60th Annual Report on Research 2015

The isotopic composition of authigenic uranium in ancient sediments has been used to reconstruct paleoredox conditions in the oceans. The rationale is that during sediment deposition in anoxic, euxinic, suboxic, and oxic sediments, the extents to which uranium isotopes are fractionated differ. Because the residence time of U in the ocean is longer than the mixing timescale by two orders of magnitude, the global ocean is homogeneous in terms of U concentration and isotopic composition. Thus, at steady state, the uranium isotopic composition of seawater should be influenced by the areal extents of sediments formed under different conditions.

In year one of the proposal, we put the U isotope system to a test by (1) precisely determining the U isotopic composition of the ocean, and (2) comparing this value with the isotopic composition predicted by commonly used U budgets. This first order test showed that, indeed, the ²³⁸U/²³⁵U ratio is a proper tracer of global redox conditions and allowed us to constrain the extent of modern oceanic anoxia (Tissot and Dauphas, 2015).

The remaining of year one and part of year two were spent developing a sequential leaching protocol for carbonate samples that would allow us to access the original U isotopic composition of the sample. There are two reasons for this approach (1) the concentration of U in carbonates is low (ppm level or less) and a large mass of sample must be digested for high precision measurements, and (2) recent work on uranium isotopes in carbonates showed that anoxic conditions in pore water can lead to authigenic uranium enrichment in the carbonate, shifting the isotopic composition recorded in bulk rock away from the seawater value (Romaniello et al., 2013). The step leaching protocol we developed was tested on a modern coral. We found that the first and last few leachates are depleted in U while the intermediate leachates show higher than bulk concentrations (Fig. 1). Similarly, the isotopic composition of U collected in each step is not constant and can differ from the isotopic composition of modern seawater by up to 0.2 ä (four times the 2SE of the measurements). From these tests, a two step leaching was designed for carbonate samples. In the first leaching step only ~25 % of the sample is digested in order to remove easily mobilized U (i.e., U contamination), then in a second step, the bulk of the carbonate (~70 % of the sample) is digested in order to access the original U isotopic composition of the carbonate sample. The reason for leaving 5 % of the sample undigested is to avoid leaching of U from phases other than carbonates (e.g., detritic phases), whose isotopic composition will likely be different from the one of the carbonate.

Once established, this protocol was used on an array of modern carbonates (age from 0 to 1.5 My) taken along a drill core in the Bahamas bank. The drill core is 225 m deep, and samples were taken every 5 m down the core in order to assess the effect of early diagenesis on the uranium proxy and validate our methodology at the same time. For all samples, both the ²³⁸U/²³⁵U ratio and the ²³⁴U/²³⁸U ratio were determined (²³⁴U is a decay product of ²³⁸U). The abundance of ²³⁴U in a sample is a function of the decay constants of ²³⁴U and ²³⁸U (l₂₃₄ and l₂₃₈) and the time since closure of the sample. At steady-state, the decay of ²³⁴U is balanced by the creation of ²³⁴U from ²³⁸U decay, and the ²³⁴U/²³⁸U ratio in the sample is equal to the ratio of the decay constants: l₂₃₈/l₂₃₄= 5.497 ×10^-5 (i.e., the secular equilibrium ratio). We observe a strong decreasing trend in d²³⁴U values (the deviation from the secular equilibrium ratio, in permil notation) with depth, whereby the sample close to the water/sediment interface have seawater-like d²³⁴U values (~+147 ä), and the samples close to the bottom of the core have d²³⁴U value close to secular equilibrium (~ 10 ä) (Fig 2).

To first order, this is in good agreement with closure of the samples a few meters below the water/sediment interface, and return to secular equilibrium in a closed system. The fit in Fig. 2 (green curve) is based on sample ages inferred from calcareous nanofossils and planktonic foraminiferal biostratigraphic datums (Eberli et al., 1997). The main conclusion of this graph is that any diagenetic process affecting U isotopes must take place within the first 10 m of the water/sediment interface (equivalent to ~ 60 kyr). The ²³⁸U/²³⁵U ratio shows a shift relative to seawater values, and some scatter in the data (Fig. 3). The average isotopic shift recorded in the samples is ~+0.25 ä relative to seawater. Because the U system closes ~60 kyr after carbonate formation, this offset must be the result of early diagenesis (probably due to precipitation of U from porewater within the sediment) rather than late secondary processes. Given the scatter in the carbonate data, any reconstruction of past ocean redox state using U isotopes will have to rely on a large number of data points, to identify statistically meaningful trends.

Having studied the modern ocean and the effect of early diagenesis, we started analyzing a series of carbonates whose ages span as much of Earth's history as possible, with the end goal of tracking the timing of Earth's oxygenation. The sample collection consists of ~220 samples, of ages ranging from present to 3.4 Gyr. Fig. 4 presents the data acquired so far. We are in the process of measuring the remaining samples.

This grant has supported methodology development, samples characterization and high precision isotopic analyses for one of my PhD student, Francois Tissot, and two chemistry major undergraduates, Benjamin Matthew Go, Magdalena Naziemiec. A paper was already published on the work done during year one (Tissot and Dauphas 2015). The study on the drill core from the Bahamas bank (see above), and the results on older samples will give rise to two additional publications in a timeframe of a year.