Thermodynamic Model for Uranium Release from Hanford Site Tank

Jan 26, 2011 - ABSTRACT: A thermodynamic model of U solid-phase sol- ubility and paragenesis was developed for Hanford Site tank residual waste that w...
1 downloads 0 Views 2MB Size
ARTICLE pubs.acs.org/est

Thermodynamic Model for Uranium Release from Hanford Site Tank Residual Waste Kirk J. Cantrell,*,† William J. Deutsch,‡ and Mike J. Lindberg† † ‡

Pacific Northwest National Laboratory, P.O. Box 999, Mail Stop K6-81, Richland, Washington 99352, United States Geochemistry Services LLC, 724 Tyler Street, #3, Port Townsend, Washington 98368, United States

bS Supporting Information ABSTRACT: A thermodynamic model of U solid-phase solubility and paragenesis was developed for Hanford Site tank residual waste that will remain in place after tank closure. The model was developed using a combination of waste composition data, waste leach test data, and thermodynamic modeling of the leach test data. The testing and analyses were conducted using actual Hanford Site tank residual waste. Positive identification of U phases by X-ray diffraction was generally not possible either because solids in the waste were amorphous or their concentrations were not detectable by XRD for both asreceived and leached residual waste. Three leachant solutions were used in the studies: deionized water, CaCO3 saturated solution, and Ca(OH)2 saturated solution. Analysis of calculated saturation indices indicate that NaUO2PO4 3 xH2O and Na2U2O7(am) are present in the residual wastes initially. Leaching of the residual wastes with deionized water or CaCO3 saturated solution results in preferential dissolution Na2U2O7(am) and formation of schoepite. Leaching of the residual wastes with Ca(OH)2 saturated solution appears to result in transformation of both NaUO2PO4 3 xH2O and Na2U2O7(am) to CaUO4. Upon the basis of these results, the paragenetic sequence of secondary phases expected to occur as leaching of residual waste progresses for two tank closure scenarios was identified.

’ INTRODUCTION At the U.S. Department of Energy’s (DOE) Hanford Site in southeastern Washington state, millions of liters of radioactive waste generated through reprocessing of spent nuclear fuel were stored in 177 single- and double-shell underground storage tanks. Most of these tanks (149) consisted of single-shell, steel and concrete storage tanks. The waste in the single-shell tanks is presently being retrieved in order to meet Hanford Federal Facility Agreement and Consent Order - Tri-Party Agreement Action Plan Milestone M-45-00 that requires the “retrieval of as much tank waste as technically possible, with tank waste residues not to exceed 360 cubic feet (cu. ft.) in each of the 100 Series tanks, 30 cu. ft. in each of the 200 Series tanks, or the limit of waste retrieval technology capability, whichever is less”. These volumes are equivalent to approximately 1 in. of residual waste remaining in each tank at closure. As part of an ongoing project, chemical analysis, phase characterization, and leach testing of residual waste from four retrieved Hanford Site sludge singleshell tanks (SSTs) (241-C-103, 241-C-106, 241-C-202, and 241C-203), one salt cake SST (241-S-112), and one partially retrieved sludge SST (241-C-108) have been completed to support closure of the SSTs. A major component of these studies is to develop contaminant release models for the residual waste that can be used in performance assessment models to evaluate r 2011 American Chemical Society

the long-term risks to human health and the environment associated with closure of Hanford Site underground storage tanks. For ease of reading, the tank designations will not include the 241 prefix from this point forward in the article. The primary objective of this waste characterization and laboratory testing work was to develop a mechanistic approach (thermodynamic or kinetic) to describe the potential future release of U from Hanford tank residual waste that could be used in tank closure performance assessment modeling. To evaluate the release of U from the residual waste, batch leaching experiments were conducted. Positive identification of most U phases present in the waste before and after contact with leachants was generally not possible due either to their amorphous character or to low concentration. Thermodynamic modeling of leachate concentrations was performed to determine if equilibrium was attained with U phases that could be present and how these phases evolve as leaching progresses. Using this approach it was determined that an equilibrium thermodynamic model for U phase solubility could in fact be used to describe U Received: November 23, 2010 Accepted: January 6, 2011 Revised: January 4, 2011 Published: January 26, 2011 1473

dx.doi.org/10.1021/es1038968 | Environ. Sci. Technol. 2011, 45, 1473–1480

Environmental Science & Technology release from Hanford tank residual waste. With this established, reaction path modeling was conducted to predict future U release concentrations and U phase paragenesis for two Hanford Site tank closure scenarios.

’ BACKGROUND During production years at the Hanford Site (1944-1990), various chemical processes were used to separate Pu and U from spent nuclear fuel. Radioactive wastes from these chemical processes were stored in large (210 to 3800 m3) SSTs and also later in double-shell tanks. Some of these SSTs are known to have leaked. To prevent further environmental impacts, most of the liquid waste stored in the SSTs has been removed and retrieval of the remaining solid wastes is ongoing. The focus of this study is the residual wastes that remain in tanks C-202 and C-203. Process knowledge indicates that wastes from the Bismuth Phosphate Plant were major components of materials disposed to tanks C-201, C-202, C-203, and C-204.1 The main U solids in tank sludge that resulted from this waste stream were Na4UO2(CO3)3 (cejkaite) and NaUO2PO4 3 xH2O.2 Retrieval of waste from these tanks began by pumping out the supernatant.1 The solids left after this operation are referred to here as prefinal retrieval waste. Final waste retrieval for tanks C-202 and C-203 was accomplished using a vacuum system to remove as much sludge as possible.3 A high pressure water spray was used with the vacuum to break up larger particles of waste that could not be removed solely with the vacuum. After the waste retrieval process was completed, samples of the residual material were collected for analysis and testing. This material is referred to as tank residual waste. ’ MATERIALS AND METHODS Analytical and test methods used to characterize tank residual waste samples and leachates have been previously described.3 A brief overview is provided in this section. The elemental and contaminant concentrations of the wastes were measured by complete dissolution of the bulk residual waste solids using fusion-dissolution procedures and acid digestions. Digests and leachates were analyzed using a combination of methods, including inductively coupled plasma-mass spectrometry, inductively coupled plasma-optical emission spectroscopy, and several radiochemical analytical techniques. Because the two solid digestion methods require the addition of acids to fully solubilize the waste solids, they are not appropriate techniques for determining anion concentrations of the bulk residual waste. Anion concentrations of the bulk residual waste solids were estimated separately by adding results from sequential deionized water (DI) extracts of the bulk residual wastes. The anion concentrations in these extracts were measured using ion chromatography. This approach may underestimate the total quantities of anions in the sample, particularly for anions that can form insoluble precipitates. The carbon contents (total carbon and total inorganic carbon [TIC]) of the wastes were determined with a carbon analyzer. Solid-phase characterization techniques that were routinely used include X-ray diffraction (XRD) and scanning electron microscopy/energy dispersive spectrometry (SEM/EDS). XRD was used to identify crystalline phases. SEM, in combination with EDS and X-ray fluorescence element mapping techniques, were used to characterize phase associations, morphologies, particle sizes, surface textures, and compositions of solid particles in the unleached and leached residual waste samples. In some cases, synchrotron-based micro-X-ray diffraction was also used. The

ARTICLE

synchrotron-based X-ray analyses were completed on beamline ID-20 at the Advanced Photon Source (Argonne National Laboratory [ANL], Argonne, Illinois). Leachants used in these studies include DI water, Ca(OH)2 saturated solution, and CaCO3 saturated solution. The Ca(OH)2 leachant was used to simulate conditions associated with the tanks being filled with cement and is intended as a simplified representation of cement pore fluid that will contact the tank residual waste after closure. The CaCO3 saturated leachant is used to simulate a future scenario in which meteroic water infiltrates through the vadose zone, into the interior of the SSTs, and then contacts the residual waste. The aqueous leaching experiments were conducted using both single-contact experiments and sequential contact experiments. The single-contact experiments typically included 1-day and 1-month contact periods. The sequential contact experiments typically included six stages in which the leachate solution was replaced with fresh leachant after each contact period. The contact period for stages 1-5 was approximately 1 day and 30 days for stage 6. The rationale for conducting the sequential contact leaching experiments versus the single-contact leaching experiments was to determine the impact of preferential removal of more readily soluble components in the residual waste on secondary uranium phase formation and solubility, and how these change over time (or contact with increasing amounts of pore water). All the leach tests were conducted at a solution to solid weight ratio of approximately 100. Experiments were conducted at room temperature (∼20 °C). It was determined that the 1-day contact period did not reach equilibrium. Because of this, only results for the 30-day single contact and stage 6 of the sequential contact experiments will be discussed. Thermodynamic equilibrium modeling was used to calculate mineral saturation indices and to identify solid phases potentially in equilibrium with the leachate compositions and to run reaction path simulations. The saturation index is defined as SI = log (Q/Ksp), where Q is the activity product and Ksp is the mineral solubility product at equilibrium at the temperature of interest. Minerals with SI values near zero (within ( ∼0.5) are generally considered to be at or near equilibrium, more positive values are considered oversaturated, and more negative values are considered undersaturated with respect to the solution composition. Geochemist’s Workbench version 8.09 4 was used to calculate the mineral SIs for the leachates and run reaction path simulations. Previous SI calculations 3 have been revised to include thermodynamic constants for a number of additional U solid phases as well as solution-phase complexes. The thermodynamic database thermo.com.V8.R6þ.dat was used for the modeling calculations. The database was augmented to include solubility products for cejkaite [Na4(UO2)(CO3)3] and NaUO2PO4 3 xH2O,5 becquerelite [Ca(UO2)6O4(OH)6 3 8H2O],6 Na-compreignacite [Na2 (UO2)6O4(OH)6 3 7H2O],7 and Na diuranate hydrate [Na2U2 O7 3 xH2O],8 an estimated value for urancalcarite [Ca(UO2)3 (CO3)(OH)6 3 3H2O],9 leibigite [Ca2UO2(CO3)3 3 10H2O] and andersonite [Na2Ca(UO2)(CO3)3 3 5H2O],10,11 autunite [Ca (UO2)2(PO4)2],12 and soddyite [(UO2)2SiO4 3 2H2O],13 and stability constants for the dissolved species CaUO2(CO3)32-(aq) and Ca2UO2(CO3)30(aq).14

’ RESULTS Solid-Phase Characterization Results. Samples of unleached, 1 month single-contact DI water leached, 1 month single-contact 1474

dx.doi.org/10.1021/es1038968 |Environ. Sci. Technol. 2011, 45, 1473–1480

Environmental Science & Technology Ca(OH)2 leached, and 1 month single-contact CaCO3 leached postretrieval residual waste from tank C-202 and C-203 (19961) were characterized by bulk XRD.3 Samples of unleached C-203 (19887) and stage six of the sequential leached C-203 (19961) solids for all three leachants were also characterized by XRD. The XRD results indicated that all these samples contain mostly amorphous (noncrystalline) solids. All of the XRD patterns3 contained a broad diffraction profile (or hump) from approximately 10 to 30°2θ. This feature is indicative of diffraction from amorphous materials, which cannot be identified by XRD methods. Diffraction from the nitrocellulose binder contributed to this broad profile. Except for the possible presence of quartz (SiO2) in the sample of unleached residual waste (C-202), no crystalline phases other than corundum (used as a 2θ internal standard) were identified in any of the samples except for the C-202 1 month single-contact Ca(OH)2 leached sample. For the C-202 1 month single-contact Ca(OH)2 leached sample, one unidentified reflection was found in the XRD ) noted pattern. This was a low angle reflection at 15.02°2θ (5.89 Å in the XRD pattern for the 1 month single-contact Ca(OH)2 leached sample. Otherwise, there were no major unassigned reflections in the XRD patterns for the postretrieval residual waste samples, which suggests that these samples did not likely contain any major crystalline phases present at more than ∼5-10 wt % of the sample mass. Based on published tank chemistry and characterization information, quartz is not expected to be a component in these wastes. Because quartz is one of the principal minerals in Hanford sediments, its presence in the C-202 sample likely resulted from blowing dust or sediment that fell into the tank during sampling or other tank operation activities. These XRD results for C-202 and C-203 postretrieval residual waste are generally consistent with those for the water-leached preretrieval wastes from tanks C-203 and C-204.15,16 Like the C-202 and C-203 residual waste, the C-203 and C-204 waterleached preretrieval wastes contained mostly amorphous solids, and no significant quantities of any crystalline phases were  ejkaite was the primary detected in their bulk XRD patterns. C crystalline phase identified by bulk XRD in the unleached C-203 and C-204 preretrieval sludge. The XRD pattern for an unleached C-203 sample also indicated the possible presence of nitratine (soda niter, NaNO3) at a concentration estimated to be significantly less than 25% of the cejkaite concentration.15,16 Analyses by synchrotron-based μXRD indicated the possible presence of goethite [R-FeO(OH)], maghemite (γ-Fe2O3), and the Na uranates, clarkeite and/or Na2U2O7, in DI water-leached preretrieval waste from tank C-203.17 Although Deutsch et al. 15 did not identify goethite or maghemite in their bulk XRD analyses of this same C-203 preretrieval waste, they did determine the presence of Fe oxides by SEM/EDS. Similarly, the bulk XRD analyses of the postretrieval (residual) waste from tanks C-202 and C-203 did not indicate the presence of any crystalline Fe oxide phases; however, Fe-oxide particles, often containing trace amounts of Mn, Cr, and sometimes Pb, were discovered by SEM/EDS in the C-202 and C-203 postretrieval residual waste. The same samples analyzed by XRD indicated above were also examined by SEM/EDS. A detailed analysis of the SEM/EDS data is provided in Deutsch et al.,3 along with all SEM micrographs and EDS analyses collected for the postretrieval samples from tanks C-202 and C-203. The major findings are summarized here. All samples contained a combination of individual and aggregate particles from several hundred to less than a micrometer in size. The particles were nondescript and appeared to be amorphous due to a general absence of crystal faces. Because

ARTICLE

XRD analyses did not indicate the presence of any crystalline phases in these samples, it is assumed that the amorphouslooking particles are likely noncrystalline; however, without further studies at higher magnification at the submicrometer scale, such as with transmission electron microscopy (TEM), the noncrystallinity of these particles could not be verified. The SEM samples of unleached C-202 residual waste were similar in appearance to the unleached C-203 postretrieval residual waste. The residual waste consists of particles generally having one of two common compositions. One composition consists of U, Na, C, O, P, and possibly H (H is not detectable by EDS). The other composition is an Fe oxide that often contained trace amounts of Mn, Cr, and sometimes Pb. Trace levels of Si and Al were sometimes observed in the C-203 samples. Material in the unleached and leached residual waste samples typically consist of mostly particle aggregates or individual grains with fine particles adhered to their surface. As a result, EDS analyses generally indicated the presence of both U and Fe in most particles. The two general compositions of particles present in the residual waste are essentially the same as those determined for the unleached and leached C-203 samples; however, the C-202 residual waste (after leaching in DI) appears to contain more Fe oxide particles relative to the U-containing phase than the unleached samples of C-203 residual waste. For those analyses of unleached residual waste that best reflect the compositions of just the U-Na-C-O-P ( H particles (i.e., the normalized atomic % values containing little or no Fe), the Na/U ratios range from approximately 2:1 to 1:1. For the U-Na-C-O-P ( H particles in the leached residual waste with little or no Fe, the Na/U ratios were generally less than the unleached samples, typically ranging from approximately 1.5:1 to 0.4:1 with some Na/U ratios as low as 0.1:1. These results suggest several possible scenarios: (1) the leach product may contain a mixture of the original U-Na-C-O-P ( H phase and a new Na poor or Na-absent U phase where its solubility was exceeded and then precipitated during the course of the leach study, (2) the U-Na-C-O-P ( H phase may be dissolving incongruently, (3) the waste solids may contain a readily soluble Na phase that contains no U, or (4) the U-Na-C-O-P ( H phase may consist of two or more U phases having similar compositions. The EDS analyses of the Ca(OH)2 and CaCO3 leached residual waste samples indicated that the Ca content of the U-containing phase increased relative to that of unleached and DI water leached samples. This effect was pronounced in the Ca(OH)2 leached solids and minor for the CaCO3 leached solids. It was not possible from the SEM/EDS and XRD results alone to ascertain the mechanism responsible for this shift in compositions. Saturation Indices. Saturation indices calculated for the three leachants after contact with the residual waste are shown in Tables 1-3. Results for the DI water leachates are shown in Table 1, CaCO3 solution leachate results are shown in Table 2, and Ca(OH)2 leachate results are shown in Table 3. The DI water leach results in Table 1 indicate that most of the phases are  ejkaite is highly undersaturated indicating that undersaturated. C if any residual cejkaite did occur in these residual tank waste samples, it was readily dissolved. The sodium uranyl phosphate phase, NaUO2PO4 3 xH2O, is consistently near saturation in all leachates shown in Table 1, with SI values ranging from -0.48 to 0.78 and an average of 0.26. This indicates that NaUO2PO4 3 xH2O remained stable throughout the sequential leach tests. Na2UO2O7(c) is undersaturated in tank C-202 leachates 1475

dx.doi.org/10.1021/es1038968 |Environ. Sci. Technol. 2011, 45, 1473–1480

Environmental Science & Technology

ARTICLE

Table 1. Saturation Index Values Calculated for Potentially Relevant Uranium and Calcium Phases in Contact with 1 Month Single-Contact Leach Tests and Stage 6 of the Sequential Leach Tests for Tanks C-202 and C-203 Residual Waste Samples (DI water leachant)a DI Water Leach C-202 (sample 19250)

phase

a

C-203 (sample 19887)

C-203 (sample 19961)

1-month single

stage 6

1-month single

stage 6

1-month single

stage 6

contact

(30 days)

contact

(30 days)

contact

(30 days) -10.57

cejkaite Na4(UO2)(CO3)3

-8.54

-12.86

-7.92

-9.90

-7.52

Na2U2O7 (c)

-2.15

-1.23

2.02

4.46

3.61

2.91

Na2U2O7 (am)

-4.66

-3.74

-0.49

1.96

1.10

0.41

Na-compreignacite Na2(UO2)6O4(OH)6 3 7H2O schoepite UO3 3 2H2O schoepite-dehydrated UO3 3 H2O

-5.18 -1.39

1.97 0.17

-0.10 -1.16

9.97 0.74

3.58 -0.64

7.11 0.42

-1.51

-0.10

-1.43

0.47

-0.91

0.15

NaUO2PO4 3 xH2O

0.46

0.78

-0.48

0.47

-0.04

0.38

Ca-autunite Ca(UO2)2(PO4)2

(-3.81)

(-0.40)

(-6.05)

-2.88

-5.17

(-3.09)

(UO2)3(PO4)2 3 4H2O

-4.72

-0.32

-10.07

-5.74

-9.22

-4.53

becquerelite Ca(UO2)6O4(OH)6 3 8H2O

(-13.07)

(-3.15)

(-8.35)

3.00

-4.66

(0.11)

rutherfordine UO2CO3

-3.70

-2.64

-5.81

-4.29

-5.69

-4.14

liebigite Ca2UO2(CO3)3 3 10H2O andersonite Na2CaUO2(CO3)3 3 6H2O

(-9.05) (-2.22)

(-7.84) (-3.78)

(-9.17) (-1.97)

-8.57 -2.66

-8.75 -1.57

(-9.31) (-3.37)

urancalcarite Ca(UO2)3CO3(OH)6 3 3H2O

(-9.93)

(-5.19)

(-8.25)

-2.99

-6.53

(-4.43)

CaUO4

(-4.83)

(-2.70)

(-1.26)

0.57

-0.18

(-0.69)

whitlockite Ca3(PO4)2

(-14.85)

(-13.41)

(-10.88)

-11.66

-9.93

(-13.08)

hydroxylapatite Ca5(PO4)3(OH)

(-21.70)

(-19.26)

(-14.08)

-15.27

-12.37

(-17.88)

calcite

(-5.43)

(-5.35)

(-4.43)

-4.89

-4.28

(-5.34)

Parentheses indicate SI values calculated with reported Ca concentrations that were below the quantification limit.

and oversaturated in tank C-203 leachates. Na2UO2O7(am) is also undersaturated in tank C-202 leachates and near saturation or oversaturated in tank C-203 leachates. Na2UO2O7(am) is near saturation in the 1 month single-contact leachate for the tank C-203 (19887) sample and stage six of the tank C-203 (19661) sample. These results suggest that Na2UO2O7(am) is likely stable in at least some of the tank C-203 leachates. As indicated previously, poorly crystalline clarkeite [Na(UO 2 )O(OH)H2O0-1] or Na uranate (Na2U2O7) was identified by bulk XRD and μXRD in DI water leached tank C-203 preretrieval waste 15-17 and is expected to be present in the leached waste. Na-compreignacite is either undersaturated or highly oversaturated in the DI water leachates. This suggests that for the conditions of these tests, Na-compreignacite precipitation is kinetically inhibited. Both schoepite (UO3 3 2H2O) and dehydrated schoepite (UO3 3 H2O) are undersaturated in the 1 month single-contact leachates and near saturation in stage six of the sequential contact leachates, with dehydrated schoepite being closer to equilibrium than schoepite. These results suggest that, as leaching of the waste progresses and Na2UO2O7(am) becomes undersaturated and dissolves, dehydrated schoepite becomes a stable phase. Note the actual structure of the uranyl oxide hydrate phase in contact with the leachates (schoepite, dehydrated schoepite, metaschoepite, or polycrystalline solid) is not discernible from these modeling results. All that can be concluded is that the log K associated with dehydrated schoepite provides the better fit to the leachate data. CaUO4 is generally undersaturated in the DI water leachates; however, it is near saturation in the leachates for stage six of tank C-203 sample 19887 and the 1 month single-contact of tank C-203 sample

19961. This supports the case that CaUO4 may be stable for conditions that occur in some of the DI water leachates. The SI results for the CaCO3 leachates (Table 2) are very similar to those of the DI water leachates. In this case, SI values for NaUO2PO4 3 xH2O range from -1.46 to 0.99 with an average of 0.10. As was the case for the DI water leachates, the results for the CaCO3 leachates are consistent with the presence of NaUO2PO4 3 xH2O and Na2UO2O7(am) in the earlier stages of the leaching process with dehydrated schoepite becoming stable later as leaching progresses. The results indicate NaUO2PO4 3 xH2O is present throughout all stages of the sequential contact experiments for all samples. It is also apparent that Na2UO2O7(am) is only near equilibrium in the 1 month single contacts of the tank C-203 samples, and Na2UO2O7(am) dissolved completely during stages 2-5. In addition, Na2UO2O7(am) was either not initially present in the C-202 sample or any Na2UO2O7(am) that was present dissolved completely during the 1 month single-contact tests. SI results for the Ca(OH)2 leachates (Table 3) are quite different than the DI water and CaCO3 leachates. In this case, nearly all U phases are undersaturated with the exception of CaUO4. For the tank C-202 sample, both the 1 month singlecontact leachate and the stage six leachates are near equilibrium with CaUO4. For the tank C-203 samples, the 1 month singlecontact leachates are highly over saturated [SI = 1.31 and 4.89], while the stage six leachates are much less oversaturated (SI = 0.75 and 0.97). This could indicate that CaUO4 is slow to precipitate, the phase that is actually forming in our experiments is a somewhat more soluble hydrated form of CaUO4, or the actual phase has a different composition. 1476

dx.doi.org/10.1021/es1038968 |Environ. Sci. Technol. 2011, 45, 1473–1480

Environmental Science & Technology

ARTICLE

Table 2. Saturation Index Values Calculated for Potentially Relevant Uranium and Calcium Phases in Contact with 1 Month Single-Contact Leach Tests and Stage 6 of the Sequential Leach Tests for Tanks C-202 and C-203 Residual Waste Samples (CaCO3 leachant)a CaCO3 Leach C-202 (sample 19250)

phase

a

C-203 (sample 19887)

C-203 (sample 19961)

1-month single

stage 6

1-month single

stage 6

1-month single

stage 6

contact

(30 days)

contact

(30 days)

contact

(30 days)

cejkaite Na4(UO2)(CO3)3

-8.66

-14.11

-7.33

-10.64

-7.41

Na2U2O7 (c)

-2.54

-5.98

2.68

1.82

3.60

-0.44

Na2U2O7 (am)

-5.05

-8.49

0.17

-0.69

1.09

-2.95

Na-compreignacite Na2(UO2)6O4(OH)6 3 7H2O schoepite UO3 3 2H2O schoepite-dehydrated UO3 3 H2O

-6.24 -1.56

-10.91 -1.86

2.25 -0.74

5.81 0.37

4.66 -0.37

3.09 0.25

-1.85

-2.13

-1.01

0.10

-0.64

-0.02

NaUO2PO4 3 xH2O

0.15

-1.46

0.08

0.45

0.36

0.99 (-1.18)

-11.21

Ca-autunite Ca(UO2)2(PO4)2

(-4.18)

(-3.38)

(-5.06)

(-2.55)

(-4.38)

(UO2)3(PO4)2 3 4H2O

-5.45

-6.13

-8.35

-3.43

-7.60

-0.44

becquerelite Ca(UO2)6O4(OH)6 3 8H2O

(-13.87)

(-14.53)

(-6.12)

(-0.80)

(-3.59)

(-3.24)

rutherfordine UO2CO3

-3.81

-3.94

-5.21

-3.53

-5.11

-2.49

liebigite Ca2UO2(CO3)3 3 10H2O andersonite Na2CaUO2(CO3)3 3 6H2O

(-8.66) (-2.09)

(-6.09) (-3.53)

(-8.81) (-1.50)

(-8.61) (-3.05)

(-8.65) (-1.46)

(-8.60) (-3.33)

urancalcarite Ca(UO2)3CO3(OH)6 3 3H2O

(-10.18)

(-9.75)

(-7.10)

(-4.53)

(-5.96)

(-5.42)

CaUO4

(-4.80)

(-3.93)

(-1.14)

(-1.35)

(-0.47)

(-3.20)

whitlockite Ca3(PO4)2

(-14.49)

(-10.73)

(-11.33)

(-13.65)

(-10.79)

(-15.51)

hydroxylapatite Ca5(PO4)3(OH)

(-21.06)

(-14.83)

(-14.91)

(-19.05)

(-13.94)

(-22.70)

calcite

(-5.18)

(-3.83)

(-4.55)

(-5.29)

(-4.52)

(-5.83)

Parentheses indicate SI values calculated with reported Ca concentrations that were below the quantification limit.

A number of U(VI)-Ca phases can form in alkaline solutions at room temperature, including becquerelite [Ca(UO2)6O4(OH)6 3 2H2O], Ca2UO5 3 (H2O)1.3-1.7, CaUO4, Ca3UO6, and CaU2 O7.18 It was suggested that the most likely phase at high pH in contact with Ca(OH)2 is Ca2UO5 3 (H2O)1.3-1.7;18 however, it was later presumed that this phase is a hydrated form of CaUO4, with a variable water content ranging from CaUO4 to CaUO4 3 H2O.19 More recent modeling of U(VI) precipitation experiments conducted in artificial cement pore waters and low-alkali pore waters indicated that amorphous Ca-uranate [CaUO4(am)], with a solubility product of log Ksp0 = 23.1, produced good agreement between the predicted solubility limits and experimental data for all solution compositions investigated.20 Using the critically reviewed value of log Ksp0 = 15.94 for CaUO4(c) provided in the database,4,21 produced reasonably good agreement between our experimental data and modeling results. It is possible that the much higher log Ksp0 required in 20 may have been because the aqueous species CaUO2(CO3)32-(aq) and Ca2UO2(CO3)30(aq) 14 were not included in their calculations. In addition to CaUO4(s) precipitation, it is possible that incorporation or coprecipitation of some UO22þ into calcite or aragonite could have occurred in our Ca(OH)2 leaching experiments.22-24 The residual wastes used in our experiments initially contain carbonate (as cejkaite and/or Na2CO3). During contact with the Ca(OH)2 leachant, calcium carbonate precipitation could have occurred. The solubility calculations indicate that calcite is oversaturated in stage 6 of the Ca(OH)2 sequential leach tests. The impact of this process on our experiments is expected to be small because of the limited amount of carbonate in the waste

and the weak partition coefficients for calcite exhibited by U (0.01 to 0.26).25 Analysis of the SI data indicates that the general trend for U leaching from tank residual waste by CaCO3 leachant is for the initial NaUO2PO4 3 xH2O and Na2UO2O7(am) to dissolve to form schoepite with Na2UO2O7(am) dissolving preferentially to NaUO2PO4 3 xH2O. This process results in an apparent leaching of Na preferentially to U. This qualitative trend is supported by the EDS analysis results of the unleached and DI and CaCO3 leached residual waste. The general trend for U leaching from tank residual waste by the CaOH2 leachant is for the initial NaUO2PO4 3 xH2O and Na2UO2O7(am) to dissolve to form CaUO4 with NaUO2PO4 3 xH2O dissolving preferentially to Na2UO2O7(am). This process results in an apparent exchange of Ca for Na. This qualitative trend was also supported by the EDS analysis results of the unleached and Ca(OH)2 leached residual waste.

’ TANK CLOSURE SCENARIO MODELING Future release of U from tank residual waste depends on a number of variables that could potentially impact the hydrology and water chemistry that will contact the tank residual waste after closure. Two illustrative modeling examples are presented to contrast the impact of possible closure scenarios from a geochemical perspective. The first scenario considered is a case in which the tank is filled with local sand or sediment to maintain its integrity. In this case, it is assumed meteoric water that contacts the tank residual waste has a composition similar to typical Hanford Site groundwater, the infiltration rate over 10 000 years is 1.0 mm/year,26 the tank itself has no impact on the hydrology 1477

dx.doi.org/10.1021/es1038968 |Environ. Sci. Technol. 2011, 45, 1473–1480

Environmental Science & Technology

ARTICLE

Table 3. Saturation Index Values Calculated for Potentially Relevant Uranium and Calcium Phases in Contact with 1 Month Single-Contact Leach Tests and Stage 6 of the Sequential Leach Tests for Tanks C-202 and C-203 Residual Waste Samples [Ca(OH)2 leachant]a Ca(OH)2 Leach C-202 (sample 19250)

phase

C-203 (sample 19887)

C-203 (sample 19961)

1-month single

stage 6

1-month single

stage 6

1-month single

stage 6

contact

(30 days)

contact

(30 days)

contact

(30 days)

cejkaite Na4(UO2)(CO3)3

-14.91

-24.33

-11.49

-19.54

-9.60

-21.58

Na2U2O7 (c)

-7.34

-15.37

-1.46

-13.24

4.76

-12.02

Na2U2O7 (am)

-9.85

-17.88

-3.96

-13.24

2.25

-14.53

Na-compreignacite Na2(UO2)6O4(OH)6 3 7H2O schoepite UO3 3 2H2O

-31.39 -6.64

-49.93 -9.27

2.25 -4.01

-40.55 -8.09

2.32 -1.24

-43.89 -8.60

-6.91

-9.54

-4.28

-8.36

-1.51

-8.87

NaUO2PO4 3 xH2O

-9.38

-16.93

-5.17

(-13.67)

-2.26

(-14.52)

Ca-autunite Ca(UO2)2(PO4)2

(-18.33)

-28.47

(-12.24)

(-24.60)

-6.29

(-25.32)

(UO2)3(PO4)2 3 4H2O

-34.96

-49.92

-24.53

(-44.48)

-16.61

(-46.44)

becquerelite Ca(UO2)6O4(OH)6 3 8H2O

schoepite-dehydrated UO3 3 H2O

(-34.11)

-47.70

(-20.04)

-40.98

-2.60

-43.32

rutherfordine UO2CO3

-12.86

-15.91

-10.39

-15.04

-8.36

-15.89

liebigite Ca2UO2(CO3)3 3 10H2O andersonite Na2CaUO2(CO3)3 3 6H2O

(-5.10) (-3.43)

-4.61 -7.90

(-6.36) (-2.36)

-5.14 -5.77

-4.20 -0.33

-5.17 -6.81

urancalcarite Ca(UO2)3CO3(OH)6 3 3H2O

(-19.13)

-25.25

(-13.11)

-22.40

-4.73

-23.54

CaUO4

(0.39)

-0.05

(1.31)

0.75

4.89

0.97

whitlockite Ca3(PO4)2

(2.08)

1.57

(-0.53)

(2.31)

1.51

(4.08)

hydroxylapatite Ca5(PO4)3(OH)

(8.93)

9.25

(4.16)

(10.18)

7.62

(13.19)

calcite CaCO3

(1.13)

2.90

(-0.74)

2.20

-0.67

2.61

Parentheses indicate SI values calculated with Ca or PO43- concentrations that were below the quantification limit. In the case of Ca, reported values were used. In the case of PO4, detection limit values were used. a

Figure 1. Uranium and TIC concentrations (A) and the paragenetic sequence of U phases present in the waste (B) as a function of reaction progress for the sediment closure scenario. A reaction progress of 1 is equivalent to 10 000 years of infiltration at 1.0 mm/year.

or system chemistry, and waste in the tank is 1-in. thick. Using these assumptions, we applied a reaction progress model to simulate U concentrations in meteoric water as it infiltrates through the waste and the paragenesis of U phases in the residual waste over the course of 10 000 years. Figure 1 shows the U and dissolved TIC concentrations as a function of reaction progress, along with the paragenetic sequence of U phases present in the waste. A reaction progress of 1.0 is equivalent to 10 000 years for the assumed conditions. The initial high U concentration in solution (∼3.2  10-4 M) occurs as a result of high carbonate complexation of U. The high initial carbonate concentrations originate from the residual waste in the form of soluble Na2CO3

and/or cejkaite. As these very soluble carbonate phases dissolve and are flushed out, the U concentrations level out at approximately 2.6  10-5 M. The major U phases present in the waste initially are NaUO2PO4 3 xH2O and Na2UO2O7(am). As meteoric water flushes the system, Na2UO2O7(am) dissolves preferentially to NaUO2PO4 3 xH2O. A minor amount of andersonite forms briefly, early in the simulation, and is then replaced by CaUO4 and soddyite [(UO2)2SiO4 3 2H2O]. Schoepite surpasses these phases in quantitative importance as the reaction progress exceeds 0.5 and becomes more significant as the reaction progress approaches 1.0. Over the course of the simulation, 6.4% of the U in the waste dissolves. 1478

dx.doi.org/10.1021/es1038968 |Environ. Sci. Technol. 2011, 45, 1473–1480

Environmental Science & Technology

ARTICLE

Figure 2. Uranium and TIC concentrations (A) and the paragenetic sequence of U phases present in the waste (B) as a function of reaction progress for the cement closure scenario. A reaction progress of 1 is equivalent to 10 000 years of infiltration at 1.0 mm/year.

In a second closure scenario, the tank is filled with a cementitious grout or concrete. For this simulation it is assumed the tank is filled with concrete, the waste does not mix into the cement significantly, and water contacting the waste has the composition of a simulated cement pore water [0.015 M Ca(OH)2 and 10-5 M Si].20 The U and TIC concentrations and the paragenetic sequence of U phases for the cement scenario over the course of 10 000 years are shown in Figure 2. Analogous to the sediment scenario, initially high U concentrations occur in solution (∼3.2  10-4 M) because of high carbonate complexation of U. In the early part of the simulation, U concentrations decline rapidly and then rebound somewhat as a small amount of andersonite first precipitates and then dissolves. As the reaction progress continues, NaUO2PO4 3 xH2O dissolves preferentially to Na2U2O7(am). As carbonate continues to leach from the waste, U concentrations continue to decline until a plateau is reached at approximately 1.0  10-6 M. This occurs at the approximate point in the reaction progress where CaUO4 becomes the dominant phase. A dramatic reduction in U concentration occurs when Na2UO2O7(am) dissolves completely, leaving CaUO4 as the only phase to control U release concentrations. Comparison of results from the sediment closure scenario with those of the cement closure scenario indicate the cement closure scenario results in significantly reduced release of U over the course of the 10 000 years considered in the simulations, both in terms of concentrations in water contacting the waste and total mass. Over the course of the simulation, 2.0% of the U in the waste dissolves in the cement closure scenario compared to 6.4% that dissolves in the sediment closure scenario. Conditions that favor transformation of the relatively soluble U phases in tank residual waste to the highly insoluble CaUO4 phase (high Ca and high pH) are desirable for minimizing U mobility after tank closure. In addition to the phases indicated to be at or near equilibrium in the leachate experiments [NaUO2PO4 3 xH2O, Na2U2O7(am), schoepite, and CaUO4], andersonite is predicted to occur in both closure scenarios and soddyite is predicted to occur in the sediment closure scenario. Andersonite [Na2Ca(UO2)(CO3)3 3 5H2O] requires high concentrations of both Na and Ca in solution in order to remain stable. In general, andersonite was significantly undersaturated in all experimental leachates except the 1 month single-contact Ca(OH)2 leachate for tank sample C-203 (19961), which had an SI value that was near equilibrium (SI = -0.33). This suggests that in most of our leaching experiments the Na concentrations had been diluted to such an extent that formation of andersonite was not favored. This is

consistent with andersonite being stable only during the early stages of the closure scenarios. A minor amount of soddyite was predicted to occur in the sediment closure scenario as a result of the silica concentrations assumed for the meteoric water contacting the waste. Soddyite would not occur in our leaching experiments because the silica concentrations in the leachates were very low.

’ ASSOCIATED CONTENT

bS

Supporting Information. Leachate compositions used for the SI calculations, and the groundwater and waste compositions used for the reaction path simulations, are provided in Tables S1-S5. Further details of the reaction progress modeling are also discussed. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*Phone: (509) 371-7175; fax: (509) 371-7249; e-mail: kirk.cantrell@ pnl.gov.

’ ACKNOWLEDGMENT The authors acknowledge M. Connelly at Washington River Protection Solutions LLC (Richland, Washington) for providing project funding and technical guidance. Pacific Northwest National Laboratory is operated for DOE by Battelle Memorial Institute under contract DE-AC05-76RL01830. Synchrotronbased analyses were completed at the X-ray Science Division & Pacific Northwest Consortium beamline 20-ID at the Advanced Photon Source. The Pacific Northwest Consortium-Collaborative Access Team project is supported by funding from DOE’s Office of Basic Energy Sciences, the University of Washington, Simon Fraser University, and the Natural Sciences and Engineering Research Council of Canada. Use of the Advanced Photon Source is supported by DOE’s Office of Science, Office of Basic Energy Sciences, under contract DE-AC02-06CH11357. ’ REFERENCES (1) Johnson, M. E. Origin of Wastes in C-200 Series Single-Shell Tanks; RPP-15408; CH2M HILL Hanford Group: Richland, WA, 2003. (2) U.S. Atomic Energy Commission. Uranium Recovery Technical Manual; Declassified Document, HW-19140; General Electric Company: Richland, WA, 1951. 1479

dx.doi.org/10.1021/es1038968 |Environ. Sci. Technol. 2011, 45, 1473–1480

Environmental Science & Technology (3) Deutsch, W. J.; Krupka, K. M.; Lindberg, M. J.; Cantrell, K. J.; Brown, C. F.; Mattigod, S. V.; Schaef, H. T., Arey, B. W. Hanford Tanks 241-C-202 and 241-C-203 Residual Waste Contaminant Release Models and Supporting Data; PNNL-16229; Pacific Northwest National Laboratory: Richland, WA, 2007. (4) Bethke, C.; Yeakel, S. The Geochemist’s Workbench, Release 8.0, Reference Manual; Hydrogeology Program: Urbana, IL, 2009. (5) Felmy, A. R.; Xia, Y.; Wang, Z. The solubility product of NaUO2PO4 3 xH2O determined in phosphate and carbonate solutions. Radiochim. Acta 2005, 93 (7), 401–408. (6) Rai, D.; Felmy, A. R.; Hess, N. J.; LeGore, V. L.; McCready, D. E. Thermodynamics of the U(VI)-Ca2þ-Cl--OH--H2O system: Solubility product of becquerelite. Radiochim. Acta 2002, 90, 495–503. (7) Gorman-Lewis, D.; Fein, J. B.; Burns, P. C.; Szymanowski, J. E. S.; Converse, J. Solubility measurements of the uranyl oxide hydrate phases metaschoepite, compreignacite, Na-compreignacite, becquerelite, and clarkeite. J. Chem. Thermodyn. 2008, 40, 980–990. (8) Yamamura, T.; Kitamura, A.; Fukui, A.; Nishikawa, S.; Yamamoto, T.; Moriyama, H. Solubility of U(VI) in highly basic solutions. Radiochim. Acta 1998, 83, 139–146. (9) Chen, F.; Ewing, R. C.; Clark, S. B. The Gibbs free energies and enthalpies of formation of U6þ phases: An empirical method of prediction. Am. Mineral. 1999, 84, 650–664. (10) Gorman-Lewis, D.; Burns, P. C.; Fein, J. B. Review of uranyl mineral solubility measurements. J. Chem. Thermodyn. 2008, 40, 335– 352. (11) Alwan, A. K.; Williams, P. A. The aqueous chemistry of uranium minerals. Part 2. Minerals of the liebigite group. Mineral. Mag. 1980, 43, 665–667. (12) Gorman-Lewis, D.; Shvareva, T.; Kubatko, K. A.; Burns, P. C.; Wellman, D. M.; McNamara, B.; Szymanowski, J. E. S.; Navrotsky, A.; Fein, J. B. Thermodynamic properties of autunite, uranyl hydrogen phosphate, and uranyl orthophosphate from solubility and calorimetric measurements. Environ. Sci. Technol. 2009, 43, 7416–7422. (13) Gorman-Lewis, D.; Mazeina, L.; Fein, J. B.; Szymanowski, J. E. S.; Burns, P. C.; Navrotsky, A. Thermodynamic properties of soddyite from solubility and calorimetry measurements. J. Chem. Thermodyn. 2007, 39, 568–575. (14) Dong, W.; Brooks, S. C. Determination of the formation constants of ternary complexes of uranyl and carbonate with alkaline earth metals (Mg2þ, Ca2þ, Sr2þ, and Ba2þ) using anion exchange method. Environ. Sci. Technol. 2006, 40, 4689–4695. (15) Deutsch, W. J.; Krupka, K. M.; Lindberg, M. J.; Cantrell, K. J.; Brown, C. F.; Schaef, H. T. Hanford Tanks 241-C-203 and 241-C-204: Residual Waste Contaminant Release Model and Supporting Data; PNNL14903; Pacific Northwest National Laboratory: Richland, WA, 2004. (16) Krupka, K. M.; Schaef, H. T.; Arey, B. W.; Heald, S. M.; Deutsch, W. J.; Lindberg, M. J.; Cantrell, K. J. Residual Waste from Hanford Tanks 241-C-203 and 241-C-204. 1. Solids Characterization. Environ. Sci. Technol. 2006, 40, 3749–3754. (17) Deutsch, W. J.; Krupka, K. M.; Cantrell, K. J.; Brown, C. F.; Lindberg, M. J.; Schaef, H. T.; Heald, S. M.; Arey, B. W.; Kukkadapu, R. K. Advances in Geochemical Testing of Key Contaminants in Residual Hanford Tank Waste; PNNL-15372; Pacific Northwest National Laboratory: Richland, WA, 2005. (18) Atkins, M.; Beckley, A. N.; Glasser, F. P. Influence of cement on the near field environment and its specific interactions with uranium and iodine. Radiochim. Acta 1988, 44/45, 255–261. (19) Moroni, L. P.; Glasser, F. P. Reactions between cement components and U(VI) oxide. Waste Manage. 1995, 15, 243–254. (20) Tits, J.; Fujita, T.; Tsukamoto, M.; Wieland, E. Uranium(VI) uptake by synthetic calcium silicate hydrates. Mater. Res. Soc. Symp. Proc. 2008, 1107, 467–474. (21) Grenthe, I.; Fuger, J.: Konings, R. J. M.; Lemire, R. J.; Muller, A. B.; Nguyen-Trung Cregu, C.; Wanner, H. Chemical Thermodynamics, Vol. 1: Chemical Thermodynamics of Uranium; North-Holland: Amsterdam, 1992.

ARTICLE

(22) Reeder, R. J.; Nugent, M.; Lamble, G. M.; Tait, C. D.; Morris, D. E. Uranyl-incorporation into calcite and aragonite: XAFS and luminescence characterization. Environ. Sci. Technol. 2000, 34 (4), 638–644. (23) Reeder, R. J.; Nugent, M.; Tait, C. D.; Morris, D. E.; Heald, S. M.; Beck, K. M.; Hess, W. P.; Lanzirotti, A. Coprecipitation of uranium(VI) with calcite: XAFS, micro-XAS, and luminescence characterization. Geochim. Cosmochim. Acta 2001, 65 (20), 3491–3503. (24) Meece, D. E.; Benninger, L. K. The coprecipitation of Pu and other radionuclides with CaCO3. Geochim. Cosmochim. Acta 1993, 57, 1447–1458. (25) Curti, E. Coprecipitation of radionuclides with calcite: estimation of partition coefficients based on review of laboratory investigations and geochemical data. Appl. Geochem. 1999, 14 (4), 433–445. (26) Bergeron, M. P.; Connelly, M. P.; Cantrell, K. J.; Krupka, K. M.; and Fayer, M. J. Recharge and Waste Release within Engineered System in Waste Management Area C; RPP-RPT-44042, Rev. 1, Appendix A; Washington River Protection Solutions LLC: Richland, WA, 2010.

1480

dx.doi.org/10.1021/es1038968 |Environ. Sci. Technol. 2011, 45, 1473–1480