Origins of kimberlites and carbonatites during continental collision - Insights beyond decoupled Nd-Hf isotopes

During the past two decades significant progress has been made in understanding the origin and evolution of kimberlites, including relationships to other diamondiferous magma types such as lamproites and aillikites. However, the association of kimberlites and carbonatites on continental shields remains poorly understood, and two opposing ideas dominate the debate. While one school of thought argues that primary carbonatite melts transform into hybrid carbonated silicate magmas akin to kimberlites by assimilation of cratonic mantle material, others use geochemical evidence to show that carbonatite magmas can evolve from near-primary kimberlite melts within the cratonic lithosphere. The 1.15 Ga Premier kimberlite pipe on the Kaapvaal craton in South Africa hosts several kimberlite and carbonatite dykes. Reconstructions of magma compositions suggest that up to 20 wt.% CO2 was lost from nearprimary kimberlite melts during ascent through the cratonic lithosphere, but the carbonatite dyke compositions cannot be linked to the kimberlite melts via differentiation. Geochemical evidence, including mantle-like delta C-13 compositions, suggests that the co-occurring kimberlite and carbonatite dykes represent two discrete CO2-rich magma batches derived from a mixed source in the convecting upper mantle. The carbonatites probed a slightly more depleted source component in terms of Sr-Nd-Hf isotopic compositions relative to the peridotitic matrix that was more effectively tapped by the kimberlites (Sr-87/Sr-86, = 0.70257 to 0.70316 for carbonatites vs. 0.70285 to 0.70546 for kimberlites; epsilon Nd-i = + 3.0 to + 3.9 vs. + 2.2 to + 2.8; epsilon Hf-i = -2.2 to + 0.7 vs. -5.1 to -1.9). Platinum-group element systematics suggest that assimilation of refractory lithospheric mantle material by the carbonatite melts was negligible ( < 1 vol.%), whereas between 5 - 35 vol.% of digested cratonic peridotite account for the kimberlite compositions, including the low Os-187/Os-188 signature (gamma Os-i = -12.7 to -4.5). The kimberlite and carbonatite dykes show similarly strong Nd-Hf isotope decoupling (Delta epsilon Hf-i= -10.7 to -7.6 vs. -8.8 to -6.1), regardless of the variable lithospheric mantle imprints. This observation suggests a common sublithospheric origin of the negative Aellf signature, possibly linked to ancient recycled oceanic crust components in the convecting upper mantle to transition zone sources of CO2-rich magmatism. Mesoproterozoic kimberlite and carbonatite magmatism at Premier was coeval with subduction and collision events along the southern Kaapvaal craton margin during the 1,220 -1,090 Ma Namaqua-Natal orogeny associated with Rodinia supercontinent formation. Thermochronology suggests that the entire Kaapvaal craton was affected by this collisional tectonic event, and it appears that the changing lithospheric stress-field created pathways for deep-sourced kimberlite and carbonatite magmas to reach Earth's surface. We find that collisioninduced (e.g., Premier) and continental breakup-related (e.g., Kimberley) kimberlite magmas are compositionally indistinguishable, with the inference that plate tectonic processes aid solely in the creation of magma ascent pathways without a major influence on deep mantle melting beneath cratons. It follows that on-craton kimberlite magmatism in the hinterland of collision zones is not necessarily more likely to entrain large sublithospheric diamonds than kimberlite eruptions linked to continental breakup. This implies that Premier's world-class endowment with 'ultradeep' Type-II diamonds is not causally related to its setting behind an active orogenic front.