Application of Fe-Ti oxide dissolution experiments to the petrogenesis of the Ekati Diamond Mine kimberlites, Northwest Territories, Canada

Abstract:

Composition of kimberlites is ambiguous due to assimilation and fractional crystallization. We propose that the evolution history of minerals can be used to decipher the magmatic history of kimberlites. We use Fe-Ti oxides (chromite and ilmenite) from six kimberlites from the Ekati Diamond Mine and dissolution experiments to elucidate the petrogenesis of kimberlites. Experiments at 0.1 MPa and variable ƒO₂s in a diopside-anorthite melt show that the dissolution rate of ilmenite is highly sensitive to ƒO₂. No significant difference was observed in chromite. Zoning in chromite is related to the Fe-content and oxidation state of the melt. Experiments at 1 GPa explore the development of chromite surface resorption features in the system Ca-Mg-Si-H-C-O. Five kimberlites contain a low abundance of ilmenite, owing to a relatively high ƒO₂, though ilmenite constituted 65% of oxide macocrysts in one kimberlite. Chromite compositions evolve from Mg-chromite to magnesio-ulvöspinel-magnetite (MUM) in all but one kimberlite where chromite evolves to a pleonaste composition perhaps as a result of rapid emplacement. The high abundance of MUM spinel and low abundance of ilmenite in the matrix could be related to the change in the stable Ti-phase with increasing ƒO₂. Core compositions of macrocrysts vary for different mantle sources but rims converge to a composition slightly more oxidized and Mg-rich than chromite from depleted peridotite. Ilmenite commonly has rims composed of perovskite, titanite and MUM. We suggest a model where the kimberlite melt composition is controlled by the co-dissolution and co-precipitation of silicates (predominantly orthopyroxene and olivine) to explain chromite evolution in kimberlites. Resorption-related surface features on chromite macrocrysts show trigon protrusions-depressions on {111} faces and step-like features along the crystal edges resembling products of experiments in H₂O fluid. We propose predominantly H₂O magmatic fluid in Ekati kimberlites.