Misidentification of oxide phases and of twinned kyanite: implications for inferred P-T histories of the Musgrave Block, central Australia

Alfredo Camacho, and John D. Fitz Gerald

In the Musgrave Block, central Australia, Mesoproterozic granulites have been locally transformed in the Neoproterozoic to eclogite (T≈660°C and P≈12 kbar) with a clear spatial association between shear zones and eclogite formation. In these high-strain zones, peraluminous rocks contain intergrowths of kyanite and magnetite that pseudomorph granulite-facies sillimanite. A previous study misidentified the minerals that make up these intergrowths as fibrolite and ilmenite which suggested the breakdown - during decompression - of garnet in a reaction of the type: garnet + rutile = aluminosilcate + hematite + quartz. We, however, consider that the intergrowth represents transformation of sillimanite by kyanite accompanied by growth of titaniferous magnetite since Fe (sillimanite typically contains ~1% Fe) is not incorporated into the more densely packed structure of kyanite. These and additional textural relationships suggest that kyanite and magnetite are coeval with eclogite-facies deformation and did not form during decompression. Our results describe a P-T history quite different to that inferred in the previous study. This contribution shows the importance of correct mineral identification and textural interpretation and illustrates why it is essential to carefully select the best available methods and instruments when characterising phases and mineral relationships for use in P-T reconstructions.