40AR/39AR AGES OF MICAS IN GARNET PERIDOTITES FROM THE BULTFONTEIN KIMBERLITE PIPE, SOUTH AFRICA

Abstract

40Ar/39Ar apparent ages on micas hosted in xenolithic garnet peridotites previously revealed the presence of extraneous 40Ar, i.e. the apparent ages significantly exceeded the respective kimberlite emplacement age. This mainly was interpreted to reflect either the presence of a mantle fluid component contributing excess 40Ar, or alternatively, to be due to relict 40Ar in course of incomplete isotopic rehomogenisation during the kimberlite event. The latter requires low effective diffusion coefficients of Ar in phlogopites under mantle conditions to avoid total loss of argon. If true, obtained relict ages could set time constraints on the metasomatic event, that led to phlogopite formation, and thus tell us more about the development of the local cratonic lithospheric mantle. For this purpose we determined 40Ar/39Ar ages of phlogopites from four garnet peridotites of the Bultfontein kimberlite, South Africa. We applied high resolution stepwise heating analyses. All obtained age spectra are disturbed, displaying lower apparent ages in the low temperature extractions and successively rising with increasing temperature, but apparent ages are always higher than the expected kimberlite emplacement age of ca. 90 Ma. Except for one sample (02BULT5), we generally obtained a very gentle increase in age with a maximum age of 730 Ma in sample 02BULT7 and a maximum age of 1030 Ma in samples 02BULT2 and 02BULT6. The shapes of all three spectra are plateau-like, but no plateau age sensu strictu could be calculated. Therefore, obtained maximum apparent ages must be regarded as minimum ages. Remarkably, the latter two samples show a different degree of 40Ar loss displayed in different lowered ages in the low temperature extractions. Hence, the samples seem not strongly disturbed by 40Ar loss or by contributions of excess 40Ar during the kimberlite emplacement event. We therefore suggest, that obtained sample specific maximum ages could bear some significance. However, we cannot rule out the presence of excess or inherited 40Ar during phlogopite formation, that would lead to an increase in apparent ages. In this case all maximum apparent ages would be meaningless. This requires that two different samples then should have suffered the same addition of excess/inherited 40Ar relative to the measured total 40Ar. Therefore, we consider this scenario as rather unlikely. Apparently, we obtained two different age constraints for one locality. Provided, both ages are geologic meaningful we may relate the formation of the different phlogopite generations to distinct orogenic events during which metasomatic fluids could have supplied the necessary water and K, e.g. during formation of the Namaqa-Natal orogenic belt (0.9-1.2 Ga). Alternatively, the apparent ages could represent cooling ages, i.e. formation of phlogopites occured much earlier (e.g. during stabilization of the Kaapvaal craton ca. 3 Ga ago), but cooling below the closure temperature of the K-Ar system was achieved later. The age difference for samples 02BULT7 (ca. 730 Ma) on the one hand and 02BULT2, 02BULT6 (ca. 1030 Ma) on the other hand could then reflect different thermal histories of both “groups”, with the latter much faster cooled down and hence, most likely stored at shallower levels than sample 02BULT7. Accordingly, samples 02BULT2 and 02BULT6 should have shared a very similar thermal history at the time before kimberlite emplacement. Distinction of both scenarios therefore principally can be achieved by a determination of the respective source p-T-conditions. Clear differences between both samples would discard the latter scenario and favour separate formation events. Resolving this issue remains a future task.