Effects of calcium and phosphate on uranium(IV) oxidation: Comparison between nanoparticulate uraninite and amorphous U^IV–phosphate

Abstract

The mobility of uranium in subsurface environments depends strongly on its redox state, with U^IV phases being significantly less soluble than U^VI minerals. This study compares the oxidation kinetics and mechanisms of two potential products of U^VI reduction in natural systems, a nanoparticulate UO₂ phase and an amorphous U^IV–Ca–PO₄ analog to ningyoite (CaU^IV(PO₄)₂·1–2H₂O). The valence of U was tracked by X-ray absorption near-edge spectroscopy (XANES), showing similar oxidation rate constants for U^IVO₂ and U^IV–phosphate in solutions equilibrated with atmospheric O₂ and CO₂ at pH 7.0 (k_obs,UO2 = 0.17 ± 0.075 h⁻¹ vs. k_obs,U^IV_PO4 = 0.30 ± 0.25 h⁻¹). Addition of up to 400 μM Ca and PO₄ decreased the oxidation rate constant by an order of magnitude for both UO₂ and U^IV–phosphate. The intermediates and products of oxidation were tracked by electron microscopy, powder X-ray diffraction (pXRD), and extended X-ray absorption fine-structure spectroscopy (EXAFS). In the absence of Ca or PO₄, the product of UO₂ oxidation is Na–uranyl oxyhydroxide (under environmentally relevant concentrations of sodium, 15 mM NaClO₄ and low carbonate concentration), resulting in low concentrations of dissolved U^VI (<2.5 × 10⁻⁷ M). Oxidation of U^IV–phosphate produced a Na-autunite phase (Na₂(UO₂)PO₄·xH₂O), resulting in similarly low dissolved U concentrations (<7.3 × 10⁻⁸ M). When Ca and PO₄ are present in the solution, the EXAFS data and the solubility of the U^VI phase resulting from oxidation of UO₂ and U^IV–phosphate are consistent with the precipitation of Na-autunite. Bicarbonate extractions and Ca K-edge X-ray absorption spectroscopy of oxidized solids indicate the formation of a Ca–U^VI–PO₄ layer on the UO₂ surface and suggest a passivation layer mechanism for the decreased rate of UO₂ oxidation in the presence of Ca and PO₄. Interestingly, the extractions were unable to remove all of the oxidized U from partially oxidized UO₂ solids, suggesting that oxidized U is distributed between the interior of the UO₂ nanoparticles and the labile surface layer. Accounting for the entire pool of oxidized U by XANES is the likely reason for the higher UO₂ oxidation rate constants determined here relative to prior studies. Our results suggest that the natural presence or addition of Ca and PO₄ in groundwater could slow the rates of U^IV oxidation, but that the rates are still fast enough to cause complete oxidation of U^IV within days under fully oxygenated conditions.

Introduction

Uranium is a toxic and radioactive element that has been used for energy generation and military purposes for several decades, resulting in stockpiles of spent fuel, mine tailings, and enrichment process waste stored at many locations around the world. Designing appropriate storage practices and predicting the effects of accidental release requires an in-depth understanding of the coupled chemical, physical, biological, and hydrological processes that control the mobility of U in oxic and anoxic environments. Developing this understanding is currently hindered by the limited availability of mechanistic information on the transformations and speciation of U under environmentally relevant conditions (Burns et al., 2012, Bargar et al., 2013, Williams et al., 2013, Newsome et al., 2014).

The solubility, and therefore the mobility, of uranium is strongly affected by its valence state. In oxidizing environments uranium is stable as U^VI. When equilibrium with schoepite (UO₃·2H₂O) controls U^VI solubility, the resulting aqueous concentrations are on the order of 10⁻⁴ M at pH 7 (Jang et al., 2006). In reducing environments uranium is stable as U^IV. When equilibrium with uraninite (UO_2+x) controls U^IV solubility, the resulting aqueous concentrations are on the order of 10⁻⁸ M at pH 7 (Ulrich et al., 2008). The dramatic decrease in solubility in this simplified scenario is expected to significantly influence U mobility in the subsurface and has been the impetus for extensive research on (bio-) reduction of U^VI for the purpose of contaminated site remediation. However, uranium transformations in the subsurface occur in the presence of various surfaces and dissolved ions, which can affect the solubility, speciation, and outcome of reactions with U. For instance, carbonate forms strong, highly soluble complexes with U^VI and together with Ca²⁺ decreases the ability of U^VI to be reduced to U^IV (Brooks et al., 2003). Phosphate and phosphatase activity influences U^VI solubility through the formation of precipitates (Beazley et al., 2007), which can affect the rate of U^VI bioreduction (Rui et al., 2013). U^VI can interact with various mineral and biological surfaces to form stable adsorption complexes (Bargar et al., 1999, Kelly et al., 2002). The complexity in U^VI speciation described above is accounted for in most transport models. However, only an amorphous uraninite phase is typically used to model the behavior of reduced U^IV, under the assumption that the lowest solubility mineral controls U^IV dynamics (Yabusaki et al., 2007a, Li et al., 2009, Li et al., 2010). Research in the past few years has drawn this assumption into question, at least for processes occurring over the months-to-years timescale. Low levels of phosphate were shown to inhibit the formation of uraninite during U^VI reduction, resulting in non-uraninite, phosphate-complexed U^IV solid phases (Boyanov et al., 2011, Stylo et al., 2013, Alessi et al., 2014b). Investigations of U speciation in reduced sediments and soils also indicates the prevalence of non-uraninite U^IV species over uraninite, although the exact identity of U^IV species remains unclear (Bargar et al., 2013, Wang et al., 2014, Alessi et al., 2014a, Li et al., 2015). Recently, high-affinity mineral surface sites were shown to stabilize mononuclear U^IV adsorption complexes and thus inhibit uraninite formation (Latta et al., 2014). Subsequent work developed a surface complexation model and determined stability constants for such sites (Wang et al., 2015a). These developments indicate a significant gap in the description of short-term U^IV dynamics in transport models and suggest the need for the study of non-uraninite U^IV reactions in model systems so appropriate reactions can be included in the more complex field-scale models (Long et al., 2015).

Of particular importance to U stability under the changing redox conditions in natural environments is the oxidation of U^IV phases, as U^VI species have а much higher potential for remobilization. Oxidation can occur due to the influx of dissolved oxidants, as well as due to water radiolysis under anoxic conditions (Ekeroth and Jonsson, 2003, Ulrich et al., 2008). A number of studies have addressed uraninite oxidation in batch and electrochemical experiments from the perspective of nuclear fuel corrosion, providing a detailed and mechanistic understanding of the process (Shoesmith, 2000, Roth and Jonsson, 2008 and references therein). Electrochemical studies coupled with surface sensitive X-ray photoelectron spectroscopy have shown that UO₂ oxidation proceeds via surface oxidation to UO_2.33, followed by accumulation of surface UO₂²⁺ layers as oxidation further progresses (Shoesmith et al., 1989, Shoesmith, 2000). Accumulation of surface U^V/U^VI hinders the interpretation of batch and flow UO₂ oxidation experiments. Surface layers of UO_2.33 do not form under high-bicarbonate conditions because U^VI dissolves directly to solution (Shoesmith, 2000, Ulrich et al., 2009). Hence, most electrochemical studies have used high bicarbonate concentrations to provide meaningful results for UO₂ oxidation and dissolution. On the other hand, electrochemical studies only measure electron transfer reactions, and thus require supporting evidence such as spectroscopic or diffraction data to provide comprehensive information about non-redox phase transformations.

More recent studies have looked at the oxidation of uraninite from the perspective of U release following reductive remediation. Consistent with electrochemical studies on UO₂ spent fuel corrosion, several studies of UO₂ dissolution after reductive immobilization have found that carbonate increases the rate of oxidative UO₂ dissolution through promotion of U^V/U^VI surface layer detachment (Pierce et al., 2005, Ulrich et al., 2009). Interestingly, Ulrich and coworkers found that biogenic nanoparticulate uraninite dissolved at a similar surface-area-normalized rate as larger particles of synthetic UO₂, but nanoparticulate UO₂ supported a higher concentration of U in solution during oxidation, presumably due to higher surface strain in the smaller particles. Several studies to date have shown that mass transport has important effects on the mobility of U in subsurface systems, and have found that the release of U^VI from UO₂ is slowed in diffusion limited and advective systems relative to mixed batch reactors (Campbell et al., 2011, Giammar et al., 2012).

In contrast to the extensive studies of UO₂ oxidation, very limited information is available on the stability and oxidation of non-uraninite U^IV phases. U^IV in biogenic, non-uraninite solids was found more labile to dissolution than uraninite in 1 M anoxic bicarbonate extractions (Alessi et al., 2012). Interpretation of U^IV release in bicarbonate extractions, however, may be complicated by reversal of the electron transfer between U^IV and the oxidized reductant of U^VI (e.g., organic matter, electron shuttles, or Fe³⁺) (Ginder-Vogel et al., 2006, Stoliker et al., 2013), leading to U^IV oxidation and release of U^VI. Cerrato et al. (2013) studied the oxidation of a biogenic, non-uraninite phase at lower bicarbonate concentrations (100 mM) and concluded comparable oxidation rates for biogenic uraninite and non-uraninite U^IV species. A slight preference for oxidation of non-uraninite U^IV species relative to uraninite was suggested in experiments with mixtures of these two phases. Sharp et al. (2011) investigated the oxidative release of U from non-uraninite U^IV species in sediments that have been previously subjected to reducing conditions and U^VI-containing influents. The authors concluded that these U^IV species were less stable than biogenic uraninite under both anoxic and oxic conditions. Despite these initial results, the knowledge about the stability of non-uraninite U^IV phases appears to be far from a state where specific reactions can be included in reactive transport models.

Here, we investigate the rates and products of the oxidation of a non-uraninite, amorphous U^IV–phosphate phase by dissolved oxygen (DO), under environmentally relevant conditions. The results are compared to the oxidation of nanoparticulate uraninite under the same conditions. The goal was to obtain insight on the oxidation mechanisms and to estimate the relative stability of the two U^IV phases in controlled experiments. We examined the effect of Ca²⁺ and HPO₄²⁻ on U^IV oxidation; both are ions that are ubiquitous and are typically in excess of U in natural environments, thus potentially affecting uranium geochemistry. Elevated Ca²⁺ and PO₄ concentrations also create conditions that inhibit U^VI dissolution, allowing for control of the surface dissolution step in the overall oxidation process and providing information on the mechanism of oxidation. Unlike previous studies, which determined the rate of U^IV oxidation from the rate of oxidant consumption or from the rate of U^VI dissolution, here we determined the change in U^IV/U^VI content over time from X-ray absorption spectroscopy measurements. This approach provides a direct measurement of U^IV oxidation in the hydrated solids under the conditions of interest, as well as information on the reaction products and intermediates. In such a way, the U^IV oxidation reaction was studied without potential interfering factors such as rate-limiting dissolution of the oxidized products or solution conditions that may favor oxidized U dissolution but alter U speciation and reaction dynamics. In addition, potential artifacts from drying of the samples or from unaccounted oxidant consumption in parallel reactions not related to U^IV oxidation are eliminated.