Carbonate sediments represent the major form of recycled carbon to the mantle via subduction.Carbon isotopes can easily distinguish organic carbon from inorganic carbon, but degassinginduced isotope fractionation of carbon during magma eruption [1] makes it difficult to track recycled carbonates in the sources of mantle-derived lavas.Marine carbonates (δ 66 Zn=0.99±0.25‰,[2-4]) are isotopically heavier by up to ~0.8‰ than the mantle (δ 66 Zn=0.18±0.05‰,[5][6][7]).Therefore, zinc isotopes may be promising tools of tracking subducted carbonates in the mantle.A large-scale heavy Zn isotopic anomaly was observed for Cenozoic basalts from Eastern China, which is spatially consistent with the sub-horizontally stagnated paleo-Pacific slab in the mantle transition zone (MTZ) [8][9].This coupling suggests that carbonates can be deeply subducted into the MTZ.Similarly heavy Zn isotopic compositions and the spatial coupling of heavy δ 66 Zn with the stagnated Neo-Tethyan oceanic slab were also observed in Cenozoic basalts from southwest China and central Myanmar in SE Tibet [10][11].Thus, oceanic subduction has the potential to transport surface carbon into the MTZ (410-660 km) globally, at timescales that significantly exceed those of arc-trench cycle.
Abstract Intraplate basaltic volcanism commonly exhibits wide compositional ranges from silica-undersaturated alkaline basalts to silica-saturated tholeiitic basalts. Possible mechanisms for the compositional transition involve variable degrees of partial melting of a same source, decompression melting at different mantle depths (so-called ‘lid effect’), and melt-peridotite interaction. To discriminate between these mechanisms, here we investigated major-trace elemental and Sr–Nd–Mg–Zn isotopic compositions of a suite of intraplate alkaline and tholeiitic basalts from the Datong volcanic field in eastern China. Specifically, we employed Mg and Zn isotope systematics to assess whether the silica-undersaturated melts originated from a carbonated mantle source. The alkaline basalts have young HIMU-like Sr and Nd isotopic compositions, lower δ26Mg (-0·42‰ to -0·38‰) and higher δ66Zn (0·40‰ to 0·46‰) values relative to the mantle. These characteristics were attributable to an asthenospheric mantle source hybridized by carbonated melts derived from the stagnant Pacific slab in the mantle transition zone. From alkaline to tholeiitic basalts, δ26Mg gradually increases from -0·42‰ to -0·28‰ and δ66Zn decreases from 0·46‰ to 0·28‰ with decreasing alkalinity and incompatible trace element abundances (e.g. Rb, Nb, Th and Zr). The Mg and Zn isotopic variations are significantly beyond the magnitude (<0·1‰) induced by different degrees of fractional crystallization and partial melting of a same mantle source, excluding magmatic differentiation, different degrees of partial melting and the ‘lid effect’ as possible mechanisms accounting for the compositional variations in the Datong basalts. There are strong, near-linear correlations of δ26Mg and δ66Zn with 87Sr/86Sr (R2=0·75 − 0·81) and 143Nd/144Nd (R2=0·83 − 0·90), suggesting an additional source for the Datong basalts. This source is characterized by pristine mantle-like δ26Mg and δ66Zn values as well as EM1-like Sr–Nd isotopic ratios, pointing towards a metasomatized subcontinental lithospheric mantle (SCLM). Isotope mixing models show that mingling between alkaline basaltic melts and partial melts from the SCLM imparts all the above correlations, which means that the SCLM must have been partially melted during melt-SCLM reaction. Our results underline that interaction between carbonated silica-undersaturated basaltic melts and the SCLM acts as one of major processes leading to the compositional diversity in intracontinental basaltic volcanism.
Fractional crystallization plays a critical role in generating the differentiated continental crust on Earth. However, whether efficient crystal-melt separation can occur in viscous felsic magmas remains a long-standing debate because of the difficulty in discriminating between differentiated melts and complementary cumulates. Here, we found large (~1 per mil) potassium isotopic variation in 54 strongly peraluminous high-silica (silicon dioxide >70 weight %) leucogranites from the Himalayan orogen, with potassium isotopes correlated with trace elemental proxies (e.g., strontium, rubidium/strontium, and europium anomaly) for plagioclase crystallization. Quantitative modeling requires up to ~60 to 90% fractional crystallization to account for the progressively light potassium isotopic composition of the fractionated leucogranites, while plagioclase accumulation results in enrichment of heavy potassium isotopes in cumulate leucogranites. Our findings strongly support fractional crystallization of high-silica magmas and highlight the great potential of potassium isotopes in studying felsic magma differentiation.
To test the ability of Mg and Zn isotopes in discriminating between different types of mantle metasomatism and identifying deep carbon cycling, here we present a comparative study on two types of Cenozoic lavas in SE Tibet, i.e., potassic-ultrapotassic lavas and alkali basalts. The contrasting chemical compositions, Sr-Nd isotopic ratios and olivine chemistry between them suggest distinct sources in the enriched lithospheric mantle and asthenosphere, respectively. 26Mg and 66Zn values of the K-rich lavas are shifted toward slightly heavier and lighter values relative to global oceanic basalts, respectively, indicating source metasomatism by recycled siliciclastic sediments. By contrast, the alkali basalts possess remarkably light 26Mg and heavy 66Zn values that are typically characterized by marine carbonates. The coupling of high 66Zn with high Zn contents and Zn/Fe ratios further suggests a pyroxenite source containing recycled Zn-rich magnesian carbonates. This is corroborated by the similarity in major elements between the alkali basalts and experimental partial melts of CO2 + pyroxenite. Thus, Mg and Zn isotopes could be an effective tool of discriminating between silicate and carbonate metasomatism in the mantle. Notably, the occurrence of the alkali basalts is spatially consistent with a stagnant slab in the mantle transition zone (410–660 km), the latter of which is interpreted to represent the deeply subducted oceanic slab. These findings thus provide evidence for recycling of carbonates into the deep mantle, which represents a long-term circulation of subducted carbon compared with that in arc systems and has crucial significances for global deep carbon cycling.