geologic map of Thor-Odin Dome. ............................................122 Figure 3.3 Photograph of reaction textures associated with garnet and Al2SiO5 as observed in outcrop. ............................................................................................123 Figure 3.4 Field relationships of migmatites ad OCG pods. ..........................................124 Figure 3.5 Full thin-section scans of OCG showing biotite modal variability. .............125 Figure 3.6 Photomicrographs of the typical replacement texture of cordierite after orthoamphibole observed in OCG of this study. ................................................126 Figure 3.7 Compositional variation of OCG from Saturday Glacier area (a) Orthoamphibole end-member composition. .......................................................127 Figure 3.8 Photomicrograph illustrating the color change of biotite associated with garnet and symplectitic reaction textures. ...........................................................128 Figure 3.9 Photomicrographs of garnet textures in OCG. .............................................129 Figure 3.10 Garnet compositional zoning. .....................................................................130 Figure 3.11 Photomicrographs from one thin section (06ET-2B) illustrating textural heterogeneity associated with symplectitic reaction textures. ............................131 Figure 3.12 Sapphirine compositional variability. .........................................................132 Figure 3.13 Backscattered electron image of a typical symplectite associated with Al2SiO5. ...............................................................................................................133 Figure 3.14 Ternary and tetrahedral plots illustrating major element variation in mafic volcanic and Al-Mg rich rocks. ..........................................................................134 Figure 3.15 Pseudosection for OCGs assuming H2O in excess. ....................................135 Figure 3.16 T-MH2O pseudosection for OCGs from Thor-Odin dome. ........................136 xii Figure 3.17. Schematic P-T diagram summarizing the stability of hornblende with changing effective bulk composition. .................................................................137 Figure 3.18 Schematic illustrating the relative timing of bulk compositional modification and reaction texture initiation. ............................................................................138 Figure 4.1 Schematic illustrating the differences between layered corona and layered symplectitic reaction textures. ............................................................................211 Figure 4.2 Generalized geologic maps of the Shuswap metamorphic core complex and Thor-Odin dome. .................................................................................................212 Figure 4.3 Plane-polarized (PPL) photomicrographs of symplectitic reaction textures after Al2SiO5.. ......................................................................................................213 Figure 4.4 Backscattered electron (BSE) images of symplectitic reaction textures after Al2SiO5. ...............................................................................................................214 Figure 4.5 BSE images (a-b) and PPL photomicrograph (c) illustrating the preservation of the shape of the central Al2SiO5. .........................................................................215 Figure 4.6 PPL photomicrographs of typical morphologies of symplectitic assemblages. ..............................................................................................................................216 Figure 4.7 Compositional (Mg#-Mg/Mg+Fe) traverses across Cd in symplectitic reaction textures. ...............................................................................................................217 Figure 4.8 PPL photomicrographs of garnet coronal texures. .......................................218 Figure 4.9 PPL photomicrographs of the textural associations of garnet and Al2SiO5 coronal textures. ..................................................................................................219 Figure 4.10 BSE image of Sp+Cd and Sp+Pl symplectitic layers. ................................220 xiii Figure 4.11 Six slices taken from HRXCT results on a single reaction texture. ...........221 Figure 4.12 Four slices taken from HRXCT data on a core of OCG to illustrate the interconnectivity between coronal textures. .......................................................222 Figure 4.13 EBSD maps of symplectitic layers. ............................................................223 Figure 4.14 EBSD band contrast maps overlain with coloration based on spinel crystallographic orientations relative to the Y-axis of the image. ...................... 224 Figure 4.15 Orthogonal EBSD maps of a Sp+Pl symplectitic layer. .............................225 Figure 4.16 Pole figures of cordierite single crystals and associated included spinel. ..226 Figure 4.17 Pole figures of plagioclase single crystals and included spinel. ......... 227-228 Figure 4.18 Pole figures illustrating crystallographic relationships of host crystals with Al2SiO5. ...............................................................................................................229 Figure 4.19 Results of material transport calculations assuming the complete absence, or the presence of only a single symplectitic layer. ................................................230 Figure 4.20 Material transport calculations considering different reactants in the original assemblage. .........................................................................................................231 Figure 4.21 Material transport calculations of garnet reaction textures. ........................232 Figure 4.22 Qualitative isothermal-isobaric µ-µ diagrams. ...........................................233 Figure 4.23 Representative pseudosection of OCG rocks from the Thor-Odin Dome. .234 Figure 4.24 Schematics illustrating the interpreted 3D morphology of symplectitic intergrowths, the relationship of 2D slices with observed shape variation and the likely growth mechanism responsible for the observed morphology. ................235 1 Chapter 1 INTRODUCTION "Often interpretations of [reaction textures] seem to lie so close at hand, that it appears only necessary to grasp a pen and write them down on paper. But, as we have found,... the reactions between minerals have been complicated by transfer of material from greater distances, thus increasing the difficulty of an interpretation" -J.J. Sederholm 1916 GENERAL BACKROUND AND MOTIVATION Metamorphic reactions occur in response to changes in the pressure-temperature- composition (e.g. fluid, bulk-rock) and deformation states of a system in an attempt to re- equilibrate to a new set of conditions. Although the parameters controlling metamorphic reactions are well known, unraveling the contribution of each to the resultant rock textures and microstructures can be difficult. Research related to understanding and interpreting metamorphic reactions and their associated relationship to textures and microstructures in the context of tectonic evolution began first with the discovery of Barrow’s zones (Barrow, 1893) and shortly followed by the first proper delineation of mineral facies by Eskola (1914). Despite over one hundred years of advances in analytical, empirical, and theoretical techniques, since the first recognition of the 2 importance of metamorphic processes in interpreting the geologic history of a terrain, there still exist fundamental questions with respect to our complete understanding of how evolving systems affect the textural and microstructural evolution of rocks. The work presented in this thesis aims to investigate the processes and conditions of metamorphic reactions. In particular, the role of deformation, P-T evolution and evolving scales of mass transport in reaction initiation and evolution are considered in detail. The following discussion describes the general background and motivation for research related to these topics. INTERACTION OF DEFORMATION AND REACTION PROCESSES The fundamental relationship between plastic deformation of rocks and metamorphic processes has long been recognized in geologic research (e.g. Teall, 1893; Mitra, 1978; Rutter & Brodie, 1985; Rubie, 1998). The extent of interaction between the two processes is difficult to determine because, under some circumstances, they may be dependent processes. Metamorphic reactions can enhance rheological weakening, as is often the case with the development of discrete crustal and mantle shear zones (e.g. White et al., 1990; Vissers et al., 1991). Deformation can, in turn, enhance the re-equilibration of a rock volume (e.g. Wintsch, 1985). Ductile deformation processes in rocks affect reaction kinetics through the generation and subsequent mobilization of dislocations and grain boundaries (e.g. Cahn, 1957; Snow & Yund, 1987). Investigations into quantifying the positive feedback between reaction and deformation processes can be carried out in a laboratory setting because rheology, 3 reaction kinetics and active deformation mechanisms can be determined/controlled during the course of an experiment. Deformation reaction experiments have focused on polymorphic phase transitions (e.g. Meike, 1993; Schmidt et al., 2003), dehydration and rehydration reactions (e.g. Rutter & Brodie, 1988; Stünitz and Tullis, 2001) as well as in nominally anhydrous systems (e.g. de Ronde et al., 2005). Many interpretations of the metamorphic and deformational histories of terrains rely on Al-rich porphyroblasts, in particular, garnet and the Al2SiO5 polymorphs. These refractory phases have relatively large stability fields with respect to P-T space, and are relatively strong compared to matrix phases. Therefore these phases tend to preserve a record of the metamorphic reaction and deformation history of a terrain that would have been overprinted in the rock matrix. However, how plastic deformation affects these Al- rich phases, specifically the Al2SiO5 polymorphs, is not well understood. Relative amounts of apparent deformation are commonly used to infer the timing of crystallization of the Al2SiO5 polymorphs when more than one is present, and thus a rough estimate of the P-T trajectory can be inferred. If there are major differences in the relative strengths of these polymorphs, such interpretations can be made with more confidence. To this end Chapter 2 focuses specifically on the role of deformation of the Al2SiO5 polymorphs and subsequent chapters focus on the chemical and textural record of Al-rich phases. Chapter 2 of this thesis focuses on the experimental deformation of the Al2SiO5 polymorphs: andalusite, sillimanite and kyanite. The Al2SiO5 system was selected because of its importance for interpreting the P-T histories of metamorphic rocks despite the difficulty, related to the common presence of more than one of the polymorphs in a 4 single rock, of interpreting crystallization sequences (e.g. Holdaway, 1978; Garcia-Casco & Torres-Roldan, 1996; Whitney, 2002). Chapter 2 presents experiments designed to examine how strain affects polymorphic transformation in these historically difficult to transform phases. Microstructural development and implications for transformation mechanisms are also presented. P-T EVOLUTION AND APPLICATION OF PHASE-EQUILIBRIA TO TEXTURALLY COMPLEX ROCKS Rocks respond to changing P-T conditions through either continuous compositional exchange reactions (e.g. Mg-Fe exchange) or via discontinuous reactions resulting in changes in the stable phase assemblage for a given set of conditions. Observations that these reactions can have a systematic relationship with changing P or T led to empirical studies of commonly observed reactions and calibrations of their respective dP-dT dependence that were subsequently used to quantify the P-T conditions of metamorphic and mantle rocks (Goldman & Albee, 1977; O’Neil & Wood, 1979; Bohlen et al., 1983; Perkins, 1987). Empirical data also led to the development of internally consistent thermodynamic datasets (e.g. Berman, 1988; Holland and Powell 1998) allowing for modeling of phase assemblage evolution in relative large, and more geologically meaningful compositional systems. Such data sets and the development of activity- composition models for important rock forming minerals resulted in the development of quantitative petrogenetic grids. Petrogenetic grids provide the ability to explore pressure-temperature relationships of low-variant assemblages and associated reactions thereby allowing for 5 the interpretations of P-T paths responsible for the phase assemblages and textures preserved in rocks. However, with further advances of mineral a-x models, it has become clear that rock textural and phase composition evolution are dominated by higher- variance reactions than cannot be presented in traditional petrogenetic grids (Holland & Powell, 1998). Pseudosections are P-T-X projections of phase relationships based on an assumed rock bulk composition. Pseudosections allow for the simultaneous presentation of phase assemblages, phase compositions, and mode percents, which can be used to tightly constrain the P-T conditions of a rock volume – assuming that assemblages are in textural equilibrium and that the bulk-rock composition did not change during metamorphism (e.g. Connolly, 1990; Powell and Holland, 1998; White et al., 2001). However, it is clear that the assumption of constant bulk-rock composition is not possible in many cases of metamorphic evolution. Furthermore, the volume of rock that represents the effective bulk composition – the volume of rock capable of chemical communication and thus controlling the stable phase assemblages – is often difficult to determine (e.g. Stüwe, 1997). Chapter 3 focuses on the application of pseudosection analysis on texturally complex orthoamphibole-cordierite gneisses from the Thor-Odin dome, British Columbia, Canada. Orthoamphibole-cordierite rocks have a pod-shaped geometry and are interlayered with high-grade migmatites and contain coronal reaction textures that formed during the exhumation of the terrain via isothermal decompression during the Eocene (Vanderhaeghe et. al., 1999; Teyssier et al., 2001; Norlander et al., 2002). The orthoamphibole-cordierite gneisses are explored in the NCKFMASHTO chemical system 6 in an attempt to determine the conditions of metamorphism prior to decompression as well as the timing of initiation of reaction texture formation. Problems associated with the application of phase equilibria in texturally complex rocks are presented. In particular, the relationship between melting and crystallization of associated migmatites is explored. Problems associated with determining the effective bulk compositions are presented along with a discussion of the influence of large-scale diffusive processes (related to reactions between the orthoamphibole-cordierite pods and surround melt) on controlling the observed phase assemblages in these complex rocks. CHANGING SCALES OF MASS TRANSPORT AND DISEQUILIBRIA DURING EVOLVING P-T CONDITIONS The common preservation of metamorphic reaction textures and the preservation of compositional zoning in minerals is evidence that the attainment of equilibrium, in a strict sense, is rarely obtained throughout the P-T evolution of a metamorphic rock. Metamorphic reaction textures record the elaborate interplay between diffusion and reaction processes in rocks (Ridley & Thompson, 1986). Reactions lead to the development of dispersed mineral segregations as well as in coronal reaction textures. Such textures are evidence of the importance of intergranular diffusion and other kinetic processes in the dispersion or localization, rates of reaction and rates of nucleation and growth of new phases during metamorphism (e.g. Carmichael, 1969; Foster, 1981; Carlson, 1989). 7 Material transport via diffusion in rocks, and its importance in metamorphic reactions has long been known in the earth sciences (e.g. Sederholm, 1916; Korzhinskii, 1959; Thompson, 1959; Fisher, 1973). Coronal reaction textures are extraordinary examples of the kinetic dominance of diffusion in systems that have departed from equilibrium. Coronal reaction textures form when two phases become unstable associations and a reaction related to the breakdown of one or both of the phases is overstepped with respect to P, T or X (e.g. Korzhinskii, 1959; Fisher, 1977; Joesten, 1986; Carlson & Johnson, 1991). The result is the development of one or more layers of product assemblages that are stable under the constraints of chemical potential gradients developed between the two phases (e.g. Joesten, 1977; Ashworth & Birdi, 1990; Santirini-Rideout, 2009). Coronal reaction textures can be divided into two subsets: polygonal layered coronas and symplectitic layered coronas. The two subsets are only differentiated based on their observed morphology. Symplectites are complex intergrowths (typically lamellar or vermicular in form) of two or more phases. Steady state diffusion models have been developed to constrain the relative magnitude of chemical potential gradients and component diffusivities controlling the development of the observed phase assemblages in both coronal and symplectitic reaction textures. (e.g. Fisher, 1974; Joesten, 1977; Johnson & Carlson, 1990; Ashworth et al., 1998). However, such models are currently incapable of dealing with evolution in the component chemical potentials controlling diffusional growth of the coronas due to changes in the reaction structure of the bulk rock (Carlson & Johnson, 1991). Such changes are exemplified by heterogeneity within the reaction texture. 8 Chapter 4 discusses the role of changing scales of material transport and evolving P-T conditions on the formation of symplectitic reaction textures in orthoamphibole- cordierite gneisses from the Thor-Odin dome. The symplectitic reaction textures are associated with the breakdown of Al2SiO5 and contain extreme heterogeneity with respect to product assemblages. The likely causes of textural heterogeneity are presented, with emphasis on the importance of P-T evolution. In addition, crystallographic data collected via electron backscattered diffraction are presented in the context of understanding growth mechanisms responsible for the symplectitic morphology observed in these textures, with comparison to polygonal layered coronas. Together, these chapters represent a comprehensive investigation of the deformation and reaction mechanisms of Al-rich phases under evolving P-T-X-d conditions. The refractory nature of Al-rich minerals makes them well suited for preserving the reaction and deformation histories in a sample; however, the refractory characteristic of these minerals can also make such interpretations more difficult. This fact is preserved in the common development of diffusionally controlled reaction textures commonly associated with Al-rich phases. 9 REFERENCES Ashworth JR, Sheplev, V.S., Bryxina, N.A., Kolobov, V.Y. & Reverdatto, V.V., 1998. Diffusion-controlled corona reaction and overstepping of equilibrium in a garnet granulite, Yenisey Ridge, Siberia. Journal of Metamorphic Geology, 16, 231-246 Ashworth, J. R. & Birdi, J. J., 1990. Diffusion modelling of coronas around olivine in an open system. Geochimica et Cosmochimica Acta, 54, 2389–2401 Barrow, G. 1893. On an intrusion of muscovite-biotite gneiss in the east highlands of Scotland, and its accompanying metamorphism. Journal of the Geological Society of London, 49, 330-358. Bohlen, S.R., Wall, V.J., & Boettcher, A.L. 1983. Experimental investigation and applicationof garnet granulite equilibria, Contributions to Mineralogy and Petrology, 83, 52-61. Cahn, J. W., 1957. Nucleation on dislocations. Acta Met. 5, 169–172. Carlson W. D. and Johnson C. D. ( 1991 ) Coronal reaction textures in garnet amphibolites of the Llano Uplift. Amer. Mineral. 76, 756-772. Carlson, W. D., 1989. The significance of intergranular diffusion to the mechanisms and kinetics of porphyroblast crystallization. Contributions to Mineralogy and Petrology, 103, 1–24 Carmichael, D.M., 1969; On the Mechanism of Prograde Metamorphic Reactions in Quartz-Bearing Pelitic Rocks, Contributions to Mineralogy and Petrology, no. 20 p. 244-267 10 de Ronde, A. A., Heilbronner, R., Stunitz, H., Tullis, J., 2004. Spatial distribution of deformation and mineral reaction in experimentally deformed plagioclase olivine aggregates. Tectonophysics 389 (1-2), 93–109. Eskola, P. 1915. On the relations between the chemical and mineralogical composition in the metamorphic rocks of the Orijarvi region. Bulletin of the Commision of Geology, FInlande, 44, 1-277. Fisher G. W. (1973) Nonequilibrium thermodynamics as a model for diffusion-controlled metamorphic processes. Amer. J. Sci. 273, 897-924. Foster, C. T., Jr., 1986, Thermodynamic models of reactions involving garnet in a sillimanite/ staurolite schist, in Yardley, and D ; Harte, eds., Mechanisms of metamorphic reactions.: Mineralogical Magazine: London, United Kingdom, Mineralogical Society, p. 427-439 García-Casco, A., and R.L. Torres-Roldán, 1996. Disequilibrium induced by fast decompression in St-Bt-Grt-Ky-Sil-And metapelites from the Betic belt (Southern Spain), J. Petrol., 37, 1207-1239. Goldman, D.S., & Albee, A.L., 1977. Correlation of Mg/Fe partitioning between garnet and biotite with O18/O16 partitioning between quartz and magnetite. American Journal of Science, 277, 750-761. Holland, T. J. B. & Powell, R., 1998. An internally-consistent thermodynamic data set for phases of petrological interest. Journal of Metamorphic Geology, 16, 309–343. 11 Joesten R. (1986) The role of magmatic reaction, diffusion, and annealing in the evolution of coronitic microstructure in troctolitic gabbro from Risrr, Norway. Mineral. Mag. 50, 441-467. Korzhinskii D. S. (1959) Physicochemical Basis of the Analysis of the Paragenesis of Minerals. Consultants Bureau USA Meike, A., 1993. A critical review of investigations into transformation plasticity. In: Boland, J. N., Fitz Gerald, J. D. (Eds.), Defects and Processes in the Solid State: Bibliography 173 Geoscience Applications (The McLaren Volume). Vol. 14. Elsevier, Amsterdam, pp. 5–25. Mitra, G., 1978. Ductile deformation zones and mylonites: the mechanical processes involved in the deformation of crystalline basement rocks. Am. J. Sci. 278, 1057– 1084. Norlander, B.H., Whitney, D.L., Teyssier, C. & Vanderhaeghe, O., 2002. Partial melting and decompression of the Thor-Odin dome, Shuswap metamorphic core complex, Canadian Cordillera. Lithos, 6, 103-125. O’Neil, H.S., & Wood, B.J., 1979. An experimental study of Fe-Mg partitioning between garnet and olivine and its calibration as a geothermometer. Contributions to mineralogy and Petrology, 72, 337. Perkins, D. 1987. Two independent garnet-clinopyroxene-plagioclase-quartz barometers. Geological Society of America Abstracts, 19, 803. Ridley, J., and Thompson, A.B. (1986) The role of mineral kinetics in the development ofmetamorphic microtextures. Advances in Physical Geochemistry, 5, 154-193 12 Rubie, D. C., 1998. Disequilibrium during metamorphism: the role of nucleation kinetics. In: Treloar, P. J., O’Brien, P. J. (Eds.), What Drives Metamorphism and Metamorphic Reactions. Vol. 138. Geol. Soc., Spec. Pub., London, pp. 199–214 Rutter, E. H., Brodie, K. H., 1988a. Experimental ”syntectonic” dehydration of serpentinite under conditions of controlled pore water pressure. J. Geophys. Res. 93 (5), 4907–4932. Satirini-Rideout, C., Gilotti, J.A., & Foster, C.T., 2006. Forward modeling corona growth in a partially eclogitized leucogabbro, Bourbon Island, North-East Greenland. Lithos, 56, 700-717. Schmidt, C., Bruhn, D., Wirth, R., 2003. Experimental evidence of transformation plasticity in silicates: minimum of creep strength in quartz. Earth. Planet. Sc. Lett. 205, 273–280. Sederholm, J.J. (1916) On synantetic minerals and related phenomena. Bulletin de la Commission Geologique de Finlande, 48, l-148. Snow, E., Yund, R. A., 1987. The e?ect of ductile deformation on the kinetics and mechanisms of the aragonite-calcite transformation. J. Metamorph. Geol. 5 (2), 141–153. Stuewe, K., 1998, The influence of effective bulk composition on retrograde assemblage development; I, Conceptual model and relevant phase diagrams, in Treloar, J ; and Brien, eds., Proceedings; What drives metamorphism and metamorphic reactions; heat production, heat transfer, deformation and kinetics? Extended 13 abstracts.: Electronic Geology: Portsmouth, United Kingdom, Electronic Journals Limited Stunitz, H., Tullis, J., 2001. Weakening and strain localization produced by syndeformational reaction of plagioclase. Int. J. Earth Sci. 90 (1), 136–148. Teall, J. J. H., 1885. The metamorphosis of dolerite into hornblende-schist. Quart. J. Geol. Soc. London 41, 133–145. Teyssier, C. & Whitney, D.L., 2002. Gneiss domes and orogeny. Geology, 30, 1139-1142 Thompson, J. B., 1959. Local equilibrium in metasomatic proesses. In Researches in Geochemistry (ed. Abelson, P. H.), pp. 427–457. Wiley, New York Vanderhaeghe, O., Teyssier, C. & Wysoczanski, R., 1999. Structural and geochronological constraints on the role of partial melting during the formation of the Shuswap metamorphic core complex at the latitude of the Thor-Odin Dome, British Columbia. Canadian Journal of Earth Sciences, 36, 917-943. Vissers, R. L. M., Drury, M. R., Hoogerduijn Strating, E. H., van der Wal, D., 1991. Shear zones in the upper mantle; a case study in an Alpine lherzolite massif. Geology 19 (10), 990–993. White, R.W., Powell, R., & Clarke, G.L., 2002. The interpretation of reaction textures in Fe-rich metapelitic granulites of the Musgrave Block, central Australia; constraints from mineral equilibria calculations in the system K (sub 2) O-FeO- MgO-Al (sub 2) O (sub 3) -SiO (sub 2) -H (sub 2) O-TiO (sub 2) -Fe (sub 2) O (sub 3). Journal of Metamorphic Geology, 20, 41-55 14 White, S. H., Burrows, S. E., Carreras, J., Shaw, N. D., Humphreys, F. J., 1980. On mylonites in ductile shear zones. J. Struct. Geol. 2 (1-2), 175–187. Whitney, D.L., 2002. Coexisting andalusite, kyanite, and sillimanite: sequential formation of three polymorphs during progressive metamorphism near the Al2SiO5 triple point, Sivrihisar, Turkey, Am. Mineral., 84, 405-416 15 Chapter 2 DEFORMATION-INDUCED POLYMORPHIC TRANSFORMATION: EXPERIMENTAL DEFORMATION OF KYANITE, ANDALUSITE, AND SILLIMANITE Eric T. Goergen, Donna L. Whitney, Mark E. Zimmerman, and Takehiko Hiraga Modified version baded on a paper published (2008) in Tectonophysics v. 454, pp. 23-35 (Used by permission) Torsion experiments were performed on the Al2SiO5 polymorphs in the sillimanite stability field to determine basic rheological characteristics and the effect of deformation on polymorphic transformation. The experiments resulted in extensive transformation of andalusite and kyanite to sillimanite. No transformation occurred during the hot-press (no deformation) stage of sample preparation, which was carried out at similar P-T conditions and duration as the torsion experiments. Experiments were conducted on fine-grained (<15 µm) aggregates of natural andalusite, kyanite and sillimanite, at 1250°C, 300 MPa, and a constant shear strain rate of 2 x 10-4/sec to a maximum shear strain of 400%. Electron back-scattered diffraction (EBSD) analysis of the experiments revealed development of lattice-preferred orientations, with alignment of sillimanite and andalusite [001] slightly oblique to the shear plane. The kyanite experiment could not be analyzed 16 using EBSD because of near complete transformation to sillimanite. Very little strain (~30%) is required to produce widespread transformation in kyanite and andalusite. Polymorphic transformation in andalusite and kyanite experiments occurred primarily along 500 µm wide shear bands oriented slightly oblique and antithetic to the shear plane and dominated by sub-µm (100-150 nm) fibrolitic sillimanite. Shear bands are observed across the entire strain field preserved in the torsion samples. Scanning transmission electron microscope imaging shows evidence for transformation away from shear bands; e.g. fibrolitic rims on relict andalusite or kyanite. Relict grains typically have an asymmetry that is consistent with shear direction. These experimental results show that sillimanite is by far the weakest of the polymorphs, but no distinction can yet be made on the relative strengths of kyanite and andalusite. These observations also suggest that attaining high bulk strain energy in strong materials such as the Al2SiO5 polymorphs is not necessary for triggering transformation. Strain energy is concentrated along grain boundaries, and transformation occurs by a dynamic recrystallization type process. These experiments also illustrate the importance of grain-size sensitive creep at high strains in system with simultaneous reaction and deformation. INTRODUCTION Deformation affects metamorphic reactions by influencing the rate and sites of reactions (Flinn, 1965; Brodie and Rutter, 1985; Yund and Tullis, 1991; Williams, 1994; Wintsch et al., 1995; Baxter and DePaolo, 2004; Barnhoorn et al., 2005; Holyoke and Tullis, 2006; de Ronde et al., 2007), and may, in some cases, affect the ability of metamorphic minerals to reach equilibrium. Metamorphic reactions may in turn influence deformation mechanisms by changing mineral assemblages and textures in a rock. Understanding the interplay of metamorphism and deformation is important for 17 evaluating the classic assumption that matrix minerals likely represent those equilibrated at the ‘peak’ (maximum temperature) of metamorphism, for determining the crystallization sequence (i.e., pressure-temperature path) of minerals representing a reaction history, and for interpreting the pressure-temperature (P-T) conditions of deformation. These applications are all important for the interpretation of tectonic processes using assemblages and textures in deformed metamorphic rocks. However, although deformation can affect the stability and recrystallization history of matrix phases in a rock, it is difficult to interpret how phases and their composition would be affected and how their stability fields would be shifted during deformation. As a result, the role of deformation is seldom considered when utilizing traditional thermobarometric techniques. Experimental deformation of P-T sensitive metamorphic minerals such as the Al2SiO5 polymorphs is one way to evaluate the effects of strain on metamorphic processes and phase equilibria. The Al2SiO5 polymorphs (andalusite, kyanite and sillimanite) are common metamorphic minerals that comprise a simple chemical system with well-understood P-T phase relationships (Fig. 2.1) and that are commonly used to place limits on metamorphic P-T conditions. Furthermore, an important advantage of the Al2SiO5 polymorphs for reconstructing metamorphic paths and processes is the common presence of two, and more rarely, three polymorphs in a single assemblage (Kerrick, 1988; Cesare et al., 2002; Holdaway, 1978; Grambling, 1981; García-Casco and Torres-Roldán, 1996; Cavosie et al., 2002; Whitney, 2002). Coexisting polymorphs are typically thought to be a consequence of sluggish kinetics and are used to infer part of a rock’s P-T path if the 18 crystallization sequence of the polymorphs can be determined. Understanding the role of deformation in the preservation and transformation of the polymorphs during progressive metamorphism may influence our interpretation of the crystallization sequence. The role of deformation in metamorphic reactions involving the Al2SiO5 polymorphs, including polymorphic transformation, has been debated (Kerrick, 1986, 1990; Vernon, 1987; Penn et al., 1999). Field, experimental, and theoretical studies have proposed that strain energy related to high dislocation densities in minerals will affect phase equilibria, and that a strained crystal from a polymorphic system is more likely to transform (Helgeson et al., 1978; Grambling, 1981; Wenk, 1983; Doukhan et al., 1985; Vernon, 1987; Carey et al., 1992). In contrast, Kerrick (1986, 1988) found no correlation between apparent strain in naturally deformed kyanite and degree of transformation of kyanite to andalusite, and therefore concluded that strain energy has a negligible effect on the stability of the Al2SiO5 polymorphs. Nevertheless, experimental studies (Doukhan et al., 1985; this study) and examples from natural settings of shock metamorphism in Al2SiO5 polymorphs (Stahle et al., 2004) show that deformation can induce polymorphic transformation. An unresolved question is whether this is a common and significant process during metamorphism, either regionally or in high-strain zones. Therefore, although the polymorphs are not typically abundant enough in a rock to influence the bulk deformation behavior (i.e., rheology) of the rock, the significance of the polymorphs for determining metamorphic conditions and paths make them ideal for an investigation of the interaction of metamorphic reactions and deformation. An experimental approach can help resolve questions about the role of deformation in 19 metamorphic reactions because we can evaluate the effect of strain on a simple type of metamorphic reaction (polymorphic transformation) in a simple chemical system and examine reaction textures for partial to complete transformation. The purpose of this communication is to provide observations of microstructural development and phase transformations from deformation experiments on polycrystalline samples of the Al2SiO5 polymorphs. The results provide basic rheological data on the relative strengths of the three polymorphs to aid in interpreting reaction mechanisms of polymorphic transformation, deformation textures in natural rocks, and crystallization sequences in rocks containing coexisting polymorphs. AL2SIO5 POLYMORPHS: CRYSTAL CHEMISTRY, STRUCTURE, AND PHASE STABILITY The three Al2SiO5 polymorphs consist of chains of edge-shared Al-O octahedra parallel to the c-axis. Al atoms that link the chains vary in coordination in each polymorph: tetrahedral in sillimanite, 5-fold in andalusite, and octahedral in kyanite (Fig. 2.1). These structural differences are associated with the different stability ranges and physical properties of the polymorphs. Sillimanite and andalusite are both orthorhombic, but belong to different space groups (Sil: Pnnm; and: Pbnm), and kyanite is triclinic. The polymorphs are easily distinguished based on their physical and optical properties, and have well-determined crystal structures, crystal chemistry, thermodynamic properties, and phase relations (Holm and Kleppa, 1966; Holdaway, 1971; Vaughn and Weidner, 1978; Winter and Ghose, 1979; Ribbe, 1980; Salje, 1986; Kerrick, 1990; Pattison, 1992; 20 Yang et al., 1997a,b; Rao et al., 1999; Pattison, 2001; Friedrich et al., 2004; Burt et al., 2006; Ohuchi et al., 2006). Although the Al2SiO5 phase diagram has been extensively investigated by experimental studies, there have been few studies of the deformation behavior of the polymorphs. In previous experimental studies of the Al2SiO5 polymorphs, the sample materials have been gem-quality single crystals deformed in uniaxial compression (e.g., Raleigh, 1965; Boland et al., 1977; Menard et al., 1979; Doukhan and Christie, 1982; Doukhan and Paquet, 1982; Doukhan et al., 1985). In one such study, single crystals of the Al2SiO5 polymorphs were deformed by uniaxial compression, and the results examined with TEM (Doukhan and Christie, 1982; Doukhan and Paquet, 1982; Doukhan et al., 1985). An interesting result of the experiments was the creation of nm-scale ribbons of kyanite within sillimanite: evidence that polymorphic transformation was assisted by deformation. These studies also demonstrated the importance of [001] slip (parallel to c) in all 3 polymorphs, and showed that deformation induces dislocations, deformation bands, and micro-twinning (the latter in kyanite), similar to textures described from naturally deformed polymorphs (e.g., Lefebvre and Menard, 1981). EXPERIMENTAL METHODS SAMPLE PREPARATION The starting materials were synthesized from natural crystals. Crystals of each polymorphs were carefully crushed, hand-picked, and sieved to produce a grain size of ≤ 21 15 mm. Powders of each polymorph were placed in a vacuum oven at 125°C for at least 24 hours to remove any water acquired during processing. Dried samples were cold pressed to a uniaxial load of 150 MPa in a nickel jacket, and then hot pressed at 300 MPa and 1000°C for 4 hours. The hot-pressed samples of each polymorph were sliced with a diamond wafer saw into 11 mm diameter discs. EXPERIMENTAL CONDITIONS AND METHODS Deformation experiments were performed in a Paterson gas-medium apparatus under torsion at the University of Minnesota. Torsion was chosen as the experimental method for this investigation because of the inherent simple shear geometry of the method, and the ability to deform samples to much higher finite strains than axial geometry experiments. Torsion experiments have been successfully applied to a number of chemical systems and both single phase and more complex multi-phase assemblages (e.g. Pieri et al., 2001; Barnhoorn et al., 2005; Dimanov et al., 2007) Torsion also affords the ability to identify and investigate progressive strain behavior of a system during one sample run due to the geometry of the sample charge; total finite strain and strain rate increase from a theoretical value of zero in the center of the sample to the maximum value at the outer diameter of the sample charge (cf. Paterson and Olgaard, 2000). All experiments were run in the sillimanite stability field (Fig. 2.1) at temperatures of 1250°C and a confining pressure of 300 MPa with constant angular displacement. Sample temperature is measured within 3 mm of one end of the sample and 22 controlled to < 2 K of the set point for the duration of an experiment. Thermal gradients during the experiments were < 0.05 K/mm over the length of the sample determined from calibration runs performed prior to the experiments. Oxygen fugacity during both the hot- press stage and the deformation experiments was fixed on the Ni/NiO buffer by using Ni jackets. Sample charges were held between alumina pistons. The differential load/torque on the sample is measured with an internal load cell. Consequently, load measurements do not require corrections for friction on the loading column or elastic compliance of the apparatus for constant load tests, and the differential stress can be resolved to within 1 MPa. Calculation of the shear stress endured during the experiment is based on the torque measured during an experimental run and is dependent on the rheology of the material being deformed. For the experimental results in this study, we assumed power law creep behavior for all three polymorphs. Therefore, the maximum shear stress achieved at the outermost diameter of the experiment is related to torque by: € M = πd 3τ 4 3+ 1n       (1) where M is the measured torque, d is the sample diameter, τ is the shear stress and n is the stress exponent (Paterson and Olgaard, 2000). We assumed a stress exponent of 4, which corresponds to dislocation creep as the deformation mechanism active in all three polymorphs during deformation (discussed in Section 4.1). 23 The maximum shear strain rate € ˙ γ achieved during an experiment is at the edge of the sample, represented by the diameter d (assuming sample geometry and dimensions are constant throughout the experiment), and is proportional to the diameter of the sample charge: € ˙ γ = d ˙ θ 2L (2) where L is the length of the sample and θ is the constant angular displacement rate (Paterson and Olgaard, 2000). For the torsion experiments, this corresponds to a maximum strain rate of 2 x 10-4 s-1 for all experiments. MICROSTRUCTURAL CHARACTERIZATION Pre-deformed and deformed samples were analyzed using electron backscattered diffraction (EBSD) and scanning transmission electron microscopy (STEM) to determine textural and microstructural development during the experiments. Analyses took place at the Characterization Facility at the University of Minnesota. All samples were analyzed using both EBSD and STEM techniques on tangential sections (Fig. 2.2) cut as close to the outer edge of the sample as possible so as to contain the shear direction and investigate the region that endured the highest shear strain and shear stress during the experiment. In addition to the tangential section, the experimentally deformed andalusite sample was investigated using EBSD on an axial 24 section to examine the evolution of crystallographic textures and microstructures as a function of shear strain. EBSD ANALYSES Samples discs were hand polished to 0.5 µm using diamond lapping polishing film, then chemically polished with colloidal silica for 6 hours. EBSD analyses were carried out using a JEOL 6500 FEG SEM with an accelerating voltage of 20 KeV and a sample current 100 microamps. Diffraction patterns were acquired using an HKL Nordlys© detector and indexed using Oxford/HKL’s CHANNEL 5© software. Orientation maps and line scans were collected to determine the microstructure and lattice preferred orientations (LPO) developed during hot-pressing and during deformation for each sample. Orientation maps were collected at 0.7 to 1 µm steps at various scales and line scans were collected at 10 to 15 µm steps to ensure collection of one point per grain for an average of 200 grains per sample. Where orientation map data based pole figures are used in the following sections, all data were reduced to one point per grain to reduce the dominance of any larger grains producing an artificially strong LPO. Average mean angular deviation (MAD) between the acquired and calculated diffraction patterns for both orientation maps and line scans was < 1.2°. During analysis of andalusite and kyanite, the match unit for sillimanite was also included to look for mm-scale phase transformation. 25 An important criterion for describing and interpreting microstructural evolution and the prevalence of plastic deformation in an experiment is based on the strength of the LPO fabric observed in pole figures created from orientation data. Although qualitative estimates of LPO strength can be made by examining a pole figure, a quantitative measure is desirable when evaluating progressive strain behavior and comparing results between/within experiments (Bunge, 1982; Skemer et al., 2005). In this study, measurement of LPO strength was carried out using the M-index method of Skemer et al. (2005). The M-index was chosen because the method is less dependent on sample size than alternative measures of LPO strength (e.g. J-index; Bunge, 1982). Most of the data are from line scans and include ~ 200 orientations of the same number of grains. STEM IMAGING Both the pre-deformed (hot-pressed) samples and torsion experiments were analyzed to investigate phase transformation at finer scales than can be assessed by EBSD analysis. The tangential surfaces of the experiments and the pre-deformed material analyzed by EBSD were cut into thin sections, polished, ion-milled with ~3 keV Ar+ ions at low angle (4-15º) for STEM investigation. All specimens were carbon coated prior to the STEM analyses. The analyses were conducted on a FEI Tecnai G2 30 TEM/STEM equipped with a field emission gun. The microscope was operated at an accelerating voltage of 300 kV. Analyses were done primarily by dark-field imaging in STEM, with energy dispersive X-ray (EDX) spectrometry and some diffraction analysis done for phase identification. 26 RESULTS PRE-DEFORMED (HOT-PRESSED) SAMPLE CHARACTERIZATION The pre-deformed samples of all three polymorphs showed varying degrees of LPO that developed during aggregate preparation (Fig. 2.3). LPO development in the pre-deformed samples was likely attained during the cold-press stage of sample preparation. Because the cold-press is essentially a very weakly formed aggregate, no attempt was made to prepare the sample for EBSD analysis. The hot-press is unlikely to add to any inherent LPO attained during the cold-press and therefore the results described here are representative of the LPO attained during sample preparation. The LPO is not strong and shows no alignment that can be related to a specific slip system in any of the polymorphs. EBSD analyses show no evidence for phase transformation to sillimanite in either the andalusite or kyanite samples. It is important to note that the LPO that developed during sample preparation is distinct from that developed in the torsion experiments and did not affect the deformation results. The pre-deformed samples were also investigated using TEM analysis to determine whether any transformation occurred during hot-pressing. Again, neither andalusite nor kyanite samples exhibited any evidence for phase transformation or recrystallization, and both samples displayed angular grains consistent with textures produced during sample preparation. TEM imaging showed that hot-pressed materials 27 retained a significant porosity, indicating that the conditions of hot-pressing were not sufficient to induce plastic deformation necessary to close pore spaces in these materials. Kyanite grains commonly displayed fractures, kinking, and deformation twins, but preserved a consistent grain size of ~10-12 mm (Fig. 2.4 a-b). The andalusite sample showed large heterogeneities in grain size, from 2-12 mm in diameter (Fig. 2.4c). TORSION RESULTS Al2SiO5 triple stack In one experiment, discs of Al2SiO5 polymorphs were stacked in the order: andalusite (bottom), kyanite (middle), sillimanite (top) (Fig. 2.5). This experiment was designed to investigate the relative rheological characteristics of the three polymorphs, although results may be somewhat dependent on the stacking order of the polymorph sample discs. For example, the experimental results could be affected by which polymorph is adjacent to the alumina pistons. General features of the individual polymorph experiments, however, confirm the results of the triple stack experiment, so no additional triple stack experiments were attempted. Experimental conditions were 1250° C at 300 MPa and a constant shear strain rate of 2.4 ± 0.2 x 10-4 s-1 to 400% shear strain. The maximum shear stress attained during this experiment was 200 MPa; there was no evidence for brittle failure during the experiment. During the experiment, all three polymorphs developed a well defined LPO (Fig. 2.5) that is distinct from the LPO generated during hot-pressing (Fig. 1.3). The steel 28 jacket in which the experimental charge is placed acts as a strain marker, and it is clear from observations of the jacket that sillimanite took up much of the strain during the experiment (Fig. 2.6), indicating that sillimanite is the weakest of the polymorphs at the conditions of the experiment. The [001] LPO of all three polymorphs formed slightly oblique (~15°) to the shear plane (Fig. 2.5). The observed LPO patterns are in agreement with the known easy glide systems for the three polymorphs (Doukhan et al., 1985). No grain-scale phase transformation was observed by EBSD analyses either within each disc or at contacts between the discs. Andalusite An experiment in which a disc of andalusite was deformed under torsion (i.e., individually, not in a stack with other polymorphs) was carried out at 1250°C and a confining pressure of 300 MPa. Peak shear stress was 340 MPa and occurred at roughly 50% strain. The experiment was run to 400 % strain over 4 hours. There was no indication of brittle fracture within the sample or failure of the experimental assembly. EBSD and STEM analysis revealed a complex response of andalusite to strain. Unlike the andalusite in the triple stack experiment, the individual andalusite experiment contained extensive transformation of andalusite to sillimanite. The sillimanite was present as fibrolite concentrated in 500 mm thick shear bands and as mantles around relict andalusite throughout the entire sample. 29 The individual torsion experiment for andalusite resulted in a similar LPO as that generated in andalusite in the triple stack experiment (Fig. 2.7a). LPO development resulted in [001] orientations aligning slightly oblique (10-15°) to the shear plane and [010] orientations aligning roughly perpendicular to the shear plane (Fig. 2.7a), consistent with the known easy glide system for andalusite ((110)[100]). EBSD analysis of the andalusite axial (central) section shows similar LPO strength across the entire range of shear strains (Fig. 2.8). Orientation maps show no sub- grain development and weak LPO across the sample (Fig. 2.8). The axial section data presented here are located 0.3 mm from the true center axis of the sample. As a result, the data range from ~10% strain to the maximum strain reported for this experiment. The axial section nevertheless represents a large variation in percent strain, and it is clear that LPO development occurred even at relatively low amounts of strain. The LPO does not vary in strength or orientation from low to high strains recorded across the axial section (Fig. 2.8). As in the tangential (maximum strain) section, extensive transformation of andalusite to sillimanite was observed in the axial section. Fibrolite shear bands occur across the entire axial section (Fig. 2.10a), and SEM imaging shows fibrolite development at andalusite grain boundaries, across the entire axial sample, typically completely surrounding andalusite grains (Fig 2.11). These observations suggest that although the deformation and reaction mechanisms active during the experiment may have been heterogeneous, the effects of these mechanisms were distributed 30 homogeneously across the minimum to maximum strain field preserved in the axial section. Textures related to transformation of andalusite to fibrous sillimanite can be seen clearly in STEM images of the sillimanite-rich bands and throughout the sample. These shear bands were oriented antithetic to the shear direction at ~ 15-20° oblique to the shear plane (Fig. 2.12a). Sillimanite in the andalusite-rich regions and in the sillimanite bands is well aligned, and some crystals are bent (Fig. 2.12b). Fibrolite occurs as extremely small crystals (diameters of 150-250 nm), and therefore are not detected by EBSD analysis. Sillimanite outside the shear bands forms mantles around andalusite crystals. Dislocations are apparent in most andalusite grains (Fig. 2.12c). Many of the relict andalusite crystals display a slight asymmetry that preserves the shear direction (Fig. 2.12d). A quantitative determination of the extent to which andalusite transformed to sillimanite in the bulk sample is difficult given the extremes in grain size observed, but is on the order of 50-70% for the andalusite experiment. Kyanite The kyanite experiment was conducted at the same conditions and at the same strain rate as the sillimanite and andalusite individual experiments. The peak stress achieved during the experiment was 320 MPa, which occurred at 230% strain. Maximum strain reached was 400 % strain over 4.5 hours. As with the andalusite sample, there were no indications of assembly or sample failure during the experiment. 31 Almost all the kyanite transformed to sillimanite in this experiment. Attempts at acquiring EBSD data from the sample failed owing to the lack of relict kyanite of sufficient size to analyze by this method. STEM imaging revealed sillimanite shear bands, similar to those in the andalusite experiment, and a much higher degree of transformation to sillimanite outside the shear bands. The identity of phases (kyanite vs. sillimanite) was confirmed by diffraction analysis in the TEM. The shear bands in the kyanite experiment formed at nearly the same orientation and with similar antithetic orientation relationship to shear direction observed in the andalusite experiment. Bands with isoclinal folds formed oblique to the shear plane and contain fibrous sillimanite (Fig. 2.13a). The shear bands contain ~ 98% sub-mm sillimanite (fibrolite) and only minor relict kyanite (Fig. 2.13b). Regions outside the shear bands contain more relict kyanite, but are dominated by sillimanite. Kyanite outside the bands has a grain size fraction of ~100-500 nm (reduced from the original grain size of ~10-12 mm). Relict kyanite in both the shear bands and matrix contains deformation- induced twins (Fig. 2.13c), but does not exhibit as many observable dislocations as the experimentally deformed andalusite. In some grains, kink bands are observed to crosscut twin lamellae. Sillimanite The sillimanite experiment was run under a confining pressure of 300 MPa at 1250°C for 4 hours. Peak shear stress was calculated at 220 MPa and occurred at 20% shear strain. 32 Maximum shear strain reached during the experiment was 140%. The maximum shear stress for the sillimanite experiment was by far the lowest shear-stress achieved of all the polymorphs and is identical to that of the triple stack experiment, confirming sillimanite as the weakest of the polymorphs and accounting for its role in controlling the rheological behavior of the triple stack experiment. There was no indication of failure in the sample or the assembly, but some slip along the piston-sample interface may have occurred, preventing the experiment from achieving higher strain. The sillimanite experiment was complicated by the presence of ~12 modal% quartz (impurities not removed during sample preparation); no quartz was present in the sillimanite used in the triple stack experiment. In the individual experiment on sillimanite, quartz was distributed heterogeneously throughout the sample, chiefly as single grains or, more rarely, as clusters of two or three grains. Results of EBSD analysis indicate that quartz had little effect on the overall deformational behavior of sillimanite. EBSD analysis of the individual sillimanite experiment yields similar results to sillimanite deformed in the triple stack experiment. LPO development resulted in [001] orientations aligned slightly oblique (10-15°) to the shear plane, and [010] orientations aligned approximately perpendicular to the shear plane (Fig. 2.7), consistent with slip along the easy glide system (010)[001]. DISCUSSION 33 EVALUATION OF EXPERIMENTAL STRESS-STRAIN RELATIONSHIPS Flow law assumptions and stress-strain behavior Calculation of shear stress endured by sample charges during experiment in a Paterson apparatus are calculated based on an assumption of a flow law for the phase(s) under investigation (eq. 1). The assumption that the Al2SiO5 polymorphs deform via a dislocation process is based on the observation that most silicates at similar experimental conditions behave via a power-law dislocation creep mechanism. However, interpretation of the general rheological response of a material to strain is not affected by the magnitude of the stress exponent chosen. For example, varying n from 3 to 5 (dislocation creep) will affect the calculated shear stress by < 5% and thus not have a great effect on general interpretation of the data (Paterson and Olgaard, 2000). In addition, although varying the stress exponent across all reasonable values changes the calculated shear stress, it does not have any impact of the overall shape of the stress-strain curve (Fig. 2.14a) and thus the first-order stress-strain behavior of the sample is not corrupted. Without having knowledge of the Al2SiO5 polymorphs flow-laws, we chose the median stress exponent value of 4 with the knowledge that the stress-strain relationships are an estimate until the stress-exponent can be determined. The goals of the experiments presented here were to address basic rheological properties of the Al2SiO5 system, establish parameters for future experiments and determine the importance of deformation-assisted phase- transformation on the polymorphs, and therefore detailed knowledge of the flow laws of the Al2SiO5 phases is not crucial to interpretations of the data. The effective stress experienced by the andalusite and kyanite individual torsion experiments was ~ 600 MPa (roughly twice the shear stress). This value is also double 34 the confining pressure of the experiment. Traditionally, such behavior would be interpreted as evidence the material being deformed was accommodating strain via a brittle deformation mechanism. However, it has been shown that microcracking is controlled by grain size and thermal expansion anisotropy; smaller grain sizes diminish the likelihood of fracture of a material during plastic deformation (Evans, 1978; Cooper, 1990). The resultant stress-strain curves derived from experiments would also be affected by brittle behavior, and this would be detected had it occurred. Mirocracking or discrete slip would result in rapid drop in shear stress over infinitesimal increase in strain, yielding a step function behavior. The stress-strain curves for the Al2SiO5 experiments do not reach a steady state condition, but also do not exhibit strong evidence for rapid drops in shear stress that would be interpreted as a brittle response to strain (Fig. 2.14b). However, the kyanite individual torsion experiment does show some small step-like drops at higher strains (> 250%), although this is more likely related to slip along sillimanite shear bands developed during the experiment (Fig. 2.14b). We cannot discount the possibility that some brittle deformation occurred during the experiments, but, if there had been brittle deformation, evidence of microcracking or larger scale fractures has been overprinted by phase transformation of kyanite and andalusite to fibrolitic sillimanite. The formation of extremely fine grained (~500 nm diameter) fibrolite in shear bands and mantles around relict grains introduces the possibility of grain boundary sliding or other grain-size sensitive creep processes being active in the andalusite and kyanite experiments (sec. 4.3). 35 THE EFFECT OF DUCTILE DEFORMATION ON SILLIMANITE Owing to the experimental limitations of the Paterson devices, we cannot ‘reverse’ the deformation experiments and examine transformation of sillimanite to andalusite or kyanite (cf. Snow and Yund, 1987, an experimental study on the effect of deformation on the displacive aragonite-calcite polymorphic transformation). The P-T conditions of the experiments conducted with the Paterson apparatus are within the stability field of sillimanite, and therefore deformation of sillimanite at these conditions tests the role of deformation on recrystallization of sillimanite and determines the mechanical behavior of sillimanite relative to the other polymorphs. The triple stack and individual torsion experiments clearly show that sillimanite is the weakest of the three polymorphs. The starting material for the sillimanite experiments was prismatic sillimanite. Therefore, although the experiments were conducted on very fine-grained aggregates of sillimanite more similar in size (though not aspect ratio) to fibrous sillimanite, the results are likely applicable to deformation of prismatic sillimanite. Our results are consistent with activation of the (010)(001) slip system, consistent with studies of naturally deformed sillimanite (e.g., Lembregts and van Roermund, 1990). The sillimanite produced via phase transformation during the experimental deformation of kyanite and andalusite is fibrous and, in contrast to the relict andalusite and kyanite in the experiments, does not show evidence for defects at the scale of our observations. Most fibrous sillimanite grains are straight, with a very strong crystallographic and shape preferred orientation subparallel to the shear plane, although slightly bent/curved grains are present (Fig. 2.11b). 36 DEFORMATION MECHANISMS, MICROSTRUCTURAL DEVELOPMENT AND PHASE TRANSFORMATION: INTERACTION OF METAMORPHISM AND DEFORMATION Estimation of the deformation/reaction mechanisms involved in the phase transformation process(es) can be made based on the nature of the stress-strain curves themselves, as well as EBSD and STEM data. Transformation of andalusite and kyanite to fibrolitic sillimanite during the torsion experiments makes direct interpretation of the stress-strain curves (Fig. 2.14b) derived from experiments difficult to address directly, as we cannot currently comment on the timing of transformation. Furthermore, neither material (andalusite, kyanite) reached steady state behavior during experiments, and the resultant stress-strain curves display a complex response to strain (Fig. 2.14b). The curves for both kyanite and andalusite show systematic weakening with increasing strain after a maximum shear stress was achieved. Given that both kyanite and andalusite deform via a single dominant c-axis slip mechanism, the lack of steady state behavior suggests that dislocation creep did not develop as the dominant creep mechanism during the entirety of the experiment. The development of a LPO in the kyanite and andalusite experiments in conjunction with the differences between microstructures and LPO observed in the pre-deformed hot-pressed materials and their deformed counterparts (Figs. 1.3-1.6, 1.7) provides evidence that dislocation creep was active during the experiments. However, M-index values for all experiments (Fig. 2.8) illustrate that the LPO’s developed are much weaker than would 37 be predicted for a material deformed to high strains. This suggests the unique deformation/phase-transformation (reaction) relationship in these experiments provided a means to produce evolving deformation pathways as strain was accommodated. The effect of strain on the stability and resultant phase-transformation kinetics of the Al2SiO5 polymorphs has been studied previously on a largely theoretical basis. Polymorphic transformation of Al2SiO5 phases involves breaking bonds and change in the mineral structure related to the coordination of one of the Al sites. Because these transformations require a high activation energy, it is common in metamorphic rocks for a polymorphs to persist beyond its stability field; e.g., kyanite in the sillimanite field. For this reason, and because kyanite has a very different crystal structure relative to andalusite and sillimanite (which are structurally more similar to each other), multiple polymorphs may coexist in a rock. It is therefore of interest to explore whether deformation is important in eliminating kyanite and andalusite during progressive metamorphism. Although the deformation/reaction relationships observed in the experiments are complex, it is possible to make some inferences about Al2SiO5 polymorph deformation behavior and the role of deformation in polymorphic reaction (phase transformation), and to discuss the results in light of previous work on Al2SiO5 deformation. Based on theoretical and TEM investigations of dislocation densities in natural and experimental Al2SiO5 polymorphs, Kerrick (1986) concluded that strain energy did not greatly affect the stability fields of the polymorphs; i.e., the increased internal strain energies of polymorphs with dislocations are not significant enough to affect the 38 conditions at which phase equilibria predicts polymorphic transformation will occur. However, dislocation densities are difficult to quantify and can be destroyed during recrystallization, and thus using the currently observed dislocation density as a measure of the dislocation density at the time of reaction may lead to under/over estimation of the internal strain energy in a phase. The observations of microstructures and textures in the experiments presented here clearly show that phase transformation process can be deformation-induced. The role of deformation in phase transformation is further illustrated by microstructures in the andalusite axial section. Shear band development and fibrolitic mantles are observed across the entire axial section (Fig. 2.10), suggesting that phase transformation of andalusite to sillimanite occurred even at very small shear strains under the experimental P-T conditions. Shear band development also led to significant shear weakening during progressive deformation of the kyanite and andalusite experiments. The product fibrolite was extremely fine-grained and there is no evidence of syn- deformational grain growth during the experiment. The very small fibrolite grain size and the high modal abundance of deformation-induced fibrolite are consistent with grain-size sensitive flow as the dominant deformation mechanism responsible for the complex stress-strain relationships recorded during the kyanite and andalusite experiments. The shear bands form complex geometries in both the tangential and axial sections of the experiments. Tangential sections show antithetic oriented shear bands subparallel with the shear plane. These results are similar to shear band development in other experimental systems using the Paterson apparatus (Delle Piane, 2007; Holtzman et al., 39 2003). The structures in the axial section show remarkable three-dimensionality, with shear bands, in some cases discontinuous, subparallel to but anastomosing with respect to the shear plane (Fig. 2.10a). This is in contrast to observations in olivine-pyroxene-melt systems where shear bands cut at angles to the shear plane in the tangential section, but are parallel to the shear plane in the axial section (Fig. 2.10b) and did not propagate into the lower strain regions of the experiment (e.g. King et al., 2007). The significant amount of temperature overstepping of the stability fields of andalusite and kyanite and the application of shear strain at the outer edge of the sample led to a positive feedback loop of deformation and reaction producing fine grained fibrolite. This transformation likely occurred early in the experimental run. The mantles of fibrolite around andalusite (Figs. 2.12d, 2.11) bear a striking resemblance to dynamic recrystallization textures observed in naturally and experimentally deformed rocks (e.g. Tullis, 2002; De Ronde and Stünitz, 2007). We prefer a model whereby andalusite and kyanite grains experienced phase transformation aided by a dynamic recrystallization related process that also produced the weak asymmetric shape fabric. The transformation likely started at the outer most edges of the sample and, once sillimanite grains were present, the reaction propagated to the interior portions of the kyanite and andalusite sample charges. Further transformation led to more interconnectivity of fibrolite and, finally, shear band development. The interconnectivity of the fibrolite reaction products also led to a change from plastic deformation to grain size sensitive flow as the dominant deformation mechanism. The shear weakening stage of the experiments is likely related to complete penetration of shear bands through the experiment and compromised sample 40 charge integrity. Between shear bands, reaction may have been enhanced along grain boundaries of crystals (Fig. 2.14) that had been viscously deformed via a strain-activated, diffusion-controlled process (cf. De Ronde and Stünitz, 2007). High strain behavior of reactive systems display complicated microstructural and reaction behavior. This paper along with recent contributions by Delle Piane et al. (2007) and Barnhoorn et al., (2005) show the importance of fine-grained reaction product formation and stabilization at high strains. These observations suggest that grain size sensitive flow is a critical mechanism in the accommodation of large shear strains in reactive/deforming systems. Reaction products in these experiments are observed as fine-grained a weak and/or strong phase(s), which develop into shear bands in the sample and localize strain accommodation. SUMMARY Results of the present study show that deformation triggers significant transformation of andalusite and kyanite to sillimanite at laboratory conditions at which no transformation occurs in the absence of deformation at the time scale of the experiments reported here. Aggregates of each of the Al2SiO5 polymorphs were deformed together and independently to strains exceeding γ = 4 in the stability field of sillimanite. All three polymorphs experienced plastic deformation, as shown by LPO formation sub- parallel with the shear plane. The stress-strain relationships of the polymorphs show that sillimanite is the weakest of the three phases. Kyanite and andalusite individual experiments showed extensive deformation-induced transformation to sillimanite and significant shear localization along sets of shear bands. 41 Microstructures and textures in the experiments imply a multistage deformation and reaction history. Plastic deformation was likely active early in the deformation history of the kyanite and andalusite experiments. This stage was quickly followed by phase transformation to fibrous sillimanite, and led to dominance of grain size sensitive flow accommodating the high strains. Deformation enhanced transformation of andalusite to sillimanite and kyanite to sillimanite during experiments in the stability field of sillimanite, most likely by increasing dislocation density near grain boundaries, and therefore nucleation sites. In the case of the kyanite experiments, and, to a much lesser extent the andalusite experiment, grain size reduction and other microstructural effects (e.g., micro-twinning, kinking) may also have been important in affecting nucleation and growth of sillimanite. Deformation enhanced transformation of andalusite and kyanite to sillimanite, and may occur in nature in high strain zones. However, it is clear that the microstructural and mechanical observations in the Al2SiO5 system are consistent with other deformation experiments conducted in the presence of reaction. This suggests that the Al2SiO5 system represents a simple chemical analog to multi-phase deformation/reaction systems and are applicable to interpreting the general microstructural and strain localization phenomena that occur in these systems 42 REFERENCES Barnhoorn A., Bystricky M., Kunze K. Burlini L., Burg, J.P. 2005. Strain localisation in bimineralic rocks: Experimental deformation of synthetic calcite-anhydrite aggregates. Earth Planet Sc Lett 240, 748-763. Baxter, E.F. & De Paolo, D.J., 2004. Can metamorphic reactions proceed faster than bulk strain? Contrib. to Min. Pet., 146, 657-670. Boland, J.N., B.E. Hobbs, and A.C. McLaren, 1977. The defect structure in natural and experimentally deformed kyanite, Phys. Status Solidi A, 39, 631-641. Brodie, K.H. and Rutter, E.H., 1985 On the relationship between deformation and metamorphism, with special reference to the behaviour of basic rocks. In: A.B. Thompson and D.C. Rubie, Editors, Metamorphic Reactions: Kinetics, Textures and Deformation, Springer-Verlag, New York, 138–179. Bunge, H., 1982. Texture Analysis in Materials Science: Mathematical Models. Butterworths, London. 593 pp Burt, J.B., N.L. Ross, R.J. Angel, and M. Koch, 2006. Equations of state and structures of andalusite to 9.8 GPa and sillimanite to 8.5 GPa, Am. Mineral., 91, 319-326. Carey, W.J., Rice, J.M. & Grover, T.W. 1992. Petrology of aluminous schist in the Boehls Butte region of northern Idaho: geologic history and aluminum-silicate phase relations. Am. J. of Science, 292, 455-473. 43 Cavosie, A., Z.D. Sharp, and J. Selverstone, 2002. Co-existing aluminum silicates in quartz veins: A quantitative approach for determining andalusite-sillimanite equilibrium in natural samples using oxygen isotopes, Am. Mineral., 84, 417-423. Cesare, B., M.T. Gomez-Pugnaire, A. Sanchez-Navas, and B. Groberty, 2002. Andalusite-sillimanite replacement (Mazarron, SE Spain): A microstructural and TEM study, Am. Mineral., 87, 433-444. Cooper, R.F., 1990. Differential stress-induced melt migration-an experimental approach. J. Geophysics Research B. 95, 6979-6992. Delle Piane C, Burlini, L., Grobety, B., 2007. Reaction-induced strain localization: Torsion experiments on dolomite. Earth and Planetary Sc. Lett. 256, 36-46. de Ronde, A.A., Stunitz, H., 2007. Deformation-enhanced reaction in experimentally deformed plagioclase-olivine aggregates Cont. Min. Pet., 153, 699-717. Dimanov, A., Rybacki, E., Wirth, R., Dressen, G., 2007. Creep and strain-dependent microstructures of synthetic anorthite-diopside aggregates. J Struct Geol 29, 1049-1069 Doukhan, J.-C., and J.M. Christie, 1982. Plastic deformation of sillimanite Al2SiO5 single crystals under confining pressure and TEM investigation of the induced defect structure, Bull. Mineral., 105, 583-589. Doukhan, J.-C., N. Doukhan, P.S. Koch, and J.M. Christie, 1985. Transmission electron microscopy investigation of lattice defects in Al2SiO5 polymorphs and plasticity induced polymorphic transformations, Bull. Mineral, 108, 81-96. 44 Doukhan, J.-C., and J. Paquet, Plastic deformation of andalusite single crystal Al2SiO5, Bull. Mineral, 105, 170-175, 1982. Evans, A.G., 1978. Microfracture from thermal-expansion anisotropy 1 Single phase systems. Acta Metall., 26, 1845-1853. Flinn, D., 1965. Deformation in metamorphism. In: Pitcher, W.S. and Flinn, G.W., (Eds.) Controls of Metamorphism, Oliver and Boyd, Edinburgh, 46–72. Friedrich, A., M. Kunz, B. Winkler, and T. Le Bihan, 2004. High-pressure behavior of sillimanite and kyanite: compressibility, decomposition and indications of a new high-pressure phase, Z. Kristall., 219, 324-329. García-Casco, A., and R.L. Torres-Roldán, 1996. Disequilibrium induced by fast decompression in St-Bt-Grt-Ky-Sil-And metapelites from the Betic belt (Southern Spain), J. Petrol., 37, 1207-1239. Grambling, J.A., 1981. Kyanite, andalusite, sillimanite, and related mineral assemblages in the Truchas Peaks region, New Mexico, Am. Mineral., 66, 702-722. Greenwood, H.J., 1972. AlIV-SiIV disorder in sillimanite and its effect on phase relations of the aluminum silicate minerals, Geol. Soc. Am. Mem., 132, 553-571. Helgeson, H.C. Delany, J.M., Nesbitt, H.W. & Bird, D.K. 1978. Summary and critique of the thermodynamic properties of rock forming minerals. Am. J. of Science, 274, 1089-1198. Holdaway, M.J., 1971. Stability of andalusite and the aluminum silicate phase diagram. Am. J. Sci., 271, 97-131. 45 Holdaway, M.J., 1978. Significance of chloritoid-bearing and staurolite-bearing rocks in the Picuris Range, New Mexico, Geol. Soc. Am. Bull., 89, 1404-1414. Holm, J.L., and O.J. Kleppa, 1966. The thermodynamic properties of the aluminum silicates. Am. Mineral., 51, 1608-1627. Holtzman, B.K., Groebner, N.J., Zimmerman, M.E., Ginsberg, S.B., Kohlstedt, D.L., 2003. Stress-driven melt segregation in partially molten rocks. Geochem Geophy Geosy. 4, Holyoke, C.W. & Tullis, J., 2006. Formation and maintenance of shear zones. Geology, 34, 105-108. Kerrick, D.M., 1986. Dislocation strain energy in the Al2SiO5 polymorphs, Phys. Chem. Mineral., 13, 221-226. Kerrick, D.M., 1988. Al2SiO5–bearing segregations in the Lepontine Alps, Switzerland: Aluminum mobility in metapelites, Geology, 16, 636-640. Kerrick, D.M., 1990. The Al2SiO5 polymorphs, Rev. Mineral., Mineralogical Society of America, Washington, D.C. King, D.S., Kohlstedt, D.L., Zimmerman, M.E., 2007. Stress-Driven Melt Segregation and Shear Localization in Partially Molten Aggregates: Experiments in Torsion. Eos Trans, AGU, 88, Fall Meet. Suppl., Abstract T43D-07 Lambregts, P.J., and H.L.M. van Roermund, 1990. Deformation and recrystallization mechanisms in naturally deformed sillimanites. Tectonophysics, 179, 371-378. Lefebvre, A. and D. Menard, 1981. Stacking faults and twins in kyanite, Al2SiO5, Acta Crystall., A37, 80-84. 46 Menard, D., J.-C. Doukhan, and J. Paquet, 1979. Uniaxial compression of kyanite Al2O3- SiO2: deformation mechanisms of minerals and rocks, Bull. Mineral., 102, 159-162. Ohuchi, F.S., Ghose, S., Engelhard, M.H., and Baer, D.R., 2006. Chemical bonding and electronic structures of the Al2SiO5 polymorphs, andalusite, sillimanite, and kyanite: X-ray photoelectron and electron energy loss spectroscopy studies. Am. Min., 91, 740-746. Paterson, M.S., and D.L. Olgaard, 2000. Rock deformation tests to large shear strains in torsion, J. Struct. Geol., 22, 1341-1358. Pattison, D.R.M., 1992. Stability of andalusite and sillimanite and the Al2SiO5 triple point: Constraints from the Ballachulish Aureole, Scotland, J. Geol., 100, 423-446. Pattison, D.R.M., 2001. Instability of the Al2SiO5 “triple-point” assemblage in muscovite + biotite + quartz-bearing metapelites, with implications, Am. Mineral., 86, 1414- 1422. Penn, R.L., J.F. Banfield, and D.M. Kerrick, 1999. TEM investigation of Lewiston, Idaho, fibrolite: Microstructure and grain boundary energetics, Am. Mineral., 84, 152-159. Pieri M, Burlini, L., Kunze, K., Stretton, I., 2001. Rheological and microstructural evolution of Carrara marble with high shear strain: results from high temperature torsion experiments. Journ. of Struct. Geol. 23, 1393-1413. Raleigh, C.B., 1965. Glide mechanisms of experimentally deformed minerals, Science, 150, 739-741. 47 Rao, M.N., S.L. Chaplot, N. Choudhury, K.R. Rao, R.T. Azuah, W.T. Montfrooij, and S.M. Bennington, 1999. Lattice dynamics and inelastic scattering from sillimanite and kyanite Al2SiO5, Phys. Rev. B, 60, 12061-12068. Ribbe, P.H., 1980. Aluminum silicate polymorphs and other aluminum silicates. In: Ribbe, P.H. (ed.), Orthosilicates, Rev. Mineral., 5, 189-214. Salje, E., 1986. Heat capacities and entropies of andalusite and sillimanite: the influence of fibrolitization on the phase diagram of the Al2SiO5 polymorphs, Am. Mineral., 71, 1366-1371. Skemer, P., Katayama, I., Jiang, Z., Karato, S., 2005. The misorientation index: Development of a new method for calculating the strength of lattice- preferred orientation. Tectonophysics. 41, 157-167. Snow, E. and R.A. Yund, 1987. The effect of ductile deformation on the kinetics and mechanisms of the aragonite-calcite transformation, J. Metamorph. Geol., 5, 141-153. Stahle, V., R. Altherr, M. Koch, and L. Nasdala, 2004. Shock-induced formation of kyanite (Al2SiO5) from sillimanite within a dense metamorphic rocks from the Ries crater (Germany), Contrib. Mineral. Petrol., 148, 150-159. Tullis, J, 2002. Deformation of granitic rocks: Experimental studies and natural examples. In: Plastic Deformation of Minerals and Rocks. Karato, S, Wenk, H (eds). Rev. Mineral. & Geochem. 51, 51-95. Vaughn, M.T., and D.J. Weidner, 1978. The relationship of elasticity and crystal structure in andalusite and sillimanite, Phys. Chem. Mineral., 3, 133-144. 48 Vernon, R.H., 1987. Oriented growth of sillimanite in andalusite, Placitas-Juan Tabo area, New Mexico, U.S.A., Can. J. Earth Sci., 24, 580-590. Wenk, H.R. 1983. Mullite-sillimanite intergrowth from pelitic inclusions in Bergell tonalite. Neues Jahrbuch fur Mineralogie Abhandlungen, 146, 1-14. Whitney, D.L., 2002. Coexisting andalusite, kyanite, and sillimanite: sequential formation of three polymorphs during progressive metamorphism near the Al2SiO5 triple point, Sivrihisar, Turkey, Am. Mineral., 84, 405-416. Williams, M.L., 1994. Sigmoidal inclusion trails, punctuated fabric development, and interactions between metamorphism and deformation. J. Metamorphic Geology, 2, 1- 21. Winter, J.K., and S. Ghose, 1979. Thermal expansion and high temperature crystal chemistry of the Al2SiO5 polymorphs. Am. Mineral., 64, 573-586. Wintsch, R.P., Christofferson, R. and Kronenberg, A.K., 1995. Fluid-rock reaction weakening of fault zones., J. Geophys. Research B, 100, 13021-13032. Yang, H.X., R.T. Downs, L.W. Finger, R.M. Hazen, and C.T. Prewitt, 1997a. Compressibility and crystal structure of kyanite, Al2SiO5, at high pressure, Am. Mineral., 82, 467-474. Yang, H., R.M. Hazen, L.W. Finger, C.T. Prewitt, and R.T. Downs, 1997b. Compressibility and crystal structure of sillimanite, Al2SiO5, at high pressure, Phys. Chem. Mineral., 25, 39-47. Yund, R.A., and Tullis J., 1991. Compositional changes of minerals associated with dynamic recrystallization. Cont. Min. Pet., 108, 346-355. 49 Figure Captions Figure 2.1. Phase diagram for the Al2SiO5 system (Holdaway, 1971). Each field is labeled with the polymorph, its density (in g/cm3), Al coordination, and crystal system. Figure 2.2. Schematic of sample charge displaying different sections cut for analysis. Maximum strain experienced by a sample deformed in torsion increases with distance from the center of the experiment. The longitudinal tangential section is used to investigate the microstructural development and deformational characteristics at the highest strains attained during the experiment. The longitudinal axial section is advantageous because it represents the entire strain gradient from a theoretical value of zero at the center of the section to the max strain endured at the outermost diameter of the sample. Figure 2.3. Pole figures for kyanite, andalusite and sillimanite from the pre-deformed (hot-press) state. The pole figures are equal area projections (lower hemisphere) with Gaussian point smoothing of 10° full-width half maximum. Contour color densities are drawn based on multiples of uniform density (mud); the scale of intensity for all pole figures are logarithmic with low intensity of point clustering represented by blue and high intensities by red. 50 Figure 2.4. STEM dark-field images of pre-deformed kyanite and andalusite. There was no evidence in either sample of phase-transformation to sillimanite during hot-pressing. A) Image illustrating the representative grain size of kyanite. Grains are angular and highly fractured at grain boundaries illustrating the lack of plastic deformation or recrystallization during the hot-press stage of sample preparation. Dark area in the upper left of the image is the hole made during ion-milling and does not represent inherent porosity of the sample. B) Representative microstructure and grain size of andalusite prior to deformation. As with the kyanite sample, andalusite is highly angular and fractured with little evidence of plastic-deformation or recrystallization. C) Deformation twins and kink bands in kyanite developed during hot-pressing. Figure 2.5. The assembly of the Al2SiO5 triple stack experiment. The iron jacket inclosing the sample charge also acts as a strain gauge. The jacket displays shear localization occurring at the location of the sillimanite portion of the experiment suggesting sillimanite may be the weakest of the polymorphs. Figure 2.6. Pole-figures of kyanite, andalusite and sillimanite from the triple stack experiment perpendicular to the shear plane and parallel to shear direction. All experiments result in dextral shear. All three phases display strong LPO development sub-parallel with the shear-plane (horizontal line). The LPO’s developed in all three phases are also distinct from LPO developed during hot-pressing (Fig. 2.3) illustrating 51 that plastic deformation was active in all three phases during the experiment. LPO’s for all phases show the importance of c-parallel slip in the Al2SiO5 system. Figure 2.7. Pole-figures from the andalusite and sillimanite individual torsion experiments. Pole-figures are again perpendicular to the shear plane and parallel to the shear direction. [001] maxima are in a similar orientation as developed during the triple stack experiment. Orientations are consistent with (010)[001] glide. Figure 2.8. M-index calculations after Skemer et al., (2005) for the predeformed, andalusite axial, andalusite tangential and sillimanite tangential EBSD analyses. The data show fairly consistent fabric strengths in all experiments, but the LPO orientations are different between the pre-deformed and deformed state of the experiments. The axial section displays some evidence of LPO strength increasing from the center to the edge of the experiment. However the LPO orientations are consistent across the sample. This suggests that plastic deformation was active for a time across the entirety of the sample and thus across a wide range of finite strains. Figure 2.9. a) Reflected light photomicrograph of the axial section from the andalusite individual experiment. The grey lines on the image show the trends of the fibrolite shear bands across the sample. Shear band development occurs across the entire sample, but are highly concentrated near the edges of the sample. The axial section is perpendicular to the shear plane and shear direction. b) Reflected light image of the axial section of an 52 olivine+pyroxene+MORB torsion experiment. The shear bands are dominated by melt and form in similar orientations as observed in (a) but do not extend across the entirety of the axial section. Figure 2.10. SEI images of fibrolite mantles around relict andalusite. Transformation of andalusite to sillimanite occurred and is equally developed across the entire strain gradient recorded in the axial section. Figure 2.11. a) Photomicrograph (crossed polars) of the longitudinal tangential section from the andalusite experiment. Yellow line represents the shear plane. Shear bands are the high birefringence bands cutting sub-parallel and oriented antithetic to the shear direction. b) STEM dark field image of bent fibrolite present in the shear bands. c) Dislocations in relict andalusite. d) STEM dark-field image illustrating the slight asymmetry developed in some relict andalusite grains consistent with shear direction. Note the sillimanite mantles developed around the relict andalusite grains. Figure 2.12. a) Photomicrograph (crossed polars) of the longitudinal tangential section from the kyanite experiment. Yellow line represents the shear plane. As with the andalusite experiment, shear bands are the high birefringence bands cutting sub-parallel and oriented antithetic to the shear direction. b) STEM dark-field image showing the representative relationships between relict kyanite and sillimanite transformation. c) STEM dark-field image of kinked deformation twins in kyanite. 53 Figure 2.13. a) Effect of choice of stress-exponent on calculated shear stress. Max shear stress is affected by choice of stress-exponent, but does not change the shape or affect the interpretation of rheological response of the material. b) shear-stress-strain data for all experiments. The data are smoothly curved indicating little evidence for brittle behavior during the experiment. Some grain boundary sliding or shear band reorganization may be associated with the drops in shear stress during the kyanite experiment. 0.2 0.4 0.6 0.8 200 400 600 800 10000 KYANITE ANDALUSITE SILLIMANITE Pr es su re ( G Pa ) Temperature (°C) 0 Experiments Al[6]Al[5]SiO5 Al[6]Al[6]SiO5 Al[6]Al[4]SiO5 orthorhombic orthorhombictriclinic ρ = 3.61 ρ = 3.24 Figure 2.1. ρ = 3.15 54 γ=0 longitudinal tangential section 0 γmax longitudinal axial section Figure 2.2. a n 55 206 grains Max mud = 6.26 Half-width 10 Kyanite pre-deformed 175 grains Max mud = 5.55 Half-width 10 Andalusite pre-deformed Sillimanite pre-deformed 392 grains Max mud = 3.66 Half-width 10 a b c Figure 2.3 [100] [010] [001] [100] [010] [001] [100] [010] [001] [100] [010] [001] upper lower upper lower upper lower 56 a b c 500nm 200 nm 100nm Figure 2.4. 57 Andalusite Kyanite Sillimanite Figure 2.5 58 Kyanite Andalusite Sillimanite 2198 points Half-Width=10 Max mud=7.71 883 points Half-Width=10 Max mud=5.47 200 grains Half-Width=10 Max mud= 8.32 Figure 2.6. [100] [010] [001] [100] [010] [001] [100] [010] [001] [100] [010] [001] upper lower upper lower upper lower 59 1218 pts Half-Width 10 Max mud 6.3 679 pts Half-Width 10 Max mud 11.5 Andalusite Sillimanite a b Figure 2.7. [100] [010] [001] [100] [010] [001] 60 center axis of sample a d Figure 2.8. cb Experiment Sample Orientation M-Index Hot-Press (pre-deformed) andalusite N/A 0.17 sillimanite N/A 0.16 kyanite N/A 0.22 Individual Experiments andalusite axial a 0.17 " " b 0.15 " " c 0.19 " " d 0.16 " tangential 0.18 Sillaminite tangential 0.18 [100] [010] [001] a b c d 200 grains m.u.d. 61 Figure 2.9. b Melt dominated shear bands a Center Axis Shear Plane Center Axis Shear Plane Al2SiO5 Ol+Px+MORB 2mm Fibrolite shear band trends 62 And And Sil 1 µm Figure 2.10. 63 500 nm c And Sil Sil 500 nm And And And And And Sil Sil d ba 100 nm Figure 2.11. 64 Figure 2.12 a c 10 nm Ky Sil 500 nm b Ky Sil 100 nm c 65 Figure 2.13. andalusite kyanite sillimanite Al2SiO5 triplet constant twist rate torsion experiments sh ea r st re ss , τ (M Pa ) 300 200 100 0 0 1 2 3 4 shear strain, γ T = 1250 °C P = 300 MPa n = 4 (assuming dislocation creep) a b shear strain, γ sh ea r st re ss , τ (M Pa ) 300 200 100 0 0 1 2 3 4 n = 1 n = 4 n = 10 Kyanite T = 1250 C P = 300 Mpa sample dimensions d = 9.83 mm l = 2.84 mm 66 67 Chapter 3 Application of phase equilibrium analysis to compositionally and texturally complex rocks: orthoamphibole-cordierite gneiss of the Thor-Odin dome, British Columbia, Canada E.T. Goergen and D.L. Whitney Submitted to Journal of Metamorphic Geology April 2009 Orthoamphibole-cordierite gneiss (OCG) from the Thor-Odin migmatite dome, British Columbia (Canada), records a complex history of compositional subdomain formation in coronas and coeval bulk compositional modification during high-temperature decompression. Pseudosections were calculated to explore phase equilibria and P-T conditions at the onset of major decompression and reaction texture formation. The pre- decompression assemblage assumed from textural analysis is orthoamphibole + garnet + kyanite + rutile ± plagioclase ± quartz. Coronal reaction textures around Al2SiO5 (typically sillimanite after kyanite) are characterized by distinct, symplectitic two-phase assemblages of spinel + cordierite, spinel + plagioclase, corundum + cordierite and sapphirine + cordierite. Thor-Odin OCG experienced changes in bulk composition during metamorphism (e.g. introduction of K). These changes were likely related to reaction of OCG pods with 68 fluids released by crystallizing melt in host migmatite. After the initial stage of decompression-related reaction texture formation, the textural and compositional evolution of OCG occurred via diffusion-controlled reactions that produced complex multi-shelled coronas on Al2SiO5 and garnet. A series of phase diagrams were calculated to model the effect of bulk composition and aH2O variation on to determine the conditions prior to decompression and corona formation, but no reasonable set of assumptions predicted the observed assemblages, mineral compositions and modal abundance of the texturally indicated pre-decompression assemblage. This result may reflect the difficulty of modeling a system in which absolute and effective bulk composition evolved during metamorphism as a result of fluid influx and associated increases in diffusion rates and length scales. For example, all samples analyzed in this study, regardless of the extent of bulk compositional modification, show pervasive evidence for diffusionally controlled reactions that may have been influenced by chemical gradients at multiple scales. INTRODUCTION The petrogenesis and P-T histories of orthoamphibole-cordierite gneiss (OCG) have been the focus of many studies because of the unusual bulk composition of these rocks (Eskola, 1914; Schumacher, 1988; Schneiderman & Tracy, 1991; Pan & Fleet, 1995; Owen & Greenough, 2000; Peck & Smith, 2005; Pitra et al., 2008), their occurrence in many high-grade terranes (Hudson & Harte, 1985; Schneiderman & Tracy, 1991, Owen & Greenough, 2000; Buick et al., 2006), and their potential for providing reaction history 69 information in their assemblages and textures (Spear, 1980; Robinson et al., 1982; Schneiderman & Tracy, 1991; Raith et al., 2008). The typical OCG bulk composition is enriched in Mg, Al and Fe, and depleted in alkali content, a chemical signature that does not correspond to any common protolith. Determining the P-T history of OCGs is challenging for many of the same reasons that these rocks are of such great interest. Until recently, the petrogenetic evolution of OCG was not explored in chemical systems relevant to natural OCG compositions because of a lack of thermodynamic models for the orthoamphibole systems; e.g. models that accounted for the effect of Na on the stability of gedrite (Harley, 1985; Hudson & Harte, 1985; Fockenberg & Schreyer, 1994; Ouzegane et al., 1996; Fischer et al., 1999). Recently, however, Diener et al. (2007) provided a-x models of orthoamphibole, thereby allowing systematic phase equilibria analysis of these rocks by means of bulk composition-specific phase diagrams (pseudosections). One important, and sometimes overlooked, aspect of pseudosection calculation is how to properly determine the effective bulk composition of the rock to be investigated (e.g. Stüwe, 1997; Evans, 2004). This determination may include assumptions about the relative timing of the textural evolution of a sample as well estimation of the length and time-scales of diffusive processes (i.e. the scale of the system in chemical communication). The issue of effective bulk composition is particularly important for rocks that contain porphyroblasts and/or coronal reaction textures. The former sequester elements and the latter represent compositional subdomains (and in this way are similar to 70 porphyroblasts) as well as evidence for disequilibrium. The P-T evolution of texturally complex rocks therefore cannot be assessed via typical methods of calculating equilibrium P-T pseudosections for a rock’s bulk composition (e.g. Baldwin et al., 2007; White et al., 2008). As a result of their Mg-Al rich composition, OCGs commonly contain refractory Al-rich phases such as garnet and Al2SiO5 polymorphs. Such refractory phases are commonly associated with the development of coronal reaction textures, including symplectites. Many OCG, including those discussed here, contain symplectitic reaction textures after garnet, or more commonly Al2SiO5 (e.g. Moore & Waters, 1990; Schneiderman & Tracy, 1991; Höltta et al., 1997; Koshimoto et al., 2004; Buick et al., 2006; Heimann et al., 2006; Santosh et al., 2006; Shimpo et al., 2006; Tsungogae, 2006). OCG have many characteristics that make them desirable for investigating P-T evolution and reaction histories, but the presence of large porphyroblasts, complex reaction textures and associated compositional sub-domains makes it difficult to extract quantitative information through standard techniques. Furthermore, OCG in migmatite domes may have the added complexity of having experienced complex P-T paths during ascent of partially molten crust; including possible convective overturn in different lobes (subdomes) of the migmatite dome (Whitney et al., 2004). The goals of this paper are to investigate the controls on phase assemblage evolution via pseudosection analysis in the context of bulk compositional variation, with a focus on a suite of OCGs from the Thor-Odin migmatite dome (Figs. 3.1, 3.2). In addition to containing large garnet and Al2SiO5 porphyroblasts and associated coronal 71 reaction textures, some OCGs from the Thor-Odin dome are also complicated by a bulk compositional modification event indicated by replacement of orthoamphibole by biotite. The least modified bulk composition is explored in detail through a series of P-T and T- MH2O diagrams, and the relative timing of bulk chemical modification and the initiation of reaction texture formation are discussed. GEOLOGIC SETTING The Thor-Odin gneiss dome, part of the Shuswap metamorphic core complex in south- central British Columbia (Reesor & Moore, 1971; Fig. 3.1), is one of several N-S elongate migmatite domes in the Shuswap and other metamorphic core complexes of the internal (Omineca) crystalline belt of the North American Cordillera (Fig. 3.1). The Thor- Odin dome is characterized by high-grade (sillimanite-K-feldspar zone) migmatitic ortho- and paragneiss (lower unit of Fig. 1). The dome rocks are overlain by metasupracrustal rocks that are also largely in the sillimanite-K-feldspar metamorphic zone and are also in part migmatitic (middle unit/Selkirk Allochthon, Fig. 3.1) (Reesor & Moore, 1971; Carr, 1992). Previous metamorphic studies of the lower and middle units determined maximum metamorphic conditions of 800˚C at > 10 kbar (Norlander et al., 2002) for the core of the dome, and ~650˚C - 800˚C at 7-9 kbar (Ghent et al., 1977; Nyman et al., 1995) for middle unit rocks, with the higher-grade middle unit rocks located closest to the dome. The melt fraction of migmatites increases from higher to lower structural levels (Reesor & Moore, 1971; Vanderhaeghe et al., 1999; Hinchey et al., 2007), and the 72 middle unit – lower unit boundary has been proposed as a metatexite-diatexite boundary in the vicinity of Mt. Odin and Mt. Thor (Vanderhaeghe et al., 1999) (Fig. 3.2). This transition, however, is not clearly present in other parts of the dome; e.g. the northern margin. Several tectonic models have been proposed for the evolution of the Thor-Odin dome and the Shuswap metamorphic core complex as a whole (e.g. Vanderhaeghe et al., 1999; Brown, 2004; Teyssier et al., 2005; Williams & Jiang, 2005), but it is generally agreed that Cordilleran high-grade metamorphism and ductile deformation began in the Late Mesozoic and continued into the Paleogene-Eocene. Crustal melting, extensional collapse and development of Thor-Odin dome occurred in the late Eocene (Parrish et al., 1988; Vanderhaeghe et al., 2003; Johnston et al., 2000; Carr & Simony, 2006; Teyssier et al., 2005; Hinchey et al. 2006), coeval with these processes in other migmatite domes of the Omineca belt (Gordon et al., 2008; Kruckenberg et al., 2008). ORTHOAMPHIBOLE-CORDIERITE GNEISS OF THE THOR-ODIN DOME Pods and lenses of Mg-Al-rich rocks (orthoamphibole-cordierite gneiss, OCG) occur throughout the Thor-Odin and other migmatite domes of the Omineca belt (Okanogan dome, WA: Harvey & Hoisch, 1994; Kruckenberg et al., 2008; Passmore dome of the Valhalla complex: Marshall & Simandl, 2006) (Fig. 3.1). OCG in the Thor-Odin dome occur as isolated boudins or as elongate lenses that crop out along strike for tens to hundreds of meters. 73 Pods and lenses vary in thickness from a few tens of centimeters to ~10s of meters. In all cases the elongate lenses are concordant with planar fabrics developed locally and regionally, and, in the case of boudins, the boudin necks are parallel to neighboring and regional linear fabrics and fold axes. Internally, the OCG units vary from well foliated and lineated to massive, but there is no difference in mineral assemblages in the ones with or without a fabric. In addition, no relationship exists between pod or lens size and the presence of a discernible fabric. OCG from the vicinity of Mt. Odin (Fig. 3.2) are boudins within migmatite. OCG from the northern and western portions of the dome are interlayered with other non orthoamphibole bearing Mg-Al rich metasedimentary lithologies (Al2SiO5 + Q + Bi + Cd ± Gt) as well as migmatitic orthogneiss. In some localities, the OCG layers are folded, with axial hinges parallel to the regional lineation. Foliation is present near lithologic contacts and is defined dominantly by phlogopite. Linear fabrics are defined by kyanite ± sillimanite and orthoamphibole ± cordierite. Coronas are visible in outcrop (Fig. 3.3). They partially to completely replace (pseudomorph) garnet and Al2SiO5 polymorphs (kyanite/sillimanite) in both foliated and non-foliated rocks. Where foliation is present, it is overgrown by the coronal textures associated with Al2SiO5 and garnet. There is no evidence for deformation in the reaction textures, even in foliated rocks. These observations are consistent with two distinct texture-forming events in the orthoamphibole-cordierite gneiss: (1) formation of foliation and lineation concordant with similar fabric in the enveloping gneiss, and (2) formation of coronas that overprint the earlier fabric. The former was likely related to Cordilleran 74 deformation during crustal thickening and regional metamorphism, and the latter was more likely associated with later decompression during development of the Thor-Odin dome by vertical flow of partially molten crust (Teyssier & Whitney, 2002). PETROLOGY AND MINERAL CHEMISTRY Petrographic, bulk rock and mineral chemical observations have been presented previously for OCG from the Bear Paw Lake area of the Thor-Odin dome (Norlander et al., 2002; Hinchey et al., 2007; Hinchey & Carr, 2007). The focus of this paper is on OCG collected from the Saturday Glacier area (Fig. 3.2). Both regions are from the southern part of the Thor-Odin dome. Mineral compositions and qualitative X-ray maps were obtained using a JEOL JXA-8900 electron microprobe in the Department of Geology and Geophysics at the University of Minnesota. Quantitative analysis operating conditions were a 15 keV accelerating voltage, a 20 nA beam current, and a range of beam diameters (focused for minerals without volatile elements, 5 mm for minerals with volatile elements). X-ray maps were acquired using a 35 nA beam current and a beam diameter of 1 mm. Natural standards were used, and matrix correction was achieved using the ZAF correction routine. Mineral analyzed were cordierite, orthoamphibole, biotite, garnet, spinel, plagioclase, and sapphirine. Assemblages are summarized in Table 3.1. 75 SOUTHERN THOR-ODIN DOME: SAMPLES 06ET-2, 02-3.03, 05-2.54 01-2.07 Samples from the vicinity of Mts. Thor and Odin (Fig. 3.2) are from concordant OCG pods (Fig. 2.4a) that range in size from a few to 30 meters in length and that are surrounded by migmatite. The pods are cut by veins consisting primarily of coarse- grained cordierite, quartz and plagioclase (Fig. 3.4a-b). OCG contains gedrite ± anthophyllite + cordierite + garnet + biotite + kyanite/sillimanite + hercynitic spinel + corundum + plagioclase ± sapphirine ± quartz ± staurolite + apatite (2-3 modal%). Accessory phases are ilmenite, magnetite, rutile, monazite, zircon ± högbomite ± pyrite and other sulfides. Rutile and ilmenite are the dominant Fe-Ti oxides in the matrix as well as in garnet and Al2SiO5 as inclusions. Analyzed samples vary in the amount of biotite that has partially replaced orthoamphibole and in the details of assemblages in symplectitic coronal reaction textures. The least retrogressed sample analyzed from the southern part of the dome is 01-2.07 (Fig. 3.2). Orthoamphibole Orthoamphibole varies in abundance (2-45 modal %) and typically occurs in randomly oriented sprays (0.25->1 cm long). Orthoamphibole has been partially to nearly completely replaced by biotite (Fig. 3.5). Where this replacement is widespread, relict orthoamphibole occurs only within garnet or between closely spaced garnet porphyroblasts (Fig. 3.5b). Some orthoamphibole has also been partially replaced by cordierite (Fig. 3.6). Gedrite is the primary orthoamphibole in these rocks, but a continuous compositional trend to anthophyllite exists in most samples (Fig. 3.7a). 76 A-site (Na+K) occupancies in Thor-Odin OCG amphibole range from 0.5 to 1.6 p.f.u. (Fig. 3.7b) (Table 3.2). The tschermak’s exchange vectors also represent a continuous range (Fig. 3.7c) that is related to zoning at the rims of orthoamphibole grains. Orthoamphibole is optically and compositionally zoned adjacent to corona textures and, in some cases, where replaced by cordierite. From core to rim, orthoamphibole compositions increase most notably in Al (1.7 to 2.4 p.f.u.), Na (0.2 to 0.4 p.f.u.) and decrease in Si (6.8 to 6.3 p.f.u.) and Mg (XMg 0.7 to 0.6) (Table 3.2). The least retrogressed sample (01-2.07) differs from the rest of the sample suite, showing compositions reflecting inter-grown anthophyllite and gedrite. Biotite Biotite ranges from 25-44 modal % in samples 06ET-2 & 05-2.54, but comprises only ~4% of sample 01-2.07. Biotite in all samples is texturally late with respect to orthoamphibole and garnet, but its growth relative to the development of the reaction textures is ambiguous. Biotite in some cases is not in direct contact with the coronal textures; in other cases it appears to be part of the original reactant assemblage associated with coronal texture development where biotite is being consumed by the formation of the cordierite moat associated with the symplectitic reaction textures (see below for further discussion). This ambiguity in terms of the timing of growth occurs in a single thin section, and is observed in all symplectite bearing samples. Biotite displays systematic color variation from brown to green depending on textural location in most samples, with the exception of sample 01-2.07, in which no 77 compositional or color variation in biotite occurs. Brown biotite occurs adjacent to coronal reaction textures and in other matrix locations, and, in most samples, grades from brown to green proximal to coronas on garnet (Fig. 3.8). Biotite in all samples is extremely Al-rich (Table 3.3) with AlVI and AlIV occupancies ranging from 1.5-1.9 and 1.8-2.0, respectively. Biotite has an XMg ranging from 0.65-0.74, with systematic zoning related to textural placement. Brown biotite from sample 01-2.07 has consistent XMg values of 0.75. Green biotite that has partially replaced garnet is the most Fe-rich (XMg = 0.59). Biotite color variation is most likely related to Fe and Ti content (Table 3.3). Garnet Garnet occurs as large (0.5-1 cm in diameter), partially resorbed anhedral to euhedral porphyroblasts (Fig. 3.9 a, b). Garnet is surrounded by a rim of symplectitic green biotite + orthoamphibole + cordierite + ilmenite ± quartz (Fig. 3.9 c, d) in all investigated samples, with the exception of 01-2.07, in which garnet is replaced by a monophase corona of plagioclase. Garnet contains inclusions of orthoamphibole, kyanite, plagioclase, quartz, apatite, rutile ± zoisite, and tourmaline samples 06ET-2 and 02-3.03, and orthoamphibole, apatite, quartz, plagioclase, ilmenite, rutile ± pyrite in sample 01-2- 07. In some cases, where garnet and Al2SiO5 are adjacent to each other the coronal reaction zones around both phases are connected by a zone in which orthoamphibole has been replaced by fine-grained (0.5 to 1 mm) biotite and cordierite (Fig. 3.9d). 78 Garnet (Table 3.4) is typically almandine-rich (Xalm= 0.61; Xprp = 0.27; Xgrs = 0.11) with average compositions staying fairly consistent across the sample suite. In some samples that contain well-developed symplectitic reaction textures around garnet and Al2SiO5 (02.203, ET06-2), garnet commonly exhibits complex asymmetric zoning (Fig. 3.10): Xalm and Xprp decrease from rim to rim, and Xgrs and Xsps increase (Fig. 3.10a). However, the relationship of reaction textures and garnet compositional zoning is not entirely systematic. Complex zoning is not observed in all garnets from a single thin section or in all symplectite-bearing samples. Where asymmetric zoning does not occur, relatively flat profiles are present with slight Mg and Fe zoning at the rims. In the least altered sample, which lacks well-developed reaction textures (01-2.07), garnet zoning is characterized by increased Fe and decreased Mg and Ca at rims relative to the core (rim: Xalm = 0.78; Xprp = 0.13; XGrs 0.02), deviations from the overall zoning trends in this sample are attributed to the location of fractures (Fig. 3.10b). Cordierite Cordierite is associated with biotite along cleavage planes of orthoamphibole and is also present in the symplectitic reaction textures. Cordierite is unzoned in all analyzed samples, and XMg ranges from 0.76-0.85 (Table 3.5). There is slight variation in composition between matrix cordierite associated with orthoamphibole, and cordierite associated with the symplectitic reaction around garnet or Al2SiO5 with the former representing the higher XMg values and the latter representing the lower end of the XMg range. 79 Corona phases In OCG from the Saturday Glacier area, sillimanite has partially to (more typically) completely pseudomorphed kyanite (Fig. 3.11) and occurs as clusters of acicular grains within the tabular shape of the former kyanite. This is in contrast to OCG described by Norlander et al. (2002) and Hinchey & Carr (2007) in samples from the Bear Paw Lake area, where kyanite is still present and is only locally transformed to sillimanite. Sillimanite in samples from the Saturday Glacier area is only observed as pseudomorphs after kyanite, and does not occur as a separate matrix phase. The central Al2SiO5 grain is surrounded by a symplectitic corona consisting of varying amounts of vermicular hercynitic spinel (XMg 0.3-0.4) + plagioclase (XAn 0.99) + cordierite (XMg ~0.75) + corundum + ilmenite ± sapphirine (XMg 0.74). The composition of sapphirine ranges from above the 7:9:3 end member along increasing Al content toward the 3:5:1 end member (Fig. 3.12). Plagioclase (XAn = 0.90-0.99) in the inner corona shell is polycrystalline and is characterized by self-affine grain boundaries separating it from the outer cordierite shell (Fig. 3.13). Some samples lack plagioclase (e.g. sample 05-2.54) and contain only cordierite as the host phase to the vermicular spinel and other aluminous corona phases. The phases associated with the symplectitic coronal replacement are present as distinct two-phase assemblages of Spl + Crd, Spl + Pl, Crn + Crd and Spr + Crd. The modal abundance of phases in the symplectites varies greatly within and between samples. There is no relationship between the compositions or the relative modal 80 abundances of phases within the texture with the assemblages in the surrounding matrix. This variation is likely, in part, the result of a two-dimensional view of a three- dimensional texture, but there are phase relationships that are consistently observed (Fig. 3.11). For example, Spr + Crd always occurs furthest from sillimanite with respect to corundum, spinel, and plagioclase, which are located closer to sillimanite. The Crn + Crd assemblage is more variable in its location, and has no obvious systematic relationship with surrounding phases or size of the corona. Despite this non-systematic variation in assemblage within and between different reaction textures, all textures are surrounded by a moat of cordierite with a consistent thickness of 100 µm. PETROGENESIS OF ORTHOAMPHIBOLE-CORDIERITE ROCKS Since the first report of orthoamphibole-cordierite rocks by Eskola (1914), there has been much debate about the origin and significance of these rocks. The current view is that many Mg-Al-rich rocks were mafic volcanic rocks that experienced hydrothermal alteration prior to regional metamorphism (Robinson, 1982; Peck & Valley, 2000; Owen & Greenough, 2000; Buick et al., 2006). Other hypotheses invoke alteration related to genesis of sulfide ore deposits (Pan & Fleet, 1995; Roberts et al., 2003) or metasomatic alteration of semi-pelitic sediments (Dasgupta, 1999; Shimpo et al., 2006). Whatever the protolith, uncertainty also remains about the mechanisms by which the protolith composition was changed via alteration, melt extraction and/or other metasomatic processes. 81 To illustrate the major element compositional variation of the Thor-Odin OCG, major element compositions of Thor-Odin samples are compared with OCG and other Mg-Al-rich rocks from the literature (Hudson & Harte, 1985; Rheinhardt, 1987; Moore & Waters, 1990; Munz, 1990; Arnold et al., 1995; Pan & Fleet, 1995; Guiraud et al., 1996; Hölttä et al., 1997; Labotka & Kath, 2001; McClintock & Cooper, 2003; Roberts et al., 2003; Peck & Smith, 2005; Buck et al., 2006; Hinchey & Carr, 2007) as well as with bulk compositions of mafic volcanic rocks (data from PetDB http://www.petdb.org Lehnert et al., 2000) that have experienced various degrees of alteration (Fig. 3.14). Bulk compositions of 6 Thor-Odin samples were determined by XRF analysis, which was carried out at Macalester College. In addition, ICP-MS analyses were done at the University of Minnesota on 2 samples from the Bear Paw Lake area (Table 3.6). Care was taken to obtain representative analyses of these heterogeneous rocks by analyzing relative large (hand sample sized) rocks. Bulk rock major elements are plotted in AFM and AFMCaO space for comparison of OCG with different altered and unaltered mafic volcanic rock compositions. OCGs show considerable variability, primarily in Fe and Mg content. This variation, however, is distinct from that of mafic volcanic suites, and is particularly distinct from highly altered mafic volcanic rocks and ultramafic rocks. OCG are markedly higher in Al content and lower in CaO along very different trajectories than extensively altered mafic volcanic rocks (Fig. 3.14). OCG is closer in composition to fresh or slightly altered mafic volcanic rocks. However, their depletion in CaO relative to 82 altered mafic rocks confirm that, if mafic rocks were the protoliths, additional processes have operated before or during metamorphism to affect major element composition. In addition to the bulk compositional variation of Mg-Al rich rocks in general and OCG in particular, there are minor but important bulk compositional differences that occur within and between the sample suite of this study and the Thor-Odin OCG described in Norlander et al. (2002) and Hinchey & Carr (2007) (Table 2.6). Comparison of major element data shows slight variation, with the noted exception of K. Samples from the present study show an enrichment in K relative to other OCG. K enrichment and other processes that have affected the bulk composition of the OCG must be considered in an analysis of the compositional and textural development of these rocks. BULK COMPOSITION AND PHASE DIAGRAM CALCULATIONS A critical question in calculating a bulk composition-specific phase diagram (pseudosection) is how to define the appropriate bulk composition of the rock volume of interest. Pseudosections are constructed based on the assumption that the bulk composition measured is representative of a rock volume across the entire P-T space of interest. High-grade rocks experience many processes that affect the bulk composition at various scales (e.g. porphyroblast growth, partial melting, hydrothermal alteration, reaction texture formation, deformation), and it can be difficult to determine the appropriate bulk composition to use. However, with careful implementation, 83 pseudosections are powerful tools for defining P-T-X constraints for systems in which other methods may not be applicable. Southern Thor-Odin dome OCG are complicated by (1) the varying amounts of biotite replacement of orthoamphibole, and (2) the development of symplectitic coronal reaction textures. The former represents a compositional shift at the bulk rock scale (i.e., introduction of K), whereas the latter represents the development of compositional subdomains related to reaction texture formation. These textural features indicate an evolving compositional history that, if it can be deciphered, records useful information about the petrologic and tectonic history of the rock. DETERMINATION OF BULK COMPOSITION FOR USE IN PSEUDOSECTIONS The Thor-Odin OCG fall within a relatively small range in terms of abundance of most major elements, with the exception of samples from the Saturday Glacier area, which vary in alkali content, in particular K2O (Table 3.6), owing to the higher modal proportion of biotite in these rocks. Thor-Odin OCGs that occur in pods tend to have more biotite replacement than OCGs that occur in layers, such as those in the Bear Paw Lake region (Norlander et al., 2002; Hinchey & Carr, 2007). Textures involving biotite in these rocks are consistent with the interpretation that biotite replaced orthoamphibole after the thermal maximum but before or during reaction texture formation. To compare the samples from this study with those from the Bear Paw Lake region, a bulk composition was recalculated by subtracting biotite from the assemblage 84 and substituting the same modal proportion of orthoamphibole that it likely replaced (taking into account density differences). The result (Table 3.7) illustrates that OCGs that exhibit extensive biotite replacement were similar in bulk composition to those in the Bear Paw Lake region prior to modification. Thus the Bear Paw Lake bulk compositions were used as a proxy for the composition of OCGs from the Saturday Glacier region prior to biotite replacement. PSEUDOSECTION CALCULATIONS Calculations were performed in the NCKFMASHTO system with THERMOCALC version 3.31 (Powell & Holland, 1998, updated April 2008) using the November 2004 updated internally consistent dataset (Holland & Powell, 1998; filename tc-ds55.txt). The NCKFMASHTO system was chosen because it best represents the compositions considered in this study as well as the recent development of a-x models that incorporate Fe3+ allow for a more realistic representation of phase topology as well as the ability to assess Fe-Ti oxide relationships. Pseudosections were constructed to identify the P-T conditions for stability of the texturally inferred pre-decompression assemblage of orthoamphibole + garnet + kyanite + biotite(minor) + rutile ± plagioclase ± quartz (plagioclase and quartz presence is based on garnet inclusion assemblages). Phases considered in the calculations are biotite, garnet and silicate melt (White et al., 2007), chlorite (Holland et al., 1998), clino- and orthoamphibole (Diener et al., 2007, with thermodynamic updates in Diener et al., 2008), orthopyroxene (White et al., 2002), 85 plagioclase-K-feldspar (Holland & Powell, 2003), spinel-magnetite, ilmenite-hematite (White et al., 2002), cordierite, epidote, staurolite and talc (Holland & Powell, 1998), and muscovite-paragonite (Coggon & Holland, 2002). The Al2SiO5 polymorphs, rutile, quartz, corundum and H2O were treated as end-members. Silicate melt was used in constructing the phase diagrams, but its stability in these Al-rich compositions was only found at temperatures greater than the range important to this study (>850°C), even when assuming H2O saturation, as a result of the high concentrations of Al2O3 (>22 wt %) in the bulk composition. The addition of Fe2O3 to the phase equilibria analysis is difficult without the aid of quantitative analysis to determine the amount of Fe3. In a general study of orthoamphibole rocks, Diener et al. (2008) utilized a € XFe 3+ value of 0.17 (based on data from Lal & Shuka, 1975). Use of this value predicts widespread stability of hematite for an average OCG bulk composition. High contents of Fe3+ in Diener et al. (2008) were assumed for the general case of OCG evolution due to the inferred metasomatic origin of this rock type. OCG in the present study, however, do not contain hematite other than texturally late fracture filling and oxide residues on some minerals. Magnetite occurs in some samples but is restricted to coronal reaction textures associated with garnet. Magnetite formation was likely related to the local boundary conditions during reaction texture formation and may not reflect bulk rock conditions. Rutile is the dominant Fe-Ti oxide in the symplectitic reaction textures, and the matrix Fe-Ti oxides consist of both rutile and ilmenite. The textural relationship of these phases is hard to determine as both phases contain lamellae of the other. To estimate Fe3+, a series of T-X diagrams were 86 calculated from published pseudosections and bulk compositions similar to those of this study to find the maximum value of molar O at which hematite and magnetite are absent or limited in the P-T range of interest. This was achieved using an € XFe 3+ value of 0.085 (Table 3.8). Another issue for pseudosection calculations is determining the appropriate water content of the bulk composition. As a first case, a P-T pseudosection was calculated assuming H2O in excess (Fig. 3.15). This approach predicts the near ubiquitous presence of coexisting clinoamphibole (hornblende) + orthoamphibole, although clinoamphibole does not occur in the Thor-Odin OCG. A hornblende-absent phase field (labeled ‘A’ in Fig. 3.15) predicts the presence of chlorite and absence of garnet, but this is also inconsistent with petrographic observations in the OCG. The persistence of chlorite to high temperatures is most likely a consequence of the assumption of water saturation, as € aH2O strongly controls the stabilization of water- rich phases (Guiraud et al., 2001). The extended stability, however, may also be related to the thermodynamically inconsistent equipartition constraint in the current chlorite a-x model (Holland & Powell, 2006). A T-MH2O pseudosection was calculated to explore the effect of H2O on phase equilibria in these rocks (Fig. 3.16). These calculations predict that a large amount of water (> 24 molar mol%) is needed to saturate the OCG bulk composition. Epidote would be stable in this bulk composition at low temperatures and high molar % H2O. Epidote occurs as inclusions in garnet, possibly indicating that these rocks were H2O-rich during an earlier stage of metamorphism but were successively 87 dehydrated during progressive metamorphism and perhaps partial melting. The OCG may also have been dehydrated during earlier (Paleoproterozoic) metamorphism prior to the Cordilleran metamorphic event explored in this paper. Although lowering the molar water content in the calculation removes the high temperature stability of chlorite, this does not solve the issue of the wide stability range of hornblende. Even at very low water activities, hornblende is stable except in phase fields in which plagioclase takes up excess Ca and destabilizes hornblende. It is likely that the CaO content of OCG in this study artificially stabilizes clinoamphibole (hornblende) in the calculations. This stabilization is artificial in that much of the CaO in the bulk composition is sequestered in garnet (Xgrs ~0.12) and in plagioclase in the reaction textures and therefore may not be part of the effective bulk composition. Hornblende may have been stable at some point in the history of these rocks, but there is no textural evidence that it was stable during the part of the metamorphic history of interest to this study; i.e. during texture formation. The absence of clinoamphibole is also in accord with textural evidence that the interpreted peak assemblage had a much higher variance than that of an average OCG (Diener et al., 2008). The control of Ca on hornblende stability was investigated by subtracting garnet core volumes from the bulk composition to better model the effective bulk composition. Garnet core volumes and the associated modal abundance of the cores were determined from garnet zoning maps determined by electron microprobe and by image analysis on a selection of ten representative slices obtained via high-resolution X-ray computed 88 tomography at the University of Texas-Austin. The average modal abundance was calculated and combined with the average garnet core composition to determine the composition to be removed from the bulk rock. Two different modal amounts of garnets were used (10% and 6%). The 10% estimate was made by nearly completely removing garnet from each image slice. This method overestimated the diameter of high-calcium cores in the garnets. This was done in an attempt to take into account Ca in apatite as well as in plagioclase present in the reaction textures. The 6% estimate only removed the cores of garnets from the images. The new composition was then renormalized to 100 mol%. A series of P-T pseudosections were calculated by using the MH2O diagram to construct T- X diagrams at decreasing water contents and with different CaO contents (Table 3.8) to provide reliable topological relationships as well as solution model compositional starting points for phase diagram construction. The result illustrates the strong control of CaO content on the stability of hornblende (Fig. 3.17). A decrease in the stability of hornblende is countered by an increase in plagioclase and orthopyroxene stability. Lower water contents also lead to a reduction in the high-pressure stability of hornblende and an increase in plagioclase stability. With decreasing CaO and H2O contents, the hornblende-out phase boundary becomes hooked shaped, and an arc of hornblende-free phase assemblages develops at moderate P and T (Fig. 3.17). H2O strongly controls the high temperature stability of hornblende, whereas CaO influences its low-pressure stability (as seen in the curvature of the hornblende-out phase field; Fig. 2.17). 89 Orthoamphibole + kyanite stability in the absence of hornblende, the primary assemblage inferred for these rocks, was not modeled in this analysis because hornblende persists to high pressure. The omission of the hornblende solution model was not considered because the bulk compositions used do not preclude its stability. In addition, phases not present can, in theory, be used to determine limits on a rock’s P-T conditions and should not be omitted for arbitrary reasons. A reduction of system size was also not considered because it would hinder the determination of P-T conditions by limiting direct comparison of mineral compositions with the model predictions. The observations of the phase topology predicted in the NCKFMASHTO system suggests one of the following scenarios to explain inconsistencies between the textural/chemical observations with model predictions: 1) the assumption of the high- pressure assemblage stable prior to decompression is incorrect or incomplete, 2) the chosen bulk compositions are not representative of the effective bulk composition and/or 3) the a-x models may over/underpredict the stability of some phases. DISCUSSION Mineral assemblages in OCG are very sensitive to bulk composition: small differences in bulk composition result in different assemblages. Furthermore, literature values for major element composition of OCG show considerable variability (Fig. 3.14); e.g. SiO2 ranges from ~36 to 70 wt%, resulting in variability among OCG suites. 90 These complexities are not unique to OCG, but the distinctive bulk composition of OCG and the limited previous exploration of their phase equilibria make quantitative comparisons of different studies and tectonic settings difficult. Fortunately, however, recent developments in a-x models for orthoamphibole provide an opportunity for exploring the phase topology of these rocks (Diener et al., 2007, 2008). In their calculation of pseudosections for orthoamphibole-cordierite gneiss, Diener et al. (2008) used a bulk composition determined from a selection of OCG reported in the literature. Their selected composition has relatively high ferric iron and SiO2 contents (63 mol% SiO2) in comparison to the compositions presented here. Most of the Diener et al. (2008) pseudosections predict clinoamphibole-free orthoamphibole + cordierite assemblages only at moderate temperature and P < 6 kbar. However, many OCG do not contain clinoamphibole either as a matrix phase or as inclusions in other minerals, are not as Fe3+ rich as the Diener et al. bulk composition (as inferred from oxide minerals where only rutile and ilmenite are the most commonly reported), and have experienced higher-pressure metamorphism without evidence of a second amphibole present (e.g. Schneiderman & Tracy, 1991; Hölttä, 1997; Owen & Greenough, 2000; Koshimoto et al., 2004; Peck & Smith, 2005). In addition, the textural evidence in Thor- Odin OCG suggests that the relationship between orthoamphibole and cordierite is one of replacement rather than equilibrium. Although the OCG composition in Diener et al. (2008) may be representative of the average value from the literature, it is clear that the variation in bulk compositions, including Fe3+ contents, in OCG may deviate significantly from the average model values. In addition, many OCG had complex 91 metamorphic histories (including polymetamorphism) that affected their effective bulk composition, water content, and fO2. Therefore, although the Diener et al. studies represent an important step forward in quantitative approaches to understand the petrologic evolution of OCG, the range of composition and reaction processes common to many OCGs may make direct comparisons to their average OCG composition difficult until future studies are able to illustrate a more complete range of phase topologies in these rocks. The Thor-Odin OCGs are compositionally and texturally complex, but there are systematic characteristics among the OCG pods that allow the interpretation of metamorphic processes leading up to corona development. Despite limitations of the phase equilibria approach to modeling the P-T evolution of these rocks, it is possible to infer some aspects of their metamorphic evolution from the calculated phase diagrams and textural observations. TIMING OF BULK CHEMICAL MODIFICATION Many Thor-Odin OCG experienced an apparent hydrothermal back reaction with surrounding migmatites, causing a bulk compositional shift during Cordilleran metamorphism. This event was not widespread but was concentrated in OCG associated with high melt fraction migmatite host rocks (diatexite). The protracted melting history interpreted for the migmatitic host rocks of the Thor-Odin dome (Hinchey et al., 2006) could have led to multiple pulses of melting/extraction and fluid interaction (e.g. Gordon 92 et al., 2008; Kruckenberg et al., 2008) that may in part have been responsible for the heterogeneity of the OCG. The replacement of orthoamphibole by biotite indicates a significant change in the chemical evolution of some of the Thor-Odin OCG and is evidence for the chemical interaction of the OCG with the surrounding migmatite. Pod-shaped OCGs are more commonly associated with high melt fraction migmatite than those in layers, and OCG pods contain more biotite than the layered OCG. In OCG adjacent to diatexite, a biotite- rich layer developed around the pods at the contact with migmatite (Fig. 3.4b); this textural feature is also consistent with chemical interaction between the OCG and host rocks. Thor-Odin OCG were not saturated with respect to H2O at maximum temperature conditions. Observations at the outcrop and thin section scale are consistent with the development of an MH2O gradient between migmatite/melt and the adjacent pod, and introduction of H2O and K2O from migmatite to the pods, triggering biotite growth. This hydration of the pods, possibly related to crystallization of melt in the migmatites, was also likely responsible for the intrusion of cordierite + quartz + plagioclase veins that cut through the OCG pods. This interaction of surrounding migmatite and interlayered lithologies is similar to hydrothermal back reactions between melt and restite in migmatite (e.g. Kriegsman, 2001; Thompson, 2001; Brown, 2002). P-T CONDITIONS 93 A previous study of the P-T history of the Thor-Odin dome used a combination of internally consistent thermobarometric techniques (TWQ) and traditional empirical thermobarometry on OCG and garnet-amphibolites (Norlander et al., 2002). Results indicated conditions of 10 kbar and 750 °C. Norlander et al. (2002) proposed that the rocks had experienced isothermal decompression based on the presence of symplectitic reaction textures and associated univariant equilibria derived from petrogenetic grids for assemblages in the reaction textures (e.g. cordierite rims on kyanite). The presence of symplectitic textures has long been associated with, and, in some cases used as evidence for, the occurrence of isothermal decompression. However, the magnitude and trajectory of decompression necessary for this type of symplectite formation is not well understood. Overstepping of reactions results in the development of chemical potential gradients leading to diffusion-driven reactions. Symplectites are not necessarily representative of large amounts or rapid rates of decompression, as the boundary conditions for their formation are closely tied to assemblages present prior to texture formation and the availability of H2O. It is particularly difficult to quantify the magnitude of decompression in the Thor- Odin OCG. Kyanite and orthoamphibole have been partially replaced by cordierite, consistent with decompression at elevated temperature, but the steep slope of the cordierite-out phase boundary in OCG leads to large changes in pressure with increasing temperature (~4-1 kbar of decompression at 650 and 750°C, respectively). Any further decompression could be recorded in the symplectitic reaction textures, but obtaining quantitative information is difficult. Furthermore, given the occurrence of the Thor-Odin 94 OCG as pods in the core of a high melt fraction migmatite dome, it is likely that they did not experience a monotonic decrease in pressure. TIMING OF REACTION TEXTURE INITIATION Reaction textures represent a reaction frozen in progress and cannot typically be interpreted using a pseudosection because: (1) a reaction texture by definition represents disequilibrium open system behavior, and (2) once a reaction texture is developed, the bulk composition relevant to its continued formation is distinct from that of the bulk rock. In combination, this makes determining any ‘bulk composition’ related to the reaction texture difficult. Careful observation of reaction textures may reveal the stable assemblage at the initiation of reaction texture formation. As soon as a coronal reaction texture develops, the initial reaction no longer drives the progress of the reaction, and reactions at the boundaries between the product and reactants (and the progress of texture formation) are driven by diffusion via chemical potential gradients across the product phase and at the interfaces separating the different two-phase assemblages (e.g. Carlson & Johnson, 1990; Ashworth et al., 1998). It is the first growth of the product phase cordierite that is of interest in this study, as cordierite first appearance represents the last point (in terms of pressure, temperature, and composition/chemical potential) that the reaction texture phases were in ‘equilibrium’ with the bulk rock. Understanding this first stage also helps place reaction 95 texture growth in a tectonic context, as well as providing boundary conditions for further study of the resultant texture. Prior to corona formation, the matrix assemblage consisted of orthoamphibole + garnet + kyanite + biotite(minor) + rutile ± plagioclase ± quartz (Fig. 3.18). A consistent feature for all of the pseudosections calculated for the OCG in this study is the close relationship between the appearance and absence (in terms of modal proportions) of cordierite and Al2SiO5 across wide regions of P-T space (Fig. 3.15b). In addition, the modal proportion of orthoamphibole decreases as the Al2SiO5 mode decreases to zero, illustrating a similar relationship as the cordierite forming-reaction that initiated the reaction texture. Although quantitative estimates for the P-T conditions for this event cannot be discerned from the phase diagrams, it is clear that the reaction of orthoamphibole and Al2SiO5 to form cordierite was the original (overstepped) reaction that initiated corona formation. The influx of fluids from the crystallizing host migmatite drove the replacement of orthoamphibole by biotite and likely in close succession with reaction texture initiation and evolution (Fig. 3.18). This initial stage of melt crystallization and subsequent fluid influx may also have aided the kinetics of kyanite transformation to sillimanite, as kyanite is preserved in reaction textures from the Bear Paw Lake region (Norlander et al., 2002; Hinchey & Carr, 2007) but not in the Saturday Glacier area. These two regions represent similar structural levels in the dome and seem to have experienced similar P-T conditions and paths. The major difference is in the amount of biotite replacement of orthoamphibole, and therefore in the amount of interaction of OCG pods with aqueous 96 fluid. Differences in the phase relationships between these two areas, which are spatially close on a regional-scale, may also be related to the complex trajectories experienced by pods entrained in an ascending dome of partially molten crust. The complex compositional and textural evolution of the Thor-Odin orthoamphibole-cordierite rocks is dramatic evidence for the effect of heterogeneous (in time and space) bulk composition on the phase relations during metamorphism, and therefore the difficultly in determining P-T conditions and paths, especially for rocks in which evidence for metamorphic reactions is apparently well preserved in corona structures. Difficulties are encountered when modeling the phase topology of texturally complex OCGs as a result of uncertainty in defining the effective bulk composition controlling phase stability in these rocks. The common development of coronal textures in OCG may enhance the already complex task of properly interpreting the effective bulk composition because the diffusive system controlling corona development may be at much larger scale than is typically considered and this could possibly exert some control on the stable phase assemblages and associated mineral chemistries relatively far (cm- scale) from the corona. Nevertheless, recent advances in a-x models for orthoamphibole and other phases in OCG provide an opportunity for evaluation of the P-T-X history of this important rock type, but in some cases where there is pervasive coronal textures development it may not possible to calculate a phase diagram that accurately predicts even the phase relationships prior to coronal texture formation of the Mg-Al-rich assemblages. 97 REFERENCES Arnold, J., Sandiford, M., & Wetherley, S., 1995. Metamorphic events in the eastern Arunta Inlier, Part 1. Metamorphic Petrology. Precambrian Research, 71, 183- 205. Ashworth, J.R., Sheplev, V.S., Bryxina, N.A., Kolobov, V.Y. & Reverdatto, V.V., 1998. Diffusion-controlled corona reaction and overstepping of equilibrium in a garnet granulite, Yenisey Ridge, Siberia. Journal of Metamorphic Geology, 16, 231-246. Baldwin, J.A., Powell, R., Williams, M.L. & Goncalves, P., 2007. Formation of eclogite, and reaction during exhumation to mid-crustal levels, Snowbird tectonic zone, western Canadian Shield. Journal of Metamorphic Geology, 25, 953-974. Brown, M., 2002. Retrograde processes in migmatites and granulites revisited. Journal of Metamorphic Geology, 20, 25-40. Brown, R.L. (2004): Thrust-belt accretion and hinterland underplating of orogenic wedges; an example from the Canadian Cordillera. In Thrust Tectonics and Hydrocarbon Systems (K.R. McClay, ed.). Am. Assoc. Petroleum Geol., Mem. 82, 51-64. Buick, I.S., Hermann, J., Williams, I.S. & Gibson, R.L., 2006. A SHRIMP U–Pb and LA- ICP-MS trace element study of the petrogenesis of garnet–cordierite– orthoamphibole gneisses from the Central Zone of the Limpopo Belt, South Africa. Lithos, 88, 150-172. Carlson, W.D., & Johnson, C.D., 1991. Coronal reaction textures in garnet amphibolites of the Llano Uplift. American Mineralogist, 76, 756-772. 98 Carr, S.D. and Simony, P.S. 2006. Ductile thrusting vs. channel flow in the southeastern Canadian Cordillera: Evolution of a coherent crystalline thrust sheet. In Channel flow, ductile extrusion and exhumation of lower-mid-crust in continental collision zones. Edited by R. Law, M. Searle, and L. Godin. Geological Society, London, Special Publications Carr, S.D., 1992. Tectonic setting and U-Pb geochronology of the Early Tertiary Ladybird leucogranite suite, Thor-Odin – Pinnacles area, southern Omineca Belt, British Columbia. Tectonics, 11, 258-278. Coggon R, & Holland TJB 2002 Mixing properties of muscovite-celadonite- ferroceladonite-paragonite micas and revised garnet-phengite thermobarometers. Journal of Metamorphic Geology, 20,683–696 Dasgupta S., Sengupta P., Ehl J., & Raith M.M., 1999. Petrology of gedrite-bearing rocks in mid-crustal ductile shear zones from the Eastern Ghats Belt, India. Journal of Metamorphic Geology, 17, 765-778. Diener, J.FA., Powell, R., & White, R.W., 2008. Quantitative phase petrology of cordierite–orthoamphibole gneisses and related rocks. Journal of Metamorphic Geology, 25, 795-814. Diener, J.FA., Powell, R., White, R.W., & Holland, T.J.B., 2007. A new thermodynamic model for clino- and orthoamphiboles in the system Na2O–CaO–FeO–MgO– Al2O3–SiO2–H2O–O. Journal of Metamorphic Geology, 24, 631-656. 99 Evans, T.P., 2004. A method for calculating effective bulk composition modification due to crystal fractionation in garnet-bearing schist: implications for isopleth thermobarometry. Journal of Metamorphic Geology, 22, 547-557. Fischer, H., Schreyer, W., & Maresch W.V., 1999. Synthetic gedrite: a stable phase in the system MgO-Al2O3-SiO2-H2O (MASH) at 800˚C and 10 kbar water pressure, and the influence of FeNaCa impurities. Contributions to Mineralogy and Petrology, 136, 184-191. Fockenberg & Schreyer, 1994. Stability of yoderite in the absence and in the presence of quartz – an experimental study in the system MgO-Al2O3-Fe2O3-SiO2-H2O. Journal of Petrology, 35, 1341-1375. Ghent, E.D., Nicholls, J., Stout, M.Z., Rottenfusser, B., 1977. Clinopyroxene amphibolite boudins from Three Valley Gap, British Columbia. Canadian Mineralogist, 15, 269 – 282. Gordon SM, Whitney DL, Teyssier C, & Grove M, 2008. Timescales of migmatization, melt crystallization, and cooling in a Cordilleran gneiss dome: Valhalla complex, southeastern British Columbia. Tectonics, 27, TC4010. Guiraud, M., Powell, R., & Cottin, J.-Y., 1996. Hydration of orthopyroxene-cordierite- bearing assemblages at Laouni, Central Hoggar, Algeria. Journal of Metamorphic Geology, 14, 467-476. Guiraud, M., Powell, R. & Rebay, G., 2001. H2O in metamorphism and unexpected behaviour in the preservation of metamorphic assemblages. Journal of Metamorphic Geology, 19, 445–454. 100 Harvey, J.L., Hoisch, T.D., 1994. Sapphirine-bearing amphibolites in the Okanogan Complex, Washington: thermobarometry and tectonic implications. Abstracts with Programs—Geol. Soc. Am. 26, 57. Hinchey, A.M. & Carr, S.D., 2007. Protolith composition of cordierite-gedrite basement rocks and garnet amphibolite of the Bearpaw Lake are of the Thor-Odin Dome, Monashee Complex, British Columbia, Canada. The Canadian Mineralogist, 45, 607-629. Hinchey, A.M., Carr, S.D., McNeill, P.D., & Rayner, N., 2006. Paleocene–Eocene high- grade metamorphism, anatexis, and deformation in the Thor–Odin dome, Monashee complex, southeastern British Columbia. Canadian Journal of Earth Sciences, 43, 1341-1365. Holland, T. J. B. & Powell, R., 1998. An internally-consistent thermodynamic data set for phases of petrological interest. Journal of Metamorphic Geology, 16, 309–343. Holland, T. J. B. & Powell, R., 2006. Mineral activity–compo-sition relations and petrological calculations involving cation equipartition in multisite minerals: a logical inconsistency. Journal of Metamorphic Geology, 24, 851–861. Hölttä, P., 1997. Geochemical characteristics of granulite facies rocks in the Archean Varpaisjarvi area, central Fennoscandian Shield. Lithos, 40, 31-53. Hudson, N., & Harte, B., 1985. K (sub 2) O-poor, aluminous assemblages from the Buchan Dalradian, and the variety of orthoamphibole assemblages in aluminous bulk compositions in the amphibolite facies. American Journal of Science, 285, 224-266. 101 Johnston, D.H., Williams, P.F., Brown, R.L., Crowley, J.L., and Carr, S.D. 2000. Northeastward extrusion and extensional exhumation of crystalline rocks from the Monashee complex, southeastern Canadian Cordillera. Journal of Structural Geology, 22: 603-625. Koshimoto S, Tsunogae T, Santosh M., 2004.Sapphirine and corundum bearing ultrahigh temperature rocks from the northern domain of Palghat-Cauvery Shear System, southern India. Journal of Mineralogy & Petrology Science, 99,298–310. Kriegsman LM 2001 Partial melting, partial melt extraction and partial back reaction in anatectic migmatites. Lithos 56, 75–96. Kruckenberg, S.C., Whitney, D.L., Teyssier, C., Fanning, C.M., & Dunlap, W.J., 2008. Paleocene-Eocene migmatite crystallization, extension, and exhumation in the hinterland of the northern Cordillera: Okanogan dome, Washington, USA. Geological Society of America Bulletin, 120, 912-929. Kruse, S., McNeill, P. D. & Williams, P. F., 2004. A geological compilation map of the Thor-Odin dome. www.unb.ca/fredericton/science/geology/monashee. Labotka, T.C., & Kath, R.L., 2001. Petrogenesis of the contact-metamorphic rocks beneath the Stillwater Complex, Montana. Geological Society of America Bulletin, 113, 1312-1323. Lal, R. K. & Shukla, R. S., 1975. Genesis of cordierite–gedrite–cummingtonite rocks from the northern portion of the Khetri Copper Belt, Rajasthan, India. Lithos, 8, 175–186. 102 Lehnert, K., Su, Y., Langmuir, C., Sarbas, B., & Nohl, U. 2000. A global geochemical database structure for rocks. Geochem. Geophys. Geosyst. 1, doi:10.1029/1999GC00026 Marshall, D., & Simandl G., 2006, Phase relations and metamorphism in the sapphirine bearing granulites of the Valhalla complex, Slocan Valley, BC, Geol. Assoc. Can./Mineral. Assoc. Can., Annual Meet. Abstr. Program, 31, 96. McClintock, M.K., & Cooper, A.F., 2003. Geochemistry, mineralogy, and metamorphic history of kyanite-orthoamphibole-bearing Alpine Fault mylonite, South Westland, New Zealand. New Zealand Journal of Geology & Geophysics, 46, 47- 62. Moore, J.M., & Waters, D.J., 1990. Geochemistry and origin of cordierite- orthoamphibole/orthopyroxene-phlogopite rocks from Namaqualand, South Africa. Chemical Geology, 85, 77-100. Munz, I.A., 1990. Whiteschists and orthoamphibole-cordierite rocks and the P-T-t path of the Modum Complex, South Norway. Lithos, 24, 181-200. Norlander, B.H., Whitney, D.L., Teyssier, C. & Vanderhaeghe, O., 2002. Partial melting and decompression of the Thor-Odin dome, Shuswap metamorphic core complex, Canadian Cordillera. Lithos, 6, 103-125. Nyman, M. W., Pattison, D. R. M. & Ghent, E. D., 1995. Melt extraction during formation of K-feldspar+sillimanite migmatites, west of Revelstoke, British Columbia. Journal of Petrology, 36, 351-372. 103 Ouzegane, K, Djemai, S., & Guiraud, M., 1996. Gedrite-garnet-sillimanite-bearing granulites from Amessmessa area, south In Ouzzal, Hoggar, Algeria. Journal of Metamorphic Geology, 14, 739-753. Pan, Y., & Fleet, M.E., 1995. Geochemistry and origin of cordierite-orthoamphibole gneiss and associated rocks at an Archean volcanogenic massive sulphide camp: Manitouwadge, Ontario, Canada. Precambrian Research, 74, 73-79. Parrish, R.R., Carr, S.D., and Parkinson, D.L. 1988. Eocene extensional tectonics and geochronology of the southern Omineca Belt, British Columbia and Washington. Tectonics, 7, 181-212. Peck, W.H., & Smith, M.S., 2005. Cordierite-gedrite rocks from the Central Metasedimentary Belt boundary thrust zone (Grenville Province, Ontario): Mesoproterozoic metavolcanic rocks with affinities to the Composite Arc Belt. Canadian Journal of Earth Science, 42, 1815-1828. Peck, W.H., & Valley, J.W., 2000. Genesis of Cordierite-Gedrite Gneisses, Central Metasedimentary Belt boundary thrust zone, Grenville Province, Ontario, Canada. The Canadian Mineralogist, 38, 511-524. Powell, R. & Holland, T. J. B. & Worley, B., 1998. Calculating phase diagrams involving solid solutions via non-linear equations, with examples using THERMOCALC. Journal of Metamorphic Geology, 6, 173–204. Raith, M.M., Rakotondrazafy, r., & Sengupta, P., 2008. Petrology of corundum-spinel- sapphirine-anorthite rocks (sakenites) from the type locality in southern Madagascar. Journal of Metamorphic Geology, 25, 647-667. 104 Reesor J.E., and Moore Jr., J.M. 1971. Petrology andstructure of Thor-Odin gneiss dome, Shuswap metamorphic complex, British Columbia. Geological Survey of Canada, Paper 195, pp. 149. Rheinhardt, J., 1987. Cordierite-Anthophyllite rocks from north-west Queensland, Australia; metamorphosed magnesian pelites. Journal of Metamorphic Geology, 5, 451-472. Reesor, J.E. & Moore, J.M., Jr., 1971. Petrology and structure of the Thor-Odin Gneiss Dome, Shuswap Metamorphic complex, British Columbia. Geological Survey of Canada Bulletin, 195. Roberts, M. D., Oliver, N. H. S., Fairclough, M. C., Ho¨ltta¨, P. S. & Lahtinen, R., 2003. Geochemical and oxygen isotope signature of sea-floor alteration associated with a polydeformed and highly metamorphosed massive sulphide deposit, Ruostesuo, central Finland. Economic Geology, 98, 535–556. Schneiderman, J. & Tracy, R.J., 1991. Petrology of orthoamphibole-cordierite gneisses from the Orijarvi area, southwest Finland. Amercian Mineralogist, 76, 942-955. Schumacher, J.C., 1988. Stratigraphy and geochemistry of the ammonoosuc volcanics, central Massachusetts and southwestern New Hampshire. American Journal of Science, 288, 619-663. Shimpo M, Tsunogae T, & Santosh M 2006 First report of garnet-corundum rocks from Southern India: implications for prograde high-pressure (eclogite-facies?) metamorphism. Earth Planet Science Letters 242:111–129. 105 Teyssier, C. & Whitney, D.L., 2002. Gneiss domes and orogeny. Geology, 30, 1139- 1142. Teyssier, C., Ferré, E., Whitney, D.L., Norlander, B., Vanderhaeghe, O., and Parkinson, D. 2005. Flow of partially molten crust and origin of detachments during collapse of the Cordilleran orogen. Submitted to Geological Society of London, Symposium volume Channel flow, ductile extrusion and exhumation of lower- mid-crust in continental collision zones. In High-strain zones: Structures and Physical Properties. Edited by: Bruhn, D. and Burlini, L. Geological Society of London Special Publication, 245: 39-64. Thompson, A.B., 2001. Clockwise P – T paths for crustal melting and H2O recycling in granite source regions and migmatite terrains. Lithos, 56, 33-45. Vanderhaeghe, O. & Teyssier, C., 1997. Formation of the Shuswap metamorphic core complex during late-orogenic collapse of the Canadian Cordillera: Role of ductile thinning and partial melting of the mid- to lower-crust. Geodinamica Acta, 10, 42- 58. Vanderhaeghe, O., Teyssier, C. & Wysoczanski, R., 1999. Structural and geochronological constraints on the role of partial melting during the formation of the Shuswap metamorphic core complex at the latitude of the Thor-Odin Dome, British Columbia. Canadian Journal of Earth Sciences, 36, 917-943. Vanderhaeghe, O., Teyssier, C., McDougall, I., and Dunlap, D.W. 2003. Cooling and exhumation of the Shuswap metamorphic core complex constrainedby 40Ar/39Ar thermochronology. Geological Society of America Bulletin, 115: 200-216. 106 White R.W., Powell R., & Holland T.J.B., 2007. Progress relating to calculation of partial melting equilibria for metapelites. Journal of Metamorphic Geology, 25, 511-527. White, R.W., Powell, R., & Baldwin, J.A., 2008. Calculated phase equilibria involving chemical potentials to investigate the textural evolution of metamorphic rocks. Journal of Metamorphic Geology, 26, 181-198. White, R.W., Powell, R., & Clarke, G.L., 2002. The interpretation of reaction textures in Fe-rich metapelitic granulites of the Musgrave Block, central Australia; constraints from mineral equilibria calculations in the system K (sub 2) O-FeO-MgO-Al (sub 2) O (sub 3) -SiO (sub 2) -H (sub 2) O-TiO (sub 2) -Fe (sub 2) O (sub 3). Journal of Metamorphic Geology, 20, 41-55. Williams, P.F., & Jiang, D. 2005. An investigation of lower crustal deformation: evidence for channel flow and its implications for tectonics and structural studies. Journal of Structural Geology, 27: 1486-1504. 107 FIGURE CAPTIONS Figure 3.1. Regional map of the Shuswap metamorphic core complex (modified from Parish et al., 1988). Figure 3.2. Simplified geologic map of Thor-Odin Dome modified after Kruse et al. (2006). Sample locations important to this study are Saturday Glacier, Bear Paw Lake, and Kelly Peak, located in the southern portion of the dome. Figure 3.3. Photograph of reaction textures associated with garnet and Al2SiO5 as observed in outcrop. Mineral abbreviations used in this paper: Gt-garnet, Sil-Sillimanite, Ky-kyanite, Q-quartz, Oam-orthoamphibole (general), Ged-gedrite-rich orthoamphibole, Anth-anthophyllite-rich orthoamphibole, Bi-biotite, Cd-cordierite, Sp-spinel, Spr- sapphirine, Pl-plagioclase, Crn-corundum, Ru-rutile, Ilm-ilmenite, Hem-hematite. Figure 3.4. (a) Photograph showing the typical field relationship between concordant OCG pods and host migmatite. A biotite-rich zone developed between the pod and migmatite. (b) Cd + Q vein that crosscuts an OCG pod. 108 Figure 3.5. Full thin-section scans from (a,b) Saturday Glacier (06ET-2B, 02-3.03) and (c) Bear Paw Lake. The modal amount of biotite is consistently higher in the Saturday Glacier region. Figure 3.6. Photomicrographs of the typical replacement texture of cordierite after orthoamphibole observed in OCG of this study. a) Cordierite commonly replaces Oam along cleavage planes. b) where this replacement is more evolved, an intergrowth texture with optically continuous fragments of Oam develops. All Oam fragments are optically continuous across the entire intergrowth texture (Circled region). Figure 3.7. Compositional variation of OCG from Saturday Glacier area (a) Orthoamphibole end-member composition. All samples contain both anthophyllite and gedrite. Sample 01-02.07 contains a complete range from gedrite to anthophyllite, which are intergrown. (b) Edenite substitution vector. (c) Mg-tschermaks exchange vector. Figure 3.8. Photomicrograph illustrating the color change of biotite associated with garnet (green) and symplectitic reaction textures (brown). Figure 3.9. Photomicrographs of garnet textures in OCG. (a) subhedral resorbed (02- 3.03). (b) euhedral (01-2.07). (c) subhedral resorbed garnet showing relationship to an adjacent Al2SiO5 reaction texture (symp = symplectite), and (d) Partial thin section scan illustrating interaction between coronas centered on garnet and Al2SiO5. 109 Figure 3.10. Garnet zoning. (a) garnet surrounded by well-developed corona structures commonly displays a complex rim (sample 06ET-2). The traverse is from a garnet surrounded with a symplectitic corona of biotite + cordierite reaction texture replacing the garnet. (b) garnet traverse from the least altered sample (01-2.07), illustrating complex core composition variation related to inclusions and fractures. Figure 3.11. Photomicrographs from one thin section (06ET-2B) illustrating textural heterogeneity associated with symplectitic reaction textures. Figure 3.12. Sapphirine compositional variability Figure 3.13. Backscattered electron image of a typical symplectite associated with Al2SiO5. Plagioclase appears in many samples as complex flame structures with no consistent thickness or apparent relationship with surrounding matrix phase compositions or assemblages. Figure 3.14. Ternary and tetrahedral plots illustrating major element variation in mafic volcanic and Al-Mg rich rocks. A=Al2O3-(Na2O+K2O); F= FeO-6(1-Mg#)K2O; M=MgO-6(Mg#)K2O; modified after Rheinhardt (1987). All plots show a distinctive variation of both Mg-Al rich rocks and gedrite-cordierite rocks from the literature. (a) AFM diagram; gedrite-cordierite rocks and Al-Mg rich rocks plot with slightly more 110 variation than the Altered and Fresh mafic volcanics. (b) AFMCaO tetrahedral plot illustrating the strong depletion in CaO in the Mg-Al rich and gedrite-crd samples. The depletion is similar to the totally altered mafic volcanics and serpentinites, but with much more enrichment in Al2O3. Data for non Mg-Al rich compositions and names of mafic rock types are from PetDB http://www.petdb.org. Figure 3.15. (a) Pseudosection for OCGs Thor-Odin dome assuming H2O in excess. Hornblende is the modally dominant amphibole across all P-T ranges except at low-P, high-T. The strongly curved phase boundary bounding the hb oam bi chl gt ilm ru q phase field is caused by a local low (bulls-eye shaped) in the predicted Al content of orthoamphibole toward the anthonphyllite end-member in this region. The high-pressure behavior of the Oam solvus has not been explored in detail, however the Oam compositions in this area approach anthophyllite but are still compositionally distinct and locatd on the gedrite side of the solvus. Oam compositions in this region of the phase diagram return to gedrite-rich compositions going in all directions from this phase field. Some divariant assemblages associated with the Cd-in phase field between 5 and 6.5 kbar are not labeled for readability. (b) plot of mode % versus pressure at 750°C (at the same P scale as the pseudosection for direct comparison). Modal amounts of phases change most drastically in the P range considered on a decompression path at the cordierite-in phase boundary. 111 Figure 3.16. T-MH2O pseudosection for OCG using the water-saturated composition from Fig. 2.15. Rocks of these bulk compositions require high molar water contents to saturate the system across all temperatures and pressures. Clino- and orthoamphibole destabilize at very similar MH2O conditions at the pressure considered in this diagram. Opx becomes the modally dominant Mg-bearing phase on the H2O-depleted side of the diagram. Figure 3.17. Schematic P-T diagram summarizing the stability of hornblende with changing effective bulk composition. The shaded regions represent areas lacking Hb as a stable phase, starting from the limits of Hb stability in Fig. 2.15 (the darkest shaded area) to the composition with the lowest water content of 0.94 mol % (Table 2.8). The thin dashed line represents Opx stability in Fig. 2.15. The thick dashed line is the high- pressure limit of stability of Opx from the composition with the lowest H2O content. The high-pressure stability of Opx is strongly controlled by the entrance of Q into the assemblage (dash-dot line). Although H2O is a major control on Hb stability, CaO controls the curvature of the Hb-out reaction. Figure 3.18. Schematic illustrating the relative timing of bulk compositional modification and reaction texture initiation. Prior to decompression the phase assemblage was modally dominated by orthoamphibole. The Crd-in phase field in dry OCGs is at relatively high (>6-8.5 kbar) pressures (at T’s ranging from 650-800°C), therefore the first appearance of cordierite likely followed shortly after decompression was initiated in 112 these rocks. Crossing the crd-in phase field also coincides with the departure from equilibrium of reaction textures (around garnet and Al2SiO5). The ambiguity of the relative timing of biotite replacement suggests that the initial stages of bulk compositional modification coincided with the early stages of decompression and reached maximum extent after coronal textures had fully formed. Continued coronal development and influx of fluids may have led to the texturally heterogeneous nature of the reaction textures in the OCG rocks. Ta bl e 3 .1. M in er al as se m bl ag es fo r o rth oa m ph ib ol e-c or di er ite gn eis se s f ro m T ho r-O di n Do m e Bu lk ro ck as se m bl ag e Sa m pl e Ga rn et Ge dr ite Ky an ite * Sil lim an ite * Bi ot ite Co rd ier ite Ilm en ite Ru til e Ap ati te 06 ET -2         05 -5 .24          02 -3 .03         01 -2 .07 **        Al 2Si O 5 D om ain A ss em bl ag es Sa m pl e Sp in el + Co rd ier ite Sa pp hi rin e + Co rd ier ite Sp in el + An or th ite Co ru nd um + Co rd ier ite Ilm en ite M ag ne tit e Ru til e 06 ET -2      05 -5 .24      02 -3 .03       01 -2 .07 N /A N /A N /A N /A N /A N /A N /A *W he n bo th p re se nt , k ya ni te an d sil lim an ite ap pe ar in p ar tia lly tr an sfo rm ed A l 2S iO 5 p or ph yr ob las ts an d ne ve r a s d ist in ct ph as es **0 1- 2.0 7 c on tai ns qu ar tz an d pl ag io cla se in th e b ul k r oc k a ss em bl ag e. Th es e p ha se s a re p re se nt as in clu sio ns in ga rn et in al l s am pl es 113 Table 3.2. Representative microprobe analyses of orthoamphibole Sample 02-3.03 (Core) 02-3.03 (Rim) 01-2.07** (Host Oam) 01-2.07** (Oam lamellae) 06ET-02 (Core) 06ET-02 (Rim) SiO2 46.75 54.10 46.49 51.02 45.72 54.53 TiO2 0.19 0.05 0.49 0.21 0.20 0.07 Al2O3 13.70 2.19 13.97 7.76 15.56 4.50 FeO 15.34 18.55 18.78 18.92 14.43 16.04 MnO 0.17 0.37 0.23 0.19 0.13 0.14 MgO 19.65 21.44 16.91 19.61 19.27 22.39 CaO 0.59 0.30 0.69 0.46 0.51 0.45 Na2O 1.24 0.12 1.37 0.59 1.52 0.46 K2O 0.01 0.01 0.01 0.01 0.01 0.00 Total 97.08 97.12 99.00 98.76 97.36 98.58 Normalized to 23 Oxygens Si 6.62 7.74 6.62 7.22 6.48 7.60 AlIV 1.38 0.26 1.38 0.78 1.52 0.40 AlVI 0.91 0.11 0.96 0.51 1.09 0.34 Ti 0.02 0.01 0.05 0.02 0.02 0.01 Fe3+* 0.09 0.10 0.00 0.06 0.00 0.00 Fe2+ 1.72 2.12 2.24 2.18 1.71 1.87 Mn 0.02 0.04 0.03 0.02 0.02 0.02 Mg 4.15 4.57 3.59 4.13 4.07 4.65 Ca 0.09 0.05 0.10 0.07 0.08 0.07 Na 0.34 0.03 0.38 0.16 0.42 0.12 K 0.00 0.00 0.00 0.00 0.00 0.00 XMg 0.70 0.67 0.62 0.66 0.70 0.71 * Fe3+ Calulations were carried out using the Σ15 scheme from Spear (1993) **Oam end-member compositions are present as Anthophyllite lamellae within Gedrite-rich hosts without systematic zoning as seen in other samples Cr2O3 was analyzed, but below detection limits in OAM 114 Table 3.3. Representative microprobe analyses of biotite Biotite Green Bi Brown Bi 02-3.03 SiO2 37.07 37.55 37.43 TiO2 0.88 1.77 0.46 Al2O3 21.00 19.46 19.75 FeO 10.30 11.68 13.22 MnO 0.08 0.02 0.06 MgO 17.72 16.52 15.99 CaO 0.00 0.00 0.02 Na2O 0.55 0.74 0.52 K2O 8.68 8.23 8.19 F 0.00 0.00 0.09 Cl 0.00 0.06 0.06 Total 96.29 96.03 95.78 Normalized to 22 Oxygens Si 5.30 5.41 5.45 AlIV 2.70 2.59 2.55 AlVI 0.85 0.72 0.83 Ti 0.10 0.19 0.05 Fe* 1.23 1.41 1.61 Mn 0.01 0.00 0.01 Mg 3.78 3.55 3.47 Ca 0.00 0.00 0.00 Na 0.15 0.21 0.15 K 1.58 1.51 1.52 F 0.00 0.00 0.04 Cl 0.00 0.02 0.01 *reported as all Fe2+ 06ET-2 115 Table 3.4. Representative microprobe analyses of garnet Sample 02-3.03 (Core) 02-3.03 (Rim) 01-2.07 (Core) 01-2.07 (Rim) 06ET-2 (Core) 06ET-2 (Rim) SiO2 38.69 38.86 38.16 39.09 38.47 38.76 TiO2 0.02 0.00 0.02 0.00 0.01 0.00 Al2O3 22.41 22.57 22.33 22.13 22.58 22.83 FeO 27.64 27.08 28.02 26.24 26.48 26.54 MnO 0.62 0.20 1.11 0.51 1.74 1.45 MgO 6.92 8.01 5.88 9.31 6.16 8.25 CaO 4.69 4.19 4.98 2.89 4.99 2.50 Total 100.98 100.92 100.49 100.16 100.41 100.33 Normalized to 12 Oxygens Si 2.97 2.97 2.98 2.99 2.98 2.98 Al 2.03 2.03 2.03 2.00 2.06 2.07 Ti 0.00 0.00 0.00 0.00 0.00 0.00 Fe3+ 0.02 0.03 0.01 0.02 0.00 0.00 Fe2+ 1.76 1.70 1.81 1.66 1.74 1.73 Mn 0.04 0.01 0.08 0.03 0.11 0.09 Mg 0.79 0.91 0.69 1.06 0.71 0.94 Ca 0.39 0.34 0.41 0.24 0.41 0.21 XMg* 0.31 0.35 0.28 0.39 0.29 0.35 XAlm 0.59 0.57 0.61 0.56 0.58 0.58 XSps 0.01 0.00 0.03 0.01 0.04 0.03 XPyp 0.27 0.31 0.23 0.35 0.24 0.32 XGrs 0.13 0.12 0.14 0.08 0.14 0.07 XMg = Mg/(Mg+Fe) Cr2O3 was analyzed but below detection limits 116 Table 3. 5. Representative microprobe analyses of cordierite, spinel, sapphirine, plagioclase Cordierite Cordierite Spinel Spinel Sapphirine Plagioclase Plagioclase Textural placement After Orthoamphibole Coronal after Al2SiO5 Symplectitic w/ Pl Symplectitic w/ Pl Symplectitic w/ Crd In corona after Al2SiO5 After Grt in 01- 2.07 SiO2 wt% 48.77 48.12 0.57 0.20 10.60 44.06 56.15 TiO2 0.00 0.01 0.95 0.05 0.06 - - Al2O3 34.15 34.46 62.88 65.08 66.66 36.59 28.14 FeO 3.57 4.94 22.34 22.79 7.17 0.37 0.33 MnO 0.06 0.11 0.05 0.06 0.05 - - MgO 11.61 10.71 11.77 12.36 14.48 - - CaO 0.01 0.11 - - 0.02 19.30 9.98 Na2O 0.16 0.05 - - 0.00 0.52 5.99 K2O 0.00 0.01 - - 0.00 0.01 0.05 Cr2O3 700 C, and typically > 750-800°C) (e.g., Schenk, 1983; Harley, 1985, 1986; Droop, 1989; Harley et al., 1990; Carlson & Johnson, 1991; Messiga & Bettini, 1990; Clarke, 1991; Davies & Warren, 1992; Elvevold, 2000; Patel, 2001; Johnson et al., 2004), and in mafic plutonic complexes that were emplaced at shallow crustal levels and experienced rapid cooling (Mongoltip & Ashworth, 1983; Ashworth, 1992; Turner, 1992; Audibert, 1993; Grantham et al., 1993; Ashworth et al., 1998; Ashworth & Chambers, 2000). Although decompression is strongly implicated in the formation of symplectites and layered coronal structures in many granulites and eclogites, other P-T paths, including nearly isobaric heating or cooling, are possible (Pitra & De Waal, 2001; Brown, 2002). Bulk compositional change at constant or evolving P-T conditions can also induce the formation of coronal textures (e.g. Joesten, 1986). In all cases, coronal structures develop in a system that has experienced perturbation (typically interpreted as rapid) from equilibrium in terms of one or more of the intensive variables acting on the system, i.e. pressure, temperature or composition (via changes in the conjugate intensive variable µ, chemical potential). As a result of coronal textures being representative of disequilibrium, interpreting the conditions or mechanisms of their formation cannot be addressed through traditional 143 phase-equilibria techniques, i.e. equilibrium phase-diagrams, geothermobarometry (cf. White et al., 2008). Essentially, once a finite thickness of the product phase(s) has developed and separated the original reactants, the reactive system leaves equilibrium and further growth is kinetically driven, necessitating an evaluation utilizing non-equilibrium thermodynamics. The development of methods to investigate non-equilibrium processes in metamorphic rocks can be divided into two approaches: forward modeling, and physical modeling (or inverse modeling), after the terminology of White et al. (2008). The fundamental basis for both styles of modeling is the assumption that reaction rates leading to corona formation are limited by diffusion, which is typically assumed to be grain boundary diffusion. Forward modeling entails the quantitative investigation of reaction textures through the utilization of chemical potential (µa-µb) space, where µa and µb represent the chemical potentials of the components controlling the phase-assemblages observed. This approach was pioneered by Korzhinskii (1959) and Thompson (1959), who presented a qualitative construct to understand the development of layered structures in hydrothermal systems. The forward modeling approach is also advocated by White et al. (2008), who assumed that current a-x models in large chemical systems are representative and thus can be used to interpret phase assemblage and coronal evolution to place quantitative constraints on the chemical potentials controlling the system. However, the choice of chemical potentials in this analysis is based on assumptions regarding what components were perfectly mobile (chemical potential gradient exists on a scale larger than the texture itself and any gradients that arise during reaction are 144 eradicated or rapidly adjusted), mobile (the components exerting control on the phase assemblage(s) developed, which have intermediate diffusivities or inert (components that have mobilities that are static, either relatively or in an absolute sense) (Korzhinskii, 1959). Determination of the identity of these different categories of components is not always straightforward and can be difficult to assess from the textures themselves (Brady, 1975). The physical modeling approach utilizes the observed assemblages, their compositions, and the thicknesses and assemblages in successive layers preserved in the coronal structure. The thermodynamic and kinetic basis for this technique was developed by Fisher (1975) and modified for the study of layered reaction textures by Joesten (1977). Physical modeling is based on the assumption that the layers observed in the texture existed from first initiation and then grew in both directions at layer boundaries (Fig. 4.1), commensurate with the relative diffusivities of components in the system (Joesten, 1977). The phases in each layer, their compositions, layer thicknesses, and the stoichiometry of reactions occurring at each layer boundary are ultimately controlled by the chemical potential gradients acting on the system. This approach has been successfully applied to a number of coronal textures (e.g. Grant, 1988; Carlson & Johnson 1991; Ashworth et al., 1998). However, similar to the forward modeling approach, the physical modeling approach is based on assumptions that can be difficult to assess. To apply this method, one of the following conditions must be met: a) at least one component is immobile, b) the location of the original interface is known, or c) the system evolved with constant volume. Additionally, this technique cannot be used if the 145 reactants providing the components diffusing into the system changed during corona formation (Carlson & Johnson, 1991). Although both of these approaches can be used to model the compositions and modal amounts of phases in both layered coronas and symplectitic coronas, neither technique predicts the morphology of phases (i.e. polygonal or symplectitic). Layered and symplectitic coronas share commonalities in their formation; the kinetics governing their formation are very similar, but little is known in geological systems of what controls the morphology of phases forming during reaction. In addition, coronal reaction textures in general and symplectitic coronas, in particular, can preserve a spatially inconsistent variation in thickness of layers and/or a series of different phase assemblages, further complicating the interpretation of their formation. Inconsistencies such as these suggest that while the fundamentals of the processes controlling the development of layered and symplectitic coronas are known, there still remain questions regarding what variables control the development of layered versus symplectitic coronas. In addition, it is clear that the scale of diffusive flux in systems can exist at or evolve to different length-scales during corona formation, and this may explain the textural complexities that lead to difficulties in properly interpreting the major fluxes governing texture initiation and evolution. The purpose of this paper is to present a comprehensive textural characterization dataset in order to develop a semi-quantitative assessment of the variables controlling the formation and textural evolution of texturally heterogeneous symplectitic reaction textures. In the specific case considered, reaction textures are developed around Al2SiO5 146 in orthoamphibole-cordierite gneisses (OCG) from the Thor-Odin Dome British Columbia, Canada (Fig. 4.2). Data will be presented from compositional and modal analysis of textures to assess the major fluxes moving into and out of the reacting zone. High-resolution X-Ray computed tomography (HRXCT) data are presented to discuss the 3D arrangement of textures within a rock volume and investigate changing reactants and associated length-scales of diffusion controlling corona texture evolution. Finally, electron backscattered diffraction (EBSD) analysis of corona phases is used to investigate crystallographic relationships between products and reactants, and provide a basis to evaluate possible crystallographic controls of the development of symplectitic morphology. These data will be integrated with previous work on the evolution of the OCG rocks from the Thor-Odin dome to place the development and evolution of symplectitic textures into a tectonic context. GEOLOGIC SETTING The Thor-Odin gneiss dome represents one in a series of migmatite-cored domes located in the Shuswap metamorphic core complex (Fig. 4.2a). The lowest structural level of the Thor-Odin dome is composed of high-grade (sillimanite-K-feldspar zone) migmatitic ortho- and paragneiss that are overlain by a sequence of metasupracrustal rocks. Maximum metamorphic conditions occurred in the migmatitic core of the dome (800˚C at > 10 kbar; Norlander et al., 2002), whereas the overlying metasupracrustal rocks experienced maximum conditions of ~650˚C - 800˚C at 7-9 kbar (Ghent et al., 1977; 147 Nyman et al., 1995). High-grade metamorphism occurred beginning at 56 Ma and was coeval with migmatization (Parrish et al., 1988; Vanderhaeghe et al., 2003; Hinchey et al., 2006). Zircon geochronology data suggests that melting was occurred over a relatively short time-span (lasting for 2-3 Ma, Hinchey et al. 2006) and may have been episodic, as has been observed in other migmatite domes in the Shuswap metamorphic core complex (e.g. Valhalla dome: Gordon et al., 2008; Okonogan dome: Kruckenberg et al., 2008). High-grade metamorphism was followed by isothermal decompression (Norlander et al., 2002), but the magnitude of decompression is difficult to constrain (Chapter 2). Although there exists some disagreement in the literature with respect to the tectonic evolution of the Thor-Odin dome (cf. Vanderhaeghe et al., 1999; Brown, 2004; Teyssier et al., 2005; Williams & Jiang, 2005), the majority of studies suggest that Cordilleran high-grade metamorphism and ductile deformation began in the Late Mesozoic and continued into the Paleogene-Eocene. Crustal melting, extensional collapse and development of Thor-Odin dome occurred in the late Eocene (Parrish et al., 1988; Vanderhaeghe et al., 2003; Johnston et al., 2000; Carr & Simony, 2006; Teyssier et al., 2005; Hinchey et al. 2006), coeval with these processes in other migmatite domes of the Omineca belt (Gordon et al., 2008; Kruckenberg et al., 2008). The migmatites of the Thor-Odin dome contain a variety of interlayered lithologies, including calc-silicate, mafic gneiss (amphibolites and OCG), and metapelitic units. Samples described in this paper are OCG from the Saturday Glacier area of the dome (Fig. 4.2b). OCG from this locality are typically pods that are concordant with the 148 surrounding gneissic fabric preserved in the host migmatites. The OCG pods have been variably affected by a hydrothermal alteration event related to the crystallization of the surrounding migmatite (Chapter 2), as indicated by the replacement of orthoamphibole by biotite. MINERAL CHEMISTRY AND DESCRIPTION OF TEXTURES Orthoamphibole-cordierite gneiss is present throughout the Thor-Odin dome. All OCG contain symplectitic coronas centered on Al2SiO5 and, in the Saturday Glacier region (Fig. 4.2b), also on garnet. The Saturday Glacier region is of particular interest because, in addition to containing spectacular examples of symplectitic coronas, it preserves evidence for mass transport at multiple scales: at the meter-scale between migmatites and OCG, and at the cm-scale between symplectitic coronas centered on Al2SiO5 and garnet (Chapter 2). Mass redistribution at multiple scales is recorded in the textural heterogeneity of reaction textures and in the evidence for chemical interaction of coronal reaction textures associated with both Al2SiO5 and garnet. The phase assemblage interpreted to be stable in orthoamphibole-cordierite gneisses prior to decompression of the Thor-Odin dome was Gt + Oam + Ky + Bi(minor) + Q ± Pl. OCG are also characterized by the prevalence of accessory phases apatite, rutile, ilmenite, zircon and monazite. The OCG of the Saturday Glacier region have been variably affected by a hydrothermal-style back reaction with the surrounding migmatite (Chapter 2). The most important features of the textural associations and mineral compositions of the symplectitic coronas and associated phases of OCG from the Thor- 149 Odin dome, as well as description of the textural and morphological heterogeneity of the symplectitic coronas, are presented in the following sections. Mineral compositions and qualitative X-ray maps were obtained using a JEOL JXA-8900 electron microprobe in the Department of Geology and Geophysics at the University of Minnesota. Quantitative analysis operating conditions were a 15 keV accelerating voltage, a 20 nA beam current, and a range of beam diameters (focused for minerals without volatile elements, 5 mm for minerals with volatile elements). X-ray maps were acquired using a 35 nA beam current and a beam diameter of 1 mm. Natural standards were used, and matrix correction was achieved using the ZAF correction routine. Mineral analyzed were cordierite, orthoamphibole, biotite, garnet, spinel, plagioclase, and sapphirine. CORONAL TEXTURES CENTERED ON AL2SIO5 Symplectitic coronal textures associated with Al2SiO5 are present in all OCG from the Saturday Glacier region. Coronal textures occur in OCG in which the matrix is dominated by orthoamphibole or biotite (Fig. 4.3), and there is no difference in coronal phase assemblages or compositions in orthoamphibole vs. biotite-dominated OCG. Coronal textures after Al2SiO5 are defined by a moat of polygonal cordierite surrounding one to three layers of different two-phase symplectitic assemblages (for this discussion the vermicular phase will be listed first, and the “host” phase will be listed second). The most common symplectitic assemblages are Sp+Cd, and Sp+Pl(An95). 150 When present as successive layers, Sp+Cd is always outboard (relative to Al2SiO5) of Sp+Pl (Fig. 4.4). Crn+Cd occurs in many textures, but is not as common as Cd or Pl + Sp assemblages (Fig. 4.3-4.6). The Crn+Cd association is observed either as the outermost layer in the texture (where Crn occurs as relatively large block shaped crystals) or directly replacing the central aluminosilicate (Figs. 4.3a-b, 6c) (where Crn has developed a vermicular structure). In very rare cases, the assemblage Crn+Pl occurs (Fig. 4a) but it is unclear if this was a primary association or a result of later modification of the texture. Spr+Cd assemblages are rare, but when present form as the outermost layer of the symplectitic portion of the corona (Fig. 4.3a). Phase compositions within the symplectitic corona textures are fairly homogeneous within and between samples (Table 4.1). Orthoamphibole is zoned from gedrite cores to anthophyllite rims. Biotite adjacent to the symplectitic coronas is mostly homogeneous (Table 4.1), but there is compositional variation in some samples related to whether biotite is close to garnet (green biotite) or reaction textures (brown). Spinel (XMg = 0.48) and plagioclase (An95) are unzoned. In addition, spinel compositions are the same whether spinel is intergrown with cordierite or plagioclase (Table 4.1). Cordierite is the only coronal phase that exhibits compositional zonation, and that the zoning is very slight (Fig. 4.7). Fe and Mg co-vary across the cordierite moat; Mg increases and Fe decreases from the outer boundary of the reaction texture to the symplectitic assemblage (Fig. 4.8a-c). Although compositional zoning is weak (variation typically ranges from 0.74-0.77 XMg), it is observed in all samples. Relative Mg enrichment is largest where the cordierite moat is in direct contact with Sp+Pl 151 symplectites (Fig. 4.7b). Where cordierite is present in the symplectitic assemblage, the composition varies around an average cordierite composition (XMg ~0.77 ± 0.5; Fig. 4.8c)/ Non-systematic and relatively large compositional swings may be an artifact related to the difficulty of analyzing a single phase in a fine-grained complex intergrowth. CORONAL TEXTURES CENTERED ON GARNET Reaction textures associated with garnet occur in all analyzed samples from the Saturday Glacier region. Unlike the reaction textures associated with Al2SiO5, reaction textures associated with garnet do not display complex textural and phase assemblage heterogeneity. The coronal textures consist of a two-phase intergrowth of Bi+Cd. The reaction texture is defined by randomly oriented biotite that is finer grained than matrix biotite and intergrown with polygonal cordierite (Fig. 4.8). Plagioclase is rare in the garnet coronal textures, and is a minor phase where present. The amount of garnet replacement is difficult to assess in some samples, although there are several examples where the shape and size of the original garnet can be inferred from the extent of coronal development (Fig. 4.8). There are many examples of interaction between reaction textures around garnet and Al2SiO5, either from close spatial relationships (Fig. 4.9a) or as a result of the intersection and apparent interconnectivity of the reaction textures (Fig. 4.9b). Some garnet porphyroblasts exhibit slight zoning at the rims and others exhibit complex asymmetric zoning profiles (Chapter 2). Cordierite in the garnet-centered 152 coronas is not zoned (XMg= 0.75; Table 4.1). Biotite in the symplectitic coronas after garnet ranges from green to, less commonly, brown in color, corresponding to zoning with respect to Ti and Fe (Table 4.1). Where zoning is present, biotite is most Ti-rich in contact with garnet. 2D MORPHOLOGY AND QUANTITATIVE TEXTURAL ASSESSMENT OF SYMPLECTITIC CORONAS A striking characteristic of all symplectitic coronas centered on Al2SiO5 in the Thor-Odin OCG is the preservation of the shape of the central Al2SiO5 porphyroblast (Figs. 3, 5). Both the cordierite moat and, in most cases, the outline of the symplectitic intergrowths, regardless of phase assemblage, closely mimic the present shape of the central porphyroblast. Imposed on this observed textural consistency within and between samples is the heterogeneous nature of the spatial distribution of the two-phase assemblages. Although there is some consistency in the observed sequence of two-phase assemblages within the reaction band (i.e. the common observation of the layer sequence Cd|Crn+Cd|Sp+Cd|Sp+Pl from matrix to the Al2SiO5 interface) the relative areal extent and spatial association of each of these two-phase assemblages can vary widely. For example, the entire thickness of the symplectitic portion of the texture may consist entirely of Crn+Cd in one location, but the two-phase assemblage that occupies the symplectitic layer changes around the perimeter of the corona to Sp+Cd (e.g. Fig. 4.5b, c) and then to Sp+ Pl. There is no evidence for a systematic textural relationship between 153 the surrounding matrix phases and the presence or absence of one or more of the two- phase assemblages in the symplectitic layer of the reaction texture. Similarly, there are wide ranges of the total thickness of the corona and the relative thicknesses of the two-phase assemblages. Total coronal thicknesses range from 500 µm to >1 mm. Whereas total coronal thicknesses vary, the thickness of the cordierite moat (mean thickness is 120 µm ± 30 µm) is nearly constant within and between samples. As a result of the heterogeneity of the symplectitic layers and associated assemblages, it is difficult to determine average thicknesses, as the standard deviation is commonly on the same order of the average measurements (Table 4.2). Modal amounts of the two-phase symplectitic assemblages are remarkably consistent within and between samples. Modal percentages for minerals in each sample were calculated using ImageJ, an open-source image analysis software package freely available via the National Institutes of Health (Table 4.2). Sp+Crd modal distribution is consistently ~40 & 60 mode % respectively. The Sp+Pl layers also show a 40% Sp to 60% host phase relationship. Crn+Crd is consistently 30 and 70 mode percent respectively. The modal distributions in the phase assemblages containing Sp are consistent regardless of the grain size or shape of Sp vermicules. Vermicular spinel, and to a lesser extent, corundum display two different grain shapes in thin section. Spinel displays either the vermicular structure commonly observed in symplectitic textures (Fig. 4.10) or occurs as elongate (lamellar) single crystals (Fig. 4.6b). Spinel vermicules are also commonly interconnected (Fig. 4.10). Corundum is observed as either complex block-shaped grains, or as vermicular intergrowths (Fig. 4.5b- 154 c). Symplectitic layers also display evidence for textural maturation during or subsequent to reaction texture formation. Grain coarsening is evident in both spinel and corundum. Grain thicknesses of both vermicular and ‘lamellar’ spinel increase away from the central Al2SiO5 (Fig. 4.6b & 4.10). 3D MORPHOLOGY (HRXCT) The above petrographic analysis suggests interconnectivity of coronal textures associated with garnet and Al2SiO5. This analysis also provided a two-dimensional view of the textural heterogeneity within symplectitic reaction textures after Al2SiO5. HRXCT data was collected to investigate the three-dimensional textural relationships between and within coronal reaction textures preserved in OCGs. Three-dimensional imaging on OCG rocks was conducted at the HRXCT facility at the Department of Geological Sciences at the University of Texas-Austin. The ultra-high-resolution (UHR) subsystem using a 225- Kv microfocal source (maximum spatial resolution of 5µm) was used for analysis of mm- scale samples of single symplectite textures and the High-energy subsystem using a 450- Kv tungsten X-ray source was used to obtain 3D data on larger samples of OCGs. HRXCT data was collected on three samples (02-2.03, 06ET-2B, and 06ET-2H). Ultra- high-resolution data was collected on single reaction textures cut from 02-2.03 and 06ET- 2B as well as a three-inch long by one inch diameter core taken from 06ET-2B to investigate corona- and intermediate-scale textural relationships. Hand sample sized 155 specimens of 02-2.03 and 06ET-2B were scanned in the High-energy system to characterize bulk-rock textures. Data collected from HRXCT analysis is in the form of a series of grey-scale images. The grey-scale distribution approximately correlates with density (Denison et al., 1997). As a result, large discontinuities in the density of phases present in a rock volume must be present to efficiently visualize the phases individually. Garnet and other relatively large and high-density phases, are typically easy to indentify and therefore their three-dimensional characteristics (e.g. spatial distribution, size, shape etc..) can be quantified (e.g. Hirsch et al., 2000; Whitney et al., 2008). Even when using the UHR system on the smallest possible diameter samples, imaging the three-dimensional structures of spinel vermicules within the symplectitic assemblages was not possible. The spinel vermicules range in size from a few to tens of microns in diameter, putting them at the edge of the spatial resolution of the technique. This fact, along with the relatively high-density of spinel obscured the fine-detail of the morphologies due to beam hardening (Denison et al., 1997). These two factors also had the added affect of obscuring phase contrast between whether spinel was intergrown with either cordierite or plagioclase. However, general details related to the total thickness and textural heterogeneity of samples can still be determined. Symplectitic reaction textures are tabular in three-dimensions. The thickness of the cordierite moat is extremely consistent (Fig. 4.11a-f) along the long-axis of the texture. As observed petrographically. Again, in keeping with petrographic observations, the heterogeneity in thickness and identity of the two-phase assemblages is variable with 156 respect to Crn+Cd and Sp+Cd assemblages in three-dimensions (Sp+Pl cannot be differentiated in the HRXCT data). Hand-sample sized specimens also had details related to density, which precluded consistent three-dimensional visualization. However, the problems at this scale were related to the lack of sufficient phase contrast, rather than the effect of high-density phases obscuring textural information as was the case above. Garnet can be easily identified, however, the matrix and coronal assemblages have overlapping grey-scale values, making them difficult to identify, and thus visualize, using available software. The purpose of collecting data at this scale was to indentify any textural relationships between coronal reaction textures. Petrographically, this relationship is defined dominantly by cordierite (e.g. Fig. 4.9b). In an attempt to overcome the phase contrast issues, data were thresholded to remove grey scale values that were in between garnet and cordierite as these had the highest and lowest grey scale values respectively. The agreement between thresholded values and grey scale images was good however visualizations of the 3D cordierite structures connecting garnet and Al2SiO5 coronas proved to difficult to produce. However, a representative series of slices of HXRCT data illustrate the interconnectivity of these reaction textures (Fig. 4.12a-d). CRYSTALLOGRAPHIC REALTIONSHIPS IN SYMPLECTITIC CORONAS 157 BACKGROUND AND MOTIVATION A defining characteristic of symplectitic coronas is the complex vermicular morphology of the intergrown phases, in contrast to the polygonal character of mono- and polyphase coronas (e.g. Ashworth and Birdi, 1990; Alverez-Velaro et al., 2007). Coronas form as a consequence of a rock volume attempting to adjust to a new set of conditions and achieve equilibrium by minimizing the energy of the system (typically via diffusion and elimination of chemical potential gradients in the case of coronal structures). However, during the development of symplectitic intergrowths, the dissipation of chemical energy results in the growth of phases displaying high degrees of curvature (e.g. Fig. 4.4b & 4.10) and high surface area to volume ratios. Such complex phase morphologies may be preferred by the nucleation and growth mechanism active during reaction. Lamellar intergrowths in undercooled steels develop as a result of “cooperative growth” of ferrite and cementite, suggesting that the nucleation and growth of both phases reached a kinetically favorable geometry (i.e. interlamellar spacing, and relative crystallographic orientations) that was preserved with the further development via branching mechanisms (Hillert, 1962). An interesting observation of this research was that perlite textures form as crystallographically distinct colonies. The observations and model of Hillert (1962) have been used to quantitatively link microstructural observations and rates of reaction and the degree of undercooling in both material science (Thompson & Howell, 1988; Capdevila et al., 2002) and geological applications (e.g. olivine exsolution: Ashworth and Chambers, 2000). Although perlite forms via exsolution, it is diffusionally controlled at the ferrite-cementite interface-scale, and similar textural, and 158 in some cases, crystallographic observations can be made with net-transfer symplectitic intergrowths. This section provides the first use of electron backscatter diffraction (EBSD) to investigate symplectitic intergrowths. EBSD is advantageous because it provides rapid and non-destructive in situ crystallographic analysis of multi-phase materials, allowing for acquisition of larger-scale textural information not possible using other techniques. A drawback of the technique is the inability to constrain misorientation and other crystallographic relationships at the boundary between different phases and the orientation of phase boundaries in the third-dimension, necessary for quantifying surface energy of the boundary in a single thin-section. However, one can still quantify the presence of crystallographic alignment between phases. EBSD ANALYSIS OF SYMPLECTITIC CORONAL TEXTURES Detailed EBSD analyses were carried out on six representative samples and eight distinct symplectitic textures within them using a JEOL 6500 FEG-SEM at the Institute of Technology Characterization Facility at the University of Minnesota. Some analyses were obtained at the Mineral and Materials characterization facility in the Dept. of Geology & Geophysics at the University of Wyoming. Analyses were carried out using a 20 keV accelerating voltage an ~80 nA sample current. Orientation indexing was achieved using Channel™ software and EBSD detector of Oxford/HKL technology. The fine-grained nature of the symplectitic textures along with the optical dominance of 159 spinel precluded a detailed analysis of the grain-scale structure of the textures. As a result, EBSD maps were collected using a 0.7 µm step size. GRAIN-SCALE STRUCTURE AND GENERAL CRYSTALLOGRAPHIC TRENDS EBSD analysis allows for unprecedented access to the grain-scale structure of reaction textures. Such data are important for understanding the dynamics of their growth and evolution. Morphological observations from EBSD maps obtained for reaction textures from the Thor-Odin dome are consistent with petrographic observations e.g. coarsening of vermicular spinel from the Al2SiO5 porphyroblast outward to the cordierite moat (Figs. 4b & 6a). EBSD analysis adds a second level of observation: crystallographic orientations, relationships and general trends therein can be examined in detail and compared directly to the overall morphological structure of the reaction texture. Band contrast maps provide general characterization of a sample’s overall grain structure (e.g. Fig. 4.13a). For example, one can identify grain, and, in some cases, phase boundary structure, but cannot, by the use of these maps alone, identify high and low angle boundaries. The reaction texture in Fig 4.13a contains both plagioclase and cordierite hosted spinel vermicules. It is evident from the band contrast map that there exists a complex boundary structure, particularly in the Sp+Cd layer of this example. The lower map illustrates that the majority of cordierite grains in the Sp+Cd layer have c-axes that are oriented more or less perpendicular with the layer boundary of the cordierite moat. Large cordierite single crystals bound by high-angle grain boundaries host the 160 majority of spinel vermicules, with the remainder of vermicules present along boundaries separating the cordierite grains. Within these large single crystals is a complex structure of low-angle boundaries that emanate from the included spinel vermicules. This sub-grain structure is largely absent where extremely fine-grained spinel is present (e.g. Sp+Pl layer in Fig. 4.13a). Similar relationships are observed in Sp+Pl symplectitic layers (Fig. 4.13b). The same complex sub-grain structure related to the presence of spinel in cordierite also exists within plagioclase grains. However, the plagioclase single crystals tend to be of larger in size (on the order of >100 µm) than is typically observed in Cd+Pl symplectitic layers (~20-50 µm). The high-angle boundaries at Pl-Pl interfaces are dominated by 180° rotations about the [010] corresponding to the albite law twinning (Smith, 1974). The geometry of albite twins are typically very complex and of small area relative to the dominant orientation. The general crystallographic orientation trends of the spinel vermicules in both plagioclase and cordierite hosts also have interesting characteristics related to their host phase. Spinel is characterized by clusters of vermicules with the same crystallographic orientation (Fig. 4.14a & b). The areal extent of these clusters is variable, however clusters of single orientations of spinel vermicules are typically bounded by single crystals of the host phase. Single orientations of the spinel vermicules hosted by both plagioclase and cordierite can be interpreted in two ways. Either the observed orientations represent complex (possibly dendritic) single crystals of spinel, or the 161 crystallographic orientation of spinel is crystallographically controlled by the host phase. Neither interpretation, however, is mutually exclusive. Two orthogonal EBSD maps were obtained in an attempt to evaluate the 3D structure of the spinel vermicules (Fig. 4.15). As described in the previous section, spinel vermicules range from having an elongate (lamellar) appearance with the common observation of curved and branched structures to being small isolated crystals. One way to achieve this dichotomy is that the isolated grains represent the intersection of the elongate spinel grains or possibly a 3D dendrite. EBSD analyses cannot distinguish between these possibilities. CRYSTALLOGRAPHIC RELATIONSHIPS BETWEEN HOST AND VERMICULAR PHASES To investigate the likely growth mechanisms active during corona growth, it is necessary to determine crystallographic alignment relationships between products and reactants, and, where applicable, between intergrown phases. These data also provide first order observations of crystallographic control on the orientation of vermicular phases within the host crystals. Such observations may give insight into the kinetic and thermodynamic processes that control the complex morphology of symplectitic intergrowths. There is evidence of some component of crystallographic control of the host phase on the orientation of spinel vermicules. Clusters of single orientation spinel vermicules within cordierite and anorthite were chosen as a subset and the pole figures of the spinel and the respective host phase were plotted to investigate the presence of any 162 crystallographic relationships between phases within the symplectitic layers (Fig. 4.16 & 4.17). Crystallographic control on the orientation of spinel by the orientation of cordierite in the Sp+Cd symplectitic layers is evident, but inconsistent within and between samples. Crystallographic alignment between spinel and cordierite ranges from the {111} of spinel being strongly aligned with the [001] of spinel (Fig. 4.16a-c), to these axes being 10-15° out of parallel (Fig. 4.17d-e), to the two phases showing no correspondence of the {111} & [001] (Fig. 4.17f). The first case is the most common relationship, and is observed in the majority of single-crystal cordierite hosted spinel clusters. It may be that the observed non-alignment, which can be as small as 10-15° deviations in {111}sp and [001]cd orientations is complicated by rotations along sub-grain boundaries within cordierite single crystals. Qualitatively, it appears that more elongate spinel shows stronger alignment with its respective host phase, however, more data are needed to quantify this relationship. Trasmitted electron microscopy would help address this issue by investigating the lattice scale interactions at spine-cordierite interfaces. A similar range of observations is present in the Sp+Pl layers. The {111} of spinel is aligned with the [001] axis of plagioclase. As is the case above, this alignment ranges from a strong correlation (Fig. 4.17a-b), to these axes being on the order of 10-30° out of alignment (Fig. 4.17c-e). Plagioclase is a triclinic phase, so some variation in alignment is easier to explain than in the case of the orthorhombic cordierite. Errors in crystallographic data obtained from the EBSD technique are on the order of 1-2° (Prior et al., 1999), so errors associated with the technique alone cannot 163 explain the observed variation. Alternatively, the observed variation in crystallographic alignment between the vermicular and host phases may point to the vermicules existing as single crystals within the host phase. In addition, the complex subgrain structures developed within both cordierite and plagioclase may have formed as adjustments within the crystal structure of plagioclase or cordierite and small rotations along these boundaries could lead to apparent non alignment of the host and vermicular phases as the texture began to mature and coarsen. CRYSTALLOGRAPHIC RELATIONSHIPS BETWEEN CORONA LAYERS AND AL2SIO5 The central Al2SiO5 crystal is dominantly sillimanite, which replaced kyanite during decompression (Norlander et al., 2002). Sillimanite has a very complicated microstructure (Fig. 4.13a) made up of small elongate prisms, but has a consistent [001] orientation with rotations of the other axes about the c-axis at the vermicular scale (Fig. 4.18a-b). Cordierite present in the outer moat and in the symplectitic layers ranges from having the same orientation (Fig 4.18a) to having distinct orientations (Fig 4.18b). Although there are examples of cordierite c-axes aligned with the c-axes of sillimanite, this relationship does not appear to be systematic. In addition, although there exist some data suggesting an alignment of the respective {110} planes of cordierite present in some examples, this has also been found to not be consistent in all samples. This suggests that a different nucleation and growth mechanism other than epitaxy or topotaxy was active during growth. 164 ESTIMATES OF MATERIAL TRANSPORT A fundamental tenet in the study of coronal reaction textures is that there exist diffusional fluxes redistributing mass within a reaction zone (e.g. within and between layers) as well as fluxes into and out of this zone. The calculation of material transport into the reaction zone (total thickness of corona), however, can be ambiguous without an adequate reference frame (Thompson, 1959; Johnson & Carlson, 1990). In order to define a reference frame to compare product and reactant compositions and use these to determine an estimate of material transport, two of three conditions must be met (Thompson, 1959): 1) a priori knowledge of the conservation, gain or loss of a component(s), 2) the position of the interface between reactants, or 3) volume changes during reaction. Reaction textures do not, however, typically provide consistent or unambiguous evidence for or against any of these conditions. The most common assumption used to quantify material transport is that of constant volume and/or the conservation of one or more components (e.g. Thompson, 1985; Ashworth & Birdi, 1998; Ashworth et al., 2001). However, the assumptions used to assume the conservation of any component may be arbitrary and based on an additional set of assumptions related to the relative component grain boundary diffusivities that are, in most cases, poorly understood. To overcome these issues, a method was developed to systematically investigate the amount of material transport independent of knowledge of either of the first two conditions stated above, and assuming 165 constant volume (Johnson & Carlson, 1990; Carlson & Johnson, 1991). This method utilizes the phase compositions and modal distribution of products within the full thickness of the reaction zone. The position of the original reaction interface and the degree of component conservation (or transport) is determined by systematically varying the ratio of the reactants participating in the reaction. A similar but modified method was used in the present study. A direct application of the method of Carlson & Johnson (1991) to the symplectitic coronal textures in the Thor-Odin dome was not possible due to the textural heterogeneity. For example, although the relative modal amounts of the two-phase symplectitic assemblages was fairly constant within and between samples, the presence or absence of one or more of the two-phase assemblages as well as their relative thicknesses was highly variable. In addition, the matrix surrounding the coronal textures may consist of orthoamphibole, biotite, or both, making a determination of the reactants difficult. Calculation of apparent material transport from outside of the reaction zone (termed boundary fluxes for convenience of discussion) was developed to allow for the systematic investigation of the effect of heterogeneity on estimating material transport in reaction textures. This was achieved by systematically changing the relative thickness (or presence/absence) of the two-phase symplectitic layers as well as the identity and composition of the matrix reactant. Modal proportions of the two-phase assemblages and the cordierite moat and their respective phase compositions were integrated to determine a representative bulk composition (normalized to 24 oxygens) of the entire thickness of 166 the reaction zone. The relative proportion of each of the layer assemblages was varied systematically. The bulk composition of the entire reaction texture was then compared with an inferred reactant assemblage to determine the net molar change necessary to achieve the product bulk composition. The relative ratio of reactants was changed at 10% intervals to represent different locations of the initial reaction interface. Different reactant assemblages were also investigated (Oam(ged)+ Al2SiO5; Oam(anth)+ Al2SiO5; and Bi+ Al2SiO5) based on the compositional and textural variations observed in most samples. Calculations were carried out on symplectitic coronas around Al2SiO5 and garnet. ESTIMATES OF MATERIAL TRANSPORT IN SYMPLECTITIC CORONAS ASSOCIATED WITH AL2SIO5 The method described above was used to determine, at a first-order level, how the estimated material boundary fluxes are affected by changing product assemblages (Figs. 19-20). These figures relate the percent change (in moles/24 oxygens) of the net gain or loss of components to the assumed ratio of original reactants and choice of layer assemblages. The system is effectively isochemical with respect to an individual component where its curve crosses zero percent change. A completely isochemical system would be one where all components converge to zero at any point along the abscissa. The relative boundary material fluxes can then be determined if the position of the original interface can be specified in the analysis. For simplicity, we will first consider examples of coronal reaction textures consisting of only a cordierite moat or a 167 combination of the consistently observed cordierite moat and a single layer of one of the symplectitic assemblages (Sp+Cd, Sp+Pl, Crn+Cd) and assuming a reaction assemblage of Oam(Ged)+Al2SiO5. It is clear that the reaction textures are not isochemical with respect to all components for any choice of the location of the original interface. This is especially explicit in Figure 19a where only the presence of the cordierite moat is considered. Figures 19b-d illustrate the evolving magnitudes of material transport in cases where a cordierite moat is present with a single symplectitic layer. A cordierite moat with a symplectitic assemblage of Sp+Cd (Fig. 4.19b) and Sp+Pl (Fig. 4.19c) is the most commonly observed progression where only one symplectitic assemblage is present. The crossover of the Si and Al curves (essentially signifying the change of Si dominated to Al dominated phases) occurs at a position of the original interface corresponding to where the end (relative to the outer boundary with the matrix phase) of the cordierite moat. Although calculations were carried out for a full range of ratios of the product layers, this relationship was constant. In addition, the percent change of Al and Si, in all calculations, at their crossover point was between 6 and 9%. This range can be considered effectively isochemical, which would mean, choosing the crossover point as the location of the original interface, the amount of Si and Al were not evolved at the texture boundaries during corona formation and evolution. The calculation containing Crn+Cd (Fig. 4.19d) does not show the relationship of the Si-Al crossover corresponding to the end of the cordierite moat, but would signify an original interface just outboard of the change from Cd to Crn+Cd. 168 There are also indications of the isochemical behavior of other components that correspond to the Al-Si crossover. The texture that contains a Sp+Cd symplectitic layer, at the Si-Al crossover, shows no isochemical behavior of other components. Significant transport of Fe and Mg into the coronal domain relative to the original bulk composition is evident (Fig. 4.19b). The coronas containing a single layer of either Sp+Pl or Crn+Cd, however show evidence for conservation of Mg and Fe respectively at the Si-Al crossover as well as significant transport of Ca and Fe in the Sp+Pl example (Fig. 4.19c) and Mg in the Crn+Cd example (Fig. 4.19d). This suggests that changing the identity of, or access to, components transported from outside the boundaries of the corona may affect the resultant phase assemblage. The above results are from calculations that all assumed a gedrite-rich orthoamphibole composition as the reacting matrix phase. However, some orthoamphibole grains are zoned at the rim toward the anthophyllite side of the solid solution, and biotite is commonly present at the boundaries of the coronal textures. The results of calculations investigating different reactant assemblages reveal that the Si-Al crossover moves systematically when considering any product assemblage or number of layers. The crossover moves toward to the left when considering an anthophyllite-rich orthoamphibole as an original reactant, representing an increase in the amount of Al2SiO5 present prior to reaction (Fig. 4.20a) and the crossover moves to the right, representing a decrease in the amount of Al2SiO5 present prior to reaction when biotite is chosen as an original reactant (Fig. 4.20b). The consistent relationship of the location of the Si-Al crossover with the end of the outermost cordierite moat when the 169 gedrite-rich orthoamphibole is chosen as a reactant may indicate that coronal development occurred when gedrite was the dominant matrix phase. In addition, similar relationships of the Si-Al crossover corresponding to the observed end of the cordierite moat are observed when investigating multi-layer examples of the coronal reaction textures. Calculations considering multiple layers of symplectitic assemblages were carried out by systematically varying the proportions of the Sp+Cd and Sp+Pl layers while fixing the thickness of the cordierite moat consistent with textural observations. Analytical composition data and modal proportions of the phases in the corona textures presented in this study show that both phases and their relative proportions are consistent within samples regardless of layer thickness or spatial relationship to other layers. As a result, varying proportions of layers present in the product assemblage results in gradual variation from one of the end member cases (i.e. either Sp+Cd and Sp+Pl) (compare Figs 4.20c-e with Figs 4.19b-c). More Fe and Ca are brought in from outside the texture boundaries and Mg is relatively conserved in the Sp+Pl dominated cases, while Fe, and especially Mg evolve to necessitating larger degrees of transport into the texture boundaries in the Sp+Cd dominated layers. In all cases, the Si-Al crossover is consistent with the input thickness of the cordierite moat. The above calculations were made with the assumption of constant volume within the reaction zone during the growth and evolution of the corona structure. Whereas large volume changes may occur in dehydration reactions (e.g. Thompson, 1975), near constant volume during diffusion-controlled reactions may be a reasonable 170 approximation for reactions not involving large amounts of H2O volitilization (Carmichael, 1987). In addition, no pertibation of matrix phases surrounding the reaction textures was observed. Calculations were carried out to obtain first-order estimations of volume change during corona formation (Table 3). These data show that the largest volume changes would occur in relation to changing the proportions of the reactant phases (orthoamphibole and Al2SiO5). Material transport estimations suggest the original reactant ratio was ~30% orthoamphibole to 70% Al2SiO5. If this assessment is true, than the overall volume change can be described as a very small decrease in volume (Table 3). In addition, even if small volume changes occur during reaction (< ± 10-20%) this only results in similar magnitude shifts in component transport and does not affect the above interpretations (Carlson & Johnson, 1991). ESTIMATES OF MATERIAL TRANSPORT IN CORONA TEXTURES ASSOCIATED WITH GARNET Coronal textures after garnet are composed of a consistent 60 : 40 ratio of cordierite to biotite. Textural evidence of the garnet reaction textures suggests that the present outline of the corona corresponds to the original interface between reactants. The same calculations described above were carried out on garnet coronas to investigate this inference as well as to estimate magnitudes of material transport in and out of the reaction domain. 171 The results of these calculations (Figs. 21a-c) show a similar crossover relationship of Si and Al in these textures. The crossover in the case of these reactions however, is located at the boundary corresponding to the entirety of the present reaction texture being occupied by garnet regardless of the assumption of the reacting phases. In fact, changing the reactant present in the matrix does little to change the magnitude or direction of transport of any of the components considered here. Interestingly the garnet coronas give off Fe and Ca, components that are needed to form the symplectitic reaction textures associated with Al2SiO5, and are relatively scarce, especially in the case of Ca, in other phases. Although the heterogeneity observed in the Thor-Odin reaction textures adds a level of complexity to the standard interpretation of coronal texture formation, this analysis suggests there are relatively simple shifts in the evolution of material balance that may exist as a second order control superimposed on the diffusionally controlled growth of these textures. In the following section, these data will be integrated with two and three-dimensional textural observations, interpretations of the morphology and crystallographic relationships of symplectitic phases to examine the evolution of coronal textures at multiple scales. DISCUSSION Layered coronal symplectitic reaction textures from the Thor-Odin dome are defined by a heterogeneous distribution of product assemblages about the central reactant Al2SiO5; in addition a crystal morphological discontinuity is present, marked by the change from 172 polygonal-shaped crystals to vermicular-shaped crystals. Coronal textures are typically interpreted to form during a single stage of growth occurring in response to a change in state variables (P-T-µ). However, the development of extreme assemblage heterogeneity could point to a reaction system responding to evolving conditions during continued coronal growth. The heterogeneity of symplectitic coronal reaction textures in orthoamphibole- cordierite gneiss from the Thor-Odin dome is the result of an evolving system, with local compositional heterogeneities superimposed on the larger-scale bulk-rock compositional and reaction evolution of these rocks. Observations from multiple scales can be integrated to develop a semi-quantitative understanding of the dynamics controlling the phase assemblage and morphological heterogeneity of symplectitic coronal textures, including the role of small-scale changes in boundary fluxes. The calculated boundary flux evolution can then be placed in the context of evolving system variables (i.e. µ, X, P, and T) and evolving system size, with emphasis on how changes in these parameters relate to the heterogeneity and how tectonic processes may have influenced the thermodynamic and kinetic evolution of the system. Superimposed on the heterogeneity of assemblages is the contrast in crystal growth mechanisms separating the cordierite moat and the symplectitic layers. Controlling parameters of layered coronas and symplectitic coronas are the same, but result in a very different morphological characteristics. Traditional petrographic analysis, combined with EBSD data allows unprecedented access to the crystallographic and sub- grain-scale characteristics of both morphological zones. Thus allowing for interpretation 173 of likely growth mechanisms responsible for the development of symplectitic layers and possibilities for why vermicular-shaped are developed over polygonal-shaped grains characteristic of layered coronas. CONTROLS ON TEXTURAL HETEROGENEITY Evolving boundary fluxes and reactants In steady-state diffusion models of layered corona formation, changing boundary fluxes affect the resulting thicknesses and phase assemblages (Carlson & Johnson, 1991). To a first-order, material transport calculations for the Thor-Odin symplectitic coronas suggest a similar relationship. The implication of this result is that while steady-state diffusion may be reached during corona formation, as evident by the consistent compositions of phases within layers, drastic changes in the composition or phase assemblage that differ significantly from local sources suggest shifts in the origin of fluxes related to changing reactants – a situation that cannot be accounted for in steady-state derived models (Johnson & Carlson, 1990; Carlson & Johnson, 1991). For example, all of the phase assemblages in coronal layers around an Al2SiO5 polymorph (Cd, Sp+Cd, Crn+Cd, Sp+Pl) require boundary fluxes outside of the reaction zone however the necessary components for Cd, Sp+Cd ,Crn+Cd, can be wholly derived from phases immediately adjacent to the corona. In contrast, the formation of Sp+Pl (An95-99) layers necessitates a relatively large flux of CaO, which is not found in appreciable concentration in any of the surrounding phases except in cases where garnet is within a few mm of an Al2SiO5 174 polymorph. Regardless, components from outside the reaction zone must have been transported from the surrounding volume even for the coronal assemblages that can be derived entirely from local sources of diffusing components. The presence or absence and relative thickness of any one of the two-phase symplectitic layer assemblages is extremely variable. The only constant is the presence of a cordierite moat enveloping the entire reaction texture. If each two-phase assemblage is considered as an end-member case and the Si-Al crossover is assumed to be the location of the original interface, the calculated boundary fluxes necessary to develop the observed bulk composition vary considerably. For example, the layer sequence Cd|Crn+Cd requires addition of Mg and conservation of Fe, Si, Al; the sequence Cd|Sp+Pl requires addition of Fe and Ca, and conservation of Mg, Si and Al (Fig 19c-d). The sequence of Cd|Sp+Cd requires the addition of both Mg and Fe, with only Si and Al conserved within the reaction layer (Fig. 4.19b). Changes in the relative proportions of any of these assemblages results in systematic variation between these end-member examples. This analysis suggests that the composition and identity of reactants and sources of boundary fluxes evolved during reaction texture formation – this is most obvious with the development of the Sp+Pl assemblages. We can investigate the role of changing sources and evolving chemical potential gradients for texture formation through investigation of chemical potential space (µ-µ). Qualitative isothermal-isobaric µ-µ diagrams were constructed from orthogonal ternary composition space in the systems MAS, FAS, AFM, and ACF in order to also explore the effect of choosing different chemical potentials (Korzhinskii 1959). These reduced 175 systems were chosen because they adequately represent the dominant components in the phases. Three-phase fields in three-component composition space become points in µ-µ space, two-phase tie lines are preserved as lines, and points become single-phase fields. The topology of each m-m diagram is based solely on the compositions of the phases and their relative position in ternary composition space (Fig. 4.22). The topology was constructed to reflect the observed product phase relationships. The slopes and positions of tie-lines in µ-µ space are most strongly controlled by changing phase composition, whereas their lengths and invariant relationships will change with changing P or T in accordance with the stability of the phases. Although a qualitative representation, the µ-µ diagrams can be used to evaluate the general directional shifts responsible for the textural heterogeneity. The µ-µ diagrams (Fig. 4.22a-e) illustrate the dependence the layer sequences of the symplectitic coronas on changing conditions regardless of the identity of assumed controlling potentials. For example, two possible gradients with respect to µMgO and µAlO3/2 required to form either of the end-member Cd|Crn+Cd or the Cd|Sp+Cd layer sequences (Fig. 4.22a-e). The cordierite field is crossed in both cases, thus explaining the consistent presence of a cordierite moat in all combinations of layer sequences. After crossing the cordierite field, the gradients intersect and evolve along either of the respective two-phase tie-lines toward the sillimanite stability field. The layer sequence of Cd|Crn+Cd|Sp+Cd is commonly observed in symplectitic textures from the Thor-Odin dome (e.g. Figs 4.3-4.6). A shift in either of the chemical potentials (caused by compositional zoning in orthoamphibole i.e. anthophyllite-rim to gedrite-core or 176 changing P-T conditions causing the movement of the Crn+Sil assemblage toward high µAlO3/2 is required to form the observed sequence (Fig. 4.22a-e). The same changes are required regardless of the chemical potentials considered in MAS, FAS and AFM systems (Fig. 4.22a-d). The presence of plagioclase in the symplectitic layer sequence necessitates expanding composition space to include CaO. The drastic change in the composition of layer sequences marked by the appearance of Sp+Pl symplectitic layers can easily be explained by a shift in the chemical potentials controlling the diffusive growth of the textures to include CaO (Fig. 4.22e). As expected, Cd|Sp+Cd layer sequences can be achieved with gradients invariant to µCaO. However, to produce the commonly observed Cd|Sp+Cd|Sp+Pl sequences, the system must evolve with respect to µCaO (Fig. 4.22e). Based on the phases present in the OCG that preserve these textures, garnet is the likely source of the CaO fluxes responsible for plagioclase formation. Steady-state models of corona formation are based on the assumption that all the preserved layers formed with infinitesimal thickness and grew to their observed thicknesses based on component diffusivities (Fisher, 1973; Joesten, 1977, 1986). Clearly, such a steady-state approach does not apply in the context of the Thor-Odin reaction textures. The layer boundary Sp+Cd|Sill changed to Sp+Pl|Sill where CaO had access to the reaction texture. Essentially, both interfaces represent the same reaction front in terms of the nomenclature of steady-state models. Garnet, as a source of CaO, is not problematic as it is present in all samples of OCG explored in this study. However, garnet is commonly heterogeneously distributed within the OCG studied. Because CaO 177 must have diffused over large distances (on the order of 1 to 5 cm) relative to the size of the reaction texture (0.5 -1 mm), an increase in the scale of the system must have occurred. The expansion of the size of the diffusive system is interpreted as the second control on the reaction texture heterogeneity. Increased size of the system and interaction of coronal reactions Textural evidence is also suggestive of a relationship between the coronas associated with garnet and those associated with Al2SiO5. Textural associations between the two coronal textures range from direct contact between garnet- and Al2SiO5-centered symplectitic reaction textures (Fig. 4.9a) to the development of spectacular reaction zones consisting of Bi+Cd connecting the two reaction textures (Fig. 4.9b). HRXCT data confirm that the observed interconnectivity is systematic throughout single samples and that one garnet- centered corona can be interconnected with many Al2SiO5-centered coronas, either through immediate interaction, or via a network of multiple Al2SiO5-centered coronas (Fig. 4.12a-d). Owing to the size and mode percent of Al2SiO5 and garnet, a corona network developed. A corona network is comprised of individual coronas that communicated with each other during corona formation. However, the occurrence of Sp+Pl layers in all samples that contain garnet with an appreciable grossular content (Xgrs >0.05) suggests that component transport occurred across a length-scale of up to 3 cm. This was achieved via diffusion through the observed Bi+Cd reaction zones from garnet to the Al2SiO5- centered coronas and through the corona networks. 178 The common development of plagioclase flame structures and variability of its presence radially around a single Al2SiO5 porphyroblast can be explained by the intermittent access to CaO. Diffusion pathways taken by CaO were complex and not necessarily orthogonal to the layer boundaries that had already developed (as the paths for MgO, AlO3/2, and in some cases FeO would be). Textural heterogeneities related to the variable phase assemblages and thickness of symplectitic layers can be explained by compositional heterogeneity of reactants at the local scale (e.g. Cd|Crn+Cd & Cd|Sp+Cd layers) and the gradual influx of CaO into the Al2SiO5-centered symplectitic coronas via diffusion from garnet breakdown. The stability field of garnet consistently persists to lower-pressures with respect to the stabilization of cordierite and the destabilization of Al2SiO5 on decompression (e.g. Fig. 3.16) and regardless of bulk composition in the OCG system (Diener et al., 2008). Therefore, some of the heterogeneity may be explained by evolution along a P-T path, most likely isothermal decompression. Once the cordierite-in phase boundary was overstepped on decompression, the Al2SiO5 system left the equilibrium condition and corona development was initiated (Fig. 4.23). On continued decompression, garnet was destabilized and Bi+Cd corona formation was initiated as well providing sources for excess FeO and CaO (Fig. 23). The Sp+Cd layers may, in part, be a reflection of the expansion from locally sourced components for diffusion controlled growth to the development of the interconnected corona network as garnet became destabilized on decompression (Fig. 4.23). In this regard the sequence Cd|Crn+Cd would represent the most ‘closed’ diffusion system and the sequence Cd|Sp+Pl would represent the most 179 ‘open’ system. Combinations of all three-layer assemblages represent the evolution of system size with continued corona development. Transport of components over the relatively large length-scales can be explained by the presence of fluids from the migmatite surrounding the OCG. There is evidence for a staged fluid influx event as well as chemical/fluid interaction with migmatite (Chapter 2). Limited access to fluid during corona texture formation could also be related to the restricted diffusion pathways and variable development of plagioclase within reaction textures. The fluid event culminated in the widespread formation of biotite after gedrite as well as quartz+cordierite veins, which crosscut the OCG pods. Both of these fluid- related events post-date corona formation. CONTROLS ON MORPHOLOGY AND GRAIN/INTERFACE-SCALE PROCESSES IN SYMPLECTITIC REACTION TEXTURES Layered coronas and layered symplectitic coronas differ, in terms of the kinetics driving their formation, only in the morphology of product assemblages. Layered coronas can consist of polygonal two-phase assemblages (e.g. Johnson & Carlson, 1990; Ashworth & Birdi, 1990; Satirini-Rideout et. al., 2007) and therefore the presence of two-phase assemblages is not sufficient to explain morphological differences. However, a survey of corona symplectitic reaction textures shows there is a systematic relationship with the symmetry of host-vermicular phases (Table 4.4). Vermicular phases are consistently of higher symmetry than the host phases, while polygonal two-phase assemblages tend to 180 have the same or similar symmetries. The symmetry contrast is not likely a first order control on developing a vermicular morphology, but likely a coincidence of the phases deemed stable is unclear. However, such a disparity in symmetry may be necessary for a host phase to include the intergrown phase. The development of different crystal morphologies during reaction is controlled by a combination of interface thermodynamics and kinetics as well as the details of the crystallography of the phase(s) considered. Quantifying the former is out of the scope of this paper, however the latter will be used along with other grain- and interface-scale observations from petrographic and EBSD analyses to discuss possible scenarios for the formation of symplectitic textures. There are two distinct phase morphologies within the layered symplectitic coronas associated with Al2SiO5: polygonal grains in the cordierite moat and two-phase vermicular intergrowths. The cordierite moat typically consists of a single layer of polygonal. The majority of grain boundaries within the polygonal cordierite layer are perpendicular to the interfaces separating the polygonal layer from the matrix and symplectitic assemblages (Figs. 13 & 14). The second characteristic morphology is the complex two-phase vermicular intergrowths of the symplectitic layer assemblages. Similar to the grain boundary orientations observed in the polygonal cordierite layer, the majority of grain and phase boundaries within each symplectitic layer are oriented perpendicular to layer boundaries (e.g. Fig. 4.13a). However, the phase, grain and subgrain boundaries exhibit a much more complex, anastomosing architecture within the symplectitic layers. Such disparities in morphology can be developed even in simple systems where differences in the 181 crystallographic character and kinetic processes at opposing interfaces result in different crystal morphologies during subsequent growth. For example, diffusion experiments on the MgO-Al2O3 system developed both fine-grained polygonal as well as elongate spinel between the reacting phases (Watson & Price, 2002). While not directly applicable, it is clear that a different set of variables at a reacting interface can greatly affect the growth mechanism and resultant morphology. EBSD data reveal that spinel grew within single crystals of a host phase (either plagioclase or cordierite). The single crystal host phases form cellular networks that grow out from the reaction front (Fig. 4.13-4.14). The vermicular spinel within these single crystal hosts ranges from small grains to elongate parallel to the shape of the host crystal. It is interpreted that the range of spinel shapes observed is related to the fact that a thin section displays different slices through a dendritic single crystal. The single crystal colonies likely grew in multiple directions giving the wide range of vermicular shapes (Fig. 4.24). The cellular structure formed by the network of single crystal hosted dendrites is of similar character to structures formed during eutectoid type reactions (e.g. discontinuous precipitation e.g. Hillert, 1962; Solorzano & Purdy, 1984; Thomson & Howell, 1998; Caballero et al., 2001). Discontinuous precipitation has been hypothesized for the growth of symplectitic textures through exsolution of orthopyroxene and olivine (Boland & Otten, 1990; Ashworth, 1991), based in part on circumstantial evidence, and in the case of net-transfer symplectites (Field, 2008). Discontinuous precipitation results in the formation at a reaction interface of a supersaturated phase with respect to a component and a less- 182 supersaturated phase. In analogy to the symplectitic reaction textures presented here, spinel would represent the supersaturated phase (with respect to Al) and cordierite and plagioclase would represent the less-supersaturated phase. A discontinuous precipitation- like reaction may develop if AlO3/2 diffusion is sufficiently slow. Al/Si ratios of the symplectitic assemblages are on the order of the Al/Si ratio of AL2SiO5 (Table 3), suggesting that while, Al diffusion was occurring, Al did not diffuse very far relative to MgO, FeO and CaO. This may result in the system conforming with the reaction dictated by the larger-scale chemical potential gradients and in the case of these textures, the slow diffusivity of AlO3/2 by creating a microstructure that limits the length-scale necessary for AlO3/2 diffusion and forming the vermicular intergrowth structure (Fig. 4.24). The complex vermicular and branching shapes formed in a similar manner as in eutectoid transformations in metals, but are further constrained by the relative complexity of the silicate structure of the host phase forcing the spinel into more complex vermicular shapes to adhere to the modal percents dictated by reaction stoichiometry. An alternative explanation related to the likelihood that AlO3/2 had a relatively slow diffusivity is related to microstructural self-organization to enhance grain-boundary area and thus the effective grain-boundary diffusivity of the system. Natural and synthetic systems that are far from equilibrium commonly display characteristics that suggest self- organization (e.g. Orteleva et al., 1987; Brown & Solar, 1998; Renard et al., 1998; Page et al., 2003). The development of symplectitic textures and the associated intergrown single crystals may be a result of the system self-organizing to maximize boundaries to enhance grain boundary diffusion. It is clear that, AlO3/2 did not diffuse at great length- 183 scales, however, it must have diffused at some scale to balance reactions at internal and external layer boundaries as well to participate in the coarsening evident in the external portions of the symplectitic layers. The resultant complex microstructure may have developed in order to provide the shortest path for AlO3/2 diffusion in order to conform to the layered structure controlled by longer length-scale chemical potential gradients. Examples of coronas that did not develop symplectitic microstructures do not necessitate long-range diffusion of typically slow diffusing cations (Al, Si). For example, troctolites that contain coronal textures between olivine and plagioclase contain a layer assemblage of Pl|Grt|Cpx+Opx|Ol. SiO2 and AlO3/2 do not have to diffuse across the majority of the reaction zone to form this corona (Johnson & Carlson, 1990). However, Lang et al. (2004) present symplectitic textures in similar rocks that contain layer assemblages of Ol|Opx|Amph|Sp+Amph|Pl. Here, AlO3/2 must diffuse further out into the texture to form the observed corona. The differences in systems that develop simple layered coronas consisting of only polygonal grains compared to those that develop symplectitic layers could be a simple case of slight variation in the relative magnitudes in the difference between fast and slow diffusivities of components or a combination with rate-limiting interfacial mechanisms. More directly, the cause could be related to reaction between phases providing components to the corona that have very different diffusivities. CONCLUSIONS 184 Layered symplectitic reaction textures associated with Al2SiO5 preserved in orthoamphibole-cordierite rocks from the Thor-Odin gneiss dome display remarkable heterogeneity in terms of phase assemblages and morphologies. Variation in the presence or absence, as well as respective thicknesses of two-phase symplectitic assemblages, is controlled by changes in the relative magnitudes of boundary fluxes into the reaction zone, both from local heterogeneity as well as an increase in the size of the diffusive system providing components to the reaction. This is exemplified by two and three- dimensional analysis indicating the intimate involvement of garnet and its associated corona with those developed around Al2SiO5 creating an interconnected corona network that channeled and redistributed components throughout the bulk rock. Phase-equilibria suggest that the relative destabilization of Al2SiO5 and garnet occur at different pressure conditions along an isothermal path. Therefore the formation of the respective coronal textures could also have been staggered. The involvement of garnet in the formation of Sp+Pl symplectitic layers points to the evolution of the symplectitic textures along an isothermal decompression path. EBSD data and associated microstructural observations show strong correlation between single-crystal cellular network developments and textures derived from discontinuous precipitation. Textural and crystallographic data suggest that a modified discontinuous precipitation growth mechanism may be the cause of symplectitic intergrowths. The development of either layered coronas or symplectitic coronas is related to the relative diffusivities of mobile components. 185 REFERENCES Alverez-Velaro, A.M., Cesare, B., & Kriegsman, L.M., 2007. Formation of spinel- cordierite-feldspar-glass coronas after garnet in metapelitic xenoliths: reaction modelling and geodynamic implications. Journal of Metamorphic Geology, Ashworth JR, Sheplev, V.S., Bryxina, N.A., Kolobov, V.Y. & Reverdatto, V.V., 1998. Diffusion-controlled corona reaction and overstepping of equilibrium in a garnet granulite, Yenisey Ridge, Siberia. Journal of Metamorphic Geology, 16, 231-246. Ashworth, J. R. & Birdi, J. J., 1990. Diffusion modelling of coronas around olivine in an open system. Geochimica et Cosmochimica Acta, 54, 2389–2401 Ashworth, J. R., 1993, Fluid-absent diffusion kinetics of Al inferred from retrograde metamorphic coronas: American Mineralogist, v. 78, no. 3-4, p. 331-337.; Ashworth, J. R., and Chambers, A. D., 2000, Symplectic reaction in olivine and the controls of intergrowth spacing in symplectites: Journal of Petrology, v. 41, no. 2, p. 285-304. Ashworth, J. R., Reverdatto, V. V., Kolobov, V. Y., Lepetyukha, V. V., Sheplev, V. S., and Ashworth, J. R., Sheplev, V. S., Khlestov, V. V. & Ananyev, V. A., 2001. Geothermobarometry using minerals at non-equili- brium: a corona example. European Journal of Mineralogy, 13, 1153–1161. Audibert, N., Betrand, P., Hensen, B. J., Kienast, J. R. & Ouzegane, K. (1993) Cordierite- K-feldspar-quartz-orthopyroxene symplectite from southern Algeria; new 186 evidence for osumilite in high-grade metamorphic rocks. Mineralogical Magazine, 57 (387), p. 354-357 Boland, J.N., and Otten, M.T. (1985) Symplectitic augite; evidence for discontinuous precipitation as an exsolution mechanism in Ca-rich clinopyroxene. Journal of Metamorphic Geology, 3(1), 13-20. Brady J.B. (1975) Reference frames and diffusion coefficients. Amer. J. Sci. 275, 954- 983. Brown, M. & Solar, G. S., 1999. The mechanism of ascent and emplacement of granite magma during transpression: a syntectonic granite paradigm. Tectonophysics, 312, Brown, M., 2002. Retrograde processes in migmatites and granulites revisited. Journal of Metamorphic Geology, 20, 25-40. Brown, R.L. (2004): Thrust-belt accretion and hinterland underplating of orogenic wedges; an example from the Canadian Cordillera. In Thrust Tectonics and Hydrocarbon Systems (K.R. McClay, ed.). Am. Assoc. Petroleum Geol., Mem. 82, 51-64 Bryxina, N. A., 1998, Textures of diffusion-controlled reaction in contact- metamorphosed Mg-rich granulite, Kokchetav area, Kazakhstan: Mineralogical Magazine, v. 62, no. 2, p. 213-224. Caballero, F. G., Capdevila, C., and De Andres, C. G., 2001, Modelling of isothermal formation of pearlite and subsequent reaustenitisation in eutectoid steel during 187 continuous heating: Materials Science and Technology (UK), v. 17, no. 6, p. 686- 692 Capdevila, C., Caballero, F. G., and De Andres, C. G.,Kinetics model of isothermal pearlite formation in a 0.4C-1.6Mn steel: Acta Materialia (USA), v. 50, no. 18, p. 4629-4641. Carlson W. D. and Johnson C. D. ( 1991 ) Coronal reaction textures in garnet amphibolites of the Llano Uplift. Amer. Mineral. 76, 756-772. Clarke, G.L., & Powell, R. (1991) Decompressional coronas and symplectites in granulites of the Musgrave Complex, central Australia. Journal of Metamorphic Geology, 9, 441-450. Davies, H.L., and Warren, R.G. (1992) Eclogites of the D'Entrecasteaux Islands. Contributions to Mineralogy and Petrology, 112(4), 463-474 Denison, C., Carlson, W. D. & Ketcham, R. A., 1997. Three-dimensional quantitative textural analysis of metamorphic rocks using high-resolution computed X-ray tomography.Part I. Methods and techniques. Journal of Metamorphic Geology, 15, 29–44. Diener, J.FA., Powell, R., & White, R.W., 2008. Quantitative phase petrology of cordierite–orthoamphibole gneisses and related rocks. Journal of Metamorphic Geology, Droop, G.T.R. (1989) Reaction history of garnet-sapphirine granulites and conditions of Archaean high-pressure granulite-facies metamorphism in the Central Limpopo mobile belt, Zimbabwe. Journal of Metamorphic Geology, 7, 383-403. 188 Elvevold, S., & Gilotti, J.A. (2000) Pressure-temperature evolution of retrogressed kyanite eclogites, Weinschenk Island, North-East Greenland Caledonides. Lithos, 53, 127-147 Field, S.W. 2008. Diffusion, discontinuous precipitation, metamorphism, and metasomatism: The complex history of South African upper-mantle symplectites. American Mineralogist, 2008, 4, 618-631. Fisher G. W. (1973) Nonequilibrium thermodynamics as a model for diffusion-controlled metamorphic processes. Amer. J. Sci. 273, 897-924. Fisher G. W. (1975) The thermodynamics of dillusion-controlled metamorphic processes. In Mass Transport processes in ceramics. (editors A. R. Cooper and A. Ghent, E.D., Nicholls, J., Stout, M.Z., Rottenfusser, B., 1977. Clinopyroxene amphibolite boudins from Three Valley Gap, British Columbia. Can. Mineral. 15, 269 – 282 Gordon SM, Whitney DL, Teyssier C, & Grove M, 2008. Timescales of migmatization, melt crystallization, and cooling in a Cordilleran gneiss dome: Valhalla complex, southeastern British Columbia. Tectonics, 27, TC4010. Grant S. M. (1988) Diffusion models for corona formation in metagabbros from the Western Grenville Province, Canada. Contrib. Mineral. Petrol. 98, 49-63. Grantham, G.H., Thomas, R.J., Eglington, B.M., de Bruin, D., Atanasov, A., and Evans, M.J., 1993, Corona textures in Proterozoic olivine melanorites of the Equeefa Suite, Natal Metamorphic Province, South Africa: Mineralogy and Petrology, v. 49, p. 91-102. 189 Harley, S. L., 1985, Garnet-orthopyroxene bearing granulites from Enderby Land, Antarctica; metamorphic pressure-temperature-time evolution of the Archaean Napier Complex: Journal of Petrology, v. 26, no. 4, p. 819-856, Harley, S.L. (1986) A sapphirine-cordierite-garnet-sillimanite granulite from Enderby Land, Antarctica; implications for FMAS petrogenetic grids in the granulite facies. Contributions to Mineralogy and Petrology, 94(4), 452-460 Harley, S.L., Hensen, B.J., and Sheraton, J.W. (1990) Two-stage decompression in orthopyroxene-sillimanite granulites from Forefinger Point, Enderby Land, Antarctica; implications for the evolution of the Archaean Napier Complex. Journal of Metamorphic Geology, 8(6), 591-613 Hillert, M. (1972) Theories of Growth During Discontinuous Precipitation. Metallurgical Transactions A 2729-274 Hinchey, A.M., Carr, S.D., McNeill, P.D., & Rayner, N., 2006. Paleocene–Eocene high- grade metamorphism, anatexis, and deformation in the Thor–Odin dome, Monashee complex, southeastern British Columbia. Canadian Journal of Earth Sciences, 43, 1341-1365. Hirsch, D. M., Ketcham, R. A. & Carlson, W. D., 2000. An evaluation of spatial correlation functions in textural analysis of metamorphic rocks. Geological Materials Research, 2, 1–42Prior, D. J., Wheeler, J., Brenker, F. E., Harte, B. & Matthews, M., 2000. Crystal plasticity of natural garnets: new microstructural evidence. Geology, 28, 1003–1006. 190 Holland, T. J. B. & Powell, R., 1998. An internally-consistent thermodynamic data set for phases of petrological interest. Journal of Metamorphic Geology, 16, 309–343. Joesten R. (1977) Evolution of mineral assemblage zoning in diffusion metasomatism. Geochim. Cosmochim. Acta 41, 649-670. Joesten R. (1986) The role of magmatic reaction, diffusion, and annealing in the evolution of coronitic microstructure in troctolitic gabbro from Risrr, Norway. Mineral. Mag. 50, 441-467. Johnson C.D. and Carlson W.D. (1990) The origin of olivine-plagioclase coronas in metagabbros from the Adirondack Mountains, New York. J. Metam. Geol. 8, 697- 717 Johnson, T., Brown, M., Gibson, R., & Wing, B. (2004) Spinel–cordierite symplectites replacing andalusite: evidence for melt-assisted diapirism in the Bushveld Complex, South Africa. Journal of Metamorphic Geology, 22, 529-545 Johnston, A.D., & Stout, J.H. (1984) Development of orthopyroxene-Fe/Mg ferrite symplectites by continuous olivine oxidation. Contributions to Mineralogy and Petrology, 88, 196-202. Keller, L.M., Wunder, B., Rhede, D., and Wirth, R.,. 2008. Component mobility at 900°C and 18 kbar from experimentally grown coronas in a natural gabbro. Geochemica et Cosmochimica Acta, 72, 4307-4322. Korzhinskii D. S. (1959) Physicochemical Basis of the Analysis of the Paragenesis of Minerals. Consultants Bureau USA 191 Kruckenberg, S.C., Whitney, D.L., Teyssier, C., Fanning, C.M., & Dunlap, W.J., 2008. Paleocene-Eocene migmatite crystallization, extension, and exhumation in the hinterland of the northern Cordillera: Okanogan dome, Washington, USA. Geological Society of America Bulletin, 120, 912-929 Kruse, S., McNeill, P. D. & Williams, P. F., 2004. A geological compilation map of the Thor-Odin dome. www.unb.ca/fredericton/science/geology/ monashee. Lang, H.M., Wachter, A.J., Peterson, V.L., & Ryan, J.G., 2004. Coexisting clinopyroxene/spinel and amphibole/spinel symplectites in metatroctolites from the Buck Creek ultramafic body, North Carolina Blue Ridge. American Mineralogist, 89, 20-30. Messiga, B., and Bettini, E. (1990) Reactions behaviour during kelyphite and symplectite formation; a case study of mafic granulites and eclogites from the Bohemian Massif. European Journal of Mineralogy, 2(1), 125-144 Milke, Dohmen, R., Becher, H-W., Wirth, R., 2007. Growth kinetics of enstatite reaction rims studied on nano-scale, Part I: Methodology, microscopic observations and the role of water. Contributions to Mineralogy and Petrology, Mongkoltip, P., and Ashworth, J. R., 1983, Quantitative estimation of an open-system symplectite-forming reaction; restricted diffusion of Al and Si in coronas around olivine: Journal of Petrology, v. 24, no. 4, p. 635-661 Norlander, B.H., Whitney, D.L., Teyssier, C. & Vanderhaeghe, O., 2002. Partial melting and decompression of the Thor-Odin dome, Shuswap metamorphic core complex, Canadian Cordillera. Lithos, 6, 103-125. 192 Nyman, M. W., Pattison, D. R. M. & Ghent, E. D., 1995. Melt extraction during formation of K-feldspar+sillimanite migmatites, west of Revelstoke, British Columbia. Journal of Petrology, 36, 351-372. Ortoleva, P., Merino, E. Moore, C., & Chadam, J., 1987. Geochemical Self-organization I: Reaction-transport feedbacks and modeling approach. American Journal of Science, 287, 979-1007. Page, K., Maini, P.K., & Monk, N.A.M., 2003. Pattern formation in spatially heterogeneous Turing, reaction-diffusion models. Physica D: Nonlinear Phenomena, 181, 80-101. Parrish, R.R., Carr, S.D., and Parkinson, D.L. 1988. Eocene extensional tectonics and geochronology of the southern Omineca Belt, British Columbia and Washington. Tectonics, 7: 181-212. Passchier, C. W., and Trouw, R. A. J., 1996, Microtectonics: New York, Springer-Verlag, 290 p Patel, S.C., Behera, S., Mohanty, A., Singh, A.K., & Patel, S.K. (2001) Metamorphic history of sapphirine- and relict-kyanite-bearing Mg-Al rich rocks at Hatimunda Hill near Junagarh, Kalahandi District, Orissa. Journal of the Geological Society of India, 57, 417-427 Pitra, P. & De Waal, S. A. (2001) High-temperature, low-pressure metamorphism and development of prograde symplectites, Marble Hall Fragment, Bushveld Complex (South Africa). Journal of Metamorphic Geology, 19 (3), 311-325. 193 Renard, F., Gratier, J-P. Ortoleva, P. Brosse, E., & Bazin, B., 1998. Self-organization during reactive fluid flow in a porous medium. Geophysical Research Letters, 25, 385-388. Satirini-Rideout, C., Gilotti, J.A., & Foster, C.T., 2006. Forward modeling corona growth in a partially eclogitized leucogabbro, Bourbon Island, North-East Greenland. Lithos, 56, 700-717. Solorzano, I.G. & Purdy, G.R. 1984. Interlamellar spacing in discontinuous precipitation. Metallurgical Transactions A, 15, 1055-1063. Teyssier, C., Ferré, E., Whitney, D.L., Norlander, B., Vanderhaeghe, O., and Parkinson, D. 2005. Flow of partially molten crust and origin of detachments during collapse of the Cordilleran orogen. Submitted to Geological Society of London, Symposium volume Channel flow, ductile extrusion and exhumation of lower- mid-crust in continental collision zones.InHigh-strain zones: Structures and Physical Properties. Edited by: Bruhn, D. and Burlini, L. Geological Society of London Special Publication, 245: 39-64 Thompson, J. B., 1959. Local equilibrium in metasomatic proesses. In Researches in Geochemistry (ed. Abelson, P. H.), pp. 427–457. Wiley, New York Thompson, S.W., & Howell, P.R., 1988. On the early stages of pearlite formation in hypoeutectoid steels. Scripta Metallurgica, 22, 1775-1778. Turner, S.P. & Stüwe, K. (1992) Low-pressure corona textures between olivine and plagioclase in unmetamorphosed gabbros from Black Hill, South Australia. Mineralogical Magazine, 56, 503-509 194 Vanderhaeghe, O., Teyssier, C. & Wysoczanski, R., 1999. Structural and geochronological constraints on the role of partial melting during the formation of the Shuswap metamorphic core complex at the latitude of the Thor-Odin Dome, British Columbia. Canadian Journal of Earth Sciences, 36, 917-943. Watson EB, Price JD (2001) Kinetics of the reaction MgO + Al2O3 fi MgAl2O4 and Mg-Al interdiffusion in spinel at 1200– 2000°C and 1.0-4.0 GPa. Geochem Cosmochim Acta 66:2123–2138 White, R.W., Powell, R., & Baldwin, J.A., 2008. Calculated phase equilibria involving chemical potentials to investigate the textural evolution of metamorphic rocks. Journal of Metamorphic Geology, 26, 181-198. Whitney, D.L., Goergen, E.T., Ketcham, R.A., & Kunze, K., 2008. Formation of garnet polycrystals during metamorphic crystallization, 28, 365-383. Williams, P.F., and Jiang, D. 2005. An investigation of lower crustal deformation: evidence for channel flow and its implications for tectonics and structural studies. Structural Geology, 27: 1486-1504 195 FIGURE CAPTIONS Figure 4.1. Schematic illustrating the differences between layered corona and layered symplectitic reaction textures. Both reaction textures are diffusionally controlled, with the only difference being purely morphological. Symplectitic reaction textures can form with or without an additional polygonal product layer. Figure 4.2. a) Generalized geologic map of the Shuswap metamorphic core complex. The Thor-Odin dome is located in the center along a line of several domes within this terrain. b) Simplified geologic map after Kruse et. al. (2006). Samples described in this paper are from the Saturday Glacier region of the dome, which is located in the southernmost extent of the Thor-Odin dome. The southern portions of the dome also contain the most evidence for high-grade melting in the form of high melt-fraction migmatites (diatexites). Figure 4.3. Plane-polarized (PPL) photomicrographs of symplectitic reaction textures after Al2SiO5 (var. Sil) to illustrate the absence of significant differences in symplectitic assemblages with respect to surrounding matrix being composed dominantly of orthoamphibole or biotite. a-b) Note the similar phase assemblages present regardless of texture surrounded dominately by Oam (a) or Bi (b) and the blocky-shapes of Crn in these textures. c) Photomicrograph showing the most common reaction texture, and 196 absence of Crn+Cd assemblage. This texture also contains Sp+Pl assemblages, but these cannot be distinguished optically (Sp*). Mineral abbreviations used in this paper: Gt- garnet, Sil-Sillimanite, Ky-kyanite, Q-quartz, Oam-orthoamphibole (general), Ged- gedrite-rich orthoamphibole, Anth-anthophyllite-rich orthoamphibole, Bi-biotite, Cd- cordierite, Sp-spinel, Spr-sapphirine, Pl-plagioclase, Crn-corundum, Ru-rutile, Ilm- ilmenite, Hem-hematite, Symp-general reference to symplectitic reaction texture. Figure 4.4. Backscattered electron (BSE) images of symplectitic reaction textures after Al2SiO5. a) BSE image illustrating the inconsistent presence and thicknesses of Sp+Pl symplectitic layers in a single texture. b) BSE image of typical flame structures observed separating Sp+Cd and Sp+Pl layers. Figure 4.5. BSE images (a-b) and PPL photomicrograph (c) illustrating the preservation of the shape of the central Al2SiO5 (Blue outline) porphyroblast by both the edge of the symplectitic assemblages (white outline) and the Cd moat itself. See text for interpretations. Figure 4.6. PPL photomicrographs of typical morphologies of symplectitic assemblages. a) vermicular (worm-like) morphology. b) lamellar/elongate morphology c) Example of the rare occurrence of Crn+Cd assemblages in the center of the reaction texture. Note the vermicular morphology of the Crn+Cd in comparison to blocky-shaped Crn in Fig. 3 & 5. 197 Figure 4.7. Compositional (Mg#-Mg/Mg+Fe) traverses across Cd in symplectitic reaction textures. a) Traverse across a cordierite moat from surrounding biotite to Sp+Cd symplectitic layer. b) Traverse across cordierite moat from biotite to the contact with Sp+Pl symplectitic layer. c) Travers across both cordierite moat and Sp+Cd symplectitic layer. Note in all cases that compositional zoning is of the same magnitude and compositional shift regardless of the surrounding assemblages. The heterogeneity of Cd composition within the Sp+Cd layer is likely due to some input of surrounding spinel during analysis. Cd analyses from this layer were stoichiometric, but varied in relative proportions of Fe vs Mg. Figure 4.8. PPL photomicrographs of garnet coronal texures. a) Garnet coronal textures contain intergrowths of polygonal cordierite and elongate biotite. The coronal texture closely mimics the present shape of the garnet porphyroblast. This is especially apparent in b) where angular shapes of the garnet are matched exactly by the outline of the coronal texture. Such observations have led to the interpretation that the diameter of the coronal textures is representative of the former size of the garnet porphyroblasts. Figure 4.9. PPL photomicrographs of the textural associations of garnet and Al2SiO5 coronal textures. a) Direct textural relationship of Gt and Al2SiO5 reaction textures. b) evidence for relatively long range chemical and textural communication between the two corona textures (Gt & Al2SiO5). Such textural interaction is common in all samples and verified in 3D (see Fig. 12). 198 Figure 4.10. BSE image of Sp+Cd and Sp+Pl symplectitic layers. Sp displays extremely complex shapes, particularly in the upper-half of the image. Sp is also displays long- range interconnectivity. These observations are suggestive of a 3D dendritic morphology of spinel in these symplectitic textures. Figure 4.11. Six slices taken from HRXCT results on a single reaction texture. Note the consistency of the Cd moat thickness across all slices. The total thickness of the symplectitic layers is relatively constant but the presence of Crn+Sp in this texture is extremely heterogeneous. It is difficult to image the Sp+Pl or Sp+Cd layers via this technique. Figure 4.12. Four slices taken from HRXCT data on a core of OCG to illustrate the interconnectivity between coronal textures. In these images garnet is the brightest phase, followed by the orthoamphibole-biotite matrix, cordierite is the darkest phase. Note the intricate interconnectivity between garnet and Al2SiO5 coronal textures across all slices. Diameter of images is 2.54 cm.(a) is labeled, and the other images are left unlabeled for clarity. Figure 4.13. EBSD maps of symplectitic layers. a) Top image: Band contrast map of Sp+Pl and Sp+Cd layered symplectitie. Band contrast images are based on the pattern quality of individual analyses and roughly correlate to BSE images (Sp is the brightest 199 phase in these images, the separation between Sp+Pl and Sp+Cd is noted. Lower map, coloring based on general crystallographic orientations with respect the X direction of the image (coloration based on inverse pole figure inset). Data show that the majority of cordierite in this map have their [001] aligned rougly parallel with the X-direction, which is also orthogonal to the symplectite-Al2SiO5 interface (at far left of image). Yellow lines in the image mark high-angle grain boundaries of cordierite. Note that the region is separated into several single crystals. Low angle grain boundaries are present within cordierite single crystals and associated with included spinel (dark lines in cordierite in top image). b) top and bottom images are similar to (a) but showing data collected on a symplectite containing only a Sp+Pl layer. Bottom image: Coloring of plagioclase orientations with respect to Euler axes used in indexing. In this image like colors represent similar orientations. Yellow lines indicate high-angle grain boundaries within plagioclase. These high-angle boundaries are dominated by 180° rotations about [010] (albite twinning). Similar to cordierite in (a) plagioclase is present as a number of smaller single crystals containing low angle grain boundaries associated with the spinel inclusions. Figure 4.14. EBSD band contrast maps overlayed with coloration based on spinel crystallographic orientations relative to the Y-axis of the image. a) EBSD map of Sp+Pl symplectite. Note the clustering of spinel with similar color/orientation. These clusters correspond to single crystals of plagioclase and form a cellular structure made up of individual single crystals of plagioclase including spinel with a single orientation. It is 200 interpreted that the single orientations of spinel occur, in part, because the spinel in these clusters are also single crystals. b) EBSD map of a Sp+Pl symplectite. Again, the clustering of spinel is evident in this texture. Figure 4.15. Orthogonal EBSD maps of a Sp+Pl symplectitic layer. There exist no systematic changes in the morphology of the spinel grains with respect to orientation analyzed. It is unclear from the data presented in the figure alone to determine the 3D geometry of spinel in these symplectites. Figure 4.16. Pole figures of cordierite single crystals and associated included spinel. The crystallographic axes/planes plotted as well as the orientation of the pole figure with respect to the sample plane are listed in the center of the figure. a-b) Pole figures of Cd and Sp showing strong alignment of the {111} of Sp with the c-axis of Cd. The {111}Sp/[001]Cd alignment is the dominant crystallographic relationship observed Sp+Cd layers. c-d) Deviation of the alignment shown in (a) and (b). In these pole figures the {111}Sp/[001]Cd alignment is off by ~10-15°. This misalignment is likely the result of reorientation along low-angle grain boundaries within the cordierite single crystals. e- f) Pole figures showing no crystallographic relationship between the {111}Sp and [001]Cd. Alignment of the [100]Cd and {100} of spinel is observed in (f), but this alignment is not systematic. These data are interpreted to represent crystallographic control of Cd on the orientation of Sp during growth. 201 Figure 4.17. Pole figures of plagioclase single crystals and associated included spinel. The crystallographic axes/planes plotted as well as the orientation of the pole figure with respect to the sample plane are listed in the center of the figure. a-c) similar to the Sp+Cd symplectites, Sp+Pl intergrowths show strong alignment of the c-axis of plagioclase and the {111} of spinel. d-e) Again similar to the Sp+Cd case, some Sp+Pl single crystal intergrowths show slight misalignment to no alignment of the [001]Pl and {111}Sp. Figure 4.18. Pole figures illustrating crystallographic relationships of host crystals with Al2SiO5. a) pole figures of a complete data set showing Cd moat and Cd within symplectitic intergrowths are dominated by a single orientation. Neither orientation shows any crystallographic relationship to sillimanite orientation. Plagioclase also shows no alignment with the central sillimanite porphyroblast. b) similar to (a) neither plagioclase nor cordierite show consistent alignment with the central sillimanite porphyroblast. These data suggest that epitaxial growth was not active during nucleation and growth of symplectitic textures. Figure 4.19. Results of material transport calculations assuming the complete absence, or the presence of only a single symplectitic layer. These calculations compare the bulk composition of the product assemblage with that of the original two-phase reactant assemblage (in this figure considering Ged+Als). The y-axis is the percent change in content of major element components from the original assemblage to the product assemblage. Positive values mean that material must be transferred out of the boundaries 202 of the reaction zone (the observed thickness of the product assemblage), negative values represent components that must be brought from outside the boundaries of the reaction zone. The system is isochemical with respect to an individual component where the curve crosses 0% change. A system that is completely isochemical would one where all components cross 0% change at the same location of the x-axis. The x-axis represents the position of the original interface in terms of percentage. For example 0.1 represents an original assemblage containing 90% Al2SiO5 and 10% Ged or, in other words 90 % of the reaction zone was occupied by Al2SiO5. a) Material transport calculation assuming only a cordierite moat. Calculations illustrate that the system is open with respect to all components for any choice of location of original interface. b) material transport calculation assuming the presence of a Cd from 0 to 0.4 (40%)along the x-axis and the rest of the reaction zone made up of Sp+Cd (60%). A consistent characteristic of these calculations is the correlation of location of the crossover between Al and Si and the input value of the innermost edge of the Cd moat (0.4 in b). The percent change of Al and Si are 5% at this point, which is interpreted here to be effectively isochemical. It is interpreted that this crossover also represents the position of the original interface between reactants. Assuming the location of the interface allows for interpretation of the relative magnitudes of transport into or out of the system of other components. In (b) this would mean that 25% of the Fe and 80% of the Mg would have to originated outside of the bounds of the reaction zone. c) Material transport estimates considering a Cd moat and a Sp+Pl symplectitic layer. Again, the Al-Si crossover correlates to the location of the inner edge of the Cd Moat. Calculation suggests all of the Fe and Ca (Ca not present 203 on graph because it exceeds 100 % change as a result of the extremely low concentrations of Ca in Ged) must be transported from outside the reaction zone, and that Mg is isochemical (assuming the Al-Si crossover as the position of the original interface) b) Material transport estimates considering a Cd moat and a Crn+Cd layer. Assuming the Al-Si crossover as the original interface, Fe is conserved, as well as Si and Al, while 50% of the Mg in the product assemblage must have originated outside the boundaries of the reaction zone. Figure 4.20. a-b) Material transport calculations considering different reactants in the original assemblage. The results are similar to Fig. 19 with the exception of the position of the Si-Al crossover, which in these calculations does not correlate to the inner edge of the Cd moat. The Si-Al crossover moves outboard of the inner edge of the Cd moat when considering Anth in the assemblage and moves inboard of the edge of the Cd moat when considering Bi. c-e) Material transport calculations considering multi-layered symplectitic assemblages. The position of the Si-Al crossover in these calculations is static and correlates with the inner edge of the Cd moat. Mg and Fe vary systematically based on the dominant symplectitic assemblage considered. Figure 4.21. Material transport calculations of garnet coronal reaction textures. The Si-Al crossover in these calculations correlates with the outer edge of the reaction texture, which is in keeping with petrographic observations that the extent of the reaction textures after garnet are representative of garnet’s original diameter. Only transport of Mg into the 204 reaction texture is necessary for its formation, while 70% of Fe and 100% of Ca within garnet is transported outside the boundaries of the reaction texture (a). This correlates with the magnitudes of Fe and Ca necessary to form the Sp+Pl symplectitic assemblages (Fig. 19c). There is virtually no change in the interpretations when considering different reactant assemblages (b-c). Figure 4.22. Qualitative isothermal-isobaric µ-µ diagrams calculated after the method of Korzhinskii (1959). a) µMgO-µAlO3/2 diagram illustrating the necessity of a change in conditions to explain the common layer sequence of Cd|Crn+Cd|Sp+Cd. Dashed lines represent the metastable extension of the Sp-Cd-Sill invariant. The observed layer sequence can be explained by a change in composition of Oam during progressive reaction, by changing P-T conditions moving the Sp phase field, or by changing the position of the Crn+Sp phase boundary to higher µAlO3/2 conditions. The exact position of this boundary cannot be constrained in qualitative diagrams and was placed based on the observed phase assemblage. b-d) µMgO-µSiO2, µFeO-µAlO3/2, µFeO-µMgO diagrams illustrating that the changes in conditions to produce the observed layer assemblage as are necessary in (a) regardless of the choice of mobile components. e) µCaO-µ(Fe+Mg)O diagram. Evolving conditions are also necessary to form the most commonly observed layer sequence of Cd|Sp+Cd|Sp+Pl. The evolving conditions are likely related to continued decompression (see text for discussion). Note that a Cd moat is likely to form even if Gt and Sill were in direct contact prior to reaction. 205 Figure 4.23. Representative pseudosection of OCG rocks from the Thor-Odin Dome to illustrate the general P-T relationships of phases involved in the symplectitic coronas. Prior to decompression, the phase assemblage was dominated by Gt, Al2SiO5, and Oam (1). On decompression the entrance of Cd and destabilization of Sil (blue field) is crossed (2). This marks the initiation of coronal texture development around Al2SiO5 and its departure from equilibrium, most likely resulting in the formation of the Cd moat and Crn+Cd symplectite. Continued decompression results in evolution in µ-µ space resulting in the formation of Sp+Cd layers. Finally the Gt-out phase boundary is crossed (3) and results in the formation of Bi+Cd coronas and the development of interconnected coronal networks linking Gt and and Al2SiO5 providing Fe and Ca to the symplectitic textures associated with Al2SiO5. Figure 4.24. Schematics illustrating the interpreted 3D morphology of symplectitic intergrowths, the relationship of 2D slices with observed shape variation and the likely growth mechanism responsible for the observed morphology. It is interpreted that the spinel clusters are present as 3D dendrites within single crystals of cordierite and plagioclase. The single crystal packages grew in multiple directions with respect to the observed 2D plane of a thin-section. Different slices through a single crystal dendrite can explain the observed variation in spinel morphology. Finally, a process similar to discontinuous precipitation, active at the Al2SiO5-symplectite interface, may also help explain the observed morphologies. Al2SiO5 must un-mix and diffuse AlO3/2 and AlSiO7/2 across the interface. The majority of data and observations presented here 206 suggest that diffusion was limited with respect to both Si and Al. In an attempt to limit the length-scale of diffusion for these components the system may have evolved similar to discontinuous precipitation at this interface (see text for discussion). Ta bl e 4 .1. R ep re se nt ati ve m icr op ro be an aly se s o f o rth oa m ph ib ol e, co rd ier ite , s pi ne l, s ap ph iri ne , p lag io cla se , a nd bi ot ite Ph as e Or th oa m ph ib ol e Or th oa m ph ib ol e Co rd ier ite Co rd ier ite Co rd ier ite Sp in el Sp in el Pl ag io cla se Bi ot ite Bi ot ite Ga rn et Ga rn et Te xu ra l Pl ac em en t Co re Ri m Af ter Or th oa m ph ib ol e Co ro na l w /S p w / Pl w / Cr d w /S p Ad jac en t t o Sy m pl ec tit e Gr t-C or on a Co re Ri m SiO 2 w t% 45 .72 54 .53 48 .77 48 .12 49 .26 0.5 7 0.2 0 44 .06 37 .43 37 .55 38 .47 38 .76 Ti O 2 0.2 0 0.0 7 0.0 0 0.0 1 0.0 0 0.9 5 0.0 5 - 0.4 6 1.7 7 0.0 1 0.0 0 Al 2O 3 15 .56 4.5 0 34 .15 34 .46 34 .50 62 .88 65 .08 36 .59 19 .75 19 .46 22 .58 22 .83 Fe O 14 .43 16 .04 3.5 7 4.9 4 5.4 3 22 .34 22 .79 0.3 7 13 .22 11 .68 26 .48 26 .54 M nO 0.1 3 0.1 4 0.0 6 0.1 1 0.1 4 0.0 5 0.0 6 - 0.0 6 0.0 2 1.7 4 1.4 5 M gO 19 .27 22 .39 11 .61 10 .71 10 .52 11 .77 12 .36 - 15 .99 16 .52 6.1 6 8.2 5 Ca O 0.5 1 0.4 5 0.0 1 0.1 1 0.0 4 - - 19 .30 0.0 2 0.0 0 4.9 9 2.5 0 N a 2O 1.5 2 0.4 6 0.1 6 0.0 5 0.0 5 - - 0.5 2 0.5 2 0.7 4 10 0.4 1 10 0.3 3 K 2 O 0.0 1 0.0 0 0.0 0 0.0 1 0.0 1 - - 0.0 1 8.1 9 8.2 3 To tal 97 .36 98 .58 98 .31 98 .51 99 .94 98 .57 10 0.5 4 10 0.8 5 95 .74 96 .04 No rm ali ze d t o 23 O 23 O 18 O 18 O 18 O 4 O 4 O 8 O 22 O 23 O Si 6.4 8 7.6 0 4.9 2 4.8 8 4.9 3 0.0 2 0.0 1 2.0 2 5.4 5 5.4 1 2.9 8 2.9 8 Al 1.5 2 0.4 0 4.0 6 4.0 5 4.0 6 1.9 8 2.0 0 1.9 7 0.0 5 0.1 9 2.0 6 2.0 7 Ti 0.0 2 0.0 1 0.0 0 0.0 0 0.0 0 0.0 2 0.0 0 - 3.3 9 3.3 1 0.0 0 0.0 0 Al VI 1.0 9 0.3 4 - - - - - - 0.0 0 0.0 0 Fe 2+ 1.7 1 1.8 7 0.3 0 0.4 2 0.4 5 0.5 0 0.5 0 0.0 1 1.6 1 1.4 1 1.7 4 1.7 3 M n 0.0 2 0.0 2 0.0 0 0.0 1 0.0 1 0.0 0 0.0 0 - 0.0 1 0.0 0 0.1 1 0.0 9 M g 4.0 7 4.6 5 1.7 5 1.6 2 1.5 7 0.4 7 0.4 8 - 3.4 7 3.5 5 0.7 1 0.9 4 Ca 0.0 8 0.0 7 0.0 0 0.0 1 0.0 0 - - 0.9 5 0.0 0 0.0 0 0.4 1 0.2 1 N a 0.4 2 0.1 2 0.0 3 0.0 1 0.0 1 - - 0.0 5 0.1 5 0.2 1 K 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 - - 0.0 0 1.5 2 1.5 1 0.2 9 0.3 5 0.5 8 0.5 8 X M g 0.7 0 0.7 1 0.8 5 0.7 9 0.7 8 0.4 8 0.4 9 - 0.6 8 0.7 2 0.0 4 0.0 3 X A n - - - - - - - 0.9 5 0.2 4 0.3 2 0.1 4 0.0 7 Cr 2O 3 c on ten t w as be lo w d ete cti on li m its in al l p ha se s 207 Table 4.2. Measurements of symplectitic layer assemblages Total thickness of reaction texture (µm)* Average s.d. n 678.64 57.92 30.00 Thickness of cordierite moat Average s.d. n number of textures 168.07 20.47 30.00 22.00 Thickness of Sp+Cd** Average s.d. n number of textures 253.50 19.09 30.00 17.00 Thickness of Sp+Pl** Average s.d. n number of textures 174.62 55.56 30.00 15.00 *Variation in thickness is also a function of 2D limitations on measureing 3D structures **Measurements of symplectitic layer assemblages was only done where the phase assemblage was the only symplectitic layer present 208 Table 4.3. Estimated volume change calculations for symplectitic reaction coronas* Original Assemblage Proportion of Reactants Product layers Proportion of layers Proportion within two-phase assemblage Molar Volume of original assemblage (J/bar) Oam+Als 30%:70% Cd|Sp+Cd 30%|70% (40%Sp+60%Cd) 19.78 T(C) 700 725 750 775 kbar 5 1.85 1.86 1.87 1.88 6 1.84 1.85 1.87 1.88 7 1.84 1.85 1.86 1.87 8 1.84 1.85 1.86 1.87 Oam+Als 30%:70% Cd|Sp+Cd 30%|70% (50%Sp+50%Cd) 19.78 700 725 750 775 5 3.19 3.20 3.21 3.22 6 3.18 3.19 3.21 3.22 7 3.18 3.19 3.20 3.21 8 3.17 3.18 3.19 3.21 Oam+Als 30%:70% Cd|Sp+Cd 40%|60% (40%Sp+60%Cd) 19.78 700 725 750 775 5 1.08 1.09 1.10 1.11 6 1.08 1.09 1.10 1.11 7 1.08 1.09 1.10 1.11 8 1.07 1.08 1.10 1.11 Oam+Als 40%:60% Cd|Sp+Cd 30%|70% (40%Sp+60%Cd) 17.67 700 725 750 775 5 -0.26 -0.25 -0.24 -0.23 6 -0.26 -0.25 -0.24 -0.23 7 -0.26 -0.25 -0.24 -0.23 8 -0.26 -0.25 -0.24 -0.23 *Calculations were based on end-member thermodynamic data from Holland and Powell, 1998 Al/Si ratios for symplectitic assemblages (Based on 24 Oxygens) Cd Moat Crn+Cd Crd+Sp An+Sp Als 0.84 1.89 1.74 1.92 2.00 209 Ta bl e 4 .4. Su rv ey of cr ys tal lo gr ap hy of sy m pl ec tit ic ph as e a ss em bl ag es Lo ca tio n Ro ck Ty pe Co ro na or Sy m pl ec tit e Ce nt ra l r ea cta nt ph as e o f t ex tu re Ve rm icu le Cr ys tal St ru ctu re H os t C ry sta l S tru ctu re Re fer en ce Th or -O di n Do m e B C Oa m -C rd G ne iss Sy m pl ec tit e Al 2S iO 5 Sp (C ub ic) C rn (H ex ) Cr d( Or th o) P l(T ri) Go er ge n & W hi tn ey (i n re vi ew ) H og ga r, A lg er ia Gr an ul ite A l-M g r ich Sy m pl ec tit e Ge d, G rt, Q tz, Il m Op x( Or th o) Sp (C ub ic) Cd (O rth o) Ou ze ga ne et al ., 1 99 6 N or th ca ro lin a b lu e r id ge M eta tro cto lit e Sy m pl ec tit e Pl ag io cla se Sp l(C ub ic) Cp x( m on o) A m ph (m on o) La ng et al ., 2 00 4 Ar ge nt in a Ga bb ro Sy m pl ec tit e Pl ag io cla se Sp l(C ub ic) Cp x( m on o) A m ph (m on o) Cr uc ian i e t a l., 20 08 So ut h A fri ca M ela no rit e Bo th Pl ag io cla se Sp l(C ub ic) Cp x( m on o) A m ph (m on o) Gr an th am et al ., 1 99 3 H un gr y Sp L he rz ol ite Sy m pl ec tit e Af ter G rt (D ec om po sit io n Sp l(C ub ic) Cp x( M on o) Fa lu s e t a l., 20 07 En de rb y L an d A nt ar tic a M eta pe lit e Co ro na Af ter q+ Sp r N /A N /A W hi te et al. , 2 00 8 So ut h A fri ca Lh er zo lit e Sy m pl ec tit e N ot cl ea r O l? Sp (C ub ic) Op x( Or th o) , C px (M on o) , G rt( Cu bi c) Am ph (M on o) Fi eld 20 08 Ca na da Sn ow bi rd Ec lo gi te Sy m pl ec tit e Ky an ite Op x( Or th o) Sp (C ub ic) C rn (H ex ) Sp r(M on o) Pl (T ri) Ba ld w in et al ., 2 00 7 N or w ay Ga bb ro Bo th Cp x Qt z( Tr ig ) Am ph (M on o) As hw or th et al , 1 99 2 N or w ay Gr an iti c Bo th Ol Op x( Or th o) Gr t(C ub ic) M ar kl et al ., 1 99 8 Co lo ro do Oa m -C rd G ne iss Sy m pl ec tit e Sil l Sp l(C ub ic) C rn (H ex ) Cd (O rth o) H eim an n et al. , 2 00 6 Sp ain N VP Pe lit e Co ro na Gr t N /A N /A Al ve re z- Va ler o e t a l., 20 07 Te xa s L lan o U pl ift Am ph ib ol ite bo th Ga rn et Ca rls on & Jo hn so n 19 91 H or om an C om pl ex , J ap an Lh er zo lit e Sy m pl ec tit e N /A Sp (C ub ic) Op x( Or th o) , C px (M on o) M or ish ita & A ra i, 2 00 3 Pa ys d e L eo n Fr an ce Ec lo gi te Sy m pl ec tit e Ky Cr n( H ex ), Sp (C ub ic) , Sp r(M on o) Pl (T ri) Go da rd & M ab it, 19 98 Su lu U H P Ch in a Ec lo gi te Sy m pl ec tit e Gr t, K y 1) Sp (C ub ic) , 2) Pr g( M on o) , 3) Di (M on o) Pl (T ri) Ya ng et al ., 2 00 4 Gr ee nl an d Ec og ite G ab br o Co ro na Cp x N /A N /A Sa rti ni -R id eo ut et al ., 2 00 7 Cz ec k R eb ub lic (M ol da nu bi an ) Ec lo gi te Sy m pl ec tit e 1) Om ph , 2 )K y/ Gr t Di (M on o) , S p( Cu bi c) Bt (M on o) Pl (T ri) Fa ry ad et al ., 2 00 6 So ut he rn In di a UH T Oa m -C rd Sy m pl ec tit e Ky Sp (C ub ic) ,Sp r(M on o) Cr d( Or th o) Ka na za w a e t a l., 20 08 W es ter n Ca na di an Sh iel d M afi c G ra nu lit e Sy m pl ec tit e Gr t Op x( Or th o) Pl (T ri) M ah an et al ., 2 00 8 So ut he rn In di a Al -M g- ric h Gr an ul ite Sy m pl ec tit e co m pl ex Op x( Or th o) Sp r(M on o) Sp (C ub ic) Cd (o rth o) Ra ith et la ., 1 99 7 N or w ay Ga bb ro Sy m pl ec tit e Ol Sp (C ub ic) Op x( Or th o) Ga rd ne r a nd R ob in s, 19 74 210 Reactants Reactants Products Products Layered Corona Polygonal Products Layered Symplectitic Corona Vermicular Products Figure 4.1 211 Middle Unit Lower Unit THOR-ODIN DOME B.C. Canada Cranberry Mountain Saturday Glacier Bear Paw Lake Mt. Odin Mt. Thor Blanket Mountain Mount Begbie CANADA USA 122º 48º 114º 48º Okanogan- Kettle Priest River Coast Belt 122º 54º 114º54º B.C. Alberta WA MT 50 km N Limit of Cordilleran deformation Valhalla Southern Rocky Mountain Trench Shuswap MCC Thor-Odin (Fig. 1b) Frenchman’s Cap Malton Foreland Fold & Thrust Belt Intermontane Belt Selkirk Allochthon a b Figure 4.2212 a1 mm Oam Bi Spr+Cd Sp* Crn+Cd Cd b c 1 mm 1 mm Bi Bi Sp* Cd Sill Sill Sill Figure 4.3 Sp* Crn+Cd Cd 213 ab Cd Bi Sill Sill Cd Crn+Cd Pl Sp Pl+Sp 500 µm 10 µm Figure 4.4 214 ab c Sill Cd Bi Crn+Cd Sill Sill Sp* Pl+Sp Oam Pl+Sp Cd Cd Oam Bi 500 µm 1 mm 1 mm Crn+Cd Crn+Cd Figure 4.5 215 ab c Sp* Crn+Cd Cd 1 mm Sill Sp* Cd Cd Sp* Bi Bi 200 µm 200 µm Figure 4.6 216 0.75 0.76 0.77 0.78 0.79 0.75 0.76 0.77 0.78 0.79 0.74 0.75 0.76 0.77 0.78 0.74 0.75 0.76 0.77 0.78 0.74 0.75 0.76 0.77 0.74 0.75 0.76 0.77 180 µm 170 µm 420 µm GB GB Cd Moat Cd+Sp M g # M g # Cd Moat Bi-Crd+Spl Cd Moat Bi-An+Spl Cd Traverse Bi-Sill a b c Figure 4.7 217 Figure 4.8 a b Bi Bi Gt Gt Gt Gt Gt Bi+Cd 500 µm 0.5 mm Bi+Cd Symp 218 ab Bi Gt Gt Bi+Cd Gt Bi Sill Sill 500 µm 0.5 mm Symp Symp Figure 4.9 219 Sill Cd Sp Pl Figure 4.10 220 a b c d e f Oa m Cd Bi Sil l Sil l Sil l Sil l Sil l Sil l Oa m Oa m Oa m Oa m Bi Bi Bi Bi Bi Oa m Cr n+ Cd Cr n+ Cd Cr n+ Cd Cr n+ Cd Sp /P l+ Cd Fi gu re 4. 11 221 Figure 4.12 a b c d Symp Gt Oam+Bi 222 200 µm 200 µm 001 010 100 X b Figure 4.13 a Sp+CdSp+Pl Sil Crn+Cd Sp+Pl Sil 223 001 010 100 Y 001 111 101 Y 001 111 101 Y a b Figure 4.14 200 µm 200 µm 224 Fi gu re 4. 15 225 H os t ( Co rd ier ite ) Ve rm icu lar P ha se (S pi ne l) Or th or ho m bi c ( m m m ) Iso m etr ic (m 3m ) H os t ( Co rd ier ite ) Ve rm icu lar P ha se (S pi ne l) Or th or ho m bi c ( m m m ) Iso m etr ic (m 3m ) [1 00 ] [0 10 ] [0 01 ] {1 00 } {11 0} {11 1} [1 00 ] [0 10 ] [0 01 ] {1 00 } {11 0} {11 1} a b c d H os t ( Co rd ier ite ) Ve rm icu lar P ha se (S pi ne l) Or th or ho m bi c ( m m m ) Iso m etr ic (m 3m ) [1 00 ] [0 10 ] [0 01 ] {1 00 } {11 0} {11 1} e f Y0 X0 Y0 X0 Y0 X0 Fi g u re 1 6 226 Fi g u re 4 .1 7 H os t ( Pl ag io cla se ) ( 1 ) H os t ( Pl ag io cla se ) ( 1 ) H os t ( Pl ag io cla se ) ( 1 ) Ve rm icu lar P ha se (S pi ne l) Iso m etr ic (m 3m ) {1 00 } {11 0} {11 1} Ve rm icu lar P ha se (S pi ne l) Iso m etr ic (m 3m ) {1 00 } {11 0} {11 1} Ve rm icu lar P ha se (S pi ne l) Iso m etr ic (m 3m ) {1 00 } {11 0} {11 1} Y0 X0 Y0 X0 Y0 X0 [1 00 ] [0 10 ] [0 01 ] [1 00 ] [0 10 ] [0 01 ] [1 00 ] [0 10 ] [0 01 ] a b c 227 Fi g u re 4 .1 7 co n t. H os t ( Pl ag io cla se ) ( 1 ) Ve rm icu lar P ha se (S pi ne l) Iso m etr ic (m 3m ) {1 00 } {11 0} {11 1} Y0 X0 [1 00 ] [0 10 ] [0 01 ] H os t ( Pl ag io cla se ) ( 1 ) Ve rm icu lar P ha se (S pi ne l) Iso m etr ic (m 3m ) {1 00 } {11 0} {11 1} Y0 X0 [1 00 ] [0 10 ] [0 01 ] ed 228 a b M oa t & In ter gr ow th C or di er ite Sil lim an ite Pl ag io cla se M oa t & In ter gr ow th C or di er ite Sil lim an ite Pl ag io cla se Fi gu re 4. 18 [1 00 ] [0 10 ] [0 01 ] [1 00 ] [0 10 ] [0 01 ] [0 01 ] [0 10 ] [1 00 ] [1 00 ] [0 10 ] [0 01 ] [1 00 ] [0 10 ] [0 01 ] [1 00 ] [0 10 ] [0 01 ] u l u l u l u l u l u l 229 Position of Original Interface 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 -100 -75 -50 -25 0 25 50 75 100 Ged-Als (100% Cd) Si Al Fe Mg 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Ged-Als (40%Cd 60% Sp+Cd) 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Ged-Als (40Cd60Crn+Cd) 1 -100 -75 -50 -25 0 25 50 75 100 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 M ol e % C ha ng e Ged-Als (30%Cd 70%Sp+Pl) M ol e % C ha ng e Position of Original Interface a b c d Figure 4.19 230 -1 00-7 5 -5 0 -2 5025507510 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Po sit io n of O rig in al In ter fa ce An th -A ls (3 9C d6 1C dS p) Si Al Fe M g 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Bi -A ls (3 9C d, 61 Cd +S p) Mole % Change -1 00-7 5 -5 0 -2 5025507510 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Ge d- Al s ( 33 Cd , 1 7S p+ Cd , 5 0S p+ Pl ) 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Po sit io n of O rig in al In ter fa ce Ge d- Al s ( 33 Cd , 2 7S p+ Cd , 4 0S p+ Pl ) 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Ge d- Al s ( 33 Cd , 4 0S p+ Cd , 2 7S p+ Pl ) 1 1 1 Mole % Change a b c d e Fi gu re 4. 20 231 Fi gu re 4. 21 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 -1 00-7 5 -5 0 -2 5025507510 0 Si4 + Al 3+ Fe 2+ M g2 + Ca 2+ Gt (R im )-B i ( 60 % Cd 40 % B i) Gt (C or e)- Ge d (6 0% Cd 40 % B i) Gt (R im )-G ed (6 0% Cd 40 % B i) Mole % Change Po sit io n of O rig in al In ter fa ce a b c 232 µ A lO 3 /2 µMgO C d C rn O am Si ll Sp M A S C rn Si ll C d O am Sp µ A lO 3 /2 µFeO FA S Sp C rn Sill C d O am µ S iO 2 µMgO M A S O am C d Sp C rn G t Pl µCaO µ F eM g A C F Si ll a b c d Sp C rn Si ll O am C d µFeO µ M g O A FM e Fi g u re 4 .2 2 233 -Cd -Si l P T -Gt 5 6 7 8 9 10 675 700 725 750 775 800 825 850 -Sp 1 2 1 2 3 3 Figure 4.23 234 Fi gu re 4. 24 Cd M oa t 1 2 3 1 2 3 3- D M or ph ol og y o f s ym pl ec tit ic in ter gr ow th s Sh ap e v ar iat io n w ith or ien tat io n of a s in gl e h os t-i nt er gr ow th ce ll M gO + F eO AlO3/2 Al2SiO5 Cd Sp (M gF e)A l 2O 4 (M gF e) 2 Si 5 Al 4O 18 (M gF e)A l 2O 4 Al SiO 7/ 2 M gO + F eO M od ifi ed d isc on tin uo us p re cip ita tio n 235 236 Chapter 5 CONCLUSION TO THE THESIS Refractory phases, in particular, the Al2SiO5 polymorphs and garnet, commonly form as porphyroblasts in crustal rocks and are resistant to deformation and, in some cases, reaction. The result of these characteristics is the metastable persistence of phases outside their equilibrium P-T fields, as is commonly observed with the Al2SiO5 polymorphs, or in the development of metamorphic reaction textures commonly associated with both the Al2SiO5 polymorphs and garnet. Although these characteristics typically result in complex microstructural and textural relationships, it is the observed textural complexity itself that allows for the interpretation of the controls on the deformational and reaction processes as well as the ability to place the observed textures and microstructures into a tectonic context. The work presented in this thesis illustrates the utility of complex microstructural and reaction texture development in placing constraints on the effect of changing P-T-X-d conditions on rocks that contain refractory phases. In particular, this work illustrates the effect of deformation on the initiation and progression of a simple metamorphic reaction - polymorphic transformation - in the Al2SiO5 system. Additionally, the work presented here illustrates the importance of the evolution of the length-scales and pathways of diffusion leading to the development of heterogeneous coronal symplectitic reaction 237 textures in orthoamphibole-cordierite rocks. Although this work was concentrated on specific bulk compositions, the observations and interpretations presented here have applications to our general understanding of metamorphic and deformation processes. THE EFFECT OF DEFORMATION ON REACTIONS The feedback relationship between metamorphic and deformation processes is well established (e.g. Rutter & Brodie, 1985, Rubie, 1998). The extent of the dominance of one process over the other is difficult to assess because the processes can be mutually dependent. However deformation experiments on the Al2SiO5 polymorphs (Chapter 2) suggest that strain in some form is necessary for polymorphic transformation at least at the laboratory time scale. No transformation occurred in experiments in the absence of deformation, whereas near complete transformation was evident in deformed samples. Deformation processes affect reaction through enhancing mobility of dislocations, grain boundaries and, locally, diffusion of chemical components (Snow & Yund, 1987). Such enhancements may be necessary where the activation energy for first releasing chemical components from a phase and subsequent diffusion is relatively high. In particular, deformation may play an important role in initiating reaction in systems that have little or no intergranular fluid thereby limiting grain boundary diffusion and/or are composed of refractory phases where strong bonds impede disassociation and limit reactivity. Once kinetic barriers are overcome, reaction may occur rapidly and result in extensive strain localization and strain weakening of the rock aggregate. For example, in 238 deformation experiments on the Al2SiO5 polymorphs, deformation was localized into shear bands composed of sillimanite after either kyanite or andalusite. These shear bands extended across the entire sample charge of the experiment, indicating that either transformation occurred even at very low strains, or that once transformation occurred at higher strains, the formation of seed crystals of the stable phase, sillimanite, provided the necessary energy to overcome kinetic barriers and the reaction propagated from the high strain parts of the sample toward the center lower strain regions. Strain localization associated with reaction is commonly observed in nature and experiment (e.g. White et al., 1990; Vissers et al., 1991; Newman et. al., 1999; Holyoke & Tullis, 2006; de Ronde and Stünitz, 2007). The work presented in this thesis suggests that, in many instances, reaction associated strain localization may be related to initially sluggish reaction kinetics, and/or rapidly overstepping reactions where the application of strain is responsible for reaction initiation and subsequent propagation of reaction across the sample volume. Further experiments within the respective stability fields of the polymorphs are necessary to fully define their rheologic differences as well as determine the dependence of the amount of polymorphic transformation on the degree of reaction overstepping versus the amount of strain endured by the sample to fully define the relationship of strain to transformation. CONTROLS ON REACTION TEXTURE DEVELOPMENT AND HETEROGENEITY 239 The development of metamorphic reaction textures in rocks is associated with either rapid change with respect to pressure, temperature and composition, and/or the inability of a rock volume to equilibrate due to relatively slow kinetic processes. Strain can aid in the initiation and equilibration of a rock volume. However, where strain is negligible or partitioned to weaker surrounding lithologies, as is the case with the orthoamphibole- cordierite rocks (Chapters 3 and 4), disequilibrium can occur at multiple scales and result in the development of metamorphic reaction textures. Layered coronal and symplectitic reaction textures are an example of largely diffusionally controlled structures that form in response to overstepping of mineral equilibria. Symplectitic coronal reaction textures consist of complex, typically vermicular, intergrowths between two or more phases. Symplectitic coronas in OCG associated with sillimanite porphyroblasts from the Thor-Odin dome, British Columbia, Canada, display extreme heterogeneity with respect to the distribution of phase assemblages and thicknesses of symplectitic layers about the central sillimanite. This heterogeneity is related to an increase in the size of the diffusive system during symplectitic corona growth. In addition, the work presented in this thesis suggests that the evolution in the size of the diffusive system was related to evolution along a decompression path. The changing size of the diffusive system is explicitly shown in two and three dimensions by the development of interconnected coronal networks linking nearly all reaction textures associated with both garnet and sillimanite. These observations illustrate that diffusion during the development and evolution of coronal reaction textures was focused and not equally distributed about the circumference of reacting porphyroblasts as is typically 240 interpreted. These observations have a wide-range applicability to the study of metamorphic petrology in general. The symplectitic coronal reaction textures presented in this study are an excellent example of how changing length-scales of diffusion are recorded in phase assemblage heterogeneity and interconnectivity linking different reactions in a rock volume during the progression of reaction. However, these observations are interesting not only for their implications for metamorphic reaction dynamics, but also for how effective bulk composition of a rock evolves with progressive reaction and movement along a P-T path. The diffusive structure of a system controls the extent that a rock volume can communicate chemically. The result of the limited, but interconnected diffusive structure and by extension the effective bulk composition of the system could explain inconsistencies between thermodynamic equilibrium phase diagram predictions of stable phase assemblages and what is actually preserved (e.g. Chapter 3). Reaction dynamics in natural samples can be difficult to resolve due to complex textural relationships and microstructures. However, detailed analysis reveals that the complexities themselves are clues to identifying the controls on the initiation and evolution of reactions and associated textural relationships. Although we have been able to provide evidence that reaction textures can form and evolve along a P-T path rather than only forming statically, we cannot quantify the rate at which the reaction texture formed or the rate of the P-T path responsible for its formation. Future studies of reaction textures should include an experimental component to understand the how varying different variables affects the resultant texture as well as 241 quantifying intergranular diffusivities in multicomponent-multiphase systems. Such data is necessary to fully develop the potential of reaction textures to be used to reconstruct the rate of both reaction and tectonic processes related to their formation. In addition, detailed experimental studies can also investigate how the product phase morphology evolves with changing thermodyanamic and kinetic parameters. EBSD data presented in this study shows a systematic relationship between intergrown phases with respect to their symmetry and mutual crystallographic alignment. These data can be used to aid in the development of growth models for coronal texture formation. 242 References Alverez-Velaro, A.M., Cesare, B., & Kriegsman, L.M., 2007. Formation of spinel- cordierite-feldspar-glass coronas after garnet in metapelitic xenoliths: reaction modelling and geodynamic implications. Journal of Metamorphic Geology, Arnold, J., Sandiford, M., & Wetherley, S., 1995. Metamorphic events in the eastern Arunta Inlier, Part 1. Metamorphic Petrology. Precambrian Research, 71, 183- 205. Ashworth JR, Sheplev, V.S., Bryxina, N.A., Kolobov, V.Y. & Reverdatto, V.V., 1998. Diffusion-controlled corona reaction and overstepping of equilibrium in a garnet granulite, Yenisey Ridge, Siberia. Journal of Metamorphic Geology, 16, 231-246 Ashworth, J. R. & Birdi, J. J., 1990. Diffusion modelling of coronas around olivine in an open system. Geochimica et Cosmochimica Acta, 54, 2389–2401 Ashworth, J. R., 1993, Fluid-absent diffusion kinetics of Al inferred from retrograde metamorphic coronas: American Mineralogist, v. 78, no. 3-4, p. 331-337.; Ashworth, J. R., and Chambers, A. D., 2000, Symplectic reaction in olivine and the controls of intergrowth spacing in symplectites: Journal of Petrology, v. 41, no. 2, p. 285-304. Ashworth, J. R., Sheplev, V. S., Khlestov, V. V. & Ananyev, V. A., 2001. Geothermobarometry using minerals at non-equili- brium: a corona example. European Journal of Mineralogy, 13, 1153–1161. 243 Ashworth, J.R., Sheplev, V.S., Bryxina, N.A., Kolobov, V.Y. & Reverdatto, V.V., 1998. Diffusion-controlled corona reaction and overstepping of equilibrium in a garnet granulite, Yenisey Ridge, Siberia. Journal of Metamorphic Geology, 16, 231-246. Audibert, N., Betrand, P., Hensen, B. J., Kienast, J. R. & Ouzegane, K. (1993) Cordierite- K-feldspar-quartz-orthopyroxene symplectite from southern Algeria; new evidence for osumilite in high-grade metamorphic rocks. Mineralogical Magazine, 57 (387), p. 354-357 Baldwin, J.A., Powell, R., Williams, M.L. & Goncalves, P., 2007. Formation of eclogite, and reaction during exhumation to mid-crustal levels, Snowbird tectonic zone, western Canadian Shield. Journal of Metamorphic Geology, 25, 953-974. Barnhoorn A., Bystricky M., Kunze K. Burlini L., Burg, J.P. 2005. Strain localisation in bimineralic rocks: Experimental deformation of synthetic calcite-anhydrite aggregates. Earth Planet Sc Lett 240, 748-763. Barrow, G. 1893. On an intrusion of muscovite-biotite gneiss in the east highlands of Scotland, and its accompanying metamorphism. Journal of the Geological Society of London, 49, 330-358. Baxter, E.F. & De Paolo, D.J., 2004. Can metamorphic reactions proceed faster than bulk strain? Contrib. to Min. Pet., 146, 657-670. Bohlen, S.R., Wall, V.J., & Boettcher, A.L. 1983. Experimental investigation and applicationof garnet granulite equilibria, Contributions to Mineralogy and Petrology, 83, 52-61. 244 Boland, J.N., and Otten, M.T. (1985) Symplectitic augite; evidence for discontinuous precipitation as an exsolution mechanism in Ca-rich clinopyroxene. Journal of Metamorphic Geology, 3(1), 13-20. Boland, J.N., B.E. Hobbs, and A.C. McLaren, 1977. The defect structure in natural and experimentally deformed kyanite, Phys. Status Solidi A, 39, 631-641. Brady J.B. (1975) Reference frames and diffusion coefficients. Amer. J. Sci. 275, 954- 983. Brodie, K.H. and Rutter, E.H., 1985 On the relationship between deformation and metamorphism, with special reference to the behaviour of basic rocks. In: A.B. Thompson and D.C. Rubie, Editors, Metamorphic Reactions: Kinetics, Textures and Deformation, Springer-Verlag, New York, 138–179. Brown, M. & Solar, G. S., 1999. The mechanism of ascent and emplacement of granite magma during transpression: a syntectonic granite paradigm. Tectonophysics, 312, Brown, M., 2002. Retrograde processes in migmatites and granulites revisited. Journal of Metamorphic Geology, 20, 25-40. Brown, R.L. (2004): Thrust-belt accretion and hinterland underplating of orogenic wedges; an example from the Canadian Cordillera. In Thrust Tectonics and Hydrocarbon Systems (K.R. McClay, ed.). Am. Assoc. Petroleum Geol., Mem. 82, 51-64 245 Bryxina, N. A., 1998, Textures of diffusion-controlled reaction in contact- metamorphosed Mg-rich granulite, Kokchetav area, Kazakhstan: Mineralogical Magazine, v. 62, no. 2, p. 213-224. Buick, I.S., Hermann, J., Williams, I.S. & Gibson, R.L., 2006. A SHRIMP U–Pb and LA- ICP-MS trace element study of the petrogenesis of garnet–cordierite– orthoamphibole gneisses from the Central Zone of the Limpopo Belt, South Africa. Lithos, 88, 150-172. Bunge, H., 1982. Texture Analysis in Materials Science: Mathematical Models. Butterworths, London. 593 pp Burt, J.B., N.L. Ross, R.J. Angel, and M. Koch, 2006. Equations of state and structures of andalusite to 9.8 GPa and sillimanite to 8.5 GPa, Am. Mineral., 91, 319-326. Caballero, F. G., Capdevila, C., and De Andres, C. G., 2001, Modelling of isothermal formation of pearlite and subsequent reaustenitisation in eutectoid steel during continuous heating: Materials Science and Technology (UK), v. 17, no. 6, p. 686- 692 Cahn, J. W., 1957. Nucleation on dislocations. Acta Met. 5, 169–172. Capdevila, C., Caballero, F. G., and De Andres, C. G.,Kinetics model of isothermal pearlite formation in a 0.4C-1.6Mn steel: Acta Materialia (USA), v. 50, no. 18, p. 4629-4641. Carey, W.J., Rice, J.M. & Grover, T.W. 1992. Petrology of aluminous schist in the Boehls Butte region of northern Idaho: geologic history and aluminum-silicate phase relations. Am. J. of Science, 292, 455-473. 246 Carlson W. D. and Johnson C. D. ( 1991 ) Coronal reaction textures in garnet amphibolites of the Llano Uplift. Amer. Mineral. 76, 756-772. Carlson, W. D., 1989. The significance of intergranular diffusion to the mechanisms and kinetics of porphyroblast crystallization. Contributions to Mineralogy and Petrology, 103, 1–24 Carlson, W.D., & Johnson, C.D., 1991. Coronal reaction textures in garnet amphibolites of the Llano Uplift. American Mineralogist, 76, 756-772. Carmichael, D.M., 1969; On the Mechanism of Prograde Metamorphic Reactions in Quartz-Bearing Pelitic Rocks, Contributions to Mineralogy and Petrology, no. 20 p. 244-267 Carr, S.D. and Simony, P.S. 2006. Ductile thrusting vs. channel flow in the southeastern Canadian Cordillera: Evolution of a coherent crystalline thrust sheet. In Channel flow, ductile extrusion and exhumation of lower-mid-crust in continental collision zones. Edited by R. Law, M. Searle, and L. Godin. Geological Society, London, Special Publications Carr, S.D., 1992. Tectonic setting and U-Pb geochronology of the Early Tertiary Ladybird leucogranite suite, Thor-Odin – Pinnacles area, southern Omineca Belt, British Columbia. Tectonics, 11, 258-278. Cavosie, A., Z.D. Sharp, and J. Selverstone, 2002. Co-existing aluminum silicates in quartz veins: A quantitative approach for determining andalusite-sillimanite equilibrium in natural samples using oxygen isotopes, Am. Mineral., 84, 417-423. 247 Cesare, B., M.T. Gomez-Pugnaire, A. Sanchez-Navas, and B. Groberty, 2002. Andalusite-sillimanite replacement (Mazarron, SE Spain): A microstructural and TEM study, Am. Mineral., 87, 433-444. Clarke, G.L., & Powell, R. (1991) Decompressional coronas and symplectites in granulites of the Musgrave Complex, central Australia. Journal of Metamorphic Geology, 9, 441-450. Coggon R, & Holland TJB 2002 Mixing properties of muscovite-celadonite- ferroceladonite-paragonite micas and revised garnet-phengite thermobarometers. Journal of Metamorphic Geology, 20,683–696 Cooper, R.F., 1990. Differential stress-induced melt migration-an experimental approach. J. Geophysics Research B. 95, 6979-6992. Dasgupta S., Sengupta P., Ehl J., & Raith M.M., 1999. Petrology of gedrite-bearing rocks in mid-crustal ductile shear zones from the Eastern Ghats Belt, India. Journal of Metamorphic Geology, 17, 765-778. Davies, H.L., and Warren, R.G. (1992) Eclogites of the D'Entrecasteaux Islands. Contributions to Mineralogy and Petrology, 112(4), 463-474 de Ronde, A. A., Heilbronner, R., Stunitz, H., Tullis, J., 2004. Spatial distribution of deformation and mineral reaction in experimentally deformed plagioclase-olivine aggregates. Tectonophysics 389 (1-2), 93–109. de Ronde, A.A., Stunitz, H., 2007. Deformation-enhanced reaction in experimentally deformed plagioclase-olivine aggregates Cont. Min. Pet., 153, 699-717. 248 Delle Piane C, Burlini, L., Grobety, B., 2007. Reaction-induced strain localization: Torsion experiments on dolomite. Earth and Planetary Sc. Lett. 256, 36-46. Denison, C., Carlson, W. D. & Ketcham, R. A., 1997. Three-dimensional quantitative textural analysis of metamorphic rocks using high-resolution computed X-ray tomography.Part I. Methods and techniques. Journal of Metamorphic Geology, 15, 29–44. Diener, J.FA., Powell, R., & White, R.W., 2008. Quantitative phase petrology of cordierite–orthoamphibole gneisses and related rocks. Journal of Metamorphic Geology, Diener, J.FA., Powell, R., White, R.W., & Holland, T.J.B., 2007. A new thermodynamic model for clino- and orthoamphiboles in the system Na2O–CaO–FeO–MgO– Al2O3–SiO2–H2O–O. Journal of Metamorphic Geology, 24, 631-656. Dimanov, A., Rybacki, E., Wirth, R., Dressen, G., 2007. Creep and strain-dependent microstructures of synthetic anorthite-diopside aggregates. J Struct Geol 29, 1049-1069 Doukhan, J.-C., and J. Paquet, Plastic deformation of andalusite single crystal Al2SiO5, Bull. Mineral, 105, 170-175, 1982. Doukhan, J.-C., and J.M. Christie, 1982. Plastic deformation of sillimanite Al2SiO5 single crystals under confining pressure and TEM investigation of the induced defect structure, Bull. Mineral., 105, 583-589. 249 Doukhan, J.-C., N. Doukhan, P.S. Koch, and J.M. Christie, 1985. Transmission electron microscopy investigation of lattice defects in Al2SiO5 polymorphs and plasticity induced polymorphic transformations, Bull. Mineral, 108, 81-96. Droop, G.T.R. (1989) Reaction history of garnet-sapphirine granulites and conditions of Archaean high-pressure granulite-facies metamorphism in the Central Limpopo mobile belt, Zimbabwe. Journal of Metamorphic Geology, 7, 383-403. Elvevold, S., & Gilotti, J.A. (2000) Pressure-temperature evolution of retrogressed kyanite eclogites, Weinschenk Island, North-East Greenland Caledonides. Lithos, 53, 127-147 Eskola, P. 1915. On the relations between the chemical and mineralogical composition in the metamorphic rocks of the Orijarvi region. Bulletin of the Commision of Geology, FInlande, 44, 1-277. Evans, A.G., 1978. Microfracture from thermal-expansion anisotropy 1 Single phase systems. Acta Metall., 26, 1845-1853. Evans, T.P., 2004. A method for calculating effective bulk composition modification due to crystal fractionation in garnet-bearing schist: implications for isopleth thermobarometry. Journal of Metamorphic Geology, 22, 547-557. Field, S.W. 2008. Diffusion, discontinuous precipitation, metamorphism, and metasomatism: The complex history of South African upper-mantle symplectites. American Mineralogist, 2008, 4, 618-631. Fischer, H., Schreyer, W., & Maresch W.V., 1999. Synthetic gedrite: a stable phase in the system MgO-Al2O3-SiO2-H2O (MASH) at 800˚C and 10 kbar water pressure, 250 and the influence of FeNaCa impurities. Contributions to Mineralogy and Petrology, 136, 184-191. Fisher G. W. (1973) Nonequilibrium thermodynamics as a model for diffusion-controlled metamorphic processes. Amer. J. Sci. 273, 897-924. Fisher G. W. (1975) The thermodynamics of dillusion-controlled metamorphic processes. In Mass Transport processes in ceramics. (editors A. R. Cooper and A. Flinn, D., 1965. Deformation in metamorphism. In: Pitcher, W.S. and Flinn, G.W., (Eds.) Controls of Metamorphism, Oliver and Boyd, Edinburgh, 46–72. Fockenberg & Schreyer, 1994. Stability of yoderite in the absence and in the presence of quartz – an experimental study in the system MgO-Al2O3-Fe2O3-SiO2-H2O. Journal of Petrology, 35, 1341-1375. Foster, C. T., Jr., 1986, Thermodynamic models of reactions involving garnet in a sillimanite/ staurolite schist, in Yardley, and D ; Harte, eds., Mechanisms of metamorphic reactions.: Mineralogical Magazine: London, United Kingdom, Mineralogical Society, p. 427-439 Friedrich, A., M. Kunz, B. Winkler, and T. Le Bihan, 2004. High-pressure behavior of sillimanite and kyanite: compressibility, decomposition and indications of a new high-pressure phase, Z. Kristall., 219, 324-329. García-Casco, A., and R.L. Torres-Roldán, 1996. Disequilibrium induced by fast decompression in St-Bt-Grt-Ky-Sil-And metapelites from the Betic belt (Southern Spain), J. Petrol., 37, 1207-1239. 251 Ghent, E.D., Nicholls, J., Stout, M.Z., Rottenfusser, B., 1977. Clinopyroxene amphibolite boudins from Three Valley Gap, British Columbia. Canadian Mineralogist, 15, 269 – 282. Goldman, D.S., & Albee, A.L., 1977. Correlation of Mg/Fe partitioning between garnet and biotite with O18/O16 partitioning between quartz and magnetite. American Journal of Science, 277, 750-761. Gordon SM, Whitney DL, Teyssier C, & Grove M, 2008. Timescales of migmatization, melt crystallization, and cooling in a Cordilleran gneiss dome: Valhalla complex, southeastern British Columbia. Tectonics, 27, TC4010. Grambling, J.A., 1981. Kyanite, andalusite, sillimanite, and related mineral assemblages in the Truchas Peaks region, New Mexico, Am. Mineral., 66, 702-722. Grant S. M. (1988) Diffusion models for corona formation in metagabbros from the Western Grenville Province, Canada. Contrib. Mineral. Petrol. 98, 49-63. Grantham, G.H., Thomas, R.J., Eglington, B.M., de Bruin, D., Atanasov, A., and Evans, M.J., 1993, Corona textures in Proterozoic olivine melanorites of the Equeefa Suite, Natal Metamorphic Province, South Africa: Mineralogy and Petrology, v. 49, p. 91-102. Greenwood, H.J., 1972. AlIV-SiIV disorder in sillimanite and its effect on phase relations of the aluminum silicate minerals, Geol. Soc. Am. Mem., 132, 553-571. Guiraud, M., Powell, R. & Rebay, G., 2001. H2O in metamorphism and unexpected behaviour in the preservation of metamorphic assemblages. Journal of Metamorphic Geology, 19, 445–454. 252 Guiraud, M., Powell, R., & Cottin, J.-Y., 1996. Hydration of orthopyroxene-cordierite- bearing assemblages at Laouni, Central Hoggar, Algeria. Journal of Metamorphic Geology, 14, 467-476. Harley, S. L., 1985, Garnet-orthopyroxene bearing granulites from Enderby Land, Antarctica; metamorphic pressure-temperature-time evolution of the Archaean Napier Complex: Journal of Petrology, v. 26, no. 4, p. 819-856, Harley, S.L. 1986. A sapphirine-cordierite-garnet-sillimanite granulite from Enderby Land, Antarctica; implications for FMAS petrogenetic grids in the granulite facies. Contributions to Mineralogy and Petrology, 94(4), 452-460 Harley, S.L., Hensen, B.J., and Sheraton, J.W. (1990) Two-stage decompression in orthopyroxene-sillimanite granulites from Forefinger Point, Enderby Land, Antarctica; implications for the evolution of the Archaean Napier Complex. Journal of Metamorphic Geology, 8(6), 591-613 Harvey, J.L., Hoisch, T.D., 1994. Sapphirine-bearing amphibolites in the Okanogan Complex, Washington: thermobarometry and tectonic implications. Abstracts with Programs—Geol. Soc. Am. 26, 57. Helgeson, H.C. Delany, J.M., Nesbitt, H.W. & Bird, D.K. 1978. Summary and critique of the thermodynamic properties of rock forming minerals. Am. J. of Science, 274, 1089-1198. Hillert, M. 1972 Theories of Growth During Discontinuous Precipitation. Metallurgical Transactions A 2729-274 253 Hinchey, A.M. & Carr, S.D., 2007. Protolith composition of cordierite-gedrite basement rocks and garnet amphibolite of the Bearpaw Lake are of the Thor-Odin Dome, Monashee Complex, British Columbia, Canada. The Canadian Mineralogist, 45, 607-629. Hinchey, A.M., Carr, S.D., McNeill, P.D., & Rayner, N., 2006. Paleocene–Eocene high- grade metamorphism, anatexis, and deformation in the Thor–Odin dome, Monashee complex, southeastern British Columbia. Canadian Journal of Earth Sciences, 43, 1341-1365. Hirsch, D. M., Ketcham, R. A. & Carlson, W. D., 2000. An evaluation of spatial correlation functions in textural analysis of metamorphic rocks. Geological Materials Research, 2, 1–42Prior, D. J., Wheeler, J., Brenker, F. E., Harte, B. & Matthews, M., 2000. Crystal plasticity of natural garnets: new microstructural evidence. Geology, 28, 1003–1006. Holdaway, M.J., 1971. Stability of andalusite and the aluminum silicate phase diagram. Am. J. Sci., 271, 97-131. Holdaway, M.J., 1978. Significance of chloritoid-bearing and staurolite-bearing rocks in the Picuris Range, New Mexico, Geol. Soc. Am. Bull., 89, 1404-1414. Holland, T. J. B. & Powell, R., 1998. An internally-consistent thermodynamic data set for phases of petrological interest. Journal of Metamorphic Geology, 16, 309–343. Holland, T. J. B. & Powell, R., 2006. Mineral activity–compo-sition relations and petrological calculations involving cation equipartition in multisite minerals: a logical inconsistency. Journal of Metamorphic Geology, 24, 851–861. 254 Holm, J.L., and O.J. Kleppa, 1966. The thermodynamic properties of the aluminum silicates. Am. Mineral., 51, 1608-1627. Hölttä, P., 1997. Geochemical characteristics of granulite facies rocks in the Archean Varpaisjarvi area, central Fennoscandian Shield. Lithos, 40, 31-53. Holtzman, B.K., Groebner, N.J., Zimmerman, M.E., Ginsberg, S.B., Kohlstedt, D.L., 2003. Stress-driven melt segregation in partially molten rocks. Geochem Geophy Geosy. 4, Holyoke, C.W. & Tullis, J., 2006. Formation and maintenance of shear zones. Geology, 34, 105-108. Hudson, N., & Harte, B., 1985. K (sub 2) O-poor, aluminous assemblages from the Buchan Dalradian, and the variety of orthoamphibole assemblages in aluminous bulk compositions in the amphibolite facies. American Journal of Science, 285, 224-266. Joesten R. (1977) Evolution of mineral assemblage zoning in diffusion metasomatism. Geochim. Cosmochim. Acta 41, 649-670. Joesten R. (1986) The role of magmatic reaction, diffusion, and annealing in the evolution of coronitic microstructure in troctolitic gabbro from Risrr, Norway. Mineral. Mag. 50, 441-467. Johnson C.D. and Carlson W.D. (1990) The origin of olivine-plagioclase coronas in metagabbros from the Adirondack Mountains, New York. J. Metam. Geol. 8, 697- 717 255 Johnson, T., Brown, M., Gibson, R., & Wing, B. (2004) Spinel–cordierite symplectites replacing andalusite: evidence for melt-assisted diapirism in the Bushveld Complex, South Africa. Journal of Metamorphic Geology, 22, 529-545 Johnston, A.D., & Stout, J.H. (1984) Development of orthopyroxene-Fe/Mg ferrite symplectites by continuous olivine oxidation. Contributions to Mineralogy and Petrology, 88, 196-202. Johnston, D.H., Williams, P.F., Brown, R.L., Crowley, J.L., and Carr, S.D. 2000. Northeastward extrusion and extensional exhumation of crystalline rocks from the Monashee complex, southeastern Canadian Cordillera. Journal of Structural Geology, 22: 603-625. Keller, L.M., Wunder, B., Rhede, D., and Wirth, R.,. 2008. Component mobility at 900°C and 18 kbar from experimentally grown coronas in a natural gabbro. Geochemica et Cosmochimica Acta, 72, 4307-4322. Kerrick, D.M., 1986. Dislocation strain energy in the Al2SiO5 polymorphs, Phys. Chem. Mineral., 13, 221-226. Kerrick, D.M., 1988. Al2SiO5–bearing segregations in the Lepontine Alps, Switzerland: Aluminum mobility in metapelites, Geology, 16, 636-640. Kerrick, D.M., 1990. The Al2SiO5 polymorphs, Rev. Mineral., Mineralogical Society of America, Washington, D.C. King, D.S., Kohlstedt, D.L., Zimmerman, M.E., 2007. Stress-Driven Melt Segregation and Shear Localization in Partially Molten Aggregates: Experiments in Torsion. Eos Trans, AGU, 88, Fall Meet. Suppl., Abstract T43D-07 256 Korzhinskii D. S. (1959) Physicochemical Basis of the Analysis of the Paragenesis of Minerals. Consultants Bureau USA Koshimoto S, Tsunogae T, Santosh M., 2004.Sapphirine and corundum bearing ultrahigh temperature rocks from the northern domain of Palghat-Cauvery Shear System, southern India. Journal of Mineralogy & Petrology Science, 99,298–310. Kriegsman LM 2001 Partial melting, partial melt extraction and partial back reaction in anatectic migmatites. Lithos 56, 75–96. Kruckenberg, S.C., Whitney, D.L., Teyssier, C., Fanning, C.M., & Dunlap, W.J., 2008. Paleocene-Eocene migmatite crystallization, extension, and exhumation in the hinterland of the northern Cordillera: Okanogan dome, Washington, USA. Geological Society of America Bulletin, 120, 912-929 Kruse, S., McNeill, P. D. & Williams, P. F., 2004. A geological compilation map of the Thor-Odin dome. www.unb.ca/fredericton/science/geology/monashee. Labotka, T.C., & Kath, R.L., 2001. Petrogenesis of the contact-metamorphic rocks beneath the Stillwater Complex, Montana. Geological Society of America Bulletin, 113, 1312-1323. Lal, R. K. & Shukla, R. S., 1975. Genesis of cordierite–gedrite–cummingtonite rocks from the northern portion of the Khetri Copper Belt, Rajasthan, India. Lithos, 8, 175–186. Lambregts, P.J., and H.L.M. van Roermund, 1990. Deformation and recrystallization mechanisms in naturally deformed sillimanites. Tectonophysics, 179, 371-378. 257 Lang, H.M., Wachter, A.J., Peterson, V.L., & Ryan, J.G., 2004. Coexisting clinopyroxene/spinel and amphibole/spinel symplectites in metatroctolites from the Buck Creek ultramafic body, North Carolina Blue Ridge. American Mineralogist, 89, 20-30. Lefebvre, A. and D. Menard, 1981. Stacking faults and twins in kyanite, Al2SiO5, Acta Crystall., A37, 80-84. Lehnert, K., Su, Y., Langmuir, C., Sarbas, B., & Nohl, U. 2000. A global geochemical database structure for rocks. Geochem. Geophys. Geosyst. 1, doi:10.1029/1999GC00026 Marshall, D., & Simandl G., 2006, Phase relations and metamorphism in the sapphirine bearing granulites of the Valhalla complex, Slocan Valley, BC, Geol. Assoc. Can./Mineral. Assoc. Can., Annual Meet. Abstr. Program, 31, 96. McClintock, M.K., & Cooper, A.F., 2003. Geochemistry, mineralogy, and metamorphic history of kyanite-orthoamphibole-bearing Alpine Fault mylonite, South Westland, New Zealand. New Zealand Journal of Geology & Geophysics, 46, 47- 62. Meike, A., 1993. A critical review of investigations into transformation plasticity. In: Boland, J. N., Fitz Gerald, J. D. (Eds.), Defects and Processes in the Solid State: Bibliography 173 Geoscience Applications (The McLaren Volume). Vol. 14. Elsevier, Amsterdam, pp. 525. 258 Menard, D., J.-C. Doukhan, and J. Paquet, 1979. Uniaxial compression of kyanite Al2O3-SiO2: deformation mechanisms of minerals and rocks, Bull. Mineral., 102, 159-162. Messiga, B., and Bettini, E. (1990) Reactions behaviour during kelyphite and symplectite formation; a case study of mafic granulites and eclogites from the Bohemian Massif. European Journal of Mineralogy, 2(1), 125-144 Milke, Dohmen, R., Becher, H-W., Wirth, R., 2007. Growth kinetics of enstatite reaction rims studied on nano-scale, Part I: Methodology, microscopic observations and the role of water. Contributions to Mineralogy and Petrology, Mitra, G., 1978. Ductile deformation zones and mylonites: the mechanical processes involved in the deformation of crystalline basement rocks. Am. J. Sci. 278, 1057– 1084. Mongkoltip, P., and Ashworth, J. R., 1983, Quantitative estimation of an open-system symplectite-forming reaction; restricted diffusion of Al and Si in coronas around olivine: Journal of Petrology, v. 24, no. 4, p. 635-661 Moore, J.M., & Waters, D.J., 1990. Geochemistry and origin of cordierite- orthoamphibole/orthopyroxene-phlogopite rocks from Namaqualand, South Africa. Chemical Geology, 85, 77-100. Munz, I.A., 1990. Whiteschists and orthoamphibole-cordierite rocks and the P-T-t path of the Modum Complex, South Norway. Lithos, 24, 181-200. 259 Norlander, B.H., Whitney, D.L., Teyssier, C. & Vanderhaeghe, O., 2002. Partial melting and decompression of the Thor-Odin dome, Shuswap metamorphic core complex, Canadian Cordillera. Lithos, 6, 103-125. Nyman, M. W., Pattison, D. R. M. & Ghent, E. D., 1995. Melt extraction during formation of K-feldspar+sillimanite migmatites, west of Revelstoke, British Columbia. Journal of Petrology, 36, 351-372. O’Neil, H.S., & Wood, B.J., 1979. An experimental study of Fe-Mg partitioning between garnet and olivine and its calibration as a geothermometer. Contributions to mineralogy and Petrology, 72, 337. Ohuchi, F.S., Ghose, S., Engelhard, M.H., and Baer, D.R., 2006. Chemical bonding and electronic structures of the Al2SiO5 polymorphs, andalusite, sillimanite, and kyanite: X-ray photoelectron and electron energy loss spectroscopy studies. Am. Min., 91, 740-746. Ortoleva, P., Merino, E. Moore, C., & Chadam, J., 1987. Geochemical Self-organization I: Reaction-transport feedbacks and modeling approach. American Journal of Science, 287, 979-1007. Ouzegane, K, Djemai, S., & Guiraud, M., 1996. Gedrite-garnet-sillimanite-bearing granulites from Amessmessa area, south In Ouzzal, Hoggar, Algeria. Journal of Metamorphic Geology, 14, 739-753. Page, K., Maini, P.K., & Monk, N.A.M., 2003. Pattern formation in spatially heterogeneous Turing, reaction-diffusion models. Physica D: Nonlinear Phenomena, 181, 80-101. 260 Pan, Y., & Fleet, M.E., 1995. Geochemistry and origin of cordierite-orthoamphibole gneiss and associated rocks at an Archean volcanogenic massive sulphide camp: Manitouwadge, Ontario, Canada. Precambrian Research, 74, 73-79. Parrish, R.R., Carr, S.D., and Parkinson, D.L. 1988. Eocene extensional tectonics and geochronology of the southern Omineca Belt, British Columbia and Washington. Tectonics, 7, 181-212. Passchier, C. W., and Trouw, R. A. J., 1996, Microtectonics: New York, Springer-Verlag, 290 p Patel, S.C., Behera, S., Mohanty, A., Singh, A.K., & Patel, S.K. (2001) Metamorphic history of sapphirine- and relict-kyanite-bearing Mg-Al rich rocks at Hatimunda Hill near Junagarh, Kalahandi District, Orissa. Journal of the Geological Society of India, 57, 417-427 Paterson, M.S., and D.L. Olgaard, 2000. Rock deformation tests to large shear strains in torsion, J. Struct. Geol., 22, 1341-1358. Pattison, D.R.M., 1992. Stability of andalusite and sillimanite and the Al2SiO5 triple point: Constraints from the Ballachulish Aureole, Scotland, J. Geol., 100, 423- 446. Pattison, D.R.M., 2001. Instability of the Al2SiO5 “triple-point” assemblage in muscovite + biotite + quartz-bearing metapelites, with implications, Am. Mineral., 86, 1414-1422. Peck, W.H., & Smith, M.S., 2005. Cordierite-gedrite rocks from the Central Metasedimentary Belt boundary thrust zone (Grenville Province, Ontario): 261 Mesoproterozoic metavolcanic rocks with affinities to the Composite Arc Belt. Canadian Journal of Earth Science, 42, 1815-1828. Peck, W.H., & Valley, J.W., 2000. Genesis of Cordierite-Gedrite Gneisses, Central Metasedimentary Belt boundary thrust zone, Grenville Province, Ontario, Canada. The Canadian Mineralogist, 38, 511-524. Penn, R.L., J.F. Banfield, and D.M. Kerrick, 1999. TEM investigation of Lewiston, Idaho, fibrolite: Microstructure and grain boundary energetics, Am. Mineral., 84, 152-159. Perkins, D. 1987. Two independent garnet-clinopyroxene-plagioclase-quartz barometers. Geological Society of America Abstracts, 19, 803. Pieri M, Burlini, L., Kunze, K., Stretton, I., 2001. Rheological and microstructural evolution of Carrara marble with high shear strain: results from high temperature torsion experiments. Journ. of Struct. Geol. 23, 1393-1413. Pitra, P. & De Waal, S. A. (2001) High-temperature, low-pressure metamorphism and development of prograde symplectites, Marble Hall Fragment, Bushveld Complex (South Africa). Journal of Metamorphic Geology, 19 (3), 311-325. Powell, R. & Holland, T. J. B. & Worley, B., 1998. Calculating phase diagrams involving solid solutions via non-linear equations, with examples using THERMOCALC. Journal of Metamorphic Geology, 6, 173–204. Raith, M.M., Rakotondrazafy, r., & Sengupta, P., 2008. Petrology of corundum-spinel- sapphirine-anorthite rocks (sakenites) from the type locality in southern Madagascar. Journal of Metamorphic Geology, 25, 647-667. 262 Raleigh, C.B., 1965. Glide mechanisms of experimentally deformed minerals, Science, 150, 739-741. Rao, M.N., S.L. Chaplot, N. Choudhury, K.R. Rao, R.T. Azuah, W.T. Montfrooij, and S.M. Bennington, 1999. Lattice dynamics and inelastic scattering from sillimanite and kyanite Al2SiO5, Phys. Rev. B, 60, 12061-12068. Reesor, J.E. & Moore, J.M., Jr., 1971. Petrology and structure of the Thor-Odin Gneiss Dome, Shuswap Metamorphic complex, British Columbia. Geological Survey of Canada Bulletin, 195. Renard, F., Gratier, J-P. Ortoleva, P. Brosse, E., & Bazin, B., 1998. Self-organization during reactive fluid flow in a porous medium. Geophysical Research Letters, 25, 385-388. Rheinhardt, J., 1987. Cordierite-Anthophyllite rocks from north-west Queensland, Australia; metamorphosed magnesian pelites. Journal of Metamorphic Geology, 5, 451-472. Ribbe, P.H., 1980. Aluminum silicate polymorphs and other aluminum silicates. In: Ribbe, P.H. (ed.), Orthosilicates, Rev. Mineral., 5, 189-214. Ridley, J., and Thompson, A.B. (1986) The role of mineral kinetics in the development ofmetamorphic microtextures. Advances in Physical Geochemistry, 5, 154-193 Roberts, M. D., Oliver, N. H. S., Fairclough, M. C., Ho¨ltta¨, P. S. & Lahtinen, R., 2003. Geochemical and oxygen isotope signature of sea-floor alteration associated with a polydeformed and highly metamorphosed massive sulphide deposit, Ruostesuo, central Finland. Economic Geology, 98, 535–556. 263 Rubie, D. C., 1998. Disequilibrium during metamorphism: the role of nucleation kinetics. In: Treloar, P. J., O’Brien, P. J. (Eds.), What Drives Metamorphism and Metamorphic Reactions. Vol. 138. Geol. Soc., Spec. Pub., London, pp. 199–214 Rutter, E. H., Brodie, K. H., 1988a. Experimental ”syntectonic” dehydration of serpentinite under conditions of controlled pore water pressure. J. Geophys. Res. 93 (5), 4907–4932. Salje, E., 1986. Heat capacities and entropies of andalusite and sillimanite: the influence of fibrolitization on the phase diagram of the Al2SiO5 polymorphs, Am. Mineral., 71, 1366-1371. Satirini-Rideout, C., Gilotti, J.A., & Foster, C.T., 2006. Forward modeling corona growth in a partially eclogitized leucogabbro, Bourbon Island, North-East Greenland. Lithos, 56, 700-717. Schmidt, C., Bruhn, D., Wirth, R., 2003. Experimental evidence of transformation plasticity in silicates: minimum of creep strength in quartz. Earth. Planet. Sc. Lett. 205, 273–280. Schneiderman, J. & Tracy, R.J., 1991. Petrology of orthoamphibole-cordierite gneisses from the Orijarvi area, southwest Finland. Amercian Mineralogist, 76, 942-955. Schumacher, J.C., 1988. Stratigraphy and geochemistry of the ammonoosuc volcanics, central Massachusetts and southwestern New Hampshire. American Journal of Science, 288, 619-663. Sederholm, J.J. (1916) On synantetic minerals and related phenomena. Bulletin de la Commission Geologique de Finlande, 48, l-148. 264 Shimpo M, Tsunogae T, & Santosh M 2006 First report of garnet-corundum rocks from Southern India: implications for prograde high-pressure (eclogite-facies?) metamorphism. Earth Planet Science Letters 242:111–129. Skemer, P., Katayama, I., Jiang, Z., Karato, S., 2005. The misorientation index: Development of a new method for calculating the strength of lattice- preferred orientation. Tectonophysics. 41, 157-167. Snow, E., Yund, R. A., 1987. The effect of ductile deformation on the kinetics and mechanisms of the aragonite-calcite transformation. J. Metamorph. Geol. 5 (2), 141–153. Solorzano, I.G. & Purdy, G.R. 1984. Interlamellar spacing in discontinuous precipitation. Metallurgical Transactions A, 15, 1055-1063. Stahle, V., R. Altherr, M. Koch, and L. Nasdala, 2004. Shock-induced formation of kyanite (Al2SiO5) from sillimanite within a dense metamorphic rocks from the Ries crater (Germany), Contrib. Mineral. Petrol., 148, 150-159. Stuewe, K., 1998, The influence of effective bulk composition on retrograde assemblage development; I, Conceptual model and relevant phase diagrams, in Treloar, J ; O, and Brien, eds., Proceedings; What drives metamorphism and metamorphic reactions; heat production, heat transfer, deformation and kinetics? Extended abstracts.: Electronic Geology: Portsmouth, United Kingdom, Electronic Journals Limited Stunitz, H., Tullis, J., 2001. Weakening and strain localization produced by syndeformational reaction of plagioclase. Int. J. Earth Sci. 90 (1), 136–148. 265 Teall, J. J. H., 1885. The metamorphosis of dolerite into hornblende-schist. Quart. J. Geol. Soc. London 41, 133–145. Teyssier, C. & Whitney, D.L., 2002. Gneiss domes and orogeny. Geology, 30, 1139-1142 Teyssier, C., Ferré, E., Whitney, D.L., Norlander, B., Vanderhaeghe, O., and Parkinson, D. 2005. Flow of partially molten crust and origin of detachments during collapse of the Cordilleran orogen. Submitted to Geological Society of London, Symposium volume Channel flow, ductile extrusion and exhumation of lower- mid-crust in continental collision zones.InHigh-strain zones: Structures and Physical Properties. Edited by: Bruhn, D. and Burlini, L. Geological Society of London Special Publication, 245: 39-64 Teyssier, C., Ferré, E., Whitney, D.L., Norlander, B., Vanderhaeghe, O., and Parkinson, D. 2005. Flow of partially molten crust and origin of detachments during collapse of the Cordilleran orogen. Submitted to Geological Society of London, Symposium volume Channel flow, ductile extrusion and exhumation of lower- mid-crust in continental collision zones. In High-strain zones: Structures and Physical Properties. Edited by: Bruhn, D. and Burlini, L. Geological Society of London Special Publication, 245: 39-64. Thompson, A.B., 2001. Clockwise P – T paths for crustal melting and H2O recycling in granite source regions and migmatite terrains. Lithos, 56, 33-45. Thompson, J. B., 1959. Local equilibrium in metasomatic proesses. In Researches in Geochemistry (ed. Abelson, P. H.), pp. 427–457. Wiley, New York 266 Thompson, S.W., & Howell, P.R., 1988. On the early stages of pearlite formation in hypoeutectoid steels. Scripta Metallurgica, 22, 1775-1778. Tullis, J, 2002. Deformation of granitic rocks: Experimental studies and natural examples. In: Plastic Deformation of Minerals and Rocks. Karato, S, Wenk, H (eds). Rev. Mineral. & Geochem. 51, 51-95. Turner, S.P. & Stüwe, K. (1992) Low-pressure corona textures between olivine and plagioclase in unmetamorphosed gabbros from Black Hill, South Australia. Mineralogical Magazine, 56, 503-509 Vanderhaeghe, O. & Teyssier, C., 1997. Formation of the Shuswap metamorphic core complex during late-orogenic collapse of the Canadian Cordillera: Role of ductile thinning and partial melting of the mid- to lower-crust. Geodinamica Acta, 10, 42- 58. Vanderhaeghe, O., Teyssier, C. & Wysoczanski, R., 1999. Structural and geochronological constraints on the role of partial melting during the formation of the Shuswap metamorphic core complex at the latitude of the Thor-Odin Dome, British Columbia. Canadian Journal of Earth Sciences, 36, 917-943. Vanderhaeghe, O., Teyssier, C., McDougall, I., and Dunlap, D.W. 2003. Cooling and exhumation of the Shuswap metamorphic core complex constrainedby 40Ar/39Ar thermochronology. Geological Society of America Bulletin, 115: 200-216. Vaughn, M.T., and D.J. Weidner, 1978. The relationship of elasticity and crystal structure in andalusite and sillimanite, Phys. Chem. Mineral., 3, 133-144. 267 Vernon, R.H., 1987. Oriented growth of sillimanite in andalusite, Placitas-Juan Tabo area, New Mexico, U.S.A., Can. J. Earth Sci., 24, 580-590. Vissers, R. L. M., Drury, M. R., Hoogerduijn Strating, E. H., van der Wal, D., 1991. Shear zones in the upper mantle; a case study in an Alpine lherzolite massif. Geology 19 (10), 990–993. Watson EB, Price JD (2001) Kinetics of the reaction MgO + Al2O3 fi MgAl2O4 and Mg-Al interdiffusion in spinel at 1200– 2000°C and 1.0-4.0 GPa. Geochem Cosmochim Acta 66:2123–2138 Wenk, H.R. 1983. Mullite-sillimanite intergrowth from pelitic inclusions in Bergell tonalite. Neues Jahrbuch fur Mineralogie Abhandlungen, 146, 1-14. White R.W., Powell R., & Holland T.J.B., 2007. Progress relating to calculation of partial melting equilibria for metapelites. Journal of Metamorphic Geology, 25, 511-527. White, R.W., Powell, R., & Baldwin, J.A., 2008. Calculated phase equilibria involving chemical potentials to investigate the textural evolution of metamorphic rocks. Journal of Metamorphic Geology, 26, 181-198. White, R.W., Powell, R., & Clarke, G.L., 2002. The interpretation of reaction textures in Fe-rich metapelitic granulites of the Musgrave Block, central Australia; constraints from mineral equilibria calculations in the system K (sub 2) O-FeO- MgO-Al (sub 2) O (sub 3) -SiO (sub 2) -H (sub 2) O-TiO (sub 2) -Fe (sub 2) O (sub 3). Journal of Metamorphic Geology, 20, 41-55. White, S. H., Burrows, S. E., Carreras, J., Shaw, N. D., Humphreys, F. J., 1980. On mylonites in ductile shear zones. J. Struct. Geol. 2 (1-2), 175–187. 268 Whitney, D.L., 2002. Coexisting andalusite, kyanite, and sillimanite: sequential formation of three polymorphs during progressive metamorphism near the Al2SiO5 triple point, Sivrihisar, Turkey, Am. Mineral., 84, 405-416 Whitney, D.L., Goergen, E.T., Ketcham, R.A., & Kunze, K., 2008. Formation of garnet polycrystals during metamorphic crystallization, 28, 365-383. Williams, M.L., 1994. Sigmoidal inclusion trails, punctuated fabric development, and interactions between metamorphism and deformation. J. Metamorphic Geology, 2, 1-21. Williams, P.F., & Jiang, D. 2005. An investigation of lower crustal deformation: evidence for channel flow and its implications for tectonics and structural studies. Journal of Structural Geology, 27: 1486-1504. Winter, J.K., and S. Ghose, 1979. Thermal expansion and high temperature crystal chemistry of the Al2SiO5 polymorphs. Am. Mineral., 64, 573-586. Wintsch, R.P., Christofferson, R. and Kronenberg, A.K., 1995. Fluid-rock reaction weakening of fault zones., J. Geophys. Research B, 100, 13021-13032. Yang, H., R.M. Hazen, L.W. Finger, C.T. Prewitt, and R.T. Downs, 1997b. Compressibility and crystal structure of sillimanite, Al2SiO5, at high pressure, Phys. Chem. Mineral., 25, 39-47. Yang, H.X., R.T. Downs, L.W. Finger, R.M. Hazen, and C.T. Prewitt, 1997a. Compressibility and crystal structure of kyanite, Al2SiO5, at high pressure, Am. Mineral., 82, 467-474. 269 Yund, R.A., and Tullis J., 1991. Compositional changes of minerals associated with dynamic recrystallization. Cont. Min. Pet., 108, 346-355. 270 Appendix Representative microprobe analyses of phases in orthoamphibole-cordierite rocks from the Thor-Odin dome Re pr es en tat iv e m icr op ro be an aly se s o f b io tit e SiO 2 38 .05 37 .67 38 .53 37 .72 37 .65 7 37 .69 8 37 .82 6 Ti O 2 1.3 0 1.4 7 1.5 7 1.4 3 1.5 7 1.5 86 1.2 98 Al 2O 3 18 .72 19 .88 18 .94 19 .44 19 .61 6 19 .66 2 19 .83 Fe O 14 .94 14 .08 14 .02 14 .25 12 .36 4 12 .17 6 12 .60 4 M nO 0.0 3 0.0 3 0.0 4 0.0 0 0.0 31 0.0 06 0.0 24 Cr 2O 3 0.0 3 0.0 7 0.0 7 0.0 4 0.0 0 0.0 0 0.0 0 M gO 17 .17 15 .74 15 .90 15 .35 15 .91 8 15 .61 2 15 .76 1 Zn O 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 Ca O 0.0 1 0.3 6 0.5 5 0.1 2 0 0.0 14 0.0 38 N a 2O 0.0 3 0.3 8 0.3 1 0.1 5 0.3 85 0.3 4 0.3 91 K 2 O 5.9 0 7.7 8 7.4 3 8.0 2 8.9 86 8.9 8.8 49 to tal 96 .17 97 .46 97 .36 96 .51 96 .52 7 95 .99 4 96 .62 1 sa m pl e n o. 00 -5 .13 00 -5 .13 00 -5 .13 00 -5 .13 02 -3 .03 02 -3 .03 02 -3 .03 No rm ali ze d t o 2 2 O xy s Si4 + 5.4 7 5.3 9 5.5 0 5.4 5 5.4 3 5.4 6 5.4 5 Ti 4+ 0.1 4 0.1 6 0.1 7 0.1 6 0.1 7 0.1 7 0.1 4 Al 3+ 3.1 7 3.3 5 3.1 9 3.3 1 3.3 3 3.3 5 3.3 7 Fe 2+ 1.8 0 1.6 9 1.6 7 1.7 2 1.4 9 1.4 7 1.5 2 M n2 + 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 Cr 3+ 0.0 0 0.0 1 0.0 1 0.0 0 0.0 0 0.0 0 0.0 0 M g2 + 3.6 8 3.3 6 3.3 8 3.3 1 3.4 2 3.3 7 3.3 8 Zn 2+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 Ca 2+ 0.0 0 0.0 6 0.0 8 0.0 2 0.0 0 0.0 0 0.0 1 N a+ 0.0 1 0.1 0 0.0 9 0.0 4 0.1 1 0.1 0 0.1 1 K+ 1.0 8 1.4 2 1.3 5 1.4 8 1.6 5 1.6 4 1.6 3 su m 15 .35 15 .53 15 .45 15 .49 15 .61 15 .56 15 .60 271 Re pr es en tat iv e m icr op ro be an aly se s o f b io tit e SiO 2 37 .69 37 .54 37 .77 37 .35 37 .41 37 .07 37 .30 37 .39 Ti O 2 1.4 4 1.2 8 1.5 4 1.1 2 1.1 4 0.8 8 0.9 8 1.0 6 Al 2O 3 19 .72 20 .08 20 .33 19 .77 20 .28 21 .00 20 .79 20 .59 Fe O 11 .77 11 .42 10 .81 11 .90 12 .28 10 .30 9.9 6 10 .45 M nO 0.0 1 0.0 5 0.0 2 0.0 4 0.0 6 0.0 8 0.0 4 0.0 5 Cr 2O 3 0.0 0 0.0 5 0.0 1 0.0 4 0.0 2 0.0 0 0.0 0 0.0 0 M gO 16 .52 16 .90 17 .23 16 .51 16 .39 17 .72 17 .53 17 .07 Zn O 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 Ca O 0.0 0 0.0 0 0.0 1 0.0 0 0.0 3 0.0 0 0.0 0 0.0 0 N a 2O 0.8 0 0.8 1 0.5 1 0.5 1 0.5 9 0.5 5 0.4 5 0.4 2 K 2 O 8.2 4 8.1 2 7.9 0 8.4 3 8.5 0 8.6 8 8.9 3 8.8 7 Cl 0.0 3 0.0 5 0.0 3 0.0 5 0.0 5 n. a. n. a. n. a. F 0.1 0 0.2 0 0.1 0 0.1 0 0.0 0 n. a. n. a. n. a. to tal 96 .26 96 .41 96 .20 95 .77 96 .73 96 .29 95 .97 95 .89 sa m pl e n o. 06 ET -2 D 06 ET -2 D 06 ET -2 D 06 ET -2 G 06 ET -2 G 06 ET -3 D 06 ET -3 H 06 ET -3 K No rm ali ze d t o 2 2 O xy s Si4 + 5.4 2 5.3 9 5.3 9 5.4 1 5.3 7 5.3 0 5.3 5 5.3 7 Ti 4+ 0.1 6 0.1 4 0.1 7 0.1 2 0.1 2 0.1 0 0.1 1 0.1 1 Al 3+ 3.3 4 3.4 0 3.4 2 3.3 8 3.4 3 3.5 4 3.5 1 3.4 9 Fe 2+ 1.4 2 1.3 7 1.2 9 1.4 4 1.4 7 1.2 3 1.1 9 1.2 6 M n2 + 0.0 0 0.0 1 0.0 0 0.0 0 0.0 1 0.0 1 0.0 0 0.0 1 Cr 3+ 0.0 0 0.0 1 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 M g2 + 3.5 4 3.6 1 3.6 7 3.5 7 3.5 1 3.7 8 3.7 5 3.6 6 Zn 2+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 Ca 2+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 N a+ 0.2 2 0.2 2 0.1 4 0.1 4 0.1 6 0.1 5 0.1 2 0.1 2 K+ 1.5 1 1.4 9 1.4 4 1.5 6 1.5 6 1.5 8 1.6 3 1.6 3 su m 15 .62 15 .63 15 .52 15 .63 15 .65 15 .70 15 .67 15 .64 272 Re pr es en tat iv e m icr op ro be an aly se s o f b io tit e SiO 2 38 .29 37 .07 37 .30 37 .54 37 .94 37 .59 Ti O 2 1.0 4 0.8 8 0.9 8 0.9 8 1.0 4 1.0 6 Al 2O 3 19 .24 21 .00 20 .79 20 .11 20 .37 19 .90 Fe O 9.8 9 10 .30 9.9 6 10 .11 10 .03 10 .02 M nO 0.0 3 0.0 8 0.0 4 0.0 5 0.1 2 0.0 4 Cr 2O 3 17 .92 17 .72 17 .53 18 .03 17 .80 17 .97 M gO 0.0 1 0.0 0 0.0 0 0.0 3 0.0 2 0.0 1 Zn O 0.3 9 0.5 5 0.4 5 0.4 7 0.4 4 0.4 9 Ca O 8.7 9 8.6 8 8.9 3 8.7 7 8.6 4 8.4 8 N a 2O 0.0 0 0.0 2 0.0 3 0.0 0 0.0 3 0.0 0 K 2 O 0.0 0 0.2 4 0.0 7 0.2 0 0.0 8 0.1 5 to tal 95 .61 96 .44 96 .03 96 .20 96 .47 95 .64 sa m pl e n o. SK 45 9- 6 SK 45 9- 7 SK 45 9- 8 RS 45 -1 2 RS 45 -1 3 RS 45 -1 4 No rm ali ze d t o 2 2 O xy s Si4 + 5.6 3 5.4 1 5.4 6 5.4 9 5.5 4 5.5 3 Ti 4+ 2.3 7 2.5 9 2.5 4 2.5 1 2.4 6 2.4 7 Al 3+ 0.9 7 1.0 2 1.0 5 0.9 5 1.0 4 0.9 7 Fe 2+ 0.1 1 0.1 0 0.1 1 0.1 1 0.1 1 0.1 2 M n2 + 0.9 3 1.1 5 1.0 7 1.1 2 0.9 3 0.9 9 Cr 3+ 0.2 9 0.1 0 0.1 5 0.1 2 0.2 9 0.2 4 M g2 + 0.0 0 0.0 1 0.0 0 0.0 1 0.0 1 0.0 1 Zn 2+ 3.9 3 3.8 5 3.8 3 3.9 3 3.8 7 3.9 4 Ca 2+ 0.0 0 0.0 0 0.0 0 0.0 1 0.0 0 0.0 0 N a+ 0.1 1 0.1 5 0.1 3 0.1 3 0.1 3 0.1 4 K+ 1.6 5 1.6 2 1.6 7 1.6 3 1.6 1 1.5 9 su m 16 .00 16 .00 16 .00 16 .00 16 .00 16 .00 273 Re pr es en tat iv e m icr op ro be an aly se s o f c or di er ite SiO 2 47 .65 48 .10 48 .07 48 .55 49 .56 49 .42 48 .89 48 .24 Ti O 2 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 3 0.0 0 Al 2O 3 33 .04 33 .09 33 .22 34 .24 34 .55 33 .98 33 .13 32 .77 Fe O 4.8 5 4.6 3 4.7 0 3.6 8 3.7 0 3.6 8 4.2 7 4.4 2 M nO 0.0 4 0.0 5 0.0 7 0.0 5 0.0 8 0.0 6 0.0 3 0.0 0 Cr 2O 3 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 M gO 10 .44 10 .77 10 .58 11 .74 11 .90 11 .83 11 .40 11 .22 Zn O 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 Ca O 0.0 3 0.0 3 0.0 5 0.0 6 0.0 1 0.0 1 0.0 1 0.0 2 N a 2O 0.2 7 0.1 6 0.2 1 0.1 6 0.1 3 0.1 7 0.1 7 0.1 9 K 2 O 0.0 6 0.0 0 0.0 1 0.0 0 0.0 0 0.0 0 0.0 1 0.0 2 to tal 96 .37 96 .82 96 .91 98 .48 99 .92 99 .16 97 .93 96 .88 sa m pl e n o. RS -5 6 RS -5 6 RS -5 6 02 -3 .03 D 02 -3 .03 D 02 -3 .03 D 00 -5 .14 00 -5 .14 No rm ali ze d t o 1 8 O xy s Si4 + 4.9 1 4.9 3 4.9 2 4.8 6 4.8 9 4.9 2 4.9 4 4.9 3 Ti 4+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 Al 3+ 4.0 1 3.9 9 4.0 1 4.0 4 4.0 2 3.9 8 3.9 4 3.9 4 Fe 2+ 0.4 2 0.4 0 0.4 0 0.3 1 0.3 1 0.3 1 0.3 6 0.3 8 M n2 + 0.0 0 0.0 0 0.0 1 0.0 0 0.0 1 0.0 0 0.0 0 0.0 0 Cr 3+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 M g2 + 1.6 0 1.6 4 1.6 1 1.7 5 1.7 5 1.7 5 1.7 2 1.7 1 Zn 2+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 Ca 2+ 0.0 0 0.0 0 0.0 1 0.0 1 0.0 0 0.0 0 0.0 0 0.0 0 N a+ 0.0 5 0.0 3 0.0 4 0.0 3 0.0 2 0.0 3 0.0 3 0.0 4 K+ 0.0 1 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 su m 11 .00 11 .00 11 .00 11 .00 11 .00 11 .00 11 .00 11 .00 274 Re pr es en tat iv e m icr op ro be an aly se s o f c or di er ite SiO 2 48 .65 48 .72 48 .00 48 .16 48 .72 49 .36 49 .14 49 .08 Ti O 2 0.0 1 0.0 2 0.0 0 0.0 4 0.0 5 -0 .02 -0 .02 -0 .02 Al 2O 3 33 .78 33 .93 33 .79 33 .39 32 .25 33 .76 33 .77 33 .48 Fe O 4.4 2 4.6 0 5.5 2 5.8 6 5.4 4 4.0 7 3.7 4 3.9 1 M nO 0.0 5 0.1 0 0.0 5 0.0 3 0.0 5 0.1 0 0.0 4 0.0 8 Cr 2O 3 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 M gO 10 .29 10 .16 9.6 6 9.6 6 9.2 5 10 .22 10 .05 10 .00 Zn O 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 Ca O 0.0 4 0.0 5 0.0 0 0.0 3 0.3 4 0.0 0 0.0 0 0.0 1 N a 2O 0.2 4 0.3 3 0.2 1 0.1 7 0.1 7 0.0 3 0.0 5 0.0 5 K 2 O 0.0 3 0.0 2 0.0 0 0.0 2 0.3 3 0.0 0 -0 .01 0.0 1 to tal 97 .50 97 .93 97 .24 97 .36 96 .60 97 .53 96 .76 96 .59 sa m pl e n o. 06 ET -3 I 06 ET -3 I 06 ET -3 I 06 ET -3 E 06 ET -3 E 06 ET -3 B 06 ET -3 B 06 ET -3 B No rm ali ze d t o 1 8 O xy s Si4 + 4.9 6 4.9 6 4.9 2 4.9 4 5.0 4 5.0 3 5.0 4 5.0 5 Ti 4+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 Al 3+ 4.0 6 4.0 7 4.0 8 4.0 4 3.9 3 4.0 5 4.0 9 4.0 6 Fe 2+ 0.3 8 0.3 9 0.4 7 0.5 0 0.4 7 0.3 5 0.3 2 0.3 4 M n2 + 0.0 0 0.0 1 0.0 0 0.0 0 0.0 0 0.0 1 0.0 0 0.0 1 Cr 3+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 M g2 + 1.5 7 1.5 4 1.4 8 1.4 8 1.4 3 1.5 5 1.5 4 1.5 3 Zn 2+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 Ca 2+ 0.0 0 0.0 1 0.0 0 0.0 0 0.0 4 0.0 0 0.0 0 0.0 0 N a+ 0.0 5 0.0 6 0.0 4 0.0 3 0.0 3 0.0 1 0.0 1 0.0 1 K+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 4 0.0 0 0.0 0 0.0 0 su m 11 .03 11 .04 11 .00 11 .00 11 .00 11 .00 11 .00 11 .00 275 Re pr es en tat iv e m icr op ro be an aly se s o f c or di er ite SiO 2 41 .17 42 .03 49 .54 49 .64 49 .55 46 .91 49 .28 Ti O 2 0.0 1 0.0 0 0.0 0 0.0 0 0.0 0 0.0 5 0.0 1 Al 2O 3 48 .30 30 .76 34 .44 34 .49 34 .61 35 .64 34 .49 Fe O 2.1 3 6.2 4 3.2 4 3.0 1 5.0 2 5.7 6 4.8 8 M nO 0.0 4 0.0 4 0.0 5 0.0 4 0.1 2 0.1 2 0.1 0 Cr 2O 3 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 M gO 6.5 0 13 .75 11 .60 11 .38 10 .55 10 .41 10 .71 Zn O 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 Ca O 0.0 8 0.0 0 0.0 2 0.0 1 0.0 5 0.1 5 0.0 4 N a 2O 0.1 6 0.1 1 0.2 3 0.2 5 0.0 4 0.0 3 0.0 3 K 2 O 0.1 3 0.0 1 0.0 0 0.0 2 0.0 0 0.0 0 0.0 0 to tal 98 .51 92 .94 99 .13 98 .85 99 .94 99 .07 99 .54 sa m pl e n o. 01 -2 .07 01 -2 .08 RS -4 5 RS -4 5 06 ET -2 06 ET -2 06 ET -2 No rm ali ze d t o 1 8 O xy s Si4 + 4.1 1 4.4 3 4.9 2 4.9 5 4.9 4 4.7 6 4.9 3 Ti 4+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 5 0.0 8 0.3 6 Al 3+ 5.6 8 3.8 2 4.0 4 4.0 5 4.0 5 4.0 5 4.0 5 Fe 2+ 0.1 8 0.5 5 0.2 7 0.2 5 0.4 2 0.4 9 0.4 1 M n2 + 0.0 0 0.0 0 0.0 0 0.0 0 0.0 1 0.0 1 0.0 1 Cr 3+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 M g2 + 0.9 7 2.1 6 1.7 2 1.6 9 1.5 7 1.5 7 1.6 0 Zn 2+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 Ca 2+ 0.0 1 0.0 0 0.0 0 0.0 0 0.0 1 0.0 2 0.0 0 N a+ 0.0 3 0.0 2 0.0 4 0.0 5 0.0 1 0.0 1 0.0 1 K+ 0.0 2 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 su m 11 .00 11 .00 11 .00 11 .00 11 .02 11 .11 11 .03 276 Re pr es en tat iv e m icr op ro be an aly se s o f g ar ne t SiO 2 38 .04 37 .93 37 .87 38 .73 38 .48 38 .56 38 .39 38 .76 Ti O 2 0.0 0 0.0 5 0.0 8 0.0 0 0.0 0 0.0 1 -0 .02 0.0 3 Al 2O 3 22 .21 22 .56 22 .20 22 .44 22 .30 22 .66 22 .21 22 .23 Fe O 30 .52 29 .59 29 .65 27 .68 27 .08 27 .16 26 .89 26 .12 M nO 0.4 9 0.5 1 0.5 2 0.4 0 0.2 6 0.2 6 0.6 6 0.6 2 Cr 2O 3 0.0 0 0.0 3 M gO 8.1 4 8.8 8 8.7 9 7.3 6 7.4 4 7.5 5 8.2 6 8.5 7 Zn O 0.0 0 0.0 0 Ca O 0.7 1 0.7 1 0.7 5 4.4 0 4.5 0 4.2 8 3.9 0 3.7 3 to tal 10 0.1 1 10 0.2 1 99 .87 10 1.0 1 10 0.0 7 10 0.4 8 10 0.3 1 10 0.0 9 sa m pl e n o. 00 -5 .13 00 -5 .13 00 -5 .13 02 -3 .03 B 02 -3 .03 B 02 -3 .03 B 01 -2 .07 01 -2 .07 No rm ali ze d t o 1 2 O xy s Si4 + 2.9 6 2.9 4 2.9 5 2.9 7 2.9 8 2.9 7 2.9 6 2.9 8 Ti 4+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 Al 3+ 2.0 4 2.0 6 2.0 4 2.0 3 2.0 3 2.0 5 2.0 2 2.0 2 Fe 2+ 1.9 9 1.9 2 1.9 3 1.7 8 1.7 5 1.7 5 1.7 3 1.6 8 M n2 + 0.0 3 0.0 3 0.0 3 0.0 3 0.0 2 0.0 2 0.0 4 0.0 4 Cr 3+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 M g2 + 0.9 4 1.0 2 1.0 2 0.8 4 0.8 6 0.8 7 0.9 5 0.9 8 Zn 2+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 Ca 2+ 0.0 6 0.0 6 0.0 6 0.3 6 0.3 7 0.3 5 0.3 2 0.3 1 su m 8.0 2 8.0 3 8.0 3 8.0 1 8.0 1 8.0 0 8.0 3 8.0 1 277 Re pr es en tat iv e m icr op ro be an aly se s o f g ar ne t SiO 2 39 .51 39 .35 39 .08 39 .34 39 .68 38 .92 38 .80 38 .98 Ti O 2 0.0 1 0.0 0 0.0 0 -0 .06 0.0 2 0.0 0 -0 .02 -0 .03 Al 2O 3 22 .84 22 .80 22 .95 23 .09 22 .99 22 .49 21 .86 22 .52 Fe O 24 .12 24 .19 23 .81 24 .08 24 .19 28 .31 28 .47 29 .12 M nO 1.4 0 1.4 7 1.3 5 1.3 3 1.2 4 0.7 5 0.7 1 0.8 1 Cr 2O 3 -0 .01 0.0 4 -0 .04 -0 .02 0.0 2 M gO 8.5 6 8.8 0 8.9 2 9.5 4 9.4 5 7.3 3 7.6 5 7.3 5 Zn O 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 Ca O 4.4 8 4.5 0 4.1 1 3.9 6 3.7 2 3.1 4 3.1 2 3.1 5 to tal 10 0.9 1 10 1.1 6 10 0.1 9 10 1.2 6 10 1.3 2 10 0.9 3 10 0.5 9 10 1.8 9 sa m pl e n o. 06 ET -2 D 06 ET -2 D 06 ET -2 D 06 ET -2 D 06 ET -2 D 06 ET -2 G 06 ET -2 G 06 ET -2 G No rm ali ze d t o 1 2 O xy s Si4 + 2.9 9 2.9 8 2.9 8 2.9 7 2.9 9 2.9 9 3.0 0 2.9 8 Ti 4+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 Al 3+ 2.0 4 2.0 3 2.0 6 2.0 5 2.0 4 2.0 4 1.9 9 2.0 3 Fe 2+ 1.5 3 1.5 3 1.5 2 1.5 2 1.5 2 1.8 2 1.8 4 1.8 6 M n2 + 0.0 9 0.0 9 0.0 9 0.0 8 0.0 8 0.0 5 0.0 5 0.0 5 Cr 3+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 M g2 + 0.9 7 0.9 9 1.0 1 1.0 7 1.0 6 0.8 4 0.8 8 0.8 4 Zn 2+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 Ca 2+ 0.3 6 0.3 7 0.3 4 0.3 2 0.3 0 0.2 6 0.2 6 0.2 6 su m 7.9 8 8.0 0 7.9 9 8.0 1 7.9 9 7.9 9 8.0 1 8.0 1 278 Re pr es en tat iv e m icr op ro be an aly se s o f O rth oa m ph ib ol e SiO 2 48 .08 47 .74 47 .60 46 .75 52 .53 45 .77 46 .33 47 .36 Ti O 2 0.1 3 0.1 5 0.1 7 0.1 9 0.2 1 0.3 5 0.3 4 0.1 4 Al 2O 3 11 .47 13 .35 12 .69 13 .70 4.9 3 16 .32 15 .66 15 .43 Fe O 14 .60 14 .83 15 .45 15 .34 18 .25 15 .69 15 .57 14 .11 M nO 0.0 9 0.1 4 0.1 5 0.1 7 0.1 4 0.1 6 0.0 9 0.1 4 Cr 2O 3 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 M gO 20 .83 20 .21 19 .97 19 .65 20 .26 18 .51 18 .57 19 .73 Zn O 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 Ca O 0.5 8 0.5 0 0.6 2 0.5 9 0.3 4 0.4 4 0.4 2 0.2 2 N a 2O 0.8 9 1.0 3 1.0 4 1.2 4 0.4 6 1.5 5 1.5 6 1.4 2 K 2 O 0.0 1 0.0 2 0.0 1 0.0 1 0.0 0 0.0 1 0.0 0 0.0 0 to tal 96 .68 97 .97 97 .70 97 .63 97 .11 98 .80 98 .54 98 .55 sa m pl e n o. 02 -3 .03 02 -3 .03 02 -3 .03 02 -3 .03 00 -5 .24 00 -5 .24 00 -5 .24 RS -5 6A No rm ali ze d t o 2 3 O xy s Si4 + 6.7 2 6.5 8 6.6 0 6.4 7 7.4 8 6.2 6 6.3 6 6.4 6 Ti 4+ 0.0 1 0.0 2 0.0 2 0.0 2 0.0 2 0.0 4 0.0 4 0.0 1 Al 3+ 1.8 9 2.1 7 2.0 7 2.2 4 0.8 3 2.6 3 2.5 3 2.4 8 Fe 2+ 1.7 1 1.7 1 1.7 9 1.7 8 2.1 7 1.8 0 1.7 9 1.6 1 M n2 + 0.0 1 0.0 2 0.0 2 0.0 2 0.0 2 0.0 2 0.0 1 0.0 2 Cr 3+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 M g2 + 4.3 4 4.1 5 4.1 3 4.0 6 4.3 0 3.7 8 3.8 0 4.0 1 Zn 2+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 Ca 2+ 0.0 9 0.0 7 0.0 9 0.0 9 0.0 5 0.0 7 0.0 6 0.0 3 N a+ 0.2 4 0.2 8 0.2 8 0.3 3 0.1 3 0.4 1 0.4 1 0.3 8 K+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 su m 15 .00 15 .00 15 .00 15 .00 15 .00 15 .00 15 .00 15 .00 279 Re pr es en tat iv e m icr op ro be an aly se s o f O rth oa m ph ib ol e SiO 2 51 .66 45 .81 47 .01 50 .35 47 .11 52 .34 44 .57 46 .78 Ti O 2 0.1 8 0.1 0 0.1 4 0.1 7 0.3 6 0.2 5 0.4 6 0.1 5 Al 2O 3 6.8 4 16 .56 15 .36 8.9 0 12 .69 11 .22 14 .29 16 .93 Fe O 16 .62 14 .46 14 .31 18 .51 19 .46 17 .78 20 .89 12 .96 M nO 0.3 1 0.0 8 0.0 9 0.2 4 0.2 3 0.2 3 0.2 9 0.2 1 Cr 2O 3 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 M gO 21 .06 19 .23 19 .64 18 .84 17 .15 14 .75 15 .09 20 .05 Zn O 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 Ca O 0.4 8 0.2 6 0.2 0 0.4 7 0.5 6 0.4 8 0.6 4 0.2 5 N a 2O 0.5 4 1.3 9 1.2 3 0.7 0 1.1 3 0.7 7 1.3 8 1.1 4 K 2 O 0.0 0 0.0 0 0.0 0 0.0 1 0.0 3 0.0 2 0.0 0 0.0 0 to tal 97 .69 97 .89 97 .97 98 .20 98 .71 97 .86 97 .60 98 .47 sa m pl e n o. RS -5 6A RS -5 6A RS -5 6A 01 -2 .07 01 -2 .07 01 -2 .07 01 -2 .07 01 -2 .07 No rm ali ze d t o 2 3 O xy s Si4 + 7.2 4 6.2 9 6.4 6 7.0 8 6.5 9 7.4 9 6.3 5 6.3 7 Ti 4+ 0.0 2 0.0 1 0.0 1 0.0 2 0.0 4 0.0 3 0.0 5 0.0 2 Al 3+ 1.1 3 2.6 8 2.4 9 1.4 8 2.0 9 1.8 9 2.4 0 2.7 1 Fe 2+ 1.9 5 1.6 6 1.6 4 2.1 8 2.2 8 2.1 3 2.4 9 1.4 7 M n2 + 0.0 4 0.0 1 0.0 1 0.0 3 0.0 3 0.0 3 0.0 3 0.0 2 Cr 3+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 M g2 + 4.4 0 3.9 4 4.0 2 3.9 5 3.5 8 3.1 5 3.2 0 4.0 7 Zn 2+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 Ca 2+ 0.0 7 0.0 4 0.0 3 0.0 7 0.0 8 0.0 7 0.1 0 0.0 4 N a+ 0.1 5 0.3 7 0.3 3 0.1 9 0.3 1 0.2 1 0.3 8 0.3 0 K+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 su m 15 .00 15 .00 15 .00 15 .00 15 .00 15 .00 15 .00 15 .00 280 Re pr es en tat iv e m icr op ro be an aly se s o f O rth oa m ph ib ol e SiO 2 46 .30 46 .53 52 .72 45 .67 48 .11 Ti O 2 0.1 9 0.2 8 0.2 5 0.1 9 0.2 4 Al 2O 3 15 .42 15 .45 6.4 0 16 .75 13 .81 Fe O 13 .94 14 .31 15 .58 14 .56 14 .09 M nO 0.1 2 0.0 6 0.1 0 0.1 2 0.0 9 Cr 2O 3 0.0 0 0.0 3 -0 .01 0.0 0 0.0 6 M gO 19 .87 19 .65 22 .27 19 .21 20 .48 Zn O 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 Ca O 0.4 8 0.4 5 0.4 6 0.5 9 0.4 6 N a 2O 1.2 6 1.4 4 0.6 0 1.6 5 1.0 4 K 2 O 0.0 1 0.0 1 -0 .01 -0 .01 0.0 2 to tal 97 .59 98 .21 98 .35 98 .72 98 .39 sa m pl e n o. 06 ET -2 D 06 ET -2 D 06 ET -2 G 06 ET -2 G 06 ET -2 G No rm ali ze d t o 2 3 O xy s Si4 + 6.5 2 6.5 2 7.3 7 6.3 9 6.7 1 Ti 4+ 0.0 2 0.0 3 0.0 3 0.0 2 0.0 3 Al 3+ 2.5 6 2.5 5 1.0 5 2.7 6 2.2 7 Fe 2+ 1.6 4 1.6 8 1.8 2 1.7 0 1.6 4 M n2 + 0.0 1 0.0 1 0.0 1 0.0 1 0.0 1 Cr 3+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 1 M g2 + 4.1 7 4.1 1 4.6 4 4.0 1 4.2 6 Zn 2+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 Ca 2+ 0.0 7 0.0 7 0.0 7 0.0 9 0.0 7 N a+ 0.3 4 0.3 9 0.1 6 0.4 5 0.2 8 K+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 su m 15 .35 15 .36 15 .16 15 .43 15 .27 281 Re pr es en tat iv e m icr op ro be an aly se s o f p lag io cla se SiO 2 44 .07 43 .55 44 .01 43 .96 47 .60 46 .75 46 .26 46 .78 Al 2O 3 36 .55 36 .64 36 .76 36 .87 33 .75 34 .27 34 .85 35 .04 Fe O 0.3 9 0.6 0 0.3 2 0.3 0 0.1 6 0.1 9 0.2 7 0.2 4 Ca O 19 .12 18 .50 19 .28 19 .46 17 .25 17 .34 17 .91 18 .21 N a 2O 0.4 9 0.5 2 0.5 4 0.4 2 1.8 7 1.6 9 1.3 2 1.4 0 K 2 O 0.0 0 0.0 4 0.0 1 0.0 0 0.0 3 -0 .01 0.0 3 0.0 3 to tal 10 0.6 1 99 .84 10 0.9 1 10 1.0 1 10 0.6 6 10 0.2 2 10 0.6 5 10 1.7 1 sa m pl e n o. 02 30 3- 5 02 30 3- 6 02 30 3- 7 02 30 3- 8 06 ET -2 G 06 ET -2 G 06 ET -2 G 06 ET -2 G No rm ali ze d t o 8 O xy s Si4 + 2.0 2 2.0 1 2.0 1 2.0 1 2.1 7 2.1 4 2.1 2 2.1 2 Al 3+ 1.9 8 2.0 0 1.9 8 1.9 9 1.8 1 1.8 5 1.8 8 1.8 7 Fe 2+ 0.0 1 0.0 2 0.0 1 0.0 1 0.0 1 0.0 1 0.0 1 0.0 1 Ca 2+ 0.9 4 0.9 2 0.9 4 0.9 5 0.8 4 0.8 5 0.8 8 0.8 8 N a+ 0.0 4 0.0 5 0.0 5 0.0 4 0.1 7 0.1 5 0.1 2 0.1 2 K+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 su m 5.0 0 5.0 0 5.0 0 5.0 0 5.0 0 5.0 0 5.0 0 5.0 0 282 Re pr es en tat iv e m icr op ro be an aly se s o f p lag io cla se SiO 2 56 .06 56 .27 56 .76 56 .08 59 .00 59 .42 66 .73 66 .21 Al 2O 3 27 .83 27 .72 27 .99 27 .76 26 .09 26 .28 21 .21 21 .13 Fe O 0.1 2 0.1 8 0.2 2 0.3 2 0.0 7 0.0 6 -0 .03 0.0 6 Ca O 9.9 2 9.7 5 9.9 9 10 .26 6.9 5 7.1 5 1.6 4 1.7 7 N a 2O 6.0 8 5.8 4 5.7 4 5.9 9 7.4 5 6.5 4 10 .16 9.9 6 K 2 O 0.0 5 0.0 5 0.0 3 0.0 2 0.1 2 0.1 2 0.5 5 0.4 5 to tal 10 0.0 6 99 .79 10 0.7 3 10 0.4 4 99 .68 99 .57 10 0.2 6 99 .57 sa m pl e n o. 01 -2 .07 01 -2 .07 01 -2 .07 01 -2 .07 06 ET -3 I 06 ET -3 I 06 ET -3 E 06 ET -3 E No rm ali ze d t o 8 O xy s Si4 + 2.5 2 2.5 4 2.5 4 2.5 1 2.6 4 2.6 5 2.9 2 2.9 2 Al 3+ 1.4 7 1.4 7 1.4 8 1.4 6 1.3 8 1.3 8 1.0 9 1.1 0 Fe 2+ 0.0 0 0.0 1 0.0 1 0.0 1 0.0 0 0.0 0 0.0 0 0.0 0 Ca 2+ 0.4 8 0.4 7 0.4 8 0.4 9 0.3 3 0.3 4 0.0 8 0.0 8 N a+ 0.5 3 0.5 1 0.5 0 0.5 2 0.6 5 0.5 7 0.8 6 0.8 5 K+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 1 0.0 1 0.0 3 0.0 3 su m 5.0 0 5.0 0 5.0 0 5.0 0 5.0 0 4.9 5 4.9 8 4.9 7 283 Re pr es en tat iv e m icr op ro be an aly se s o f s ap ph iri ne SiO 2 11 .04 10 .50 10 .34 10 .75 10 .38 10 .31 10 .27 Ti O 2 0.0 0 0.0 0 0.0 0 0.0 4 0.0 7 0.0 2 0.0 5 Al 2O 3 66 .03 67 .31 67 .63 66 .72 67 .49 67 .17 66 .98 Fe O 6.6 3 6.6 1 6.9 9 7.0 2 6.8 3 7.0 3 6.7 2 M nO 0.0 7 0.0 6 0.0 0 0.0 2 0.0 4 0.0 4 0.0 3 Cr 2O 3 0.0 3 0.0 0 0.0 8 0.0 9 0.0 0 0.0 6 0.0 2 M gO 13 .83 14 .25 14 .51 14 .44 14 .38 14 .47 14 .24 Zn O 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 Ca O 0.7 8 0.3 5 0.0 4 0.0 0 0.0 4 0.0 1 0.1 6 to tal 98 .41 99 .07 99 .58 99 .08 99 .22 99 .10 98 .48 sa m pl e n o. 02 -3 .03 02 -3 .03 02 -3 .03 02 -3 .03 02 -3 .03 02 -3 .03 06 ET -2 No rm ali ze d t o 2 0 O xy s Si4 + 1.3 3 1.2 6 1.2 3 1.2 9 1.2 4 1.2 3 1.2 4 Ti 4+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 1 0.0 0 0.0 0 Al 3+ 9.3 8 9.4 9 9.5 0 9.4 2 9.5 0 9.4 8 9.5 0 Fe 2+ 0.6 7 0.6 6 0.7 0 0.7 0 0.6 8 0.7 0 0.6 8 M n2 + 0.0 1 0.0 1 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 Cr 3+ 0.0 0 0.0 0 0.0 1 0.0 1 0.0 0 0.0 1 0.0 0 M g2 + 2.4 9 2.5 4 2.5 8 2.5 8 2.5 6 2.5 8 2.5 6 Zn 2+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 Ca 2+ 0.1 0 0.0 4 0.0 1 0.0 0 0.0 0 0.0 0 0.0 2 su m 13 .98 14 .00 14 .02 14 .00 14 .00 14 .02 14 .01 284 Re pr es en tat iv e m icr op ro be an aly se s o f s pi ne l SiO 2 0.0 1 0.0 3 0.5 7 0.2 0 0.2 0 1.8 8 0.0 2 Ti O 2 -0 .03 0.0 4 0.9 5 0.0 5 0.0 5 0.0 0 -0 .02 Al 2O 3 60 .57 62 .36 62 .88 65 .08 65 .08 62 .60 60 .73 Fe O 33 .21 30 .91 22 .34 22 .79 22 .79 21 .89 32 .00 M nO 0.3 0 0.3 1 0.0 5 0.0 6 0.0 6 0.0 1 0.3 4 Cr 2O 3 0.1 0 0.0 0 0.0 7 0.0 5 0.0 5 0.0 2 0.0 8 M gO 6.2 5 6.8 1 11 .77 12 .36 12 .36 12 .63 7.2 2 Zn O 0.0 0 0.0 0 0.0 0 0.0 0 0.3 2 0.5 4 0.2 0 to tal 10 0.4 1 10 0.4 6 98 .64 10 0.5 9 10 0.9 1 99 .59 10 0.5 9 sa m pl e n o. 02 -3 .03 02 -3 .03 06 ET -2 D 06 ET -2 D 06 ET -2 G 06 ET -2 G 06 ET -2 H No rm ali ze d t o 4 O xy s Si4 + 0.0 0 0.0 0 0.0 2 0.0 1 0.0 1 0.0 5 0.0 0 Ti 4+ 1.9 8 2.0 1 1.9 8 2.0 0 2.0 0 1.9 4 1.9 7 Al 3+ 0.0 0 0.0 0 0.0 2 0.0 0 0.0 0 0.0 0 0.0 0 Fe 2+ 0.7 7 0.7 1 0.5 0 0.5 0 0.5 0 0.4 8 0.7 4 M n2 + 0.0 1 0.0 1 0.0 0 0.0 0 0.0 0 0.0 0 0.0 1 Cr 3+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 M g2 + 0.2 6 0.2 8 0.4 7 0.4 8 0.4 8 0.5 0 0.3 0 Zn 2+ 0.0 0 0.0 0 0.0 0 0.0 0 0.0 1 0.0 1 0.0 0 su m 3.0 1 3.0 0 2.9 8 2.9 9 2.9 9 2.9 8 3.0 1 285