Subducting oceanic crust is a critical driver of chemical and lithological heterogeneity within the deep mantle. Zinc isotopes (δ66ZnJMC-Lyon) have been widely utilized to trace recycled oceanic crust and surficial materials. However, the lower oceanic crust rocks are not well studied given limited sampling, making it unclear if the δ66Zn estimated from mid-ocean ridge basalts (MORBs) is valid to represent that of the bulk oceanic crust. In this study, we report δ66Zn data for a series of gabbro cumulates (n = 37) from the ultraslow-spreading Southwest Indian Ridge, alongside MORBs (n = 12) from the fast-spreading East Pacific Rise and the slow-spreading South Mid-Atlantic Ridge. MORBs across various spreading rates exhibit homogeneous δ66Zn values (0.27 ± 0.05 ‰, 2sd), in agreement with previous studies. In contrast, the gabbros with several magma episodes exhibit a significant variation of δ66Zn from 0.11 ‰ to 0.34 ‰, with an average of 0.22 ± 0.11 ‰ (2sd). Magma differentiation to form oceanic crust can partly explain such variation (∼ 0.10 ‰), and post-cumulus modification further extends this range. Given the voluminous lower oceanic crust, the weighted δ66Zn estimated for the bulk oceanic crust (0.23 ± 0.03 ‰) is lower than the MORB-based value, which reconciles with the predictions from mantle partial melting. Therefore, recycled oceanic crust with significantly varied Zn isotopic compositions would lead to heterogeneous δ66Zn of the metasomatized mantle. The overall lower δ66Zn value underscores the importance of surficial materials other than oceanic crust, such as carbonates, to explain high δ66Zn of many mantle-derived magmas.
Read full abstract