Diamond provenance studies from 40Ar/ 39Ar dating of clinopyroxene inclusions: An example from the west coast of Namibia

Abstract

The west coast of Namibia is host to substantive detrital diamond deposits located in onshore and offshore beach gravels, desert deflation deposits and lower Orange river terraces. The origin of the Namibian diamonds is controversial, with some studies favouring derivation from distal Cretaceous/Jurassic kimberlites on the Kaapvaal craton, and others arguing that most diamonds originated from proximal Dwyka glacial deposits (~ 300 Ma), which incorporated diamonds from older (≥ 500 Ma), pre-Karoo kimberlites. Previous studies have demonstrated that clinopyroxene inclusions extracted from their host diamonds give 40Ar/ 39Ar ages approaching the time of source kimberlite eruption. This behaviour is attributed to diffusion of argon to lattice defect sites and the diamond/inclusion interface region during mantle residence, with subsequent loss of the latter component on cleaving of the diamond to release the inclusion(s). In this study, we measured 40Ar/ 39Ar ages of extracted clinopyroxene inclusions from Namibian detrital diamonds, in order to determine potential kimberlite sources, craton erosion histories and palaeo-drainage evolution in southern Africa. 40Ar/ 39Ar step-heating data were obtained for eclogitic and peridotitic clinopyroxene inclusions from 50 Namibian diamonds. Low temperature steps produced older apparent ages than high temperature (fusion) steps, consistent with partial retention of pre-eruption argon in defect sites. With one exception, fusion steps yielded younger ages, ranging from 62 ± 30 Ma to 1441 ± 700 Ma. The majority (80%) of inclusions have fusion ages < 300 Ma, indicating that most Namibian detrital diamonds originated from post-Dwyka (< 300 Ma) kimberlites. Six inclusion aliquots (13%) produced ages unique to Cretaceous Group I kimberlites, confirming erosion of diamonds from these sources. The proportion of diamonds sourced from Group II kimberlites is uncertain, although forward modelling suggests roughly equal quantities from Group I and Group II localities. A lesser population of ages close to ~ 240 Ma could indicate minor contributions from syn-Karoo kimberlites (e.g., Jwaneng cluster, Botswana); however derivation from younger kimberlites is considered equally plausible. Based on the ages of known southern African kimberlites, the current results suggest that very few (< 11%), if any, Namibian diamonds were sourced from Dwyka glacial deposits or older sediments. Estimates of erosion levels for southern African kimberlites suggest that Cretaceous Group I and Group II kimberlites in South Africa are the most likely sources of the Namibian diamond deposits. Palaeo-drainage reconstructions are consistent with diamond transport to the Namibian coast by the palaeo-Karoo and modern Orange river systems. If the ~ 240 Ma inclusion ages represent contributions from the Jwaneng kimberlites, this would support models for an earlier Kalahari river with headwaters in southern Botswana.