Page 8
PROGRESS AND RESULTS OF DOE'S RESEARCH AND DEVELOPMENT PROGRAM
This section provides an evaluation of DOE's R&D program on the three alternative processes for the removal of cesium and one alternative process for the removal of strontium and actinides from the high-level waste at SRS. The committee has used the conclusions and recommendations from the final report of the 2000 NRC committee (NRC, 2000) as a starting point in its evaluation and has reviewed the R&D results generated since that report was issued to determine if the issues raised by that committee have been addressed adequately.
This section is organized as follows: For each alternative process, a brief recapitulation of the relevant conclusions and recommendations of the 2000 NRC committee is provided. This is followed by an analysis of current R&D efforts under way at SRS, other national laboratories, and academic laboratories to address the 2000 NRC committee's findings and recommendations. The present committee presents findings that identify any technical uncertainties that it believes warrant further consideration and, where appropriate, makes recommendations to address them. In addition, the committee provides its conclusions on the state of resolution of technical uncertainties and its impact on the downselection process.
The four processes under primary consideration by DOE include one process for removal of strontium and actinides from high-level waste with a nonelutable ion exchange process utilizing one or more sodium titanate compounds and three candidate processes for cesium removal: caustic side solvent extraction, crystalline silicotitante non-elutable ion exchange, and small tank tetraphenylborate precipitation. 2
2A fourth option, direct disposal in grout, was previously considered by SRS. Westinghouse Savannah River Company eliminated direct grout because the schedule uncertainty due to public and regulatory approval and potential litigation did not meet a 2010 schedule requirement based on available tank farm space. Fortenberry, J. K. 1998 (November 20). Memorandum to G. W. Cunningham regarding the SRS Report for Week Ending November 20, 1998. http://www.dnfsb.gov.
Page 9
SMALL TANK TETRAPHENYLBORATE PRECIPITATION
The small tank tetraphenylborate precipitation process uses a sodium tetraphenylborate (NaTPB) reagent to remove cesium from the high-level waste salt solutions. The processing approach is fairly straightforward: The NaTPB reagent is added to a batch of salt solution and stirred to promote the formation of a cesium tetraphenylborate (CsTPB) precipitate, which is subsequently separated from solution by filtration. The precipitate is washed to remove unreacted NaTPB and excess salt and is then sent to the Defense Waste Processing Facility (DWPF) for further processing 3 and immobilization in glass. The decontaminated salt solution is sent to the on-site saltstone facility.
The STTP option is an engineered version of the in-tank precipitation (ITP) process that was originally designed and demonstrated for cesium removal at Savannah River. The ITP process was designed to be carried out in an existing high-level waste tank at Savannah River, but the process was abandoned after a large benzene excursion was observed during the startup of processing operations in Tank 48 in 1995 (NRC, 2000, pp. 44-49). 4 The STTP process, as currently designed, will be carried out in specially designed tanks that are smaller than the existing tanks at Savannah River to reduce contact time between the salt solutions and the NaTPB reagent, which should reduce the decomposition of tetraphenylborate and the generation of benzene and also to allow the safe handling and abatement of any benzene that is generated during processing.
The process, if implemented, will be carried out on 67 production batches of waste obtained by transferring supernate and dissolved salt cake from one or more million-gallon high-level waste tanks. SRS will obtain detailed compositional data on each waste batch and will run additional tests to confirm its compatibility with the selected processing option before the batch is processed.
2000 NRC Committee Recommendations
The 2000 NRC committee had several concerns regarding the
3The CsTPB precipitate is treated with acid, which allows for controlled release of benzene. The benzene is separated and burned at an on-site incinerator. The remaining aqueous stream contains boron, cesium, and potassium salts, which are suitable for vitrification in the DWPF. 4The benzene excursion was produced by decomposition of TPB, most likely through reactions with a metal catalyst in the tank waste. This excursion has been difficult to explain, and subsequent lab tests have not been able to reproduce it.Page 10
STTP option. First and most important, the committee determined that considerable effort was needed to identify the mechanism of tetraphenylborate decomposition, including the estimation of bounding TPB catalytic decomposition rates. In addition, a number of other issues were identified:
-
whether the process can achieve the required decontamination factors (DFs) 5 of 7,700 (average) and 40,000 (upper bound) for Cs-137,
-
whether the composition of the process stream to the vitrification process in the Defense Waste Processing Facility is acceptable for making glass,
-
whether washing and recycle of the process stream can minimize the amount NaTPB required for this option, 6
-
whether foaming of the waste after treatment with NaTPB could block transfer lines or result in poor separation of CsTPB from solution, and
-
whether cycle times and products associated with the decomposition of CsTPB during precipitate hydrolysis processing are compatible with existing processes.
The 2000 NRC committee recommended that as part of the effort to bound catalytic decomposition rates, SRS should develop robust testing protocols to process moderately sized samples of real waste from each of the blended batches from the high-level waste (HLW) tanks at SRS using NaTPB. The 2000 NRC committee also recommended that tests on moderately sized samples of real waste be implemented as soon as possible to help assess the viability of the STTP option.
Current Research and Development Results
SRS appears to be making significant progress to resolve the issues raised in the 2000 NRC committee report, and the present committee commends the STTP team for its research accomplishments. In particular, SRS is making good progress to (1) further elucidate the general features of the catalytic decomposition of sodium tetraphenylborate, using various analytical techniques (nuclear magnetic resonance [NMR], extended x-ray
5Decontamination factor (DF) is the ratio of feed solution contaminant concentration (in this case cesium-137) to the contaminant concentration of the solution after treatment by the cesium removal process.6Design of the small tank option has focused on minimizing the amount of NaTPB reagent used in the cesium separation process. Thus, work has concentrated on the establishment of conditions that optimize the recovery of excess NaTPB during washing and the recycling of this reagent.
Page 11
absorption fine structure [EXAFS], and transmission electron microscopy [TEM]) to help define the nature of a Pd/Hg catalyst system and the TPB decomposition intermediates, and (2) design, construct, and demonstrate equipment for preliminary testing of real waste in 0.5-, 2.0-, and 20-liter samples. The committee offers comments on this work below.
Mechanism of TPB Decomposition
In 1997, SRS personnel uncovered the possible role of metal catalysts in TPB decomposition and subsequent release of benzene. SRS personnel have indicated that Pd/Al2O3 in the presence of mercury (especially diphenylmercury [HgPh2]) can function as a catalyst system for TPB decomposition in both simulants 7 and in real waste samples. Consultants hired by SRS have provided a plausible, but speculative, mechanism for Pd/Hg-catalyzed TPB decomposition that shares some of the features of catalytic Suzuki coupling (Miyaura and Suzuki, 1995).
Analysis by 11B NMR spectroscopy has revealed some additional features of the catalytic decomposition of TPB by Pd(0) on alumina/HgPh2, including the observed induction period. 8 Mercury may participate in catalyst activation by facilitating the nucleation and growth of palladium or Pd/Hg nanoclusters. The reactivity of nanoclusters is a function of surface area. The presence of mercury may enhance the reactivity of paladium particles as a catalyst for TPB hydrolysis. EXAFS and TEM analyses completed on simulant solids reveal face-centeredcubic (fcc) palladium nanoclusters and palladium-rich fcc clusters with mercury atoms surrounded by palladium atoms; both nanoclusters appear to be stabilized against aggregation by other components in HLW simulants.
Although a catalyst system for TPB has been identified, observed decomposition rates, using the best estimates of SRS tank waste concentrations of palladium and mercury, are still too low to account for the 1995 Tank 48 excursion. The average rate for the Tank 48 excursion, estimated at 10 mg of benzene per liter-hour of salt solution, was achieved using 26 mg/L of the catalyst Pd(0) on alumina in combination with Hg(II). It should be noted, however, that palladium concentrations in the tanks are currently estimated to range from 0.01 to 1.5 mg/L. The incubation periods have not been reproduced consistently with simulants or real waste, and the synergetic effects of mercury or diphenylborate are
7Simulants are laboratory-developed materials that mimic the properties of real waste samples. They are similar to real waste samples except that they do not contain radioactive species.
8The Tank 48 benzene excursion occurred 3 months after NaTPB was added to the waste. This incubation period is thought to be important for catalyzing TPB to produce benzene.
Page 12
not yet clear. Moreover, it is puzzling that no significant TPB decomposition was observed in any of the tests conducted with real waste by SRS, even after 6 months, except for one run conducted with an added catalyst simulant. Thus, although progress has been made, there is still an insufficient understanding of the TPB decomposition process to allow rates to be bounded with a high level of confidence. This is especially true given the compositional variability of the tank waste to be processed. 9
Finding: The process by which TPB decomposes is not completely understood and is not predictable, either mechanistically or empirically. At least two issues remain: (1) bounding rates for the generation of benzene, and (2) possible DF loss due to unanticipated rapid TPB decomposition.
The committee believes that both of these issues can be engineered around, 10 but that cost and schedules may be impacted. Specifically, the lack of understanding and predictability of the TPB decomposition process presents the possibility that one or more batches of waste may not be treatable without unplanned-for tank waste blending to dilute catalysts present in the waste. 11 A commitment by SRS to put into place preprocessing testing to confirm the suitability of each batch of feed would significantly reduce the potential impacts and effects of untreatable batches, although it would leave unresolved the issue of how such batches should be processed. SRS's plans to isolate roughly million-gallon batches of real waste for testing prior to processing by these types of procedures partially circumvent the need for a full understanding of the reasons behind the 1995 excursion in Tank 48.
Recommendation: Given the lingering uncertainties in the catalytic decomposition of TPB, the committee believes that additional research should be undertaken to improve the predictability of TPB decomposition if STTP survives downselection.
9A discussion of SRS tank waste compositions is provided in NRC (2000). Concentrations of many key radionuclides (e.g., cesium and actinides) in the tanks vary by several orders of magnitude, and concentrations of trace constituents that might act as catalysts are poorly known.
10As discussed in NRC (2000), the use of specially designed processing tanks will allow for much better control of reaction rates and benzene handling than is possible using the existing underground tanks as was planned for the ITP process.
11Given the variability in tank waste compositions, it may be possible to blend an unacceptable batch with waste from another tank to dilute the catalyst concentrations.
Page 13
Waste Foaming
Waste foaming was first observed in laboratory tests on real waste samples and is caused by the entrainment of air during stirring of NaTPB-waste slurries (see NRC, 2000, p. 49). Foaming is of concern because of the potential for clogging process transfer lines and inhibiting the separation of CsTPB from the waste slurry. Antifoaming and defoaming agents 12 have been identified and tested using HLW simulants. Although an effective antifoaming agent (IITB52) 13 has been identified, its stability is limited, its delivery system has not been established, and the downstream consequences of its use for processing have not yet been evaluated.
A number of antifoam issues remain. SRS plans to evaluate the effects of radiation on the performance of IITB52 by conducting a series laboratory experiments using irradiated and unirradiated samples. Additionally, SRS is planning to conduct process simulation studies on the IITB52 antifoam agent to determine its effects on downstream processes. This agent will also be utilized in a test on real waste.
Recommendation: Further research on antifoaming agents, including diluents, 14 agent stability under processing conditions, and their impact on downstream processing is recommended. Determinations of the extent to which different batches of the antifoaming agent will perform under operating conditions with real waste also should to be made.
This testing will allow process performance to be established systematically and evaluated under conditions that safely bracket the acceptable conditions for planned processing operations.
Real Waste Tests
Tests of the STTP process of approximately 6 months' duration using real waste have been developed and completed on samples from six different tanks. Tests conducted without added Pd(0) on alumina/HgPh2 showed little TPB decomposition and achieved sufficient cesium DFs (>40,000). However, these tests were conducted on samples of waste
12These agents are chemical compounds that reduce the viscosity of the waste slurry, thereby inhibiting the development of air bubbles.
13IITB52 is a water-soluble liquid mixture of esters with a density of 1.01 g/mL.
14The diluent used for IITB52 is wash water.
Page 14
supernate only and did not include dissolved salt cake.
Twenty-liter continuous stir tank reactor 15 demonstrations with real waste have been run successfully, and rapid cesium and strontium decontaminations have been demonstrated. 16 In one test with added catalyst simulant (7.8 mg/L Pd(0)/Al2O3 + 85 mg/L Hg(II)), acceptable cesium, strontium, and uranium DFs were achieved and maintained. 17 Benzene monitoring and abatement have been demonstrated.
Finding: Based on the real waste tests, the STTP process appears to meet cesium DF requirements for SRS tank waste.
Recommendation: SRS should continue to refine preprocessing testing protocols, and if STTP is selected as the primary or backup option, plans should be made to process moderately sized samples from each of the proposed processing batches using MST and TPB and the selected antifoaming agent before the process is implemented.
This preprocessing is recommended to mitigate the possibility of unplanned-for tank waste blending to dilute TPB decomposition catalysts present in the waste.
Committee Conclusions on the STTP Process
The STTP process has remaining technical uncertainties, but engineering solutions can probably be found for most of these potential processing problems. However, because of the unpredictability of the decomposition rate of TPB, there remains the risk that one or more of the 67 HLW production batches will require additional or special treatment before it can be processed using this option.
15A continuous stir tank reactor maintains the same concentration of reactants throughout the tank by stirring.
16DFCs>40,000; DFSr ~100.
17The DF for cesium was maintained between 10,000 and 40,000.
Page 15
CRYSTALLINE SILICOTITANATE ION EXCHANGE
The ion exchange process for removing ionic species from aqueous solutions has been in commercial use for more than 100 years. Although the underlying technology is well established, ion exchange for cesium removal from high-level waste presents many challenges. The ion exchange material must withstand both high alkalinity and high radiation fields while exhibiting selectivity for cesium in the presence of much higher concentrations of sodium and potassium. A promising ion exchange material, crystalline silicotitanate (CST), has been the subject of extensive R&D efforts at numerous laboratories over the last 30 years and is the material of choice at SRS for separation of cesium from high-level waste by ion exchange.
In the current design for this process, CST will be packed into three 5-foot-diameter by 16-foot-long columns arranged in series. The columns will be cooled by the flow of process liquids through the column, which will remove heat produced by radioactive decay. The salt solutions will be filtered and then pumped through the first column, known as the lead column, at moderate pressure, where most of the cesium is expected to be exchanged. The solution will exit that column into a gas separation apparatus, where radiolytic gas (mainly hydrogen) will be removed. The solutions will then be pumped through the second and third columns, known as the middle and guard columns, for further decontamination if necessary, and the decontaminated solutions will be sent to the saltstone facility for immobilization in grout.
CST is nonelutable, so once the lead column is loaded with cesium it will be removed and sent to the DWPF for processing. The loaded CST will be removed from the column, size-reduced by grinding, mixed with glass frit, sampled, and transferred to the glass vitrifier. Once the lead column is removed from the processing facility, the middle column becomes the new lead column while the guard column becomes the new middle column. A column loaded with fresh CST is installed as the guard column. The facility is designed with valves and jumpers so that column positions can be swapped with a minimum of handling.
2000 NRC Committee Recommendations
Concerns by the 2000 NRC committee regarding CST as a process for cesium removal centered on material consistency and column design. The committee noted that CST displayed wide variability in performance, which could probably be traced to manufacturing variability, variations in
Page 16
pretreatment 18 of the material at SRS, or in the testing used to characterize the materials. As a result, the 2000 NRC committee recommended that efforts be made to ensure that a consistent and reproducible material could be obtained for use, and that uniform CST pretreatment and testing protocols be developed.
The 2000 NRC committee further concluded that the column design for the CST process may not be adequate for the thermal loadings and radiation fields expected with real waste. In addition, the committee determined that possible problems with radiolytic gas generation inside the ion exchange column, which could disrupt the flow of liquid and reduce the efficiency of the ion exchange process, had not been resolved. As a result, the 2000 NRC committee recommended that the ion exchange column design be reevaluated. In addition, the committee recommended that an R&D effort be undertaken to study factors that are important for process and column operation and design.
Current Research and Development Results
The R&D program in place for CST is designed to address issues of concern identified by the 2000 NRC committee. Research thrusts related to the CST process include the following:
-
the chemical and thermal stability of manufactured CST, including chemical pretreatment requirements;
-
the effects of gas generation in CST ion exchange columns; separation of radiolytic gas from the liquid process streams during transfers between columns; and handling, size reduction, and sampling of CST in the preparation of the DWPF feed;
-
the performance of CST in columns, the kinetics of sorption related to temperature and feed composition, the capacity of CST to load cesium, modeling of CST performance, and alternatives to conventional column designs and parameters; and
-
the impact of additional titanium from CST on the DWPF glass product. 19
18The form of CST for use with tank wastes is described in the literature as a sodium salt, but it is manufactured and distributed in a protonated form at pH 3; pretreatment is needed to convert it to the sodium salt. The 2000 NRC committee concluded that this pretreatment may contribute to the observed instability of CST.
19Because CST is nonelutable, it must be incorporated into the high-level waste glass made at the DWPF. For further discussion of the possible impact of additional titanate from CST on DWPF glass, see the discussion on additional titanium later in this section on CST.
Page 17
CST Performance and CST Pretreatment Technologies
The tank wastes at SRS are caustic and contain significant amounts of aluminum. Previous studies showed that when CST, which is delivered to SRS in an acid form, is contacted with caustic waste solutions, the rise of pH leads to Al(OH)3 precipitation and immediate plugging of the ion exchange columns. Preconditioning of the CST columns with aluminum-free NaOH solutions appeared to be an easy answer to the problem. This work showed, however, that column plugging can be caused by precipitates other than aluminum hydroxide.
Testing at ORNL and SRS has shown that niobium introduced into CST during manufacturing could be dissolved and reprecipitated as several niobium phases (sodium-niobiate, sodium-sulfate-halide, cancrinite) at elevated temperature and thereby plug the ion exchange column. As a consequence, manufacturing processes for CST, including preconditioning to remove leachable niobium, have been modified. These modifications resulted in removal of much (>95 percent) of the leachable niobium added during manufacture and the removal of about 40 percent of the leachable silica. The low residual niobium eliminated precipitation of hydrous niobium oxides as a cause of column plugging.
Silica leaching from CST also has been found to impact column performance through the formation of an aluminosilicate precipitate, even after leachable silica is removed from the CST as described previously. These precipitates include cancrinite and minor amounts of sodalite. In addition, the precipitation of aluminosilicates on CST was observed in waste simulant experiments (described below) at elevated temperatures (50-120 °C) and/or on prolonged exposure (86 days) to the simulants.
Although fundamental understanding of the formation and impact of secondary minerals and phases has been enhanced through research, particularly with respect to the formation of new niobium phases, significant questions remain regarding aluminosilicate precipitation. The absence of useful predictive models and basic understanding of the events leading to precipitation of aluminosilicates has inhibited work on alternative designs for the CST process. This is discussed in more detail below.
Finding: SRS does not appear to have a clear and comprehensive understanding of the mechanism of aluminosilicate precipitation on CST. This issue poses potentially high technical risk for this candidate process.
Page 18
CST is manufactured by only one company in the world, UOP. This supplier has worked cooperatively with SRS to improve the manufacturing process to remove the niobium-based impurity phase and decrease the amount of leachable silica. Nevertheless, if this supplier were to cease manufacturing of CST, it might prove difficult for SRS to identify an alternate source of this material. In addition, it is unlikely that a new supplier would be familiar with the manufacturing process developed by the present supplier, parts of which are proprietary. As a result, it could take time to develop new and reproducible manufacturing and testing protocols, which could result in substantial delays in the program.
Finding: The reliance on a single supplier for CST poses potentially high schedule risks for this candidate process.
Chemical and Thermal Stability of Cesium-Loaded CST
R&D by SRS on the characteristics of cesium-loaded CST has the following two objectives: (1) to understand the principles and resolve issues related to temperature changes during processing in order to determine the time and temperature profile at which irreversible desorption of cesium from CST occurs after CST is added to waste simulants, 20 and (2) to determine the reason for the apparent reduced cesium capacity of CST following long-term exposure to cesium-bearing simulated waste solutions. In these experiments, cesium contained in a nonradioactive waste simulant is loaded onto CST at room temperature.
CST shows a slight loss of capacity after exposure to cesium solutions for more than 9 months. Loading of CST with cesium from simulated waste was found to be proportional to carbonate concentration, but the reason for this effect is not understood. Irreversible sorption of cesium at temperatures greater than 50 °C remains an unexplained phenomena, and factors that determine the reversibility and irreversibility of cesium sorption are not yet clearly understood. SRS representatives told the committee that this problem can be eliminated by controlling the operating temperature of the columns so that it does not exceed 50 °C.
Finding: Thermal stability most likely will not pose an insurmountable problem if operating temperatures remain low and CST columns are changed before they lose their sorptive capacity. However, the operating margins are small.
20As discussed in NRC (2000), at temperatures above 50 °C, cesium has been observed to undergo irreversible deposition from CST. The reasons for this desorption are unknown.Page 19
The effect of organic impurities and minor components on cesium sorption on CST was also examined by SRS to determine the effect of organic compounds, carbonate, oxalate, and peroxide on the rate of cesium loading on CST. Organic impurities, and oxalate, did not affect column performance. Cesium loading on CST was found to be proportional to carbonate content, as noted previously. High concentrations of peroxide were observed to decompose CST, although SRS does not expect this to be a problem at anticipated operating conditions at SRS (see below).
Finding: Significant progress has been made in understanding the effects of chemical impurities on the performance of CST.
CST Process Columns
R&D has been focused on determining whether gas generated by water radiolysis within a CST ion exchange column can affect cesium sorption. In addition, the performance of the CST column, including the performance of alternative column configurations, has also been examined. Measurements of cesium sorption in the presence and absence of radiolytic gas showed no significant differences. Radiolysis will generate hydrogen in concentrations that are expected to exceed the explosive limit. SRS notes that the process columns are expected to be vented and that gas generation is a manageable safety matter.
Tests of the effects of gas generation in the tall columns that are part of the current conceptual design were conducted using hydrogen peroxide. 21 These studies showed that hydrogen peroxide reacts with CST to liberate silica and titanium and form aluminosilicate precipitates on the surfaces of the CST particles. The formation of these precipitates plugged the columns and hindered hydraulic removal of the CST. SRS calculated that the amount of hydrogen peroxide formed by radiation in a loaded process column would be much lower than the amount used for these tests and, therefore, precipitate formation and column plugging from this source would not represent a significant risk during actual processing operations.
SRS has also demonstrated that disengagement of gas during transfers of liquids between ion exchange columns would be adequate to reduce the transfer of gas from the lead to the middle column and therefore would represent a low risk. In light of the severe operational consequences attending interactions of CST and hydrogen peroxide—the plugging of an ion exchange column could lead to the loss of fluid flow
21Hydrogen peroxide will be formed by water radiolysis in a column loaded with cesium.Page 20
and the buildup of heat—these reactions could be a high risk during actual processing operations.
SRS is also evaluating alternative methods of contacting CST with the process solution. Current column designs are based in part on CST transfer requirements to the DWPF. Other designs, such as shorter columns or moving beds, may be difficult to implement reliably.
Finding: The present conceptual design for this process requires large sorption columns in series that are potentially subject to blockage, precipitate formation, and gas bubble formation. These phenomena may disrupt flow and sorption of cesium in the columns.
As noted above, these phenomena have been studied individually and some of the problems have been solved. However, the committee has not seen evidence that SRS is attempting to integrate the solutions to these various problems into a relevant process simulation.
Effect of Additional Titanium from CST on the DWPF
The tolerance of DWPF glass to the addition of titanium from CST and also from monosodium titanate (MST) expected to be used for removal of strontium and alpha emitters (discussed below) has been reexamined by SRS. The presence of titanium can lead to crystallization in glass, which in turn can increase the liquidus temperature. 22 Crystallization in the DWPF melter can lead to processing problems.
Glasses were cooled at different rates to study their product consistency (Edwards, 2001). These glasses contained CST and MST (plus a simulated sludge) in amounts consistent with operations where real waste would be treated by both CST and MST. The measured values of these glasses consistently fall above the predicted values in the models used to predict durability. On the basis of these test results, SRS concluded that the very good durability of the CST-containing glasses implies that durability may not be the limiting factor for waste loading in the CST option for cesium removal, or the MST option for actinide or strontium removal.
Finding: The presence of titanium in presently estimated amounts does not appear to negatively impact the quality of the DWPF glass product.
22The temperature at which there is an equilibrium between the glass and the primary crystalline phase.
Page 21
If the final flowsheet for the CST or the MST process adds more titanium than currently envisioned, then the durability of the glass may have to be reevaluated.
Discussion
The need for information to allow process design and application to proceed has not been met adequately, as noted in the preceding discussion. Although alternative column designs and gas evolution issues have generally been identified, the committee did not learn of design selections or specifics of the management of safety issues associated with radiolytic gas formation and management. Suggested remedies for the plugging of columns observed in small-scale experiments have not yet been tested on a larger scale or with real waste. Most of the experimentation with waste simulants failed to include trace elements likely to be found in real waste.
Finding: Information needed to evaluate the risk of applying CST to cesium removal from the treated supernate has not been fully developed, although many of the issues have been identified. Therefore, the technical uncertainties remaining for the application of CST—including column plugging, resistance to hydraulic transfer, irreversible desorption, and column system technologies—will constitute a high risk for the use of this process for cesium removal.
Recommendation: If CST is selected as either the primary or the backup option, the technical uncertainties identified above must be addressed, particularly alternative column designs to mitigate aluminosilicate buildup and radiolytic gas formation.
Committee Conclusions on the CST Process
Of the three cesium separation processes under consideration, it is the committee's judgment that CST has the most technical uncertainties and the highest technical risks.
Page 22
CAUSTIC SIDE SOLVENT EXTRACTION
As a general method of separation, solvent extraction is a mature technology that has been used in the nuclear industry for more than 50 years, although primarily with acidic process streams. At SRS, the goal of this process is to extract cesium ions from the aqueous waste stream into a immiscible solvent, thereby reducing the radionuclide content of the aqueous phase and enabling it to be disposed in the saltstone facility. The caustic process under development at SRS would be the first plant-sized caustic solvent extraction unit at Savannah River.
The basic principle of CSSX is to use a sparingly soluble diluent material that carries an extractant that will complex with cesium ions in the caustic solution. Separated cesium can then be stripped back into an aqueous phase ready for transfer to DWPF. Following cesium extraction, the solvent is scrubbed with dilute caustic to remove other salts from the solvent stream. The solvent is then contacted in a countercurrent flow with a dilute acid stream to transfer cesium to the acid stream (in the strip stages). The solvent is then scrubbed or purged to remove degradation products prior to recycling to the front of the process.
2000 NRC Committee Recommendations
The 2000 NRC committee found that although solvent extraction in general is a well-developed process, the technical maturity of the proposed solvent extraction process for the removal of cesium from high-level waste at SRS lagged significantly behind that for the two competing processes (STTP and CST). 23 As a result, the majority of the 2000 NRC committee's conclusions and recommendations focused on operational concerns.
These operational concerns were numerous. The committee questioned whether the solvent system could be cleaned and reused successfully. The buildup of particulate matter at the interface of the two solvent phases, common in solvent extraction, was considered by the previous committee to be a possibility in the process at SRS and could limit separation of these phases. The impact of changes in the feed
23The solvent used in cesium extraction studies at SRS is a multicomponent system. BoBCalixC6 is a calixarene crown ether that appears to work through a combination of effects to generate a cavity that preferably incorporates cesium relative to other ions. BoBCalixC6 is present in the CSSX solvent system in 0.01 M concentration. A diluent modifier, Cs-7SB, increases cesium extraction and is present at 0.5 M. An inhibitor, tri-N-octylamine, that inhibits impurity effects, is present at 0.001M. Isopar L is the solvent for these three active organic constituents of the extractant.
Page 23
composition, pH, temperature, and ratios of the solvent constituents on the separation efficiency and capacity of the solvent system was also of concern. Other areas of concern were whether sophisticated control systems would be required to maintain a steady-state, high-DF operation, and the effect of ions in tank waste on separation efficiencies. The 2000 NRC committee also questioned whether a reliable supply (and supplier) of the calixarene crown ether (BobCalixC6) used in the CSSX process at SRS would become available.
As a result of these concerns, the 2000 NRC committee had three recommendations regarding the use of this process. First, a “cold” demonstration of this process on a modest scale was recommended. The purpose of this demonstration was to address as many of the aforementioned operational concerns as possible and, in particular, to identify any possible “showstoppers” that would preclude the use of this process at SRS. Second, the committee recommended that design of a hot laboratory demonstration process, using real tank waste, on a scale sufficient to define the final process should begin as soon as the cold tests demonstrated a high degree of confidence in the feasibility of the process. Finally, the committee recommended that work begin immediately on defining the production capability and economics for commercial quantities of the calixarene crown ether.
Current Research and Development Results
SRS appears to have implemented a robust R&D program that addresses many of the operational concerns expressed by the 2000 NRC committee. A demonstration of the complete CSSX process flowsheet using simulated waste was conducted, and real waste tests are scheduled for the spring of 2001. The R&D program is addressing factors such as the chemical and physical properties of the solvent, the stability of the solvent system, batch DF extraction and stripping, and solvent commercialization. Details are provided below.
CSSX Proof of Concept
The overall CSSX process, including solvent recycle, has been demonstrated using simulants on the bench scale. SRS personnel reported that under these conditions, the desired goal of a DF of greater than or equal to 40,000 was met, as was a cesium concentration factor of about
Page 24
15. 24 The R&D program has examined effects of impurities and trace components and effective temperature control, and SRS personnel report that no significant problems were encountered.
The operation of the centrifugal contactors (designed to provide efficient mixing between the solvents followed by separation of the phases) was also studied. SRS personnel determined that due to a problem with stage efficiency, a better understanding of the multistage hydraulic performance of the 2-cm centrifugal contactor used in the bench-scale tests was needed, but because this issue is mainly a result of the small size of the contactor, it will not be an issue in plant scale operations. 25
Solvent Chemical Stability
Research on the chemical stability of the CSSX solvent system examined the solvent and how it works, phase behavior of the primary solvent components, and distribution performance 26 and solvent cleanup. The overall process was reported by SRS personnel to be very effective. The CSSX solvent system was found to be stable to precipitation of solids for at least a year. The lower limit of the formation of dense liquid organic phases, a detriment common in solvent extraction processes, was found to be at 20 °C, well below the planned operating temperature for this process (about 35 °C). The distribution of cesium between phases was found to be reproducible for the various steps in the CSSX process.
Research on the effects of impurities on the solvent also has been conducted. The solvent has been cycled through multiple extraction, scrubbing, and stripping batch processes using a waste simulant to determine whether impurity buildup would degrade stripping performance. An impurity buildup was observed to occur, but a dilute sodium hydroxide wash was found to be effective for cleaning the solvent system and removing these impurities. The waste simulants included noble metals and organic compounds that are expected to be encountered in real waste. These were not observed to degrade the solvent system or its performance.
Thermal stability tests were also undertaken to examine the operational limits of the solvent system. For each step of the CSSX process (extraction, scrubbing, and stripping), the stability of the solvent in contact with the waste simulant was analyzed with external heating to 35-60 °C. SRS determined that even at 110 days at the planned maximum
24The cesium concentration factor is the cesium concentration in the aqueous strip effluent [EW] divided by the cesium concentration in the feed simulant [FS]: [EW]/[ FS].
25Plant-size contactors will be much larger and, as a result, will not have these hydraulic problems.
26Distribution performance measures the cesium distribution ratio between the organic and aqueous phases over the progressive steps in the CSSX process (extraction, scrubbing, and stripping).
Page 25
operating temperature of 35 °C, performance remains good. SRS personnel also reported that the solvent does not undergo unacceptable degradation due to nonradiation effects such as the presence of noble metals.
Solvent Radiological Stability
The first purpose of solvent irradiation experiments is to determine the radiolytic stability of the solvent, identify decomposition products, and assess their impact on solvent performance. A secondary focus of these experiments is to compare simulant and real waste test data and to compare internal and external irradiation results. The radiolysis program involves four experiments: (1) cobalt-60 external irradiation of the solvent; (2) solvent self-irradiation with 137Cs-spiked simulant; (3) contactor hydraulic performance with 137Cs-spiked simulant; and (4) solvent self-irradiation with real waste from SRS.
Results of the radiolytic stability tests indicated no significant degradation of the solvent system. External irradiation of the solvent using cobalt-60 produced only minor decomposition. 27 Irradiation of the solvent using a 137Cs-spiked simulant has not identified any radiolytic concerns. 28 In both sets of experiments, third-phase formation or formation of particulates at the solvent interface was not observed, and cesium decontamination efficiency was still found to be within the expected range.
Results from the contactor hydraulic performance test have not identified any radiolytic concerns. Physical observations did not identify any dose-related impacts, and chemical analysis of the samples from this test were in progress at the time of the committee's briefings.
Further tests of the solvent are planned in the spring of 2001 using real waste. Batch extraction tests, designed to measure the solvent performance using real waste from the SRS tank farms, will attempt to operate the process for 28 solvent turnovers, and demonstrate a decontamination factor of >15,000.
Finding: R&D on the CSSX process has, so far, encountered no significant technical obstacles, and there do not appear to be any technical obstacles to
27Approximately 10 percent of BOBCalixC6 was lost at a 16-Mrad dose (equivalent to exposure of the solvent system to SRS real waste for 160 years), and about 10 percent of TOA was lost at a 6-Mrad dose (equivalent to exposure of the solvent system to SRS real waste for 60 years).
28In these experiments, the solvent system received exposure to cesium-137 equivalent to a decade of plant operation of the process with SRS real waste.
Page 26
scale-up of the CSSX process to plant-scale operations.
The research performed over the last year has addressed most of the issues identified by the 2000 NRC committee and appears to confirm the viability of this process at the laboratory scale. The present committee has been impressed with the unimpaired performance of this process under very rigorous testing. The strong success of the CSSX process in this R&D program is unusual for most process development.
Finding: Successful bench-scale demonstration of the complete process with actual tank waste is critical for qualifying the CSSX process for serious consideration in the down-selection process.
This demonstration, if done well, will show whether the CSSX process can remove cesium from real waste at levels sufficient for saltstone requirements and whether pilot-scale testing is warranted.
Solvent Preparation and Commercialization
A key technology issue that impacts successful use of the CSSX process is the availability of the solvent at plant-scale quantities. Commercial suppliers of all solvent components, including BOBCalixC6, have been identified by SRS. BOBCalixC6 with greater than 97 percent purity has been prepared successfully by an outside manufacturer, and its production scale and cost are well within the economic and schedule parameters (given the projected loss rates) for the use of this process at SRS. 29 Additional suppliers of BOBCalixC6 have been identified by SRS, and a U.S. patent protects the government's rights to grant a license to manufacture this material. No other component of the solvent system appears to present an economic or scheduling burden to the use of this process.
Finding: No significant economic obstacles for scale-up of the CSSX process appear to have been encountered so far.
Recommendation: If this process remains a viable candidate for cesium removal, monitoring of the cost and potential suppliers of the reagents for this process should continue.
29For a plant-scale charge of 4,000 liters of solvent, 46 kg of BOBCalixC6 would be required. The cost for this amount of BOBCalixC6 should be approximately $5 million, well within the $8 million budgeted for this reagent.
Page 27
Discussion
After being far behind in level of development compared to the other two processes a year ago, there has been a major acceleration of work on the CSSX process, and the development gap between it and the others has been reduced markedly. The committee is impressed by the overall quality of the science on this alternative. The seamless integration of research from several laboratories is especially impressive.
Finding: The main source of concern regarding the viability of the CSSX process continues to be the stability of the solvent system. The process itself should be relatively straightforward to scale up, at least from a mechanical standpoint, because the process hardware has been proven in nuclear applications and there are no solid-handling steps.
The potential concerns include chemical and radiolytic stability, the possible detrimental effects of its breakdown products on the DF and the cesium concentration factor, and the possibility of performance-degrading effects of trace components in real waste that have not been included in the simulated feed tests.
Recommendation: Extensive testing of the most performance-critical components of the solvent (e.g., composition, pH, and temperature ranges the solvent would most likely encounter) should continue in parallel with the bench-scale process test using real plant waste in order to give the greatest possible assurance that the required separation performance can be achieved and maintained with any waste composition likely to be encountered. Successful completion of this program will allow concerns about the solvent system to be characterized as low risk.
Committee Conclusions on the CSSX Process
Unless tests with actual waste encounter new problems, the CSSX option for cesium separation presents, at present, the fewest technical uncertainties of any of the three cesium separation alternatives.
Page 28
ACTINIDE AND STRONTIUM REMOVAL
The removal of strontium and actinides (especially plutonium and neptunium) is an important step in the high-level waste processing flowsheet at SRS. As presently envisaged, strontium and actinides will be removed from the salt solutions in all three of the cesium processing options discussed in this report. At present, the use of monosodium titanate is the method of choice. Although the mechanisms for strontium and actinide removal by MST are not well understood, it is presumed that an ion exchange reaction of the sodium ions in the MST takes place, primarily with cations in higher oxidation states (e.g., strontium, plutonium, neptunium, and uranium) but also, to a lesser extent, with monovalent cesium and potassium cations.
The three cesium removal options are designed to process waste streams that have been treated to remove actinides and strontium. SRS plans to remove these radionuclides at the “front end” of processing operations with CST and CSSX by batch contact of the waste solution with finely powdered MST. 30 Incoming salt solution from storage tanks containing entrained sludge solids is pretreated with MST to adsorb strontium and plutonium. The resulting slurry is filtered using a cross-flow filter, and the MST and sludge solids are to be sent to the DWPF for vitrification.
2000 NRC Committee Recommendations
The 2000 NRC committee found that two major issues had to be resolved before SRS could successfully implement MST for actinide and strontium removal: (1) whether strontium and actinide removal could be accomplished within saltstone limits and throughput rates required by the DWPF, and (2) whether the MST concentrations used to remove strontium and actinides would exceed compatibility limits for DWPF glass. The 2000 NRC committee recommended that R&D be performed to resolve these issues, that requirements reliable sources for the manufacture of MST be established, and that SRS look at alternatives to MST.
Current Research and Development Results
The 2000 NRC committee questioned the assumption that strontium-alpha separation is to precede cesium removal. This challenge provides the basis of some of the R&D activities described below.
30With STTP, the MST actinide and strontium removal process is carried out concurrently with cesium removal.
Page 29
Actinide and Strontium Removal Requirements
Removal of alpha emitting elements and strontium is required to meet saltstone waste acceptance requirements, which can be defined in terms of the average and the highest bounding activities of the process streams. 31 Further, the throughput rates must meet downstream feed requirements, especially to keep the DWPF operational.
One important focus of the R&D program is elucidating actinide and strontium removal rates for MST. Simulated waste solutions at 5.6 M Na+ containing known quantities of strontium, plutonium, uranium, and neptunium were used and, at controlled temperature, samples were drawn and analyzed for sorbate concentrations after removal of solids by filtration. The results are dependent on temperature and ionic strength. Based on the experiments with waste simulants, SRS reported the following:
-
MST removal of strontium is adequate to meet saltstone requirements. The experimental DF reported for strontium was about 150, which exceeds the maximum required DF of 26,
-
MST removal of neptunium is inadequate to meet saltstone requirements for waste in some of the tanks. The experimental DF reported for neptunium is 3.47, which is well below the maximum required DF of 33, and
-
MST removal of plutonium is also inadequate to meet saltstone requirements for waste in some of the tanks. The experimental DF reported for Pu is 11.3, slightly below the average required DF for the tank waste of 12, and well below the maximum required DF of 55 for tanks with the highest plutonium concentrations.
SRS also reported that saltstone limits probably could be met by blending waste from different tanks to reduce the required DFs for neptunium and plutonium. To this end, SRS plans to prepare 67 separate batches for MST and cesium processing by blending together waste from different tanks. Careful blending will allow SRS to dilute the high radionuclide concentrations in the “problem” tanks, thereby reducing the DFs required to meet saltstone requirements. In fact, SRS personnel reported that by careful blending and MST processing, they can produce batches that meet saltsone requirements for neptunium and strontium, and
31The required DF (average/bounding) for plutonium/americium is 12/55, for uranium it is 1/1, for neptunium it is 1/33, and for strontium it is 5/26. Bounding values are decontamination factors for tanks with the highest concentration of the radionuclide in question.
Page 30
that these same batches can be processed by MST to achieve required DFs for plutonium.
An issue related to the use of MST for actinide and strontium removal is the possibility that colloidal plutonium could exist in the tank waste. Such material would not be chemically removed by the MST, and colloidal particles would be too small to be removed by filtration. However, filtration experiments using varying membrane pore sizes of samples from several waste tanks did not indicate the presence of colloidal plutonium.
Finding: The blending of tank waste to produce 67 process batches and treatment by MST appears to meet the saltstone requirements for neptunium and strontium decontamination. Based on the information received by the committee, MST appears to be adequate to separate Pu, as long as there is no colloidal plutonium in the waste, but with little margin to meet saltstone requirements.
A technical uncertainty that remains to be resolved is the kinetics of sorption on MST. This issue is particularly important because of the additional titanium present in MST (and also from the CST), when in the waste feed at the DWPF, may exceed acceptable limits on the amount of titanium in the glass waste form.
Finding: The maximum quantity of titanate allowable in the process stream to meet the titanium levels acceptable in the vitrified waste form remains a technical uncertainty. 32
Solid-Liquid Separation Studies
Once MST solids have been added to the salt solutions to sorb strontium and actinides, these solids (along with any sludge solids in the waste) must be separated from the liquids and transferred to the DWPF. Filtration is currently the baseline process for solids removal, and several studies have been performed to elucidate filtration performance. The objectives of these studies are (1) to confirm baseline sludge and MST cross-flow filtration performance 33 at pilot scale with simulated wastes,
32SRS has conducted durability tests on CST- and MST-loaded glasses.
33In cross-flow filtration, the process stream flow is tangential to the filter surface, thereby minimizing the buildup of solids on the surface that reduces filter efficiency.
Page 31
and (2) to evaluate alternative solid-liquid separation technologies for their potential to reduce facility size.
The main objective of the pilot-scale filtration studies is to measure filtration rates for slurries containing simulated sludge and MST in large, prototypic equipment. The need for backpulsing 34 was evaluated and the extremes of the system were tested. A target cross-flow filter permeate flux of 0.22 gallons per minute per square foot is desired for concentrating slurries with up to 5 weight percent insoluble solids. The observed fluxes meet or exceed these design assumptions. Filtration experiments have been carried out at the pilot scale with highly promising preliminary results, and appear to indicate that filtration can be achieved within the requisite parameters required for all three candidate cesium removal technologies.
A focus of the R&D program has been to investigate several ways to increase MST and sludge filtration rates. Additives, both flocculants and antifoamants, have been found that improve filtration. Cross-flow filter tests with flocculants have shown a 1.3-fold improvement in filter flow rate over baseline. Alternative filtration technologies are under investigation but have not yet provided significant results.
Ongoing studies include tests using real waste samples, including cross-flow flux and rheology measurements and tests of flocculants with and without MST. Additional pilot-scale filtration testing will involve filtration tests using sludge only and MST only for two waste compositions, and a potential test using a flocculant. Planned experiments on alternative filtration techniques include settling and decanting tests, high shear filtration (centrifugal) tests, and centrifuge evaluation.
Alternatives to MST for Actinide and Strontium Removal
In November 1999, SRS personnel reported to the 2000 NRC committee that alternative processes for actinide and strontium removal were being investigated in case MST fails to meet expectations. One of the most promising alternatives is a precipitation method for strontium and actinide removal using sodium permanganate. In this process, removal of actinides occurs upon the precipitation of hydrated manganese oxide following the sequential addition of strontium nitrate, calcium nitrate, and sodium permanganate to the highly alkaline waste solutions. The likely mechanism for actinide removal involves adsorption, inclusion, and occlusion in the hydrous manganese oxide matrix. The sodium
34Backpulsing is a method for removing particles that have collected in the pores and on the surface of the filter membrane using a periodic reversal of the transmembrane pressure.
Page 32
permanganate process achieved higher DFs for strontium than MST (199 versus 150), for plutonium (30.4 versus 11.3), for uranium (1.88 versus 1.14), and for neptunium (7.90 versus 3.47).
Two other alternative adsorbents to MST, sodium nonatitanate (ST) and SrTreat® (proprietary), were also compared to MST for their ability to remove strontium and actinides. Both SrTreat and ST exhibit rates of removal for strontium that are similar to MST (approximately 0.15 µg/L of strontium remain in solution after treatment versus approximately 5 µg/L for MST after 107 hours). For plutonium removal, ST has a similar rate to MST (approximately 3.5 µg/L of Pu remain versus 7 µg/L after 107 hours), while the rate for SrTreat® was significantly lower (approximately 90 µg/L of plutonium remain after 107 hours). Similarly, ST and MST had almost identical rates of neptunium removal (approximately 60 µg/L of neptunium remain after 107 hours), while the rate for SrTreat was again much lower (approximately 300 µg/L of neptunium remain after 107 hours). In general, ST showed a behavior parallel to that of MST. It was concluded by SRS personnel that MST kinetics are adequate to meet the baseline preconceptual design for each processing alternative and that the ST sorbent process and a manganese-based precipitation process provide promising backups for MST.
Finding: Two alternate precipitation processes are competitive with MST. These employ sodium nonatitanate, which behaves similarly to MST, and sodium permanganate.
Recommendation: The backup processes, sodium nonatitanate and the sodium permanganate-based precipitation process, should be studied further. The R&D program for these two processes should be based on that developed for MST and should continue until MST processing can be demonstrated to meet the saltstone, DWPF throughput, and DWPF glass requirements.
If one of these backup processes is found to be superior to MST, its substitution for MST will have to be done soon so as not to delay the implementation of the cesium removal processes. Resolution of the choice for this process is largely independent of the choice of a cesium separation process.
Page 33
Committee Comments on the MST Process
All of the cesium separation processes depend upon a separate step to remove strontium, neptunium, and plutonium. Currently, that step uses MST. Because the success of this step is essential to all three of the processes for cesium separation, the committee believes that continued R&D on alternate processes for the removal of actinides and strontium is essential until MST processing can be demonstrated to meet the saltstone, DWPF throughput, and DWPF glass requirements.
