Multistage diamond formation, mantle uplift and changing geothermal regimes recorded by inclusions in Kimberley diamonds

Abstract Thermobarometry of composite peridotitic mineral inclusions in De Beers Pool diamonds (Kimberley, South Africa) has yielded puzzling results. Most non-touching inclusions record higher temperatures than touching inclusions, but both types record conditions colder than the Kimberley xenolith geotherm. Scenarios previously proposed to explain this discrepancy (lithosphere cooling after diamond formation, cooling of discrete diamond-growth pathways by slab-derived fluids, and diamond formation under various thermal regimes) fail to fully account for the observed thermobarometric outcomes. We propose an alternative scenario based on elastic theory of inclusion–host systems, which reconciles the contrasting pressure–temperature ( P–T ) estimates. Forward model calculations show that P–T conditions similar to those estimated for the touching inclusions can result from the development of overpressures on the inclusions. Our model requires initial diamond formation under conditions colder than a 35-mW/m 2 geotherm, followed by mantle uplift (~ 60 km, possibly multi-stage) and reequilibration on the Late Cretaceous xenolith geotherm (~ 40 mW/m 2 ). The initial cold conditions could be promoted by foundering of shallow lithospheric materials. Consequent development of exsolution textures could favor entrapment of composite orthopyroxene–garnet inclusions in these early forming diamonds. The subsequent large uplift may be the result of Archean and possibly, in part, later tectonic events. The diamonds with the ‘warmer’ non-touching inclusions belong to one or more generations of uncertain age, which formed on a relaxed geotherm that was distinctly colder (~ 37 mW/m 2 ) than the xenolith geotherm. Our proposed scenario may offer a generic explanation for sporadic cases of ‘cold’ touching inclusions reported at other localities.