Uranophane dissolution and growth in CaCl2–SiO2(aq) test solutions

Available online at www.sciencedirect.com

Geochimica et Cosmochimica Acta 72 (2008) 4508–4520 www.elsevier.com/locate/gca

Uranophane dissolution and growth in CaCl2–SiO2(aq) test solutions James D. Prikryl * Center for Nuclear Waste Regulatory Analyses Southwest Research InstituteÒ, 6220 Culebra Road, San Antonio, TX 78238, USA Received 4 March 2008; accepted in revised form 26 June 2008; available online 8 July 2008

Abstract The dissolution and growth of uranophane [Ca(UO2)2(SiO3OH)25H2O] have been examined in Ca- and Si-rich test solutions at low temperatures (20.5 ± 2.0 °C) and near-neutral pH (6.0). Uranium-bearing experimental solutions undersaturated and supersaturated with uranophane were prepared in matrices of 102 M CaCl2 and 103 M SiO2(aq). The experimental solutions were reacted with synthetic uranophane and analyzed periodically over 10 weeks. Interpretation of the aqueous solution data permitted extraction of a solubility constant for the uranophane dissolution reaction ðlog K osp ¼ 11:01  0:09Þ and standard state Gibbs free energy of formation for uranophane (DGof ¼ 6196:1  0:5 kJ mol1). Ó 2008 Elsevier Ltd. All rights reserved.

1. INTRODUCTION Yucca Mountain, Nevada, is being evaluated as a potential geologic repository site for disposal of high-level nuclear waste (HLW). Most of the HLW to be housed at Yucca Mountain is spent nuclear fuel (SNF) derived from nuclear reactors. SNF is 95% UO2 with the remainder composed of highly radioactive fission and activation products and actinides (Oversby, 1994; Johnson and Shoesmith, 1988). Long-term leaching studies of unirradiated UO2 and SNF designed to simulate conditions at a potential Yucca Mountain repository indicate that the calcium uranyl silicate uranophane [Ca(UO2)2(SiO3OH)25H2O], along with the alkali uranyl silicates boltwoodite [K(UO2)(SiO3OH)1.5H2O] and Na-boltwoodite [Na(UO2)(SiO3OH)1.5H2O], are the dominant end products of spent fuel alteration (Wronkiewicz et al., 1992, 1996; Finch et al., 1999). Uranophane is also the dominant end product of the oxidative weathering of uraninite (UO2+x, an analog to UO2 in SNF) hosted in silicic volcanic rocks at the Nopal I U deposit in Chihuahua, Mexico—a natural analog to the potential HLW repository at Yucca Mountain (Pearcy et al., 1994). *

Fax: +1 210 522 5184. E-mail address: [email protected]

0016-7037/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.gca.2008.06.022

In a HLW geologic repository, incorporation of radionuclides in secondary uranyl phases formed during SNF alteration is a mechanism for potentially limiting released radionuclide concentrations and, thus, doses from groundwater exposure. Recent performance assessment calculations for the planned Yucca Mountain repository indicate that groundwater pathway dose is dominated by Np-237 at times beyond 15,000–60,000 years (DOE, 2002; Mohanty et al., 2002). This result is due to a combination of Np-237 radiological and geochemical characteristics including its abundance in SNF (400–600 ppm), long half-life (2.14  106 years), and potential high mobility, as NpO2 þ , in chemically oxidizing groundwater. Crystal chemical considerations and experimental data on the synthesis of uranyl compounds from solutions containing Np5+ indicate that Np may be incorporated in the structure of uranophane (Burns et al., 1997, 2004; Douglas et al., 2005; Klingensmith and Burns, 2007; Forbes and Burns, 2008). Dating of uranophane from the Nopal I U deposit has yielded a radiometric determined age (U–Pb isochron method) of 3.2–3.4 Ma (Pickett and Murphy, 1997). U-series disequilibrium measured on a single uranophane from the Nopal I U deposit yielded an age of 320 ± 30 ka (Pickett and Leslie, 2005). Due to its long-term chemical stability and potential to incorporate Np, uranophane may limit

Uranophane dissolution and growth

the concentration of Np in solution during SNF degradation and, therefore, influence the long-term performance (i.e., 104 to 106 years) of the planned Yucca Mountain repository. Despite the potential significance of uranophane at Yucca Mountain, the thermodynamic parameters required for predictive modeling and interpretations of chemical interactions among uranophane and groundwater are poorly known. Previous efforts to measure uranophane solubility have produced inconsistent results. Nguyen et al. (1992) and Perez et al. (2000) conducted dissolution experiments using synthesized uranophane and reported solubility products ðlog K osp Þ for uranophane of 9.4 ± 0.5 and 11.7 ± 0.6, respectively. Casas et al. (1994) reported a log K osp for uranophane of 7.8 ± 0.8 derived from dissolution experiments using natural uranophane samples. Natural uranophane samples tend to be finely crystalline and intimately intergrown with other uranyl phases and mineral impurities, such as iron oxides/hydroxides. The presence of impurities results in intrinsic uncertainties in uranophane composition, which affects the interpretation of aqueous solubility data. The complexities in evaluating experimental results of dissolution studies carried out using natural uranophane samples are discussed by Casas et al. (1994). In this study, experiments were conducted to investigate the dissolution and potential growth of uranophane in chemically oxidizing, Ca- and Si-rich solutions predicted to bracket uranophane solubility. This study intends to add to previous work on uranophane solubility, particularly considering the disparity between published solubility constants for uranophane. This study is distinguished from previous studies by attempting to approach equilibrium from both below and above saturation. Based on preliminary data on uranophane dissolution, Prikryl and Murphy (2004) reported provisional thermodynamic data for uranophane dissolution [i.e., reaction quotients (log Qs) ranging from 10.54 to 11.06]. In this paper, the data on uranophane dissolution reported by Prikryl and Murphy (2004) are supplemented with additional data on uranophane dissolution and evaluated together with data on uranophane growth. To avoid uncertainties in using natural uranophane samples, the uranophane solubility experiments in this study were performed using synthesized uranophane. Evaluation of aqueous solution chemistry data from the uranophane solubility experiments permits extraction of a solubility constant and standard state Gibbs free energy of formation for uranophane. 2. EXPERIMENTAL MATERIALS AND METHODS 2.1. Uranophane synthesis Uranophane was synthesized based on the method of Vochten et al. (1997). The method entails initial synthesis of boltwoodite followed by conversion to uranophane by exchange of Ca2+ for K+ in the crystalline structure of the uranyl silicate. To synthesize boltwoodite, 9.28 g of uranyl nitrate hexahydrate [UO2(NO3)26H2O] and 1.33 g of KCl were dissolved in 200 mL of deionized (DI) water. The resultant

4509

solution was adjusted with 1 M KOH to a pH of 11.5. The solution was transferred into the teflon liner of a 2-L Parr pressure reactor (Parr Model 4522, Parr Instruments, Moline, IL). Approximately 32 g of ultrasonically cleaned natural rock crystal quartz fragments (Wards Geology, Rochester, NY) were added to the mixture in the teflon liner. The quartz fragments had widths or diameters >0.5 cm (prepared by breaking pieces of rock crystal quartz in an agate mortar and pestle). The reaction vessel was heated to 185 °C, and the contents were allowed to react for 4 days. After cooling to room temperature, the resultant mixture was filtered through a Spectra/MeshÒ fluorocarbon filter (Spectrum, Houston, TX) having 210-lm mesh openings to remove the quartz fragments. The resultant solid consisted of pale yellow acicular crystals and a whitecolored, gel-like precipitate. This precipitate was separated from the yellow crystals by washing with 5% acetic acid, centrifugation, and supernatant decantation. This process was repeated three times to adequately dissolve and remove the gel-like precipitate. The mixture was then vacuum filtered and washed several times with DI water to remove the acetic acid. The remaining solid consisted of fine, pale-yellow needles of boltwoodite, as confirmed by chemical analysis and X-ray diffraction (XRD). To obtain uranophane, approximately 1-g fractions of the synthesized boltwoodite were placed in 100-ml teflon vessels with 50 ml of 2 M CaCl2. The vessels were sealed and placed in a 125 °C oven for 15 days. Every 3 days during the 15-day period, the vessels were removed, the supernatant was decanted, and a new 50-ml aliquot of 2 M CaCl2 was added to the vessels. After 15 days, the resultant mixtures were vacuum filtered and washed several times with DI water. The remaining solids consisted of fine (generally <50 lm in length) yellow acicular crystals (Fig. 1A). XRD analysis confirmed the acicular crystals to be uranophane. XRD did not detect the presence of residual boltwoodite in the synthesized solids. An XRD pattern for the synthetic uranophane is shown in Fig. 1B, together with a reference pattern for natural uranophane collected from a mine site, Van Horn, TX, USA (Blanchard, 1987). The synthetic uranophane XRD pattern was obtained using an automated (RADIX) Siemens D–500 X-ray diffractometer (CuKa radiation, Ni filter, 40 kV, 37 mA; scan 5–40° 2-theta at 0.02° step; count time 1.0 s). Correspondence between the sample and reference patterns is good. Two small peaks (at 26.9° and 28.2° 2-Theta) appear in the pattern for the synthesized sample, but are absent in the reference pattern. However, these peaks are reported in a reference pattern for natural uranophane collected from the Ruggles pegmatite, Grafton Center, NH, USA (Frondel et al., 1956) and appear in uranophane synthesized by Nguyen et al. (1992). The magnitude and sharpness of the peaks in the XRD pattern indicate that the synthesized uranophane is well crystallized. Chemical analyses of the synthesized uranophane were performed using a whole rock procedure that entails complete dissolution of the sample in 0.1 M HCl followed by inductively coupled plasma-atomic emission spectrometry (ICP-AES) for major cations (e.g., Ca, K, and Si) and

4510

J.D. Prikryl / Geochimica et Cosmochimica Acta 72 (2008) 4508–4520

Fig. 1. (A) Scanning electron photomicrograph of uranophane synthesized for use in the solubility experiments. (B) X-ray diffraction (XRD) pattern of the synthesized uranophane compared to a reference pattern for natural uranophane from a mine stie, Van Horn, TX, USA (Blanchard, 1987).

ICP mass spectrometry (ICP-MS) for uranium. Using measured analytical contents, stoichiometric coefficients for Ca:Si:U in the synthesized uranophane were calculated to be 0.95(±0.01):1.99(±0.01):1.99(±0.07). When compared to the coefficients for Ca:Si:U expected for ideal uranophane (1:2:2, respectively), the synthesized uranophane is slightly deficient in Ca. 2.2. Dissolution and precipitation experiments Solubility experiments were designed to approach uranophane equilibrium in both undersaturated and supersaturated solutions. Estimation of the standard state Gibbs free energy of formation of uranophane by a theoretical method of prediction (Chen et al., 1999), together with data for aqueous species from the Data0.com.R2 database (Wolery, 1992), was used to estimate the equilibrium solution chemistry for uranophane. The Data0.com.R2 database adopts properties of aqueous uranium species from Grenthe et al. (1992). Based on this estimate, the EQ3NR geochemical modeling code version 7.2b (Wolery, 1992) was used to determine starting solution chemistries for the solubility experiments. Experimental solutions were prepared in matrices of 102 M CaCl2 and 103 M SiO2(aq). Test solutions for uranophane dissolution (i.e., solutions undersaturated with uranophane) were prepared to have initial U concentrations of 0.0 and 107 M. The initial concentrations of U in experimental solutions for uranophane precipitation (i.e., solutions supersaturated with uranophane) were determined from the results of the uranophane dissolution experiments (reported hereinafter). These results indicated that experimental solutions were in near-equilibrium with uranophane in the latter stages of the dissolution experiments (i.e., after test solutions had reacted with uranophane for 6 weeks). Based on these late stage solution compositions, experimental solution compositions estimated to be supersaturated with uranophane were prepared to have initial U concentrations of 106 and 2  106 M. Before addition of uranophane, the pHs of test solutions were adjusted to 6.0 by addition of 1.3  104 g CaCO3

to the dissolution test solutions and 3.5  105 g CaCO3 to the precipitation test solutions. After addition of CaCO3, test solutions were allowed to equilibrate with atmospheric CO2(g) for 2 weeks. The solubility experiments were carried out by reacting measured amounts of these solutions (200.0 ± 0.3 g) with measured amounts of synthetic uranophane (0.500 ± 0.001 g) in 250-ml polypropylene bottles. Experiments were conducted at room temperature (20.5 ± 2.0 °C) under atmospheric PCO2 conditions. Solutions were continuously agitated during the experiments using a gyratory shaker. For the dissolution experiments, two 5-ml aliquots of the experimental solutions were taken at 24 h and then at 1-week intervals for 10 weeks. For the precipitation experiments, two 5-ml aliquots of the experimental solutions were taken at 24 and 48 h and then at 1-week intervals for 10 weeks. Before each sampling, the test solutions were centrifuged (4.0  104 m/s2 for >15 min) to remove suspended solids. Test solution weights were measured before and after sampling to track loss of solution due to sampling and evaporation. The pH of one of the 5-ml aliquots was measured immediately after sampling using a ROSSTM semi-micro glass combination pH electrode (Thermo Orion, Beverly, MA, USA). This electrode had an uncertainty of ±0.03 pH units. The other 5-ml aliquot was acidified to pH < 2.0 by addition of concentrated HNO3. Concentrations of cations in the acidified aliquots were determined by ICP-AES; U concentrations were measured by ICPMS. Based on comparison of expected and measured values of calibration standards, measurement of Ca and Si by ICPAES had an uncertainty of ±2%, while measurement of U by ICP-MS had an uncertainty of ±10%. 3. RESULTS 3.1. Experimental data 3.1.1. Dissolution experiments The pHs and concentrations of Ca, Si, and U measured in dissolution test solutions D-0A and D-0B with initial U concentrations of 0.0 M and in dissolution test solutions

Uranophane dissolution and growth

D-7A and D-7B with initial U concentrations of 107 M are plotted as a function of reaction time in Figs. 2 and 3, respectively. The values plotted at time 0 in Figs. 2 and 3 represent solution pH and concentrations prior to addition of uranophane.

4511

Reaction of synthesized uranophane with the dissolution test solutions was characterized by relatively constant pH over the first week of the tests followed by a general decrease in pH from week 2 to 5 (Figs. 2 and 3). The decrease in solution pHs in the early parts of the dissolution tests

Fig. 2. The pH and concentrations of Ca, Si, and U in uranophane dissolution test solutions D-0A and D-0B plotted as a function of time. Error bars show analytical uncertainties in pH and concentration measurements. Best fit curves computed from regression analyses of measured concentration data are shown for Ca, Si, and U. Spurious Ca and U concentrations in the test solutions noted in Section 3.1.1 were not included in the regression analyses.

4512

J.D. Prikryl / Geochimica et Cosmochimica Acta 72 (2008) 4508–4520

Fig. 3. The pH and concentrations of Ca, Si, and U in uranophane dissolution test solutions D-7A and D-7B plotted as a function of time. Error bars show analytical uncertainties in pH and concentration measurements. Best fit curves computed from regression analyses of measured concentration data are shown for Ca, Si, and U. Spurious Ca and U concentrations in the test solutions noted in Section 3.1.1 were not included in the regression analyses.

was independent of initial dissolved uranyl concentrations, suggesting a surface phenomenon such as hydroxide sorption or Ca2+ exchange with H+. Over the remainder of the dissolution tests (i.e., from week 6 to 10), solution pH remained relatively constant, ranging from 5.86 to 5.92. Small up and down variations in solution pH in the latter

parts of the tests (i.e., from week 5 to 10) tended to be within the analytical error expected for pH measurement (±0.03 pH units). An initial increase in the Ca content of the uranophane dissolution test solutions is observed at the 24-h sampling interval (Figs. 2 and 3) and suggests rapid dissolution of

Uranophane dissolution and growth

uranophane. After the initial increase, Ca contents generally decreased at the 1-week sampling interval. Over the remainder of the dissolution tests (i.e., from week 1 to 10), a general trend of increasing Ca content was observed with the exception of outliers at the week 3 sampling interval of test D-0A and the week 10 sampling interval of test D-7B. The trend of increasing Ca content in the dissolution test solutions was caused by the effects of evaporation. Measurement of test solution masses before and after sampling indicated that, over the 1-week sampling intervals, dissolution test solutions lost on average 0.65 ± 0.15 g of mass. The effects of evaporation on Ca contents would be greatest in the latter stages of the tests as the mass of the test solutions progressively decreased due to sampling (i.e., weekly removal of 10 g of solution for pH and cation analysis). Small up and down variations along the increasing trends of Ca content in the dissolution tests are generally within the analytical error expected in the Ca measurements (±2%). Si contents in the test solutions generally decreased over the first 3 or 4 weeks of the tests and then generally increased over the remainder of the tests (i.e., from week 5 to 10) (Figs. 2 and 3). The general trend of decreasing Si content in the early stages of the dissolution tests strongly suggests precipitation of a Si-bearing phase or phases. The general trend of increasing Si content in the latter stages of the dissolution tests is attributed to the effects of evaporation. As with Ca, the effects of evaporation on Si contents would be greatest in the latter stages of the tests as the mass of test solutions progressively decreases due to sampling. Small up and down variations along the general trends of decreasing and then increasing Si content in the dissolution tests are generally within the analytical error expected in the Si measurements (±2%). Rapid dissolution of uranophane is indicated by an initial increase in the U content of the dissolution test solutions at the 24-h sampling interval (Figs. 2 and 3). From 24 h to week 3, an overall increase was observed in the U contents of the tests solutions. With the exception of several spuriously high measurements (e.g., weeks 4 and 10 of test D-0A, weeks 4 and 7 of test D-0B, week 7 of test D-7A, and weeks 6 and 10 of test D-7B), U contents over the remainder of the tests (i.e., from week 3 to 10) varied over a relatively narrow concentration range (i.e., 1.84  107 to 3.06  107 M). Although test solutions were centrifuged prior to sampling to remove suspended solids, the spuriously high U contents suggest that uranophane colloids may have been present in the sampled experimental solutions. Disregarding these spurious measurements, up and down variations in U contents of each test solution over the duration of the dissolution tests sometimes exceeded variations in U that might be expected to result from analytical uncertainty alone (±10%). These variations are attributed to analytical and sampling errors, the relatively low U contents of the dissolution test solutions, and the relatively high sorption coefficients for U at near-neutral pH. 3.1.2. Precipitation experiments The pHs and concentrations of Ca, Si, and U measured in precipitation test solutions P-1A and P-1B, which had U

4513

concentrations of 106 and 2  106 M, respectively, are plotted as a function of reaction time in Fig. 4. The values plotted at time 0 represent solution pH and concentrations prior to addition of uranophane. Reaction of synthesized uranophane with precipitation test solution P-1A was characterized by an increase in pH over the first 24 h of reaction followed by a decrease in pH over the first week of the test (Fig. 4). The pH of test solution P-1A increased from week 1 to 4 and then remained relatively constant over the remainder of the test (pH range of 5.94– 5.96 from week 5 to 10). In test solution P-1B, a trend of decreasing solution pH was observed over the first 2 weeks of the tests (Fig. 4). Over the remainder of the tests (i.e., from week 3 to 10), the pH of test solution P-1B remained relatively constant, ranging from 5.78 to 5.84. Small up and down variations in solution pH in the latter parts of the precipitation tests (i.e., from week 4 to week 10) were within the analytical error expected for pH measurement (±0.03 pH units). The complex trend in the pH behavior of test solution P-1A suggests that it may not have been in equilibrium with atmospheric PCO2 prior to adding the synthesized uranophane. As a result of the significant pH increase over the first 24 h of the test, the pH of test solution P-1A was consistently higher than the pH of test solution P-1B. Precipitation of a Ca-bearing phase is indicated by initial decreases in the Ca content of the precipitation test solutions at the 24- and 48-h sampling intervals (Fig. 4). From the 1-week sampling interval, a general trend of increasing Ca content is observed. As in the dissolution tests, the trend of increasing Ca content resulted from the effects of evaporation. Measurement of test solution masses before and after sampling indicated that, over the 1-week sampling intervals, precipitation test solutions lost on average 0.33 ± 0.12 g of mass. The small up and down variations observed along the trends of increasing Ca content in the precipitation tests were generally within the analytical error expected in the Ca measurements (±2%). Si contents in the precipitation test solutions decreased over the first 2 weeks of reaction and then remained relatively constant over the remainder of the tests (Fig. 4). Precipitation of a Si-bearing phase or phases is strongly suggested by the trend of decreasing Si content over the first 2 weeks of the tests. The relatively constant Si content of solutions in the latter stages of the precipitation tests suggests a balance between evaporation and Si precipitation. Again, small up and down weekly variations in the Si content of the precipitation test solutions are within the analytical error expected in the Si measurements (±2%). A decrease in the U content of the precipitation test solutions is observed at the 24- and 48-h sampling intervals (Fig. 4), and along with initial decreases in Ca and Si, suggests rapid precipitation of uranophane. With the exception of several outliers (e.g., weeks 8 and 10 of test P-1A and weeks 5, 6, 9, and 10 of test P-1B), U contents increased slightly but steadily over the remainder of the tests. Disregarding spuriously high U measurements, U concentrations in the precipitation tests ranged from 2.25  107 to 5.00  107 from week 3 to week 10. The steady increase in the U contents of the precipitation tests can again be attributed to the effects of evaporation. As in the dissolution tests, up

4514

J.D. Prikryl / Geochimica et Cosmochimica Acta 72 (2008) 4508–4520

Fig. 4. The pH and concentrations of Ca, Si, and U in uranophane precipitation test solutions P-1A and P-1B plotted as a function of time. Error bars show analytical uncertainties in pH and concentration measurements. Best fit curves computed from regression analyses of measured concentration data are shown for Ca, Si, and U. Spurious U concentrations in the test solutions noted in Section 3.1.2 were not included in the regression analyses.

and down variations in the U contents of the precipitation test solutions sometimes exceeded variations in U that might be expected to result from analytical uncertainty (±10%) and are attributed to analytical and sampling errors, the relatively low U contents of the test solutions, and relatively high U sorption coefficients at near-neutral pH.

3.2. Mass transfer relations Thermodynamic and kinetic interpretation of the uranophane solubility data requires knowledge of mass transfer (i.e., moles of Ca, Si, and U released or precipitated) as a function of time and solution chemistry. The

Uranophane dissolution and growth

4515

cumulative release in a dissolution or precipitation experiment is given by nI;R ðts Þ ¼ mI ðts ÞW ðts Þ þ nI;E ðts Þ  mI ðt0 ÞW ðt0 Þ

ð1Þ

where nI,R(ts) is the net number of moles of a component I (e.g., Ca, Si, or U) released to solution at the time of sampling (ts) (which is negative for net precipitation), mI(ts) is the molality of I in the solution at time ts, mI(t0) is the molality of I in solution at the start of the experiment (t0), W(ts) is the mass of solution prior to sampling at ts, W(t0) is the mass of solution at time t0, and nI,E(ts) is the number of moles of I extracted in all solution samples removed at all times tp prior to time ts, which is given by nI;E ðts Þ ¼ RmI ðtp ÞW E ðtp Þ

ð2Þ

where mI(tp) is the molality of I in solution taken at time tp, and WE(tp) is the mass of solution extracted in the sample taken at time tp. Measurements of experimental solution masses before and after sampling provide values of W and WE, which allow effects of variations in solution mass due to sampling and evaporation to be explicitly accounted for in the cumulative mass transfer calculations. Because of the variable solution mass, cumulative moles released corresponding to each sampling time and analytical measurement depend on measured concentrations for all earlier experimental times. Spurious measurements in combination with errors in the concentrations of Ca, Si, and U resulting from the sampling and analysis of the test solutions at earlier experimental times will therefore adversely affect calculated values of the total mass released for all subsequent times. To minimize these effects, the spurious Ca and U concentrations in the test solutions noted in Section 3.1 were removed, and then regression analyses were performed on the Ca, Si, and U concentration data. Calculated concentrations resulting from the regression analyses were then used to calculate solute masses extracted and cumulative moles of Ca, Si, and U released at each sampling time. The best fit curves computed by the regression analyses are shown in Figs. 2–4. 3.2.1. Dissolution experiments Results of the mass transfer calculations for Ca, Si, and U (i.e., moles of Ca, Si, and U released or precipitated as a function of time and solution chemistry) in the uranophane dissolution tests are illustrated in Fig. 5. The mass transfer data indicate net release of Ca and U and net precipitation of Si over the duration of the tests. Over the first half of the dissolution tests (i.e., from 24 h to week 5), mass transfer of Ca and U was characterized by progressive release and mass transfer of Si was characterized by progressive precipitation. In the latter stages of the dissolution tests (i.e., from week 6 to week 10), mass transfer of Ca, Si, and U tended to be relatively steady, suggesting that test solutions and solids were near equilibrium. 3.2.2. Precipitation experiments Results of the mass transfer calculations for Ca, Si, and U in the uranophane precipitation experiments are illustrated in Fig. 6. The mass transfer data indicate net precipitation of Ca, Si, and U over the duration of the experiments. Over the first part of the precipitation tests (i.e., from 24 h to week

Fig. 5. Net moles of Ca, Si, and U released to solution in the uranophane dissolution tests plotted as a function of time.

5), mass transfer of Ca and Si was characterized by progressive precipitation. In contrast, mass transfer of U in the first part of the precipitation tests was characterized by initial precipitation at the 24- and 48-h sampling intervals followed by progressive release of U up to the 5-week sampling interval. The U mass transfer data suggest that precipitation test solutions were driven to undersaturation with respect to uranophane in the very early parts of the tests (i.e., up to 48 h) as a result of rapid precipitation of uranophane and possibly another uranyl phase (e.g., soddyite [(UO2)2SiO42H2O], metaschoepite [UO32H2O], becquerelite [Ca(UO2)6O4

4516

J.D. Prikryl / Geochimica et Cosmochimica Acta 72 (2008) 4508–4520

CaðUO2 Þ2 ðSiO3 OHÞ2  5H2 O þ 6Hþ $ Ca2þ þ 2UO2 2þ þ 2SiO2 ðaqÞ þ 9H2 O:

ð3Þ

The corresponding reaction quotient (Q) for uranophane is defined by   2 6 ð4Þ Q ¼ Ca2þ UO2 2þ ½SiO2 ðaqÞ2 ½Hþ  where the square brackets represent thermodynamic activities corresponding to a standard state of a one molal solution referenced to infinite dilution. Using measured test solution pH and calculated Ca, Si, and U concentrations based on regression analyses of measured test solution chemistries, and assuming equilibrium with atmospheric PCO2, activities of the aqueous species in Eq. (4) were determined using the EQ3NR Version 7.2b geochemical code with the Data0.com.R2 database (Wolery, 1992). Reaction quotients for uranophane dissolution were then calculated using Eq. (4). Calculated log Qs as a function of reaction time are plotted in Fig. 7. Log Qs over the duration of the dissolution tests ranged from 10.62 to 11.08 and over the duration of the precipitation tests from 10.22 to 11.10 (Fig. 7). Because of their high initial concentrations, the activities of Ca2+ and SiO2(aq) were relatively constant over the duration of the tests and had little influence on variations in log Q values. Variations in log Qs were influenced primarily by changes in solution pH (i.e., [H+]) and the activity of UO2 2þ . Variations in U concentrations, which could significantly affect calculated log Q values, were limited by using U values derived from regression analysis to calculate UO2 2þ activities. Error bars showing the possible range in calculated log Q values resulting from analytical error in pH measurements (±0.03 pH units) are displayed for test D-0A in Fig. 7. This possible error range could be expected for all of the test solutions. Up and down variations in calculated log Q values in the latter parts of the tests (i.e., from week 6 onward), which are assumed to be near equilibrium, could be attributed to errors in pH measurement alone. 4.2. Evaluation of soddyite precipitation

Fig. 6. Net moles of Ca, Si, and U released to solution in the uranophane precipitation tests plotted as a function of time.

(OH)68H2O], or CaUO4). In the latter stages of the precipitation tests (i.e., from week 6 to 10), mass transfer of Ca, Si, and U tended to be relatively steady, suggesting that tests solutions and solids were near equilibrium. 4. DISCUSSION 4.1. Thermodynamic calculations The equilibrium reaction for uranophane dissolution can be written as

Net precipitation of Si in relation to net release of Ca and U in dissolution test solutions suggests precipitation of a secondary Si-bearing phase or phases. Likewise, after initial precipitation of Ca, Si, and U in the precipitation tests, continued precipitation of Si in contrast to release of Ca and U from week 1 to week 4 (Fig. 5) also suggests precipitation of a secondary Si-bearing phase or phases. Silica concentrations in the test solutions are higher than the solubility of amorphous silica, suggesting that precipitation of Si relative to U release could be due to silica precipitation. Casas et al. (1994) characterized recovered solids from dissolution experiments on natural uranophane samples by scanning electron microscopy (SEM) that showed the formation of the uranyl silicate soddyite [(UO2)2SiO42H2O] on the surface of strongly corroded uranophane. Soddyite has also been observed in the progression of secondary uranyl phases formed by corrosion of synthetic UO2 and spent UO2 fuel under conditions designed to simulate a potential Yucca Mountain repository and by weathering of uraninite

Uranophane dissolution and growth

Fig. 7. Calculated log Qs for the uranophane dissolution in the dissolution and precipitation test solutions plotted as a function of time. Error bars show analytical uncertainties in pH measurements (±0.03 pH units) for test solution D-0A. This error range could be expected for all test solutions.

at the Nopal I U deposit in Mexico (Wronkiewicz et al., 1992, 1996; Pearcy et al., 1994; Finch et al., 1999). In addition, calculation of equilibrium phase relations among uranyl minerals indicates that the soddyite stability field

4517

encroaches on the uranophane field at the temperatures (20.5 ± 2.0 °C) and chemistries of the dissolution and precipitation test solutions (Murphy, 2007). Using SEM to compare solids recovered from the dissolution and precipitation test containers in this study, no textures characteristic of strong leaching or soddyite growing from the surface of uranophane crystals were revealed (Fig. 8). In addition, particulate matter other than uranophane captured on polycarbonate filters with submicron openings did not exhibit morphologies that would indicate formation of soddyite (i.e., blocky, tabular cyrstals characteristic of the orthorhombic, dipyramidal system) as a discrete precipitate (Fig. 8). XRD analyses were also unable to identify soddyite as a minor phase in the recovered solids. However, the insensitivity of XRD to phases present in amounts less than approximately 5% by volume makes detection of minor phases difficult. Although soddyite was not identified by SEM and XRD analysis, dissolution and precipitation test solution chemistries were evaluated with respect to the solubility limit of soddyite to further evaluate the potential for soddyite precipitation in the experiments. In Fig. 9, the log activities of UO2 2þ =ðHþ Þ2 versus SiO2(aq) in dissolution and precipitation test solutions are plotted with respect to solubility limits of soddyite based on experimental data of Nguyen et al. (1992), a recalculation of the experimental data of Nguyen et al. (1992) based on a new set of complexation reactions and ionic strength corrections (Giammer and Hering, 2002), and theoretical prediction (Chen et al., 1999). Fig. 9 illustrates that the activities in the dissolution

Fig. 8. SEM photomicrographs showing (A) unreacted synthetic uranophane, (B) uranophane recovered from test D-0B, (C) uranophane recovered from test P-1A, and (D) particles other than uranophane recovered from test P-1A.

4518

J.D. Prikryl / Geochimica et Cosmochimica Acta 72 (2008) 4508–4520 Table 1 Aqueous phase U(VI) reactions and associated thermodynamic stability constants used to derive log K osp for uranophane

Fig. 9. Logarithmic activity diagram of UO2 2þ =ðHþ Þ2 versus SiO2(aq) illustrating the position of the uranophane dissolution and precipitation test solutions with respect to the solubility limit of soddyite. Solubility limits for soddyite are based on experimental data of Nguyen et al. (1992), a recalculation of the experimental data of Nguyen et al. (1992) based on a new set of complexation reactions and ionic strength corrections (Giammer and Hering, 2002), and theoretical prediction (Chen et al., 1999).

Speciation reaction

Log K (I = 0)a

UO2 2þ þ H2 O ¼ UO2 OHþ þ Hþ UO2 2þ þ 2H2 O ¼ UO2 ðOHÞ2 ðaqÞ þ 2Hþ UO2 2þ þ 3H2 O ¼ UO2 ðOHÞ3  þ 3Hþ UO2 2þ þ 4H2 O ¼ UO2 ðOHÞ4 2 þ 4Hþ 2UO2 2þ þ H2 O ¼ ðUO2 Þ2 OH3þ þ Hþ 2UO2 2þ þ 2H2 O ¼ ðUO2 Þ2 ðOHÞ2 2þ þ 2Hþ 3UO2 2þ þ 4H2 O ¼ ðUO2 Þ3 ðOHÞ4 2þ þ 4Hþ 3UO2 2þ þ 5H2 O ¼ ðUO2 Þ3 ðOHÞ5 þ þ 5Hþ 3UO2 2þ þ 7H2 O ¼ ðUO2 Þ3 ðOHÞ7  þ 7Hþ 4UO2 2þ þ 7H2 O ¼ ðUO2 Þ4 ðOHÞ7 þ þ 7Hþ UO2 2þ þ HCO3  ¼ UO2 CO3 ðaqÞ þ Hþ UO2 2þ þ 2HCO3  ¼ UO2 ðCO3 Þ2 2 þ 2Hþ UO2 2þ þ 3HCO3  ¼ UO2 ðCO3 Þ3 4 þ 3Hþ 3UO2 2þ þ 6HCO3  ¼ ðUO2 Þ3 ðCO3 Þ6 6 þ 6Hþ 2UO2 2þ þ HCO3  þ 3H2 O ¼ ðUO2 Þ2 CO3 ðOHÞ3 þ 4Hþ 3UO2 2þ þ HCO3  þ 3H2 O ¼ ðUO2 Þ3 OðOHÞ2 ðHCO3 Þþ þ 4Hþ 3UO2 2þ þ HCO3  þ 4H2 O ¼ ðUO2 Þ3 ðOHÞ5 CO2 þ þ 4Hþ 11UO2 2þ þ 6HCO3  þ 12H2 O ¼ ðUO2 Þ11 ðCO3 Þ6 ðOHÞ12 2 þ 18Hþ

5.21 10.31 19.22 33.03 2.71 5.63 11.93 15.59 31.05 21.95 0.66 3.75 9.43 8.06 11.22

a

and precipitation test solutions tended toward a limit corresponding to undersaturation with soddyite when compared to the solubility limits of Nguyen et al. (1992), Giammer and Hering (2002), and Chen et al. (1999). 4.3. Extraction of LogK osp for uranophane dissolution Solution pH and mass transfer of Ca, Si, and U were relatively constant in the latter parts of the dissolution and precipitation tests (i.e., from week 6 to 10), suggesting steady state conditions. For the last five sampling intervals, the range of log Qs for uranophane in the dissolution tests was 10.79–11.06 and the range of log Qs for uranophane in the precipitation tests was 10.96–11.10. The log Q midpoint of the overlap between these two ranges is 11.01, and this value was taken as the best approximation for the solubility product ðlog K osp Þ for uranophane. The standard deviation in the log Qs over the last five sampling intervals for the combined dissolution and precipitation tests was calculated to be 0.09. Over the duration of the dissolution and precipitation tests, Ca2+ and SiO2(aq) accounted for 99% or more of Ca and Si in solution. Aqueous phase U(VI) reactions used to derive the log K osp for uranophane and their associated stability constants (log K, I = 0) are listed in Table 1. From these, one can determine that the dominant uranyl species in solution was UO2(OH)2(aq) followed in succession by UO2OH+, UO2 2þ , and UO2CO3(aq). These four species accounted for 99% or more of UO2 2þ in solution. The range in the proportion of the major uranyl species in each test solution under presumed steady state conditions (i.e., from week 6 to 10) is listed in Table 2. Using a log K osp for uranophane dissolution of 11.01 ± 0.09 and standard state Gibbs free energy data from Guillaumont et al. (2003) for components of the

9.71 9.62 25.73

Source: Grenthe et al. (1992).

Table 2 Proportions of major uranyl species in dissolution and precipitation test solutions from week 6 to 10 UO2OH+

UO2 2þ

UO2CO3(aq)

Dissolution tests D-0A 0.780–0.788 D-0B 0.775–0.788 D-7A 0.761–0.775 D-7B 0.761–0.788

0.144–0.148 0.143–0.151 0.151–0.159 0.143–0.158

0.045–0.049 0.045–0.051 0.051–0.057 0.045–0.057

0.017 0.017 0.017 0.017

Precipitation tests P-1A 0.798–0.802 P-1B 0.741–0.751

0.135–0.137 0.168–0.163

0.039–0.040 0.061–0.066

0.017 0.016

UO2(OH)2(aq)

uranophane reaction in Eq. (3), the standard Gibbs free energy of formation for uranophane and the uranophane dissolution reaction can be determined. These values are 6196.1 ± 0.5 kJ mol1 for uranophane and 62.8 ± 0.5 kJ mol1 for the uranophane dissolution reaction. Based on log K osp of 9.4 ± 0.5 and 11.7 ± 0.6 determined from uranophane dissolution experiments, Nguyen et al. (1992) and Perez et al. (2000) derived free energies of formation for uranophane of 6210.6 ± 7.6 and 6192.3 ± 3.4 kJ mol1, respectively. Using the log K osp of 7.8 ± 0.8 determined by Casas et al. (1994) from dissolution experiments using natural uranophane samples and free energy data for reactants in Eq. (3) from Guillaumont et al. (2003), a free energy of formation for uranophane of 6214.5 ± 4.5 kJ mol1 was calculated. Chen et al. (1999) developed a semi-empirical method for estimating free energies and enthalpies of formation for uranyl phases by summing contributions of the component

Uranophane dissolution and growth

Fig. 10. Logarithmic activity diagram of UO2 2þ =ðHþ Þ2 versus SiO2(aq) versus Ca2+/(H+)2, illustrating the position of test solution chemistries for the last five sampling intervals with respect to the solubility limit of uranophane based on experimental data from this study; the solubility limits of uranophane based on solubility products reported by Nguyen et al. (1992), Casas et al. (1994) and Perez et al. (2000); and the theoretical solubility limit of uranophane predicted by Chen et al. (1999).

polyhedral units of the minerals. According to this model, when silica concentrations exceed 104 M, uranyl silicate phases such as uranophane are predicted to be the thermodynamically favored solids. Based on the method, Chen et al. (1999) estimated a free energy of formation of uranophane of 6189.2 kJ mol1. A logarithmic activity diagram of UO2 2þ =ðHþ Þ2 versus SiO2(aq) versus Ca2+/(H+)2, illustrating the position of test solution chemistries for the last five sampling intervals from the dissolution and precipitation tests, is plotted in Fig. 10. The solubility limit of uranophane based on the results of this study (i.e., calculated using a log K osp of 11.01); the solubility limits for uranophane based on solubility products reported by Nguyen et al. (1992), Casas et al. (1994), and Perez et al. (2000) (i.e., log K osp of 9.4, 11.7, and 7.8, respectively); and the solubility limit of uranophane estimated by Chen et al. (1999) are also plotted on the activity diagram. The estimated uranophane solubility limit of Chen et al. (1999) was calculated using an equilibrium constant for uranophane dissolution of 12.2, which was derived from Eq. (3) and the estimated standard state free energy of formation of uranophane (6189.2 kJ mol1) reported by Chen et al. (1999) and free energy data for other reactants from Guillaumont et al. (2003). 5. CONCLUSIONS Uranophane solubility experiments were designed to approach uranophane equilibrium from both undersaturation

4519

and supersaturation in Ca- and Si-rich solutions at low temperatures (20.5 ± 2.0 °C) and near-neutral pH (6.0). Sampling schemes were chosen to track the mass transfer of uranophane components (i.e., moles of Ca, Si, and U) as a function of time and solution chemistry. Trends indicated that pH and elemental release of all three uranophane components (i.e., Ca, Si, and U) were relatively steady in the latter stages of the tests (e.g., at 7 to 10 weeks), suggesting that test solutions were in near equilibrium with uranophane. Based on test solution chemistries in the last five sampling intervals of the dissolution and precipitation tests, a log K osp of 11.01 ± 0.09 for uranophane dissolution was approximated. The assumed precipitation of uranophane in this study, as determined by mass transfer data, is consistent with the widespread occurrence of uranophane in oxidizing low temperature U deposits (Pearcy et al., 1994). Determination of a reversed solubility for uranophane (i.e., an approach to uranophase equilibrium from supersaturated conditions) places an additional constraint on the thermodynamic data needed to predict the stability of uranophane, especially under conditions (e.g., low temperature and near-neutral pH) that are expected to exist over the period of regulatory concern at a potential Yucca Mountain repository. ACKNOWLEDGMENTS The author acknowledges and thank W.M. Murphy, who contributed substantially to the conceptual design and interpretation of the results of this experimental study. D. Pickett, R. Pabalan, G. Wittmeyer, and two anonymous reviewers provided constructive reviews. Assistance from N. Naukam in preparing the document is gratefully acknowledged. This paper describes work performed by the Center for Nuclear Waste Regulatory Analyses (CNWRA) and its contractors for the U.S. Nuclear Regulatory Commission (USNRC) under Contract No. NRC–02–07–006. The activities reported here were performed on behalf of the USNRC Office of Nuclear Material Safety and Safeguards, Division of High-Level Waste Repository Safety. This paper is an independent product of CNWRA and does not necessarily reflect the view or regulatory position of the USNRC.

REFERENCES Blanchard F. (1987) Joint Committee on Powder Diffraction Standards (JCPDS), Grant-in-Aid Report. Burns P. C., Ewing R. C. and Miller M. L. (1997) Incorporation mechanisms of actinide elements into the structures of U6+ phases formed during the oxidation of spent nuclear fuel. J. Nucl. Mater. 245, 1–9. Burns P. C., Deely K. M. and Skanthakumar S. (2004) Neptunium incorporation into uranyl compounds that form as alteration products of spent nuclear fuel: implications for geologic repository performance. Radiochim. Acta 92, 151–159. Casas I., Bruno J., Cera E., Finch R. J. and Ewing R. C. (1994) Kinetic and thermodynamic studies of uranium minerals. Assessment of the long-term evolution of spent nuclear fuel. SKB Technical Report 94-16. Chen F., Ewing R. C. and Clark S. B. (1999) The Gibbs free energy and enthalpies of formation of U6+ phases: an empirical method of prediction. Am. Mineral. 84, 650–664.

4520

J.D. Prikryl / Geochimica et Cosmochimica Acta 72 (2008) 4508–4520

DOE (2002) Final environmental impact statement for a geologic repository for the disposal of spent nuclear fuel and high-level radioactive waste at Yucca Mountain, Nye County, Nevada. DOE/EIS-0250, Department of Energy, Office of Civilian Radioactive Waste Management, North Las Vegas, Nevada. Douglas M., Clark S. B., Friese J. I., Arey B. W., Buck E. C., Hanson B. D., Utsunomiya S. and Ewing R. C. (2005) Microscale characterization of uranium (VI) silicate solids and associated neptunium (V). Radiochim. Acta 93, 265–272. Finch R. J., Buck E. C., Finn P. A. and Bates J. K. (1999) Oxidative corrosion of spent UO2 fuel in vapor and dripping groundwater at 90 °C. In Scientific Basis for Nuclear Waste Management XIX Materials Research Society Symposium Proceedings, vol. 556 (eds. D. J. Wronkiewicz and J. H. Lee). Materials Research Society, Warrendale, PA, pp. 431–438. Forbes T. Z. and Burns P. C. (2008) Synthesis, structure, and infrared spectroscopy fo the first Np5+ neptunyl silicates, Li6(NpO2)4(H2Si2O7) (HSiO4)2(H2O)4 and K3(NpO2)3(Si2O7). Inorg. Chem. 47, 705–712. Frondel C., Riska D. and Frondel, J. W. (1956) X-ray powder data for uranium and thorium minerals. U.S. Geological Survey Bulletin 1036-G, p. 70. Giammer D. E. and Hering J. G. (2002) Equilibrium and kinetic aspects of soddyite dissolution and secondary phase precipitation in aqueous suspension. Geochim. Cosmochim. Acta 66(18), 3235–3245. Grenthe I., Fuger J., Konings R. J. M., Lemire R. J., Mueller A. B., Nguyen-Trung C. and Wanner H. (1992). Chemical thermodynamics of uranium. Nuclear Energy Agency, Organisation for Economic Co-operation and Development. Elsevier, Amsterdam, the Netherlands. Guillaumont R., Fanghanel T., Fuger J., Grenthe I., Neck V., Palmer D. A. and Rand M. H. (2003) Update on the chemical thermodynamics of uranium, neptunium, plutonium, americium and technetium. In Chemical Thermodynamics Series 5 (eds. F. J. Mompean, M. Illemassene, C. Domenech-Orti and K. Ben-Said). Elsevier, p. 970. Johnson L. H. and Shoesmith D. W. (1988) Spent fuel. In Radioactive Waste Forms for the Future (eds. W. Lutze and R. C. Ewing). Elsevier, pp. 635–698. Klingensmith A. L. and Burns P. C. (2007) Neptunium substitution in synthetic uranophane and soddyite. Am. Mineral. 92, 1946– 1951. Mohanty S., Codell R., Menchaca J. M., Janetzke R., Smith M., LaPlante P., Rahimi M. and Lozano A. (2002) System-level performance assessment of the proposed repository at Yucca Mountain using the TPA version 4.1 code. CNWRA 2002–05, Rev. 1, Center for Nuclear Waste Regulatory Analyses, San Antonio, Texas. Murphy W. M. (2007) Hydrothermal phase relations among uranyl minerals at the Nopal 1 analog site. In Scientific Basis for Nuclear Waste Management XIX Materials Research Society Symposium Proceedings (eds. D. Dunn, C. Poinssot and B. Begg) vol. 985, Paper no. 0985-NN01-10.

Nguyen S. N., Silva R. J., Weed H. C. and Andrews, Jr., J. E. (1992) Standard Gibbs free energy of formation at the temperature 303.15 K of four uranyl silicates: soddyite, uranophane, sodium boltwoodite, and sodium weeksite. J. Chem. Thermodynam. 24, 359–376. Oversby V. M. (1994) Nuclear waste materials. In Materials Science and Technology, A Comprehensive Treatment, Nuclear Materials 10B (ed. B. R. T. Frost). VCH, Weinheim, Germany, pp. 392–442. Pearcy E. C., Prikryl J. D., Murphy W. M. and Leslie B. W. (1994) Alteration of uraninite from the Nopal I deposit, Pen˜a Blanca District, Chihuahua, Mexico, compared to degradation of spent nuclear fuel in the proposed U.S. high-level nuclear waste repository at Yucca Mountain, Nevada. Appl. Geochem. 9, 713– 732. Perez I., Casas I., Martin M. and Bruno J. (2000) The thermodynamics and kinetics of uranophane dissolution in bicarbonate test solutions. Geochim. Cosmochim. Acta 64(4), 603–608. Pickett D. A. and Leslie B. W. (2005) Mobilization of uranium from the Nopal I analog and implications for nuclear waste repository performance. In Annual Meeting Geological Society of America. Geological Society of America Abstracts with Programs, vol. 37, No. 7, p. 268. Pickett D. A. and Murphy W. M. (1997) Isotopic constraints on radionuclide transport at Pena Blanca. In Seventh EC Natural Analogue Working Group Meeting Proceedings, EUR 17851 EN (eds. H. Von Maravic and J. Smellie), pp. 113–122. Nuclear Science and Technology Series. European Communities, Stein am Rhein, Switzerland: Luxembourg. Prikryl J. D. and Murphy W. M. (2004) The dissolution of uranophane in CaCl2–SiO2(aq) test solutions. In Scientific Basis for Nuclear Waste Management XXVIII Materials Research Society Symposiun Symposium Proceedings (eds. J. M. Hanchar, S. Stroes-Gascoyne and L. Browning). Materials Research Society, Warrendale, PA, pp. 189–194. Vochten R., Blanton N., Peeters O., Van Springel K. and Van Haverbeke L. (1997) A new method of synthesis of boltwoodite and of formation of sodium boltwoodite, uranophane, sklodowskite and kasolite from boltwoodite. Can. Mineral. 35, 735– 741. Wolery T. (1992) EQ3/6, A software package for geochemical modeling of aqueous systems. UCRL-MA-110662 PT1, Lawrence Livermore National Laboratory, Livermore, CA. Wronkiewicz D. J., Bates J. K., Gerding T. J., Veleckis E. and Tani B. S. (1992) Uranium release and secondary phase formation during unsaturated testing of UO2 at 90 °C. J. Nucl. Mater. 190, 107–127. Wronkiewicz D. J., Bates J. K., Wolf S. F. and Buck E. C. (1996) Ten-year results from unsaturated drip tests with UO2 at 90 °C: Implications for the corrosion of spent nuclear fuel. J. Nucl. Mater. 238, 78. Associate editor: Eric H. Oelkers