Inclusions in diamonds constrain thermo-chemical conditions during Mesozoic metasomatism of the Kaapvaal cratonic mantle

Fluid/melt inclusions in diamonds, which were encapsulated during a metasomatic event and over a short period of time, are isolated from their surrounding mantle, offering the opportunity to constrain changes in the sub-continental lithospheric mantle (SCLM) that occurred during individual thermo-chemical events, as well as the composition of the fluids involved and their sources. We have analyzed a suite of 8 microinclusion-bearing diamonds from the Group I De Beers Pool kimberlites, South Africa, using FTIR, EPMA and LA-ICP-MS. Seven of the diamonds trapped incompatible-element-enriched saline high density fluids (HDFs), carry peridotitic mineral microinclusions, and substitutional nitrogen almost exclusively in A-centers. This low-aggregation state of nitrogen indicates a short mantle residence times and/or low mantle ambient temperature for these diamonds. A short residence time is favored because, elevated thermal conditions prevailed in the South African lithosphere during and following the Karoo flood basalt volcanism at ∼180 Ma, thus the saline metasomatism must have occurred close to the time of kimberlite eruptions at ∼85 Ma. Another diamond encapsulated incompatible-element-enriched silicic HDFs and has 25% of its nitrogen content residing in B-centers, implying formation during an earlier and different metasomatic event that likely relates to the Karoo magmatism at ca. 180 Ma. Thermometry of mineral microinclusions in the diamonds carrying saline HDFs, based on Mg–Fe exchange between garnet–orthopyroxene (Opx)/clinopyroxene (Cpx)/olivine and the Opx–Cpx thermometer, yield temperatures between 875–1080 °C at 5 GPa. These temperatures overlap with conditions recorded by touching inclusion pairs in diamonds from the De Beers Pool kimberlites, which represent the mantle ambient conditions just before eruption, and are altogether lower by 150–250 °C compared to P–T gradients recorded by peridotite xenoliths from the same locality. Oxygen fugacity (fO₂) differs as well. The fO₂ calculated for the saline HDF compositions (Δlog⁡fO_2(FMQ)=−2.47 to −1.34) are higher by about a log unit compared with that recorded by xenoliths at 4–7 GPa. We conclude that enriched saline HDFs mediated the metasomatism that preceded Group I kimberlite eruptions in the southwestern Kaapvaal craton, and that their ‘cold and oxidized’ nature reflects their derivation from a deep subducting slab. This event had little impact on the temperature and redox state of the Kaapvaal lithosphere as a reservoir, however, it likely affected its properties along limited metasomatized veins and their wall rock. To reconcile the temperature and oxygen fugacity discrepancy between inclusions in diamonds and xenoliths, we argue that xenoliths did not equilibrate during the last saline metasomatic event or kimberlite eruption. Thus the P–T–fO₂ gradients they record express pre-existing lithospheric conditions that were likely established during the last major thermal event in the Kaapvaal craton (i.e. the Karoo magmatism at ca. 180 Ma).