Preservation of an Archaean whole rock Re-Os isochron for the Venetia lithospheric mantle: evidence for rapid crustal recycling and lithosphere stabilisation at 3.3 Ga

Abstract

Re-Os and platinum group element analyses are reported for peridotite xenoliths from the 533 Ma Venetia kimberlite cluster situated in the Limpopo Mobile Belt, the Neoarchaean collision zone between the Kaapvaal and Zimbabwe Cratons. The Venetian xenoliths provide a rare opportunity to examine the state of the cratonic lithosphere prior to major regional metasomatic disturbance of Re-Os systematics throughout the Phanerozoic. The 32 studied xenoliths record Si-enrichment that is characteristic of the Kaapvaal lithospheric mantle and can be subdivided into five groups based on Re-Os analyses. The most pristine group I samples (n = 13) display an approximately isochronous relationship and fall on a 3.28 ± 0.17 Ga (95% conf. int.) reference line that is based on their mean TMA age. This age overlaps with the formation age of the Limpopo crust at 3.35-3.28 Ga. The group I samples derive from ∼50 to ∼170 km depth, suggesting coeval melt depletion of the majority of the Venetia lithospheric mantle column. Group II and III samples have elevated Re/Os due to Re addition during kimberlite magmatism. Group II has otherwise undergone a similar evolution as the group I samples with overlapping 187Os/188Os at eruption age: 187Os/188OsEA, while group III samples have low Os concentrations, unradiogenic 187Os/188OsEA and were effectively Re-free prior to kimberlite magmatism. The other sample groups (IV and V) have disturbed Re-Os systematics and provide no reliable age information. A strong positive correlation is recorded between Os and Re concentrations for group I samples, which is extended to groups II and III after correction for kimberlite addition. This positive correlation precludes a single stage melt depletion history and indicates coupled remobilisation of Re and Os. The combination of Re-Os mobility, preservation of the isochronous relationship, correlation of 187Os/188Os with degree of melt depletion and lack of radiogenic Os addition puts tight constraints on the formation and subsequent evolution of Venetia lithosphere. First, melt depletion and remobilisation of Re and Os must have occurred within error of the 3.28 Ga mean TMA age. Second, the refractory peridotites contain significant Re despite recording >40% melt extraction. Third, assuming that Si-enrichment and Re-Os mobility in the Venetia lithospheric mantle were linked, this process must have occurred within ∼100 Myr of initial melt depletion in order to preserve the isochronous relationship. Based on the regional geological evolution, we propose a rapid recycling model with initial melt depletion at ∼3.35 Ga to form a tholeiitic mafic crust that is recycled at ∼3.28 Ga, resulting in the intrusion of a TTG suite and Si-enrichment of the lithospheric mantle. The non-zero primary Re contents of the Venetia xenoliths imply that TRD model ages significantly underestimate the true depletion age even for highly depleted peridotites. The overlap of the ∼2.6 Ga TRD ages with the time of the Kaapvaal-Limpopo collision is purely fortuitous and has no geological significance. Hence, this study underlines the importance of scrutiny if age information is to be derived from whole rock Re-Os analyses.