GARNET AND STAUROLITE BEHAVIOUR IN THE NELSON AUREOLE, BC: IMPLICATIONS FOR PHASE EQUILIBRIUM MODELLING AND REACTION KINETICS

This study examines the textures and compositional zoning of garnet and staurolite in metapelites going upgrade through the the Nelson aureole, BC, and compares these features with predictions from phase equilibrium modelling. The mineral assemblage sequence is satisfactorily modelled by a 3.8 kbar isobaric trajectory through a P-T phase diagram section, calculated for average Nelson metapelite, in either of the chemical systems MnKFMASH or MnNCKFMASHT, using the Holland & Powell (2003) data base. Garnet appears to grow in accordance with phase equilibrium predictions, but does not dissolve as much as predicted, showing negligible consumption during staurolite growth and upgrade of terminal staurolite consumption, where it is predicted to be completely consumed. Phase equilibrium modelling assuming purely fractional behaviour with no dissolution (i.e., garnet is allowed to grow but not dissolve) only partially solves the problem, because in this scenario garnet is not predicted to grow upgrade of the garnet zone, which contradicts evidence from Ca zoning and textures. An intermediate scenario in which garnet only partially dissolves may be closest to reality. Staurolite shows a different range of behaviour. In different rocks from the same outcrop and in individual, relatively homogeneous, thin sections, staurolite may show evidence for substantial reaction to form andalusite in some domains but appear wholly unreactive in others. Bulk compositional differences are unlikely to be the cause. Unreacted staurolite may persist several 100s metres (equivalent to many 10s °C) upgrade of where it first starts to break down in the andalusite stability field (~ the equilibrium boundary?), reacting to form sillimanite and new garnet in a domain in which phase equilibrium modelling predicts garnet to be consumed. All of these observations indicate significant kinetic inhibitions to dissolution of garnet and staurolite, with implications for geothermobarometry, mineral-out isograds and bathograd analysis. X-ray mapping of the Nelson garnets, combined with Fe-Ti oxide phase analysis, challenges another sacred cow, namely Mn enrichment in garnet rims (almost universally ascribed to garnet resorption, but in some Nelson rocks instead a prograde feature linked to ilmenite breakdown in sulphide-rich samples).