²³⁸U/²³⁵U in calcite is more susceptible to carbonate diagenesis

The uranium isotopic composition (δ²³⁸U) of bulk marine calcium carbonates has been extensively explored as a promising paleoredox proxy to track the extent of global oceanic anoxia in deep time. Multiple studies have examined whether primary calcium carbonates can directly capture seawater δ²³⁸U and whether bulk measurements of recent and ancient carbonates preserve seawater U isotope signatures. Here we assess the role of diagenesis in altering δ²³⁸U signatures in carbonates sediments that have a primary calcitic mineralogy at the Paleocene-Eocene Thermal Maximum (PETM), an interval with rapid global warming and oceanic deoxygenation at ∼56 million years ago. Although primary abiotic and biogenic calcium carbonates (aragonite and calcite) can directly capture seawater δ²³⁸U with small offsets (<0.1‰) relative to modern seawater, diagenetic alteration of Bahamian shallow-water platform carbonate sediments that have a predominantly primary aragonitic mineralogy resulted in significantly larger offsets (up to 0.6‰). Since U concentration in primary aragonite is at least one order of magnitude higher than primary calcite (>1 ppm vs. <0.1 ppm), δ²³⁸U in calcite should be even more susceptible to diagenesis than that in aragonite. We find strong evidence of this effect in analysis of δ²³⁸U in PETM shallow-water carbonate sediments from Drilling Project (ODP) Hole 871C (Limalok Guyot, Pacific Ocean). Our results reveal large fluctuations in bulk carbonate δ²³⁸U from −0.69 to +0.71‰ around the PETM boundary but consistently heavier δ²³⁸U (between −0.14 and +0.47‰) than modern seawater outside of this interval. The significantly lighter δ²³⁸U values than modern seawater were interpreted to result from the operation of a Mn oxide shuttle. The heavier δ²³⁸U values are most likely caused by authigenic reductive accumulation of U(IV) in pore waters below the sediment-water interface. We found that carbonate δ²³⁸U values higher than modern seawater tend to increase with increasing U/Ca. This relationship is well-explained by an authigenic reductive accumulation model that simply assumes addition to primary calcite during diagenesis of calcitic cements containing isotopically heavier U(IV). Our work confirms expectations that δ²³⁸U in primary calcite is more susceptible to the amount of diagenetic cementation compared to primary aragonite, and that variations of δ²³⁸U in carbonate sediments with a primary calcitic mineralogy would more dominantly reflect the local redox state of depositional and early diagenetic environments. It is essential to identify the original carbonate mineralogy, the diagenetic history, and constrain the redox state of local deposition environments of sedimentary carbonate rocks when applying bulk carbonate δ²³⁸U as a global proxy for oceanic anoxia in deep time.