Uranium isotope cycling on the highly productive Peruvian margin

S Bruggmann,G.J Gilleaudeau,S.J Romaniello,S Severmann,D.E Canfield,A.D Anbar,F Scholz,R Frei

doi:10.1016/j.chemgeo.2021.120705

Abstract

Uranium isotopes (δ238U values) in ancient sedimentary rocks (shales, carbonate rocks) are widely used as a tool to reconstruct paleo-redox conditions, but the behaviour of U isotopes under modern non-sulfidic anoxic vs. oxic conditions remains poorly constrained. We present U concentration and isotope data for modern sediments from the Peruvian margin, a highly productive open ocean environment with a range of redox conditions. To investigate U in different host fractions of the sediment (reactive, silicate, and HNO3-soluble fraction), we conducted a series of sequential extractions. Detrital-corrected authigenic U isotope compositions (δ238Uauth) in sediments deposited beneath an oxic water column show little deviation from the dissolved seawater U source, while anoxically deposited sediments have δ238Uauth values that are up to 0.4‰ heavier compared to seawater δ238U. Under anoxic, non-euxinic conditions, the U isotope offset between sediment and seawater is larger compared with oxic, but significantly smaller when compared with euxinic conditions from the literature. The results from sequential extractions show that the reactive sediment fraction records more pronounced differences in δ238Ureactive than δ238Uauth values depending on the oxidation state of the overlying water column. Furthermore, we found a strong correlation between total organic carbon (TOC) and both U concentrations (Uauth) and δ238Uauth values (R2 = 0.70 and 0.94, respectively) at the persistently anoxic site that we examined. These correlations can be caused by several processes including U isotope fractionation during microbially-mediated U reduction at the sediment-water interface (diffusive U input), during sorption onto and/or incorporation into organic matter in the water column (particulate U input) and diagenetic redistribution of U, or a combination of these processes. Our data show that several factors can influence δ238U values including oxidation state of U, the presence or absence of hydrogen sulfide and organic matter. These findings add new constraints to the degree of U isotope fractionation associated with U incorporation into sediments in different low-oxygen environments, thus aiding in interpretation of ancient paleo-redox conditions from U isotope data.

Highlights

The oceans underwent substantial changes in their redox state from the Archean Eon to the present (e.g. Diamond and Lyons, 2018), and they are bound to change in the future
The results from sequential extractions show that the reactive sediment fraction records more pronounced differences in δ238Ureactive than δ238Uauth values depending on the oxidation state of the overlying water column
We found a strong correlation between total organic carbon (TOC) and both U concentrations (Uauth) and δ238Uauth values (R2 = 0.70 and 0.94, respectively) at the persistently anoxic site that we examined

Summary

Introduction

The oceans underwent substantial changes in their redox state from the Archean Eon to the present (e.g. Diamond and Lyons, 2018), and they are bound to change in the future (e.g., deoxygenation; Keeling et al, 2010; Schmidtko et al, 2017). Reconstructing the extent of ocean anoxia in the past will reveal important information for future pre dictions. Trace metals such as Mo, V, Cr, and U are powerful tools to record redox changes in marine environments, because their solubility decreases under reducing conditions. Hexavalent U in oxic seawater occurs as the stable uranyl carbonate ion (UO2-CO3), or as Ca/Mg-UO2-CO3 complexes, at concentrations between 13 and 14 nmol kg− 1 (Ku et al, 1977; Langmuir, 1978; Endrizzi and Rao, 2014). Incorpora tion into organic particles has been recognised as a mechanism for U delivery to the sediment, especially in high productivity environments such as the Peruvian margin (Anderson et al, 1989a; McManus et al, 2006; Zheng et al, 2002a)

Methods

Results

Discussion

Conclusion