Tracking halogen recycling and volatile loss in kimberlite magmatism from Greenland: Evidence from combined F-Cl-Br and δ37Cl systematics

Abstract

Kimberlite magmatism occurs as a result of volatile fluxed melting of the convecting upper mantle underlying cratonic lithospheric mantle regions. During passage to the Earth's surface, proto-kimberlite magma can interact with, and assimilate, variably enriched cratonic mantle producing hybrid melts consisting of asthenospheric and cratonic mantle components including contributions from metasomatic domains. The halogen elements (F, Cl, Br, I) and chlorine isotope ratios (37Cl/35Cl) are increasingly used as tracers of recycled crustal materials within the Earth's mantle yet are only rarely reported in analyses of kimberlites. As a result, the origin and distribution of halogens in kimberlite magmas is poorly constrained. Here, we present novel, combined elemental (F, Cl, Br) and isotopic (δ37Cl) halogen data for 14 fresh kimberlite samples from the North Atlantic Craton (NAC) of West and South-West Greenland.The F composition of kimberlites from the NAC appears to be controlled by melting in the convecting upper mantle with minimal effect from interaction with metasomatized lithospheric mantle or volatile loss during or after emplacement. By contrast, Cl and Br in the studied samples have undergone significant devolatilization during kimberlite dyke emplacement and post-emplacement processes, whereby up to 99% of the original halogen budget was removed.Whilst all the studied kimberlites broadly follow the same geochemical pattern, there exists some regional variability in their halogen systematics. The northern NAC kimberlite samples from Majuagaa have mantle-like δ37Cl values of −0.2 to −0.5‰ [versus SMOC (standard mean ocean chloride)]. In contrast, kimberlite dykes from Nigerdlikasik and Pyramidefjeld near the southern craton margin display positive δ37Cl values of +0.4‰ to +1.3‰, in addition to a relative Cl and Br enrichment, which is consistent with the assimilation of recycled crust-derived halogens by the kimberlite magmas. The data support a scenario in which recycled halogens were sampled either from within an OIB-type reservoir in the convecting mantle or through interaction with subduction-modified lithospheric mantle reservoir during eruption. We prefer a scenario in which the ascending kimberlite magmas assimilated Cl-rich, metasomatized regions within cratonic mantle lithosphere.