Environ. Sci. Technol. 2006, 40, 3755-3761
Residual Waste from Hanford Tanks 241-C-203 and 241-C-204. 2. Contaminant Release Model KIRK J. CANTRELL,* KENNETH M. KRUPKA, WILLIAM J. DEUTSCH, AND MICHAEL J. LINDBERG Pacific Northwest National Laboratory, P.O. Box 999, Mail Stop K6-81, Richland, Washington 99352
Release of U and 99Tc from residual sludge in Hanford waste tanks 241-C-203 and 241-C-204 at the U.S. Department of Energy’s (DOE) Hanford Site in southeastern Washington state was quantified by water-leaching, selective extractions, empirical solubility measurements, and thermodynamic modeling. A contaminant release model was developed based on these experimental results and solidphase characterization results presented elsewhere. Uranium release was determined to be controlled by two phases and occurred in three stages. In the first stage, U release is controlled by the solubility of cˇ ejkaite, which is suppressed by high concentrations of sodium released from the dissolution of NaNO3 in the residual sludges. Equilibrium solubility calculations indicate the U released during this stage will have a maximum concentration of 0.021 M. When all the NaNO3 has dissolved from the sludge, the solubility of the remaining cˇ ejkaite will increase to 0.28 M. After cˇ ejkaite has completely dissolved, the majority of the remaining U is in the form of poorly crystalline Na2U2O7 [or clarkeite Na[(UO2)O(OH)](H2O)0-1]. In contact with Hanford groundwater this phase is not stable, and becquerelite becomes the U solubility controlling phase, with a calculated equilibrium concentration of 1.2 × 10-4 M. For Tc, a significant fraction of its concentration in the residual sludge was determined to be relatively insoluble (20 wt % for C-203 and 80 wt % for C-204). Because of the low concentrations of Tc in these sludge materials, the characterization studies did not identify any discrete Tc solids phases. Release of the soluble fraction of Tc was found to occur concomitantly with NO3-. It was postulated that a NaNO3-NaTcO4 solid solution could be responsible for this behavior. The Tc release concentrations for the soluble fraction were estimated to be 2.4 × 10-6 M for C-203 and 2.7 × 10-5 M for C-204. Selective extraction results indicated that the recalcitrant fraction of Tc was associated with Fe oxides. Release of the recalcitrant fraction of Tc was assumed to be controlled by dissolution of Fe oxide in the form of ferrihydrite. Based on this assumption and measured values for the ratio of recalcitrant Tc to total Fe in each bulk sludge, the release concentration of the recalcitrant fraction of Tc was calculated to be 3.9 × 10-12 M for C-203 and 10.0 × 10-12 M for C-204.
* Corresponding author phone: (509)376-2136; fax: (509)376-5368; e-mail:
[email protected]. 10.1021/es0511568 CCC: $33.50 Published on Web 05/17/2006
2006 American Chemical Society
Introduction Beginning in 1944 and continuing throughout the cold war era, 177 large underground tanks located at the Hanford Site in the state of Washington were used to store radioactive waste. These wastes included supernatants and precipitated sludge originating from the processing of spent nuclear fuels, research and development programs, and various tank-waste management activities. Subsequent evaporation and waste management processes resulted in the formation of precipitated sodium salts within many of the tanks (1). As part of ongoing tank decommissioning efforts, as much tank waste as possible is being removed for treatment and final disposal prior to the decommissioning of each underground storage tank. After waste removal, some residual sludge will remain in these tanks. As part of the Resource Conservation and Recovery Act (RCRA) Corrective Action Program and DOE Order 435.1, a closure performance assessment must be conducted to evaluate the long-term risk to human health resulting from tank closure (2). As a result of the complex and continuously evolving process history, unplanned process releases, and waste management activities (including secondary processing of waste for targeted removal of specific waste components and evaporation of waste liquids), the compositions of solid phases and liquids present in the residual sludge within the tanks are highly uncertain. In addition, the wastes have been aged in a hot [predicted to be as high as 137 °C (3)], caustic, and extremely radioactive environment. As a result, direct characterization of the tank sludge must be conducted in order to develop conceptual models and reliable source terms for contaminant release to be used in tank closure performance assessments. Reported here is a synopsis of some of the most relevant results determined with total fusion analyses, water leaches, selective extractions, and empirical solubility measurements of residual sludges from tanks C-203 and C-204. Also discussed are thermodynamic modeling results for solutions from the empirical solubility experiments. These results in combination with results from X-ray diffraction (XRD), synchrotron X-ray microdiffraction (µXRD), and scanning electron microscopy/energy dispersive spectroscopy (SEM/ EDS) studies on C-203 and C-204 sludge samples described in the companion paper (4) were used to develop a contaminant release model for use in tank closure performance assessments. The sludge characterization data presented here are the first to be reported for tanks C-203 and C-204. Of the 177 tanks slated for closure, the C-203 and C-204 sludge samples were among the first to be available for detailed characterization. These results along with characterization results from other Hanford tank sludge samples to be collected in the future will be used to evaluate the overall applicability of the release models developed for tanks C-203 and C-204 to other tanks in the Hanford tank farm complex.
Methods Analytical Methods. The bulk composition of sludge solids was determined using a KOH-KNO3 fusion method developed internally at PNNL. Deutsch et al. (5) also determined bulk compositions using an acid digestion procedure based on a modified version of U.S. Environmental Protection Agency (EPA) SW 846 Method 3050B (6). For some elements, the fusion method provided slightly higher results and was used here to represent total concentrations. The KOH-KNO3 fusion method used to determine the total elemental VOL. 40, NO. 12, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3755
composition of the sludge is generally not appropriate for determining anion concentrations due to the compositions of the acids used in the method. The anion compositions were measured separately in solutions obtained by water leaching of the solids. The reader is referred to ref 5 for a detailed description of these and the other methods that follow. Following digestion of the sludge, major cations (including Al, Si, Ca, Mg, Na, K, Fe, and Mg) were analyzed by inductively coupled plasma-optical emission spectroscopy (ICP-OES). Analysis of trace metals (including Cr, Mo, As, Se, Cd, Ag, Pb, 99 Tc, and U isotopes) was performed by inductively coupled plasma-mass spectrophotometry (ICP-MS). Anion analysis was completed with ion chromatography. Fluoride, acetate, formate, chloride, nitrite, bromide, nitrate, carbonate, sulfate, oxalate, and phosphate were separated on a Dionex AS17 column with a gradient elution technique from 1 mM to 35 mM NaOH and measured using a conductivity detector. Prior to analysis, digestates were either centrifuged and then filtered, or carefully decanted and then filtered. Filtration was conducted with 0.45-µm pore size membranes. Solution pH values were measured using a solid-state pH electrode and a pH meter calibrated with buffers bracketing the expected range. Selective Extraction Experiments. Selective extractions were conducted on sludges from tanks C-203 and C-204 to facilitate the identification of solid phases containing U and 99Tc and to evaluate how these contaminants would be released into water. Characterization studies by XRD, µXRD, and SEM/EDS and analyses by fusion and acid digestion of sludge solids were used to design a sequence of the selective extractions (5). Sequential selective extractions were conducted with 0.3 g of sludge and 30 mL of each extractant. The sludge and each extractant were combined in a 50 mL polypropopylene centrifuge tube and placed on a shaker table for a typical contact period of 24 h. After the contact period the tube was centrifuged at 4000 rpm for 20 min. The supernatant was then carefully decanted and filtered through a 0.45 µm membrane. Steps 1 and 2 of the extraction process utilized double deionized (DDI) water as the extractant. DDI water is expected to remove readily soluble salts along with readily leachable uranium and 99Tc. Two DDI steps were used to increase the likelihood of complete dissolution of the readily soluble salts. In step 3, a buffer solution of 0.1 M acetic acid/0.1 M K-acetate (pH ∼ 4.6) was used for removal of any carbonate phases that were not removed in the DDI extractions [modified from that described in ref 7]. Steps 4 and 5 consisted of two ethanol extractions for removal of tributyl phosphate (TBP). Because all the U was accounted for in the other extracts, it was concluded that little or no U was associated with the TBP, and solutions from the ethanol extractions were not analyzed. Step 6 of the sequential extraction process consisted of an 8 M HNO3 extraction that was expected to remove a majority of the residual material contained within the Fe and Al oxides (for the sake of simplicity, all Al and Fe oxides, oxyhydroxides, and hydroxides will be referred to here as “oxides”). The last step of the sequence was a hot concentrated HNO3 extraction intended to dissolve recalcitrant residuals not dissolved in step 6. A second four-step set of sequential extractions was conducted to differentiate between 99Tc associated with recalcitrant Al containing phases from that associated with Fe oxides. A buffer solution of 0.02 M HF/0.01 NaF (pH ∼ 2.9) was used in these extractions to preferentially dissolve Al containing phases. Equilibrium thermodynamic modeling suggests that this solution should readily dissolve Al oxides while dissolving negligible quantities of Fe oxides. Higher buffer concentrations were avoided to preclude the precipitation of F-containing solids such as cryolite (Na3AlF6). The following sequence of extractions was used for this 3756
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 12, 2006
analysis on aliquots of the sludge samples. Steps 1 and 2 were DDI water extracts for soluble salt removal. Steps 3 and 4 utilized the 0.02 M HF/0.01 M NaF extractant (pH approximately 2.9) for removal of Al containing phases. For each extraction step, 0.3 g of sludge and 30 mL of extractant were combined and placed on a shaker table. The DDI extractions were contacted for 24 h each, and the HF/NaF extractions were contacted for 2 h each. Each extract was analyzed for Al, Fe, Cr, Si, and 99Tc. Empirical Solubility Measurements. XRD and SEM/EDS characterization of C-203 and C-204 tank sludges indicated that the majority of the U in the sludges was in the form of the mineral cˇejkaite [Na4UO2(CO3)3] (4, 5). The results of water-leaching experiments (5) indicated that nearly all the U in the C-203 and C-204 sludges was soluble at a solidto-solution ratio of 1:100 and dissolved in less than 24 h at room temperature (final solution pH values near 9.5). As a result of these findings, it was necessary to empirically measure the solubility of cˇejkaite to develop a U release model. A more detailed experimental study to determine the thermodynamic solubility constant for cˇejkaite was beyond the scope of this study. To estimate the solubility of cˇejkaite in C-203 sludge in contact with water, a series of water extraction experiments was conducted at higher solid-tosolution ratios than those used in the initial water-leach experiments (5). Solid-to-solution ratios of 1:1 and 2:1 were used to prevent the complete dissolution of the cˇejkaite during the extractions and to ensure that equilibrium with cˇejkaite would be achieved. In addition to the sludge samples, solubility measurements were also completed on some large yellow nuggets found in and separated from the C-203 sludge sample [Figure 2.3 (5)]. These nuggets were composed primarily of an intergrowth of needles and rods of cˇejkaite [Figure S-6, Supporting Information of ref 4]. Two solubility experiments were performed with the nuggets, one with a whole nugget and another with crushed nugget material. In the case of the nugget solubility experiments, a solid-tosolution ratio of approximately 1:2 was used. All solubility experiments used DDI water as the solvent so the results could be compared to those from the other batch leaching experiments. To minimize the impact of common ion effects that could result from dissolution of soluble salts in the sludge (other than cˇejkaite), the solubility determinations were conducted using a series of multiple contacts. After each equilibration period, most of the supernatant was removed for analysis and replaced with fresh deionized water. All solubility experiments were run at ambient temperature for 24 h, except the first C-203 experiment which lasted for 4 days (due to scheduling requirements). Based on results from singlecontact water-leach tests and periodic water replenishments tests (5), it was determined 24 h was adequate to achieve steady state. For the C-203 sludge samples, the water was removed and replaced with fresh DDI water and reequilibrated a total of four times. For the nugget samples, two sequential contacts were performed. All experiments were conducted in 50 mL centrifuge tubes and equilibrated on a shaker table for the indicated time period. After the prescribed contact period, the tubes were centrifuged at 4000 rpm for 30 min. Upon removal, the supernatants were analyzed for U, 99Tc, major cations and anions, and pH. Thermodynamic Modeling. The solubilities of relevant U phases and ferrihydrite (used for the release model for 99Tc) were calculated under various geochemical conditions using the React module in Geochemist’s Workbench 6.0 (GWB) (8). The thermodynamic databases supplied with the software were modified to include reaction constants relevant to the solution compositions of interest. At high ionic strength, the thermo_phrqpitz.dat database was used and modified to include the parameters shown in Table 1. The parameters
TABLE 1. Pitzer Ion-Interaction Parameters and Standard State Equilibrium Constants [25 °C (298 K), I ) 0.0] Used for Solubility Calculations ion interaction/reaction Na+
- UO2(CO3)3 Na+ - UO2(CO3)34+ Na - UO2(CO3)34NO3- - UO2(CO3)34Na4UO2(CO3)3(s) T UO2(CO3)34- + 4Na+ 2Ca2+ + UO22+ + 3CO32- T Ca2UO2(CO3)3(aq) Ca(UO2)6O4(OH)6‚8H2O + 14H+ T Ca2+ + 6UO22+ + 18H2O 4-
include the solubility constant for cˇejkaite, the Pitzer interaction parameters (9, 10) for (Na+ - UO2(CO3)34-) determined by Felmy et al. (11), and an estimated Pitzer ion interaction parameter (θ) for (NO3- - UO2(CO3)34-). Because a measured Pitzer ion interaction parameter for (NO3- UO2(CO3)34-) is not currently available, a value was estimated by assuming it was equal to that of (Cl- - UO2(CO3)34-) determined by Felmy et al. (11). Data for the formation and interactions of (UO2)3(CO3)66- were not included or required for the carbonate concentrations of interest to this study. The source of the experimental solubility data used by Felmy et al. (11) to determine the solubility contant and Pitzer ion interaction parameters includes 25 solubility data points (12) covering a wide range of aqueous media (NaClO4, Na2SO4, NaCO3, and NaCl) and concentrations. A previous analysis using specific ion interaction theory (13) used six of these same solubility data points, measured in NaClO4 at ionic strengths of 1.7-3.6 m, to extrapolate a solubility constant to zero ionic strength. The thermodynamic constants, calculated using the Pitzer ion interaction model (11), are more consistent with our solubility data. Thermodynamic calculations made at high ionic strength required concentrations in the molal (m) scale. To convert the molar (M) concentration data to the molal (m) scale, the density of each solution was required. The solutions of interest were composed of various concentrations of NaNO3, NaHCO3, and Na2CO3. The densities of these solutions were estimated by assuming the solution was composed of pure NaNO3 equal of its total sodium concentration. The density of the hypothetical NaNO3 solution was determined from concentration property tables (14). Because NaNO3, NaHCO3, and Na2CO3 solutions at equal sodium concentration have very similar densities, this method for estimating solution density was considered sufficiently accurate. For low ionic strength solubility calculations (non-cˇejkaite solubility calculations), the thermodynamic database thermo.com.v8.r6t.dat provided with GWB 6.0 was used. This database was modified to include the thermodynamic stability constant for the neutral aqueous complex Ca2UO2(CO3)3(aq) and the solubility constant for becquerelite [Ca(UO2)6O4(OH)6‚8H2O(s)] (Table 1).
Results and Discussion Selective Extraction Results. Results of the first set of selective extraction experiments (SE1) for C-203 and C-204 sludge samples are shown in Table 2. The data include results for 99Tc, U, Fe, and Al. The data are presented in terms of percent extracted relative to the total concentrations determined from analysis of the bulk unleached sludge by the fusion method (dry weight basis). Also included in Table 2 are the total percentages determined by addition of the concentrations measured from each extraction step. For the C-203 samples, the 99Tc results are rather uncertain due primarily to the very low 99Tc concentrations in the sample extracts, which are mostly at or below the detection limit. Despite the uncertainties of the 99Tc results, the results from the first water extracts (80% for C-203 and 25% for C-204)
parameter β0 β1 R1 θ log K log K log K
value
reference
0.61 18.2 2.0 -0.13 -4.08 29.8 41.4
11 11 11 estimated 11 15 16
TABLE 2. Averaged Selective Extraction Results as a Weight Percentage of the Total Concentrations (Based on Fusion Results) for C-203 and C-204 Sludge Samplesa average weight % sample number SE1-203-S1 SE1-203-S2 SE1-203-S3 SE1-203-S6 SE1-203-S7 total SE1-204-S1 SE1-204-S2 SE1-204-S3 SE1-204-S6 SE1-204-S7 total
extractant C-203 deionized water deionized water acetate buffer 8M HNO3 hot concn HNO3 C-204 deionized water deionized water acetate buffer 8M HNO3 hot concn HNO3
99Tc
U
Fe
Al
80
