Meeting room "Glova", 4 floor, Integrated Research Building
Abstract
Although majoritic garnet - the major Al mineral phase at ~200-700 km Earth's depth - is experimentally proven to decompose at shallower levels to garnet-pyroxene intergrowths, interpretation of natural analogues in mafic/ultramafic surface rocks is controversial. Suggested scenarios for the breakdown of the majorite component include decompression within sub-lithospheric mantle, cooling within lithospheric mantle and pressure release after tectonic ultra-high pressure (UHP) metamorphism. Here we performed decompression experiments to constrain the time and temperature conditions and by that the geodynamic environment necessary for the exsolution process to take place.
Polycrystalline supersilicic garnet (Si ~3.55 cations per 12 oxygen) was synthesised with a grain size of <=5 μm from powdered glass with a 'pyrolite minus olivine' composition at dry conditions, 18 GPa and 1600℃ during 140 min using a multi-anvil apparatus. Subsequently, the recovered synthetic garnet and the powdered glass starting material both were subjected to 10 GPa and 1450℃ for times between 0 h and 12 h to breakdown the solid solution mineral and to produce the corresponding equilibria mineral phases simultaneously. BSE images and X-ray maps from recovered disequilibria samples show corona textures of two different, micron sized mineral phases with contrasting Al content surrounding relict garnet cores. XRD spectra from polycrystalline areas and TEM diffraction patterns from single crystals identify coronas to be composed of Ca pyroxene and garnet. FE-SEM EDS and preliminary TEM EDS suggest relict garnet is highly supersilicic (Si ~3.50 cpfu) and compositionally zoned in divalent cations, coronitic Ca pyroxene is tschermakitic (XTs 0.10) and coronitic garnet is highly supersilicic too (Si ~3.45 cpfu). Equilibria samples contrast with virtually Tschermak's-free Ca pyroxene (XTs 0.02) and lower supersilicity in garnet (Si ~3.2 cpfu) both up to 5 μm in size. Transformed volumes estimated from peak ratios of XRD spectra indicate corona formation occurred to ~40 vol.% rapidly until the experimental target temperature was reached, but did not change afterwards.
Collectively, the data implies a two-stage breakdown process. Rapid phase transformation of supersilicic garnet rims to tschermakitic pyroxene and new garnet slightly lower in supersilicity forms a coronitic decomposition microstructure during decompression. Subsequent chemical equilibration of the mineral phases to ambient pressure-temperature conditions will be diffusion driven and therefore strongly depend on temperature and cooling rates rather than on decompression rates. Sluggish Si-Al interdiffusion in garnet perfectly explains the coupled occurrence of lamellae type and corona type exsolution microstructures after supercilicic garnet precursors in chemically equilibrated natural analogues. Dodson closure temperatures show that cooling rates of UHP metamorphic settings and of kimberlite magma with mantle xenolith cargo both suppress chemical equilibration of pyroxene precipitates 10 μm and larger in size from supersilicic garnet. By implication, slow cooling rates and/or long time integrated residence at upper mantle temperatures are necessary to explain majoritic garnet breakdown products occurring in mantle fragments enclosed in UHP metamorphic terrains and in kimberlites. As a consequence, mineral microstructures after majoritic garnet precursors are related to sub-lithospheric mantle decompression and sub-continental lithospheric mantle evolution.
For inquiry：Taku Tsuchiya TEL:（089）927-8198
E-mail takut@sci.ehime-u.ac.jp
