Sources and mobility of carbonate melts beneath cratons, with implications for deep carbon cycling, metasomatism and rift initiation

Kimberlite and carbonatite magmas that intrude cratonic lithosphere are among the deepest probes of the terrestrial carbon cycle. Their co-existence on thick continental shields is commonly attributed to continuous partial melting sequences of carbonated peridotite at >150 km depths, possibly as deep as the mantle transition zone. At Tikiusaaq on the North Atlantic craton in West Greenland, approximately 160 Ma old ultrafresh kimberlite dykes and carbonatite sheets provide a rare opportunity to study the origin and evolution of carbonate-rich melts beneath cratons. Although their Sr–Nd–Hf–Pb–Li isotopic compositions suggest a common convecting upper mantle source that includes depleted and recycled oceanic crust components (e.g., negative Δε _Hf coupled with >+5‰δ ⁷Li), incompatible trace element modelling identifies only the kimberlites as near-primary low-degree partial melts (0.05–3%) of carbonated peridotite. In contrast, the trace element systematics of the carbonatites are difficult to reproduce by partial melting of carbonated peridotite, and the heavy carbon isotopic signatures (−3.6 to −2.4‰δ ¹³C for carbonatites versus −5.7 to −3.6‰δ ¹³C for kimberlites) require open-system fractionation at magmatic temperatures. Given that the oxidation state of Earth's mantle at >150 km depth is too reduced to enable larger volumes of ‘pure’ carbonate melt to migrate, it is reasonable to speculate that percolating near-solidus melts of carbonated peridotite must be silicate-dominated with only dilute carbonate contents, similar to the Tikiusaaq kimberlite compositions (e.g., 16–33 wt.% SiO ₂). This concept is supported by our findings from the North Atlantic craton where kimberlite and other deeply derived carbonated silicate melts, such as aillikites, exsolve their carbonate components within the shallow lithosphere en route to the Earth's surface, thereby producing carbonatite magmas. The relative abundances of trace elements of such highly differentiated ‘cratonic carbonatites’ have only little in common with those of metasomatic agents that act on the deeper lithosphere. Consequently, carbonatite trace element systematics should only be used with caution when constraining carbon mobility and metasomatism at mantle depths. Regardless of the exact nature of carbonate-bearing melts within the mantle lithosphere, they play an important role in enrichment processes, thereby decreasing the stability of buoyant cratons and promoting rift initiation – as exemplified by the Mesozoic–Cenozoic breakup of the North Atlantic craton.