Kinetics of Na- and K- uranyl arsenate dissolution

Abstract

We integrated aqueous chemistry analyses with geochemical modeling to determine the kinetics of the dissolution of Na and K uranyl arsenate solids (UAs_(s)) at acidic pH. Improving our understanding of how UAs_(s) dissolve is essential to predict transport of U and As, such as in acid mine drainage. At pH 2, Na_0.48H_0.52(UO₂)(AsO₄)(H₂O)_2.5(s) (NaUAs_(s)) and K_0.9H_0.1(UO₂)(AsO₄)(H₂O)_2.5(s) (KUAs_(s)) both dissolve with a rate constant of 3.2 × 10⁻⁷ mol m⁻² s⁻¹, which is faster than analogous uranyl phosphate solids. At pH 3, NaUAs_(s) (6.3 × 10⁻⁸ mol m⁻² s⁻¹) and KUAs_(s) (2.0 × 10⁻⁸ mol m⁻² s⁻¹) have smaller rate constants. Steady-state aqueous concentrations of U and As are similarly reached within the first several hours of reaction progress. This study provides dissolution rate constants for UAs_(s), which may be integrated into reactive transport models for risk assessment and remediation of U and As contaminated waters.

Graphical Abstract

Introduction

Uranium (U) and arsenic (As) are toxic elements that co-occur in geologic deposits, mining settings, and drinking water sources (Chou et al., 2007; Keith et al., 2013; Neiva et al., 2016; Das et al., 2018; Yadav et al., 2020; Corkhill et al., 2017). Elevated levels of U and As have been observed in communities located in the Northern Plains and southwestern United States (Hoover et al., 2017, 2018; Blake et al., 2015; Jones et al., 2020; Sobel et al., 2021); Native American communities are concerned about the release of these contaminants to their sources of water (Hoover et al., 2017; Blake et al., 2015). U and As are also commonly found co-occurring in acid mine drainage (AMD) (Majzlan et al., 2014; Kipp et al., 2009; Groudev et al., 2008). AMD is a common condition in sites affected by mining legacy, where metal sulfide minerals are oxidized resulting in formation of acidic sulfate-rich drainage that can mobilize toxic metals and metalloids (Akcil et al., 2006). The motivation of this study relates to the co-occurrence and predominance of uranyl and arsenate in mine waste sites exposed to ambient, oxic environments as stated in previous publications from our research group (Blake et al., 2015, 2017, 2019).

Only a few studies have measured the solubility of environmentally relevant uranyl arsenates solids (UAs_(s)) (Chernorukov and Karyakin, 1995; Zhiltsova et al., 1987; Chernorukov et al., 2009, 2012; Nipruk et al., 2011, 2020; Meza et al., 2023), which can impact the mobility of U and As in water systems (Gonzalez-Estrella et al, 2020; Tian et al., 2022; Blake et al., 2019). To our knowledge, the rates of UAs_(s) dissolution in water systems have not been directly measured. Understanding the kinetics of dissolution of UAs_(s) in aqueous solutions is essential to gain fundamental knowledge of U and As mobilization under environmentally relevant conditions.

Several uranyl arsenates (along with uranyl phosphates) occur in the meta-autunite group (Aⁿ⁺[(UO₂)(TO₄)] (H₂O) where Aⁿ⁺ is a mono-, di- or trivalent cation and T = P or As) (Frost et al., 2005). The thermodynamics of uranyl arsenates containing H⁺, Na⁺, Li⁺, K⁺, Rb⁺, Cs⁺, NH₄⁺, Sr²⁺, and Cu²⁺ have been studied using high-temperature calorimetry and solubility experiments (Nipruk et al., 2011; Dzik et al., 2018). A recent study by our group characterized the thermodynamics and solubility of UAs_(s) containing Na and K (Meza et al., 2023), but the kinetics of dissolution of these minerals in water is not well-characterized. Further understanding of the influence of chemical equilibrium and kinetics on the mobilization of U and As is necessary for improved interpretation and prediction of the fate and transport of these toxic elements in the environment.

The objective of this study is to determine the UAs_(s) dissolution kinetics influencing the aqueous mobilization of U and As in acidic waters. Dissolution experiments were performed using well-characterized Na uranyl arsenate Na_0.48H_0.52(UO₂)(AsO₄)(H₂O)_2.5(s) (hereafter referred to as NaUAs_(s)) and K uranyl arsenate K_0.9H_0.1(UO₂)(AsO₄)(H₂O)_2.5(s) (hereafter referred to as KUAs_(s)) at a pH of approximately 2 and 3 at room temperature. We determined dissolution kinetic rate constants for NaUAs_(s) and KUAs_(s) in acidic waters using static batch experiments and applying the TST rate law using the CrunchFlow code (Steefel et al., 2015). This study provides new insights into the kinetics of UAs_(s) dissolution at acidic pH conditions, which may be integrated into reactive transport models for improving risk assessment and developing remedial strategies.

Materials and methods

Synthesis and characterization of uranyl arsenate solids.

The same uranyl arsenate solids (NaUAs_(s) and KUAs_(s)) that were synthesized and characterized in our previous study (Meza et al., 2023) were utilized in this study as well.. In brief, NaUAs_(s) and KUAs_(s) were synthesized at room temperature using the diffusion method of Dzik et al. (2017) where 0.5M aqueous solutions of uranyl nitrate UO₂(NO₃)₂(H₂O)₆ and arsenic pentoxide (As₂O₅) were allowed to diffuse into a 0.05M barrier solution containing NaNO₃ or KNO₃. This method produced high-purity and well-crystallized material that was easily recovered. The NaUAs_(s) and KUAs_(s) solids were ground in a mortar and pestle and characterized using a Bruker D8 Advance DaVinci diffractometer with Bragg-Brentano geometry and CuKα radiation. As observed in Meza et al. (2023), the diffraction patterns of NaUAs_(s) and KUAs_(s) solids displayed sharp profiles with all peaks assignable to the NaUAs_(s) and KUAs_(s) target phases.

Elemental composition of NaUAs_(s) and KUAs_(s) were also identified using an Orbis micro-XRF (X-ray fluorescence) spectrometer, which can measure K, U, and As concentrations, but not Na concentrations due to its low mass. NaUAs_(s) and KUAs_(s) were analyzed before dissolution experiments, and KUAs_(s) was analyzed following dissolution experiments. Unfortunately, not enough NaUAs_(s) sample was collected and stored following dissolution experiments to analyze using micro-XRF. KUAs_(s) samples prior to and following dissolution experiments contained 36%±2% (1 standard deviation) As, 32%±3% U, and 32%±5% K. NaUAs_(s) samples also had a U to As ratio of nearly one.

Dissolution experiments.

To measure the rate of dissolution before the system reached equilibrium, we avoided complete dissolution of UAs_(s) by using an excess amount of uranyl arsenate solids. While dissolution experiments are typically performed as flow-through experiments to avoid the solution chemistry from evolving, this initial study of UAs_(s) dissolution kinetics utilized batch experiments to prevent radioactive contamination as U dissolved. Geochemical modeling accounted for the approach to chemical equilibrium and possible back-reactions (see below). All dissolution experiments were conducted in triplicate in batch experiments in three Teflon Nalgene^® bottles at 25°C. Initial experiments added ~100 mg of uranyl arsenate solid to ~71mL of 18 MΩ H₂O, which would result in ~2.9 mmol L⁻¹ of U and As if the solid completely dissolved. Later experiments used the same ratio of solid to water, but were made more efficient by adding ~40 mg of UAs_(s) to ~29 mL of water. The batch reactors were closed and agitated at room temperature (20 °C) at 60 rpm in an analog rotisserie tube rotator (Scologex MX-RL-E, Rocky Hill, CT, US), which provided a gentle mixing of the solution with the UAs_(s). The solutions should have remained oxic throughout the experiment since the reactors were opened and closed in an air atmosphere and no reductants were present. Once equilibrium was achieved in the experiments, U and As concentrations remained at stable concentrations, which suggests no redox reactions occurred. As the UAs_(s) dissolved, aliquots were extracted from each triplicate reactor at selected times (0 min, 15 min, 30 min, 1h, 2h, 4 h, 6 h, 8 h, 12 h, 1 d, 8 d, 15 d, 22 d, 31 d). Analyte concentrations were corrected for the loss of solution during aliquot extraction by multiplying the measured analyte concentrations by the initial volume of solution before extraction and dividing it by the final volume after extraction. In Meza et al. (2023), we determined that chemical equilibrium was reached within our experimental period of 31 days. Therefore, we ran our experiments for 31 days as well to achieve equilibrium conditions by the end of the experimental period. Aliquots were filtered through 0.1 μm MilliporeSigma Millex filters and then were diluted with 2% HNO₃ prior to ICP-OES (inductively coupled plasma optical emission spectroscopic) analysis. We conducted a blank experiment at pH 2 to determine the background concentrations of U, As, Na, and K during the experimental and analysis process. Aliquots were taken at 0h, 1h, 4h, 1d, and 8d to test if these metals were leaching into solution over time.

Dissolution experiments were performed at a pH of approximately 2 and 3. For experiments conducted at pH 2, we added a small quantity of concentrated HNO₃ to reach the desired pH. The pH was measured when sampling occurred using a Thermo Orion Ross micro electrode calibrated daily with 3 National Institute of Standards and Technology (NIST) standards (pH 4, 7, and 10) (Blake et al., 2019; Gonzalez-Estrella et al., 2020; Meza, et al., 2023). U and As in solution naturally tend to buffer at a pH of ~2 and ~3. Arsenate deprotonates from H₃AsO_4(aq) to H₂AsO₄⁻ at a pH of approximately 2.2, while the uranyl arsenate complex UO₂H₂AsO₄⁺ deprotonates to UO₂HAsO_4(aq) at roughly pH 3.1 (Nipruk et al., 2011; Nipruk et al., 2020). We also avoided the complicating influence of carbonate on U speciation at these acidic pH’s.

Solution analyses.

The aqueous concentrations of U, As, Na, and K from sampled aliquots were measured using a PerkinElmer Optima 5300DV ICP-OES. This instrument was calibrated with Fischer Scientific Standards with a five-point calibration curve, and quality assurance and control measures were taken to ensure quality results. The ICP-OES has an instrument detection limit of 0.33 μM, 1.26 μM, 1.28 μM, and 3.00 μM for As, U, K, and Na, respectively. The detection limit of our analysis method was ten times that of the instrument detection limit, or 3.3 μM, 12.6 μM, 12.8 μM, and 30.0 μM for As, U, K, and Na, respectively. Our dissolution experiment of NaUAs_(s) and our blank experiment both at pH 2 revealed limitations in analyzing Na and K. The measured concentrations of these alkali metals were highly variable (Tables S1 and S2) in these experiments, which was likely due to difficulties in the detection of Na and K at analytical stages, which causes high signal variability. The aqueous concentrations of U, As, Na, and K for all dissolution experiments are illustrated in the SI (Table S1 and Figures S1, S2, S3, and S4). We could not exclusively dedicate clean equipment for the levels of detection required for Na and K for this study, so unexpected and unrealistic concentrations were observed. Therefore, we decided to base our quantitative analyses for this study only on U and As concentrations as a function of pH.

Estimating uranyl arsenate dissolution rates.

Given the increase in U and As concentrations, we estimated the dissolution rates of Na- and K- uranyl arsenates using the following:

$$R_{UA{s}_{\left(s\right)}}=\frac{dC}{dt}*\frac{V}{\mathit{A}*m}$$ (1)

where R (mol m⁻² s⁻¹) is the reaction rate per unit surface area, dC/dt is the change in concentration per second (mol/L/s), V is the volume of solution (L), A is the specific surface area (m² g⁻¹) of the uranyl arsenate solid, and m is the mass of solid (g) added to solution. The ratio of the volume of solution to the mass of uranyl solid added (V/m) in each experiment was 0.72 L/g. The specific surface areas of Na- (1.46 m² g⁻¹) and K- (0.77 m² g⁻¹) uranyl arsenates were measured by gas absorption analysis in a previous study (Meza et al., 2023). The change in concentration over time was calculated using least squares linear regression of both the U and As concentrations over only the first two sampling times (15 minutes and 30 minutes) in order to primarily capture the kinetics of the system and minimize the inherent slowdown of dissolution rates as the systems neared equilibrium. However, these calculated rates likely slightly underestimated the dissolution rates due to some equilibrium effects. The two rates obtained for both As and U dissolution were similar in all experiments studied, so these rates were averaged for our initial estimates of uranyl arsenate dissolution rates, which we then utilized in our kinetic modeling.

Uranyl arsenate dissolution kinetic modeling.

We considered a reaction in a batch reactor,

$$R_{UA{s}_{\left(s\right)}}=k_{UA{s}_{\left(s\right)}}*a_{H^{+}}^n*A*\left(1-\frac{Q}{{\text{K}}_{\text{eq}}}\right)$$ (2)

where the reaction rate was given by:

where R (mol g⁻¹ s⁻¹) is the reaction rate, $a_{H^{+}}^n$ is the activity of H⁺, reflecting the reaction dependence on pH, k_UAs(s) (mol m⁻² s⁻¹) is the rate constant per unit surface area of UAs_(s), A is the specific surface area of the solid (m² g⁻¹), Q (--) is the activity product, and K_eq (--) is the equilibrium constant. In this paper, we will at times compare mineral dissolution rates as slower or faster reaction rates, but we are technically comparing rate constants. This rate law follows from transition state theory (Lasaga, 1981; Lasaga, 1984; Aagaard and Helgeson, 1982), where the rate is influenced by how far the system is from equilibrium. If Q = K_eq, the system is in equilibrium, if Q > K_eq, precipitation occurs, and if Q < K_eq, dissolution occurs. In our experiments, the initial aqueous concentrations of U, As, Na, and K are 0 mM, so Q << K_eq and dissolution occurs, which is controlled by the surface area and rate constant of the UAs_(s). As the experiments progress, U and As concentrations increase, so the systems become more controlled by equilibrium. The chemical data collected at later time points in all experiments of this study are consistent with model calculations of chemical equilibrium. Therefore, by the end of the experiments, the systems are effectively in equilibrium with respect to the solubility of U and As solids. Since Na and K concentrations were not well-constrained, our modeling of UAs_(s) were based exclusively on U and As concentrations as a function of pH. This limitation affects the value of the activity product and negatively impacts our ability to calculate equilibrium constants highly accurately. The accurate determination of equilibrium constants is beyond the scope of this study, given that here we aimed to estimate the dissolution kinetics of UAs_(s). Na and K concentrations did not significantly affect estimated kinetic rate constants (Supporting Information).

Modeling of UAs_(s) dissolution was performed using the open-source computer code CrunchFlow (Steefel et al., 2015), which was used for simple zero-dimensional simulations of the congruent dissolution reactions for UAs_(s) as well as simultaneous reactions involving secondary minerals. The primary species used during modeling were UO₂²⁺, AsO₄³⁻, Na⁺, K⁺, H⁺, HCO₃⁻, and NO₃⁻. These species formed secondary species, including uranyl arsenates (Rutsch et al., 1999), through aqueous species equilibration (Table S4). Initially, the model contained a trivial amount of UO₂²⁺, AsO₄³⁻, Na⁺, and K⁺. The initial concentration of H⁺ was controlled by the initial pH of each experiment (Table 1), and NO₃⁻ was used to balance the positive charge of H⁺ in the acidic system. The low concentration of HCO₃⁻ in each simulation was controlled by CO₂ in air (400 ppm) equilibrating with the aqueous solutions. In addition, the model domain contained a volume fraction of either 0.000364 or 0.000416 of NaUAs_(s) or KUAs_(s), respectively (Table 1). The surface area of these minerals were updated by CrunchFlow during dissolution to account for particle shrinking. The aqueous concentrations of UO₂²⁺, AsO₄³⁻, Na⁺, and K⁺ increased as UAs_(s) dissolved. In CrunchFlow, reaction rate constants are input as the logarithm of the reaction rate. The reaction rates calculated by least squares regression were used for the CrunchFlow reaction rate constants, but they were adjusted in a positive direction to the nearest second digit of the logarithm to account for the underestimation of the regression method. For example, if the least squares rate was 2.9 × 10⁻⁷ mol m⁻² s⁻¹, then the logarithm was −6.54, and we used a logarithm of −6.5 (or rate of 3.16 × 10⁻⁷ in the model (Table 1). During the NaUAs_(s) experiment at pH 3, trögerite was found to precipitate. The reaction rate constant for trögerite in this experiment was selected to be slightly slower than that of NaUAs_(s) since NaUAs_(s) first dissolved before the effects of trögerite were observed (see below).

Table 1.

Modeled reaction rate constants obtained under various initial conditions.

Initial pH	Equilibrium constant (log Keq)	Uranyl arsenate reaction	Initial volume fraction	Specific surface area (m2/g)	kUAs(s) (mol/m2/s)
1.9	−23.51	Na0.48H0.52(UO2)(AsO4)(H2O)2.5 → 0.48Na++0.52H++UO22++AsO43−+2.5H2O	0.000364	1.46	3.16 × 10−7
3.0	−23.51	Na0.48H0.52(UO2)(AsO4)(H2O)2.5 → 0.48Na++0.52H++UO22++AsO43− +2.5H2O	0.000364	1.46	6.31 × 10−8
	−45.63	(UO2)3(AsO4)2(H2O)12a → 3UO22++2AsO43−+12H2O	0.000001	1.46b	3.16 × 10−8
2.0	−23.87	K0.9H0.1(UO2)(AsO4)(H2O)2.5 → 0.9K++0.1H++UO22++AsO43−+2.5H2O	0.000416	0.77	3.16 × 10−7
3.2	−23.87	K0.9H0.1(UO2)(AsO4)(H2O)2.5 → 0.9K++0.1H++UO22++AsO43−+2.5H2O	0.000416	0.77	2.00 × 10−8

^a:log K_eq as reported by Nipruk et al. (2011)

^b:specific surface area of trögerite not measured by gas absorption analysis, but assumed to be similar to NaUAs_(s)

The equilibrium constants used for the reaction rates of these minerals (Table 1) have been previously measured by our group (Meza et al., 2023). The stoichiometry of NaUAs_(s) [Na_0.48H_0.52(UO₂)(AsO₄)(H₂O)_2.5(s)] and KUAs_(s) [K_0.9H_0.1(UO₂)(AsO₄)(H₂O)_2.5(s)], which include H⁺ in addition to their respective alkali metals, were previously determined (Table 1) (Meza et al., 2023). The specific surface area of NaUAs_(s) (1.46 m² g⁻¹) and KUAs_(s) (0.77 m² g⁻¹) measured by gas absorption analysis in our previous study were used for our kinetic modeling (Meza et al., 2023). The known mass, specific surface area, molecular weight, and molar volume of UAs_(s) added to experimental solutions were used to calculate the volume fraction of the UAs_(s) used in the model simulations (Table 1). Trögerite [(UO₂)₃(AsO₄)₂(H₂O)_12(s)], which formed during the NaUAs_(s) experiment at pH 3 (see below), did not have a measured specific surface area, so we utilized the same specific surface area as measured for NaUAs_(s). A minimal amount of trögerite was added to the simulation to allow the trögerite to nucleate and begin precipitating. While these modeling constraints may be contrived, the rate constant of trögerite precipitation (which is not a focus of this study) may be changed to obtain the same precipitation rate and reach an equilibrium state within several days.

Effect of initial Na and K concentrations on dissolution kinetic rate constants.

Due to limitations in analyzing Na and K concentrations, we utilized U and As concentrations as a function of pH to determine dissolution kinetic rate constants. However, the concentration of Na and K in these experiments affects the activity product (Q) of the aqueous solution and could impact UAs_(s) dissolution rates. For our primary models, we assumed that the initial Na or K concentrations in these experiments were 0 mM. However, our measured Na and K concentration data, which was unreliable, had non-zero concentrations of Na and K in initial or early samples. Therefore, we performed dissolution kinetic modeling of our four experiments with the worst case scenarios of Na and K concentrations to test the impact on dissolution kinetics. We used initial Na concentrations of 3 mM and 0.5 mM for the NaUAs_(s) experiments at pH 1.9 and pH 3.0, respectively, and initial concentrations of 1 mM and 0.1 mM for the KUAs_(s) experiments at pH 2.0 and pH 3.2, respectively. No other parameters were changed in the simulations. Overall, we saw a decrease in the equilibrium concentrations of U and As to compensate for the higher alkali levels (Figures S5–S8). However, we did not observe a significant change in the kinetics of these models.

Results and discussion

Determination of aqueous dissolution rates for NaUAs solids.

When NaUAs_(s) dissolved at pH 2, aqueous concentrations of As and U increased over the first six hours of the experiment before these concentrations mostly plateaued at an equilibrium concentration of 1.7 mM (Figure 1A). The pH of the solution remained relatively stable, starting at 1.89 and stabilizing at 2.03 (Figure 1B). The least squares estimate of NaUAs_(s) dissolution per unit surface area was 2.5 × 10⁻⁷ mol m⁻² s⁻¹. We utilized a slightly higher reaction rate constant for NaUAs_(s) dissolution of 3.16 × 10⁻⁷ mol m⁻² s⁻¹, which agreed well with both the U and As concentrations as well as the pH.

Figure 1.

A) NaUAs dissolution experiments at pH 1.9 vs modeling with Crunchflow (continuous lines) presenting concentrations of U (open circles), and As (open triangles). B) pH of the NaUAs dissolution experiments vs modeling with Crunchflow in time.

During the dissolution of NaUAs_(s) at pH 3, aqueous concentrations of As and U initially increased similarly to the experiment at pH 2 before differences emerged. Over the first hour of the experiment, As and U concentrations increased at the same rate (Figures 2A, 2B). These concentrations were lower than those observed in the pH 2 experiment due to the lower solubility of UAs_(s) at higher pH. After two hours of dissolution, As and U concentrations clearly diverged with higher As concentrations as observed by the lower U/As ratio (Figures 2A, 2B). Following eight days of reaction, equilibrium was reached with U concentrations at ~0.13 mM and As concentrations of ~0.73 mM.

Figure 2.

A) NaUAs dissolution experiments at pH 3.0 vs modeling with Crunchflow (continuous lines) presenting concentrations of U (open circles), and As (open triangles). B) U/As ratio of the NaUAs experiments vs modeling with Crunchflow in time. C) pH of the NaUAs dissolution experiments vs modeling.

The dissolution of NaUAs_(s) at pH 3 was clearly impacted by the formation of a secondary mineral with a greater mole fraction of U than As. In nearly identical experiments that measured the solubility of NaUAs_(s) at pH 3, we observed the same behavior where As concentrations increased significantly more than U (Meza et al., 2023). X-ray diffraction data demonstrated that this behavior was due to the formation of trögerite [(UO₂)₃(AsO₄)₂(H₂O)_12(s)] (Meza et al., 2023), which becomes more insoluble at less acidic pH (Nipruk et al., 2020). Another study investigating the stability of uranyl arsenates independently confirmed the formation of trögerite as a secondary phase at a pH above 2 (Nipruk et al., 2020). During the dissolution of NaUAs_(s) at pH 3, U and As concentrations increase until they reach the solubility limit of trögerite, at which point trögerite forms and controls U and As concentrations. NaUAs_(s) dissolved at a rate of 5.4 × 10⁻⁸ mol m⁻² s⁻¹ according to least squares regression. Our geochemical model, which includes reactions with both NaUAs_(s) and trögerite, agrees well with the measured U and As concentrations and pH (Figures 2A, 2B). We utilized a reaction rate constant of 6.31 × 10⁻⁸ mol m⁻² s⁻¹ for the dissolution of NaUAs_(s), which was approximately one quarter of the rate of reaction at pH 2. Trögerite formed at a slower rate, which led to the slow increase in the concentration of As over the first several days of the experiment.

Determination of aqueous dissolution rates for KUAs solids.

The dissolution behavior of KUAs_(s) at pH 2 was quite similar to that of NaUAs_(s) at pH 2. Aqueous concentrations of As and U increased over the first eight hours of the experiment before these concentrations reached equilibrium at 2.1 and 2.2 mM, respectively (Figure 3A). The pH of this experiment increased slightly from 1.8 at one hour to 1.9 at the conclusion of the experiment (Figure 3B). KUAs_(s) dissolved at a rate of 2.9 × 10⁻⁷ mol m⁻² s⁻¹ according to least squares regression. In our model, which agreed well with concentration and pH data, we used the same reaction rate constant for KUAs_(s) dissolution (3.16 × 10⁻⁷ mol m⁻² s⁻¹) as we did for NaUAs_(s) dissolution at pH 2.

Figure 3.

A) KUAs dissolution experiments at pH 2.0 vs modeling with Crunchflow (continuous lines) presenting concentrations of U (open circles), and As (open triangles). B) pH of the KUAs dissolution experiments vs modeling with Crunchflow in time.

At pH 3, KUAs_(s) dissolved slightly to reach steady-state U and As concentrations of only 0.07 and 0.08 mM, respectively (Figure 4A). According to least squares, KUAs_(s) dissolved at a rate of 1.9 × 10⁻⁸ mol m⁻² s⁻¹. The concentrations of As and U were slightly lower than those measured in similar previous experiments (Meza et al., 2023), which is why the model, which uses an equilibrium constant based off of these experiments, slightly overestimates As and U concentrations. This small discrepancy is not the focus of this kinetics study, but may be due to high K⁺ concentrations. However, challenges in the analytical detection of K⁺ using ICP-OES did not allow further investigation.

Figure 4.

A) KUAs dissolution experiments at pH 3.2 vs modeling with Crunchflow (continuous lines) presenting concentrations of U (open circles), and As (open triangles). B) pH of the KUAs dissolution experiments vs modeling with Crunchflow in time.

The U and As concentrations fluctuated between two hours and one day of experimental progress with slightly lower U concentrations and higher As concentrations than expected (Figure 4A). It is possible that trögerite formed at this time, which then redissolved as the experiment proceeded. This pattern may be explained by the pH of the experiment, which also varied slightly, increasing from 3.2 to 3.35 during the first day and then decreasing to 3.15 over the remaining time (Figure 4B). To model this experiment, we used a reaction rate constant of 2.00 × 10⁻⁸ mol m⁻² s⁻¹ for KUAs_(s) dissolution, which was 6% of the rate of KUAs_(s) dissolution at pH 2. However, with the low U and As concentrations observed in this experiment, the system neared equilibrium within two hours versus six hours at pH 2.

Insights about dissolution kinetics of uranyl arsenate solids.

The dissolution rates of NaUAs_(s) and KUAs_(s) were quite similar, particularly at pH 2. Our fitted rate constants were identical at 3.2 × 10⁻⁷ mol m⁻² s⁻¹ at pH 2. We observed a clear relationship between dissolution rate and pH with slower rates for both NaUAs_(s) (rate constant of 6.3 × 10⁻⁸ mol m⁻² s⁻¹) and KUAs_(s) (rate constant of 2.0 × 10⁻⁸ mol m⁻² s⁻¹) at pH 3. The change in dissolution rates versus pH is consistent for the uranyl arsenates in this study and those observed in analogous uranyl phosphate minerals. For example, Na meta-autunite dissolved at a rate of 1.6 × 10⁻¹⁰ mol m⁻² s⁻¹ at pH 2.0 and 4.0 × 10⁻¹² mol m⁻² s⁻¹ at pH 3.0. Ca meta-autunite similarly dissolved at a rate of 1.3 × 10⁻¹⁰ mol m⁻² s⁻¹ at pH 2.0 and 3.0 × 10⁻¹² mol m⁻² s⁻¹ at pH 3.0 (Wellman et al., 2007). Our results similarly observed roughly an order-of-magnitude decrease in dissolution rate with an increase of 1 pH unit. However, uranyl arsenate solids dissolved at a rate more than three orders of magnitude faster than uranyl phosphate minerals. The presence of arsenate within the UAs_(s) crystal lattice may promote rapid dissolution. Different types of arsenate containing minerals including the Co-bearing mineral erythrite (Co₃(AsO₄)₂(H₂O)_8(s)) and the Ni-bearing mineral annabergite (Ni₃(AsO₄)₂(H₂O)_8(s)) also dissolve rapidly at pH 2 with As concentrations reaching mM levels within the first hour of reaction (Zhu et al., 2013).

Environmental implications.

The experiments in this study were conducted in acidic and oxic conditions. These experiments lie in the lower pH range of acid mine drainage, which can result in waters with a pH between 2 and 6 (Bigham and Cravotta, 2016). The experiments in this study contained only the solutes dissolving from NaUAs_(s) or KUAs_(s) and/or a minor contribution of nitrate from nitric acid added in pH 2 experiments. For this initial study, our experiments attempted to measure the kinetics of UAs_(s) dissolution in simplified acidic conditions. NaUAs_(s) and KUAs_(s) dissolved rapidly compared to uranyl phosphates. Aqueous U and As reached steady-state concentrations within several hours of the solids reacting with acidic waters. Uranyl arsenate minerals in mining settings should rapidly dissolve when in contact with AMD waters. If acidic waters are in contact with uranyl arsenate minerals for an extended period of time, U and As concentrations should reach the solubility limit for UAs_(s). Our previous study observed that UAs_(s) also precipitates quickly, so these minerals could limit U and As concentrations in AMD waters by removing U and As above UAs_(s) solubility limits. The solubility constants and reaction rate constants obtained for NaUAs_(s) and KUAs_(s) may be integrated into reactive transport modeling for improved predictions of U and As fate and transport in contaminated settings. The conditions tested in this study represent an initial step toward understanding the kinetics of uranyl arsenate dissolution. Additional studies should investigate experimental conditions more relevant to AMD conditions involving more side-reactions triggered by high concentrations of sulfate, iron, and other co-occurring metals.

UAs_(s) is less soluble at pH 3 than pH 2, so uranyl arsenate solids are likely less soluble at circumneutral pH. For a wide range of uranyl-bearing minerals including autunite, compreignacite, uranophane, and soddyite, mineral dissolution rates are controlled by pH and total carbonate concentrations (Reinoso-Maset et al., 2020; Reinoso-Maset et al., 2017; Gudavalli et al., 2018; Perez et al., 1996; Perez et al., 2000). Dissolution rates of uranyl minerals appear to be slowest at slightly acidic to neutral pH (5–7) (Reinoso-Maset et al., 2020), and increase at more alkaline pH. At higher total carbonate concentrations, U can form strong aqueous uranyl-carbonate complexes that enhance dissolution rates (Reinoso-Maset et al., 2020). In neutral waters, UAs_(s) may further control U and As concentrations. However, circumneutral waters typically contain bicarbonate as well, which would increase U concentrations by forming strong uranyl-carbonate complexes.

The focus of the present study was on oxic conditions given that uranyl arsenate minerals in mine waste are exposed to ambient atmospheric conditions. However, in addition to carbonate concentrations and pH conditions, shifts in redox conditions highly control uranyl and arsenate concentrations in contaminated waters (Coyte and Vengosh, 2020). Stream sediments and aquifers often contain organic matter and other reductants, which can induce microbial or abiotic reduction of U and As (Gorny et al., 2015; Kumar et al., 2020; Janot et al., 2016). Uranyl reduction to U(IV) typically leads to precipitation or strong adsorption of U (Blake et al., 2015). While reduction of arsenate to arsenite decreases arsenate concentrations, this reaction can promote desorption of arsenite from iron minerals (Kumar et al., 2020; Coyte and Vengosh, 2020). If waters become undersaturated in reference to UAs_(s) as uranyl and arsenate are reduced, UAs_(s) could act as a source of U and As release. Redox shifts can also promote numerous other competing reactions, such as reductive dissolution of iron minerals that may release U and As (Kumar et al., 2020). In this case, uranyl and arsenate may become oversaturated and UAs_(s) precipitation could occur. Further study of UAs_(s) dissolution and precipitation at variable redox conditions, pH conditions, and total carbonate concentrations would be valuable to determine how uranyl arsenate minerals influence the concentration of U and As in typical drinking waters.

Conclusions

This study observed that uranyl arsenate minerals rapidly dissolve under acidic conditions. At pH 2, NaUAs_(s) and KUAs_(s) have a dissolution rate constant of 3.2 × 10⁻⁷ mol m⁻² s⁻¹, which is approximately three orders of magnitude faster than uranyl phosphate analog minerals. At pH 3, the dissolution rate of NaUAs_(s) was lower (rate constant of 6.3 × 10⁻⁸ mol m⁻² s⁻¹). As equilibrium was approached, it was controlled by the secondary precipitation of trögerite. KUAs_(s) also dissolved slower at pH 3 (rate constant of 2.0 × 10⁻⁸ mol m⁻² s⁻¹) than pH 2, which is consistent with other uranyl-bearing minerals. The presence of uranyl arsenate minerals in AMD settings may promote U and As concentrations to reach the solubility limit for UAs_(s). The reaction rate constants obtained in this study may be integrated into reactive transport modeling for improved understanding of the fate and transport of U and As in contaminated waters.

Highlights:

• Uranyl arsenates dissolve significantly faster than analogous uranyl phosphates.

• K- and Na- uranyl arsenates dissolve slower at pH 3 than pH 2.

• Reactive transport models can use these rates to predict [U] and [As].