High-pressure-high-temperature melting experiments on a SiO ₂ -poor aphanitic kimberlite from the Wesselton Mine, Kimberley, South Africa

Phase relations in a SiO₂-poor aphanitic Group I kimberlite from the Wesselton mine, South Africa, were determined at 10-40 kbar and 1000-1525 °C. Experiments were done first with no additional H₂O or CO₂, equivalent to the initial amounts in the rock of 6.20 wt% H₂O and 4.77 wt% CO₂, and second with sufficient CO₂ added to bring the total to 10.34 wt% CO₂. These amounts are equivalent to a mole fraction of CO₂ (X_CO2) of 0.24 and 0.52, respectively. Oxygen fugacities are difficult to predict but were likely less than the MW and greater than the IW buffer assemblages. All experiments are suprasolidus and apparently vapor absent. In addition to liquid, runs at X_CO2 = 0.24 produced the following assemblages with decreasing pressure from 40 to 10 kbar and temperatures from 1400 to 1000 °C: olivine, olivine + spinel, olivine + spinel + clinopyroxene ± calcite ± perov-skite, and olivine + spinel + monticellite + calcite + perovskite. At X_CO2 = 0.52, runs at 20-35 kbar produced olivine + spinel, olivine + clinopyroxene + spinel, and olivine + clinopyroxene + spinel + calcite. Above 35 kbar, the latter assemblage is replaced by olivine + clinopyroxene + dolomite. Comparison of the compositions of the minerals in the rock with those produced in the experiments indicates good agreement except for spinels that are highly oxidized in the experimental products relative to those in the rock. The run assemblages are comparable with the minerals in the aphanitic kimberlite except for clinopyroxene, which is absent in the rock, and ilmenite, apatite, and phlogopite, which are present in the rock but not in the experimental runs. Based on the X_CO2 = 0.24 experiments, the inferred P-T path of ascent of the aphanitic kimberlite magma above about 40-km depth may have exceeded the temperature (1250-1300 °C) at which clinopyroxene is stable for this composition. The antipathy of clinopyroxene and monticellite in the X_CO2 = 0.24 experiments may be used to explain the incompatibility of these minerals in kimberlite and suggests that monticellite need not be a product associated with crustal processes.

The aphanitic Wesselton kimberlite has lower SiO₂ and MgO and higher CaO than other Group I kimberlites and, when plotted on the CMS system, lies on the CaO-rich side of the olivine-clinopyroxene join whereas more SiO₂-rich Group I kimberlites fall close to this join. These compositional differences may account for the absence of orthopyroxene in any of the experimental assemblages and possibly for the fact that calcite is the only carbonate mineral below 35 kbar and occurs only at high X_CO2 conditions. Most Group I kimberlite compositions are represented by the model carbonated lherzolite system (oliv-ine-clinopyroxene-orthopyroxene-dolomite) for which a pseudo-eutectic melting relationship involving enstatite may occur. The absence of orthopyroxene in the present experiments does not preclude the more SiO₂-undersaturated Wesselton kimberlite magma from being the product of crystal fractionation at mantle depths of a more SiO₂-rich kimberlite magma derived by partial melting of a carbonated lherzolite source. However, the absence of orthopyroxene in the experiments does indicate that the aphanitic kimberlite magma could have evolved from a source devoid of orthopyroxene and with calcite as the carbonate phase. The present experiments suggest that the aphanitic Wesselton kimberlite is not an evolved species but may represent a primitive kimberlite magma.