Testing the robustness of the carbonate uranium isotope proxy: Evidence from a partial dissolution study

Abstract

Uranium isotopes in carbonates have been used extensively to determine past extent of global oceanic anoxia. However, local spatial and temporal variations in reducing conditions and/or subsequent diagenesis can result in isotopic variations of U (²³⁸U/²³⁵U) that complicate the extrapolation of U isotopic variation in ancient carbonate sediments to global estimates of ocean anoxia. The fidelity of the U isotopic signal can also be impacted by the co-dissolution of U-containing non‑carbonate phases, such as aluminosilicates, manganese-oxide coatings, and iron-containing minerals, that might introduce an independent U isotopic signal that alters or biases the global ocean-derived δ²³⁸U value recorded by the carbonate component. Leaching protocols and/or precleaning steps have been introduced to avoid the dissolution of these non‑carbonate phases and extract the primary carbonate signal; however, these methods can lengthen the sample preparation time while also potentially introducing contamination. The work presented here suggests that a one-step partial dissolution procedure using excess 0.08 M nitric acid or 2 M acetic acid results in Ca, Mg, and Mn concentrations that are similar to concentrations determined from total dissolution methods. This approach precludes the complete dissolution of Sr, U, Fe, and Al (and occasionally Mg and Mn) in insoluble non‑carbonate phases. When targeting the dissolution of only ∼80 % of the sample, the release of Fe and Al from non‑carbonate phases is minimized. This partial dissolution approach still results in release of sufficient U to permit interpretation of past reducing conditions. Measured δ²³⁸U values are nearly invariant regardless of acid type or concentration, suggesting that the dissolution of the non‑carbonate phases does not adversely affect the U isotopic signal.