Calcium isotopic signatures of carbonatite and silicate metasomatism, melt percolation and crustal recycling in the lithospheric mantle

Ca isotopes can be strongly fractionated at the Earth's surface and thus may be tracers of subducted carbonates and other Ca-rich surface materials in mantle rocks, magmas and fluids. However, the δ ^44/40 Ca range in the mantle and the scope of intra-mantle isotope fractionation are poorly constrained. We report Ca isotope analyses for 22 mantle xenoliths: four basalt-hosted refractory peridotites from Tariat in Mongolia and 18 samples from the Obnazhennaya (Obn) kimberlite on the NE Siberian craton. Obn peridotites are Paleoproterozoic to Archean melting residues metasomatised by carbonate-rich and/or silicate melts including unique xenoliths that contain texturally equilibrated carbonates. δ ^44/40 Ca in 15 Obn xenoliths shows limited variation (0.74–0.97‰) that overlaps the value (0.94 ± 0.05‰) inferred for the bulk silicate Earth from data on fertile lherzolites, but is lower than δ ^44/40 Ca for non-metasomatised refractory peridotites from Mongolia (1.10 ± 0.03‰). Bulk δ ^44/40 Ca in four Obn peridotites containing metasomatic carbonates ranges from 0.81 ± 0.08‰ to 0.83 ± 0.06‰ with similar values in acid-leachates and leaching residues, indicating isotopic equilibration of the carbonates with host rocks. We infer that (a) metasomatism tends to decrease δ ^44/40 Ca values of the mantle, but its effects are usually limited (≤0.3‰); (b) Ca isotopes cannot distinguish “carbonatite” and “silicate” types of mantle metasomatism. The lowest δ ^44/40 Ca value (0.56‰) was obtained for a phlogopite-bearing Obn peridotite with a very high Ca/Al of 8 suggesting that the greatest metasomatism-induced Ca isotope shifts may be seen in rocks initially low in Ca that experienced significant Ca input leading to high Ca/Al. Two Obn peridotites, a dunite (melt channel material) and a veined spinel wehrlite, have high δ ^44/40 Ca values (1.22‰ and 1.38‰), which may be due to isotope fractionation by diffusion during silicate melt intrusion and percolation in the host mantle. Overall, we find no evidence that recycling of crustal carbonates may greatly affect Ca isotope values in the global mantle or on a regional scale.